
<#FROWN:J01\>
Cosmology, Clustering and Superclustering
William C. Saslaw
On large scales, the probability that galaxies occupy a given size volume of space is not random, like the toss of a coin, but is related to the presence of nearby galaxies. Positions of galaxies depend on one another. Details of this dependence may provide important clues to the origin and evolution of the universe. Modern systematic searches for these clues started in the 1930s with Edwin P. Hubble's galaxy counts, accelerated in the 1950s with the development of new statistical techniques, and surged in the 1970s and 1980s as computers became powerful enough to handle large amounts of data and calculate complex simulations of clustering physics.
There are two main descriptions of galaxy clustering. The first is essentially pictorial, derived by searching the sky for filaments, voids, and overdense regions. Galaxies in projected high-density regions having similar redshift distances are likely to form a physically related cluster. If such a cluster has been bound gravitationally for a large fraction of the present Hubble time, it evolves fairly independently of its surrounding galaxies. Occasionally several contiguous large clusters form a supercluster. This may result partly from the initial positioning of clusters and partly from their subsequent motions.
Most galaxies are not in large physically bound clusters. They are, however, clustered in a more general statistical sense. Rather than examining individual clusters, this second description looks for statistical departures from a Poisson distribution (which is the uniform random spatial distribution of objects whose positions are entirely independent of one another, i.e., totally uncorrelated). In a Poisson distribution there will be some regions where galaxies are strongly clustered just by chance, and other regions which are unusually empty, also by chance. The observed numbers and statistics of such regions may then be compared with those expected from various theories.
STATISTICAL MEASURES OF CLUSTERING
The most useful and informative statistics can be measured objectively from observations and related analytically to physical processes of clustering and computer experiments. Low-order correlation functions are one example. The two-point correlation function <*_>xi<*/>(r) in its simplest form for a homogeneous isotropic system is defined by
<O_>formula<O/>
Here n(r)dr is the average number of galaxies between radial distance r and r+dr from any given galaxy. The overall average number density of galaxies in the entire system, or over a very large volume of the universe, is <*_>unch<*/>. For a random Poisson distribution, <*_>xi<*/>(r)=0; so n(r) is determined just by <*_>unch<*/> and geometry. Therefore <*_>xi<*/>(r) helps measure departures from the Poisson state. Higher-order correlation functions use the relative positions of three or more galaxies for a more refined description that, unfortunately, is harder to measure observationally. The first accurate observations of <*_>xi<*/>(r) for galaxies, made by H. Totsuji and T. Kihara in 1969, gave a power law of the form <O_>formula<O/> on scales <O_>formula<O/> (for a Hubble constant of 50 km s<sp_>-1<sp/> Mpc<sp_>-1<sp/>). Large clusters of galaxies are observed to have a similar two-point correlation function if each cluster is represented by a single point, but this result is much more uncertain. On small scales <*_>xi<*/>(r)>>1, so the observed clustering is highly nonlinear; that is, correlations dominate for r <*_>unch<*/>10 Mpc.
Another observed simple objective clustering statistic is the distribution function f(N). This is the probability for finding N galaxies in a volume of size V, or projected onto the sky in an area of size A. If the distribution is statistically homogeneous then it will not depend significantly on the shape of the volumes or areas, provided they are sufficiently large and numerous to give a fair average sample. Recent analyses of the area counts of galaxies show that they have a distribution of the form
<O_>formula<O/>
where <*_>unch<*/>=<*_>unch<*/>V is the average number in a volume V for an average number density <*_>unch<*/>. The quantity b is a measure of clustering and is related to gravitational correlations. The observed value of b is 0.70<*_>unch<*/>0.05 for galaxies whose separations are typically 1-10 Mpc. For large clusters with separations of <*_>approximate-sign<*/>10-50 Mpc, b=0.3<*_>unch<*/>0.1. For a sample of faint radio sources with separations <*_>unch<*/>50 Mpc, b=0.0, which is a random Poisson distribution. The f(N) distribution for N=0 gives the probability that a region is a void with no galaxies at all.
Other statistics applied to galaxy clustering include minimal spanning trees (the shortest line connecting all the galaxies in a sample), topological patterns formed by contour maps of regions with the same density, and multifractal analyses (a single fractal dimension does not adequately describe galaxy clustering), which are related to how the average number density of galaxies around a given galaxy changes with distance from the galaxy. These other statistics also yield valuable information. Unlike <*_>xi<*/>(r) and f(N), however, they have not yet been related generally to an underlying dynamical theory. Some specialized computer experiments have examined their behavior.
THEORIES OF CLUSTERING
To understand the observed statistics we need to know the initial conditions for clustering as well as a physical theory for its subsequent evolution. Initial conditions may indicate properties of the early universe before galaxies formed and perhaps even close to the Big Bang. Some possibilities are that galaxy clustering started from a random Poisson distribution, or from a state with local clustering or from large-scale coherent structures. No clear observational evidence for any particular initial state has been found. On small scales the clustering processes themselves tend to destroy this evidence, while on large scales it is difficult to detect.
Different types and distributions of dark matter may also be important for forming galaxies and clusters. For example, massive neutrinos, other weakly interacting massive particles, cosmic strings, quark nuggets, or other currently speculative objects of various high-energy theories may influence galaxy clustering if they exist in sufficient quantity.
All known forms of matter gravitate and gravitation promotes clustering. Therefore, astronomers have developed analytical theories and examined many computer simulations to describe the gravitational clustering of galaxies. Results for different models are then compared with <*_>xi <*/>(r) and f(N). The models generally differ in their initial conditions, the role of dark matter, and the time available for clustering.
Computer simulations calculate the gravitational orbits of many thousands of particles, each one representing a galaxy, in the background of the expanding universe and any dark matter present. The orbits are found either by integrating the thousands of equations of motion - each with thousands of terms - directly, or by averaging the gravitational forces in different ways to simplify the problem. Averaging sacrifices detailed information in order to include a larger number of galaxies.
Computer models which start with strong structure on scales of tens of megaparsecs frequently do not agree with the observed correlation and distribution functions. Those that do, often agree only for a short span of their evolution. On the other hand, models with fairly homogeneous initial distributions, such as an uncorrelated Poisson state, evolve gravitationally to agree reasonably well with the observations and remain in agreement as they continue to evolve. In other words, they relax into the observed state and remain there rather than just pass through it. This may make them more aesthetically pleasing, although it does not guarantee they are correct. For example, some models in which galaxies have formed and clustered very recently may conflict with the uniformity of the cosmic microwave background.
<O_>figure&caption<O/>
Gravitational theory predicted the observed form of f(N) given earlier for relaxed statistically homogeneous clustering of point masses in a slowly expanding universe. The value of b is essentially the ratio of gravitational correlation energy (representing departures from a uniform Poisson distribution) to the kinetic energy of galaxies' peculiar velocities (representing departures from the Hubble flow). Computer experiments such as the example in Fig. 1 show that this relaxed state is a very good description for universes which start with no or little large-scale correlation, have <O_>formula<O/>, and have expanded by more than several times their initial radius. As differences with these conditions become greater, agreement with the observed f(N) decreases. These conditions also lead to two-point correlation functions in reasonable agreement with the observations provided, for example, that clustering started at redshift z<*_>approximate-sign<*/>8 in <*_>OMEGA<*/><sb_>0<sb/>=1 models and z<*_>approximate-sign<*/>30 in <*_>OMEGA<*/><sb_>0<sb/>=0.1 models. Therefore, gravitational clustering starting from fairly simple initial conditions seems likely to account for the objective statistical evidence now available. These include large underdense regions and filamentary structures, some of which could have formed just by chance concentration of independently clustering regions or their boundaries. When more subtle statistics are developed further and related to dynamical evolution, perhaps they will reveal clear evidence for other processes such as primordial explosions, or large-scale initial structures.
Cosmology, Cosmic Strings
Neil Turok
One of the most active areas of current research in physics and astronomy is the search for a theory of the formation of structure in the universe.
Historically, this is a result of the success of two different theories and the attempt to combine them. In astronomy, the Hot Big Bang model of the universe has three big successes. It successfully explains the expansion of the universe, the relic microwave background radiation, and the abundances of the light elements today. The weakness of the standard Hot Big Bang model is that it says nothing about how structure in the universe (galaxies and cluster of galaxies) could have originated.
In high-energy physics, the idea that the underlying theory of particles and their interactions has a high degree of symmetry which is broken at low energies forms the basis of the Weinberg-Salam model of the electroweak interactions. Over the last decade many predictions of this model have been confirmed, the discovery of the W and Z particles being the most dramatic. Based on the idea of symmetry breaking, theories which unify all the forces except gravity (grand unified theories), and theories including gravity (superstring theories) have been developed. Unfortunately, at present there are many different theories, and few ways of testing them.
One idea, which emerged from particle physics in the early 1980s, was that the same process which broke the symmetry between the particles and forces might break the spatial symmetry of the universe, producing the structure we see today. This is physically a very reasonable idea. In fact, similar processes happen in everyday substances. Most liquids are quite homogeneous and isotropic, which is not surprising, because there is nothing in the description of atoms and their interactions that singles out a particular direction or place in space as different from any other. Cool the liquid, however, and it freezes. The crystal structure of the solid picks a particular direction; but in different regions different directions are chosen. The result is that (unless the process happens very slowly) the solid is formed full of defects where there is a mismatch between neighboring crystalline regions.
<O_>figure&caption<O/>
Cosmic strings are very similar to these defects. They are predicted to occur by some grand unified theories and superstring theories. In the very early universe, at high temperature the fields in these theories are random, just as the atoms of a liquid. The density is quite uniform. As the universe cools below a certain temperature, some of the fields 'freeze.' As they do so, defects are formed, just as in solids.
In different theories, the defects may be at points, along lines, or in sheets. Cosmic strings are the linelike defects: They have special properties which make them well suited to forming structure later in the universe. For topological reasons they cannot have ends; they must form closed loops or continue on forever. The only way to change the length of a string is if it crosses itself and reconnects the other way (Fig. 1), chopping off a loop. This means that if one starts with some strings which wander right across the universe, there is no way to get rid of them completely. At best one can progressively chop more and more of the long string off into loops, and the loops can then radiate away (see Fig. 1). Thus some of the strings formed at very early times (around 10<sp_>-34<sp/> s after the Big Bang in most models predicting strings) survive right up to today.
<#FROWN:J02\> We have attempted to make these bibliographies as complete and comprehensive as possible. The intensity and broadening compilations cover the published literature form November 1983 through June 1990, and they include several earlier references overlooked in the previous compilations. The pressure-induced line shift compilation includes all references published through June 1990 that were available to us. We have also added a number of more recent references as they came to our attention in the course of preparing the present bibliography.
II. INTENSITIES
Detailed discussions of intensity parameters, units, and measurement techniques have been given in Refs. [1-2] and will not be repeated here. However, in order to properly interpret the values given in Tables II and III and in the individual references, we must discuss the difference between the vibrational band intensity, S<sb_><*_>nu<*/><sb/>, and the integrated band intensity S<sb_>Band<sb/>. S<sb_><*_>nu<*/><sb/> is related to the squared vibrational transition moment by
<O_>formula<O/>
where h is Planck's constant, c is the speed of light, <*_>nu<*/><sb_>0<sb/> is the wavenumber of the band center, N is the total number of molecules of the absorbing gas per cubic centimeter per atmosphere, Q<sb_><*_>nu<*/><sb/> is the vibrational partition function, and <O_>formula<O/> is the vibrational transition moment. S<sb_>Band<sb/> is defined as the summation of individual line intensities S<sp_>B<sp/><sb_>A<sb/> for all possible rotational transitions within the vibrational band; that is,
<O_>formula<O/>
where R<sp_>B<sp/><sb_>A<sb/> is the rotational factor, given by
<O_>formula<O/>
where S'<sb_>AB<sb/> is the dimensionless quantity called the line strength by Herzberg [3], E<sb_>A<sb/> is the energy of the lower state A, k is Boltzmann's constant, T is the gas temperature in degrees Kelvin, <*_>nu<*/><sb_>AB<sb/> is the wavenumber of the transition from lower state A to upper state B, and Q<sb_>r<sb/> is the rotational partition function. In Eqs. (1-3) we are using the approximation <O_>formula<O/>, where Q is the total internal partition sum; this approximation is valid except at very high temperatures.
The F-factor in Eq. (2) accounts for the effects of centrifugal distortion, Fermi- and Coriolis-type interactions, and other perturbations. For a rigid rotor, the F-factor would be 1, and S<sb_><*_>nu<*/><sb/> and S<sb_>Band<sb/> would be equal (assuming all possible rotational transitions are included in the summation). However, if the F-factor is significantly different from 1 for most of the rotational transitions, as the case for many CO<sb_>2<sb/> bands, S<sb_>Band<sb/> is usually larger than S<sb_><*_>nu<*/><sb/>. S<sb_>Band <sb/> values determined from the summation of individual line-intensity measurements using high-resolution techniques are more appropriate for comparison to low-resolution band intensity measurements (e.g., [4]). Therefore, where a given reference reports both S<sb_><*_>nu<*/><sb/> and S<sb_>Band<sb/> values, (e.g., [5]) we have chosen to include only the S<sb_>Band<sb/> values in Table 2. Unless the reported intensity is designated as S<sb_><*_>nu<*/><sb/> or S<sb_>Band<sb/> in the footnotes to the table, the reader should obtain more detailed information from the original reference.
Table I gives the various units in which band intensities have been reported in our literature survey, along with the multiplicative factors used to convert these units to our standard units (cm<sp_>-2<sp/> atm<sp_>-1<sp/> at 300K) for comparison purposes. However, because these multiplicative factors do not account for the temperature dependence of the partition functions or the exponential factors shown in Eq. (3), the converted values may not represent the true band intensity at 300K, particularly for transitions whose lower state is not the ground state. Goldman, Dang-Nhu and Bouanich [6] give an excellent brief discussion of the temperature dependence of line intensities.
Several of the references reporting intensity measurements of bands of CHCl<sb_>3<sb/>, CDCl<sb_>3<sb/>, CHCl<sb_>2<sb/>F, CHClF<sb_>2<sb/>, CHD<sb_>2<sb/>F, CHF<sb_>3<sb/>, and CDF<sb_>3<sb/> give the results not as S<sb_><*_>nu<*/><sb/> or S<sb_>Band<sb/> as defined in Eqs. (1-3), but as a quantity denoted as G. This G is identical to the <*_>GAMMA<*/> originally defined by Golike et al. [7] and is usually reported in units of (length)<sp_>2<sp/> mole<sp_>-1<sp/> or (length)<sp_>2<sp/> molecule<sp_>-1<sp/>. As previously stated in [2], G (or <*_>GAMMA<*/>) may be related to S<sb_>Band<sb/> approximately by S<sb_>Band<sb/> = G<*_>nu<*/><sb_>0<sb/>, where <*_>nu<*/><sb_>0<sb/> is the wavenumber of the observed band center.
To compare measurements by investigators in different laboratories and apply their results to atmospheric or astrophysical measurements, we must also understand the corrections applied for the isotopic composition of the measured gas samples. Unfortunately, in many papers, the isotopic composition is not specified, and the interpretation of 'natural abundance' is left to the reader. Johns [8] discussed in detail the difference between the 'natural abundance' of <sp_>13<sp/>C in atmospheric CO<sb_>2<sb/> and that in commercially supplied CO<sb_>2<sb/>. In our conversion of reported intensities to a consistent set of units, we have attempted to rescale each intensity value to represent the band intensity for a 100% pure sample of a single isotope. Where no specific isotopic abundance information has been given in the original reference, we have used the 1986 HITRAN abundances [9].
III. COLLISION BROADENING
For most of the range of pressures typically encountered in the terrestrial and planetary atmospheres, infrared vibration-rotation lines have predominantly the Lorentz, or collision-broadened line shape, in which the width of the line is linearly dependent on the gas pressure (for a fixed temperature). At very low gas pressures, the line shapes follow the Doppler profile, where the line width depends only on the temperature, mass of the molecule, and wavenumber of the transition. In the intermediate range where the Doppler and Lorentz profiles both contribute significantly, the line shape is most often modeled by the Voigt profile, which is the convolution of these two profiles. Detailed expressions for these three line profiles have been given in [2] and will not be repeated here. The dependence of the collision-broadened halfwidth on pressure and temperature is usually expressed as
<O_>formula<O/>
where b<sb_>L<sb/>(p,T) is the collision-broadened halfwidth of the line at measured pressure p and temperature T, b<sp_>0<sp/><sb_>L<sb/>(T<sb_>0<sb/>) is the collision-broadened halfwidth of the line at the reference pressure (usually 1 atm) and temperature T<sb_>0<sb/> (usually 296 K). According to the classical theory, the exponent n should have the value 0.5; however, many of the experimental results cited in Table IV of the present compilation and in [2], show quite different values for n. The experimentally determined values of this exponent appear to vary according to the absorbing and perturbing gases and the vibrational and rotational quantum numbers.
In recent years measurements of Lorentz broadening coefficients and their temperature dependence have become more numerous. There have been more measurements of individual line widths within a band, and very few studies reporting simply an average broadening coefficient (or n value) for an entire band. Recognizing that temperatures in the terrestrial upper atmosphere and in planetary atmospheres are quite different from ambient values in the laboratory, several investigators have pursued measurements of collision-broadened gas-phase spectra at low temperatures down to about 150K. Other investigators, driven by requirements to measure the spectra of combustion exhaust gases in situ, have recorded laboratory collision-broadened spectra at high temperatures.
Another are of increased interest has been the examination of line shapes that depart from the usual three forms (Lorentz, Doppler, Voigt) most commonly used to model laboratory, atmospheric, planetary, or astronomical spectra. Collisional narrowing is observed for spectral lines whose Lorentz halfwidth b<sb_>L<sb/> is very small. The Galatry line shape [10] is usually used to model the profiles of absorption lines affected by collisional narrowing in addition to Doppler and Lorentz broadening. Other phenomena that have been examined extensively include non-Lorentzian profiles in the far wings of very strong absorption lines (particularly for H<sb_>2<sb/>O and CO<sb_>2<sb/>) and line mixing (also called line coupling) in regions of very closely spaced lines such as Q-branches. References for laboratory studies of collisional narrowing, far-wing line shapes, and line mixing have also been included in Table IV.
IV. PRESSURE-INDUCED LINE SHIFTS
Shifts in the positions of vibration-rotation lines due to collisions with other gases have become a concern for certain types of infrared measurements. Since ambient pressures in the terrestrial and planetary atmospheres range from several Torr to several atmospheres, pressure-induced line shifts must be considered in spectroscopic studies of these atmospheres. Knowledge of line shifts is also important for certain transitions of CO<sb_>2<sb/>, CH<sb_>4<sb/>, and other gases that are used for stabilization of laser frequencies.
Prior to the mid-1980s, very few measurements of pressure-induced line shifts appeared in the literature; most of these were for diatomic molecules such as CO, HBr, HCl, HF, H<sb_>2<sb/>, and HD. More recent studies have provided extensive information on shifts in transitions of larger molecules, including CO<sb_>2<sb/>, H<sb_>2<sb/>O, N<sb_>2<sb/>O, O<sb_>3<sb/>, NH<sb_>3<sb/>, and CH<sb_>4<sb/>. However, some of these studies report measurements of shifts for only a small number of lines within a band.
Theoretical prediction of pressure-induced line shifts has been successful only for selected types of molecules such as polar diatomic molecules [11, 12]. Adequate models for pressure-induced shifts in bands of other molecules of interest for atmospheric studies, such as O<sb_>3<sb/> or CH<sb_>4<sb/>, have not yet appeared. Only a very few measurements of pressure-induced line shifts at temperatures far above or below room temperature have been reported, and the form of the temperature dependence of the shifts is not certain. A number of investigators, including Grossmann and Browell [13, 14], have modeled pressure-induced line shifts at low temperatures using a relation similar to that used for the Lorentz halfwidths:
<O_>formula<O/>
where <*_>delta<*/>(T) is the shift coefficient (in cm<sp_>-1<sp/>/atm) at the measured temperature T, <*_>delta<*/>(T<sb_>0<sb/>) is the shift coefficient at the reference temperature T<sb_>0<sb/>, and the exponent n' is empirically determined. However, the above expression is not valid in cases where the line shift changes sign with temperature. Other investigators, such as Houdeau and Boulet [15], have developed more rigorous theoretical models to account for the temperature dependence of both halfwidths and shifts.
V. EXPLANATION OF TABLES AND TABLE REFERENCES
The tables continue in the format of Refs. [1-2] with one exception. For carbon dioxide and its isotopic variants, we have adopted the vibrational band notation of Rothhman and Young [16], which is also used in the HITRAN and GEISA spectroscopic line parameters compilations [9, 17]. This notation is widely used by numerous investigators. Briefly, each vibrational level of CO<sb_>2<sb/> is designated by an integer whose five digits correspond to the sequence <O_>formula<O/>, where <*_>nu<*/><sb_>1<sb/>, <*_>nu<*/><sb_>2<sb/>, and <*_>nu<*/><sb_>3<sb/> are the vibrational quantum numbers, l is the degeneracy index of the <*_>nu<*/><sb_>2<sb/> vibrational mode, and r indicates the level's ranking in a Fermi resonance polyad. Thus, the ground state is designated by 00001, the <*_>nu<*/><sp_>1<sp/><sb_>2<sb/> level by 01101, the <*_>nu<*/><sb_>1<sb/> level by 10001, and the 2<*_>nu<*/><sp_>0<sp/><sb_>2<sb/> level by 10002. For each transition the level designations are given with the upper level first. For example, the <sp_>12<sp/>C<sp_>16<sp/>O<sb_>2<sb/> laser transition at 961 cm<sp_>-1<sp/>, designated <*_>nu<*/><sb_>3<sb/>-<*_>nu<*/><sb_>1<sb/> in the previous compilations, is indicated as 00011-10001 in the present work, and the <O_>formula<O/> laser transition at 1064 cm<sp_>-1<sp/> is indicated as 00011-10002.
The literature values and references for intensities of vibration-rotation transitions are given in Tables II and III. All of these values are for electric dipole allowed transitions measured in the gas phase with a few exceptions. These include the 'forbidden' 03301-00001 band of <sp_>12<sp/>C<sp_>16<sp/>O<sb_>2<sb/> at 2003 cm<sp_>-1<sp/>, liquid-phase measurements of several N<sb_>2<sb/>O bands that were published along with gas-phase measurements, and quadrupole and electrically-induced dipole vibration-rotation transitions of H<sb_>2<sb/> and N<sb_>2<sb/>. Purely rotational and electronic transitions are not included in the compilation.
As in the previous compilations, Table II contains the data for molecules with 2,3,4, and 5 atoms. The molecular data are sorted first by number of atoms, then by molecular formula in alphabetical order, then by isotopic form in order of increasing molecular weight. If the isotopic forms are not individually spectified, the reader should assume that the data refer to the most abundant isotope of the molecule. The first page of Table II serves as an index to the succeeding pages of the table. The bands for each molecule are presented in order of increasing wave number, with the assignments and band centers (in cm<sp_>-1<sp/>) given in the first column of the table. In the second column, headed Reference Value, are listed the integrated band intensity values as reported in the original references. Any unusual features of the reported measurements are indicated by footnotes in the relevant sections of the table.
<#FROWN:J03\>If this is indeed the case the data in Fig. 5 may be skewed upward as r/R decreases, since they could still reflect, in part, the highly aligning nature of the converging extensional flow at the tube entrance.
At the highest shear rate studied in this work, 1270 s<sp_>-1<sp/>, a value of 0.66 was found for S<sb_>11<sb/>, indicating essentially perfect fiber alignment (S<sb_>11<sb/> = 2/3).
A priori predictions may be obtained using Eqs. (3)-(8) together with the parameters given in Table III. That shown for S<sb_>11<sb/> in Fig. 5 is seen to portray the data quite well within their experimental uncertainty, and this was equally true at the other conditions studied. This is the most gratifying aspect of the analysis: the ability to make at least approximate predictions of fiber orientation using only the measured apparent viscosity function.
(2) The parameter S<sb_>12<sb/>, as noted earlier, is a measure of the skewness of the orientations about the flow direction. Intuitively one would expect that, for fibers aligned in a shearing flow, this parameter should be very close to zero, and all 14 sets of experimental data, including that shown in Fig. 5, bear this out. The theoretical prediction is thus simply incorrect.
(3) The components S<sb_>22<sb/> and S<sb_>33<sb/> are smaller than S<sb_>11<sb/>, as expected, and are consequently of lesser interest. The theoretical predictions are seen to be of the same general magnitude as the data but the agreement is not at all exact. Again, both data and theory show very little change with shear rate.
In conclusion, while further evaluation of the theory will be needed before fiber orientations can be predicted with confidence the orientation parameter of greatest interest, S<sb_>11<sb/>, and its variation with shear rate, appear to be predicted well. Further, very low levels of shearing rate appear to be sufficient for substantial fiber alignment in these concentrated systems.
Flow behavior and morphology of extruded polymers
Die swell and melt fracture are ordinary phenomena which occur during the extrusion of unfilled polymers from dies of any geometry. Die swell generally increases with extrusion rate, and melt fracture occurs at a shear stress near 0.1 MPa, as reported by Ramamurthy (1986) and others. For extrusion form a tube, die swell changes the diameter of the extruded filament and melt fracture may produce surface matteness as well as both longitudinal and radial ripples. Neither die swell nor melt fracture, however, significantly changes the shape of a cross section of the extruded filament from the circular shape induced by the die, except at very high stress levels well beyond those at which incipient irregularities are formed.
The extrusion of fiber-filled polyethylenes and polypropylenes from a circular die involves reduced die swell (as compared to the unfilled polymers) and the occurrence of gross surface irregularities not related to the phenomenon of melt fracture. The photographs in Figs. 6 and 7 illustrate the surface irregularities which occur during the extrusion of polypropylene containing 10 and 40 wt % glass fibers. These polymers were extruded at 230<*_>degree<*/>C from a 0.05 in. i.d. by 2 in. long capillary at the apparent shear rates shown in the figures. Similar morphologies were observed for the fiber-filled polyethylenes.
The photographs in Figs. 6 and 7 show that extraordinary surface irregularities and longitudinal variations in the diameters of the filaments occurred at the lowest extrusion rates. Numerous fibers protruded from the surface of the filament giving them a fuzzy appearance, and the orientations of the protruding fibers seemed to be random. As extrusion rates increased, the surface irregularities faded and the diameters of the filaments became much more uniform. The measured shear stresses for the filaments in Figs. 6 and 7 obtained at 4.6 and 115 s <sp_>-1<sp/> were well below the melt fracture stress of 0.1 MPa, and the shear stresses at 1272 s<sp_>-1<sp/> were only slightly above the critical value. Melt fracture, therefore, did not seem to be a contributing factor in the observed surface irregularities, nor was any indication of sudden changes in surface irregularity observed at the critical stress of 0.1 MPa.
<O_>figures&captures<O/>
Wu (1979), Crowson et al. (1980), and Knuttson and White (1981) have reported similar surface morphologies for fiber-filled polymer melts. Knuttson and White reported that the surface of fiber-reinforced polycarbonate extrudates was significantly rougher under all conditions than that of the unfilled polymer, but the smoothness of the extrudates increased with extrusion rate. They attributed the increased smoothness to increased fiber orientation in the direction of flow, although we have noted only modest changes for our materials. Crowson et al. observed a similar improvement in smoothness with increased extrusion rate, and they found that shear flow produced a decrease in fiber alignment parallel to the flow direction. Wu in his study of glass fiber-filled poly(ethylene terephthalate), observed three distinct regimes of extrudate surface morphology: a smooth surface at low shear rates, irregular surface at medium shear rates, and a somewhat smoother extrudate at high rates. Wu suggested that the observed morphologies arose from normal stress effects and rotation of fibers in the velocity field. Thus we see that a comprehensive and consistent understanding of physical concepts leading to such extrudate morphologies has not been developed.
Cross sections of the extruded glass fiber-filled polypropylenes are shown in Figs. 8 and 9 for fiber loadings of 10 and 40 wt %, respectively. The cross sections in these figures are digitized representations obtained from photomicrographs. The sections are perpendicular to the axis of the filaments, and the shear rates correspond to those in Figs. 6 and 7. The points in Figs. 8 and 9 represent the locations of individual, well-aligned fibers. The most striking observation to be made from Figs. 8 and 9 is that the shapes of the cross sections are extremely distorted at low shear rates and become largely circular at high rates.
<O_>figures&captures<O/>
The same shear rate dependence of morphology was observed for suspensions extruded from capillaries with lengths of 1, 2, and 3 in. The distortion disappeared, however, when the materials were extruded through an orifice. Cross sections of extrudates of the 10 wt % suspension in polypropylene obtained from the orifice are shown in Fig. 10. The orifice had the same diameter (0.05 in.) as the capillaries for which surface irregularities were observed, and it retained the 90<*_>degree<*/> convergence at the entrance. For the orifice cross sections in Fig. 10, there appears to be slight reversal in the distortion of the cross-sectional shape; i.e., the shape is quite circular and smooth at the lowest rate and becomes more jagged at the highest rate.
<O_>figure&caption<O/>
Factors affecting surface morphology
A number of factors were considered for the development of a plausible mechanism for the observed surface morphologies. As pointed out in the previous section, melt fracture did not appear to play a role since the shear stresses at which distortion occurred were too low. Other factors considered by Becraft (1988) were die swell, thermal connection of the polymer, surface effects between the polymer and metal, flexing of fibers near the surface of the extruded filaments, expansion of gas bubbles in the material, regions of local fiber alignment, stress variations resulting from fiber concentration inhomogeneities, and velocity profile rearrangements at the tube exit. The more interesting and pertinent of these considerations are as follows.
Die swell
Die swell was significantly reduced in the fiber-filled polymers and it was unlikely that the large distortions observed in the extrudates resulted from swelling of the polymer. Die swell increased with extrusion rate even for the filled polymers. Therefore, if swelling were the relevant mechanism, extrudate distortion should have increased with shear rate, not decreased as observed.
Fiber flexing at the surface of extrudates
The glass fibers in the polyethylenes and polypropylenes are somewhat flexible, and the flexibility increases with temperature. This creates the opportunity, with a large enough fiber flexural modulus, for misaligned or bent fibers near the surface of the extruded polymers to 'spring out' while the material is still in a molten state. During this straightening process the fiber may entrain polymer and conceivably cause the nonuniformities observed in the cross sections. The photograph in Fig. 11 illustrates entrainment of polymer by a fiber protruding from the surface. This photograph was taken at the surface of a distorted extrudate using polarized light and a magnification of 90x. An envelope of polymer surrounds the protruding fiber and a surface irregularity has been created by the entrained polymer. Protruding fibers in all of the samples showed similar entrainment of polymer.
To pursue this possible origin of the surface morphology the dynamics of flow from the tube was observed using a Panasonic VHS camcorder (model PV200). In the recorded images at low shear rates, fibers were visibly springing out from the surface of the extrudates immediately after the material emerged from the capillary. All fiber motion appeared to be complete within one or two diameters of the tube exit. It was not possible, however, to determine from the flow visualization whether the springing out of fibers was a consequence of long fiber flexibility or of rotation of rigid fibers. A random side-to-side motion of the exiting filaments was also evident, and was produced by forces originating at or prior to the tube exit. At higher shear rates, the frequency of the side-to-side motion increased dramatically, and the movement resembled that of a vibrating string although with no obvious periodicity.
Fiber flexing and 'spring back' of bent fibers implies the occurrence of bent fibers within the polymer melt during flow into and through the tube. No such bent fibers were ever found in any of these extrudates, possibly because cooling of the emerging extrudate in the air was slower than the rate of springback of the fibers. To check, the extrudates were also extruded into a cold glycol-water bath held within one diameter of the capillary exit (a closer positioning was not possible). Although perfect quenching could not be achieved significant differences between quenched and unquenched extrudates could be found. The quenched extrudates had only much shorter fibers protruding from the surface, and examination of quenched sections under polarized light revealed numerous bent fibers within the frozen polymer. Clearly, then, fiber flexing contributes to the observed extrudate irregularities.
Fiber concentration inhomogeneities
Evidence of local variations in fiber concentration appears in the cross section of Figs. 8 -10. Close examination of these cross sections reveals that there are regions that seem to be completely free of fibers, and other regions which have high densities of fibers. Such variations in fiber concentration may cause point-to-point variations in viscosity and a variable deformation rate profile across the diameter of the molten polymer as it flows. The effects would be greatest at low deformation rates when the local viscosity variations due to differences in fiber concentration (Figs.2 - 4) are greatest; at increasing shear rates the viscosity differences diminish and a smoother extrudate would appear, in attractive agreement with that observation. Radial fiber migrations were reported for poly(ethylene terephthalate)-glass systems by Wu (1979) but, in the present work on polypropylene-glass fiber systems careful measurements revealed no such large-scale changes. However, local concentration fluctuations were indeed of large magnitude: in the 10 wt % material the volume fractions of fibers varied between 0 and 0.15 and, for the 40 wt % system, between 0.09 and 0.35. Using the data of Kitano et al. (1981) to estimate the local viscosity variations tenfold changes are computed. To determine more definitively the probable effects of large viscosity variations on extrudate morphology, an experiment was conducted using a 50/50 wt % blend of high density polyethylene and polypropylene. The viscosities of these two immiscible polymers varied by nearly an order of magnitude at low deformation rates and a temperature of 180 <*_>degree<*/>C. The blend was prepared by chopping extruded strands of the individual polymers into fragments 1 mm or less in length and mechanically mixing an equal mass of each of the solid fragments. The blend was extruded through a 2 in. long capillary with an L/D of 40, and cross sections of the extrudates were examined for distortion.
<#FROWN:J04\>We also derive, for the case of random sphere percolation, an exact Smoluchowski equation describing the scaling behavior of the mean cluster densities. In Sec. III we focus on the function n<sb_>c<sb/>(k), which gives the mean number of clusters, per unit volume, containing exactly k particles. We obtain an exact differential equation for this quantity and discuss its solution. We also give an efficient general numerical method for solving this equation and, thus, for obtaining cluster-size distributions. In Sec. IV we extend these methods to clustering in ionic systems. In Sec. V, we show how to calculate the mean volume of a cluster as a function of cluster size. To do this, we first extend earlier research by Fanti et al. [20] to give a recipe for calculating the mean volume of a cluster. We then use the ghost-field method developed in previous sections to obtain the full distribution of mean cluster volumes. Section VI gives our conclusions. In the Appendix, we give a closed set of Ornstein-Zernike equations that are satisfied by the asymmetric (ghost-field-dependent) connectedness functions. We discuss the difficulties in obtaining the mean-spherical approximation (MSA) for this system. Finally, we discuss the relationship between this system of equations and similar systems encountered in other random media problems, including correlations of a liquid in porous media and structure of the gel phase in a percolating medium.
II. GENERATING FUNCTION FORMALISM FOR CLUSTER STATISTICS
In this section we show that including a color field, or ghost field, in a continuum percolation model allows one to develop generating functions for the basic quantities describing such a model. We develop these generating functions using the continuum-Potts-model formalism already applied to percolation without a ghost field. Certain basic relations between these generating functions are also developed. In particular, we show that the virial theorem for percolation models provides an exact equation of Smoluchowski type describing the growth and coalescence of clusters as the density is increased.
The one-state limit of the continuum Potts model [21-23] gives a very general correlated, continuum percolation model. We have previously [12] discussed in detail this relationship, which is the continuum generalization of the mapping originally developed for lattice percolation by Fortuin and Kastelyn [14]. Here we add a ghost field to the Hamiltonian in order to develop the basic quantities describing this model as generating functions. The ghost field [16, 18] is an artificial external field of constant magnitude that acts on particles in all but one species (we choose species 1 for definiteness). Since the correspondence between Potts-model quantities and percolation quantities has been developed in detail previously, we do not repeat this work here. We rather focus on the additional structure created by including the ghost field in the CPM.
The asymmetric CPM is a generalization of the Widom-Rowlinson model [24]; it describes a mixture of s different species which can undergo phase separation at sufficiently low temperatures. In the one-state limit, this phase-separation transition becomes a percolation transition. This mapping exhibits continuum percolation as an analytic continuation of a many-body system of a standard type, and allows the methods of liquid-state physics to be applied to it. The Hamiltonian of this system is
<O_>formula<O/>.
The first term is a repulsive interaction [we will take <*_>nu<*/>(x)>0] acting only between particles of different species (we use the symbol <*_>sigma<*/><sb_>i<sb/> to denote the species of particle i). The second term is a species-independent potential acting between each pair of particles. In the one-species limit, the thermodynamic quantities describing this model become the basic quantities describing a very general model of clustering and percolation in the continuum. This model consists of particles with interparticle correlations induced by the potential <*_>phi<*/>(x) and bonds connecting each pair of particles with a separation-dependent bond probability p<sb_>b<sb/>(x) given by
<O_>formula<O/>.
The last term in (2.1) is a ghost field which acts equally on particles of all species except those in one distinguished species (we choose species 1 for definiteness; henceforth we will use subscripts i and j to denote species other than 1). This term can be absorbed into the fugacities characterizing different species; if particles of species 1 have activity z<sb_>1<sb/>, then particles of species i have activity
<O_>formula<O/>,
with H=<*_>beta<*/>h, <*_>beta<*/>=1/kT the inverse temperature and h the ghost field from (2.1). If particles in species 1 are present with density <*_>rho<*/><sb_>1<sb/>, those in every other species have density <*_>rho<*/><sb_>i<sb/>, a nontrivial function of the fugacities z<sb_>1<sb/>,z<sb_>i<sb/>. The one-state limits of these quantities are
<O_>formula<O/>,
where <*_>unch<*/> is the percolation density, while
<O_>formula<O/>,
where n<sb_>c<sb/>(K) is the density, or number per unit volume, of clusters containing exactly k particles. The quantity <*_>unch<*/><sb_>i<sb/>(H) is a generating function for the cluster densities n<sb_>c<sb/>(k). Furthermore, if the field H were pure imaginary, we could recover the individual cluster densities from the function <*_>unch<*/><sb_>i<sb/>(H) by performing an inverse Fourier transform. The convergence of the series (2.5) for purely imaginary values of H follows from the exponential decay, as a function of cluster size k, of the {n<sb_>c<sb/>(k)}. We will make this program explicit in Sec. III. Here, we will assume that inversions [25] of this kind are in fact possible and focus on presenting formal results. That is, we will give the other basic generating functions provided by the CPM mapping and state the results of performing the inverse transform just described. Note that, according to (2.5), setting H to zero gives<O_>formula<O/> [we treat here only the regime below the percolation transition, and thus all particles are contained in finite clusters; see, however, the discussion below Eq. (A18) in the Appendix].
The other low-order moments of the generating function are also given by basic CPM quantities. The one-state limit of the pressure is
<O_>formula<O/>.
We can also define the H-dependent version of the mean cluster size,
<O_>formula<O/>.
We give an integral equation satisfied by this function below. The one-species limits of the distribution functions are also generating functions for useful quantities, as we now show. As before, we will use subscripts i and j to denote two different species other than species 1, which we have distinguished by applying a ghost field. In the one-species limit s<*_>arrow<*/>1, the CPM distribution functions give connectedness and blocking functions, denoted g<sb_>c<sb/>(x,H) and g<sb_>b<sb/>(x,H), respectively, according to
<O_>formulae<O/>.
The zero-field limits of these functions, which we write simply as g<sb_>b<sb/>(x) and g<sb_>c<sb/>(x), respectively, have a direct physical interpretation: The connectedness function g<sb_>c<sb/>(x) is the probability density associated with finding two particles with separation x both in the same cluster. Similarly, the blocking function g<sb_>b<sb/>(x) is the probability density associated with finding two particles with separation x but in different clusters. These two functions are related by
<O_>formula<O/>,
with g<sb_>t<sb/>(x) the thermal correlation function associated with the potential interaction <*_>phi<*/>(x). For the nonzero H field, the H-dependent CPM distribution functions, in the s<*_>arrow<*/>1 limit, give generating functions for blocking and connectedness distribution functions restricted to clusters of specified size, which we denote <O_>formula<O/>, <O_>formula<O/>, according to
<O_>formulae<O/>.
Here <O_>formula<O/> is the probability density associated with finding two particles separated by a distance x, but in different clusters, one of size m, one of size n. Also, <O_>formula<O/> is the probability density associated with finding two particles separated by distance x and both in the same m-particle cluster. An explicit system of Ornstein-Zernike equations obeyed by the correlation functions of an asymmetric CPM is presented in the Appendix. Our strategy is to solve these equations in the s<*_>arrow<*/>1 limit, and use their solutions, as indicated in (2.11) and (2.12) to recover the H-dependent connectedness functions. The function g<sb_>ij<sb/> is itself useful as a generating function. It is valuable as an intermediate quantity because it obeys the relation <O_>formula<O/>. It is also important because it obeys the Ornstein-Zernike equation
<O_>formula<O/>.
Here the symbol <*_>unch<*/> denotes a convolution. A basic equation relating the above generating functions is the virial theorem for the CPM, and its s<*_>arrow<*/>1 limit, the virial theorem for percolation. This theorem has been developed [12] for the general correlated percolation model defined below Eq. (2.1). However, for simplicity, we discuss here only the case of random sphere percolation, in which randomly located particles are directly connected if and only if they are closer together than a fixed distance a. The virial theorem for the zero-field case of this model is
<O_>formula<O/>.
For the model of sphere percolation, n<sb_>c<sb/> depends only on the dimensionless variable <O_>formula<O/>. Thus we can rewrite
<O_>formula<O/>.
This equation can be given a direct physical interpretation by identifying the particle diameter a with the time variable in a particular growth process. Specifically, we imagine randomly distributed particles with zero diameter at time t=0, with their diameters growing at a constant rate so as to remain proportional to the time. As the particles grow, they overlap, and form successively larger clusters. Now we focus on a specific particle at the moment when its diameter has just grown to size a. The quantity <O_>formula<O/> gives the density of particles that are in different clusters from the particle in question, and are about to become directly connected to it. The factor 4<*_>pi<*/>a<sp_>2<sp/> gives the total surface area on which this can happen, while the factor of 1/2 prevents double counting. Such coalescence events are precisely the ones that decrease the total number of clusters.
We can extend this result to show that the cluster densities {n<sb_>c<sb/>(k)} for random continuum percolation obey an exact equation of Smoluchowski type. To see this, we write down the virial theorem for our asymmetric Potts model, analytically continue the ghost field H to purely imaginary values, and then perform an inverse Fourier transform. It is convenient to express the result in terms of reaction rates describing the fusion and breakup of clusters. To do this, we define an H-dependent blocking function g<sb_>b<sb/>(x,H) in terms of the function <*_>rho<*/><sb_>b<sb/>(x,H) already defined:
<O_>formula<O/>.
This function, evaluated at contact, is a generating function for reaction rates according to
<O_>formula<O/>.
In terms of these quantities, we can write the following equation of Smoluchowski type for the cluster densities:
<O_>formula<O_>.
We pause to note the connection between this equation and the form of Smoluchowski equation which is furnished by a mean-field treatment. If we replace the function g<sb_>b<sb/>(x,H) on the right-hand side (RHS) of Eq. (2.16) by the function g<sb_>b<sb/>(x), i.e., if we ignore the H dependence of this function, the summations over the subscripts m and n in the terms on the RHS of Eq. (2.18) will be eliminated. If we set m=0 and n=0 in these terms, a mean-field Smoluchowski equation results. The probabilistic interpretation of such an equation is immediate: The first term on the RHS gives the rate at which pairs of smaller clusters coalesce to give k clusters. The second term on the RHS gives the rate at which k clusters are removed by fusing into larger clusters. Equation (2.18) thus has the form of a generalized Smoluchowski equation [26]; we emphasize that this equation is exact. Here the contact values of blocking functions play the role of rate constants. These contact values can be derived for random sphere percolation by using an asymmetric version of the scaled-particle theory for percolation [13]. In the general case a variety of integral equation methods have been developed for calculating these quantities, as we discuss in Sec. III and also in the Appendix.
III. GENERAL ALGORITHM FOR CLUSTER-SIZE DISTRIBUTIONS
In this section we discuss the cluster-size distribution in a very general correlated continuum percolation model. We first use the formalism of Sec. II to develop a differential equation for the quantity <O_>formula<O/>, which, as we have shown, is the generating function for the cluster densities n<sb_>c<sb/>(k). The RHS of this differential equation involves the volume integral of the H-dependent connectedness function <*_>rho<*/><sb_>c<sb/>(x,H). Numerical solution of this equation is computationally demanding; we discuss efficient methods for solving this equation.
The analog, for percolation, of the compressibility theorem is the theorem giving the mean cluster size as an integral over the two-point connectedness function. The probabilistic argument that gives this theorem has a form specific to k clusters which we now describe.
<#FROWN:J05\>Thus, each finger is related by a rotation of 3 x 32<*_>degree<*/> and a translation of about 10 A<*_>circlet<*/> (3 x 3.3 A<*_>circlet<*/>) along the DNA axis. Unlike the recognition helix of the helix-turn-helix motif, the <*_>alpha<*/> helix in the zinc finger is tipped at an angle away from the major groove. The <*_>beta<*/> sheet is on the back of the helix away from the base pairs and is shifted towards one side of the major groove. The two strands of the <*_>beta<*/> sheet have different roles. The first <*_>beta<*/> strand does not make contact with the DNA, whereas the second <*_>beta<*/> strand is in contact with the sugar phosphate backbone along one strand of DNA.
The zinc finger peptide makes 11 hydrogen bonds with the bases in the major groove. Six amino acid side chains interact with the base pairs, two from each zinc finger. However, five of the six residues that are binding to the nucleic acids are arginine residues. One of these arginines immediately precedes the <*_>alpha<*/> helix in each of the three fingers, and it also includes the residues from either the second, the third, or the sixth residues in the <*_>alpha<*/> helix. All of these form hydrogen bonds with the G-rich strand of the consensus binding site. Figure 7 shows the interactions of the various arginine residues and one histidine residue from the three fingers. It can be seen that five of the six interactions are those postulated at an earlier time in an analysis of protein-nucleic acid recognition (Fig. 3).
Each zinc finger interacts with a subset of three bases. All of them interact with the first base in the subset. However, two of them interact with the third base of the subset and one with the second base. The detailed interactions are listed in Table II and are also shown in Fig. 7. Some of the arginine residues binding to guanine are also stabilized by aspartic acids that occur as the second residue in the <*_>alpha<*/> helices. It is likely that this side-chain interaction helps to stabilize the binding of the arginine to the guanine residue, although it is not found in all of the fingers. The recognition system is relatively simple. The residue immediately preceding the <*_>alpha<*/> helix contacts the third base on the primary strand of the substrate at the 5' end. The third residue on the <*_>alpha<*/> helix can contact the second base on the primary strand, and the sixth residue can contact the first base. In this structure, each zinc finger came in contact with only two bases of each three-base subset. It is not known whether it would be possible to have other zinc fingers in which all three bases are recognized.
The DNA is essentially in the form of a B-type helix with small distortions. There are small changes in base-pair twist going from one site to the next, although the overall conformation is similar to normal B-DNA. Each of the three fingers binds in a similar orientation and has similar contacts with the three base pairs of the DNA. In a formal sense, the relationship of one zinc finger to the next is in the form of a translational rotation or a screw operation that tracks the inside of the major groove of the double helix.
Even though the zinc finger uses the <*_>alpha<*/> helix for recognition, it has several unique features that differentiate it from the helix-turn-helix interactions. First, of all, the zinc finger complex is formed out of modular units that can be repeated a large number of times. This entails the possibility of recognizing very long stretches of DNA by simply having a larger number of zinc fingers. As mentioned, some proteins have very large numbers of zinc fingers and may actually use this. The second characteristic is that the contacts seem to be largely with one strand only of DNA, in this particular case with the purine-rich or guanine-rich strand. The recognition depends largely on interaction with the bases, and there are fewer hydrogen bonds with the DNA backbone than are seen in the other structures.
Even though studies of other protein-DNA complexes appear not to have a recognition code, the zinc finger complex appears to have a recognition code which is largely based on the arginine-guanine contacts, at least for the Zif complex. It remains to be seen whether these types of contacts will be found more specifically when the structures of more zinc fingers have been done. Solution of this structure will make it possible to synthesize new zinc finger binding domains with different nucleotide binding specificities. Thus, it will be possible to explore the full gamut of interactions found in this widely used recognition motif in eukaryotic systems.
THE GLUCOCORTICOID RECEPTOR ALSO CONTAINS ZINC IONS
The glucocorticoid receptor has the property of binding a hormone, such as estrogen or another steroid, and is then translocated from the cytoplasm to the nucleus, where it binds to specific DNA sequences, called glucocorticoid response elements (GREs). The binding affects transcription of the genes. A large number of these exist; they include receptors for steroid hormones, retinoids, vitamin D, thyroid hormones, and others. Members of this superfamily have a common amino acid sequence organization with discrete domains that are used for binding DNA as well as zinc. All of these nuclear receptors are characterized by a pattern containing eight cystenes and, in the glucocorticoid receptor, these cystenes coordinate two zinc ions in a tetrahedral manner. The structure of the glucocorticoid receptor bound to DNA has been determined recently by Luisi et al. (1991). Unlike the typical zinc fingers, the glucocorticoid receptor forms a distinct globular binding domain and does not occur in a long series of modular units, as is often found in the typical zinc finger DNA binding.
The structure of the glucocorticoid receptor bound to DNA reveals that the receptor dimerizes onto a DNA molecule that contains two repeats of the glucocorticoid response element sequence, each with the major groove facing in the same direction. Each of the proteins forms a compact globular structure in which the two zinc ions serve to nucleate the formation of a conformation in which an <*_>alpha<*/> helix is positioned in the major groove of B-DNA and thereby has sequence-specific binding. The glucocorticoid receptor conformation may be looked upon as a conformation similar in some respects to the helix-turn-helix conformation, except that zinc ions are used in maintaining the stable fold of the protein rather than the helix interactions found in the helix-turn-helix system. A number of interactions are found between the glucocorticoid receptor and the DNA. However, three of them are interactions with bases that are important for determining sequence specificity. One of the most important of these is an arginine 466 that binds to guanine 4 using the system of arginine-guanine interactions, which has been described above using two hydrogen bonds. Another hydrophobic interaction involves valine 462 interacting with the methyl group of thymine 5 while a lysine 461 forms a single hydrogen bond to N7 of guanine 7 as well as to a water molecule, which in turn binds to O6 of guanine and O4 of thymine in an adjacent base pair. If arginine 466 is replaced by lysine or glycine, the protein no longer functions in vivo. Arginine 466 and lysine 461 are absolutely conserved in the superfamily of nuclear receptors; their targeted bases, guanine 4 and guanine 7, also occur consistently in all the known sequences of the hormone response elements (Luisi et al., 1991).
The major difference between the zinc-containing glucocorticoid response element and the traditional zinc finger is the fact that the latter conformation is stabilized individually by an extensive hydrophobic core as well as by the zinc ion. Furthermore, it assumes this conformation independent of the presence or absence of DNA. In contrast, experiments with the glucocorticoid receptor show that it only condenses as a dimer in the presence of the DNA. The dimerization is stabilized both by the DNA as well as by contacts between the protein.
The arginine-guanine interaction, which played so predominant a part in five of the six interactions seen in the three modules of the traditional zinc finger structure, also plays a role in interactions with the glucocorticoid response element. However, in this case, only one of the three interactions that are sequence-determining involves the arginine-guanine interaction.
ECO RI ENDONUCLEASE BINDING TO DNA
Restriction endonucleases are very important tools in molecular biology since they cleave DNA molecules at specific sequences. One of the widely used enzymes was obtained from E. coli and is called Eco RI endonuclease. It cleaves DNA at a specific double-stranded sequence (d(GAATTC)). Eco RI contains 276 amino acids, and it has been crystallized with a fragment of DNA containing 13 base pairs. The solved structure revealed a complex interaction between a globular protein and a DNA double helix (McClarin et al., 1986; Kim et al., 1986). The DNA recognition motif consists of a parallel bundle of four <*_>alpha<*/> helices penetrating the major groove of the DNA. There, amino acids at the end of the <*_>alpha<*/> helix interact with the bases in the major groove. Although <*_>alpha<*/> helices are employed, this motif differs from the interaction seen both in the helix-turn-helix proteins and the zinc fingers. In this case, a cluster of very long <*_>alpha<*/>-helical segments interact with the DNA at their ends. In addition, a segment of extended polypeptide chain runs along the major groove of the DNA roughly parallel to the DNA backbone. This is anchored at one end by one of the recognition helices, and it has several contacts with bases. Among the interactions that are described is one involving arginine 200 binding to guanine in a manner similar to that described in Fig. 3.
We do not know whether this structure is likely to be a general structure for the recognition of DNA by restriction endonucleases. However, one of the interesting projects arising from solution of this protein-DNA complex is the possibility of modifying side chains to alter recognition modes so that one might be able to make restriction enzymes with altered cleavage specificities using the Eco RI framework for carrying this out. Further work will be necessary before we know whether this is a general recognition motif for other enzymes as well. However, it is important to emphasize that the mode of interaction is quite distinct from that seen in any other protein-nucleic acid cocrystal.
<*_>beta<*/> SHEET DNA BINDING PROTEINS
The methionine repressor controls its own gene as well as structural genes for enzymes involved in the synthesis of methionine. It is a protein with 104 amino acids and forms stable dimers in solution. The structure had been determined by Phillips and colleagues, and it consists of two highly intertwined monomers that form a two-stranded antiparallel <*_>beta<*/> sheet with one strand coming from each monomer (Rafferty et al., 1989). This <*_>beta<*/> sheet forms a protrusion on the surface of the molecule. A similar structure has been deduced for the Arc repressor based on NMR studies (Kaptain et al., 1985). Phillips has also solved the structure of the Met repressor bound to a synthetic DNA fragment containing 18 base pairs (S. Phillips, personal communication). The structure of the Met dimer is not changed greatly by binding to the DNA, which is largely in the B conformation. The two-stranded <*_>beta<*/> sheet of the repressor is found in the major groove of the DNA with side chains from the <*_>beta<*/> strands interacting with base pairs within the operator sequences. These interactions are the base sequence-specific interactions. The DNA itself is somewhat kinked in the center of the operator sequence. That has the effect of narrowing the major groove slightly so that it can form closer bonding to the side chains of the <*_>beta<*/> sheet.
The two-stranded <*_>beta<*/> sheet is thus another DNA binding motif which, unlike the others mentioned above, does not use an <*_>alpha<*/> helix for recognition but rather an extended polypeptide chain.
The listing of protein structural motifs that are involved in recognizing DNA sequences (Table I) is necessarily incomplete.
<#FROWN:J06\>
II. EXPERIMENT
The molecular beam time-of-flight mass spectrometer (TOFMS) used in this study is identical to that in Paper I, to which the reader is referred. A given C<sb_>6<sb/>H<sb_>6<sb/>-(CH<sb_>3<sb/>OH)<sb_>n<sb/> neutral cluster size is selected for photoionization using resonance enhancement through the S<sb_>0<sb/>-S<sb_>1<sb/> transitions of the C<sb_>6<sb/>H<sb_>6<sb/> chromophore in the cluster. Resonant two-photon ionization scans utilize the unfocused, doubled output of an excimer-pumped dye laser. Typical peak UV laser powers are 3x10<sp_>5<sp/> W/cm<sp_>2<sp/>. Laser power studies indicate that the observed product ions result from two-photon processes. Relative product yields are determined from scans over the resonant features of the reactant complex while simultaneously monitoring ion signals from all relevant product mass channels using a 100 MHz digital oscilloscope.
III. RESULTS
In Paper I, we reported on the spectroscopy of neutral C<sb_>6<sb/>H<sb_>6<sb/>-(CH<sb_>3<sb/>OH)<sb_>n<sb/> clusters. In that case, attention was focused on R2PI spectra taken monitoring unreactive ion mass channels [C<sb_>6<sb/>H<sb_>6<sb/>-(CH<sb_>3<sb/>OH)<sb_>n<sb/>]<sp_>+<sp/>. Figures 1-3 present a series of scans over the same 6<sp_>1<sp/><sb_>0<sb/> region including scans monitoring mass channels which arise from intracluster ion-molecule chemistry. As is readily apparent from these spectra, resonant two-photon ionized C<sb_>6<sb/>H<sb_>6<sb/>-(CH<sb_>3<sb/>OH)<sb_>n<sb/> clusters with n<*_>unch<*/>3 [Eq. (1)] react by dissociative electron transfer (DET) to form (CH<sb_>3<sb/>HO)<sb_>n<sb/><sp_>+<sp/> [Eq. (3)], while those with n<*_>unch<*/>4 also undergo dissociative proton transfer (DPT) to form H<sp_>+<sp/>(CH<sb_>3<sb/>OH)<sb_>n<sb/> ions [Eq. (4)]. This intracluster ion chemistry is completely absent from the resonantly photoionized C<sb_>6<sb/>H<sb_>6<sb/>-(H<sb_>2<sb/>O)<sb_>n<sb/> clusters and from C<sb_>6<sb/>H<sb_>6<sb/>-(CH<sb_>3<sb/>OH)<sb_>n<sb/> with n<*_>unch<*/>2 [Fig. 1(b)]. In all cases, the ion chemistry occurs in competition with fragmentation of the cluster [Eq. (2)] via loss of one (or sometimes two) methanol molecules.
<O_>formulae<O/>
The dissociative electron transfer channel to form [(CH<sb_>3<sb/>OH)<sb_>n<sb/><sp_>+<sp/>] is quite unexpected since only protonated methanol cluster ions are observed in either electron bombardment or photoionized pure methanol clusters. While the resolution of our reflectron TOFMS is easily capable of distinguishing protonated from unprotonated clusters (<O_>formula<O/>), we have as an additional check carried out resonant two-photon ionization (R2PI) scans using CH<sb_>3<sb/>OD which confirm that the major product is [(CH<sb_>3<sb/>OD)<sb_>n<sb/><sp_>+<sp/>] and not [(CH<sb_>3<sb/>OD)<sb_>n<sb/>D]<sp_>+<sp/> or [(CH<sb_>3<sb/>OD)<sb_>n<sb/>H]<sp_>+<sp/>.
Figures 4 and 5 present the relative product yields for the observed fragmentation, DET, and DPT channels following resonant ionization through the 6<sp_>1<sp/><sb_>0<sb/> and <O_>formula<O/> transitions of the 1:2, 1:3, 1:4, and 1:5 clusters. These yields are obtained by integrating the peak intensities of the 6<sp_>1<sp/><sb_>0<sb/> or <O_>formula<O/> ion signals in the relevant product channels in order to avoid nonresonant contributions from the signal. The measurements at <O_>formula<O/> increase the maximum internal energy in the cluster ion (determined by the photon energy) by 5.2 kcal/mol from those using the 6<sp_>1<sp/><sb_>0<sb/> transition as the intermediate state.
<O_>figure&caption<O/>
IV. DISCUSSION
One factor which plays a major role in determining the presence and efficiency of a given product channel is the energetic threshold for the channel relative to the internal energy of the photoionized reactant cluster. In photoionization, a distribution of ion internal energies is produced which reflect both direct ionization and autoionization processes in the cluster. In the free C<sb_>6<sb/>H<sb_>6<sb/> molecule, R2PI through the 6<sp_>1<sp/><sb_>0<sb/> transition reaches 8 kcal/mol above the ionization threshold. In that case, the largely <*_>DELTA<*/><*_>nu<*/>=0 Franck-Condon factors between the S<sb_>1<sb/> state of the neutral and the ground states of the ion result in the electron taking away most of the excess energy as kinetic energy.
<O_>figure&caption<O/>
In clusters, on the other hand, the nature and strength of the intermolecular forces change significantly upon ionization, often leading to very different lowest-energy structures for the neutral and ionic clusters. This is particularly true for the present clusters containing polar methanol molecules as solvent, where the good Franck-Condon factors between the S<sb_>1<sb/> state and the ion are to regions of the ionic potential energy surface far above the adiabatic ionization threshold for the cluster. As we saw in Paper I, the efficient fragmentation of the photoionized clusters is one result of the high ion internal energies produced. Here, since the same photoionization process also initiates the intracluster ion chemistry, we expect to form a distribution of reactant cluster ion internal energies which favors ion internal energies near the maximum allowed by the photon energy.
Figures 6-8 present energy level diagrams reflecting our best estimates of the energies of fragmentation, DET, and DPT product thresholds relative to the maximum ion energies produced by photoionization. Experimentally observed product channels are highlighted by placing them in boxes. In the figures, the zero of the energy level scales is taken to be the energy of the <O_>formula<O/> asymptote. Table I collects the heats of formation of the relevant species. The thermochemistry of the protonated methanol clusters is known with good accuracy by virtue of several studies of these clusters. The neutral cluster binding energies are not known from experiment. We estimate them using the calculations of Paper I after approximate correction for zero point energy effects.
<O_>figure&caption<O/>
An upper bound for the energy of the DET channel [Eq. (3)] on our relative scale is determined by the threshold for reaction (5)
<O_>formula<O/>
whose thermochemistry is determined by well-known heats of formation for all species. If the CH<sb_>3<sb/>O radical is formed instead of CH<sb_>2<sb/>OH, the upper bound for DET would be 10 kcal/mol higher. We assume that the observation of [(CH<sb_>3<sb/>OH)<sb_>n<sb/>]<sp_>+<sp/> ions in the present experiment arises because there is an energetic barrier to breakup of this ion to <O_>formula<O/>. The shaded region in the diagrams place some reasonable bounds on the height of such a barrier (0-10 kcal/mol) and thus loosely brackets the threshold for DET.
<O_>figure&caption<O/>
A. The 1:2 cluster
As Fig. 4 indicates, the only observed product channel in one-color R2PI of the 1:2 cluster is fragmentation via loss of a single CH<sb_>3<sb/>OH to form <O_>formula<O/>. Fragmentation is 86% <*_>unch<*/> 5% efficient using the 6<sp_>1<sp/> level of the S<sb_>1<sb/>state as the intermediate state and 80% <*_>unch<*/> 10% efficient at 0<sp_>0<sp/>, despite our use of laser powers for which three-photon contributions to the ion signals are negligible. The small change in the percentage of fragmentation accompanying the 3 kcal/mol increase in the two-photon energy confirms the notion that most of the cluster ions have an internal energy near the two-photon energy in Fig. 4, well in excess of the <O_>formula<O/> dissociation asymptote in both the O<sp_>0<sp/> and 6<sp_>1<sp/> scans. Loss of CH<sb_>3<sb/>OH from the ionic cluster occurs on a time scale fast compared to movement of the cluster ion in the ion source (t<1<*_>mu<*/>s), since no asymmetry is observed in the arrival time profile of the parent or fragment. Fragmentation occurs almost exclusively by loss of a methanol monomer rather than a methanol dimer, despite the fact that the latter channel should be accessible to many of the clusters. This favoring of evaporative loss of monomer units has been a trademark of cluster fragmentation in many types of clusters.
<O_>figures&captions<O/>
The lack of ion-molecule chemistry in the <O_>formula<O/> cluster appears to be a direct consequence of energetic constraints. Dissociative electron transfer, which is the dominating reaction channel in higher clusters, is predicted to be endothermic in R2PI through 6<sp_>1<sp/> (Fig. 6). This is the case because the ionization potential of CH<sb_>3<sb/>OH is some 37 kcal/mol above that for C<sb_>6<sb/>H<sb_>6<sb/> (Table I), so that even with the stabilization of CH<sb_>3<sb/>OH<sp_>+<sp/> provided by a second methanol molecule, formation of [(CH<sb_>3<sb/>OH)<sb_>2<sb/>]<sp_>+<sp/> is still endothermic. Similarly, the high proton affinity of C<sb_>6<sb/>H<sb_>5<sb/> (which clearly precludes proton transfer of C<sb_>6<sb/>H<sb_>6<sb/><sp_>+<sp/> to CH<sb_>3<sb/>OH) is nearly thermoneutral for transfer to (CH<sb_>3<sb/>OH)<sb_>2<sb/> (see Table I and Fig. 6). Hence it is not surprising that no reaction products beside fragmentations are observed in the 1:2 clusters.
B. The 1:3 cluster
The <O_>formula<O/> cluster ion is the smallest sized 1:n cluster to undergo intrascluster ion-molecule chemistry. The appearance of [(CH<sb_>3<sb/>OH)<sb_>3<sb/>)]<sp_>+<sp/> (i.e., mass 96) is consistent with the energy level diagram of Fig. 7 which shows that the asymptote for DET is now well below the maximum ion internal energy produced in R2PI. Nevertheless, DET is still only a minor channel (6%) which competes only poorly with fragmentation via loss of CH<sb_>3<sb/>OH (94%).
1. The formation of [(CH<sb_>3<sb/>OH)<sb_>3<sb/>]<sp_>+<sp/>
The formation of a reaction product at mass 96 is notable for being exclusively the unprotonated methanol cluster ion. As mentioned earlier, when pure methanol clusters are either electron bombardment or photoionized, only protonated cluster ions are observed due to exothermic ion-molecule reactions of the type
<O_>formulae<O/>
The bimolecular analogs of these reactions
<O_>formula<O/>
are exothermic by 24 and 14 kcal/mol, respectively. Both these facts suggest that the initially unprotonated cluster ions would be inherently unstable with respect to loss of CH<sb_>2<sb/>OH or CH<sb_>3<sb/>O. However, in recent experiments by Vaidyanathan et al., electron bombardment ionization of Ar/CH<sb_>3<sb/>OH heteroclusters has successfully produced significant quantities of unprotonated <O_>formula<O/> ions via intracluster Penning ionization involving high-lying states of the argon neutrals. The present study offers a second example of the formation of unprotonated methanol cluster ions, this time mediated by C<sb_>6<sb/>H<sb_>6<sb/>. The attachment of the methanol clusters to a C<sb_>6<sb/>H<sb_>6<sb/> chromophore offers an extremely gentle means of producing unprotonated M<sb_>n<sb/><sp_>+<sp/> clusters by photoionizing the cluster with only 9.60 eV energy via the C<sb_>6<sb/>H<sb_>6<sb/> chromophore. In the case of the 1:3 cluster, the unprotonated M<sb_>3<sb/><sp_>+<sp/> products can be formed completely free from interference from protonated clusters. This mechanism thus provides a route producing the novel M<sb_>n<sb/><sp_>+<sp/> cluster ions for subsequent spectroscopic and mass spectrometric study.
The formation of the DET product M<sb_>3<sb/><sp_>+<sp/> (M = methanol) provides supporting evidence for the neutral C<sb_>6<sb/>H<sb_>6<sb/>-M<sb_>3<sb/> cluster possessing a structure composed of hydrogen-bonded methanols all attached to the same side of the benzene ring (Paper I). It seems unlikely that DET to form M<sb_>3<sb/><sp_>+<sp/> would occur from an initial geometry in which methanol molecules are on both sides of the benzene ring rather than as an aggregate on the same side.
The structure of the mass 96 ion is not determined in the present work. However, it seems likely that the M<sb_>3<sb/><sp_>+<sp/> cluster exists either as <O_>formula<O/> or as <O_>formula<O/> in which proton transfer within the methanol cluster has occurred with cluster energy insufficient to allow its breakup to form the protonated M<sb_>2<sb/>H<sp_>+<sp/> ion. Loss of the C<sb_>6<sb/>H<sb_>6<sb/> molecule form the cluster provides the means by which the M<sb_>3<sb/><sp_>+<sp/> ion is stabilized below its M<sb_>2<sb/>H + CH<sb_>2<sb/>OH/CH<sb_>3<sb/>O dissociation asymptote.
The only energetically allowed channel which competes successfully with DET is fragmentation to form <O_>formula<O/>, which dominates the product distribution (94%). No parent 1:3 cluster ions are observed in our experiment. Thus the observed processes are the following:
<O_>formulae<O/>
One could imagine two limiting cases for the competition between DET and fragmentation. In one limit, the energy dependence of the rate constants is either small enough or similar enough in the two channels that the observed product distribution directly reflects the relative magnitudes of the rate constants for fragmentation and DET; i.e., the product yields are kinetically controlled. In this case, the rate of fragmentation would provide and 'internal clock' for the DET reaction if it could be determined by other means.
In the second limit, the observed product distribution will reflect energetic constraints rather than kinetics. This would occur if k<sb_>frag<sb/>(E) is much greater then k<sb_>DET<sb/>(E) for energies where fragmentation can occur. Then all cluster ions with internal energies above the fragmentation threshold would undergo fragmentation, while the remaining, lower energy cluster ions would react via DET. More complete control over the ion internal energies produced or detected (e.g., via photoelectron-photoion coincidence measurements) will be required before the energy dependences of the reaction rates can be determined unambiguously.
2. Reactions which are not observed
The formation of unprotonated M<sb_>3<sb/><sp_>+<sp/> ions is unusual in a second respect - it occurs to the exclusion of several other energetically open channels, most notably those involving dissociative proton transfer [Eq. (4)]. Careful searches for the DPT products <O_>formula<O/> and <O_>formula<O/>, or the CH<sb_>2<sb/>OH loss channel to produce <O_>formula<O/> place upper bounds on these channels at less than 1% of the 1:2 fragment channel (Fig. 4), nearly ten times less than the observed yield of DET products.
Again, energetics could play a significant role in suppressing the proton transfer channels if the rate constants for the product channels were sensitively dependent on their exothermicity.
<#FROWN:J07\>
Let's briefly examine what this might do. If the acceleration is a, the change in the potential energy is <O_>formula<O/> and the total kinetic energy gained is <O_>formula<O/>. Here the rate of growth is<*_>tau<*/><sp_>-1<sp/> and <*_>delta<*/>z is the displacement. Then the growth time scale is given by
<O_>formula<O/>
where k<*_>approximate-sign<*/>z<sp_>-1<sp/> is the wave number for the disturbance (assuming that the wavelength is of order z). The instability is guaranteed if the density gradient is in the opposite direction to the local acceleration. Now by noting that any acceleration is locally equivalent to gravitation (this is after all the basis of general relativity), we can replace a by <*_>DELTA<*/>P, the pressure jump across the shock front at the wind-wind interface. We could also use g<sb_>eff<sb/>, the effective gravity including radiation pressure. Let's concentrate on the first choice. Across a shock, which moves with a speed v<sb_><*_>SIGMA<*/><sb/> into the surrounding gas, the pressure is given by
<O_>formula<O/>
For a perfect gas this is larger on the postshocked side and the pressure jump is <O_>formula<O/>. The interface is unstable to the formation of knots and blobs. As we have already discussed, as these rise, they tend to show induced vorticity and develop the characteristically mushroom-shaped structures so familarfamiliar from both violent explosions (like nuclear blasts) and simulations of supernova envelopes. This same condition will certainly prevail during the formation of planetary nebulae, and the wind-wind collision should generate stringy and bloblike structures because of this instability. If optically thick, they will shadow the material in the slow wind, and the result is that low ionization regions should be mixed into the H II region. One expectation is that in any of the blastlike expansions, be they Wolf-Rayet winds, planetary nebulae, or supernova remnants, it should be possible to see the effects of the wind collision preserved in these structures.
8.7 Accretion Disks in Astrophysics

8.7.1 Some Observational Motivations
We have invoked binary systems frequently in this chapter, and it is therefore appropriate to end it with a discussion of a most interesting confluence of rotation and viscosity, namely accretion disks around massive objects. But first, before launching into an extended expos of the properties of the disks themselves, let's examine the conditions under which rotating accretion flows may arise.
For many binaries, especially ones of long period, this is precisely what is observed. The more evolved star really is the more massive. In the case of Algol (<*_>beta<*/> Persei), an extremely well-studied star and the first discovered eclipsing binary, just the opposite is observed. Algol consists of a G giant and a B main sequence star. The mass ratio is <*_>approximate-sign<*/>3, but in favor of the main sequence B star. The orbital period is short, less than 3 days. the light curve data and modeling the equilibrium shapes of the stars show that the red subgiant completely fills its Roche surface. This is the limiting surface for tidal interaction, given approximately by
<O_>formula<O/>
where a is the semimajor axis given by Kepler's law <O_>formula<O/> for circular orbits, q is the mass ratio, and M is the total mass of the system. The reason for the peculiar mass of the red giant is that it has been significantly altered by mass loss from the binary system and by mass transfer onto the main sequence star. The fact that this is still going on, in both this and related semidetached systems (the term comes from the fact that only one of the stars is in contact with the critical surface), means that accretion flows onto the companion have played a role in the orbital dynamics and that this has fed back into the stellar evolution through the alteration of the mass and boundary conditions on the stars. The observation, for a number of these stars, of emission lines which are formed in a Keplerian disk surrounding one of the components adds fuel to the argument, although in Algol it does not appear that an extensive accretion disk is observed.
8.7.2 Flow through the Inner Lagrangian Point
First, a binary system, that is, a close system, is one that is not spherically symmetric. The presence of the companion star, as well as the rotation of the mass-losing star due to spin-orbit coupling, produces immediate departures from sphericity. To see what happens to the mass transfer at the inner Lagrangian point, also called L<sb_>1<sb/>, we need to consider the flow of material in a potential that switches sign at some point in the flow (see Fig. 8.4).
Let's look back at the spherical case for a moment. The mass loss is driven by the combined effect of pressure gradient and retardation due to gravity. At some point, where the outward driving becomes strong enough relative to gravity, the material coasts at the sound speed and then accelerates as the gravitational acceleration continues to fall off. In other words, the reason the effective gravity vanishes is that at some point, g is balanced by <*_>unch<*/>p. But what about other possible cases? We've already seen that if g is balanced by <*_>unch<*/>p<sb_>rad<sb/>, or at some point g(1 - <*_>GAMMA<*/>) vanishes, then the material becomes supersonic and a strong wind. Both of these depend on the presence of a pressure gradient to do the job of producing a sign change in the acceleration. The alternative is to say that the gravity itself reverses sign, something impossible for a single star but normal for a close binary. Put differently, imagine that the potential is taken to be
<O_>formula<O/>
in the vicinity of the L<sb_>1<sb/> point. The problem can be rendered one dimensional if we assume that we look at the flow only in the vicinity of the Lagrangian point and that the system is not so rapidly rotating that the Coriolis deviation of the flow is large compared with rectilinear flow. This means that <*_>OMEGA<*/><u', where the prime denotes the spatial derivative of the velocity.
<O_>figure&caption<O/>
The fact that the equation of motion can be written as
<O_>formula<O/>
means that we can apply the same condition that we had in the spherical Parker wind solution. Take a look at the flow through a small region around the L<sb_>1<sb/> point. The equipotentials on either side of L<sb_>1<sb/>, along the line of centers, give a local critical point to the flow. This is because there is a local maximum in the gravitational field. Perpendicular to the line of centers the gradient has a local minimum. (See Fig. 8.5.) The L<sb_>1<sb/> point is a saddle point in the gravitational potential; the gravitational acceleration changes sing on crossing this point. For the case of one-dimensional flow, this has the same effect as the changing gravitational acceleration relative to the pressure gradient. The critical condition for stream formation is similar to the Parker solution; that is, u = a<sb_>s<sb/> at L<sb_>1<sb/>. Thereafter, as the gravitational field increases toward the secondary, the flow accelerates. As material is forced through this point, it has the same effect as the passage through the g<sb_>eff<sb/> = 0 point in a spherical wind. The pressure gradient does not vanish, so the material is accelerated and the sound speed is reached, after which the flow is ballistic toward the secondary. Stream formation is important because it transfers material with high specific angular momentum toward the accreting star. Once in the vicinity of the companion, the matter forms an accretion disk, the details of which we shall now discuss.
<O_>figure&caption<O/>
8.7.3 Some Consequences of Mass Transfer
The observation of tidal distortion leads immediately to important hydrodynamic consequences. The existence of the Roche surface in general, and the limiting radius in particular, is the result of the three-body problem. The inner Lagrangian point, L<sb_>1<sb/>, is the point along the line of centers at which the effective gravitational acceleration vanishes. However, since the pressure does not vanish even though the gravitational acceleration does, matter will be forced to exit through this region and begin to flow to the other star. In effect, we have set up the de Laval nozzle problem from Chapter 1. The variation in the effective gravity acts like the nozzle (although without many material walls) to accelerate the flow through the sonic point and eventually to hypersonic speeds. The matter carries some net angular momentum because the L<sb_>1<sb/> point is generally not at the center of mass, and therefore the deviation of the flow and its acceleration toward the companion produce an accretion disk.
The stream must eventually rid itself of this excess angular momentum before it can accrete onto the companion even for direct impact. Several mechanisms are available, probably all of which operate somewhere in the universe. One is turbulent viscosity. That is the one we shall mainly deal with here. Another is magnetic breaking. If the material forms a disk that becomes Kelvin-Helmholtz unstable at the boundary of a stellar magnetosphere, blobs may be formed that accrete onto the companion. The process is certainly not well understood but can be simulated for neutron star accretion and is well established as a scenario. For direct impact, the stream may submerge and pump angular momentum into a deeply generated boundary layer.
The final mechanism is spiral shocks. Since the matter falls into the disk with excess angular momentum and drives spiral waves in the disk, these may form stable circulating structures that serve to deviate the flow onto the companion and dissipate energy and momentum. (See Fig. 8.6.) Presently, however, the details of the accretion process are the most schematic parts of accretion disk theory. This is a pity, because these details contain virtually all of the essential physics.
If the mass-accreting star is a compact object, like a white dwarf, neutron star, or black hole, the disks reach temperatures considerably higher than for main sequence stars. The simple reason is that the gravitational well around the accreting object is very deep and the energy source for the dynamics is consequently greater. Any dissipative processes fed by the global circulation will therefore find a very large reservoir of energy which can be tapped, and the resulting emission of radiation can take place in the ultraviolet or even the x-ray. It was the latter wavelength region, observed with satellites like Uhuru and Einstein, that first signaled the presence of accretion disks around neutron stars like Hercules X-1 = HZ Her and black holes like Cygnus X-1 = HD 226868. The added discovery that the optical counterparts of these and other galactic x-ray sources are spectroscopic binaries and the observation of optical emission lines which tracked the compact star clinch the argument for accretion powering the radiation.
<O_>figure&caption<O/>
The observation of strong emission lines arising in a very compact region in the cores of active galaxies like quasars and Seyferts also indicates that accretion can occur on a scale of extremely massive but otherwise 'single' collapsed objects. These galaxies have emission line widths indicative of velocities of order 1000 to 10<sp_>4<sp/> km s<sp_>-1<sp/> coming from a region less than 1 pc across. The temperature of the regions, indicated by the presence of extremely high radiation rates for x-rays, argues that accretion flow around a black hole is the likely source of the observed luminosity.
In light of these observations, and because the range of physics required for an understanding of such flows touches on virtually all aspects of astrophysical hydrodynamics, we will discuss accretion disk theory at some length.
8.7.4 Heating the Disks: Dissipation and Viscous Torques
We first take up the question of the effect of the generation of energy by the shearing within the disk due to its differential rotation. Recall that the viscous energy dissipation rate is given by
<O_>formula<O/>
where T<sb_>ij<sb/> is the stress tensor and <*_>sigma<*/><sb_>ij<sb/> is the shear. The shear for an axi-symmetric system is given by <O_>formula<O/>. The negative sign in the second term comes from <O_>formula<O/>. In light of the previous discussion, the disk is Keplerian, and since the angular frequency is given by <O_>formula<O/>, the shear is <O_>formula<O/>. Since the shear varies with radius for such a disk, so does the rate of energy generation.
<#FROWN:J08\>
Global Warming and Potential Changes in Host-Parasite and Disease-Vector Relationships
ANDREW DOBSON AND ROBIN CARPER
I. INTRODUCTION
Parasitology has always been a discipline in which purely academic studies of the evolution of parasites and their life cycles have progressed as a necessary complement to the study of the pathology and control of the major tropical diseases of humans and their livestock. Indeed, the most striking feature of parasitology is the diversity of parasites in the warm tropical regions of the world and the frightening levels of debilitation and misery they cause. Determining how long-term climatic changes will affect the distributions of different parasites and pathogens at first seems a daunting task that almost defies quantification. Nevertheless, as parasitologists have always been concerned with the influence of climatological effects on different parasite species, it is possible to begin to speculate on the ways that global warming might affect the distributions of some specific tropical diseases. Similarly, the study of parasite population dynamics has developed within a solid theoretical framework (Anderson and May 1979, May and Anderson 1979). This permits the development of quantitative speculation in more general studies concerned with how parasite-host interactions may respond to perturbation.
This chapter addresses both general questions about the response of parasite-host systems to long-term climatic changes and the specific response of one particular pathogen, Trypanosoma, to the changes in climate predicted for the next hundred years.
A. Macroparasites and Microparasites
Current estimates suggest that parasitism of one form or another may be the most common life-history strategy in at least three of the five major phylogenetic kingdoms (May 1988, Toft 1986). The enormous array of pathogens that infect humans and other animals may be conveniently divided on epidemiological grounds into microparasites and macroparasites (Anderson and May 1979, May and Anderson 1979). The former include the viruses, bacteria, and fungi and are characterized by their ability to reproduce directly within individual hosts, their small size and relatively short duration of infection, and the production of an immune response in infected and recovered individuals. Mathematical models examining the dynamics of microparasites divide the host population into susceptible, infected, and recovered classes. In contrast, the macroparasites (the parasitic helminths and arthropods) do not multiply directly within an infected individual but instead produce infective stages that usually pass out of the host before transmission to another host. Macroparasites tend to produce a limited immune response in infected hosts; they are relatively long-lived and usually visible to the naked eye. Mathematical models of the population dynamics of macroparasites have to consider the statistical distribution of parasites within the host population.
B. Direct and Indirect Life Cycles
A second division of parasite life histories distinguishes between those species with monoxenic life cycles and those with heteroxenic life cycles. The former produce infective stages that can directly infect another susceptible definitive host individual. Heteroxenic species utilize a number of intermediate hosts or vectors in their transmission between definitive hosts. The evolution of complex heteroxenic life cycles permits parasite species to colonize hosts from a wide range of ephemeral and permanent environments, while also permitting them to exploit host populations at lower population densities than would be possible with simple direct transmission (Anderson 1988, Dobson 1988, Mackiewicz 1988, Shoop 1988). However, heteroxenic life cycles essentially confine the parasite to areas where the distribution of all the hosts in the life cycle overlap. Shifts in the distribution of these host species due to climatic changes, will therefore be important in determining the areas where parasites may persist and areas where parasites may be able to colonize new hosts.
C. Aquatic and Terrestrial Hosts
Climatic changes are likely to have different effects on aquatic and terrestrial environments (chapter 24). The heteroxenic life cycles of some parasite species often allow them to utilize hosts sequentially from either type of habitat. It is thus important to determine the different responses of the terrestrial and aquatic stages of a parasite's life cycle to climatic change. That, along with an examination of other parasite responses to climatic change, demands a quantitative framework within which to discuss parasite life-history strategies.
II. PARASITE LIFE-HISTORY STRATEGIES
The complexities of parasite host population dynamics may be reduced by the derivation of expressions that describe the most important epidemiological features of a parasite's life cycle (Anderson and May 1979, May and Anderson 1979, Dobson 1988). Three parameters are important in describing the dynamics of a pathogen: the rate it will spread in a population, the threshold number of hosts required for the parasite to establish, and the mean levels of infection for the parasite in the host population.
Basic reproductive rate of a parasite, Ro: The basic reproductive rate, Ro, of a microparasite may be formally defined as the number of new infections that a solitary infected individual is able to produce in a population of susceptible hosts (Anderson and May 1979). In contrast, Ro for a macroparasite is defined as the number of daughters that are established in a host population following the introduction of a solitary fertilized female worm. In both cases the resultant expression for Ro usually consists of a term for the rates of parasite transmission divided by an expression for the rate of mortality of the parasite in each stage in the life cycle (Dobson 1989). Increases in host population size or rates of transmission tend to increase Ro, and increases in parasite virulence or other sources of parasite mortality tend to reduce the spread of the pathogen through the population.
Threshold for establishment, H<sb_>T<sb/>: The threshold for establishment of a parasite, H<sb_>T<sb/>, is the minimum number of hosts required to sustain an infection of the pathogen. An expression for H<sb_>T<sb/>, may be obtained by rearranging the expression for Ro to find the population density at which Ro equals unity. This may be done for both micro- and macroparasites with either simple or complex life cycles. The resultant expressions suggest that changes in the parameters that tend to increase Ro tend to reduce H<sb_>T<sb/>, and vice versa. Although many virulent species require large populations to sustain themselvlesthemselves, reductions in the mortality rate of transmission stages may allow parasites to compensate for increased virulence and maintain infections in populations previously too small to sustain them.
Mean prevalence and burden at equilibrium: It is also possible to derive expressions for the levels of prevalence (proportion of the hosts infected) and incidence (mean parasite burden) of parasites in the host populations. In general, parameters that tend to increase Ro also tend to give increases in the proportion of hosts infected by a microparasite and increases in the mean levels of abundance of any particular macroparasite (Anderson and May 1979, May and Anderson 1979, Dobson 1988). Most important, increases in the size of the host population usually lead to increases in the prevalence and incidence of the parasite population (fig. 16.1).
These expressions, which characterize the most important features of a parasite's interaction with its host at the population level, can be used to ascertain how parasites with different life cycles will respond to long-term climatic changes. This may best be undertaken by determining which stages of the life cycles are most susceptible to climatic variation and by quantifying the response of those stages to climatic change.
<O_>figure&caption<O/>
III. EFFECT OF TEMPERATURE ON PARASITE TRANSMISSION RATES
The physiology of adult parasites is intimately linked with the physiology of their hosts. Providing the hosts can withstand environmental changes, it seems unlikely that the within-host component of the parasite life cycle will be significantly affected. However, any form of increased stress on the host may lead to increase in rates of parasite-induced host mortality (Esch et al. 1975). In the absence of data from the specific experimental studies that could throw considerable light on these relationships, this study will concentrate on the effect of changes in meteorological factors on the free-living infective stages of different groups of parasites.
A. Parasites with Aquatic Transmission Stages
Several detailed laboratory studies have examined the effect of temperature on the transmission success of parasites with aquatic infective stages. The parasitic trematodes are probably the most important class of parasites to utilize an aquatic stage for at least part of their life cycle. The data presented in figure 16.2 are for an echinostome species that is a parasite of ducks. Increased temperature leads to increased mortality of the larval infective stages of the parasite. It also leads to increased infectivity of the larval stage. The interaction between larval infectivity and survival means that net transmission efficiency peaks at some intermediate temperature but remains relatively efficient over a broad range of values (16<*_>degree<*/>-36<*_>degree<*/>C for Echinostoma liei cercariae; fig. 16.2). These synergistic interactions between the different physiological processes determining survival and infectivity allow the aquatic parasites to infect hosts at a relatively constant rate over the entire spectrum of water temperatures that they are likely to experience in their natural habitats (Evans 1985).
B. Poikilothermic Hosts
The effect of temperature on the developmental rate of parasites in both aquatic and terrestrial hosts has been examined for several of the major parasites of humans in the tropics. In contrast to the effect on transmission efficiency, increases in temperature usually lead to reduced development times for parasites that utilize poikilothermic hosts (fig. 16.3). As with many physiological processes, a 10<*_>degree<*/> increase in temperature seems to lead to a halving of the developmental time. This may allow parasite populations to build up rapidly following increases in temperature.
C. Parasite Populations in Thermal Cooling Streams
The expressions for Ro and H<sb_>T<sb/>, derived in the first part of this chapter, suggest that increases in transmission efficiency and reductions in development time induced by temperature changes allow parasites to establish in smaller populations and grow at more rapid rates. This is observed to some extent in a pair of long-term studies that compare the parasite burdens of mosquito fish (Gambuis affinis) populations in artificially heated and control sections of the Savannah River in South Carolina. The data for the trematode Ornithodiplostomum ptychocheilus show significant differences between heated and ambient sites during the earlier period of the study when temperature differences were most pronounced. Infection by the parasites starts several months earlier each year in the thermally altered sites (fig. 16.4). However, infection rates decline in the summer in the artificially heated sites when populations of hosts decline in response to high water temperatures (Camp et al. 1982). This effect may be compounded by the movement of the waterfowl that act as definitive hosts for the parasite. These birds tend to prefer the warmer water in winter and cooler water in the summer. Similar but less clearly defined patterns are observed in the data for Diplostomum scheuringi from the same site (Aho et al. 1982).
These studies illustrate the important role of host population density in the response of a parasite's transmission rate to thermal stress, while also demonstrating the ability of parasites to capitalize on improved opportunities for transmission and to establish whenever opportunities arise. Obviously the data are open to several interpretations, but they do emphasize the importance of long-term experiments in determining the possible effects of global warming on the distribution of parasites.
D. Terrestrial Hosts
The survival rates of the infective stages of the parasites of most terrestrial species tend to decrease with increasing temperature (fig. 16.5a). Although little evidence is available to determine how the infectivity of these larvae is affected by temperature, rates of larval development tend to increase with increasing temperatures (fig.16.5b). These two processes again interact synergistically - as an increase in temperature depresses survival, development speeds up - allowing the parasite to establish at a broad range of environmental temperatures. In contrast to parasites that utilize aquatic hosts, parasites of terrestrial hosts have transmission stages that are susceptible to reduced humidity, and these stages are highly susceptible to desiccation (Wallace 1961). To compensate for reduced opportunities for transmission during periods of severely adverse climate, parasites of terrestrial hosts have evolved adaptations such as hypobiosis, the ability to remain in a state of arrested development within the relatively protected environment provided by their hosts until such time as transmission through the external environment proves more effective.
<#FROWN:J09\>
EXTRACELLULAR SIGNAL-REGULATED KINASES IN T CELLS
Anti-CD3 and 4<*_>beta<*/>-Phorbol 12-Myristate 13-Acetate-Induced Phosphorylation and Activation
CHARLES E. WHITEHURST, TERI G. BOULTON, MELANIE H. COBB, AND THOMAS D. GEPPERT
Extracellular signal-regulated kinases (ERK) 1 and 2 are growth factor- and cytokine-sensitive serine/threonine kinases that are known to phosphorylate microtubule-associated protein 2 and myelin basic protein. The current studies examined whether ERK1 and/or ERK2 was present in T cells and whether they were phoshorylated and activated as a consequence of T cell activation. The data demonstrated that both ERK1 and ERK2 were present in Jurkat cells and peripheral blood T cells. In T cells, ERK2 was more prevalent than ERK1. The concentrations of ERK1 and ERK2 were not altered by stimulating the cells for 16 h with immobilized anti-CD3 mAb or anti-CD3 mAb and phorbol myristate acetate. mAb to CD3 and phorbol myristate acetate stimulated an increase in ERK1 and ERK2 MBP kinase activity. Anti-CD3 mAb triggered an increase their phosphate content which was detectable at 2 min but reached a maximum at 5 min. A portion of the increase in phosphate was caused by an increase in phosphotyrosine. We also examined the rate of ERK2 degradation. ERK2 was stable for up to 36 h, and its degradation was unaffected by the activation state of the cells. The data demonstrate that ERK1 and ERK2 are part of an anti-CD3 mAb-stimulated signal transduction cascade that is downstream of protein kinase C and, therefore, suggest that these kinases play an important role in T cell activation.
T cell activation is triggered by an interaction between the TCR and a complex formed by the antigenic peptide and a class I or II MHC molecule. The recognition of Ag by the TCR leads to a variety of early biochemical changes (1-4). Among the earliest occurring within seconds of engagement of the TCR, is the phosphorylation of a variety of substrates on tyrosine residues (5). The kinase or kinases responsible for these first phosphorylation events have not been identified, although a member of the src family of tyrosine kinase, fyn, appears to be physically associated with the CD3 complex (6). The subsequent steps in the cascade of reactions have not been well characterized. Temporally, activation of tyrosine phosphorylation precedes activation of phospholipase C (PLC) (5). Further, inhibitors of tyrosine kinases block the activation of PLC, demonstrating that tyrosine phosphorylation is necessary even for the stimulation of PLC (5, 7). As in other systems, PLC releases two products, inositol phosphates, that elevate intracellular calcium, and diacylglycerol, that activates and translocates protein kinase C (PKC) to the membrane, where it is in proximity to membrane substrates (2-4, 8, 9). In addition to these events triggered by engagement of the TCR complex, cross-linking CD4 or CD8 molecules on T cells leads to the activation of a tyrosine kinase, p56<sp_>1ck<sp/>, or the src family (10, 11). Based on burgeoning evidence from studies of T cell activation and by analogy to other receptor signaling systems, it is believed that these early biochemical phenomena result in the activation of additional cascades of protein kinases which are responsible for the subsequent cellular events. In other systems, the vast majority of these subsequent phosphorylations occur on serine and threonine residues.
Of the data suggesting that other kinases are stimulated in T cells, the most convincing is for a MAP2 kinase activity that is increased by cross-linking the TCR complex (12-14) or by the phorbol ester, PMA (13). The stimulation of this serine/threonine kinase activity by a mAb to the TCR complex is diminished but not eliminated by an inhibitor of PKC (H7) or by prior depletion of PKC (13). Cross-linking CD4 on T cells stimulates the tyrosine phosphorylation of a 43-kDa protein that co-migrates with a protein with MAP2 kinase-like activity (12, 14). Cross-linking CD3 together with CD4 results in a greater increase in MAP2 kinase activity than cross-linking CD3 alone (14). This combination of stimuli also delivers a more effective activation signal than cross-linking the TCR complex alone (15-17), indicating a correlation between MAP2 kinase activity and T cell activation. Taken together these findings suggest that MAP2 kinase plays an important role in T cell activation.
Recently, cDNA have been cloned and sequenced from human, rat, mouse, and Xenopus libraries which encode two proteins with insulin and nerve growth factor-sensitive MAP2/MBP kinase activity (18-23). These two proteins are 90% identical within their catalytic domains and have been named extracellular signal-regulated kinase (ERK) 1 and 2 because of the wide variety of extra-cellular signals which stimulate their activity (13, 14, 18, 20, 24-32). Moreover, they are highly conserved across species (19). The activation of these two serine/threonine kinases is correlated with increases in the phosphorylation of their tyrosine and threonine residues (20). Phosphatases specific for either tyrosine (CD45) or serine/threonine (2a) residues partially inactivate ERK1 and ERK2 (25), indicating that full activation of these kinases requires both tyrosine and serine/threonine phosphorylation.
The current studies were carried out, therefore, to determine whether ERK1 and/or ERK2 is present in T cells and to determine whether they are phosphorylated and/or activated as a consequence of activation. We found that both ERK1 and ERK2 are present in Jurkat cells and peripheral blood T cells and that ERK2 is more abundant. The concentration of both ERKs was not altered by stimulating the cells for 16 h with immobilized anti-CD3 mAb or anti-CD3 and PMA. Both anti-CD3 and PMA stimulated the phosphorylation of ERK1 and ERK2 in Jurkat cells and induced an increase in ERK1 and ERK2 MBP kinase activity. The data suggest that ERK1 and ERK2 may play an important role in T cell activation. Moreover, the size, kinetics of activation, and substrate specificity of ERK1 and ERK2 are similar to properties of kinases with anti-CD3-inducible MAP2 kinase-like activity noted in previous reports. Thus, the studies strongly support the hypothesis that ERK1 and ERK2 are the kinases responsible for this activity.
MATERIALS AND METHODS
Reagents, mAb, and mitogens. PMA (Sigma Chemical Co., St. Louis, MO) and phorbol 12,13-dibutyrate (PDB; Sigma) were dissolved in ethanol and added to the cultures at a final concentration of 10 ng/ml for PMA and 20 ng/ml for phorbol 12,13-dibutyrate, and 0.1% ethanol. The mAb used were OKT3 and 64.1 (American Type Culture Collection, Rockville, MD), IgG2a mAb directed at the CD3 molecular complex on mature T cells; TS1/18, an IgG1 mAb directed at the <*_>beta<*/>-chain of CD18; and P1.17, an IgG2a mAb or irrelevant specificity. All mAb were purified over a column of Sepharose 4B coupled with staphylococcal protein A and used at final concentrations as indicated in the figure legends. PHA (Wellcome Reagents Division, Burroughs Wellcome Co., Research Triangle Park, NC) was used at a final concentration of 0.5 <*_>mu<*/>g/ml. Orthophosphate (H<sb_>3<sb/>PO<sb_>4<sb/>; 285Ci/mg P) and <sp_>35<sp/>S-labeled amino acids (Tran<sp_>35<sp/>S-label; >1000 Ci/mmol) were purchased from ICN Biomedicals, Inc., Costa Mesa, CA.
Cell preparation. PBMC were obtained from healthy adult volunteers by centrifugation of heparin-treated venous blood over sodium diatrizoate/Ficoll gradients (Isolymph; Gallard Schlesinger Chemical Manufacturing Corp., Carle Place, NY). Cells were washed once in HBSS and twice in saline before further processing. T cells were prepared from PBMC by isolating the cells forming rosettes with neuraminidase-treated SRBC and passing them over a nylon wool column to deplete B cells and macrophages. Jurkat cells, a malignant T cell line, were generously provided by Dr. Arthur Weiss (33).
Antipeptide antisera. Antisera raised to peptides at the carboxyl terminus (837; IFQETARFQPGAPEAP) and subdomain XI (691; KRITVEEALAHPYLEQYYDPTDE) of rat ERK1 had characteristics described previously (25). Antiserum A249 was raised to a peptide at the carboxyl terminus of rat ERK2 (ELIFEETARFQPGYRS) using the same methodologies described previously (25). The sequences of human ERK1 and ERK2 are identical to the rat sequences in the region of the peptides used to derive 691 and A249 (19). The sequence of the carboxyl terminus of human ERK1 differs slightly from that of rat ERK1 (IFQETARFQPGVLEAP).
Preparation of cell lysates and immunoblotting. After T cells or Jurkat cells were treated with the various stimuli as indicated in the figure legends the cells were homogenized or sonicated in homogenization buffer (20 mM Tris, pH 7.5, 20 mM p-nitrophenyl phosphate, 1 mM EGTA, 50 mM NaF, 50 <*_>mu<*/>M sodium vanadate, 3 <*_>mu<*/>g/ml aprotinin, 1 mM leupeptin, 1 mM PMSF, and 10 mM iodoacetamide), the lysates clarified by centrifugation, and the soluble protein concentrations were determined using the bicinchoninic acid/Cu<sp_>2+<sp/> reagent (Micro BCA kit; Pierce Chemical Co., Rockford, IL). Equivalent amounts of protein/lane were analyzed by electrophoresis on 10 to 14% SDS-polyacrylamide gels, transferred to nitrocellulose or Immobilon-P (Millipore), and immunoblotted with a mAb to phospho-tyrosine (hybridoma 4G10; Upstate Biotechnology Ind., Lake Placid, NY) or antiserum 691 or A249. Reactive proteins were visualized with goat anti-rabbit or mouse IgG (Cappel) conjugated to horseradish peroxidase and 4-chloro-1-napthol, or with chemiluminescent procedure per manufacturers' specifications (Amersham Corp., Arlington Heights, IL). For chemiluminescent detection, the films were preflashed before exposure.
Radiolabeling of Jurkat cells and immunoprecipitations. Jurkat cells (40 x 10<sp_>6<sp/> cells/ml) were preincubated in phosphate-free RPMI 1640 containing 1% BSA for 50 to 60 min at 37<*_>degree<*/>C, [<sp_>32<sp/>P]orthophosphate was added (1.0 mCi/ml), and the cells were incubated an additional 50 to 60 min at 37<*_>degree<*/>C. After washing twice in ice-cold 0.9% NaCl, the cells were resuspended in RPMI 1640, stimulated as described in the figure legends, pelleted, and lysed. For immunoprecipitation with A249, the cells were lysed in homogenization buffer containing 1% NP-40 and 2 mM MgCl<sb_>2<sb/>, pH 8.0, for 10 min at 4<*_>degree<*/>C. Insoluble debris was removed by centrifugation. For immunoprecipitation with 837 the proteins were denatured either by boiling the lysates in SDS (0.2%) or by lysing the cells directly in 1% SDS. With the latter technique the lysate was later diluted so that the concentration of SDS was <0.2% for immunoprecipitation. For immunoprecipitations, the antisera were either added directly to the supernatants (10 <*_>mu<*/>l/sample), incubated for at least 1 h at 4<*_>degree<*/>C, and then precipitated with SPA-agarose (30 <*_>mu<*/>l of packed SPA-agarose/sample) for at least 1 hour at 4<*_>degree<*/>C, or alternatively, saturating amounts of the antisera were preabsorbed to SPA-agarose, and this complex was added to the lysates (30 <*_>mu<*/>l of packed preabsorbed SPA-agarose/sample and incubated in the same manner. The results of using these two methods were comparable. The immunoprecipitates were washed four times in homogenization buffer containing 1% NP-40, 0.1% SDS, and 150 mM NaCl, boiled in gel loading buffer containing <*_>beta<*/>-ME, and analyzed by SDS-PAGE and autoradiograhy. To correlate the molecular weights of the proteins immunoprecipitated by anti-serum 837 with those detected by antiserum 691 on an immunoblot, 837 immunoprecipitates or a whole cell lysate were electrophoresed on each side of a molecular weight standard on an SDS-PAGE gel. The gel was then sliced in half down the middle of the lane containing molecular weight standards, and the half of the gel containing the 837 immunoprecipitates was dried and used for autoradiography. The proteins on the other half of the gel were transferred to a polyvinylidene difluoride membrane for immunoblotting with antiserum 691. Exposures from the autoradiograph and immunoblot were then aligned using the molecular weight markers. Sequential immunoprecipitations were performed by immunoprecipitating with A249 as described earlier, boiling the immunoprecipitate in 1% SDS, and then diluting the supernatant to 0.2% SDS using 1% NP-40 homogenization buffer. 837 immunoprecipitates were then performed as described above.
Chromatography. Jurkat cells (200 x 10<sp_>6<sp/> cells/group) were stimulated for 5 min as indicated in the figure legends, pelleted, and then resuspended in 1 ml of homogenization buffer at 4<*_>degree<*/>C (20 mM Tris, pH 7.5, 1 mM EGTA, 50 mM NaF, 20 mM p-nitrophenyl phosphate, 1 mM sodium vanadate, 1 mM PMSF). The cells were sonicated with two 10-s bursts using a Vibracell (model ASI; Sonics and Materials Inc., Danbury, CT) and the cellular debris pelleted by centrifugation for 10 min at 15,000 x g. The supernatants were stored at -80<*_>degree<*/>C. For chromatography, the sonicates were thawed on ice, diluted with 3 vol. of water, and loaded onto a Mono Q HR5/5 column equilibrated with buffer A (50 mM <*_>beta<*/>glycerophosphate, pH 7.3, 1 mM EGTA, 0.1 mM sodium orthovanadate, and 0.1 <*_>mu<*/>M pepstatin).
<#FROWN:J10\>Such trade-offs would make genetic load and the cost of natural selection greater than they would be with functional independence.
Wallace (1987, 1989; Reeve et al. 1988) recently renewed his effort to lay the problem to rest with arguments and evidence that the traditional genetic-load argument, such as mine above, is based on faulty logic and a misunderstanding of the dynamics of viability selection during the culling of an age cohort. He emphasizes that much culling must take place in every population because of the universal Malthusian factor of over-production of offspring. He envisions a world in which an individual dying as a result of some genetic deficiency is thereby making room for a better endowed individual. If one does not die the other one would. Given that populations do remain finite, the argument fits the facts. Yet the conceptual problem remains, especially in low-fecundity species like our own, because the genetic-load arithmetic makes us expect far more culling than is actually found, and far more than the population could bear.
Wallace's (1987) illustrative model of cohort culling is the competition between seedlings in an experimental tray (e.g. Schmidt and Ehrhardt 1990). The space available will permit only a limited plant biomass to develop, and this will be produced by the small number of successful contenders. The great majority will do very little growing and gradually die out. The possibly small fraction that survives to maturity will be enormously variable in size and fitness. This result seems to be unaffected by levels of genetic load in the seeds used. If a thousand viable seeds are sown in one tray and a thousand with 90% lethal genotypes in another, the two trays may produce about the same number of ultimate survivors, total biomass, and phenotypic fitness variation.
I would suggest another kind of experiment as more relevant to the problem Haldane had in mind. Sow only 100 of the viable seeds in the first tray and 100 with a high incidence of genetic load in the other, and also in each tray sow 900 seeds of competing species. I would expect the 100 viables to win much more representation in their tray than the ten viables and 90 lethals in the other, and this result would be a closer parallel to what usually takes place in the culling of a plant cohort in nature. For most animals, Wallace's experimental model is even less realistic. Only sessile invertebrates meet intense and inescapable competition from near neighbors. Wallace's seedling experiment would be broadly applicable to animal populations only with competition for social status, West Eberhard's (1983) social selection, of which sexual selection would be a special case. Social status is a resource that can seldom be appropriated by a member of a different species.
The central problem with Wallace's model, which he calls soft selection, is that it implies unrealistically strong density dependence with an age cohort. Most populations in nature are extremely sparse. The tendency for field ecologists to study organisms that are abundant enough to study may greatly bias our impressions. Populations in nature are seldom dense enough to cause any obvious resource depression (Tilman 1982). The most convincing examples of resource depression result from exploitation by many species, such as the reduction of marine invertebrate biomass on mudflats from concerted onslaughts of many species of migratory bird (Schneider 1978). Even the extraordinarily dense populations commonly studied by field ecologists, e.g. by Andrewartha and Birch (1954), usually show numerical changes that look like random fluctuation and seldom give clear evidence of density effects in short-term studies. Individual survival must be mainly a matter of chance, partly a matter of many kinds of adaptive performance, and only to a minor degree affected by density-dependent competition with conspecifics.
Near neighbors, sessile or motile, will often be of different species in diverse natural communities, and the death of one individual will often allow the survival of a member of a competing species. In such situations we most clearly confront the challenge of genetic load in relation to population survival, the problem that worried Haldane. If too large a dose of its population's genetic load causes one individual to die, it is likely to mean that the abundance of that species will be reduced by one. A reduced genetic load would make it more likely for the population to survive in competition with other species. Every (1-s) that enters into a fitness calculation means a finite deficiency in some sort of adaptive performance. Any such deficiency implies not only adverse selection within a population, but also a decreased representation of the population in the community. Dudash's (1990) experiments nicely confirm this expectation. Fitness differences between her inbred and outbred seedlings were much greater in the field than in greenhouse monoculture. The expectation is that the natural populations that are still available for study should be those that have extremely low levels of genetic load, and this is not what is found.
Wallace (1989) claims that Haldane and others have been needlessly worried about a mere "computational artifact." They arbitrarily assign a fitness of 1 to a favored genotype and of 1-s to an unfavored competitor. If instead we used 1+s and 1 we would not calculate such low fitness values for so many multi-locus genotypes. This is true, but the change is merely cosmetic. We would still get the same variation in fitness and be faced with the same problem of how fit the average individual can possibly be. Also, the traditional notation is more realistic. A rare favorable mutation may be said to have a fitness of 1+s, but this implies a deficiency in the ancestral gene pool. If a mutation can improve some character by some fraction s, that character must have been suboptimal. How could the ancestral population have survived with suboptimal genotypes at a large number of loci if there were competing populations with a lower genetic load?
I think the time has come for renewed discussion and experimental attack on Haldane's dilemma.
10.2 Paradoxes of sexuality
Sexual reproduction by its existence and in many special aspects is a complex of puzzles on which many books have been written (e.g. Bradbury and Andersson 1987; Stearns 1987; Michod and Levin 1988). The main theoretical challenge is in the cost of meiosis, but this is a matter already getting attention from able investigators. I can do no better than refer readers to Maynard Smith (1984b), Eberhard (1985), Felsenstein (1985a), Bierzychudek (1989), Hamilton et al. (1990), Parts II and III of Stearns (1987) and Chapters 4-9 in Michod and Levin (1988). Another major challenge is in resolving the data of life-history diversity in the frequency and the developmental and ecological correlates of sexual phases. The problem here is not so much logical as logistic. The diversity is overwhelming in relation to the time and money that a few thousand interested biologists can devote to it.
Of the many recombination-related difficulties that I could discuss I will echo Maynard Smith (1988a) and choose the one that best serves as a kind of text-book illustration of an evolutionary anomaly, the absence of sexual reproduction throughout the rotifer order Bdelloidea. This is anomalous because it clearly violates the principle of Muller's ratchet, which seems a logically tight line of reasoning from well established premises (but see Gabriel (1989)). Muller (1964) was the first to recognize that an asexual lineage "incorporates a kind of ratchet mechanism, such that it can never get to contain, in any of its lines, a load of mutation smaller than that already existing in its at present least loaded lines." It can acquire a higher load of mutation simply by the occurrence of a new one in a least loaded line. So exclusively asexual reproduction leads inevitably to a degeneration of the genome, in the sense of its being ever more ruled by chemical stability, and ever less informative as to what has succeeded in the past. This must always lead to rapid extinction on an evolutionary time scale. For a recent quantitative study of Muller's ratchet, see G. Bell (1988).
Muller's ratchet explains the phylogenetic distribution of asexual species in most major types of eukaryotes. There is a fair number of exclusively clonal species, but never any entirely clonal genera or higher categories. Asexual species arise from time to time, but Muller's ratchet must lead them to extinction long before they can produce any appreciable taxonomic diversification. The loss of sexuality seems to be a classic example of an evolutionary step that is opposed by clade selection (Van Valen 1975). Unfortunately the general rule of conformity to expectations of Muller's ratchet has some exceptions. The whole rotifer order Bdelloidea (Meglitch (1967) calls them a class), with its several families and many genera and species, is composed entirely of parthenogenetic females (Pennak 1978). There are also a few other noteworthy violators of the theory, such as the freshwater gastrotrich order Chaetonotoidea (Meglitch 1967; Pennak 1978).
Another difficulty that surely deserves more attention is the scarcity of adaptively flexible sex determination (for a comprehensive review, see Bull, (1983)). Sex determination in most animals is genetic and is fixed at conception. Only a few have sex determination as a facultative response to information perceived during development. A neatly understandable example (Conover and Heins 1987) is provided by a fish, the Atlantic silverside, which spawns every spring in shallow waters along the Atlantic coast of the United States and Canada. Eggs spawned early in the season become females; those later on become males. Temperature provides the cue, so that development below a certain threshold causes female development, warmer water male development. This mechanism gives females a longer growing season and larger size for the following spring's spawning. The close relationship between size and fecundity in fishes makes large size more important to female fitness than to male.
This environmental sex determination is adaptive in a way that extend the size-advantage model used to explain the occurrence of protandry vs. protogyny among sequential hermaphrodites (Ghiselin 1969; Warner 1975). The silverside is almost entirely semelparous, and this would rule out sequential hermaphroditism as a viable life history. As predicted (Conover and Heins 1987), the temperature threshold that determines sex varies as expected of combined optimizing and frequency-dependent selection. It is lower in northern, higher in southern parts of the range, so that the sexes are nearly equally abundant all along the coast. Experiments by Conover and Van Voorhees (1990) show that the threshold can be changed by selection in the laboratory.
Besides the greater dependence of reproductive success on size in females than in males in most animals, it is possible to think of many other ways in which it may be more adaptive for an individual in a given situation to be male or female, and to identify cues that would predict the situation during development. A clear example would be the stochastically varying sex ratios of many social groups. In a pond in which most of the frogs happen to be of sex A, it would pay a tadpole to develop into a member of sex B. It would also pay a parent to bias sex determination away from whatever is the majority in previously produced young (Taylor and Sauer 1980). This would avoid what might be called Baptista's Burden. Having Bianca instead of a son after having had Katharina was something of a challenge to his fitness. Yet despite the advantage that can be envisioned in alternating sons and daughters, each sex determination is a largely independent event in most animal populations (Williams 1979; Huck et al. 1990).
Facultative sex determination would only be expected in groups in which useful cues can be perceived prior to any major developmental commitment to maleness or femaleness, requirements discussed in detail by Bull (1983), Charnov and Bull (1977), and Korpelainen (1990). These conditions must surely obtain in many diploid insect populations. Why is adaptive sex determination, such as that found in the silverside, not widespread in many groups of insects?
<#FROWN:J11\>
The argument for conversation is strongest when functional tests, such as gene replacement or in vitro complementation, can be applied. Most often, though, we must rely on sequence comparisons. But do these sequence comparisons monitor adaptive evolution, or do they monitor genetic drift? In the case of p34 there is extensive sequence identity throughout the molecule between yeast and humans, which diverged more than a billion years ago. In the case of the conserved regulators of p34, cyclin, cdc25, and weel, the identity is limited to a small portion of the molecules and there is extensive divergence in other domains. Yet in the case of cdc25, for example, despite the large sequence divergence, the human molecule will complement yeast mutants (Sadhu et al. 1990) and in the case of weel, frog can also complement yeast (Booher and Kirschner, unpublished).In an in vitro system, sea urchin cyclin, despite its large divergence in most of the molecule from the frog cyclin, will complement a deficiency of frog cyclin. These results suggest that functional divergence of these important regulating genes has been minimal, while sequence divergence has been extensive. Since cdc25, weel, and the p34 kinase fully complement deletions of these genes in species that diverged more than a billion years ago, we can conclude that no important yeast functions are missing in the human protein. The reciprocal experiment is not possible in humans but may soon be possible in mouse (Thomas and Capecchi 1990). Therefore, while these sequences have apparently drifted extensively, they do not appear to have evolved functionally very much.
With the help of genetic tests, the list of highly conserved cellular functions has continued to grow. In some cases phylogenetic barriers have emerged, but for many systems they are minor and easily overcome. For example, the <*_>beta<*/>-adrenergic receptor that normally responds to catecholamines in heart muscle will not function in yeast to replace a related receptor that responds to mating pheromones. However, addition of one more element to the signaling system, the mammalian G<*_>alpha<*/> protein, will allow the yeast cell to respond to catecholamines and undergo the mating response (King et al. 1991). The obvious conservation of DNA structure has been matched by the conservation of histones, transcription factors, splicing enzymes, ribonucleoprotein complexes, and nuclear pores. The well-known conservation of the protein synthesis machinery has been extended to protein secretion, including components of the endoplasmic reticulum and Golgi. The major cytoskeletal proteins such as tubulin and actin have been conserved and serve similar functions in many organisms, as do the metabolic enzymes. The signaling and regulating molecules such as ras and hormone receptors, the regulating kinases such as protein kinase A, protein kinase C, S6 kinase - all have been found in every eukariotyc cell. There have been, of course, new inventions, expansion and specialization of each repertoire, but, as we shall see, many of the new demands of specialized cell types have come about by usurpation of existing components. Viewed as a computer, we would have to say that the basic hardware is similar in all eukaryotic cells; if anything has changed, it is the software.
Software Changes in the Cell Cycle
Though the proteins that regulate the cell cycle may be nearly identical in all organisms, the strategy for regulating the cell cycle is not. In the frog egg the accumulation of cyclin to a threshold initiates mitosis, and this process is independent of any transcriptional control (Murray and Kirschner)1989b). In the Drosophila embryo after cellularization, cyclin accumulation is also required for mitosis, but it is not the regulator. In this case control of the mitotic process is under control of the mitotic activator cdc25, and its expression is under transcriptional control (Edgar and O'Farrell 1990). Recent studies of cyclins specific to the G1/S transition in yeast have shown that their accumulation is under transcriptional control, but also may be under posttranslational control (Wittenberg et al. 1990; I. Herskowitz, pers. comm.). In cleaving frog eggs and sea urchin eggs, cyclin accumulation is completely unregulated. However, in the case of the G1/S cyclin in yeast, the accumulation is tied to the whole pathway of the mating pheromone response as well as to other less well understood pathways involving cell size and nutrition.
In the few well-studied cases of cell cycle control we can see both conservation and divergence. The major components are highly conserved and most are functionally interchangeable. The basic reaction pathway involving cyclins, p34 kinase, and other kinases and phosphatases is also identical. Yet the rate limiting steps and their linkage to other processes are different in different cells. This has enabled the cell cycle control mechanisms to respond transcriptionally to spatial signals, to be linked to extracellular cues, to be coupled to various homeostatic mechanisms, or to operate nearly autonomously during the rapid cleavages in the early embryo. As with modern computers, the architecture of the machine allows for many software applications. One might even say that computer hardware has evolved to allow for greater software flexibility. As we shall see, much the same can be said for cellular mechanisms.
Divergent Pathways of Photoreception
The cephalopod eye and the vertebrate eye are exquisite examples of convergent evolution. The anatomy suggests that the origins are totally independent. The vertebrate eye develops as an outgrowth of the brain; the cephalopod and insect eye develops as a peripheral ectodermal structure that grows into the brain (Young 1974). The topology of the nerves and photoreceptors is reversed. In the vertebrate eye, light passes through the nerves to the photoreceptor; in the cephalopod eye, light impinges directly on the photoreceptors. Is this anatomical convergence reflected in a totally separate origin of the biochemistry of photoreception?
The key event in photoreception, the photoisomerization of retinal-dehyde, has been widely used. In prokaryotes, where it is part of the proton pump, and in eukaryotes, where it is used as a photoreceptor, retinaldehyde, which is chemically the same in all systems, is bound to an integral membrane protein called opsin, whose polypeptide chain spans the plasma membrane seven times. There is no sequence homology between the prokaryotic opsins and the eukaryotic opsins, though overall structural similarities in the positions of the amino and carboxyl ends and the number of transmembrance helices suggest that at one time these proteins could have had a common origin (Henderson and Schertler 1990).
In eukaryotes, whether cephalopods or mammals, opsin is a 7-membrane spanning protein, and all such proteins are receptors that are thought to couple to intracellular GTP binding proteins called G-proteins. This widespread family of membrane protein receptors includes the receptors for the mating pheromones in yeast, the cAMP receptor in slime molds, and the serotonin and <*_>beta<*/>-adrenergic receptor in mammals (King et al. 1991). The receptors catalyze the exchange of GTP for GDP on the heterotrimeric G protein. Binding of GTP causes dissociation of the trimeric G protein into G<*_>alpha<*/> and G<*_>beta<*/>G<*_>gamma<*/>; these subunits interact with other cellular enzymes and regulate their functions. The invertebrate opsins, which are 7-membrane spanning integral membrane proteins, have clear sequence similarity to the vertebrate opsins (Yokoyama and Yokoyama 1989). In the central region of the molecule there is also a very strong similarity on the nucleic acid level, and throughout the molecule there is extensive similarity with a few insertions or deletions. There is no question that rhodopsin, the primary unit of photoreception, has evolved from a common precursor.
The vertebrate opsins are known to couple to a heterotrimeric G protein called transducin, which in its GTP form activates directly a cGMP phosphodiesterase. In the vertebrate photoreceptor, increased levels of cGMP open a Na<sb_>+<sb/> channel leading to increased neurotransmitter release. Therefore, the action of light causes a drop in cGMP and an inhibition of transmitter release that inactivates an inhibitory neuron, which ultimately leads to elevated electrical activity in the brain (Stryer 1988). The vertebrate photoreception system also has a means of adaptation that desensitizes the receptor after stimulation. It involves the binding of a small protein, called <*_>beta<*/>-arrestin, to the cytoplasmic domain of the receptor after a period of activation (Bennett and Sitaramayya 1988).
In invertebrates, although the initial coupling of opsin to signal transmission are similar, the complete pathway is designed differently. Drosophila is known to contain G-proteins (Guillen et al. 1990), and the structure of invertebrate opsin strongly suggests that the receptor couples to G-proteins; the exact G-protein that couples to Drosophila rhodopsin is not known. Like vertebrates, Drosophila contains a <*_>beta<*/>-arrestin molecule that is highly conserved, suggesting that Drosophila rhodopsin contains the same desensitization system as mammals (Smith et al. 1990). However, the next part of the pathway seems divergent. G proteins are known to couple to several second messenger systems, and the best evidence suggests that G protein in invertebrates (Drosophila and the horseshoe crab, Limulus) couples to a different second messenger system from that affecting cGMP phosphodiesterase (Suss et al. 1989; J.Brown, pers. comm.). Genetic approaches can be useful in delineating this second messenger pathway. Recently, Drosophila mutants have been obtained that have morphologically normal cells that do not respond to light. The gene that is defective in one of these mutants has been cloned and shown to have strong similarity phospholipase C, an enzyme involved in cell signaling (Bloomquist et al. 1988). There is evidence that Ca<sp_>++<sp/> release, mediated by inositol triphosphate, occurs during light stimulation, which suggests that in the invertebrate photoreceptors the G protein linked to opsins may activate phospholipase C and signal either Ca <sp_>++<sp/> pathways via inositol triphosphate or protein kinase C via diacylglycerol. It is also possible that the unknown G protein signals some other second messenger pathway. Downstream of this signaling system there is an increase (as opposed to the decrease in vertebrate photoreceptors) in a nonselective cation channel leading to a depolarization and secretion. Thus the invertebrate system uses the same visual pigment, an evolutionarily related receptor, a very similar desensitization system; but most likely it couples this receptor to a different G-protein-mediated system to produce the opposite electrophysiological result from the one that occurs in vertebrates. In the end the brain still gets the signal.
The lessons of the comparative physiology of vertebrate and invertebrate photoreceptors is that the basic components have been highly conserved but their linkage has developed differently. The basic input of photons is the same; the output hyperpolarization or depolarization of the photoreceptor cell is completely different. In between there has been a high degree of conservation: retinaldehyde, 7-membrane spanning receptors, G proteins, <*_>beta<*/>-arrestin, phospholipase C, nonselective cation channels; but the circuitry is different. The evolutionary invention was not in the types of proteins but in software for linking signaling and responding pathways together.
New Components and Their Evolutionary Value
Not all the remodeling of the eukaryotic cell is the equivalent of rearranging the furniture. There are, of course, new genes whose expression facilitated rapid evolutionary change. In the computer analogy these are the hardware improvements, which often provide new capacities for software innovations. As we shall see, some of these new genes may have persisted underutilized for extended periods of time, until the appropriate software mechanisms were developed to make use of them. In most cases the origins of these genes is traceable to more primitive structures that were stitched together by gene duplication and exon shuffling, but in some cases there is little clue as to their origins. It seems likely that some of these specific genes are crucial for major branches of macroevolution. Although one can tabulate many genes that would qualify as a "great moment in evolution", I will discuss only two structures dependent on new genes that are important for the major radiations within the vertebrates: myelin and feathers.
The biophysical features of nerve conduction explained by cable theory show that the rate and efficiency of nerve conduction increase with the diameter of the nerve fiber and with the decrease in the capacitance of the plasma membrane. To process complex information or to respond quickly to a predator or to capture food, rapid nerve conduction is obviously advantageous.
<#FROWN:J12\>Of those animals which had positive isolations from the oviducts, 68.8% had isolations from both oviducts.
Since a relatively high inoculating dose of chlamydiae was used to infect the guinea pigs, we wanted to determine if ascending infection into the uterus and oviducts would develop with lower numbers of organisms. Thus, a dose-response experiment was performed in which animals were inoculated with 10<sp_>4<sp/>-10<sp_>8<sp/> IFU of GPIC (Table 1). When animals were sacrificed on days 7 or 9, no dose-related differences were found in the number of animals acquiring infection in the oviducts.
Histopathologic Analysis of Endometrium
After it was established that ascending infection was a common occurrence in the genital tract, we examined the various endometria and oviducts for pathologic changes associated with infection. The histopathology of the exo-and endocervix has been previously described. The data presented is from a pool of animals killed at various times after infection and is based on the total number of a given specimen examined, rather than the total number of animals, i.e., two uterine horns and two oviducts per animal. The percentage of total endometrial tissues with specific pathologic changes is presented in Figure 2. Only 1 in 12 animals showed inflammation by day 3. Acute inflammation was the most prevalent finding at all time points from day 7-12, peaking at day 9. Peak infiltrates with lymphocytes and plasma cells were also seen at day 9 although the percentage of animals showing these findings was less than those showing acute inflammation. Fibrosis of endometrial stroma was virtually unseen. The percentage of animals displaying an inflammatory response decreased by day 12, and with the exception of a single animal on day 20 (not shown), no histopathology was identified in the endometrium after day 12 even including specimes examined on days 30 and 75-85. Of interest is the observation that on day 7, only 43.6% of the uterine horns were positive for pathology whereas 63.8% were isolation positive. Similarly, on day 9, 57.7% were positive for pathology with 66.7% positive for isolation.
<O_>caption&table<O/>
To further characterize the pathologic findings, we semiquantified the morphologic findings in endometria showing abnormalities (Figure 3). Animals were included if any pathologic parameter was positive. The predominant pathologic finding was acute inflammation particularly on days 7 and 9. Polymorphonuclear leukocytes infiltrated the glandular surface epithelium, filled the endometrial gland lumens, and were scattered throughout the superficial stroma. Chronic inflammation was also present at all timepoints, but in lesser quantity. The lymphocytes were arranged in loose aggregates in the superficial stroma with occasional transformed lymphocytes identified (Figure 4). Plasma cell infiltrates were seen from days 5 through 12 but always in smaller numbers than either polymorphonuclear leukocytes or lymphocytes. Plasma cells were scattered throughout the endometrial stroma in a patchy distribution.
<O_>figure&caption<O/>
In 12 uninfected control animals, an occasional solitary lymphocyte aggregate was seen in either the uterine fundus or horns. Infrequently, one or two endometrial glands contained a few scattered polymorphonuclear leukocytes. There were no plasma cells, fibrosis, or erosions identifed in any control animal. Thus, the pathology described earlier in infected animals was obviously a result of the infection and not associated with a normal resident endometrial response.
Histopathologic Analysis of Oviduct and Mesosalpinx
Figure 5 illustrates the percentage of mesosalpingeal tissue and oviducts that showed infiltrates of polymorpho-nuclear leukocytes, lymphocytes or plasma cells and/or fibrosis. Because many animals had only unilateral pathology, the data is presented based on the total number of tissues examined. The number of specimens in either the mesosalpinx or oviduct that had pathologic changes was low at 7 days after infection, when the isolation of organisms from the same specimens was maximum. Nevertheless, by 9 days after infection, the number of samples with pathologic changes had doubled although this number was never as great as the total number of specimens from which chlamydiae were isolated. Thus, some tissues were isolation positive but did not have detectable pathology.
<O_>figure&caption<O/>
Early in the infection (days 5-12), acute and chronic inflammatory responses as well as plasma cell infiltration were common in both the mesosalpinx and oviducts. By day 30, the acute inflammatory and plasma cell responses had markedly diminished in both tissues; however, lymphocytic infiltrates and fibrosis were still evident in the mesosalpinx. Pathologic changes in the oviducts at day 30 and beyond were minimal. The reactions in the mesosalpinx continued to decrease, but 21% of the tissues still had obvious fibrosis 75-85 days after infection, and 19% had ongoing chronic inflammation.
Semiquantification of the morphologic findings in the specimens showing abnormalities is presented in Figure 6. The early stage of the infection was characterized by an acute inflammatory reaction in both the mesosalpinx and oviducts (Figure 7). Chronic inflammation was also present early but did not reach its peak level until day 12 in the mesosalpinx as did plasma cell infiltration (Figure 8). The appearance of plasma cells corresponded to the development of antibody that normally is detectable about day 10. The development of fibrosis was primarily restricted to the mesosalpinx and was maximum at day 12 although it persisted and was still obvious as late as 75-85 days after infection. Tubal dilatation (hydrosalpinx) was apparent in 12% of the observed oviducts in the 75-85 day period, and in some cases, was marked with tubal diameters as great as 1 cm.
<O_>figures&captions<O/>
Oviduct and mesosalpingeal tissues were also stained with guinea pig anti-GPIC antibodies followed by peroxidase-labeled rabbit anti-guinea pig IgG to visualize and localize chlamydial antigen. Chlamydial antigen and inclusion bodies were commonly detected in the 7-12 day period in the epithelial cells of the oviduct (Figure 9).
<O_>figure&caption<O/>
Oviducts from uninfected animals were also examined to determine whether any inflammatory infiltrates were normally present. No acute or chronic inflammation was identified in any of 12 control animals nor were any plasma cells, fibrosis, or erosions found.
Discussion
In this study, we describe a model for chlamydial genital infection in which ascending infection to the endometrium and oviducts routinely occurs as a result of vaginal inoculation of the chlamydial agent. Other than the primate, which is limited in usefulness by expense, the guinea pig:GPIC model represents the only animal model in which ascending infection from a vaginal inoculation, analogous to the human situation, can be commonly demonstrated, even though GPIC is a member of the C. psittaci, it has been found to elicit ocular and genital infections remarkably paralleling the corresponding human disease. GPIC primarily infects superficial epithelial cells of the cervix and epithelial cells of the male urethra, but most significantly, the infection can be transmitted sexually in guinea pigs. Moreover, newborns of infected mothers acquire a conjunctival infection by passage through the birth canal and can develop a pneumonia typical of chlamydial pneumonia of the newborn when inoculated intranasally. Hormonal effects on chlamydial infection can and have been effectively studied in the guinea pig, since of all the rodents, the reproductive system of the female guinea pig most closely resembles the human with regard to their long estrous cycle (17 days), spontaneous ovulation, and actively secreting corpus luteum. Immunologically, guinea pigs develop both cell-mediated and humoral immune responses to GPIC. Analogous to humans, guinea pigs also produce significant antibody responses to the major outer membrane protein (39 kDa), the chlamydial GroEL (57 kDa), the Omp2protein (60 kDa), and lipopoly-saccharide of GPIC. Also comparable to humans is the short immunity to reinfection that occurs, with animals becoming susceptible to reinfection as early as 2 months after the resolution of a primary infection.
<O_>caption&figure<O/>
Previously, we have only noted the develoment of upper tract disease when animals have been manipulated either by treatment with cyclophosphamide or estradiol. However, in those studies, the early timepoints of the infection were not carefully evaluated or studied in a large number of animals. In the current investigation, we analyzed the tissues early after infection and found that in a high percentage of guinea pigs, chlamydiae can be isolated form the oviducts within 1 week of vaginal inoculation. The presence of organisms in the oviducts is limited in duration, with disappearance from the oviducts concomitant with the resolution of cervical infection and the development of both cell-mediated immunity and serum and secretion antibodies.
<O_>caption&figure<O/>
The ascending nature of the infection is confirmed by the fact that no chlamydiae could be isolated from the oviducts on day 3 and only a few on day 5 despite the presence of organisms in the cervix of virtually all animals. Isolation from the uterus paralleled that of the oviducts although animals became positive earlier in the uterus. Thus, several days were required for the organisms to reach both the endometrium and the oviducts. However, it is significant that organisms could be recovered from the endometrium and oviducts of almost 80% of the guinea pigs assessed at day 7. Moreover, the appearance of the bacteria in the oviducts by day 7 also represents a remarkably rapid ascending infection for a non-motile, slowly growing organism. If one estrapolates these data to the human situation, they suggest that a much higher number of women develop upper tract infection than previously believed and that this may occur quickly after infection. As stated earlier, Jones et al have reported evidence in support of these data in cases of chlamydial infections in women.
However, it is interesting that not all animals that acquire tubal infection go on to develop tubal pathology. Although 78% of the animals had tubal infection on day 7, only 45% were found to have pathologic changes in the oviducts on day 9, the time at which maximum pathology was noted. Swenson et al using the direct injection model with MoPn also observed that not all injected animals developed salpingitis. A similar phenomenon was also noted in our study with regard to isolation and pathologic changes in the endometrium. Significantly, it has been recently reported that chlamydiae could also be isolated from the fallopian tubes of women without laparoscopic evidence for salpingitis. These data would suggest that, in some cases, the immune response may be sufficiently rapid in producing those effector functions that can resolve the infection. Antibody can be detected in genital secretions of guinea pigs as early as 10 days after infection, and we have found even earlier appearance of antibody on some occasions. Cell-mediated immunity, which is also required for resolution of a chlamydial genital infection, can also be present as early as 10 days.
An alternative explanation for the variation in incidence of pathology is that other physiologic factors may be affecting their development. We have previously reported that when guinea pigs are treated with estradiol in either pharmacologic or physiologic doses, a markedly enhanced infection is noted in the cervix with a significantly increased number of animals developing hydro-salpinx. Moreover, the infection is prolonged when compared with untreated controls. Sweet et al have also noted that the onset of acute salpingitis in women occurred significantly more often within 7 days following the beginning of menses than at other times in their menstrual cycle. In addition, it has been well-described that treatment with oral contraceptives does increase the number of individuals from whom chlamydiae can be isolated. Although hormonal changes are not the only possible factor in the variance seen in tubal pathology, they certainly may have some role based on available data. Since this model resembles humans in that not all individuals develop overt salpingitis, it will be useful to investigate those endogenous or exogenous factors that alter the incidence of salpingitis. An important point, however, is that a high number of individuals do have organisms in the oviducts, and the development of overt salpingitis may be dependent on factors that prevent elimination of the bacteria from the oviducts or mediate the development of pathologic changes. Furthermore, the presence of organisms in the oviducts may not necessarily mandate the production of disease with harsh sequelae.
Finally, it was found that the pathologic changes occurring as a result of chlamydial genital infection in the guinea pig with the GPIC agent were remarkably parallel to that in human chlamydial endometritis and salpingitis.
<#FROWN:J13\>
Between Bone Tissue-Type Comparison
The second step in the tissue-type comparison was to analyze interbone variation using the bone sites selected as least and most variable - the midshaft femur and vertebral body, respectively. Waldron (1989) reported that between bones of the same skeleton, elemental levels can vary by as much as a factor of two. Elaborating upon the method employed by Tanaka et al. (1981) to compare vertebra to other bones, the Nubian and modern samples were compared for the degree of dissimilarity between cortical and cancellous tissues using a modification of the two-sample Student's t-test (Greene, 1973). Greene first developed this modified t-test as a method for comparing hominoid species for differences in the degree of sexual dimorphism expressed in the dentition. the method was further revised by Greene (1989) for comparisons "between populations and comparisons between generations within populations" (p.121). This modification was used in the present investigation as follows:
<O_>formula<O/>
where subscript: 1=Nubian midshaft femur
2=Nubian vertebral body
3=Modern midshaft femur
4=Modern vertebral body
Table 17 lists the results of these comparisons. Barium, copper, iron, potassium, magnesium, manganese, and zinc show significant (p<0.05) tissue-type differences between Nubian and modern samples. Of these, barium, iron, magnesium, manganese, and zinc are more variable in the Nubians than the moderns. Conversely, copper and potassium are more variable in the modern samples.
<O_>caption&table<O/>
The more variable elements in the Nubian bone may indicate enrichment from the soil. Except for zinc, each is found at higher concentrations in the soil than in the bone. The reduced variability of copper and potassium in the Nubian bone (compared to the moderns) may indicate depletion of these elements into the soil.
Summary of Element Selection
Seven tests were employed to identify five measures of diagenesis among the elements - 1) range overlap between the modern and Nubian samples; 2) variability among the elements using CV; 3) antagonistic/synergistic interactions between the elements using multi-element correlations; 4) analysis of bone contamination from elements in the soil; and 5) variation between tissue-types to assess enrichment/depletion of elements in bone. Based on these measures, elements were divided into those minimally, moderately and highly affected by diagenesis.
Nubian means for the midshaft femur and for all sites combined were overlayed on the modern distributions for each element. Only one element fell outside the modern range - boron. Manganese and iron varied for the combined site mean, but the Nubian femur mean was within the modern distribution.
Coefficients of variation were metrically ranked for the Nubians and moderns across all sites combined and for the midshaft femur alone. In all four rankings, calcium, phosphorus, magnesium, sodium, and strontium were the least variable 100% of the time. Vanadium was among the six least variable 87% of the time. The three most variable elements across all four rankings were boron, manganese, and zinc.
When elements were ranked across all bone sites ordinally, again calcium, phosphorus, sodium, magnesium, and strontium were the least variable for both the Nubian and modern samples. Boron and manganese were the most variable.
All of the multi-element correlations noted for the Nubians could be explained using the modern matrix or the literature, except for zinc/phosphorus. The modern and Nubian patterns matched for calcium/phosphorus, nickel/vanadium, potassium/sodium, and boron/lead. The Nubians agreed with the literature for those elements believed to be diagenetic - copper, iron, and manganese. The strong strontium/barium and sodium/magnesium correlations were also expected based on published studies (Buikstra et al., 1989).
Comparison of the Nubian bones to associated soil values and to the modern sample showed that calcium, phosphorus, strontium, and sodium were not affected by enrichment or depletion from the surrounding soil. Boron, iron, and manganese showed a pattern indicative of enrichment, and potassium of depletion. Barium, copper, magnesium, nickel, lead, and zinc were indeterminate because while Nubian bone levels were below soil values, they were equivalent to the modern samples.
Tissue-type comparisons within and between bones also showed distinct patterns. The within-bone comparison between cortical and cancellous bone showed that elements concentrated in the same tissues for both the femur and humerus. When the percent variation between tissues was compared, calcium, phosphorus, strontium, magnesium, sodium, and potassium were found in highest concentrations in cortical bone. The latter three, significantly so. All other elements concentrated in cancellous bone. According to the literature (Buikstra et al., 1989; Waldron, 1989), elements concentrating in cancellous bone are more likely the result of diagenesis than those in cortical bone.
The between-bone tissue-type comparison showed degree of dissimilarity for the least and most variable bone sites in the Nubian skeleton. Barium, iron, magnesium, manganese, and zinc were more variable among the Nubian bones, possibly indicating enrichment. Copper and potassium, were found to vary less in the Nubians than moderns, indicating potential depletion.
Other than the between bone comparison, for each measure, calcium, phosphorus, sodium, and strontium appeared the least affected by diagenesis. Magnesium also rated high in each test except the soil and between-bone comparisons. Boron and manganese were consistently ranked the most variable. Based on these comparisons, calcium, phosphorus, magnesium, strontium, and sodium are considered minimally affected; barium, copper, nickel, vanadium, iron, potassium, and zinc, moderately affected; and, boron and manganese, highly affected.
SUMMARY OF DIAGENESIS EVALUATION
Analysis of the degree of diagenesis was conducted along multiple lines of inquiry. First, the quality of the bone composition was assessed. Then bone sites were compared to determine which were the least affected in the depositional environment. Finally, elements were tested for alteration and grouped into those minimally, moderately, and highly affected by diagenesis.
Bone preservation proved exceptional by all measures except %ash. However, this discrepancy is very likely the result of experimenter error. Of the bone sites, the midshaft femur proved the least altered, based on a rank-ordering of coefficients of variation for both Nubian and modern samples. Among the elements, calcium, phosphorus, magnesium, sodium, and strontium were judged minimally affected by diagenesis. Boron and manganese were the most affected, and all other elements were moderately altered.
This analysis employed virtually all the conventional methods suggested in the literature for assessing diagenesis (Price, 1989; Buikstra et al., 1989). Several traditional tests were modified in this study, and others were introduced for future use. Whereas diagenesis cannot be measured directly, the present analysis demonstrates a strong circumstantial argument for selection of the least affected bone and elements in the Nubian remains.
CHAPTER 5:
RESULTS AND DISCUSSION II:
BIOCULTURAL RECONSTRUCTION
Numerous biocultural reconstructions of ancient human populations have been conducted utilizing chemical analyses (for reviews of this literature, see Klepinger 1984; Price et al, 1985; Gilbert, 1975; Schoeninger 1979; Sillen and Kavanagh, 1982; Price, 1989; Buikstra et al., 1989). Most promote an interdisciplinary approach to the use of elemental variation.
A principal advantage of the Nubian remains to elemental analysis is the rich biocultural context within which the elemental data can be interpreted. Price et al (1985) and Blakely and Beck (1981) recommended that chemical analyses be conducted in concert with nutritional and paleopathological investigations of ancient remains, the later stating that elemental research is "of little utility when ... findings cannot be corroborated by other sources of information, such as demographic data or pathological diagnoses" (p.422). Martin and Armelagos (1985) concurred, advising that elemental analyses will not realize their full potential "until they are used in conjunction with other techniques such as gross and microscopic analyses. A thorough understanding of nutritional deficiencies and health in prehistory will require an examination of anatomy, pathology, histology, and chemistry in a systematic analysis using multiple indicators" (p.527).
The biocultural analysis of elemental variation at Kulubnarti will proceed along two lines of inquiry. First, a general assessment by age and sex will be conducted. Then given the importance of independent lines of confirmatory evidence noted by the authors quoted above, the second phase of the investigation will focus on previous studies of aging, nutrition, and disease at Kulubnarti. Based on the results of Chapter 4, only the femur will be used in these analyses. Also, given the substantial evidence of their idagenetic alteration, boron and manganese will be excluded from further consideration.
BIOCULTURAL DIMENSIONS AT KULUBNARTI
As previously discussed, the people of the Batn el Hajar lived in villages of perhaps a dozen households dependent on small-scale farming for their livelihood. Agriculture was intensified where possible by simple irrigation systems and retaining walls which protected alluvial soils. Staple crops included sorghum, dates, millet, barley, beans, lentils, peas, and a small amount of wheat. Coprolite analysis also revealed that some fish and crocodile were consumed (Cummings, 1988). In addition, a few cattle, sheep and pigs were kept, but animal protein appears to have been a minor part of the Nubian diet (Carlson et al., 1974; Adams, 1977).
From the standpoint of dietary variation within the Kulubnarti population, the archaeological (Adams, 1977) and biological records (Van Gerven et al., 1981) overwhelmingly support a single interpretation: Kulubnarti was an egalitarian community of household producers and consumers. The few communal activities that existed were limited to production and maintenance of the village saquia (waterwheel) and the irrigation ditches and retaining walls. There is no indication of a political or economic elite with preferential access to critical resources.
BIOCULTURAL RECONSTRUCTIONS UTILIZING ELEMENTAL ANALYSIS
Elemental analyses in the past have been applied to questions of: differential access to food resources related to status (Brown, 1973; Schoeninger, 1979; Lambert et al., 1979; Blakely and Beck, 1981; Hatch and Geidel, 1985), sex (Brown, 1973; Schoeninger, 1979; Lambert et al., 1979; Price et al., 1986); changes in subsistence methods (Lambert et al, 1979; Gilbert, 1975; Jaworoski et al, 1985; Katzenberg, 1984; Price and Kavanagh, 1982; Sillen 1981; Schoeninger 1982); relative contributions of plant versus animal resources (Lambert et al, 1983; Price et al., 1986); contributions of marine versus terrestrial resources (Connor and Slaughter, 1984); patterns of weaning (Sillen and Smith, 1985); and, residence patterns (Ericson, 1985). However, inclusion of many of these parameters in the present investigation was not possible.
For example, the Kulubnarti Nubians were egalitarian, therefore, evidence for preferential access to food resources based on political or economic status was absent. Given that only individuals from the Feudal period were examined, no analysis of diachronic change in subsistence patterns was possible. Also, an examination of marine versus terrestrial resources could not be conducted because the Batn el Hajar is land-locked. Patterns of weaning could not be assessed because only adults were studied. And finally, analysis of residence patterns requires examination of elemental concentrations in the teeth and bone to determine childhood versus adult strontium levels. Teeth were not studied in the present investigation.
Lacking evidence for either political/economic stratification or temporal change, elemental variation related to nutrition and disease at Kulubnarti was analyzed from the demographic perspectives of age and sex. Price (1985) took a similar approach in his analysis of a prehistoric Amerindian population stating that "Late Archaic groups are generally regarded as egalitarian so that dietary differences associated with rank, status or position do not play a major role" (p 450). Outlined below is a review of the literature pertaining to sex and age-related variation in elemental concentrations, followed by the pertinent findings of the present investigation.
Prior to an analysis of elemental variation by age, sex, diet or disease however, it is important to describe features of those elements that most often appear in such reconstructions. These include strontium, barium, magnesium, and zinc.
Strontium
Strontium is an alkaline earth metal with an uneven distribution throughout the lithosphere (Odum, 1951). There are four naturally occurring isotopes and 14 radionuclides. Of the latter, two are the result of nuclear fission, which explains the vast body of knowledge amassed for this element.
Strontium has a similar ionization energy, ionic size, and electron configuration to calcium, and therefore behaves in a similar fashion. Both enter the foodchain at the plant level through soil and water, and the amount found in plants is directly proportional to the concentrations in the local environment. Plants do not discriminate between calcium and strontium in absorption, and therefore have a higher strontium concentration than that found at any other trophic level.
<#FROWN:J14\>In adults, the small number of patients and their uniformly poor performance status has not allowed clear identification of the high-risk group. However, certain clinical features, as well as immunologic and morphologic features, of adult ALL have been associated with poorer clinical performance.
Clinical Features
Sex, age, initial blast count, initial platelet count, and organ involvement are features that identify high-risk patients. Males do more poorly than females. Patients over 40 years of age have a shorter mean survival time than those below 40. Hepatosplenomegaly and/or lymphadenopathy at presentation has been associated with shorter remission duration, although the presence of isolated splenomegaly may indeed not be a high-risk factor.
The white blood cell count at presentation is clearly most important in predicting remission duration. Blast counts in excess of 100,000 cell/mm<sp_>3<sp/> were a dire prognostic sign in many series. Additionally, statistically significant longer remissions were demonstrated in patients presenting with an initial blast count of 10,000 cells/mm<sp_>3<sp/> or less. Survival is also adversely affected by decreased platelet counts at presentation: survival was significantly worse in patients with presenting platelet counts below 50,000/mm<sp_>3<sp/>. Finally, the presence of central nervous system involvement at the time of diagnosis was associated with a significantly shorter survival.
<O_>table&caption<O/>
Morphologic and Immunologic Subtypes
The L<sb_>3<sb/> B cell ALL has a very poor prognosis in most series. In rank order of increasingly good prognosis, the immunologic subtypes may be listed as follows: B cell ALL, T cell ALL, null cell ALL, and common ALL. Whether these immunologic subtypes are independent prognostic variables or whether they reflect differences in clinical presentation remains unclear at the present time.
Treatment
Chemotherapy with cytotoxic agents remains the cornerstone of the treatment of adult ALL. The present-day approach has been developed as a result of multiple clinical trials in both pediatric and adult patients. The therapy is divided into remission induction (restoration of the normal hematologic state with less than 5% blasts in the bone marrow aspirate), early intensification once remission is obtained, maintanencemaintenance therapy to maintain the leukemia-free state, and central nervous system prophylaxis or treatment.
Induction and Early Intensification
A comparison of various induction regimens is shown in Table 4-5. Initial complete remission rates with vincristine and prednisone in adults with ALL were approximately 40%-50%, and the remissions were of short duration. it was shown in a number of trials that the addition of daunorubicin or doxorubicin to the induction regimen increased the complete remission rate to 70%-80%. The addition of L-asparaginase to the vincristine and prednisone combination yielded similar success.
Cancer and Leukemia Group b (CALGB) showed that the addition of L-asparaginase after induction with vincristine and prednisone resulted in more long-term remissions than the simultaneous administration of the three drugs or than vincristine and prednisone used alone. Lister et al,. using vincristine, prednisone, daunorubicin, and L-asparaginase in combination reported a complete remission rate of 71%.
The concept of early intensification has been sparked by the disappointing duration of remission in the adult population. Gee et al. reported a mean remission duration of 24 months with a program incorporating cytosine arabinoside, 6-thioguanine, L-asparaginase, and carmustine after induction with vincristine, prednisone, and daunorubicin. Trials in the pediatric population by Aur et al. and Muriel-Sackman et al., using cytosine arabinoside, cyclophosphamide, and methotrexate intensification, showed no improvement in remission duration.
It appears safe to conclude that the approach to adult ALL should include intensive induction with vincristine, prednisone, and an anthracycline, with the addition of L-asparaginase soon after. An example of such a regimen is shown in Table 4-6. The role of early and late intensification remains unclear.
<O_>table&caption<O/>
Maintenance
The need for effective maintenance therapy in ALL is evidenced by the high relapse rates in patients in whom a complete marrow remission is obtained. Short remission durations were noted in studies using short intensive induction courses of chemotherapy. Similarly, patients in complete remission who discontinue therapy relapse quickly. The most commonly used maintenance schedule has been methotrexate and 6-mercaptopurine. Attempts to improve maintenance duration with the addition of intermittent vincristine, prednisone, cytosine arabinoside, and cyclophosphamide have not proved successful. Sallan et al. reported improvement in duration of remissions with the addition of doxorubicin to vincristine, methotrexate, 6-mercaptopurine, and prednisone. However, it is too early to know whether this improvement will stand the test of time. As there are so few patients remaining in complete remission, the question of duration of maintenance therapy remains unanswered in adults. In children, it appears that therapy can safely be stopped after 3 years of relapse-free maintenance therapy.
Relapse
Hematologic relapse occurs in 70%-80% of adult patients with ALL in whom a complete remission is obtained. Once in relapse, reinduction with vincristine, prednisone, anthracyclines, L-asparaginase, or cytosine arabinoside is successful in 80% of children and 50%-70% of adults. These remissions are usually short, especially in patients having relapsed during maintenance chemotherapy. New chemotherapeutic agents, such as m-AMSA, VP-16 and VM-26, are being used in refractory ALL with variable success. High-dose therapy with either methotrexate or cytosine arabinoside is also being used.
Central Nervous System Prophylaxis and Treatment
With improvement in the treatment of ALL in children, the incidence of central nervous system involvement by leukemic infiltration became significant, occurring in 50%-60% of patients in hematologic remission. The problem of central nervous system involvement was realized as early as 1965, when Frei et al. of CALGB instituted intrathecal methotrexate in the induction regimen for childhood ALL. This led to a series of trials over the years to examine the question of central nervous system prophylaxis. The St. Jude's group showed a decrease in incidence to less than 10% with the institution of whole-brain irradiation therapy in intrathecal methotrexate. The same group showed similar results using 2400 rads craniospinal irradiation. This group and others have shown that the goal of central nervous system prophylaxis can be achieved with whole-brain irradiation therapy and intrathecal methotrexate, avoiding spinal irradiation, which causes bone marrow suppression and a higher incidence of complications. Attempts to avoid central nervous system disease by incorporating drugs that cross the blood-brain barrier have been ineffective.
The problem of central nervous system leukemia was observed in the adult population as well, with an incidence of at least 40%-50%. Omura et al. showed that 2400 rads whole-brain radiotherapy combined with five doses of intrathecal methotrexate in doses of 10 mg/m<sp_>2<sp/> (15 mg maximum dose) decreased central nervous system incidence from 11/34 in an untreated group to 3/28 in a treated group. Unfortunately, no difference in the duration of hematologic remission nor survival could be seen between the two groups. The L<sb_>2<sb/> protocol series from memorial Hospital, which incorporated intrathecal methotrexate without radiotherapy, had a 66% incidence of central nervous system disease in patients with white counts above 25,000/mm<sp_>3<sp/>. It seems that the combination of whole-brain radiotherapy and intrathecal methotrexate should be advocated at this time. Investigations into intraventricular drug delivery via Ommaya reservoir are currently under way.
In patients with documented central nervous system leukemia, treatment is approached by a combination of radiotherapy and intrathecal methotrexate or cytoxine arabinoside. Clinical presentation of central nervous system leukemia is variable. Cranial nerve palsies are common due to leukemic infiltration of nerve sheaths at presentation. Diagnosis depends on cytologic examination of cytocentrifuge preparations of cerebrospinal fluid. The CNS relapse rates are high in patients in whom intrathecal therapy is discontinued once central nervous system relapse has occurred.
New Approaches to ALL
Despite advances in remission induction rates, the ability to cure ALL with conventional chemotherapy in the majority of adults remains elusive. New approaches have emerged to improve results in these patients.
Bone Marrow Transplantation
Bone marrow transplantation of 22 patients with HLA-compatible donors, aged 4-22 years, in second or subsequent remission, has resulted in long-term survival: 50% remain in remission at 15-35 months after transplant. Only 18% of those transplanted in relapse have survived. Therefore, it appears that transplant of patients in second or subsequent remission is preferable to transplant once disease is refractory. When transplantation was compared with conventional chemotherapy in patients in second or subsequent remission, 11 of 24 of the transplanted group remained disease free at 17-55 months, whereas only 1 child of 21 treated with conventional chemotherapy remained in remission for more than 2 years. Certainly, bone marrow transplantation is not available for all patients. However, if available, one should screen family members for HLA-compatible donors and should consider transplant for patients under 30 years old during second remission or possibly in first remission if they have 'high-risk' disease (i.e., a high white count at presentation, and null, B, or T cell disease).
Serotherapy and Specific Therapy
With the expansion of immunologic techniques and subtyping of ALL, passive serotherapy with anti-CALLA antibodies has been attempted in patients with CALLA-positive ALL. Initial responses have been noted; however, these have been very short lived. The role of immunologic manipulation in the therapy of ALL remains experimental and speculative. Finally, specific selective therapy has been reported with 2-deoxycoformycin, an inhibitor of adenosine deaminase in a patient with T cell ALL.
Outlook
ALL remains a fatal disease in 70%-80% of affected adults. It has been seen that the favorable results in adult ALL have more or less been confined to adolescents and young adults, with the 'older' population faring less well.
The challenges seem clear. The role of both early and late intensification requires further study. Furthermore, more precise designations of poor prognostic signs need to be made in the adult population so that more intensive treatment may be attempted. The immunologic approach to therapy is in its infancy and clearly needs development. It is hoped that these advances will come quickly and improvement in long-term survival will be realized.
Chronic Lymphocytic Leukemia
Chronic lymphocytic leukemia (CLL) was described by Dameshek as "an accumulative disease of immunologically incompetent lymphocytes." This concise definition describes the majority of patients with CLL, who present with physical findings and laboratory data that are consistent with Dameshek's theory. CLL is a disease of older persons: less than 10% of patients present before the age of 50 years. Men are about two times as likely to develop CLL as are women. As in all forms of leukemia, CLL is more common in the white than in the black population.
The diagnosis of CLL is based on the minimal requirement of an absolute and sustained lymphocytosis in the peripheral blood of no less than 15,000 cells/mm<sp_>3<sp/>. The bone marrow is hypercellular and more than 40% of the cells are lymphocytes. The lymphocytes in peripheral blood and marrow are of the small, well-differentiated type (Figure 4-4). Slowly enlarging lymph nodes and gradual enlargement of the liver and spleen due to the accumulation of neoplastic lymphocytes may occur early or late in the course of the disease. The immunologic incompetence of the expanding lymphocyte population expresses itself in hypo-gammaglobulinemia with predisposition to infections, and in the emergence of such autoimmune phenomena as the production of antibodies against host red blood cells.
Biology
Lymphocytes in Chronic Lymphocytic Leukemia
CLL primarily represents a clonal proliferation of B lymphocytes (Figure 4-2). The malignant cells have immunoglobulin molecules on their surfaces bearing the same idiotype and light chain type. The most common surface immunoglobulins are IgM and IgD. The finding of two heavy chain classes does not preclude the concept of monoclonality, as the IgD and IgM have the same idiotype specificity and presumably reflect a stage of differentiation of normal B lymphocytes. CLL lymphocytes also express the Ia antigen, the receptor for C'3, and the receptor for the Fc portion of immunoglobulin.
Studies of the clonal origin of CLL have used glucose-6-phosphate dehydrogenase (G6PD) as a marker in patients who are heterozygous at this locus. Skin tissue from these patients manifested both A and B G6PD types. However, as one would predict, the CLL B lymphocytes manifested only one type of G6PD. In contrast to the neoplastic B lymphocytes, the T lymphocytes, granulocytes, erythrocytes, and platelets displayed both enzyme types in proportions similar to those found in skin. These findings indicate that CLL involves only committed B lymphocyte progenitors.
<#FROWN:J15\>
Molecular Mechanisms of Drug Addiction
Eric J. Nestler
Laboratory of Molecular Psychiatry, Departments of Psychiatry and Pharmacology, Yale University School of Medicine, Connecticut Mental Health Center, New Haven, Connecticut 06508
Drug addiction has afflicted mankind for centuries, yet the mechanisms by which particular drugs lead to addiction, and the genetic factors that make some individuals particularly vulnerable to addiction, have remained elusive. From a clinical perspective, drug abuse continues to exact enormous human and financial costs on society, yet all currently available treatments for drug addiction are notoriously ineffective. The search for a better understanding of the neurobiological mechanisms underlying the addictive actions of drugs of abuse and of the genetic factors that contribute to addiction should be given a high priority, as this should result in crucial advances in our ability to treat and prevent drug addiction.
From the basic neuroscience perspective, study of the neurobiology of drug addiction offers a novel opportunity to establish the biological basis of a complex and clinically relevant behavioral abnormality. Many prominent aspects of drug addiction in people can be clearly reproduced in laboratory animals, in striking contrast to most other forms of neuropsychiatric illness, such as psychotic and affective disorders, animal models for which are much harder to interpret. Advances made in the study of drug addiction should provide important insights into mechanisms underlying some of these other disorders.
Three terms related to drug abuse are used commonly: tolerance, dependence, and addiction. Tolerance represents a reduced effect upon repeated exposure to a drug at a constant dose, or the need for an increased dose to maintain the same effect. Dependence is defined as the need for continued exposure to a drug so as to avoid a withdrawal syndrome (physical and/or psychological disturbances) when the drug is withdrawn. Dependence is considered a priori to result from adaptive changes that develop in body tissues in response to repeated drug exposure. The traditional distinction between physical and psychological dependence is somewhat artificial, since both are mediated by neural mechanisms, possibly even similar neural mechanisms, as will be seen below. Addiction is defined as the compulsive use of a drug despite adverse consequences. In the past, physical dependence was part of the definition of addicton. However, the requirement for physical dependence as a neccessary or sufficient aspect of drug addiction is no longer considered valid. Many drugs with no abuse potential, for example, <*_>beta<*/>-adrenergic antagonists, clonidine, and tricyclic antidepressants, can produce marked physical symptoms on withdrawal. On the other hand, many unquestionably severe abusers of some drugs have little or no physical withdrawal syndrome upon cessation of drug exposure (e.g., most marijuana or cocaine users). Similarly, not all drugs of abuse produce tolerance to all of their effects.
This article reviews the results of recent research efforts that have begun to characterize the neurobiological basis of compulsive drug use. Its major focus is on opiates and cocaine, since the addictive mechanisms underlying the actions of these drugs are the best understood.
Cellular site of drug addiction
The discovery of endogenous opiate receptors in the 1970s raised the possibility that opiate addiction might be mediated by changes in these receptors. However, a decade of research has failed to identify consistent changes in the number of opiate receptors, or changes in their affinity for opiate ligands, under conditions of opiate addiction (Loh and Smith, 1990). Changes in levels of endogenous opioid peptides also do not appear to explain prominent aspects of opiate tolerance and dependence. The discovery that cocaine and other addictive psychostimulants acutely inhibit the reuptake or stimulate the release of monoamines throughout the brain has focused study of their addictive mechanisms on the regulation of monoamine neurotransmitters and their receptors. These studies too have been disappointing because it has been difficult to demonstrate consistent long-term changes in specific neurotransmitter or receptor systems in brain regions thought to underlie psychostimulant addiction (see Clouet et al., 1988; Liebman and Cooper, 1989; Peris et al., 1990).
The failure to account for important aspects of opiate and psychostimulant addictions in terms of regulation of neuro-transmitters and receptors has shifted attention to postreceptor mechanisms. Most types of neurotransmitter receptors present in brain produce most of their physiological responses in target neurons through a complex cascade of intracellular messengers. These intracellular messengers include G-proteins (Simon et al., 1991), which couple the receptors to intracellular effector systems, and the intracellular effector systems themselves, which include second messengers, protein kinases and protein phosphatases, and phosphoproteins (Nestler and Greengard, 1984, 1989). Regulation of these intracellular messenger pathways mediates the effects of the neurotransmitter-receptor systems on diverse aspects of neuronal function, including gene expression. Given that many important aspects of drug addiction develop gradually and progressively in response to continued drug exposure, and can persist for a long time after drug withdrawal, it is likely that the regulation of neuronal gene expression is of particular relevance to addiction.
In recent years, the increasing knowledge of intracellular messenger pathways has provided an experimental framework for studies of the molecular mechanisms underlying drug addiction. These investigations have demonstrated that changes in the activity of G-proteins and the cAMP second messenger and protein phosphorylation pathway mediate important aspects of opiate, and possibly cocaine, addiction in a number of drug-responsive brain regions.
Molecular mechanisms underlying opiate tolerance, dependence, and withdrawal: studies in the locus coeruleus
The locus coeruleus (LC) of the rat has served for many years as a useful model of opiate action. The LC is the largest noradrenergic nucleus in brain, located bilaterally on the floor of the fourth ventricle in the anterior pons. It is particularly suited for biochemical and molecular investigations, as it is a relatively homogeneous brain region that has been extensively characterized anatomically and electrophysiologically.
Pharmacological and behavioral studies have indicated that modulation of LC neuronal firing rates contributes to physical aspects of opiate addiction, namely, physical dependence and withdrawal, in several mammalian species, including primates (see Redmond and Krystal, 1984; Rasmussen et al., 1990). The importance of the LC in mediating opiate addiction is highlighted by a recent study that examined the effects of local injection of an opiate receptor antagonist into various brain regions of opiate-dependent rats (Maldonado et al., 1992). The most severe opiate withdrawal syndrome was produced by antagonist injections into the LC, which, in fact, elicited a withdrawal syndrome even more severe than that seen following intracerebroventricular administration.
Acute opiate action in the LC
The mechanism of acute opiate action in the LC, based on electrophysiological and biochemical studies, is well established and is shown schematically in Figure 1 (top). Acutely, opiates decrease the firing rate of LC neurons via activation of an inward rectifying K<sp_>+<sp/> channel (Aghajanian and Wang, 1987; North et al., 1987) and inhibition of a slowly depolarizing, nonspecific cation channel (Aghajanian and Wang, 1987; M. Alreja and G. K. Aghajanian, unpublished observations). Both actions are mediated via pertussis toxin-sensitive G-proteins (i.e., G<sb_>i<sb/> and/or G<sb_>o<sb/>) (Aghajanian and Wang, 1986; North et al., 1987), and inhibition of the nonspecific cation channel is mediated by reduced neuronal levels of cAMP and activated cAMP-dependent protein kinase (Aghajanian and Wang, 1987; Wang and Aghajanian, 1990; Alreja and Aghajanian, 1991). Opiates acutely inhibit adenylate cyclase activity in the LC (Duman et al., 1988; Beitner et al., 1989), as is the case in many other brain regions (see Childers, 1991), and inhibit cAMP-dependent protein phosphorylation (Guitart and Nestler, 1989). Such regulation of protein phosphorylation presumably mediates the effects of opiates on the nonspecific cation channel through the phosphorylation of the channel itself or some associated protein. Opiate regulation of protein phosphorylation also probably mediates the effects of opiates on many other aspects of LC neuronal function, including some of the initial steps underlying longer-term changes associated with addiction.
Chronic opiate action in the LC
Upon chronic opiate treatment, LC neurons develop tolerance to the acute inhibitory actions of opiates, as neuronal firing rates recover toward pretreatment levels (Aghajanian, 1978; Andrade et al., 1983; Christie et al., 1987). The neurons also become dependent on opiates after chronic exposure, in that abrupt cessation of opiate treatment, for example, by administration of an opiate receptor antagonist, leads to an elevation in LC firing rates manyfold above pretreatment levels (Aghajanian, 1978; Rasmussen et al., 1990).
The tolerance and dependence exhibited by LC neurons during chronic opiate exposure occur in the absence of detectable changes in opiate receptors or opiate-regulated ion channels themselves (see Christie et al., 1987; Loh and Smith, 1990). This raises the possibility that intracellular messenger pathways may be involved. Indeed, over the past several years, it has been demonstrated that chronic administration of opiates leads to a dramatic upregulation of the cAMP system at every major step between receptor and physiological response (Fig. 1, bottom). Chronic opiate treatment increases levels of G<sb_>i<*_>alpha<*/><sb/> and G<sb_>o<*_>alpha<*/><sb/> (the active subunits of the G-proteins G<sb_>i<sb/> and G<sb_>o<sb/>) (Nestler et al., 1989), adenylate cyclase (Duman et al., 1988), cAMP-dependent protein kinase (Nestler and Tallman, 1988), and a number of MARPPs (morphine- and cAMP-regulated phosphoproteins) (Guitart and Nestler, 1989). Among these MARPPs is tyrosine hydroxylase (TH) (Guitart et al., 1990), the rate-limiting enzyme in the biosynthesis of catecholamines. These various intracellular adaptations to chronic opiate treatment are mediated via persistent activation of opiate receptors: the adaptations are blocked by concomitant treatment of rats with naltrexone, an opiate receptor antagonist, and are not produced by a single morphine injection.
Direct evidence for a functional role of an upregulated cAMP system in opiate addiction in the LC
The upregulated or 'hypertrophied' cAMP system in the LC can be viewed as a compensatory, homeostatic response of LC neurons to the inhibition devolving from chronic opiate treatment (Fig. 1). According to this view, opiate upregulation of the cAMP system increases the intrinsic excitability of LC neurons and thereby accounts, at least in part, for opiate tolerance, dependence, and withdrawal exhibited by these neurons (Nestler, 1990). In the opiate-tolerant/dependent state, the combined presence of the opiate and the upregulated cAMP system would return LC firing rates toward pretreatment levels, whereas removal of the opiates would leave the upregulated cAMP system unopposed, leading to withdrawal activation of the neurons. This model, which is similar to one proposed previously based on studies of cultured neuroblastoma x glioma cells (Sharma et al., 1975; Collier, 1980), is supported by several lines of evidence.
First, cAMP and agents that elevate cAMP levels excite LC neurons via the activation of cAMP-dependent protein kinase and subsequent activation of the nonspecific cation channel (Wang and Aghajanian, 1990). In fact, the spontaneous firing rate of LC neurons requires an active cAMP system and the opening of the nonspecific cation channel (Alreja and Aghajanian, 1991). Second, the time course by which certain components of the upregulated cAMP system revert to normal levels during naltrexone-precipitated opiate withdrawal parallels the rapid, early phase of the time course of recovery of LC neuronal firing rates and of various behavioral signs during withdrawal (Rasmussen at al., 1990). Third, upon bath application of naltrexone, LC neurons in brain slices obtained from morphine-dependent animals exhibit spontaneous firing rates more than twofold higher compared to LC neurons in slices from normal animals (Fig. 2A) (Kogan et al., 1992). Earlier studies failed to detect such withdrawal activation of morphine-dependent LC neurons in vitro, possibly due to the poor condition of the brain slices used and the small number and nonrandom samples of neurons examined (Andrade et al., 1983; Christie et al., 1987). Since most major afferents to the LC are severed in the brain slice preparation, the results establish that an increased intrinsic excitability caused by chronic opiate exposure contributes to opiate dependence in these cells. Fourth, LC neurons from morphine-dependent animals show a greater maximal responsiveness to cAMP analogs in vitro (Fig. 2B) (Kogan et al., 1992). Taken together, these results provide strong evidence to support the view that the opiate-induced upregulation of the cAMP system represents one mechanism by which opiates produce addictive changes in LC neurons.
Molecular mechanisms underlying opiate upregulation of the cAMP system in the LC
One of the central questions raised by these studies concerns the molecular mechanisms by which chronic opiate administration leads to upregulation of the cAMP system in LC neurons.
<#FROWN:J16\>
Spliced RNA of Woodchuck Hepatitis Virus
C. WALTER OGSTON AND DOLORES G. RAZMAN
Polymerase chain reaction was used to investigate RNA splicing in liver of woodchucks infected with woodchuck hepatitis virus (WHV). Two spliced species were detected, and the splice junctions were sequenced. The larger spliced RNA has an intron of 1300 nucleotides, and the smaller spliced sequence shows an additional downstream intron of 1104 nucleotides. We did not detect singly spliced sequences from which the smaller intron alone was removed. Control experiments showed that spliced sequences are present in both RNA and DNA in infected liver, showing that the viral reverse transcriptase can use spliced RNA as template. Spliced sequences were detected also in virion DNA prepared from serum. The upstream intron produces a reading frame that fuses the core to the polymerase polypeptide, while the downstream intron causes an inframe deletion in the polymerase open reading frame. Whereas the splicing patterns in WHV are superficially similar to those reported recently in hepatitis B virus, we detected no obvious homology in the coding capacity of spliced RNAs from these two viruses.
INTRODUCTION
Hepatitis B virus (HBV), the prototype of the hepadnavirus family, causes serious human disease and has a complex and interesting mode of replication. Hepadnaviruses make efficient use of a very small genome by use of overlapping reading frames, each of which may produce several alternative peptides (Ganem and Varmus, 1987), and their genome replication cycle involves reverse transcription of an RNA template (Summers and Mason, 1982. We would like to know how the peculiar features of hepadnavirus replication relate to their natural history and pathogenesis.
The mammalian hepadnaviruses woodchuck hepatitis virus (WHV) and ground squirrel hepatitis virus (GSHV), which are genetically close to HBV and similar in their pathogenesis, make the best experimental models for our purposes. We have chosen WHV as the object of our studies because infected woodchucks suffer from chronic hepatitis, and the majority of persistently infected woodchucks get hepatocellular carcinoma (HCC) that is histologically similar to liver cancer caused by HBV in humans (Popper et al., 1987; Korba et al., 1989).
The manner in which hepadnavirus messenger RNA codes for proteins remains a subject for research. Four genes have been identified in the DNA sequence of the mammalian hepadnaviruses as open reading frames (ORFs) diagrammed in Fig. 1. The core (C) ORF gives rise to the major nucleocapsid polypeptide as well as the secreted e antigen, and the surface (S) ORF gives rise to a total of six known glycoprotein species that are incorporated in the envelope of the virion and in related 'empty' 22-nm particles. The large polymerase (P) ORF is inferred from amino acid sequence homology to encode reverse transcriptase. The function of the X ORF in virus replication is not known, although it has been shown to be a transcriptional trans-activator in transfected cells.
The pattern of RNA transcription of the hepadnavirus genome has been reviewed by Ganem and Varmus (1987). Two abundant transcripts of 3.5 and 2.1 kb, respectively, can account for expression of the abundantly expressed C and S ORFs. A minor transcript of 2.4 kb encoding the pre-S1 ORF has been described in HBV-infected chimpanzee liver (Cattaneo et al., 1984; Will et al., 1987), but no corresponding RNA has been detected in WHV or GSHV (Ganem and Varmus, 1987).
The expression of the ORFs P and X remains to be accounted for, since no transcripts have been detected in vivo that include these ORFs at their 5' end. To solve this difficulty three hypotheses may be considered. First, these ORFs may be expressed from abundant mRNA by use of downstream AUG start codons (Kozak, 1986); second, there may be rare mRNAs that have 5' ends immediately upstream of these ORFs; and third, there may be rare spliced transcripts that express these ORFs.
The third hypothesis above is supported by the recent reports of spliced mRNAs in HBV-infected tissues and cells transfected with the HBV genome (Chen et al., 1989; Su et al., 1989; Suzuki et al., 1989, 1990; Wu et al., 1991). These workers have described two introns that encompass the region of overlap between the C and P genes of HBV. We have used polymerase chain reaction (PCR) (Kawasaki, 1990) to look for analogous splicing in WHV, and we now describe two species of spliced mRNA in WHV-infected liver, both of which could encode core-polymerase fusion proteins.
MATERIALS AND METHODS
The animals used in this study were naturally infected woodchucks (Ogston et al., 1989) and woodchucks infected as neonates (RW101, RW103, RW115, and RW116; C. W. Ogston and B. C. Tennant, unpublished). Virus from the serum of naturally infected woodchuck RW003 was used to initiate the infections of RW101 and RW103, and RW013 was the source of the inoculum for RW115 and RW116.
Liver tissue obtained at autopsy was cooled on ice, cut into slices 2-5-mm thick, and frozen in liquid nitrogen. RNA was prepared by homogenization in guanidine thiocyanate followed by cesium chloride centrifugation (Maniatis et al., 1982). Polyadenylated RNA was purified by oligo(dT) cellulose chromatography (Maniatis et al., 1982). The yield of RNA was measured by absorbance at 260 nm.
Complementary DNA was synthesized in a reaction containing 5 <*_>mu<*/>g of RNA, 100 mM Tris, pH 8.3, 7 mM MgCl<sb_>2<sb/>, 40 mM KCl, 1 mM dithiothreitol, and 0.5 <*_>mu<*/>g dT<sb_>12-18<sb/> primers (Collaborative Research), with 8 units of AMV reverse transcriptase (U.S. Biochemical). The reaction was incubated 2 hr at 42<*_>degree<*/>, then stopped by heating to 100<*_>degree<*/> for 3 min. The DNA was precipitated with ethanol, dried, and dissolved in 30 <*_>mu<*/>l of water.
The standard PCR reaction contained 100 mM Tris; pH 8.3 (at 25<*_>degree<*/>); 0.2 <*_>mu<*/>M each of dATP, dCTP, dGTP, and TTP (Pharmacia); primers at 0.08 <*_>mu<*/>M; and 0.6 units of Taq DAN polymerase (Perkin-Elmer Cetus). Amplification was carried out using a Perkin-Elmer Cetus DNA thermal cycler for 30 cycles with denaturation at 94<*_>degree<*/> for 60 sec, annealing at 50<*_>degree<*/> for 2 min, and extension at 72<*_>degree<*/> for 30 or 60 sec. Longer extension times were used in some reactions that were expected to produce amplification products greater than 1000 bp.
To generate single-stranded amplified DNA for sequencing, the amplified DNA fragment of interest was cut from an agarose gel and electroeluted. Asymmetric amplification was carried out using the same pair of primers as in the original amplification, but the concentration of one primer was reduced 32-fold (McCabe, 1990). The product of asymmetric PCR was purified by two rounds of centrifugal dialysis (Centricon 100, Amicon) to remove the primers and other components from the PCR reaction, then concentrated by centrifugal evaporation to a volume of about 7 <*_>mu<*/>l. The asymmetric DNA was sequenced directly using the Sequenase system (U.S. Biochemical) according to the manufacturer's directions. A 'nested' primer internal to the original PCR fragment was used for sequencing whenever possible.
<O_>figure&caption<O/>
The sequences were analyzed using the Beckman Microgenie software package. Sequence read from the gel was first compared to the published sequence of WHV (Galibert et al., 1982) by dot-matrix comparison. Areas of homology so identified were confirmed and then aligned.
In referring to primer sequences and experimental results, WHV map coordinates are given in nucleotides originating at the EcoRI site, as defined by the data of Galibert et al. (1982). The total length of this sequence is 3308 nt.
RESULTS
PCR can be used to detect spliced sequences in cDNA by employing primers located in different exons spanning one or more introns. Unspliced template then gives rise to an amplification product whose length corresopnds to the distance between the primers on the genomic DNA, whereas amplification of spliced template results in a shorter PCR product. Since short templates are amplified more efficiently than long ones this method is very sensitive even in the presence of excess unspliced template. We chose pairs of primers, shown in Fig. 1b, that we expected to span intron(s) analogous to those previously described in HBV.
<O_>figure&caption<O/>
When we amplified cDNA from WHV-infected liver using the primers 2110(+) and 575(-), we obtained a product of approximately 560 bp, as shown in Fig. 2a, as well as the full-length product of 1773 bp. This indicates that a segment of about 1200 nucleotides has been removed from the template. Likewise, the primers 2110(+) and 1478(-) gave a short product of 1260 bp as well as the full-length product of 2676 bp, indicating removal of about 1400 nucleotides. These two results presumably imply the existence of a single intron, since the inferred sizes of the missing segment agree within the limits of experimental error.
In the reaction using the primers 2110(+) and 1478(-), an additional product of 320 bp was amplified, indicating a template from which a total of 2300 nucleotides are missing from between the priming sites (Fig. 2a). This will be discussed further below.
To check the specificity of the amplification we carried out PCR using lower concentrations of nucleotide triphoshates. As shown in Figure 2a the short fragments described above were all amplified when the triphosphate concentration was reduced, while the full-length segments became less intense as expected. As a further control for specificity we amplified WHV DNA from the clone pWHV81-2 (Seeger et al., 1987) with both the above primer sets. No short products were obtained from this reaction, although the full-length sequence was amplified (data not shown).
To determine whether these observations are general among separate strains of WHV, we amplified cDNA from a number of experimentally and naturally infected woodchucks (Figure 2b and data not shown). The PCR products were the same size in all cases, although the relative intensity of the fragments varied. We think that this variation is due to complex interactions in the competition between more abundant unspliced template on the one hand, and rarer but shorter spliced template on the other hand. This competition will also be influenced by such factors as the degree of RNA degradation and the efficiency of reverse transcription.
To determine whether these results reflect authentic RNA splicing, the short PCR products were sequenced. We found in preliminary experiments that the best sequence data were obtained when a different 'nested' primer was used for sequencing rather than that used for amplification. This method also ensures that products derived from mispriming will not contribute to the sequence. Accordingly the primers 1961(+) and 314(-) were used to amplify a fragment of about 400 bp derived from the first spliced template described above, and this 400-bp fragment was purified by preparative electrophoresis. Single-stranded DNA was then made by asymmetric reamplification, using the primer 314(-) in excess over 1961(+) (McCabe, 1990), and the single-stranded DNA was sequenced using the 2110(+) primers, with the result shown in Fig. 3a.
The sequence obtained in this experiment was aligned to the published sequence of WHV (Galibert et al., 1982) as shown in Fig. 3b. This alignment is consistent with an intron of 1300 nucleotides between the donor site GT at bases 2151-2152 to the acceptor AG at 141-142. To confirm the sequence of this splice junction we generated positive-strand DNA template by amplification with 2110(+) in excess over 314(-) and sequenced the minus strand from the primer 314(-). The resulting complementary sequence showed the identical splice junction (data not shown).
<O_>figure&caption<O/>
The smaller (320 bp) product of amplification using the primers 2110(+) and 1478(-) (Fig. 2) could result from a doubly spliced mRNA with the 1300 nt intron and one of 1000 nt, or alternatively from a singly spliced mRNA with an intron of about 2300 nt. We first looked for the hypothetical second splice junction by amplifying cDNA with the primers 176(+) and 1478(-), but detected only the 1300-bp product derived from the unspliced sequence (data not shown). We thought that abundant cDNA derived from the 2.1-kb S-gene transcript might be overwhelming a rarer spliced template sequence, so we set out to separate the putative spliced sequence from unspliced S transcript by carrying out PCR in two steps. In the first step we used the primers 1961(+) and 1478(-) to selectively amplify spliced DNA, avoiding the 2.1-kb S transcript whose 5' end is at nt 145 (Moroy et al., 1985).
<#FROWN:J17\>
CONFUSION IN THE DETERMINATION OF DEATH: DISTINGUISHING PHILOSOPHY FROM PHYSIOLOGY
JEFFREY R. BOTKIN and STEPHEN G. POST
Two decades have now passed since the Report of the Ad Hoc Committee of the Harvard Medical School to Examine the Definition of Death. The adoption of whole brain criteria as a new standard of death proceeded quickly, supported by the American Medical Association, the American Bar Association, and the President's Commission, and has been adopted into legislation in 49 states. One of the final tasks toward the uniform adoption of the whole brain standard is the establishment of criteria for death in young infants. But despite this definitional consensus and its rapid adoption, beliefs about what it means to be dead may vary considerably in our highly pluralistic society.
Within the medical community alone there remains a startling amount of confusion about the determination of death. The redefinition of death has given rise to a distinct tension between the new definition of death based on brain criteria and common perceptions of the nature of death. Physicians, nurses, and others at the bedside may 'know' that a person is dead by established criteria, yet life still may be perceived as long as the heart continues to beat and the skin remains warm to touch. Many clinicians can rattle off criteria for obscure conditions or long differentials for anomalous lab results, yet they cannot offer specific criteria for death. More importantly, it is obvious that many health care professionals, including physicians, simply do not believe patients are dead when their brains alone have ceased to function.
Such confusion in the medical professions was clearly documented by Youngner, et al. This study documented that only 35 percent of physicians or nurses who dealt routinely with critically ill patients could correctly identify the legal and medical criteria for determining death. Further, 58 percent of the respondents did not apply a consistent concept of death when asked to justify their determination of death in hypothetical cases. The wider public is confused as well - a confusion no doubt fostered by news reports of 'brain dead' patients who are kept 'alive' by 'life-support' technology.
We commonly perceive the presence of death when the body is cold and pale, when breathing and heart cease. Little confusion about the presence of death arises in such circumstances. But contemporary technology has fostered confusion by forcing us to recognize the ambiguous nature of the moment of death. We can now restart functions such as breathing or heartbeat when once they would have been irreversibly lost. Certainly someone is not really dead with the cessation of heartbeat if the person can be revived. Yet despite common language about being 'brought back to life,' we like to think of death as a final and irreversible state, so we must now speak in terms of an 'irreversible cessation' of functions. However, what constitutes 'irreversible' will no doubt change with technology. This influence of technology on the moment of death indicates that the determination of this moment cannot be made independently of the cultural environment in which death occurs. Additionally, the contemporary ability to entirely replace cardio-respiratory functions with machines further blurs the boundaries between life and death. It was, of course, this latter technological development that was the primary force behind the adoption of the whole brain standard of death.
The basic argument we wish to make here is that the moment of death is not a specific physiologic event amenable to scientific determination. Rather, it is a moment defined by philosophic concepts - concepts that speak to what it means to be alive. Since such philosophic contentions defy objective proof, the moment of death must be seen as an event fixed by social consensus. Two implications of this realization will be discussed: first, that education for physicians and other medical professionals must address philosophic concepts if confusion about death is to be reduced; and second, that our pluralistic society must address the demand for alternative definitions of death for those who reject the philosophic foundations of the prevailing standard.
First we wish to look critically at the three competing standards of death. By 'standard' we mean a set of criteria for determining death that is based on an underlying philosophic concept. The three standards of death are the traditional heart/lung standard, the whole brain standard, and the higher brain standard. Current law recognizes the combination of the heart/lung or whole brain standards, but for the purposes of clarity, these will be discussed separately. In this discussion we will focus on our common perceptions of death and how these correlate with the philosophic concepts of death that underlie the three standards.
Advocates of the higher brain death standard argue that death occurs when cognitive function irreversibly ceases. Cognition is considered to be the essential element in personal identity, and thus the loss of cognition heralds the death of the individual. Under this standard, death would be an event occurring when the relevant brain activities cease. The philosophic appeal of this approach is that it accurately defines what most of us feel is of primary importance in human life - when all cognitive capacities are lost, we 'might as well be dead' figuratively, and, it is argued, literally. In addition, cognition is embodied in the cerebral hemispheres and so is potentially measurable by technical means (although it is not so at the present time).
From the perspective of many observers, the problem with the higher brain standard is the perception of continued 'life' despite the presence of death so defined. Individuals who have permanently lost all higher brain functions but retain brain stem functions will still breathe spontaneously, remain warm, and may have sleep/wake cycles and spontaneous movement. These patients simply do not look dead by our more intuitive understanding of the term. The conflict and confusion that arises here is due to a distinct and important difference between the death of the individual and the death of the organism. When cognition is irreversibly lost, but other physiologic processes continue, the human individual may be dead, but the human organism is not.
Which is the correct focus for our determination of death - death of the human individual or death of the human organism? There are reasoned arguments for each. One distinct advantage of determining death in terms of the organism is that such a definition could be applied across the spectrum of life. Since death is a term that is relevant to all life forms, it must therefore have some consistent biologic meaning. When we speak of death in other forms of life, we refer to the death of the organism - 'personhood' or cognition do not have any logical correlates in the majority of life forms. We may choose to define death in humans in terms of the loss of a single, key function (like cognition), but this will leave us with the task of defining such key functions for other life forms as well. Is there a single key function in fungi, barnacles, or bacteria on which our recognition of life hinges?
Even in the simplest of life forms, the definition of life has proven remarkably difficult and has been the subject of much debate. Orstan has provided a brief review of the definitions of life offered in Western thought and concludes: "It does not take much effort to realize that these definitions are little more than arbitrary lists of things the living organisms do". Thus, we would suggest that the common perception of death in an organism requires the cessation of a host of functions that comprise the perception of life. It is quite unlikely that any nonhuman life forms would be considered dead as long as there was spontaneous movement, respiration, and the flow of vital fluids.
So while the higher brain standard has the intellectual appeal of specifying the loss of a key function as the moment of transition between life and death, it will remain unconvincing for those whose intuitive or intellectual beliefs require that death signify the death of the human organism. Thus the higher brain standard is not 'wrong,' but if the popular perception of death is grounded in a concept of a living organism, the appeal of the higher brain standard may be limited as a basis for public policy.
The Whole Brain Standard
The current standard for the determination of death reads as follows:
An individual who has sustained either (1) irreversible cessation of circulatory and respiratory functions, or (2) irreversible cessation of all functions of the entire brain, including the brain stem, is dead.
The second half of the current standard represents the whole brain standard. The key concept for the whole brain standard is the loss of integration of the organism. Clearly some cellular and organ function continues for a period of time after brain or heart and lung functions cease; but the argument is that we do not care about isolated organ or cell function, we care about the integrated function of the organism. When the organism as a coordinated whole has irreversibly ceased to function, the argument goes, the organism is dead.
The brain is the logical locus for this integrative activity. Loss of whole brain activity leads to the loss of consciousness, respiratory arrest, no spontaneous movement, and no interaction with the environment. Without 'life support,' cardiac arrest soon develops and loss of all organ and cellular functions follow. Under this standard, the event of death is determined as the time at which there is irreversible loss of whole brain function.
The whole brain standard has several advantages; it can be measured with reliability, and irreversibility can be assured within reasonable limits. In addition, loss of integration is potentially applicable to other life forms (although it may be less clear how integration is determined in organisms without central nervous systems). Finally, loss of whole brain function produces an individual with many of the qualities we usually associate with death: lack of spontaneous breathing, movement, and response to the environment.
A principal difficulty with this standard is that individuals who lack whole brain function while on life support retain significant attributes of life: their skin may be warm and soft, they have a heartbeat, food continues to be metabolized, and waste products are excreted. Individuals without whole brain function may even gestate fetuses and bear children. It is precisely these attributes that lead many in the public and the medical profession to believe that such individuals are not 'really dead.'
The counterargument is that these life-like attributes of whole brain-dead patients do not count, since they do not reflect any integrated function in the organism, but exist only by technical intervention. This counterargument only makes sense if we can define what we mean by integration. According to Webster, integrate means: (1) "to make or become whole or complete," (2) "to bring (parts) together into a whole." Clearly an individual without brain function is not integrated; however, an individual without function in any major organ is not integrated. Barney Clarke was not intrinsically integrated once the artificial heart was implanted, yet he was obviously alive by virtue of technical support.
Youngner and Bartlett make this argument with a hypothetical case of a man who has had brain stem function surgically destroyed by a diabolical physician. The patient is then hooked to technical support that maintains bodily functions as well as cerebral functions: In essence, this hypothetical patient has been afforded a mechanical brain stem. Youngner and Bartlett note that the patient has "entirely lost the innate and spontaneous ability to integrate essential body systems," yet is obviously alive, at least as long as consciousness remains. It is irrelevant to our perception of life that brain stem functions are performed mechanically.
When an individual has lost all brain function yet is maintained on a ventilator, the heart, bowels, kidneys, and other organ systems may continue to function normally, at least temporarily, and these subsystems work in coordination. The individual without brain function is not wholly integrated, but neither is he or she wholly disintegrated.
<#FROWN:J18\>
We have thus shown that the Hilbert transform determines a bounded operator from L<sp_>p<sp/> into itself, for p of the form 2n. By the Riesz Interpolation Theorem, it follows then that the Hilbert transform determines a bounded operator from L<sp_>p<sp/> into itself, for <O_>formula<O/>. The proof can now be completed by appealing to Exercise 6.21 for the cases 1<p<2.
EXERCISE 6.24. (a) Show that any constant <O_>formula<O/> will satisfy inequality (6.4). Supposing that f is supported in the interval [-a,a], show that any constant <O_>formula<O/> will satisfy inequality (6.5) if <O_>formula<O/>.
(b) Establish Equation (6.6) by integrating around a large square contour.
(c) Let m be the characteristic function of an open interval (a,b), where <O_>formula<O/>. Prove that the multiplier M, corresponding to m determines a bounded operator on every L<sp_>p<sp/> for 1< p < <*_>infinity<*/>. HINT: Write m as a finite linear combination of translates of - isgn.
(d) Let m be the characteristic function of the set <O_>formula<O/>. Verify that the multiplier M corresponding to m has no bounded extension to any L<sp_>p<sp/> space for p <*_>unch<*/> 2.
CHAPTER VII
AXIOMS FOR A MATHEMATICAL MODEL OF EXPERIMENTAL SCIENCE
This chapter is a diversion from the main subject of this book, and it can be skipped without affecting the material that follows. However, we believe that the naive approach taken in this chapter toward the axiomatizing of experimental science serves as a good motivation for the mathematical theory developed in the following four chapters.
We describe here a set of axioms, first introduced by G.W. Mackey, to model experimental investigation of a system in nature. We suppose that we are studying a phenomenon in terms of various observations of it that we might make. We postulate that there exists a nonempty set S of what we shall call the possible states of the system, and we postulate that there is a nonempty set O of what we shall call the possible observables of the system. We give two examples.
(1) Suppose we are investigating a system that consists of a single physical particle in motion on an infinite straight line. Newtonian mechanics (f = ma) tells us that the system is completely determined for all future time by the current position and velocity, i.e., by two real numbers. Hence, the states of this system might well be identified with points in the plane. Two of the (many) possible observables of this system can be described as position and velocity observables. We imagine that there is a device which indicates where the particle is and another device that indicates its velocity. More realistically, we might have many yes/no devices that answer the observational questions: 'Is the particle between a and b?' 'Is the velocity of the particle between c and d?'
Quantum mechanical models of this single particle are different from the Newtonian one. They begin by assuming that the (pure) states of this one-particle system are identifiable with certain square-integrable functions and the observables are identified with certain linear transformations. This model seems quite mysterious to most mathematicians, and Mackey's axioms form one attempt at justifying it.
(2) Next, let us imagine that we are investigating a system in which three electrical circuits are in a black box and are open or closed according to some process of which we are not certain. The states of this system might well be described as all triples of 0's and 1's (0 for open and 1 for closed). Suppose that we have only the following four devices for observing this system. First, we can press a button b<sp_>0<sp/> and determine how many of the three circuits are closed. However, when we press this button, it has the effect of opening all three circuits, so that we have no hope of learning exactly which of the three were closed. (Making the observation actually affects the system.) In addition, we have three other buttons b<sb_>1<sb/>, b<sb_>2<sb/>, b<sb_>3<sb/>, b<sb_>i<sb/>, telling whether circuit i is open or closed. Again, when we press button b<sb_>i<sb/>, all three circuits are opened, so that we have no way of determining if any of the ciricuits other than the ith was closed. This is a simple example in which certain simultaneous observations appear to be impossible, e.g., determining whether circuits 1 and 2 are both closed.
The axioms we introduce are concerned with the concept of interpreting what it means to make a certain observation of the system when the system is in a given state. The result of such an observation should be a real number, with some probability, depending on the state and on the observable.
AXIOM 1. To each state <*_>alpha<*/> <*_>EPSILON<*/> S and observable A <*_>EPSILON<*/> O there corresponds a Borel probability measure <*_>mu<*/><sb_><*_>alpha<*/>, A<sb/> on R.
REMARK. The probability measure<*_>mu<*/><sb_><*_>alpha<*/>, A<sb/> contains the information about the probability that the observation A will result in a certain value, when the system is in the state <*_>alpha<*/>.
EXERCISE 7.1. Write out in words, from probability theory, what the following symbols mean.
(a) <*_>mu<*/><sb_><*_>alpha<*/>,A<sb/>([3,5]) = 0.9.
(b) <*_>mu<*/><sb_><*_>alpha<*/>,A<sb/>({0}) = 1.
AXIOM 2. (a) If A, B are observables for which <O_>formula<O/> for every state <*_>alpha<*/> <*_>EPSILON<*/> S, then A = B.
(b) If <*_>alpha<*/>, <*_>beta<*/> are states for which <O_>formula<O/> for every observable A <*_>EPSILON<*/> O, then <*_>alpha<*/> = <*_>beta<*/>.
EXERCISE 7.2. Discuss the intuitive legitimacy of Axiom 2.
AXIOM 3. If <*_>alpha<*/><sb_>1<sb/>, ... <*_>alpha<*/><sb_>n<sb/> are states, and t<sb_>1<sb/>, ... t<sb_>n<sb/> are nonnegative real numbers for which <O_>formula<O/>, then there exists a state <*_>alpha<*/> for which
<O_>formula<O/>
for every observable A. This axiom can be interpreted as asserting that the set S of states is closed under convex combinations. If the <*_>alpha<*/><sb_>i<sb/>'s are not all identical, we call this state <*_>alpha<*/> a mixed state and we write <O_>formula<O/>.
We say that a state <*_>alpha<*/> <*_>EPSILON<*/> S is a pure state if it is not a mixture of other states. That is, if <O_>formula<O/>, with each t<sb_>i<sb/> > 0 and <O_>formula<O/>, then <*_>alpha<*/><sb_>i<sb/> = <*_>alpha<*/> for all i.
EXERCISE 7.3. Discuss the intuitive legitimacy of Axiom 3. Think of a physical system, like a beaker of water, for which there are what we can interpret as pure states and mixed states.
AXIOM 4. If A is an observable, and <O_>formula<O/> is a Borel function, then there exists an observable B such that
<O_>formula<O/>
for every state <*_>alpha<*/> and every Borel set <O_>formula<O/>. We denote this observable B by f(A).
EXERCISE 7.4. Discuss the intuitive legitimacy of Axiom 4. Show that, when the system is in the state <*_>alpha<*/> and the observable A results in a value t with probability p, the observable B = f(A) results in the value f(t) with the same probability p.
EXERCISE 7.5. (a) Prove that there exists an observable A such that <O_>formula<O/> for every state <*_>alpha<*/>. That is, A is an observable that is nonnegative with probability 1 independent of the state of the system. HINT: Use <O_>formula<O/> for example.
(b) Given a real number t, show that there exists an observable A such that <O_>formula<O/> for every state <*_>alpha<*/>. That is, A is an observable that equals t with probability 1, independent of the state of the system.
(c) Show that the set of observables is closed under scalar multiplication. That is, if A is an observable and c is a nonzero real number, then there exists an observable B such that
<O_>formula<O/>
We may then write B = cA.
(d) If A and B are observables, does there have to be an observable C that we could think of as the sum A + B?
(e) In what way must we alter the descriptions of the systems in Example 1 and Example 2 in order to incorporate these first four axioms (particularly Axioms 3 and 4)?
DEFINITION. We say that two observables A and B are compatible, pairwise compatible, or simultaneously observable if there exists an observable C and Borel functions f and g such that A = f(C) and B = g(C). A sequence {A<sb_>i<sb/>} is called mutually compatible if there exists an observable C and Borel functions {f<sb_>i<sb/>} such that <O_>formula<O/> for all i.
EXERCISE 7.6. Is there a difference between a sequence {A<sb_>i<sb/>} of observables being pairwise compatible and being mutually compatible? In particular, is it possible that there could exist observables A, B, C, such that A and B are compatible, B and C are compatible, A and C are compatible, and yet A, B, C are not mutually compatible? HINT: Try to modify Example 2.
EXERCISE 7.7. (a) If A, B are observables, what should it mean to say that an observable C is the sum A + B of A and B? Discuss why we do not hypothesize that there always exists such an observable C.
(b) If A and B are compatible, can we prove that there exists an observable C that can be regarded as A + B?
DEFINITION. An observable q is called a question or a yes/no observable if, for each state <*_>alpha<*/>, the measure <*_>mu<*/><sb_><*_>alpha<*/>,q<sb/> is supported on the two numbers 0 and 1. We say that the result of observing q, when the system is in the state <*_>alpha<*/>, is 'yes' with probability <*_>mu<*/><sb_><*_>alpha<*/>,q<sb/>({1}), and it is 'no' with probability <*_>mu<*/><sb_><*_>alpha<*/>,q<sb/>({0}).
THEOREM 7.1. Let A be an observable.
(1) For each Borel subset E in R, the observable <*_>chi<*/><sb_>E<sb/>(A) is a question.
(2) If g is a real-valued Borel function on R, for which g(A) is a question, then there exists a Borel set E such that g(A) = <*_>chi<*/><sb_>E<sb/>(A).
(Note that condition 2 does not assert that g necessarily equals <*_>chi<*/><sb_>E<sb/>.)
PROOF. For each Borel set E, we have
<O_>formula<O/>,
and
<O_>formula<O/>,
which proves that <O_>formula<O/> is supported on the two points 0 and 1 for every <*_>alpha<*/>, whence <*_>chi<*/><sb_>E<sb/>(A) is a question and so part 1 is proved.
Given a g for which q = g(A) is a question, set E = g<sp_>-1<sp/>({1}), and observe that for any <*_>alpha<*/> <*_>EPSILON<*/> S we have
<O_>formula<O/>
Since both q and <*_>chi<*/><sb_>E<sb/>(A) are questions, it follows from the preceding paragraph that
<O_>formula<O/>
showing that
<O_>formula<O/>
for every state <*_>alpha<*/>. Then, by Axiom 2, we have that
<O_>formula<O/>
We now define some mathematical structure on the set Q of all questions. This set will form the fundamental ingredient of our model.
DEFINITION. Let Q denote the set of all questions. For each question q <*_>EPSILON<*/> Q, define a real-valued function m<sb_>q<sb/> on the set S of states by
<O_>formula<O/>
If p and q are questions, we say that <O_>formula<O/> if <O_>formula<O/> for all <*_>alpha<*/> <*_>EPSILON<*/> S.
If p, q and r are questions, for which <O_>formula<O/>, we say that p and q are summable and that r is the sum of p and q. We then write r = p + q. More generally, if {q<sb_>i<sb/>} is a countable (finite or infinite) set of questions, we say the q<sb_>i<sb/>'s are summable if there exists a question q such that
<O_>formula<O/>
for every <*_>alpha<*/> <*_>EPSILON<*/> S. In such a case, we write <O_>formula<O/>.
Finally, a countable set {q<sb_>i<sb/>} is called mutually summable if every subset of the q<sb_>i<sb/>'s is summable.
REMARK. As mentioned above, the set Q will turn out to be the fundamental ingredient of our model, in the sense that everything else will be described in terms of Q.
THEOREM 7.2.
(1) The set Q is a partially ordered set with respect to the ordering <*_>unch<*/> defined above.
(2) There exists a question q<sb_>1<sb/> <*_>EPSILON<*/> Q, which we shall often simply call 1, for which <O_>formula<O/> for every q <*_>EPSILON<*/> Q. That is, Q has a maximum element q<sb_>1<sb/>.
(3) There exists a question q<sb_>0<sb/> <*_>EPSILON<*/> Q, which we shall often simply call 0, for which <O_>formula<O/> for every q <*_>EPSILON<*/> Q. That is, Q has a minimum element q<sb_>0<sb/>.
(4) For each question q, there exists a question <*_>unch<*/> such that
<O_>formula<O/>
That is, every question has a complementary question.
PROOF. That Q is a partially ordered set is clear.
<#FROWN:J19\>
Consider the diagram
<O_>formula<O/>
where <*_>pi<*/> is orthogonal projection (with respect to the Hodge inner product), i is the isomorphism
<O_>formula<O/>,
and
<O_>formula<O/>.
Using the Hodge decomposition theorem, it is not hard to see that this diagram commutes, and that <*_>pi<*/> is an isomorphism for <*_>rho<*/> near 1. Hence, for such <*_>rho<*/>, det <*_>nu<*/>) can be identified with the determinant of D'<sb_><*_>rho<*/><sb/>.
The lagrangian <O_>formula<O/> gives rise to a smooth section of <O_>formula<O/> near <*_>rho<*/> = 1. For <O_>formula<O/>, this section coincides with <O/>formula<O/>. For <*_>rho<*/> = 1, it coincides with <O_>formula<O/>. Similar things are true if W<sb_>1<sb/> is replaced by W<sb_>2<sb/> or if 1 is replaced with another representation in P. Thus we have found the desired extensions of det<sp_>1<sp/>(<*_>nu<*/>) and det<sp_>1<sp/>(<*_>eta<*/><sb_>j<sb/>).
As a notational contrivance, let us define, for <O_>formula<O/>,
<O_>formula<O/>
If <O_>formula<O/>, define
<O_>formula<O/>.
So, abusing notation slightly, det<sp_>1<sp/>(<*_>nu<*/>) is a smooth bundle over all of <*_>unch<*/> and det<sp_>1<sp/>(<*_>eta<*/><sb_>j<sb/>) is a smooth section defined over all of <*_>unch<*/><sb_>j<sb/>. This will simplify notation in what follows.
(1.15) Proposition. The first Chern class c<sb_>1<sb/> (det(<*_>nu<*/>)) of det(<*_>nu<*/>) is represented by a multiple <*_>omega<*/> of <*_>omega<*/><sb_>B<sb/>. If <O_>formula<O/> is a symplectic 2-torus corresponding to a 2-dimensional symplectic summand of H<sb_>1<sb/>(F; R), then <O_>formula<O/>.
Proof: Let A be the space of connections on <O_>formula<O/>. The first step is to compute the curvature the universal S<sb_>0<sb/><sp_>1<sp/> bundle <O_>formula<O/> over A x F (cf. [AS]).
Identify A with <O_>formula<O/> via <O_>formula<O/>. Identify T<sb_>A<sb/>A with <O_>formula<O/>. Let <O_>formula<O/> and <O_>formula<O/>. Define the connection 1-form <*_>theta<*/> by
<O_>formula<O/>
The connection <*_>theta<*/> has the property that over the surface {A} x F it restricts to the connection A on <O_>formula<O/>. The curvature of <*_>theta<*/> is <O_>formula<O/>. Thus, for <O_>formula<O/>,
<O_>formula<O/>
Let G<sb_>0<sb/> be the group of smooth maps <O_>formula<O/> such that g(x<sb_>0<sb/>) = 1 for some fixed <O_>formula<O/>. G<sb_>0<sb/> acts freely on U. This action preserves the connection <*_>theta<*/>, and so leads to a connection <*_>theta<*/><sp_>b<sp/> on <O_>formula<O/>. Recall that <*_>unch<*/> can be identified with F/G<sb_>0<sb/>, where <O_>formula<O/> is the space of flat connection on <O_>formula<O/>. Thus <*_>theta<*/><sp_>b<sp/> restricts to a connection on <*_>unch<*/> x F.
The representation <O_>formula<O/> associates to <O_>formula<O/> bundle E over <*_>unch<*/> x F with curvature
<O_>formula<O/>
Hence <O_>formula<O/> and <O_>formula<O/> is represented by
<O_>formula<O/>
Thus the Chern character of E is represented by
<O_>formula<O/>
E can be thought of as a family of flat <O_>unches<O/> bundles over F parameterized by <*_>unch<*/>. Corresponding to each flat connection A is the operator D<sb_>A<sb/>, defined above, and det(<*_>nu<*/>) can be identified with the determinant of the index bundle of this family.
A flat connection A of <O_>formula<O/> determines a holomorphic structure A' on <O_>formula<O/>, and the index of the family D<sb_>A<sb/> can be identified with the index of the family <O_>formula<O/>. (F is given the complex structure compatible with its metric.) So, by Theorem 5.1 of [AS4] and the Riemann-Roch theorem, c<sb_>1<sb/>(det(<*_>nu<*/>)) is represented by the 2-dimensional part of
<O_>formula<O/>
where <O_>formula<O/> denotes integration along the fibers (i.e. along F) and T(F) is a form representing the Todd genus of F. Since <O_>formula<O/>, where <O_>formula<O/>, this is just
<O_>formula<O/>
Note that, for <O_>formula<O/> and <O_>formula<O/>,
<O_>formula<O/>.
Hence c<sb_>1<sb/>(det(<*_>nu<*/>)) is represented by the 2-form <*_>omega<*/>, where
<O_>formula<O/>
(Here we have identified <O_>formula<O/> with R via <O_>formula<O/>. Let <O_>formula<O/> be a symplectic 2-torus. Then
<O_>formula<O/>
D. Results of Newstead.
In this subsection we use results of Newstead ([N1,N2]) to show that certain diffeomorphisms of representation spaces are homologically trivial.
Let <O_>formula<O/> be a separating simple closed curve. Define
<O_>formula<O/>
Let F<sb_>1<sb/> and F<sb_>2<sb/> be the components of F cut along <*_>gamma<*/>. Define
<O_>formula<O/>
(j = 1,2). Then we can make the identifications
<O_>formula<O/>
(In the first equation, some basing of the free loop <*_>gamma<*/> is assumed.)
Let <O_>formula<O/> be a diffeomorphism such that <O_>formula<O/> and <O_>formula<O/> is the identity. <*_>tau<*/> induces diffeomorphisms of F (which is the identity on F<sb_>2<sb/>) and <O_>formula<O/>, which will also be denoted <*_>tau<*/>. In the proofs of (3.10) and (3.12), we will need the following lemma.
(1.16) Lemma. <O_>formula<O/> is the identity.
Proof: Note that since [<*_>gamma<*/>] lies in the commutator subgroup of <O_>formula<O/>, <O_>formula<O/> consists entirely of irreducible representations. Hence there is a fibration
<O_>formula<O/>
with fiber <O_>formula<O/>. Since SO(3) is a rational homology 3-sphere, there is a Gysin sequence
<O_>formula<O/>
(All coefficients are in Q. <O_>formula<O/> is the Euler class of the fibration.)
I claim that <O_>formula<O/> is an isomorphism for i = 2 or 3. This is obvious for i = 2, and injectivity is obvious for i = 3. By Proposition 2.6 of [N2]), the natural map
<O_>formula<O/>
is an isomorphism (j = 1,2). Furthermore, <O_>formula<O/> (Theorems 1 and 1' of [N2]). Therefore, by the Knneth formula, the composite
<O_>formula<O/>
is an isomorphism. (The first arrow is induced by the fibration <O_>formula<O/>.) Hence <O_>formula<O/> is onto.
By Theorem 1' of [N2] and the Knneth formula, <O_>formula<O/> is generated by classes of dimensions 2 and 3 for <O_>formula<O/>. I claim that <O_>formula<O/> is generated by classes of dimensions 2 and 3 and <O_>formula<O/>, for <O_>formula<O/>. This is clear for i<*_>unch<*/>3. Suppose it holds for <O_>formula<O/>. Let <O_>formula<O/>. Then there exists <O_>formula<O/> lying in the subring of <O_>formula<O/> generated by classes of dimensions 2 and 3 such that <O_>formula<O/>. Hence
<O_>formula<O/>
where, by inductive assumption, <O_>formula<O/> lies in the subring of <O_>formula<O/> generated by classes of dimensions 2 and 3 and <*_>chi<*/>.
So it suffices to show that <*_>tau<*/>* is the identity on <O_>formulae<O/> and <*_>chi<*/>.
Since <*_>tau<*/> acts on the fibration (1.17) in an orientation preserving fashion, <O_>formula<O/>.
Since <O_>formula<O/> is an isomorphism for i = 2 or 3, it suffices to show that <*_>tau<*/> acts trivially on <O_>formula<O/>. By the Knneth formula and the fact that <*_>tau<*/> act trivially on <O_>formula<O/>, it suffices to show that <*_>tau<*/> acts trivially on <O_>formula<O/> (i = 2,3). This is done in [AM] (Lemma VI.2.1 and VI.2.2).
Let x<sb_>1<sb/>, y<sb_>1<sb/>, ..., x<sb_>k<sb/>, y<sb_>k<sb/> be a symplectic basis of <O_>formula<O/>. This gives rise to identifications
<O_>formulae<O/>
Define <O_>formula<O/> by
<O_>formula<O/>
<O_>formula<O/> descends to a diffeomorphism of <O_>formula<O/>, which will also be denoted by <*_>sigma<*/>. In the proof of (3.59), we will need the following lemma.
(1.18) Lemma. <O_>formula<O/> is the identity.
Proof: The proof of (1.16) can be adapted to this case almost verbatim. It is left to the reader to check that the proofs of Lemmas VI.2.1 and VI.2.2 of [AM] work with <*_>tau<*/> replaced by <*_>sigma<*/>. (This amounts to observing that <O_>formula<O/> is the identity and that <*_>sigma<*/> acts on various sphere bundles in an orientation preserving fashion.)
E. Special Isotopies.
(1.19) Definition. An isotopy <O_>formula<O/> of R is called special if
1. <O_>formula<O/> is transverse to Q<sb_>2<sb/> at <O_>formula<O/> for all t.
2. <O_>formula<O/> for all t.
3. <O_>formula<O/> is symplectic, and hence (in view of 2 above) preserves the fibers of the normal bundle <*_>nu<*/>.
(1.20) Proposition. There exists a special isotopy <O_>formula<O/> of R such that <O_>formula<O/> is transverse to Q<sb_>2<sb/> (i.e. their Zariski tangent spaces are transverse at each point of <O_>formula<O/>).
Proof: Since M is a QHS, Q<sb_>1<sb/> is transverse to Q<sb_>2<sb/> at P and T<sb_>1<sb/> is transverse to T<sb_>2<sb/> in S. Choose a compactly supported isotopy of a tubular neighborhood of S<*_>unch<*/> in R which moves <O_>formula<O/> transverse to <O_>formula<O/> for each <O_>formula<O/> and is symplectic on <O_>formula<O/>. At this stage Q<sb_>1<sb/> is transverse to Q<sb_>2<sb/> in a neighborhood of S, and so we can find a compactly supported isotopy of R<*_>unch<*/> which moves Q<sb_>1<sb/> transverse to Q<sb_>2<sb/>.
F. Orientations.
Orientations will be important in what follows, so in this section we will establish orientation conventions. [Y] will denote the orientation of the space Y. If Y is singular, this means the orientation of the top stratum. If Y is a bundle, it means an orientation of the fibers (not the total space).
First of all, orient the Heegaard surface F so that [F] followed by a normal vector to F pointing into Q<sb_>2<sb/> gives [M]. This fixes an identification of <O_>formula<O/> with R, and so fixes the sign of <*_>omega<*/>.
Complex vector spaces (and almost complex manifolds) have a natural orientation: If a<sb_>1<sb/>, ..., a<sb_>n<sb/> is a basis over C, then a<sb_>1<sb/>, ia<sb_>1<sb/>, ..., a<sb_>n<sb/>, ia<sb_>n<sb/> is an oriented basis over R. Symplectic vector spaces and manifolds are oriented according to a compatible almost complex structure. Thus <*_>omega<*/> determines the orientations [R], [S],[<*_>nu<*/>] and [<*_>xi<*/>]. <*_>eta<*/><sb_>j<sb/> and <*_>unch<*/><sb_>j<sb/>, when lifted to <O_>formula<O/>, have J-complex structures. This determines [<*_>eta<*/><sb_>j<sb/>] and <*_>unch<*/><sb_>j<sb/> (as bundles over <O_>formula<O/>).
Choose orientations of <*_>unch<*/> and <*_>unch<*/> so that
<O_>formula<O/>
at points of <O_>formula<O/>. ([<*_>unch<*/>] is the orientation lifted from [S].)
In general, given a fibering <O_>formula<O/>, we choose orientations so that [E] = [B][Y]. Thus orientations on two of E, B or Y determine an orientation of the third. If G acts on X, there is a fibering <O_>formula<O/> (at least at points where G acts freely, which is enough to determine orientations). We regard spaces of unit vectors (e.g. <*_>unch<*/><sb_>j<sb/>) as quotients of spaces of non-unit vectors (e.g. <*_>theta<*/><sb_>j<sb/>) by R<sp_>+<sp/>, the positive reals. Let [SU(2)], [S<sb_>0<sb/><sp_>1<sp/>] and [R<sp_>+<sp/>] be the standard orientations. The following equations determine orientations of spaces not yet oriented. (There are some cases of over determination, and it is left to the reader to check that these cases are consistent.)
<O_>formulae<O/>
(The third equation requires some comment. <O_>formula<O/> determines an orientation of a double cover of a neighborhood of <O_>formula<O/> in Q<sb_>j<sb/>, which induces an orientation on Q<sb_>j<sb/>.)
All boundaries will be oriented according to the 'inward normal last' convention. (e.g. F is oriented as <*_>delta<*/>W<sb_>2<sb/>.) In particular, this fixes an orientation of <*_>delta<*/>F*. Hence the map
<O_>formula<O/>
is well-defined up to conjugation by elements of SU(2) (which preserves orientation). Note that <O_>formula<O/>. Thus, near regular points of R<sp_>#<sp/>, there is an identification of the normal fiber (in R*<sp_>#<sp/> with su(2), well-defined up to orientation preserving linear maps (i.e. Ad (SU(2))). Choose [R*<sp_>#<sp/>] so that
<O_>formula<O/>
([su(2)] is, of course, the standard orientation of su(2).)
Now we give an orientation convention for intersections of manifolds. Let A and B be oriented properly embedded submanifolds of an oriented manifold Y, intersecting transversely in X. Let <*_>alpha<*/> be the normal bundle of A in Y. Orient <*_>alpha<*/> so that
<O_>formula<O/>
<O_>formula<O/> is the normal bundle of X in B. Orient X so that
<O_>formula&figure<O/>
Note that this orientation convention is compatible with the convention given above for boundaries in the sense that
<O_>formula<O/>
Note also that the induced orientation of X depends on the ordering of A and B.
G. Dehn Twists and Dehn Surgery.
Let <O_>formula<O/> be a smooth function such that f(r) = 0 for r near 1 and f(r) = 2<*_>pi<*/> for r near 2. Let <O_>formula<O/>. Define <O_>formula<O/> by
<O_>formula<O/>
for <O_>formula<O/> and <O_>formula<O/> (see Figure 1.1). Note that g is the identity near <*_>delta<*/>A and the the isotopy class of g does not depend on the choice of f.
Let <*_>gamma<*/> be a simple closed curve in an oriented surface F. Let <O_>formula<O/> be orientation preserving embedding of A in F which maps the boundary components of A to curves which are isotopic to <*_>gamma<*/>. Define <O_>formula<O/> by
<O_>formula<O/>
The isotopy class of h<sb_><*_>gamma<*/><sb/> does not depend on the choice of <*_>phi<*/>. Any map in the isotopy class of h<sb_><*_>gammma<*/><sb/> is called a left-handed Dehn twist along <*_>gamma<*/>. The inverses of such maps are called right-handed Dehn twists.
Let N be a 3-manifold and let <O_>formula<O/> be a boundary component of N which is diffeomorphic to a torus. Let <O_>formula<O/> be a primitive homology class (that is, one which can be represented by a simple closed curve). Let <O_>formula<O/> be a diffeomorphism which sends <O_>formula<O/> to a curve representing a.
<O_>formula<O/>
is called the Dehn surgery of N along a. Note that N<sb_>a<sb/> does not depend on the sign of a. If N is oriented, we give N<sb_>a<sb/> the orientation induced from <O_>formula<O/>.
If <O_>formula<O/> is a knot in a 3-manifold M, then Dehn surgery on K means Dehn surgery on M\U, where U is an open tubular neighborhood of K.
<#FROWN:J20\>Let <*_>unch<*/> denote the neighborhood of <*_>unch<*/><sb_>g<sb/> parameterized by <O_>formula<O/> with <O_>formula<O/>. In this paper, we will be interested in the associated real analytic polar coordinates (Re <*_>unch<*/>, Im <*_>unch<*/>, |t|, arg t): the coordinate function arg t is S<sp_>1<sp/>-valued, but since we will soon lift to a Z-cover and consider the lift of arg t to be R-valued, here we will write arg t <*_>epsilon<*/> [0, 2<*_>pi<*/>).
Next, let <*_>unch<*/><sb_>g<sb/> denote the Deligne-Mumford compactified moduli space of stable curves and set <O_>formula<O/>. Then <*_>pi<*/> extends to a nonsurjective map <O_>formula<O/> which is a branched cover over its image. Moreover <*_>unch<*/><sb_>g<sb/> is a differentiable manifold ([BE], [Ea-Ma], [Mas]).
Now the neighborhood <*_>unch<*/> admits a Fenchel-Nielsen parameterization based on uniformizing the surfaces <O_>formula<O/>. On the surface M<sb_>t<sb/>, choose curves <*_>gamma<*/><sb_>1<sb/>, ..., <*_>gamma<*/><sb_>3g-4<sb/> so that M<sb_>t<sb/> <*_>approximate-sign<*/> {<*_>gamma<*/><sb_>0<sb/>, <*_>gamma<*/><sb_>1<sb/>, ..., <*_>gamma<*/><sb_>3g-4<sb/>} is a collection of 3-holed spheres. Let <*_>sigma<*/><sb_><*_>unch<*/>, t<sb/> denote the hyperbolic metric constistent with the conformal structure on the Riemann surface <O_>formula<O/>, and let <O_>formula<O/> denote the <*_>sigma<*/><sb_><*_>unch<*/>, t<sb/>-length of the unique <*_>sigma<*/><sb_><*_>unch<*/>, t<sb/> geodesic representing the free homotopy class [<*_>gamma<*/><sb_>i<sb/>] <*_>epsilon<*/> <*_>pi<*/><sb_>1<sb/>M<sb_>t<sb/>. Let <O_>formula<O/> denote the Fenchel-Nielsen twist angle (see [Ab] for details) of [<*_>gamma<*/><sb_>i<sb/>] of <O_>formula<O/>. Then <O_>formula<O/> provide Fenchel-Nielsen polar co-ordinates for <*_>unch<*/>; here we leave <O_>formula<O/> undefined. Also, <O_>formula<O/> should be thought of as S<sp_>1<sp/>-valued, but it will be notationally convenient to consider <O_>formula<O/>.
The local functions (Re <*_>unch<*/>, Im <*_>unch<*/>, |t|, arg t) and (<*_>unch<*/>, <*_>unch<*/>) each provide real analytic polar coordinates for <*_>unch<*/>, the former system from a conformal geometry point of view, and the latter from a hyperbolic geometry point of view. These two real analytic structures are not equivalent: in this paper, we aim to relate the two structures.
Before we state the main theorems, we need a technical definition of what we mean for a coordinate function to be real analytic in a polar coordinate system. We say that a function defined on a neighborhood of the origin in R<sp_>2<sp/> parametrized by the polar coordinate (r, <*_>theta<*/>) is sector real analytic provided that for any ray from the origin, there is a sector of the neighorhood at the origin containing that ray in which the function has a convergent expansion in the variables r and <*_>theta<*/> in that sector. The definition extends immediately to a product of planes. We prove
THEOREM A. The coordinate functions <O_>formula<O/>, are sector real analytic in (-log |t|)<sp_>-1<sp/>, arg t, and <*_>unch<*/> near t = 0; the coordinate functions <O_>formula<O/>, are also sector real analytic in (-log |t|)<sp_>-1<sp/>, arg t, and <*_>unch<*/>.
From our method, it follows
THEOREM B. The hyperbolic metrics <O_>formula<O/> are sector real analytic in (-log |t|)<sp_>-1<sp/>, arg t, and <*_>unch<*/> near t = 0.
We also conclude
THEOREM C. If (l<sb_>0<sb/>, <*_>unch<*/>, <*_>theta<*/><sb_>0<sb/>, <*_>unch<*/> are Fenchel-Nielsen coordinates for the neighborhood <*_>unch<*/>, then the coordinates |t|, <*_>unch<*/>, arg t are sector real analytic in <O_>formula<O/>, <*_>unch<*/>, <*_>theta<*/><sb_>0<sb/>, and <*_>unch<*/>.
Remarks. (1) Additional notation extends the results to surfaces with multiple nodes.
(2) We would like to place the current work in the context of [Wf] and [Wlpt].
In [Wf], we showed that the hyperbolic metrics <O_>formula<O/> were real analytic in <O_>formulae<O/>, and we provided a somewhat explicit Taylor series. There only the hyperbolic structure of the surfaces M<sb_><*_>unch<*/>, t<sb/> was considered and the proof relied on harmonic maps. We found that the harmonic maps were well suited to the study of the single real analytic polar structure (<*_>unch<*/>, <*_>unch<*/>): the maps were independent of the twist variable <O_>formula<O/>, and so any sector real analytic function of (<*_>unch<*/>, <*_>unch<*/>) was in fact real analytic in a full polar neighborhood. In the present situation, we are interested in comparing the conformal and hyperbolic geometry, and we find that we must use different maps between surfaces; the harmonic maps are unsuitable.
In [Wlpt], we provided the initial expansion for the hyperbolic metrics <O_>formula<O/> in (<*_>unch<*/>, t). The approach was based on the plumbing construction, approximate hyperbolic metrics, and solving the prescribed curvature equation. The method appears not to provide that the hyperbolic metrics are sector real analytic in (log |t|)<sp_>-1<sp/>, arg t, and <*_>unch<*/>. The motivation for the current work was to treat both the real analyticity and the complex parameterization.
2. Maps between surfaces. In this section, we define the map between surfaces that we will use, and we prove a lemma about its relevant properties.
2.1. We work on the augmented Teichmller space <*_>unch<*/><sb_>g<sb/> (see [Ab]). Briefly, the Teichmller space T<sb_>g<sb/> is an infinite cyclic cover of P<sb_>g<sb/> with covering group T(p) generated by isotopy classes of Dehn twist about <*_>gamma<*/><sb_>0<sb/>. Setting <*_>PI<*/> to be the covering map <O_>formula<O/>, consider <O_>formula<O/> where <O_>formula<O/> and let <O_>formula<O/> denote a point in T<sb_>g<sb/> lying over m(<*_>unch<*/>, t). Assign <*_>unch<*/> the coordinates <O_>unch<O/> and lift the coordinate functions on N near m(<*_>unch<*/>, t) to a neighborhood of <*_>unch<*/>. When we continue the functions <*_>unch<*/>, |t|, and arg t on T<sb_>g<sb/>, we see that <*_>unch<*/> and |t| are invariant under the action of the deck group T(p) while if <O_>formula<O/> with <O_>formula<O/>, then the arg t coordinates for <*_>unch<*/><sb_>1<sb/> and <*_>unch<*/><sb_>2<sb/> differ by an integer multiple of 2<*_>pi<*/>.
To obtain <*_>unch<*/><sb_>g<sb/> we adjoin to T<sb_>g<sb/> classes representing the noded surfaces M<sb_><*_>unch<*/>, 0<sb/> so that the coordinate functions <*_>unch<*/> and |t| are continuous. The projection <O_>formula<O/> extends to a map <O_>formula<O/> infinitely branched over the locus of points t = 0. Let <*_>unch<*/>* be the preimage of <*_>unch<*/> under <*_>unch<*/>; then <*_>unch<*/>* is sometimes called a horocyclic neighborhood in <*_>unch<*/><sb_>g<sb/>, and we have provided coordinates (<*_>unch<*/>, |t|, arg t) on <*_>unch<*/>*.
2.2. In this section, we prove
LEMMA 2.2. Fix (<*_>unch<*/><sb_>0<sb/>, |t<sb_>0<sb/>|, arg t<sb_>0<sb/>) with |t<sb_>0<sb/>| <*_>unch<*/> 0 and let <O_>formula<O/> represent a point in <*_>unch<*/><sb_>g<sb/> with coordinates (<*_>unch<*/>, |t|, arg t) near (<*_>unch<*/><sb_>0<sb/>, |t<sb_>0<sb/>|, arg t<sb_>0<sb/>). Then there exists maps <O_>formula<O/> so that <O_>formula<O/> is real analytic in <*_>unch<*/>, (-log |t|)<sp_>-1<sp/> and arg t and the radius of convergence of the power series expansion of <O_>formula<O/> about <*_>unch<*/><sb_>0<sb/>, |t<sb_>0<sb/>|, arg t<sb_>0<sb/> is independent of <*_>unch<*/><sb_>0<sb/>, t<sb_>0<sb/>.
Remark. Theorem B will follow from Lemma 2.2 by letting |t<sb_>0<sb/>| tend to zero. We are forced to take this approach rather than simply setting |t<sb_>0<sb/>| = 0 initially because we do not know the expansion of <O_>formula<O/> near the node p, and so we lack boundary conditions with which to simultaneously open up the node and uniformize.
In this approach, the size of the sector in which we are guaranteed analyticity in the theorem depends on the radius of convergence given by the lemma.
Proof. We organize our argument into four steps.
1. Constructing the maps. Important to our construction of the map <*_>zeta<*/> will be a decomposition of <O_>formula<O/> into three domains determined by its coordinates. Our original surface M<sb_>0<sb/> admitted a decomposition as
<O_>formula<O/>
where the Beltrami differentials <*_>mu<*/><sb_>i<sb/> were supported on M* and where we now further assume that M<sb_>0<sb/> <*_>approximate-sign<*/> M* is given as <O_>formula<O/>. Now define <O_>formula<O/>; here c is a small number to be determined later. The cylinder C<sb_>|t|<sb/> conformally embeds in <O_>formula<O/> and we will also use the notation C<sb_>|t|<sb/> for the embedded cylinder. Now recall that the map <O_>formula<O/> was conformal on M<sb_>0<sb/> <*_>approximate-sign<*/> M*, and define <O_>formula<O/>; we notice from our constructions that the conformal type of <O_>formula<O/> depends only on <*_>mu<*/>(<*_>unch<*/>), hence only on <*_>unch<*/>. We conclude that <O_>formula<O/> can be decomposed as
<O_>formula<O/>
Note that the middle component, <O_>formula<O/>, is a pair of cylinders whose geometry depends only on |t|; moreover the geometry of these cylinders is bounded independently of |t| for <O_>formula<O/>. We will define the map <*_>zeta<*/> on each domain and then patch the definitions together across the frontiers.
We begin by considering the cylinder C<sb_>|t<sb_>0<sb/>|<sb/>, but before we define the map <*_>zeta<*/> on C<sb_>|t<sb_>0<sb/>|<sb/>, we discuss some of the geometry of a cylinder C<sb_>|t|<sb/>. First, it will be more convenient for us to work with the conformally equivalent domain <O_>formula<O/> where we identify <O_>formula<O/> with <O_>formula<O/>. Second, we notice that A<sb_>|t|<sb/> admits the (noncomplete) hyperbolic metrics <O_>formula<O/>; in this metric the curve <O_>formula<O/> is a geodesic of length <O_>formula<O/>, and the curves <O_>formula<O/> and <O_>formula<O/> have constant geodesic curvature and are of length <O_>formula<O/>. It is often enough to focus on the subdomain <O_>formula<O/>; the domain <O_>formula<O/> is bounded by the geodesic <*_>gamma<*/><sb_>|t|<sb/> and is antiisometric to the interior of <O_>formula<O/>.
We now define the map <O_>formula<O/> by setting <O_>formula<O/> and then defining u(|t<sb_>0<sb/>|; |t|)(z) and v(|t<sb_>0<sb/>|; |t|)(z) in the coordinates on A<sb_>|t<sb_>o<sb/>|<sb/> by
<O_>formula&caption<O/>
Let <O_>formula<O/> denote the <O_>formula<O/> distance function. Then the map <O_>formula<O/> has the important property that
<O_>formula<O/>
Because u(|t<sb_>0<sb/>|; |t|)(x) is odd about <*_>gamma<*/><sb_>|t|<sb/>, we could also have defined <*_>zeta<*/> to take <O_>formula<O/> onto <O_>formula<O/> and the interior of <O_>formula<O/> onto the interior of <O_>formula<O/>; when we allow |t| to tend to zero, this will be a useful characterization of the map <*_>zeta<*/> between the domains <O_>formula<O/> and <O_>formula<O/>.
Later we will need to know <O_>formula<O/> and the Beltrami differential <O_>formula<O/>. These we compute to be
<O_>formula&caption<O/>
and
<O_>formula&caption<O/>
where <O_>formula<O/>, and <O_>formula<O/>.
It is perhaps easier to note that the first term in the brackets of (2.2) is <O_>formula<O/> and that
<O_>formula&caption<O/>
and
<O_>formula&caption<O/>
The following properties of (2.1), (2.2), and (2.3) will be crucial for our discussion:
(i)<O_>formula<O/>,
(ii)<O_>formula<O/> and <O_>formula<O/>,
(iii) for <O_>formula<O/> is bounded independently of |t<sb_>0<sb/>|, for |t<sb_>0<sb/>| small,
for |t| > 0, the map <O_>formula<O/> and for |t| = 0, the map <O_>formula<O/> on <O_>formula<O/> (also on the interior of <O_>formula<O/>, with the natural conventions) where here we define <O_>formula<O/> and most importantly of all,
(v) for fixed <O_>formula<O/>, the families <O_>formula<O/>, v(z), and Dv(z) are analytic in <O_>formula<O/> for 0<|t|<<*_>delta<*/> for some <*_>delta<*/> and can be analytically continued to a neighborhood of <O_>formula<O/>.
Next consider the components <O_>formula<O/>, which are conformally a pair of cylinders of bounded geometry; we will assume for convenience that <O_>formula<O/> does not depend on (<*_>unch<*/>, t) and that B<sb_>1<sb/> and B<sb_>2<sb/> are the two connected components. Divide each cylinder B<sb_>i<sb/> into thirds so that <O_>formula<O/>, so that B<sb_>i<sb/><sp_>1<sp/> share a frontier with C<sb_>|t|<sb/>, B<sb_>i<sb/><sp_>2<sp/> are the middle thirds, and B<sb_>i<sb/><sp_>3<sp/> share a frontier with <O_>formula<O/>. Suppose that B<sb_>i<sb/><sp_>2<sp/> is parameterized by <O_>formula<O/>; using the anticonformal equivalence of B<sb_>1<sb/> with B<sb_>2<sb/>, we may suppose that {|z| = <*_>beta<*/>} parameterizes the boundary between B<sb_>1<sb/><sp_>2<sp/> and B<sb_>1<sb/><sp_>1<sp/> and also the boundary between B<sb_>2<sb/><sp_>2<sp/> and B<sb_>2<sb/><sp_>3<sp/>. Then we define <O_>formula<O/> to be a pair of twist diffeomorphisms realizing the difference in arguments arg t - arg t<sb_>0<sb/>, i.e. in the coordinated <O_>formula<O/> we define
<O_>formula&caption<O/>
We will later define <*_>zeta<*/> on B<sb_>i<sb/><sp_>1<sp/> and B<sb_>i<sb/><sp_>3<sp/> to interpolate between the maps on C<sb_>|t|<sb/>, B<sb_>i<sb/><sp_>2<sp/>, and <O_>formula<O/>.
On the remaining subsurface <O_>formula<O/>, define <*_>zeta<*/> to be a quasiconformal map <O_>formula<O/> depending only upon <*_>unch<*/> and <*_>unch<*/><sb_>0<sb/> so that <O_>formula<O/> is the identity and <O_>formula<O/> is analytic in <*_>unch<*/>.
We are left to construct <*_>zeta<*/> on the cylindrical regions <O_>formula<O/>. Here we define <*_>zeta<*/> to interpolate between the maps defined on <O_>formula<O/> so that (i) <*_>zeta<*/> is analytic in (<*_>unch<*/>, |t|, arg t), with a real analytic continuation possible to a neighborhood of (<*_>unch<*/><sb_>0<sb/>, t = 0), (ii) <*_>zeta<*/>(<*_>unch<*/><sb_>0<sb/>, |t<sb_>0<sb/>|, arg t<sb_>0<sb/>; <*_>unch<*/><sb_>0<sb/>, |t<sb_>0<sb/>|, arg t<sb_>0<sb/>)(z) = z and (iii) <O_>formula<O/>. Such a family of maps exists because we have defined <*_>zeta<*/> on <O_>formula<O/> to have those properties and to also have derivatives which are uniformly bounded on <O_>formula<O/>. This concludes our construction of the map <O_>formula<O/>.
2. Constructing the grafted metrics. Our next goal is the construction of the hyperbolic metric <O_>formula<O/> on <O_>formula<O/> in such a way that we can analyze its dependence on <*_>unch<*/>, |t|, and arg t. Our approach is to use the method of Wolpert in [Wlpt]: we first construct a family of smooth metrics <O_>formula<O/> on <O_>formula<O/> which are hyperbolic on C<sb_>|t|<sb/> and <O_>formula<O/> but which are generally not hyperbolic on <O_>formula<O/>. We then consider the pullback metric <O_>formula<O/> which will be hyperbolic on <O_>formula<O/> but generally not hyperbolic in <O_>formula<O/>.
<#FROWN:J21\>
THE SELF-JOININGS OF RANK TWO MIXING TRANSFORMATIONS
DANIEL ULLMAN
(Communicated by R. Daniel Mauldin)
ABSTRACT. The class of rank one mixing transformations has been known for some time to have certain so-called 'exotic' properties. Specifically, any rank one mixing T has only the two trivial factors, and nothing can commute with T but powers of T itself. This much was known to Ornstein in 1969 [3]. In 1983 J. King showed that rank one mixing transformations possess an even stronger property known as minimal self-joinings (MSJ). In this note we investigate how these results can be extended to the case of rank two mixing transformations. In this class, it is possible for there to exist nontrivial factors and commuting transformations: the square of a rank one mixing transformation and certain two point extensions of a rank one mixing transformation are rank two mixing [2]. What we prove is that, other than those two kinds of rank two mixing transformations, this class also has MSJ.
Definitions. Assume throughout that T is a measure-preserving automorphism of a Lebesgue probability space (X, B, <*_>mu<*/>). By a tower for T of height h we mean a set B <*_>unch<*/> B together with its first h translates under T, provided that these translates are disjoint. B is called the base of the tower <O_>formula<O/>, and, for <O_>formula<O/>, T<sp_>i<sp/>B is called a level of the tower. (Notice that according to this terminology, a tower of height h has h +1 levels.)
The well-known Rohklin Lemma asserts that, given any measure-preserving transformation T and any <*_>epsilon<*/>>0, there is a tower M for T with <O_>formula<O/>.
All partitions are presumed to consist of a finite number of measurable sets. If P<sb_>j<sb/> is a sequence of partitions of X, we say P<sb_>j<sb/> generates the <*_>sigma<*/>-algebra B if B is the smallest <*_>sigma<*/>-algebra containing all the P<sb_>j<sb/>.
A measure-preserving transformation T is called rank n or less if there is an infinite sequence of sets of n disjoint towers <O_>formula<O/> such that the partitions P<sb_>j<sb/> of X given by
<O_>formula<O/>
generate the full <*_>sigma<*/>-algebra B. Given such a sequence, the towers with subscript j are called j-towers. Obviously, we call a transformation rank n (exactly) if it is rank n or less but not rank n - 1 or less.
It turns out that the class of rank n or less transformations can be characterized as the family of automorphisms isomorphic to a cutting and stacking transformation of the unit interval with n or fewer towers at each stage. (The n =1 version of this fact was first prove by Baxter [1]; the case of general n is not much harder.) I do not describe the details of such a construction nor refer to cutting and stacking in the sequel. Suffice it to say that this result allows us to assume that the partitions P<sb_>j<sb/> in (1) form a nested sequence, each refining the previous.
A TxT-invariant measure <*_>unch<*/> on the Cartesian product space XxX, which projects to <*_>mu<*/> in both directions, is called a self-joining of T. Any T has certain obvious ergodic self-joinings; namely, product measure <*_>mu<*/>x<*_>mu<*/> and the so-called off-diagonal measures - measures defined on measurable rectangles by <O_>formula<O/> for some integer n. If these are the only ergodic self-joinings, then T is said to have minimal self-joinings (MSJ). This definition is due to D. Rudolph [4].
The commutant of T is the group of measure-preserving transformations that commute with T modulo the (necessarily normal) subgroup of integral powers of T. A factor <*_>sigma<*/>-algebra is a T-invariant sub-<*_>sigma<*/>-algebra of B. B itself and {<*_>phi<*/>, X} are factors of any T and are called trivial. It is not hard to see that MSJ implies that T has trivial commutant and factors. For if S commutes with T and is not a power of T, <O_>formula<O/> defines a new ergodic self-joining, and if F is a nontrivial T-invariant (factor) <*_>sigma<*/>-algebra, then <O_>formula<O/> defines a self-joining not all of whose ergodic components can be product measure or off-diagonal measures.
A map T is mixing (or satisfies the mixing property) if, for all A and B <*_>unch<*/> B, <O_>formula<O/> as n goes to infinity.
THE CLASSIFICATION THEOREM
Theorem. If T is a rank two mixing transformation, then either
(1) T is the square of a rank one mixing transformation,
(2) T is a two point extension of a rank one mixing transformation, or
(3) T has MSJ.
The proof proceeds through a series of lemmas. Throughout, T acts on a space (X, B, <*_>mu<*/>). For each natural number j, X is partitioned into two towers M<sb_>j<sb/>(1) and M<sb_>j<sb/>(2), called j-towers, and a set <O_>formula<O/>. M<sb_>j<sb/> and M<sb_>j<sb/>(2) have bases B<sb_>j<sb/>(1) and B<sb_>j<sb/>(2) and heights h<sb_>j<sb/>(1) and h<sb_>j<sb/>(2), respectively. Levels of j-towers are naturally called j-levels. The partitions P<sb_>j<sb/> of X whose elements are j-levels successively refine each other and generate B. We assume that the real numbers <*_>mu<*/>(M<sb_>j<sb/>(1)) and <*_>mu<*/>(M<sb_>j<sb/>(2)), for j = 1, 2, 3,..., are bounded away from zero, by <*_>gamma<*/>, say. (Otherwise, T is rank one.) Finally, we assume that the base of the j-towers are a union of at least twenty (j+1)-levels; this can be arranged by passing to a subsequence of towers j<sb_>k<sb/>.
We begin with a lemma that guarantees that almost all points in X lie in the middle 98% of their j-tower for at least 3/4 of all j. For 0<<*_>alpha<*/><1/2 and i =1 or 2, let
<O_>formula<O/>
and set
<O_>formula<O/>.
This is the set of points on one of the j-towers that are not too close to the top or bottom.
Lemma 1. Let
<O_>formula<O/>.
Then <*_>mu<*/>S(1/100) =1.
Proof. First notice that x<*_>unch<*/>S(1/99) implies that <O_>formula<O/> for all integers n. Consequently, <O_>formula<O/>, so
<O_>formula<O/>.
To show that <*_>mu<*/>S(1/100) =1, it is enough to show that <*_>mu<*/>s(1/99) >0, since then the left-hand side of (2) is an invariant set of positive measure.
The idea is that the sets D<sb_>j<sb/>(1/99), for j = 1, 2, 3,..., are close to being independent sets. We modify them slightly (to produce independent sets) and then use the law of large numbers.
Find four natural numbers a(1), a(2), b(1), and b(2) such that, if
<O_>formula<O/>,
then
<O_>formula<O/>,
<O_>formula<O/>,
and
<O_>formula<O/>.
Condition (5) says simply that E<sb_>j<sb/> must be a union of j-levels the first of which is a subset of the base of a (j - 1)-tower and the last of which is a subset of the top level of a (j - 1)-tower. That a(1), a(2), b(1), and b(2) can be chosen satisfying (3) and (4) requires that the j-towers be many times higher than the (j - 1)-towers, which is equivalent to the assumption made at the end of the first paragraph after the statement of the theorem.
Let <*_>nu<*/> be the measure <*_>mu<*/> restricted to <O_>formula<O/> and normalized; that is,
<O_>formula<O/>
for <O_>formula<O/>. Then the sets E<sb_>j<sb/> are independent with respect to nu, since no data of the form <O_>formula<O/> or <O_>formula<O/> for k>j give any information about which j-level the point x is on.
Hence the strong law of large numbers tells us that, for <*_>nu<*/>-almost every x,
<O_>formula<O/>.
But <O_>formula<O/> and so <*_>nu<*/>S(1/99) =1. This implies that <*_>nu<*/>S(1/99) >0 and so <*_>nu<*/>S(1/100) =1 as required.<*_>square<*/>
Hereafter, we assume that the property that Lemma 1 ascribes to almost all points in fact holds for all points. We may do this, since the property is invariant.
For any pair <O_>formula<O/>, we set r<sb_>j<sb/> equal to the greatest nonpositive integer i such that either T<sp_>i<sp/>x or T<sp_>i<sp/>y is in <O_>formula<O/>. Similarly, let s<sb_>j<sb/> be the smallest nonnegative integer i such that either T<sp_>i<sp/>x or T<sp_>i<sp/>y is in the top level of a j-tower, that is, in the set <O_>formula<O/>. We say that the interval of integers between r<sb_>j<sb/> and s<sb_>j<sb/> is a frame (the j-frame) for (x, y). Note that, when j is sufficiently large, both x and y are on j-towers, and for such j, the j-frame is shorter than both h<sb_>j<sb/>(1) and h<sb_>j<sb/>(2).
Lemma 2. Suppose <*_>unch<*/> is a self-joining of T, (x, y)<*_>unch<*/>XxX, and x and y are not on the same orbit of T. Suppose also that there is a subsequence J<*_>unch<*/>N of natural numbers j such that (1) x and y lie in the middle 98% of the same j-tower, and (2) for all sets A and B in X that are unions of levels of some tower,
<O_>formula<O/>
as j goes to infinity along the subsequence J. Then <O_>formula<O/>, product measure.
Comment. Condition (2) of the lemma is certainly satisfied if (x, y) is generic for <O_>unch<O/>, which means that for any sets A and B made out of levels of some tower,
<O_>formula<O/>.
Proof. For j<*_>unch<*/>J, let F<sb_>j<sb/> be the phase shift between the j-block in which x sits and the j-block in which y sits. By this I mean that if n(x) and n(y) are the smallest nonnegative integers such that <O_>formula<O/> and <O_>formula<O/>, then <O_>formula<O/>. Since x and y are assumed to be on different orbits of T, <O_>formula<O/>, as j goes to infinity along J.
Let A and B be subsets of X made up of j<sb_>0<sb/>-levels. Fix <*_>epsilon<*/> >0. Because T is mixing, there is an integer N so that
<O_>formula<O/>.
Pick j<*_>unch<*/>J so large that
<O_>formulae<O/>
and
<O_>formula<O/>.
By switching the roles of x and y if necessary, we may assume that <O_>formula<O/>, so that Figure 1 is qualitatively correct.
<O_>figure&caption<O/>
We now calculate:
<O_>formula<O/>.
The first equality here comes from (9) and the second holds because <O_>formula<O/> if and only if <O_>formula<O/> and <O_>formula<O/>, when <O_>formula<O/>. If <O_>formula<O/>, let
<O_>formula<O/>,
the set of points on the same j-tower as x, at a height between 0 and s<sb_>j<sb/>-r<sb_>j<sb/>. Now (7) implies that <O_>formula<O/> is made entirely of complete j-levels, hence expression (10) above implies
<O_>formula<O/>.
Since C is at least one hundrethhundredth of M<sb_>j<sb/>(i) (because of condition (1) of the lemma) we see that <*_>mu<*/>C ><*_>gamma<*/>/100. (Remember that <O_>formula<O/>.) Thus,
<O_>formula<O/>,
which, owing to (6) and (8), is <O_>formula<O/>. The outcome of this calculation is that
<O_>formula<O/>,
or
<O_>formula<O/>,
where L is a constant independent of A and B. Since this inequality holds for all A and B made up of j<sb_>0<sb/>-levels, and since j<sb_>0<sb/> is arbitrary,
<O_>formula<O/>.
Finally, since <*_>unch<*/> is TxT-invariant (and normalized), we conclude that <O_>formula<O/> and the lemma is proved.<*_>square<*/>
Lemma 3. Suppose <*_>unch<*/> is an ergodic self-joining of T that is not product or off-diagonal measure. Then for all <*_>epsilon<*/> > 0, there is a natural number N so that j < N implies
<O_>formula<O/>.
Proof. Suppose that there are an <*_>epsilon<*/> > 0 and a sequence of sets <O_>formula<O/> with <O_>formula<O/>, for infinitely many j. Then <O_>formula<O/>, so lim inf<sb_>j<sb/>A<sb_>j<sb/> has a generic point (x, y) for <*_>unch<*/>. If x and y were on the same orbit, then <*_>unch<*/> would be an off-diagonal measure. If x and y were not on the same orbit, then we could infer from <O_>formula<O/> that condition (1) of lemma (2) is satisfied. The genericity of (x, y) gives us condition (2) of that lemma, which would imply that <*_>unch<*/> is product measure. The lemma is proved.<*_>square<*/>
Lemma 4. Suppose that <*_>unch<*/> is an ergodic joining of T that is not product measure. Assume that both (x, y<sb_>1<sb/>) and (x, y<sb_>2<sb/>) are generic for <*_>unch<*/> and that x and y<sb_>i<sb/> are on opposite j-towers for all sufficiently large j, for i = 1 and 2. Then y<sb_>1<sb/> = y<sb_>2<sb/>.
Proof. y<sb_>1<sb/> and y<sb_>2<sb/> are on the same j-tower for all sufficiently large j. Let r<sb_>j<sb/> and s<sb_>j<sb/> be the beginning and the end of the j-frame for (y<sb_>1<sb/>, y<sb_>2<sb/>).
Skip those j for which either y<sb_>1<sb/> or y<sb_>2<sb/> is not in D<sp_>j<sp/>(1/100). Lemma 1 assures us that we are left with infinitely many j. Along the subsequence that remains, we have <O_>formula<O/>.
<O_>figure&caption<O/>
There is a natural topology on X associated with the towers M<sb_>j<sb/>(i), and that is the one generated by levels of towers. That is, a set is open (and closed) if it is a countable union of elements of the partitions P<sb_>j<sb/>.
<#FROWN:J22\>Women between the ages of fifteen and fifty-nine now normally have a job or are looking for one. In Denmark, 75.9 percent of adult women were employed in 1987. Spain had the lowest participation rate in the EC with 37.5 percent of adult women at work (Jackson 1990: 5). Women most commonly work in the service sector. Indeed, 76.4 percent of them find their jobs in that sector (CEC 1989p: 124). Women are also most likely to work in feminized work places. They are bank tellers, nurses, teachers, and cleaners in hotels. They are unlikely to be bank managers, executives in private business, well-paid technicians, or hotel managers. Women who work in industry tend to be concentrated in a few sectors which are less well paid and more labor intensive, such as the clothing and textile industry. The situation in Germany illustrates the fact of segregation. Ninety percent of working women find their jobs in only twelve occupational categories (CEC 1989k: 87). Women also fill a disproportionate number of part-time jobs; for example, they provide 90 percent of the part-time work force in Germany (CEC 1989v: 72). According to the latest report, 28.6 percent of women's employment is part-time (CEC 1989p: 140). Working women in the EC still have not achieved equality in pay or opportunity despite laws requiring equality. Women's pay is probably 31 percent less than men's (Jackson 1990: 50). However, pay scales differ greatly among the member states. What women want and what society expects for women also varies among the member states. For example, a French woman employed in a bank has a much greater chance of reaching a managerial position than does her British counterpart. Also the French woman is much more likely to find a place in a good public nursery for her child than her British counterpart, and both the British and the French woman would expect more social acceptance for their careers than would a female bank employee in Portugal.
NATIONAL POLICIES FOR WOMEN IN THE WORK FORCE
The governments of all of the member states of the EC have laws to protect and assist working women. In addition, the governments have created high-level agencies or even ministries for women's issues. The old paternalistic laws, such as those banning women from night work, have gradually given way to more modern laws on equal treatment. Discrimination is illegal in all countries, but the definition of discrimination varies considerably, as does the quality of enforcement (Landau 1985). A 1983 French law requires employers to make an annual report about their personnel policies for women. The law provides sanctions for transgressions as well as protections for an employee making charges against an employer (France 1984: 487). In general, most governments have been more effective in banning overt discrimination than in devising policies for affirmative action.
Working mothers have more assistance from the law in EC countries than they have in the United States. Every country has a law which provides for maternity leave. Italian women are entitled to twenty weeks of paid leave but British women to only six. In all countries, employers may not fire a regular employee because of pregnancy or refuse to allow a woman to return to work following maternity leave. Parental leave is also beginning to appear in some countries. Most countries have inadequate public provisions for child care, but the French government provides a good system of public child care centers. That system may be a factor in explaining why French women are more likely to remain in the work force during childbearing years than are women in most other EC countries (OECD 1985: 34).
New public policies for women are being devised in response to unemployment and to changes in the work place. Schools are encouraging girls to consider a variety of careers. Training programs are being reformed so that women may participate more easily. A great deal of research is being conducted in order to ascertain what is needed in order to better use women in the work force. The efforts are scattered and sporadic, but governments are increasingly accepting responsibility to assure that women have the preparation needed for modern job opportunities.
THE DEVELOPMENT OF EC POLICIES FOR WOMEN
The right of the EC to formulate a policy for working women derives from the Treaty of Rome. Article 119 establishes the principle of equal pay for men and women. The preamble to the treaty as well as Articles 117 to 122 give the EC a general grant of power for social policy.
The 1970s was the decade when the EC began to address women's issues. It was a period when both the Council of Ministers and the Commission had leaders who were sympathetic to the social concerns of the day. Equality, worker's rights, and social justice were values which found their way onto the political agendas of the countries of Western Europe and onto the agenda of the EC. The Social Action Program of 1974 was the result. The program promised action "to achieve equality between men and women as regards access to employment and vocational training and advancement and as regards working conditions including pay" (CEC 1974b). The statement constituted the first elaboration of the meaning of Article 119 and laid the foundation for the EC to act over a broad range of job rights for women.
The EC quickly started to fulfill its commitment by enacting three directives on equal rights at work. They were the Equal Pay Directive of 1975, the Equal Treatment Directive of 1976, and the Social Security Directive of 1978. The meaning of each directive has been broadened by subsequent rulings of the European Court, but most member states have been remiss in enforcing them. Today the definition of equal pay in the EC means equal pay for work of equal value. Equal treatment now makes illegal all forms of sex discrimination at work including hiring, training, and promotion. Most importantly, it protects women against both direct and indirect discrimination. The directive on social security applies to both national social security systems and special occupational and supplementary schemes. It does not require uniformity among the national programs, and it allows those programs to contain special benefits for women. It protects women against provisions which are discriminatory even when the discrimination is indirect (CEC 1983).
The organization of the Commission was changed in the 1970s in response to the new interest in women's issues. The units added continue to be responsible for EC policies for women. The equal opportunities unit in the Directorate General for Employment, Social Affairs, and Education (DG V) carries on the bulk of the work. A handful of civil servants are responsible for the information gathering, analysis, and consulting necessary for preparing and overseeing policies for working women. DG V is under the direction of the commissioner who holds the portfolio for social affairs. (In 1991, the commissioner responsible for social affairs was Vasso Papandreou. She was the first woman in that position and one of the only two women ever to be a commissioner.) A women's information service operates in the Directorate General for Information, Communication and Culture (DG X). It is responsible for disseminating information about women and publishes a series called Women of Europe. The European Parliament and the Economic and Social Committee have special committees to deal with women's issues. Both institutions have advocated a strong EC policy for women. The Court of Justice has also played a role in developing the EC policy for women through a liberal interpretation of EC law.
The scope and ambition of the EC policy for working women developed in the 1970s is quite remarkable. The policy contains both legal measures to ban discrimination and nonlegal measures to facilitate the social and psychological changes necessary for true equality. Traditional family values intermingle with newer feminist concerns in a broad range of initiatives. For example, the Commission sponsored seminars to encourage bankers to be less sexist in their personnel policies. The Commission studied vocational education in order to ascertain why women remain in feminized work places. The Commission delved into the question of the relationship between family responsibilities and success in the work place. The EC used the Social Fund to provide training for women to enter jobs formerly inaccessible to them. During the formative period of the 1970s, the EC chose an activist approach which surpassed merely formulating measures essential to harmonize national policies which might have inhibited competition in the internal market.
In 1981, the disparate EC activities for women were brought together in the first action program to promote equal opportunities for women. The opening sentence of the document states, "The Community's longstanding commitment to the improvement of the situation of women has established it as a pioneer and innovator in this field" (CEC 1981). The rest of the document does not match the bold opening sentence. Discussion in the document is brief and focuses primarily on the problems working women were facing because of the recession. Member governments were given the primary responsibility of alleviating the problems.
The annex of the document contains sixteen proposals for legal and nonlegal measures to promote equality. In almost every case, the responsibility for action is divided between the EC and the member state. Frequently, the role of the EC is only to study the situation and then consider action; however, six of the proposals fit into the activist mold of the 1970s. They are:
1. An EC law on equal treatment for women in occupational, social security schemes.
2. An EC law on equal treatment for self-employed women and women in agriculture.
3. An EC law on parental leave and leave for family reasons, and on the building of public services and facilities to assist working parents.
4. Possible legislation on pregnancy and motherhood if the Commission considers it necessary.
5. Future legislation on steps needed for action to assist women in achieving equal opportunity.
6. Extension of EC action on vocational education so women can participate in new technological sectors through the Social Fund and the center for vocational training in Berlin.
The dates for the action program, 1982-1985, coincided with the period when the integration appeared stalled and economic problems took precedence. Only the first two proposals listed above became law according to the schedule given in the program (CEC 1984b); however, the other proposals remained on the agenda of the Commission.
A second action program appeared in 1985, when Europessimism was strongest. It was also the year when the EC was deeply involved with two historic documents: the White Paper on Completing the Internal Market and the Single European Act. Seen against that time, the second action program is quite remarkable. Although it contains no major new programs, it is a thoughtful and interesting document. It shows the influence of research conducted in the Commission over the past decade. Emphasis is given to the psychological dimension of discrimination. The writers of the document doubted the efficacy of laws to end discrimination. Ways to change attitudes, and not just attitudes in the work place, were needed. The basis of discrimination is in society and in the family. The sharing of family responsibility is listed as the sine qua non for true equality (CEC 1986b: 5). Many of the proposals in the document reflect this orientation, such as the proposal for a campaign to increase public awareness. Other proposals were a reiteration of some in the first action program, which still await acceptance. Four proposals were to become focal points of controversy. They are:
1. A legal instrument to facilitate action by women against employers who discriminate. The instrument was to be based on the principle of the reversal of the burden of proof.
2. A code of practice on positive actions which should guide employers and member states in order to facilitate providing equal opportunity.
3. A measure to protect working women during pregnancy and motherhood.
4. A directive on parental leave and leave for family reasons.
<#FROWN:J23\>If any of these changes are substantial, they will affect the relative values of the different ways of institutionalizing property rights. If the result of these changes is that a new property rights scheme can produce greater aggregate benefits than the existing system does, the contracting parties will consider the following cost-benefit calculation: Do these additional benefits exceed the costs of changing the present contract? If the answer is yes, the actors will enter into a new contract institutionalizing new rights (North, 1990: 67). Here we see the third important element of the transaction-costs theory: Institutional change will occur only if the resulting outcome is Pareto superior to the previous institutional arrangement.
Note that at this point we are discussing only the motivations of the actors in the individual exchange. The fact that low-cost rules may be selected by marketlike competition has not yet entered the analysis. North's reliance on transaction-costs minimization reflects the approach of the new institutional economics. In analyzing the contractual arrangements of social organization, the theory proceeds from the "working rule that low-cost organizations tend to supersede high-cost ones" (Eggertsson, 1990: 213-14, emphasis in original). When exceptions to this rule are observed, three factors should be considered.
First, there may be hidden benefits that are not readily apparent. This reduces to a claim that the apparent exception to the working rule is in fact an instance of the more general rule of maximizing benefits. Here a higher-cost institutional rule in fact produces a more beneficial exchange than does a less costly one. Second, the efforts of social actors to create low-cost institutional forms may be constrained by the interests of the state. This latter issue is closely related to issues of intentional design and reform, but it is proposed as an explanation of why the voluntary arrangements being analyzed here can be restricted by some type of formal external constraint. The point relevant to this analysis is the implicit claim that if it were not for external constraints, rational social actors would voluntarily create the least costly forms of organization. This is consistent with the third exception to the minimization-of-costs rule: uncertainty. According to this account, actors may not create the least costly rules because they lack either the capacity or the knowledge to establish them.
This minimization-cost standard is a problematic one. As the arguments in Chapter 2 demonstrate, the claim that the minimization of aggregate transaction costs is the motivation for institutional development and change is inconsistent with individual rationality. The assumption of narrow rationality on which the transaction-costs approach is based implies that strategic actors will prefer more costly rules if these rules give them greater individual benefits than those from less costly ones. Remember that in this case the explanation is not inability or incapacity but, rather, self-interest. The transaction-costs approach does not predict costly rules produced by fully informed, self-interested behavior. Here we return to some of the same explanatory problems we encountered with the theory of social conventions. The standard transaction-costs approach fails to explain either the distributional features of informal social conventions or the redistributive changes in these rules.
I should clarify this criticism in the context of current attempts by Libecap and North, among others, to incorporate distributional considerations. The standard approach has traditionally avoided the question of power asymmetries and instead has focused on the problem of transaction-costs minimization among symmetrically endowed actors (North, 1981). North (1990) acknowledges the importance of asymmetries for explaining distributional differences; Libecap (1989) explicitly addresses the complications introduced by distributional concerns in the effort to contract for socially efficient property rights. These attempts are in the spirit of the sorts of arguments I am making. But what these authors seem to have in mind are efforts at intentionally creating formal property rights. The importance of power asymmetries for explaining the emergence of informal rules governing property rights and economic organization has not been pursued; and this is the basis of my criticism. This failure can be explained by the reliance on competition as the selection mechanism for such decentralized emergence, competition thought to undercut the importance of power in individual exchange.
In regard to institutional change, the dubious relationship between a criterion of Pareto-superior change and strategic rationality has already been well exposed. Not only does the transaction-costs approach predict a Pareto-improving institutional change that might not be accepted by strategic actors (the path-dependent argument), but it also fails to explain either redistributive or Pareto-inferior change. This follows by implication from the combination of a reliance on voluntary agreement and a limitation to Pareto-superior change. If a change entails a loss to some individuals as a cost of establishing new property rights that produce greater aggregate benefits, narrowly rational actors will not voluntarily agree to the change.
The transaction-costs theory of exchange and competition has a way to resolve this weakness: a reconciliation of transaction-costs minimization (and the consequent maximization of aggregate benefits) with the pursuit of individual gain by means of compensation. We can see how the idea of compensation can be used in this way by examining Posner's (1980) interpretation of property rights and economic organization in primitive societies. He argues for an analogy between primitive economic institutions and insurance. That is, he explains the rules structuring economic activity in these primitive societies as a means of distributing risk in the community. The property rights scheme provides rules governing the enjoyment of community resources. Distributional advantages are granted to certain members of the community, and they are explained as the product of an exchange with the less advantaged, an exchange for the protection in the future if the community is struck by hard times. Here Posner reconciles distributional differences with the long-term efficiency of the property rights scheme. This insurance analogy can be seen as a form of compensation: The less advantaged are compensated for the distributional bias in the economic rights scheme by the promise of future insurance protection.
It is easy to reduce this insurance explanation to a mere functionalist assertion without some fairly rigorous empirical evidence. To see this, consider the general logic of compensation. Say that two actors currently share the benefits in the following manner: $60,000 to A and $40,000 to B. Now assume that circumstances change so that an alternative set of property rights allows the actors to produce an additional $25,000 per year but that this alternative scheme produces a new annual distribution: $50,000 to A and $75,000 to B. The logic of the new institutional economics predicts that as long as the aforementioned limitations on change are not applicable, the actors will adopt the new contract. But as long as the mechanism of change is one of voluntary agreement, my argument is that the new rules will not be adopted because they diminish A's benefits. If, however, B agreed to compensate A by giving her a side payment of at least $10,000 per year, then according to some accounts of individual rationality, it would be rational for A to agree to the changes.
Here the logic is plausible, and in the case of the intentional design of formal institutions, compensation is a possibility that must be considered. But the utility of this concept for explaining the spontaneous emergence of informal rights is highly questionable. The empirical requirements necessary to satisfy the theory are substantial, and there must be evidence of compensatory payments between the individuals in the institutional change, payments that are temporally related to these changes and are anticipated before the change. In the case of spontaneous emergence, an additional element must be shown: Either the form of compensation must be a common practice in similar exchanges in a society, or the compensation must be an element of the creation of those contractual rules that are subsequently selected out by the competitive process. Posner's insurance explanation exemplifies the weakness of most accounts of compensatory-like mechanisms: The evidence necessary to justify a compensation explanation is lacking.
Competition. But the transaction-costs approach invokes more than the actors' intentions in an individual exchange to explain the emergence of social institutions. Individual exchanges merely produce a variety of possible institutional forms, but the key selection mechanism is competition. Many explanations of institutional development and change situate the decision to establish social institutions in the context of a market or a marketlike environment. The main influence of the market on the choice of institutional form is in the competitive pressure it supposedly exerts on the institutionalization process. There are two related but conceptually distinct ways that competitive pressure can enter into this analysis.
First, as a dynamic effect, competition can be a selection mechanism that determines the survival of various institutional forms on grounds of survival and reproductive fitness. This is the logic behind Alchian's (1950) model of evolutionary competition, used in most economic analyses of institutional emergence. The existence of a large number of firms seeking profits from a common pool of consumers produces pressure for survival. Over time those firms that employ less efficient techniques lose profits to those that are more efficient. Losing profits eventually translates into extinction. As the competitive process continues, only those firms that use efficient techniques survive. Second, as a static effect, competition can undermine the actors' bargaining power in a particular interaction:
The main curb on a person's bargaining power, and the main pacifying influence on trade in general, is competition. A person has competition if the party he wants to trade with has alternative opportunities of exchange. The people who offer these alternative opportunities to his opposite party are his competitors. Competition restricts a person's bargaining power by making the other less dependent and therefore less keen on striking a bargain with him. (Scitovsky, 1971: 14)
This latter effect enters into a theory of development and change only to the extent that it establishes the environment in which rational actors seek to produce institutions.
The existence of competition raises questions for theories of institutional emergence and change. The dynamic effect forms the basis of the theory of exchange and competition. The relevant question here is whether the competitive pressure is sufficient to select out less efficient institutional rules. The static effect relates mainly to arguments such as the one in the next chapter, that would invoke the asymmetries of power in a society in order to explain institutional emergence. The relevant question there is whether the existence of competition prevents social actors from using asymmetries in power to develop institutions that produce systematic distributional consequences. This static effect is also related to the theory of exchange and competition in that it justifies ignoring power asymmetries in its own analysis. It is important to remember that competition is not an either/or phenomenon; there are degrees of competition and therefore degrees of competitive effect. The best way to answer these questions is to establish the empirical conditions under which competition affects the emergence of social institutions. Because the issues are so closely related, the analysis is best presented by addressing both of these questions. I will clarify a few of the conclusions in my later discussion of bargaining theory.
What are the prerequisites for the existence of competition? Lists of the necessary conditions for marketlike competition are numerous. According to Scitovsky's analysis (1971) of competition, the following conditions are necessary for the existence of competition in social institutions: (1) a large number of competitors in pursuit of a common pool of resources, (2) a set of institutional alternatives differentiated only by their distributional consequences, (3) full information about the availability of alternatives, and (4) low transaction costs. If an explanation of institutional change is to invoke competition as a relevant factor, these conditions must be empirically satisfied.
Discussions of competition in the market are numerous. What I want to do is limit my analysis to those issues unique to the question of competition in regard to social institutions, as this will allow to me to emphasize the difficulties of satisfying these conditions in the institutional case. Central to my argument is the following distinction: Institutions are not goods.
<#FROWN:J24\>
Why Fewer Women Become Physicians: Explaining the Premed Persistence Gap
Robert Fiorentine and Stephen Cole
Previous research indicates that the answer to the question of why fewer women become physicians lies in the 'premed persistence gap.' Women are no less likely than men to enter undergraduate premed programs, but they are less likely to complete the program and apply to medical school. This article presents data from a study designed to test four plausible explanations of the persistence gap that are consistent with the structural barriers, normative barriers, and cognitive differences theories of gender inequality. The findings do not support the 'perception of discrimination' hypothesis, the 'discouragement' hypothesis, the 'self-derogation' hypothesis, and the 'anticipated role conflict' hypothesis. Rather, the evidence suggests another explanation - the normative alternatives approach, This approach holds that contemporary gender norms offer women fewer disincentives to changing or lowering their high-status career goals when encountering hardship, self-doubt, and the possibility of failure.
INTRODUCTION
This article presents the latest results of a research program designed to understand the contemporary causes of gender inequality in the United States (Cole, 1986; Fiorentine, 1986, 1987, 1988a, 1988b). The specific area of research is medicine, one of the most prestigious and lucrative occupations.
As is true with other prestigious occupations, medicine is characterized by extensive gender inequality. Men make up about two-thirds of medical students and more than 80% of all practicing physicians in the United States (Cole, 1986; U.S. Bureau of the Census, 1987; Bickel, 1988). Men are overrepresented in the most prestigious specialties, and they are more likely to hold positions of authority in hospitals and clinics (Lorber, 1984).
Three theoretical approaches have emerged in sociology and social psychology to explain the underrepresentation of women in high-status professional and executive positions. The structural barriers approach holds that differences in occupational achievement are a consequence of sex discrimination that limits the opportunity of women. Because female gender may be a 'discrepant status' that undermines trust and certainty so highly desired by organizational players (Kanter, 1977) or because prejudicial attitudes of employers force women into the less desirable, secondary-sector jobs (cf. Bibb and Form, 1977; Bridges, 1982) or because capitalist production systematically exploits female labor (Hartmann, 1976; Eisenstein, 1979), women may encounter barriers to their mobility.
The normative barriers approach assumes that gender socialization brings young women to view the pursuit of success in a high-status career as a transgression of norms (Angrist and Almquist, 1975; Douvan, 1976). Anticipating social rejection (Horner, 1972), and believing they would be unable to fulfill the role expectations of wife and mother if they were committed to a demanding career (Angrist and Almquist, 1975; OLeary, 1974), young women place limits on their ambitions, emphasize the primacy of their domestic role, and select normatively appropriate, 'feminine' occupations.
The cognitive differences approach assumes that because cultural stereotypes do not depict women in achievement roles (cf. Weitzman et al., 1972), or because parents and others are less encouraging of female achievement (Hoffman, 1972), women have lower confidence in their ability to perform successfully in a variety of achievement situations (Lenney, 1977; Maccoby and Jacklin, 1974). With lower levels of confidence, women are more likely to attribute their successes to 'external' or 'unstable' causes such as luck or effort, and their failures to 'internal' or 'stable' causes such as low ability or task difficulty (cf. Weiner et al., 1971; Deaux, 1976; Frieze et al., 1982). As successes are discounted while failures are affirmed, women are less likely to enter into, persist with, or perform well in a wide array of achievement tasks.
In earlier research we looked closely at the structural barriers explanation of the underrepresentation of female physicians. Inasmuch as women comprise only about one-third of the students currently admitted to medical school, it could be that medical schools discriminate against female applicants. Medical schools may employ restrictive quotas (Walsh, 1977), they could subject female applicants to more rigorous admissions standards, of they may systematically, if unconsciously, devalue their credentials. But in an analysis of application and admission rates of males and females to all American medical schools (Cole, 1986), it was determined that female applicants have about the same qualifications as male applicants, and they are just as likely to be admitted to medical schools. Women account for one-third of the admittees simply because they constitute one-third of the pool of applicants.
The evidence suggests that medical schools do not overtly discriminate against female applicants. But what happens once they enter medical school? Are women admitted only to be systematically 'cooled out' after they begin their training? The evidence suggests not. Women graduate from medical school at almost the same rate as men (Journal of the American Medical Association, 1984, 1986).
If women are as likely as men to get into and through medical school, then efforts to understand why fewer women become physicians need to focus on gender differences in ambition rather than differences in opportunity. It needs to be determined why fewer women apply to medical school.
Subsequent research has demonstrated that the answer lies in the 'premed persistence gap': An equal ratio of women and men enter undergraduate premed programs, but men are twice as likely to complete the program and apply to medical school (Fiorentine, 1986, 1987). Further, it was determined that the persistence differential cannot be explained entirely by differences in academic performance, even though female premed students earn slightly lower grades in the required premed courses. For while women with a marginally competitive or a noncompetitive academic performance are more likely to relinquish their medical career goals, women with a competitive grade average of 3.50 or higher are no less likely to persist in the program and apply to medical school.
A similar persistence gap exists for male and female high school students who aspire to high-status professional occupations such as doctor, college teacher, scientist, and lawyer. Women with an A average are just as likely as their male counterparts to persist with these career aspirations, but at every lower academic level, males are from one and one-half to two times more likely to maintain their professional aspirations from the sophomore to the senior years (National Center for Educational Statistics, 1983; Fiorentine, 1986). This suggests that the premed persistence gap is explained by general social or cultural processes rather than by factors that are unique to premedical programs.
Although that earlier study demonstrated that sex differences in academic performance account for only a small portion of the premed persistence differential, it was not designed to empirically examine nonperformance explanations of the persistence gap. This article presents data from a second study that tests four plausible explanations that are consistent with the (informal) structural barriers, normative barriers, and cognitive differences theories of gender inequality.
Perception of Discrimination
Even though medical schools do not overtly discriminate against female applicants, this 'objective reality' may not be the subjective perception of many female premed students. It may be that young women erroneously assume medical schools restrict admission to all but the most academically distinguished female students. If so, then those with a highly competitive performance would be about as likely as young men to persist, but those with a less competitive performance would underestimate their chances of admission, and would be less likely to complete the premed program and apply to medical school.
Discouragement
Although there has been a great deal of change in gender expectations over the last several decades, it could be that professors, parents, peers, and lovers continue to treat the medical career goals of young women with more indifference, less encouragement, or even outright hostility. As premed programs are usually competitive, even slight differences in encouragement may lead women with marginally competitive and non-competitive performances to believe the goal of becoming a physician is unattainable or inappropriate. It would not be immediately apparent, however, why women with a competitive performance are either not discouraged by others or do not react to this discouragement by relinquishing their medical career goals. One possibility is that the negative reactions of others are not sufficiently strong to sway the most determined, who eagerly and realistically anticipate success.
Self-Derogation
Consistent with the assumption of expectancy-value and attribution theories (Atkinson, 1964; Weiner, 1986; Weiner et al., 1971), it may be that young women enter the premed program with lower expectancies of success, and attribute a less-than-competitive performance to stable, internal causes (low ability), or to stable, external causes (high task difficulty). If so, then women would be less likely to believe they could turn a marginally competitive or a noncompetitive performance into a competitive one, and consequently, less likely to persist in the premed program.
Anticipated Role Conflict
Finally, it may be that young women initially underestimate the ease of negotiating family demands and a career in medicine. Realizing they are forced to choose between their career goals and their family plans, most relinquish their career goals, while some, particularly those with a competitive academic performance, relinquish their conventional family plans.
METHOD
As in the original investigation, transcripts of all male and female nontransfer premed students entering the State University of New York at Stony Brook as freshmen between 1982 and 1985 were acquired. During four weeks in April 1986, trained interviewers attempted to locate and interview, via telephone, all persisting and defecting students, including those who had withdrawn from the university. Interviews were completed with 302 males and 240 females. Only one male student refused to participate in the survey. Interviews were completed with all persisting and defecting premed students who were currently enrolled at the time of the survey. There was no difference in the distribution of males and females by year in college. Among the males in the sample, 26% were freshmen, 26% sophomores, 25% juniors, and 22% seniors. Among females, 27% were freshmen, 26% sophomores, 25% juniors, and 22% seniors.
As typical in attrition studies, some of the students who were no longer enrolled at Stony Brook could not be located in a national search, and were not interviewed. The sample represents 73% of all males and 60% of all females who began their studies as premed students. Because females were more likely to drop out of the premed program (and not leave a forwarding address or phone number) they were more likely to be in the hard-to-find subsample. Students who were not interviewed had very low GPAs (about a D+ average), but there was no difference between males and females in this group. Inability to locate these students had the following effect on the results: (1) the sample has a higher grade average than the population of all those who began as premed students and (2) the proportion of persisters in the sample is higher than it would be if there were data on all those who began as premeds. This would not, as far as we can tell, have influenced any of the conclusions about gender differences in persistence.
Focused fact-to-fact interviews also were conducted with 23 female and 13 male premed students during the fall 1985 and spring 1986 semesters. Students in the main sample (n = 542) were categorized along lines of persistence, grades, and sex; and then a small number within these broad categories were randomly selected for interviewing.
In order to assess initial expectancies of success prior to entering the premed program, a list of high school students who planned to begin as freshmen in the fall of 1986 and were preregistered for any two of the required premed courses was acquired from the registrar's office (n = 76). Sixty-eight of these students could be located, and interviews were completed during the same four-week period in April with the 62 students who indicated they planned a 'premed' course of study.
Premed persistence was assessed by self-report measures along with actual records of medical school application acquired from the Office of Undergraduate Studies. Respondents were determined to be 'persisters' if they either had applied to medical school (premed students typically apply to medical school during the spring semester of their junior year) or indicated they planned to apply to medical school in the future.
<#FROWN:J25\>
DESIGN FOR VULNERABILITY: CUES AND REACTIONS TO FEAR OF CRIME
Jack L. Nasar
Ohio State University
Bonnie Fisher
University of Cincinnati
Fear of crime is a critical problem on university campuses. This paper describes cues in the built environment that may affect fear of crime. It develops and tests a theory about the relationship between these cues and fear, and consequent reactions. In our analysis of responses to open-ended questions, we found that fear was heightened by several site-specific cues: poor prospect for the passerby due to inadequate lighting, blocked escape for the passerby, and concealment for the offender. Respondents also reported avoidance, protective, and collective actions in response to their site-specific fears. The results suggest that reductions in fear (and actual crime) on campus may be achieved through the design of micro-level physical features.
Introduction. Fear of criminal victimization threatens the quality of life for many Americans (Gallup Poll, 1989). Almost half of the U.S. population have reported feeling unsafe in areas within a mile of their homes (National Opinion Research Center, 1987). Many citizens have been found to feel unsafe in the neighborhoods where they shop, work, go to school, and seek entertainment (Bureau of Justice Statistics, 1984; Fisher, 1991).
Fear of crime is also a significant problem on college campuses, causing faculty, staff, students, and parents to demand safer campuses (Gaines, 1989). Recent court rulings have held universities liable for foreseeable victimizations (Raddatz, 1988), and Congress has passed a law requiring colleges and universities to publicly report their crime statistics (House Report 101-883, Section 201-205, 1990). All of these factors pressure college administrators to address the safety concerns and needs of the university.
To better understand causes of fear of crime, researchers have examined demographics factors. They have found higher levels of fear among socially or physically vulnerable individuals, such as minorities, low-income people, women or the elderly, especially after dark (Box, Hale, and Andrews, 1988; Skogan and Maxfield, 1981; Warr, 1984). Most young urban women, a group prominent on many campuses, have been found to fear certain types of personal crimes, such as rape (Warr, 1985); and when individuals are fearful, they have been found to adopt different crime prevention behaviors (Gates and Rohe, 1987; Lab, 1990).
Researchers have also moved beyond demographics predictors to examine contextual causes. In this regard, we see the occurrence of victimization from two interrelated theoretical levels - one at the macro-level and the other at the micro-level. From the macro-level perspective, researchers have consistently found that neighborhood-level problems, such as 'physical and social disorders' or 'incivilities,' heighten fear of crime (Gates and Rohe, 1987; Skogan and Maxfield, 1981; Wilson and Kelling, 1982). While this macro-level perspective explains broad characteristics of fear, it misses the micro-level cues in the built environment that affect feelings of vulnerability and fear of crime at a specific location. Yet, research has shown that offenders rely on micro-level cues to select a suitable target (cf. Taylor and Gottfredson, 1986). The cues indicate opportunities for committing a crime. A sizable body of research has revealed links between the built environment, crime, and crime prevention at the street or block level (Jacobs, 1961; Jeffery, 1977), or the site level, as in housing projects (Newman, 1972; Merry, 1981) or schools (Hope, 1985).
The role of the physical environment on university campuses has been overlooked. Yet, campuses have high levels of crime and fear of crime, and they often have macro-level conditions that lead to these problems. Routine activity theory explains victimization as a function of the routine activities of victims that place them at risk (Cohen and Felson, 1979). Campuses exhibit all three risk factors: the opportunity for victimization presented by potential targets (students, faculty and staff), a supply of motivated offenders nearby (neighborhoods with poor socio-economic conditions), and poor guardianship (open access and varying levels of security among students).
With knowledge of how micro-level features affect crime and fear of crime, university officials can institute design policies to reduce the risk. They can plan, evaluate or alter campus buildings and micro-level features to promote safety. Compared to students and personnel, physical facilities are relatively permanent. Thus, while educational programs only have short-term value because they must be repeated for new students and personnel, environmental design strategies can produce long-term positive effects on crime and feelings of vulnerability.
How can campuses be designed for safety? Unfortunately, there has been little empirical evidence on the effects of the physical arrangements on crime and fear of crime on campus. Most studies of environmental strategies have concentrated on residential or commercial areas (Kuskmuk and Whittemore, 1981). This paper draws from both the macro- and micro-level perspectives to understand the cues in the outdoor environment to feelings of vulnerability and reactions to that fear. We present a theoretical model of how micro-level site features affect fear of crime and we describe the results of a study aimed at uncovering some of those cues in relation to a campus setting. In particular, we analyzed written responses to open-ended questions in relation to three buildings on a university campus. The results revealed three micro-level cues as important influences on site-specific fear of crime: refuge, escape and prospect (depending on different light levels). Furthermore, we found evidence of changes in behavior in response to the site-specific fears.
Cues and reactions to fear of crime. The built environment may go relatively unnoticed until it creates a substantial problem or pleasure. Nevertheless, the design of the built environment does influence perceptions of safety (Fisher and Nasar, 1992) which in turn may affect psychological well-being, and spatial behaviors (Nasar, 1988; Russell, 1989).
In areas where crime and fear are present, people regularly evaluate their risk of victimization by scanning their immediate environment for cues of danger (Appleton, 1975; Gates and Rohe, 1987; Goffman, 1971; Warr, 1990). Based upon their assessment of the presence or absence of such cues, they adjust their behavior accordingly. In discussing cues for alarm, Goffman (1971) observed,
Individuals seem to recognize that in some environments wariness is particularly important, constant monitoring and scanning must be sustained, and any untoward event calls forth a quick and full reaction.(242)
Individuals do not necessarily notice all signals as cues for alarm. They only attend to those cues in their immediate environment as signals for alarm. According to Goffman,
The individual's surround ... (is) that region around him within which signs for alarm ... can originate ... This is likely to be measured by means of a radius that is only yards long. His body is what he mainly can become immediately concerned about, and it is principally vulnerable ...(253-254).
As we move from place to place, our immediate surroundings continually change, changing the position or distance of different signals of alarm. Consequently, different types of situations give rise to alarm while others do not. For example, Goffman noted,
Walking along the street, the individual's surround follows him; walking around a room, it does not, or does so only in a small degree. Note, the individual is likely to concentrate his scanning for signs for alarm at the moment and place of his entering a bounded area. For often it will be at the doorway that he will have to notice alarming things if he is to notice them in time. (255-256)
When a person's proximate surroundings have a cue for alarm, the person feels vulnerable and reacts. As Goffman puts it,
... when the subject senses that something is up, his attention and concern are mobilized; adaptive behavior occurs if the alarms proves 'real,' but if reassuring information is acquired, the alarm proving false, his concentration will decay quickly. (262)
In contrast, when a cue for alarm is absent, "When the world immediately around the individual portends nothing out of the ordinary," Goffman writes that the person,
will sense that ... he is safe and sound to continue on with that activity at hand with only peripheral attention given to checking up on the stability of the environment.(239)
According to Goffman, then, individuals monitor their environment for signals of danger. When they detect a danger signal, they react accordingly. Otherwise, when appearances are "natural" or normal (23), they go about their routine.
It becomes important, then to identify what cues convey the message of alarm or danger, how such cues provoke fear, and what responses then follow. In discussing the design of vulnerability, Goffman (1971) introduced the concept of 'lurk lines'. These lines demarcate zones that lie beyond or behind the individual's line of sight. These zones have also been called 'blind' spots (Warr, 1990). Lurk lines or blind spots occur in many places - the areas behind an open door, inside an unlocked closet, or around a sharp bend in hallway. You do not necessarily have to walk through a physical obstruction to discover these areas. You can see potential blind spots, in the near distance and anticipate their likely character.
The concept of lurk lines and blind spots conforms with Appleton's (1975) discussion of prospect and refuge. Appleton argued that for evolutionary reasons humans favor places that afford prospect (an open view) and refuge (protection). In Goffman's (1971) and Warr's (1990) terms, humans dislike lurk lines or blind spots. Places with prospect and refuge offer an observation point from which humans can see, react, and if necessary, defend themselves, as well as a protective space to keep them from harm. Appleton (1975) went on to say that humans need not enter an area to determine it's prospect and refuge. They anticipate what those qualities would be if they were to enter the place. For example, observers can look at a clear-cut mountain-top in the distance and anticipate that it would offer them prospect but little refuge should they venture there.
We believe that potential offenders also value places which offer prospect and refuge. As Lorenz's (1964: 181) said, such places offer an advantage to "the hunter and hunted alike - namely to see without being seen." This agrees with views on the role of natural surveillance in reducing crime (cf. Jacobs, 1961, Jeffery, 1977; Newman, 1972); and several studies indicate that robbers favor prospect and refuge. Bank robbers have been found to target banks that have poor visibility of the inside from the outside, and clear views out to escape routes and to see who might enter (Camp, 1968). Tapes of bank robberies show that robbers select banks and follow a path that let them see as much as possible while remaining unseen (Wise, 1983). In a study of 78 tapes of sting operations, Archea (1985) found that robbers operate from locations with prospect ("enough visual access to gain and maintain control") and refuge ("low visual exposure from areas not in direct control") (p. 249). Stoks (1983) found that rape outdoors tend to occur in areas with refuge ("physically confined" spaces) and limited prospect ("barriers") nearby areas of pedestrian movement (p. 334). In sum, the research confirms prospect and refuge as relevant constructs in relation to site-specific opportunities for crime. It was our hypothesis that prospect and refuge would also affect fear of crime among passersby in outdoor settings.
In theory, humans would feel vulnerable passing by places such as alcoves and tall, dense shrubs that afford potential offenders refuge (concealment) and potential victims limited prospect. Offenders would have hiding places, into which passersby could not see in time to escape or avoid an attack. Conversely, where pedestrians have wide and deep prospect onto an area which has no places of refuge for an offender, they would feel safer because they could anticipate and avoid an attack.
Expanding the notion of prospect, we argue that insufficient lighting at night also provokes feelings of vulnerability because it limits prospect. The idea that the night may be scary is not new (Fisher and Nasar, 1992; Warr, 1985). Previous research has found that darkness represents a potent sign of danger, and combined with novelty and uncertainty can produce considerable fear (Warr, 1990). We believe, however, that it is not only darkness per se, but also its effects on prospect that effect site-specific fear of crime.
<#FROWN:J26\>
Population Decline from Two Epidemics on the Northwest Coast
ROBERT BOYD
Direct contact by Indians of the Northwest Coast culture area with Euro-Americans began very late, in 1774. Therefore, documentation of the contact process is relatively good. It appears that the introduction of Old World diseases was similarly late and coincided with contact. Although Campbell (1989), using archeological data, has made a case for penetration of the Northwest Coast by the hemispheric smallpox epidemic of the 1520s, all other evidence (archeological, historical, Indian oral tradition) for Northwest disease introduction postdates 1774. That evidence is considerable, with nearly the full list of high mortality infectious diseases introduced in rapid fire sequence, with accompanying heavy mortalities strongly suggesting that these diseases were new to the area, or at least present again after a very long absence.
Smallpox epidemics appeared roughly every generation - the first, in the late 1770s following direct contact, over the entire coast, and others in 1801, 1836-1838, 1853, and 1862-1863 in more limited regions (Boyd 1990:137-143). This interesting sequence appears to have been created by the dual factors of introduction from an outside source plus presence of a large enough pool of non-immune susceptibles in the Northwest born since the last epidemic. Evidence for depopulation from the initial outbreaks is limited to descriptions of abandoned villages and native recollections, but high mortalities from the later epidemics are well documented in the Indian censuses and estimates compiled by the Hudson's Bay Company and by early government officials. Besides smallpox, malaria was introduced to and became endemic in a sizable portion of the southern coast, and records indicate considerable population decline concurrent with its introduction. Other 'new' diseases that contributed to Northwest Coast Native American population decline included measles, influenza, dysentery, whooping cough, and (in a different way) tuberculosis and venereal diseases (Boyd 1985, 1990).
This is the situation for the region as a whole. This chapter will discuss the records on population decline associated with two localized outbreaks - on the lower Columbia in the decade following the introduction of malaria in 1830, and in the Queen Charlotte Islands during the years of the last great smallpox epidemic, 1862-1863. The two epidemics appear to qualify as virgin soil or near virgin soil; malaria had not been known in the Northwest prior to 1830, and smallpox had been absent from the Charlottes since the late 1770s. The data on depopulation in these two instances are remarkably good and may provide clues to patterns of depopulation in similar virgin soil situations in other places and at earlier dates in the Americas.
MALARIA ON THE LOWER COLUMBIA, AFTER 1830
The Lower Columbia drainage, here defined as the area from the river mouth to near The Dalles and including the basins of large tributaries such as the Willamette and Cowlitz, was home to several Indian groups. The two largest, Chinookans on the banks of the Columbia and Kalapuyans in the Willamette Valley, are also the most completely documented as far as population decline is concerned; and they are used as the base of discussion here.
In the late 1820s the population of these two peoples approximated 18,580; within a decade it had dropped to 2,433, or 13 percent of the initial number. Detailed discussions of the sources for these numbers have been given elsewhere (Boyd 1985, 1990, 1991a; Boyd and Hajda 1987); suffice it to say that the preepidemic estimates are from Meriweather Lewis and William Clark and the Hudson's Bay Company, and the postepidemic numbers are from the 1841 U.S. Exploring Expedition.
What caused this monumental decline in Lower Columbia populations? The contemporary sources (a large body of data) uniformly ascribe it to the disease called 'fever and ague' by Americans and 'intermittent fever' by the British. Despite some controversy over the identity of this disease (e.g., Cook 1955; Taylor and Hoaglin 1962), epidemiological evidence strongly suggests that it was malaria (Boyd 1975), described most often in later accounts (e.g., Townsend 1839:197; Brackenridge 1931:141, 221; J.R. Dunn 1846) as the tertian form (Plasmodium vivax and P. falciparum). The sources are unanimous in stating that this disease was new to the area, and unknown before 1830 (e.g., McLoughlin 1941:88). This points to a virgin soil situation. Malaria, as has been fully established by research (F. Dunn 1965; Wood 1975), is an Old World disease, which was introduced in sequence to different regions of the Americas, where suitable vectors, mosquitoes of the genus Anopheles, were native (Harrison 1978). Once introduced, it became endemic. In the unbroken tropical and subtropical regions of the Amazon, circum-Caribbean, and Southeast United States this introduction was early and apparently has gone unrecorded (Ashburn 1947:112-123; Friedlander 1969). In the Lower Columbia and adjacent California central valley (Cook 1955) malaria arrived late and was both observed and recorded.
The statistics from the area are unambiguous in documenting a dramatic population decline following the introduction of 'fever and ague.' What remains unclear are the dynamics of this decline. Such high mortalities are not characteristic of endemic malarial regions today, nor of those occasional instances where a combination of factors inflate mortality rates in endemic areas to numbers that merit the term 'epidemic' (Wood 1979:260). It is essential to remember that on the Lower Columbia the outbreak was a virgin soil epidemic, and virgin soil outbreaks of whatever disease are almost by definition unusually virulent (Crosby 1976). This may be part of the answer to the large decline. Three other contributing factors are: the Indian means of dealing with the alternating spells of chills and fever characteristic of the disease, secondary diseases that precipitated mortality, and fertility decline directly or indirectly associated with malaria.
A small, consistent, yet independently written body of sources from the earliest years of the epidemic ascribe the high mortality to inadequate treatment of the disease. During the fever spells, Indians would plunge into cold water; when chills came they retreated to sweat lodges. Sudden death followed. The phenomenon is securely documented; the actual dynamics of death are not (Boyd 1979). Quinine sulfate, the medicine used by Whites, was in short supply in the early 1830s (Allan 1882:79; McLoughlin 1941). Factor two: Modern malaria is a disease of complications; death, when it comes, is usually from a secondary illness superimposed on a weakened malarial constitution (Brown 1983:65). Children, in endemic regions, form the bulk of casualties (Wood 1979:258-259). Though evidence is sparse, deaths listed in the record book of an 1830s Willamette Valley mission school, for instance are ascribed to such secondary diseases (including influenza and tuberculosis) introduced to Indian children chronically ill from malaria (Shepard 1922). It seems likely that much of the mortality in later years may have been due to secondary illnesses.
These two factors, in combination, appear to have contributed to the mortality side of the population equation in the Lower Columbia valley. But it appears that there may have been a decline in fertility as well in this newly malarial population. Post-malaria Chinookan and Kalapuya censuses (Hudson's Bay Company 1838; Spalding 1851; Boyd 1985: chart 26) incorporate information on population structure, specifically age and sex. Censuses of groups within the epidemic focal area show low - below 30 - percentages of children. Elsewhere in the Northwest, among Indian populations counted by the Hudson's Bay Company but not subjected to disease, the percentages are much higher, generally in the 30-40 percent range. In the case of the 1851 counts, mortality from the 1848 measles epidemic (Boyd 1991b) decreased the percentage of children even more. But in 1838 and 1851, two underlying explanations for fertility decline also seem likely. First, in immediate postepidemic years (MacArthur 1961:8), there was a sudden drop in fertility due to disrupted marital units. Second, there was considerable malaria-caused anemia, which, among women of child-bearing age, led to spontaneous abortions and stillbirths (Wood 1979:258).
The approximate 87 percent decline in Lower Columbia Indian populations in the decade following the introduction of malaria therefore appears not to have been a simple process. Inappropriate Indian treatments precipitated deaths in cases where proper treatment would have resulted only in temporary debilitation; secondary ailments moved in and pushed weakened malarial cases on to death, particularly among the very young; and maternal anemia is likely to have caused spontaneous abortions and miscarriages, resulting in a drop in fertility. The decline was regular and cumulatively dramatic: in 1841 an American physician, resident since 1835, estimated that the population of the Willamette Valley Indians diminished by one-quarter annually (Bailey in Wilkes 1926:57), a proportion fully in keeping with before (late 1820s) and after (1841) Kalapuyan population figures.
Socially, after 1841, most Chinookan and Kalapuyan populations were approaching extinction as viable and ethnically distinct entities. Inhabitants of Chinookan winter villages, tied to their local land base, remained in place with minimal regrouping (Dart 1851) until they were reduced to a few survivors or overwhelmed by migrants from more robust populations (Sahaptin or Salishan) from the peripheries of the endemic malaria zone (Tappan 1854). The 1851 Kalapuya census showed that these people likewise remained in their local band territories, with negligible regrouping. The average band size in 1851 was a mere 53 people, and the bands were evenly spread over the entire valley. As on the Columbian, more robust outsiders (Sahaptin Klikitat) had begun to move in, and by 1851 they were equivalent in numbers to the sum total of the Willamette Valley indigenes. The Kalapuyan population nadir was reached in the 1850s, and numbers (in what was a highly mixed population) began to increase again only after enforced concentration on the Grand Ronde Reservation and regular access to Western medicines.
SMALLPOX IN THE QUEEN CHARLOTTE ISLANDS, 1862-1863
The steady population decline of the Lower Columbia Indians following the introduction of malaria was very different in nature from the sudden catastrophic drop of the Haida population of the Queen Charlottes caused by the 1860s smallpox epidemic, even though the percentage losses were comparable.
The 1862-1863 British Columbia smallpox epidemic has been documented (Yarmie 1968; Pethick 1978). Over 20,000 Indians, nearly 60 percent of the pre-epidemic total, died in British Columbia and the Alaska Panhandle (Duff 1964; Boyd 1985, 1990; 142-144). In classic fashion, the disease arrived with an infected individual on a ship from San Francisco. The ship docked at Victoria, and the disease spread rapidly to the crowded Indian encampment on the city's outskirts. Instead of quarantining the Indians, the authorities evicted them, and fleets of Haida, Tlingit, Tsimshina, and Kwakiutl traders sailed back to their homelands, taking the epidemic with them.
On the densely populated (relatively speaking) and isolated Queen Charlotte Islands, the effect was rapid and devastating. A Hudson's Bay census dating from 1839-1842 showed a total of 8,428 Haidas (table 1); there is no evidence that this number changed significantly in the next 20 years. The people were concentrated in several winter villages spread along the coastline in areas of heavy resource concentration (especially halibut, the local staple). Each village was associated with a particular matriclan.
<O_>table&caption<O/>
Census of the Haida taken 1882-1884 (Chittenden 1884; Petroff 1882) show 1,598 survivors, or a population drop of over 80 percent, nearly all of which must on present evidence be assigned to the smallpox epidemic. This exceedingly large loss was due to several factors. First, the strain of smallpox virus in the 1860s epidemic was particularly virulent; second, Haida settlement was (relatively) dense and continuous, facilitating transmission; third, the Queen Charlotte population was particularly vulnerable, as there was no segment with acquired immunity, as was the case with most other British Columbia Indian populations. Whereas Tsimshian, Tlingit, and others had experienced an outbreak 24 years prior, the Haida had not known the disease for more than 90 years (Boyd 1985; Blackman 1981:23; cf. MacDonald 1983:17). It was, in essence, a virgin soil experience for them. A fourth factor was vaccine. While sizable numbers of Tlingit and some Kwakiutl had been vaccinated in the 1830s, and the Metlakatla Tsimshian in 1862, none of the Haida had been vaccinated (Boyd 1985, chapter 4 and appendix).
<#FROWN:J27\>
The Division of Household Labor
The most obvious predictors of wives' perceptions of fairness might seem likely to be the outcomes related to the quantity of labor, such as the extent of the husband's contribution to household labor and the total amount of labor performed by the wife. Benin and Agostinelli (1988) have shown that women are relatively unaffected by increased household labor participation on the part of husbands unless the labor is specifically spent in those tasks traditionally defined as 'feminine' chores, such as cooking or cleaning. What might explain this pattern? First, the traditional male tasks involve little investment of time relative to the female tasks. Second, male involvement in female tasks is likely to have direct and visible consequences for the wife's workload. Third, since husbands' contributions to such tasks cross traditional gender lines, they will be more salient and more likely to be perceived as involving a special effort by husbands to be 'fair.' We thus expect perceptions of fairness to be affected by husbands' participation in 'female' household tasks, but not by their participation in 'male' or gender neutral tasks. Since our arguments for this effect hinge in part on the reduced need for wives' labor, we also predict a direct effect of wives' total labor on perceptions of fairness.
However, we do not expect even this aspect of husbands' household labor contribution to be related to perceptions of fairness in all families. As Thompson (1991) rightly emphasizes, women's orientations to various household outcomes vary considerably. It is still the case that many families allocate labor market participation exclusively to husbands (for the sample utilized here, about a third of the families have nonemployed wives), with household labor presumably becoming the wife's responsibility. For such families, it is unlikely that husbands' participation in housework would have much impact on perceptions of fairness. Hence, all of our analyses will be carried out separately for families in which the wife is employed and those in which she is not. We expect that the quantity of labor variables will have their clearest impact for families in which the wife is employed.
Qualitative Aspects of Household Labor
Thompson has reasonably argued that quantitative aspects of the division of labor or "the distribution of time and task" do not tap the full range of possibly important outcomes. For example, Oakley (1974) found that wives who were dissatisfied with the division of labor in their home were likely to complain about the monotony and fragmentation of their chores. Dissatisfied wives were also likely to complain about the loneliness of household labor; that is, they expressed concern over the lack of companionship during the performance of household labor. Thus, although the relationship between 'satisfaction' and perceptions of fairness may be complicated, it is reasonable to hypothesize, on the basis of Oakley's findings, that wives' perceptions of the extent to which their household labor is lonely and complicated will be related to their perceptions of the fairness of the division of labor.
Appreciation
Kessler and McCrae (1982) conclude that husbands' willingness to participate in housework is important to wives' in part because it carries a 'symbolic meaning' for wives, such that their work is recognized and appreciated by the husbands. Thompson (1991) also argues that among the most important outcomes of the division of household labor are symbolic outcomes, particularly the significance of caring. We would argue that the 'satisfaction' involved in such a demonstration will be greater when wives perceive that their work is, indeed, appreciated. We expect, therefore, that women who do not perceive their household labor to be appreciated will be more likely to feel that the division of labor is unfair, and that this effect should be evidenced whether or not she is employed.
Ideological Factors
What role might we expect gender ideology to play in women's reactions to division of house-hold labor? A number of authors (e.g., Huber and Spitze, 1983) assume that gender ideologies are an important determinant of the allocation of chores within the home, and we will be able to test this assumption. However, since task allocation certainly takes place through a negotiation process (Atkinson & Huston, 1984), and one that might be somewhat protracted (Hochschild, 1989) there will undoubtedly be a far from perfect relationship between gender ideology and the division of labor. Given the general finding that men, on average, contribute very little to house-hold labor, it is reasonable to assume that there will be some direct relationship between women's gender ideology and their assessment of fairness in the division of labor.
Additionally, however, we would expect ideology to be related to all three of the factors in Thompson's (1991) model. First, husbands' contributions to the division of household labor should be a particularly salient outcome for less traditional wives. Second, less traditional wives should be more inclined to make the cross-gender comparisons that would produce dissatisfaction with husbands' lack of participation. Third, less traditional wives will presumably be less likely to accept the traditional justification for low levels of husbands' contributions to household work. Thus, we would expect an interaction effect involving ideology and husband's contributions, such that women with egalitarian ideologies will react strongly to the extent of their husband's participation, while those with more traditional ideologies will not. And we would expect this effect to be particularly clear for those families in which both partners participate in the paid labor force.
DATA AND VARIABLES
Data for this study are taken from the 1988 National Survey of Families and Households (Sweet, Bumpass, & Call, 1988). The NSFH provides a cross-sectional national sample of 13,017 respondents aged 19 and older. The sample here is limited to 778 married white women for whom data were complete across the variables required here-in. Our smaller sample size results from the restriction of the study to only white, currently married wives for whom data on both attitudinal and household labor measures are complete. The primary limitations on sample size result from the use of only married households (68% of the total sample), women (50% of the total sample), whites (85% of the total sample), and individuals less than 65 years of age (86% of the total sample). The sample was also restricted to include only respondents who reported a weekly household labor performance of 120 hours or less.
Household labor. For each couple, the NSFH includes reports from each partner/spouse regarding their own household labor contributions and those of their partner/spouse. Each partner provides an estimate of the hours spent per week on eight specific household tasks. The tasks are: (a) preparing meals, (b) washing dishes, (c) cleaning house, (d) outdoor tasks, (e) shopping, (f) washing and ironing clothes, (g) paying bills, and (h) auto maintenance. Given the focus of this investigation on wives' reactions to the division of labor, measures of both husbands' and wives' household labor are based on reports from wives.
Two primary aspects of the quantitative division of household labor are of central interest here: how much work is done by husbands and how much by wives. First, with regard to husbands' household labor, our major focus will be on male labor in female-dominated tasks, the number of hours spent by husbands in those chores typically envisioned as 'female' tasks (meal preparation, dishes, cleaning house, and ironing and washing clothes), as measured by wives' estimations. We will also begin our analysis of the effects of male labor on wives' perceptions of fairness with a demonstration that male labor in male-dominated tasks (outdoor tasks and auto maintenance) and male labor in gender neutral tasks (shopping and paying bill) have little to do with perceptions of fairness. Second, total female labor is the total number of hours spent in all household labor per week by wives.
Qualities of household labor. Three measures of wives' assessment of the nature of their household labor were available in the NSFH. Responses were taken from the question: "How would you describe the work you do around the house? Would you say it is: (1) boring-interesting, (2) complicated-simple, (3) lonely-sociable?" Responses were taken on a 7-point scale, with the responses shown above representing the respective end points of each scale.
Appreciation. The extent to which women perceive their household labor to be appreciated is based on the same question used to assess the qualities of household labor. One of the 7-point scales presented in the NSFH was "unappreciated-appreciated."
Gender ideology. An indexed measure of wives' sex-role ideology was created from responses to the following questions: (a) It is much better for everyone if the man earns the main living and the woman takes care of the home and the family, (b) Preschool children are likely to suffer if their mother is employed, (c) Parents should encourage just as much independence in their daughters as in their sons, and (d) In a successful marriage, each partner must have the freedom to do what they want individually. Wives answered each question on the basis of a 5-point scale, ranging from strongly agree to strongly disagree. Each item was coded appropriately, with a high score indicating egalitarian sex-role orientations. Coefficient alpha for the scale was .55.
A measure of ideology focused more directly on the division of labor, family-labor ideology, was taken from responses to the question: "If a man and a wife both work full-time, they should share household tasks equally." Here, responses ranged across a 5-point scale, with a higher score indicating egalitarian family-role orientations.
Employment. The employment variable is based on a question regarding number of hours per week spent in the paid labor force. Our first intention was to separate the wives into three groups (employed full time, employed part time, and nonemployed), but for these data less than 10% of the women were employed fewer than 35 hours per week. We decided, therefore, to combine all 'working' women, and the employed group includes all women who report any hours of paid employment.
Perceptions of fairness. Wives' perception of fairness of the division of labor was measured with the following question: "How do you feel about the fairness in your relationship in each of the following areas? (household chores)." Responses were on a 5-point scale, ranging from "very unfair to me" to "very unfair to him." A code of 5 indicates "very unfair to my partner," 4 is "somewhat unfair to my partner," 3 corresponds to a response of "fair," 2 represents "somewhat unfair to me," and 1 is "very unfair to me." Since very few wives responding perceived the division of labor to be unfair to their partner (see below), we will refer to higher scores on this variable as indicating "fairness."
<O_>table&caption<O/>
Among the 496 employed wives in this sample, 61% perceive the division of labor to be fair, 28% "somewhat unfair to me," 6% "very unfair to me," and about 4% as either "somewhat" or "very" unfair to him; the mean is 2.64. Among the 282 nonemployed wives, 71% report a "fair" division of labor, 24% see it as "somewhat unfair to me," 2% as "very unfair to me," and 3% as somewhat or very unfair to him; the mean is 2.76.
RESULTS
Division of Labor
Table 1 presents the mean hours per week spent in household labor, by each spouse, across the eight household chores, as well as their percentage distribution. As shown, wives and husbands perform significantly different total amounts of labor and different types of labor. In those households in which the wife is employed, wives average 31.06 hours per week of total labor, while their husbands average only about 15.28. The disparity between male and female total labor is, of course, even greater within those households in which the wife is nonemployed. Here, wives perform approximately 42.04 hours per week of total labor, while their husbands average 12.23 hours per week. Obviously, the employment status of wives has a significant impact on their total labor in the home; their employment status has a substantially weaker association with husbands' household labor.
<#FROWN:J28\>Rather than seek integration of the best elements of each model, model-based professional organizations place behavioral scientist-practitioners in a multiple-choice stance in which they are encouraged to affiliate with one model.
Furthermore, many literature reviews, such as those published in Psychological Bulletin and Clinical Psychology Review organize their presentations around separate causal models and evaluate their comparative efficacy. This strategy also places causal models in a competitive stance and fails to acknowledge that behavior disorders may be a function of several classes and levels of causal variables, either concurrently or under different conditions, and that some variables and causal relationships in different causal models may be compatible and equally valid. As Blalock (1964) noted, this leads to "jurisdictional disputes" among proponents of models that are not necessarily incompatible.
Causal models can also be artificially contrasted by focusing on only one parameter of a behavior disorder (e.g., focusing only on the 'occurrence' of depression). This narrow focus is limiting in that it fails to recognize that different parameters of a disorder are affected by different variables. The cognitive model of depression, presented in chapters 1 and 2, provides a good example. An examination of the published longitudinal research might suggest that "beliefs of helplessness" should be minimized as an explanatory concept for the occurrence of depressive symptoms because the evidence that it reliably precedes the onset of depressive episodes is weak. However, if we focus on the magnitude or duration of depressive symptoms, the research is more supportive of a causal role for "learned helplessness" (Barnett & Gotlib, 1988).
'Alternative' causal models are sometimes simply different levels of explanation for the same phenomena. We can base a useful causal model for lamp illumination on either initiating electron flow through a resistor or on the behavior of throwing a light switch. Both can be useful, are not incompatible, and are simply different levels of explanation. Similarly, 'learned helplessness beliefs,', 'neurotransmitter deficits,' and 'social rejection experiences' need not be incompatible causal concepts for depression (Beck & Young, 1985; Boyd & Levis, 1980; Coyne, Kahn, & Gotlib, 1987; Lewinsohn & Hoberman, 1982). They may reflect only different labels, different measurement procedures, and different levels of foci for the same phenomena.
The relative validity or power of causal models may vary across domains or conditions. For example, between-spouse communication difficulties may be an important factor in marital distress, but significantly less so in the context of a supportive extended family. Similarly, the relative importance of diet and life stressors as determinants of hypertension risk may differ between European-Americans and African-Americans, or between younger and older adults. The various models, however, should not be viewed as necessarily competing but as potentially complimentary; elements of each may be necessary to account for causal relationships in different domains.
Unspecified Causal Mechanisms
Because causal inference demands a logical connection between variables, and because causal mechanisms are important in treatment decisions, the mechanisms of action of a causal variable (i.e., how or in what manner the causal variable exerts its influence) must be identified or at least amenable to reasonable hypotheses (Hyland, 1981). For example, we can hypothesize that aerobic conditioning can reduce premenstrual distress (Gannon, 1985), hypertension (Danforth, et al., 1990), depression (Doyne, Chambless, & Beutler, 1983), heart attacks (Dubbert, Rappaport, & Martin, 1987), and obesity (Brownell & Foreyt, 1985). But, as noted in chapter 1, it is important that the mechanisms of these hypothesized causal relationships be identified. 'How does exercise reduce premenstrual distress or hypertension?'
Causal relationships with unspecified mechanisms may have predictive and clinical efficacy. Consider the clinical utility of a reliably identified causal relationship between aerobic conditioning and hypertension, even if the causal mechanism is unknown. However, the utility of any hypothesized causal relationship is limited if the mechanism that underlies that relationship is not articulated. Knowledge of a causal mechanism helps us to identify the 'active components' of a causal relationship, to refine our intervention procedures, and to develop more effective and efficient interventions. Going back to our exercise example, it is useful to know that moderate levels of aerobic exercise can reduce premenstrual distress for some women (Gannon, 1985). However, if we also identify a possible causal mechanism for that relationship (e.g., that aerobic conditioning increases receptor-site sensitivity to endogenous opiates that may be released during premenstrual phases or affects prostaglandin release), we can possibly develop other interventions that operate through the same mechanism.
However, the question of 'how' an hypothesized causal variable affects the target behavior is too infrequently addressed in causal models of behavior disorders. How does childbirth sometimes lead to postpartum depression (Atkinson & Rickel, 1984)? How does a severe threat to self-esteem sometimes lead to paranoid ideation (Haynes, 1986b)? How does sexual abuse of a child sometimes lead to disturbed interpersonal relationships as an adult (Harter, Alexander, & Neimeyer, 1988)? How do self-efficacy beliefs sometimes influence a person's social interactions (Bandura, 1977a)? How does an 'accepting' and 'positive' client-therapist relationship sometimes lead to a reduction in interpersonal anxiety?
In summary, many causal models of behavior disorders have unspecified causal mechanisms. As a result, these models will (1) be less likely to evolve into more powerful models, (2) have a limited impact on our ability to explain behavior, and (3) have limited utility for the development of more effective interventions.
Unacknowledged Domains
As indicated in chapter 2, causal relationships are never unconditional. They have domains (i.e., boundaries or necessary conditions) outside of which the causal relationships are no longer valid. In psychopathology, these domains include population characteristics, internal states, developmental stages, variable values, temporal factors, and environmental contexts. Domains may apply to the types of causal variables related to the behavior disorder as well as the strength and form of causal relationships.
Many causal models of behavior disorders fail to specify their domains adequately. For example a causal model that suggests that self-efficacy beliefs affect the probability of posttreatment relapse (Marlatt, Baer, Donovan, & Kivlahan, 1988) should carefully delimit the social context in which that proposed relationship is applicable. Similarly, hormonal models of gender behavior should specify the environmental conditions in which hormonal factors will influence sexual orientation (Ellis and Ames, 1987). Also, it is important to stipulate the conditions under which stimulus pairings will produce a classically conditioned response (Rescorla, 1988; Papini & Bitterman, 1990), defensiveness and withdrawal will lead to long-term marital distress (Gottman & Krorkoff, 1989), a single exposure to a stimulus will produce a phobic response (McNally, 1987), and social support will mediate the impact of environmental stressors (Alloway & Bebbington, 1987).
Unacknowledged domains can lead to several conceptual and methodological errors. First, a causal relationship may erroneously be presumed to be absent. This can occur when the causal model is tested outside of its domain of operation. Second, domains can provide information about causal mechanisms and an unspecified domain restricts access to this information. For example, long-duration and short-duration stressors may have opposite effects on the immune system. Some studies have suggested a strengthening of the immune system during brief stress and a weakening during protracted stress (Miller, 1983), illustrating a 'chronicity' domain for the effects of the stressors. Examining causal relationships within these chronicity domains may lead to a better understanding of causal mechanisms involved in immune system deficiencies. Therefore, unacknowledged domains restrict the evolution of causal models.
Third, undefined domains for causal relationships can lead to inappropriate clinical applications. If negative outcome expectancies contribute to the maintenance of depression, but not its onset (a behavior disorder parameter domain), cognitive intervention efforts may be more effective if used during depressive episodes to reduce its duration than if used as a prevention strategy between depressive episodes.
Excessively High Level
As noted in chapters 1 and 2, causal models of a behavior disorder can be expressed at different levels. Furthermore, various levels of causal models can be clinically and empirically useful, depending on their intended application. Consider the utility of both political-sociological and biological models of post-traumatic stress disorders of Vietnam veterans. However, there is probably a hyperbolic functional relationship between the level of a causal model and its utility in that excessively low and excessively high levels often have diminished clinical utility for treating and preventing behavior disorders as well as diminished predictive and explanatory utility. However, behavioral scientists-practitioners err more frequently by proposing excessively high-level models.
As pointed out earlier the difficulty with a high-level causal variable is that it includes many lower-level variables. Therefore, higher-level variables and models do not permit distinction among the multiple possible causal variables and paths that they represent. For example, finding that 'marital distress' has a strong causal relationship to a behavior disorder, such as depression or alcoholism, has limited clinical utility because this variable does not identify the specific forms, mechanisms, or parameters responsible for the causal relationship. The active elements in 'marital distress' can include hostile verbal exchanges, physical abuse, sarcastic comments, defensiveness, withdrawal, infrequent presence in the home, feeling unloved, anxiety and emotional arousal, or lack of positive statements to the spouse. Therefore, attributing a behavior disorder to 'marital distress' provides only an array of possible causal variables and paths.
High-level behavior-disorder constructs are also problematic. A causal model for 'anxiety' is insufficiently specified because we do not know if the model applies to physiological, subjective, cognitive, or behavioral components of anxiety or to its onset, magnitude, or duration (Bernstein, Borkovec, & Coles, 1986).
Personality variables have played a prominent role in the behavioral sciences but are particularly vulnerable to the criticism of being dysfunctionally high level. Constructs such as 'locus of control,' 'hardiness,' 'authoritarianism,' 'need for dependency,' 'assertiveness,' 'self-esteem,' 'Type A behavior pattern,' 'emotional adjustment,' 'self-respect,' 'sexuality,' and 'need for achievement,' include so many possible lower-level variables that they are rendered scientifically and clinically debilitated. Personality variables suffer from an added disadvantage as causal variables because they usually imply stability across time and conditions and, therefore, are less useful as explanations of variance in the parameters of behavior disorders (Epstein, 1979, 1980).
Less molar causal variables are also open to this criticism. For example, it has been reliably demonstrated that there is a causal association between 'exercise' and a variety of behavior problems (e.g., response to stressors, hyper-tension). However, at a lower level of analysis, it is apparent that some types of exercise (i.e., weight lifting) can increase blood pressure while others (e.g., jogging) can decrease it; exercise early in the day can facilitate sleep onset while exercise later in the day can inhibit it, and the mediating effects of exercise on responses to psychosocial stressors may be related to the length of time a person has been exercising. Again, the concept of 'exercise' can be suggestive of causal relationships but requires lower-level specification to facilitate the design of intervention programs.
As with other elements of limited causal models, excessively high-level causal variables also promote inferential and measurement errors. Proposing a causal relationship between 'exercise' and response to psychological stressors, without more precisely describing the relationship, increases the chance that it will be measured at the wrong time of day, with the wrong sampling rate, or with the wrong assessment instruments.
Identifying higher-level causal variables can be a useful first step in psycho-pathology research and treatment programs. However, higher-level variables should be viewed as preliminary 'markers' for lower-level and more heuristic causal relationships. Subsequent inquiry can increase the chance of identifying causal variables and paths with greater clinical and empirical utility.
Linearity
Most causal models do not specify the mathematical form of their functional relationships. Consequently, most applications and tests of the models presume that the relationships are linear in form. However, as noted by umerousnumerous scholars and clinicians (e.g., Asher, 1976; Biddle & Marlin, 1987; Bishop, Fienberg, & Holland, 1975; Blalock, 1964; Bridgman, 1931; Grove & Andreasen, 1986, James et al., 1982; Miller, 1983), many functional relationships adhere more closely to parabolic, sine-wave, log, exponential, or other nonlinear forms.
The main drawback to an erroneous presumption that a causal relationship is linear in form is, again, an increased chance of an inferential error - underestimating the strength of a causal relationship. This can happen when statistical techniques based on a presumed linear relationship are applied to date that are nonlinear.
<#FROWN:J29\>
Conflict Talk:
Sociolinguistic Challenges to Self-Assertion and How Young Girls Meet Them
Amy Sheldon
University of Minnesota
Cultural stereotypes which interpret girls as less forceful or less assertive than boys in pursuing their own agendas, particularly during conflict episodes, are questioned. A theory of double-voice discourse is proposed to characterize a type of conflict talk that has a dual orientation, in which speakers negotiate their own agenda while simultaneously orienting toward the viewpoint of their partner. In double-voice discourse, self-assertion is enmeshed with addressee-oriented mitigation. Examples of such discourse in 3- and 4-year-old middle-class white girls' conflict talk are analyzed. Gender differences in conflict talk are seen to be contextualized variations that reflect differences in the organization of same-sex groups. Girls' use of the pretend frame in negotiating is discussed. Also highlighted is the importance of analyzing children's language in the context of conversational turns in order to develop a fuller interpretation of utterances.
This paper is an investigation of how young girls argue during social play, the kinds of tactics they use to further their own interests in disputes. As Maccoby (1986) pointed out, "we have a clearer picture of what girls' groups do not do than what they do do" (p.271). She called for "a more clearly delineated account of interaction in female social groups" (p. 271). Such an account requires description of the ways in which language creates and maintains social relationships among girls in everyday conversation.
A theory of double-voice discourse (name taken from Bahktin, 1929) is proposed to describe a linguistic style in white, middle-class, preschool girls' social interactions. The term points to a principle of dual orientation that shapes the agenda and style of their conflict talk. One of the speaker's orientations is toward her own agenda, toward the self. It asserts the speaker's own wishes and proposes activities that are in the speaker's interest. The other orientation is toward the other members of the group. As a result, self-assertion is enmeshed in an orientation toward the other. Self-assertion is thus regulated and contextualized by the speaker's relationship-centered orientation.
A theory of double-voice discourse captures the linguistic complexity and creativity with which young girls manage disputes. It also raises questions for traditional beliefs about femininity and masculinity, which treat gender as a polarity or a comparison of opposites. Such traditional views of gender are ill-conceived and inadequate to develop an account of girls' sociolinguistic interaction. Gender and context are confounded (Goodwin, 1980; Sheldon, 1990; Thorne, 1990). We can best see how talk is gendered if we take into consideration the context in which it emerges (e.g., the sex of the speakers and what the speakers are trying to accomplish). Considering children's talk from the perspective of gender has the advantage of raising methodological issues that focus attention on how we study children's language.
First, I will raise concerns about androcentric interpretations in some recent research on gender differences in children's talk. Next, I will describe the theory of double-voice discourse and relate it to children's solidarity-based task orientation. I will then discuss methodological advantages of a qualitative approach to the study of children's discourse and analyze examples of preschool girls' conflict talk to show how double-voice discourse can be read from them. I will conclude with remarks on the study of language and gender.
Gender and conflict. In most parts of American society women's conflict talk is constrained by the expectation that they will be 'nice.' 'Tough' talk, hard bargaining, and 'confrontational' talk is taboo for women. Men, however, have the license to argue in directly demanding ways. They can engage in unmitigated rivalry. Women are criticized if they speak as men do (Campbell, 1988, 1989; Coates, 1987b; Lakoff, 1975; Thorne, Kramarae, & Henley, 1983). Consequently, social talk which is considered ordinary 'assertive' talk for men is likely to be perceived as 'confrontational,' 'bossy' (or worse) when uttered by women. Does the talk of very young girls show evidence of the cultural taboo on tough talk? How do young girls express themselves in assertive ways while they keep to the cultural mandate that they not be 'too assertive'?
It is difficult to describe the full range of girls' and womens'women's talk without echoing stereotypical thinking about feminine and masculine behavior. Our thinking about conflict has an androcentric bias. As a society, we view aggression as the conflict norm. Equating conflict with aggression, however, does not fully capture how girls argue. The dispute management norm for girls is different from that for boys (Goodwin & Goodwin, 1987). Expecting conflict to have an aggressive component can prevent us from even noticing conflict in girls' groups in which a more subdues, yet nonetheless assertive, conflict style is common.
If we reconsider verbal conflict by focusing on feminine conflict styles, there are several interesting consequences. First, we must ask what agency, self-assertion, and power mean for females, because looking at conflict through an androcentric lens hides feminine agentic behavior and obscures the dynamics of feminine power. Second, we see the constructive and facilitative aspects of double-voice conflict, which is attentive to the social cost of self-assertion, in contrast to the constrictive and even destructive aspects of single-voice, baldly aggressive conflict, which is enacted with little regard to social cost. Finally, we gain a much clearer understanding of interaction in female groups.
Some problems with research on gender differences in children's conflict talk. In a review of the literature of gender differences in children's language and language development, Klann-Delius (1981) concluded that this area is in "dire need of being developed." Two recent quantitative studies have observed the way that young children's talk is gendered. Miller, Danaher, and Forbes (1986) studied over 1,000 quarrels by 24 racially and socioeconomically mixed 5- to 7-year-old children. They concluded that "boys are more concerned with and more forceful in pursuing their own agendae, and girls are more concerned with maintaining interpersonal harmony" (p. 543). What led the authors to the interpretation that boys were more "forceful" in trying to get their own way is that their dispute style was more heavy-handed than girls'. Boys used more threats and physical force. By contrast, girls were interpreted as more concerned with maintaining group harmony because they used more conflict-mitigating strategies, such as compromise, evasion, acquiescence, and clarification of intent.
This conclusion by Miller et al. (1986) is influenced by traditional views of gender hierarchy. It implies that girls are not as self-assertive as boys because (a) when measured against the masculine norm for conflict, which is coercive, girls do not use as much verbal brute force to get what they want; and (b) part of girls' agenda is to be empathic, and empathy presumably limits self-assertion. To describe the boys as "more forceful" in pursuing their own agenda is ambiguous as well as politically loaded. It is ambiguous because forceful means both 'effective' as well as "overpowering ... using force." It is political because it implies that girls are not as effective as boys in conflict situations; hence, girls are not good at verbally managing conflict, at furthering their own interests when opposed by someone. It is also a political conclusion because it values and emphasizes a masculine mode of brute force over a feminine mode of conflict mitigation. A different conclusion about feminine conflict style is possible, however; one that values the feminine conflict process and does not interpret it as weakness, as something 'less' than the masculine mode. From such a perspective, mitigation can be understood as functioning to tone down coercion and domination, to bring about adjustment and accord, and to restore group function. Girls and women are skillful at negotiating constructive conflicts. This view permits us to ask what important effects a constructive conflict process has on how girls' groups function.
To say that boys are "more forceful" persuaders, or "more assertive" (Sachs, 1987), overlooks the very important work that mitigation does to further self-assertion in the conflict process and reinforces cultural stereo-types that portray girls and women as submissive, ineffective, and weak. It equates effectiveness in conflicts with aggression. It measures self-assertion and independence too narrowly. To remedy this, I propose that the term coercive be used to describe only the heavy-handed conflict style. This frees the word forceful, meaning effective, to describe girls' (and boys') moderate conflict styles.
In another study of gender differences in children's talk, Leaper (1991) analyzed the conversations of 138 middle to upper-middle class 4- to 9-year-old children in either same- or mixed-sex dyads. Leaper's findings are consistent with those of Miller et al. (1986). Girls used more collaborative speech acts, defined as "direct" and "affiliative" (e.g., invitations to play, constructive offers, mutual affirmations). Boys used more controlling speech acts, defined as "direct" and "distancing" (e.g., insults, orders, refutations, and nonacceptance). Leaper (1991) hypothesizes,
Given that girls have been found to demonstrate more mutual coordination, responsivity, and elaboration in their conversations ... the female pairs were expected to use more affiliative speech acts ... and demonstrate more cooperative exchanges compared to male pairs. In contrast, since boys have been observed to be more demanding and domineering in their interactions than girls, ... male dyads were hypothesized to display more controlling speech acts and more domineering exchanges. (p. 800)
The girls and boys in these two studies actually used both self-assertive and supportive speech. But because there was a significant difference in the comparisons of interest, it is easy to lose sight of the similarities between girls and boys and to be left with the differences, despite the efforts of the authors to stress similarities. To address this problem in the study of conflict talk, disputes can be described, instead, in a way which better captures the complexity of girls' talk, thus revealing the imaginative and elaborate ways in which girls are self-assertive and powerful within the constraint of being relationship oriented.
Conflict is a contest of wills. Feminine conflict, because it requires the overlay of mitigation to avoid jeopardizing interpersonal harmony, asks for more sociolinguistic sensitivity than the more direct masculine conflict style. In fact, Sachs (1987) found that preschool girls have learned already how to assert themselves "with a smile." Camras (1984) studied what she identified as dominant and subordinate middle-class children in pre-school, kindergarten, and second grade. Dominant boys were much less polite than subordinate girls or boys, but they were also much less polite than dominant girls. Camras (1984) interprets these results as showing that dominant girls "are gradually socialized to mask their exercise of power during conflicts with use of polite language" (p. 263).
That girls and women are prescribed to assert themselves in a way that is responsive to others has been noted by Carol Gilligan (1987) and Jean Baker Miller (1986). This means that feminine agency functions in a different way than masculine agency, not that females are less agentic than males. Feminine self-assertion requires responsiveness to others, whereas masculine agency does not necessarily do so. The self-in-relation models of feminine groups proposed by Miller and by Gilligan also predict less hierarchical power relationships in female groups in contrast to the masculine model in which power is a relation of domination over others.
Double-voice discourse. Double-voice discourse is a talk style that is predicted by the self-in-relation model of feminine development (Gilligan, 1987; Miller, 1986). The term double refers to the perspective-taking stance of this style in which the speaker expresses a double orientation or double alignment. The primary orientation is to the self, to one's own agenda. The other orientation is to the members of the group. The orientation to others does not mean that the speaker necessarily acts in an altruistic, accommodating, or even self-sacrificing manner. It means, rather, that the speaker pays attention to the companion's point of view, even while pursuing her own agenda. As a result, the voice of the self is enmeshed with and regulated by the voice of the other.
Double-voice discourse is the norm in groups that are solidarity based. The best example of such groups are girls' (or women's) groups. Their social orientation is more often or more consistently relationship centered.
<#FROWN:J30\>
Recovery of traumatic memories leads to reintegration of the split and isolated multiples.
There is another aspect to Freud's early belief that the cure of neurosis lies in the remembrance of traumatic experiences. There is no other treatment that so stresses remembering, and there is no religion other than Judaism that makes a religious duty of remembrance of traumatic events. "You shall not forget that your forefathers were slaves in Egypt and you shall teach it to your children and to your children's children" is one of the cardinal commandments of Judaism. The Pass-over Seder is a dramatization of that traumatic event and the redemption from it, so that it will not be forgotten. The Jew must remember that his forefathers were slaves. Freud repudiated Judaism as a religion and consciously was an atheist who followed no religious practices or ceremonies; however, he never repudiated his identity as a Jew or his cultural adherence to Judaism. On the contrary, he was proud of it. I would suggest that the psychoanalytic emphasis on remembering as the essence of the cure was a return of the repressed or perhaps a return of the disavowed that was in part determined by the unconscious part of Freud's identity as a Jew. This, of course, does not affect the theoretical validity or the degree of practical utility of the cure through remembering, nor does it deny the clinical inspiration for the theory. Theories, like all psychological states and products, are, to use another Freudian concept, overdetermined; that is, they have many causes. The source of an idea has nothing to do with its value; to think so is to commit a genetic fallacy. After I wrote this, I came across Yosef Hyman Yerushalmi's brilliant and moving Freud's Moses: Judaism Terminable and Interminable (1991), in which he expresses a similar understanding of the origin of some of Freud's psychoanalytic theorizing.
Breuer, the third of Freud's spiritual fathers broke with him over the issue of sexuality. Love turned to hate, or, more accurately, the flip side of Freud's ambivalence toward fathers came to the fore, and Freud found it necessary to cross the street when he saw Breuer, his presence being so distasteful to Freud. There followed a period of lonely isolation during which Freud met and fell in love with Wilheim Fleiss, a charismatic Berlin internist to whom he was related by marriage. Freud was neither the first nor the last to be fascinated by Fleiss. Confident, successful, and uncritically admired by many, Fleiss was just what Freud needed. Brilliant, if erratic and eccentric in his ideas, Fleiss was receptive to Freud's otherwise and otherwhere unwelcome theorizing. Fleiss had a mesmerizing charm and was probably more than a little crazy. His theory that all illnesses were caused by nasal disorders, the nose being a sexual organ, has found little scientific support, nor has his belief that all natural phenomena could be accounted for by combinations and permutations of the female (28-day) and male (23-day) cycles. Fliess's pseudoscientific numerology probably owes an unconscious debt to cabalistic number mysticism - altogether, an unlikely consort for the Helmholtzian, scientifically rigorous Freud, but the heart has its reason, and a passionate relationship developed between the two men. Their contact was mostly through their correspondence, occasionally punctuated by congresses, Freud's term for their anxiously anticipated meetings, a term that suggests both grandiosity and sexuality. Reading Freud's side of their correspondence, which is all that has survived (Freud, 1985), we get a sense of intense intellectual excitement: here are two men approaching 40 who sound like adolescents who have just discovered the world of ideas, with all the passion and excitement that goes with that discovery. Of course, Fliess's excitement is an inference from Freud's letters, but it certainly appears to be there. Freud's letters to Fliess are a depiction of life of the educated Jewish middle class of late 19th-century Vienna that have all the vividness and richness of a great novel. Sentences filled with Freud's deep love of children alternate with sarcastic comments on his academic rivals, discussion of current political events, and theoretical 'drafts'. The overall effect is exhilarating. Freud's early theories about neurosis, anxiety, and the role of sexuality are all given trial balloons in the drafts he sent to Fliess. The most extensive of the drafts is Freud's 'Project for a Scientific Psychology' (1895/1950), which he abandoned and never published. It is a brilliant attempt to give a quantitative neurological explanation of psychological states and of psychopathology. It was Freud's last attempt to reduce psychology to physiology. Although he never abandoned the belief that a neurochemical explanation of mental events was possible, he himself turned to purely psychological explanations to account for both normal and pathological events. It is true that his psychological models and accounts retain a physicalistic basis, and much of Freud's theorizing is based on a "hydraulic model" of forces, pressures, flows, and blockages. It is a model based on 19th-century physics. It is also true that his theorizing becomes more and more a theory about meaning, and about relationships, and becomes truly psychological rather than pseudopsychological physics.
During Freud's almost two-decade-long relationship with Fliess, he suffered a "considerable psycho-neurosis" (Jones, 1961, p. 198) himself. Freud's emotional pain drove him to undertake his self-analysis, in which Fliess served as a sort of analyst by mail, and more important, was a transference figure eliciting all of Freud's intense feelings of love and hate for his father. Although it is unlikely that the two men were actually lovers, there is no question that Wilheim Fliess was the great love of Freud's life.
In the course of his self-analysis and his relationship with Fliess, Freud 'discovered' the Oedipus complex and wrote what is usually considered his most important work, Interpretation of Dreams (1900/1953a). In analyzing his dreams, Freud came to see that dreams have the same structure as symptoms. They too are disguised expressions of forbidden wishes. He concluded that all dreams are wish fulfillments. In the course of his self-analysis, he discovered much about himself: about his intense rivalry with and ambivalence toward his father; about his murderous feelings toward his infant brother, Julius; about his drivenness; and about his narcissistic vulnerability.
The dreams reported in Interpretation of Dreams make a unique contribution to the autobiographical literature of the West. They expand the account of self to include a new dimension. The self asleep - at least while dreaming - now becomes an integral part of self. Descartes's questions about distinguishing dreams and waking reality as a vital component of reality testing become irrelevant, and Locke's concern about the continuity of self during sleep is seen in a new light: dream consciousness is just as much consciousness, just as integral to the self, as waking consciousness. The injunction "Know Thyself" changes in meaning as the locus of self shifts to that which is not known, to the unconscious as represented in disguised and distorted forms in the dream. The self is now more unknown and unknowable, apart from undergoing the rigors of analysis, than hitherto believed. Freud's technique of dream analysis is double-edged: on the one hand, it gives us a tool for knowing the self; on the other hand, it reveals a new, unknown territory that must be reclaimed before the self can be either known or integral.
Having gone public in a unique, if partial and selective, way, Freud put an important part, by his lights the most important part, of himself up for scrutiny by any and all; and indeed his dreams have been interpreted and reinterpreted in a bewildering variety of ways, both from within and from without the psychoanalytic movement. One of the most fascinating perspectives on Freud's dreams is that of Carl Schorske (1980), who looks at their political meaning and significance and sees Freud as "regressing" from the political (adult's) to the familial (child's) world, from external reality to internal reality, because of the disintegration of the Austrian-Hungarian empire, its series of defeats in war, and growing dissension, corruption, and decadence; also, increasingly virulent anti-Semitism (Karl Lueger was installed as the anti-Semitic mayor of Vienna just as Interpretation was published) made action in the outer world increasingly futile and hopeless. Freud's dreams do indeed have many political references, and Freud like Plato before him takes the relation between social classes as representative of, or isomorphic to, the relationships of the parts of the psyche. Additionally, Freud's metaphors of self and mind are consistently political, and even sometimes military: defense, resistance, occupation, and drive.
Schorske interprets what Freud calls the manifest dream, the dream as dreamt, which Freud distinguishes from the latent dream , which is where his interest lies. In Freud's theory of the mechanism of dreams (which serves as a paradigm for his theory of mind in the sense of self) the dream thoughts that are forbidden childhood wishes, derivative of drives (instinctual energies) striving for discharge, are "converted" by the dream work into the manifest dream through the mechanisms of displacement, condensation, symbolization, visualization, and secondary revision. The task of dream interpretation is to work backwards from the manifest dream to the latent dream thoughts by listening to the dreamer's association to each dream element. Secondary revision is the mind's reworking of the dream material to give it more apparent sense and continuity than it possesses, that is, to give the dream a better story line. Dreams make use of current materials (the "day residue") but always equally, or more than equally, represent in distorted form the events and desires of childhood. Dreams are always egoistic. The censor imposes the dreamwork on the latent dream thoughts so they do not arouse so much anxiety as to wake the dreamer.
Freud has now moved from psychopathology to a normal psychological phenomenon, dreaming, and found that dreams are compromise formations in just the same way as symptoms. He is now in a position to expound a general psychology, an omni-applicable account of human nature. In the years following the Interpretation of Dreams, Freud went on to apply his paradigm to jokes, art, hallucinations, religion, and culture in general, finding each to have the same basic structure as compromises and disguised wish fulfillments.
In the famous "specimen dream of psychoanalysis", the dream of Irma's injection, Freud for the first time subjects a dream of his own to analysis. In the dream, the dreamer is in a large reception hall receiving guests, including Irma, who is a former patient who is still ill. By the time the dream ends, Irma's continued illness is blamed on at least three other persons, including one who represents Breuer, Freud interprets the dream wish as the desire to be blameless as well as to pay back some old scores. Irma in real life was Emme Eckstein, whom Fliess had operated on for "nasal neurosis" (which was plain madness), an intrusive application of his wild theory to a human being. To make matters worse, he left the packing in, which infected (long before antibiotics) and almost killed the patient, who suffered the torments of the damned and was given psychological interpretation of her difficulties by Freud. Freud told her that her symptoms were a holding onto her illness, which was a manifestation of her negative transference to him. Freud's dream was certainly an attempt to find himself guiltless by projecting blame for Irma's difficulties onto others, but Freud missed the main thrust, the deepest wish, behind the dream: to find Fliess blameless in order to protect his (Freud's) idealized love object from contamination and devaluation. Freud missed the motive power of our need for ideal objects, for perfect lovers with whom we can identify and perhaps merge. Fliess was such an ideal object for him. If Fliess was a transference object, as according to Freud's theory he had to be, then it was his father who was to be protected from the charge of injuring a woman. The childhood wish represented in distorted form in the dream was his wish that Father be perfect and blameless.
<#FROWN:J31\>
THE TIMING OF AN INTERPRETATION
A Comparative Review of an Aspect of the Theory of Therapeutic Technique
Lawrence Josephs
Interpretation in psychoanalysis is often considered to be as much an art as a science. And perhaps the most intuitive aspect of interpretive work, an aspect that is crucial to therapeutic outcome, is the timing of an interpretation. When do we interpret the transference? When do we interpret resistance? How deeply do we interpret? (And by 'deep' do we mean highly defended-against material, archaic fantasy, or early life historical events?) When would we simply empathize with the phenomenology of conscious self-experience and when should we confront and challenge resistance to discussing anxiety-laden unconscious conflict? By what criteria do we define an interpretation as premature or perhaps as too late? To what degree is the timing of an interpretation dependent on the presence of a prerequisite atmosphere of safety? How do we assess whether early interpretation of latent negative transference will either strengthen or impair the formation of a working alliance? The answers to these questions are partially based on the analyst's evaluation of the particular dynamics of the individual patient, yet to a large degree the answers to such questions derive from the analyst's theory of therapeutic technique. Although timing is in many respects dependent on the analyst's intuitive grasp of a unique analytic moment and is informed by past experience conducting analyses, the analyst's theory of therapeutic technique provides a preconscious sensibility that quietly guides the analyst's intuition as to the proper timing of an interpretation. In other words, intuition is partially derivative of a preconscious theory-derived rationale. It is preconscious in the sense that one's theory of therapeutic technique is not, for the most part, subject to repression, denial, or disavowal but nevertheless operates prereflectively, often without explicit conscious formulation.
For the purpose of this paper interpretation will be defined in the broadest sense as the provision of verbal feedback which aims at increasing the patient's self-understanding. Thus, clarification, empathic reflection, and confrontation will be considered as forms of interpretive activity in addition to the more narrow definition of interpretation that involves linking transference manifestations in the here and now to genetic constellations deriving from the there and then. The timing of an interpretation refers to the sequencing and pacing of one's interpretive activity and as such reflects a pre-conscious interpretive strategy; continually modified as it will be by the vicissitudes of the ongoing therapeutic interaction. The purpose of this paper will be to review, compare, and contrast recommendations concerning the timing of an interpretation that derive from diverse theories of therapeutic technique and offer an integrative overview in the service of resolving some of the disputes under discussion.
FOUNDATIONAL MODELS
The Topographic and Psychodynamic Approaches
The original theory of how to time an interpretation is to be found in Freud's (1911, 1912a, 1912b, 1913, 1914, 1915) seminal papers on technique which were based primarily on his topographic and psychodynamic models, having been written prior to the introduction of the structural model in 1923. Well-known psychoanalytic aphorisms informing the timing of an interpretation derive from these papers, including the principles that one should always work from surface to depth and that one should always interpret resistance before content. In these papers, resistance emerged as the key concept on which the assessment of the proper timing of an interpretation must be based. Resistance to free association was conceived of as a surface phenomenon that blocked access to greater depth. According to Freud (1914), the analyst "contents himself with studying whatever is present for the time being on the surface of the patient's mind, and he employs the art of interpretation mainly for the purpose of recognizing the resistances which appear there, and making them conscious to the patient" (p. 147). Resistance analysis was seen as the primary technique through which the analyst helped the unconscious to become conscious.
Freud discovered in his psychoanalytic work that resistance to the uncovering of unconscious contents was a ubiquitous phenomenon. Patients blocked introspective understanding of unconscious conflicts in inhibiting the flow of free association and they would not accept the analyst's interpretation of unconscious contents, be it forbidden wishes or painful memories, until the analyst had interpreted the resistance to the awareness of such warded-off contents. If one interpreted prematurely (i.e., too deeply) or too late (i.e., too superficially) resistance was exacerbated rather than ameliorated.
Expecially pertinent to the timing of an interpretation was a consideration of the dynamics of the transference. The art of correct timing largely focuses on the problem of overcoming the resistance to free association, and the major form of resistance is invariably to be found in the transference:
Over and over again, when we come near to a pathogenic complex, the portion of that complex which is capable of transference is first pushed forward into consciousness and defended with the greatest obstinacy .... These circumstances tend towards a situation in which finally every conflict has to be fought out in the sphere of transference. (Freud, 1912b, p. 104)
In terms of the relationship of the timing of an interpretation to the nature of the transference, Freud (1913) gave this recommendation:
So long as the patient's communications and ideas run on without any obstruction, the theme of transference should be left untouched. One must wait until the transference, which is the most delicate of all procedures, has become a resistance. (p. 139)
Yet, before transference can be interpreted as a resistance, there is a crucial preparatory phase that must occur if interpretation is to be effective at all:
When are we to begin making our communications to the patient? ... The answer can only be: Not until an effective transference has been established in the patient, a proper rapport with him. It remains the first aim of the treatment to attach him to it and to the person of the doctor. To ensure this, nothing need be done but to give him time. If one exhibits a serious interest in him, carefully clears away the resistances that crop up at the beginning and avoids making certain mistakes, he will of himself form such an attachment. (Freud, 1913, p. 139)
Part of what makes for an effective interpretation is not only the correct content of the interpretation but the nature of the transference at the time the interpretation is made: "The patient, however, only makes use of the instruction in so far as he is induced to do so by the transference; and it is for this reason that our first communication should be withheld until a strong transference has been established" (Freud, 1913, p. 144).
The Structural Viewpoint
Freud discovered that resistance analysis was no easy task. Although resistance is closer to the surface than the wishful impulses and painful memories whose expression resistance serves to inhibit, resistance itself is nevertheless a deeply unconscious phenomenon for which there is no ready access to awareness. In other words, there is always resistance to the awareness of resistance. In the structural model, resistance was traced to the unconscious defensive activities of the ego:
There can be no question but that this resistance emanates from his ego .... We have come upon something in the ego itself which is also unconscious, which behaves exactly like the repressed - that is, which produces powerful affects without itself being conscious and which requires special work before it can be made conscious. (Freud, 1923, p. 17)
The technical rule of interpreting resistance before content could be amended to read as a recommendation to analyze ego before id, ego defined as the unconscious mechanisms of defense.
Deviations in Technique
By the end of his career, Freud (1937) began to assess the limitations of psychoanalysis as a therapeutic modality. He had discovered that psychopathology and character structure were massively resistant to change, that patients were massively resistant to the awareness of their resistance to change, that whatever changes did transpire occurred very slowly and only with painstaking working-through of anxiety-laden conflictual issues, and that, even when change occurred, regressive developments could always undo it. Despite this tragic vision of human development Freud believed that his technical recommendations provided the optimum conditions for a 'talking cure.' Naturally, succeeding generations of analysts would attempt to improve results with technical innovations given the modest therapeutic expectations that Freud suggested.
If Freud's technical recommendations are treated as a baseline, it becomes apparent that technical innovations have deviated in basically two directions. On the one side, the central idea has been that classical technique is too conservative in addressing anxiety-laden unconscious conflicts and that resistance to the awareness of unconscious processes needs to be addressed more directly, vigorously, and systematically. On the other side, the central idea has been that classical technique has laid excessive emphasis on the analysis of defense, resistance, and latent transference in a context of abstinence, in neglect of establishing a therapeutic relationship and building psychic structure, activities that are seen as preceding and laying down the foundation for meaningful analysis of unconscious conflict. The debate has centered on matters of relative emphasis and priority, with the issues resting on a continuum from conservative to radical approaches. When classical technique has been seen as too conservative, the innovation has been to address latent transference more directly, actively, and systematically. Reich (1933) was one of the first innovators in that regard, chastising his peers for insufficient attention to latent negative transference and suggesting that nonverbal characterological resistances to treatment be consistently interpreted as manifestations of latent negative transference. Klein (1952) shared Reich's focus on systematically interpreting the latent negative transference, but felt that links to archaic unconscious fantasy could be established much more quickly than was the case utilizing classical technique. Searles (1979), Langs (1976), and Gill (1982) also advocated the primacy of interpretation of the latent negative transference, but felt that transference interpretation needed to be linked to the triggering stimuli in the here-and-now therapeutic interaction. The overall innovation in technique in all of the aforementioned suggestings is that unconscious conflict is most quickly brought to the fore through a primary and active interpretation of the latent negative transference. The difference in these approaches resides in whether transference interpretation is most efficaciously linked to character analysis, archaic fantasy, or current interpersonal transactions. The analyst's resistance to employing such innovations is usually traced to his or her fear of the patient's anger in confronting latent negative transference, fear of psychosis in interpreting archaic fantasy, and fear of admitting interpersonal involvement interpreting the here-and-now interaction.
When the critique of the classical approach is for insufficient attention to relationship building, technical recommendations center on interpretive strategies that are thought to facilitate a therapeutic relationship as well as greater emphasis on nonverbal affective attitudes that are thought to facilitate the proper ambience for treatment. Be it the ego psychologist's fostering of a working alliance and an observing ego, the object relations theorist's provision of containment or a holding environment, or the self psychologist's facilitation of a selfobject transference through empathy, the essential idea is the same: A therapeutic climate, which is partially curative in and of itself, must be established before conflict defense analysis can be fruitfully conducted. Verbal interpretation from this perspective tends to focus on the articulation of preconscious processes that are thought to be crucial in ego development and in the establishment of a cohesive sense of self.
Each approach claims to achieve maximal therapeutic benefit and possesses a critique of the inadequacies of other approaches. When the technical approach is seen as unnecessarily avoidant of addressing unconscious conflict and therefore superficial, the treatment is devalued as re-educative supportive psychotherapy that bolsters defenses rather than analyzes them. When the approach is seen as going too deep too quickly, the assumption is that treatment will result in excessive negative therapeutic reaction, intellectualized insight, false self-compliance, and a moralistic maturity ethic. The basic question in regard to interpretation boils down to one of timing: how deep to go how quickly? In working from surface to depth, must painstaking elucidation of the phenomenological surface precede deeper interpretation or is such an approach only working with the manifest content and avoiding a more active, direct, and systematic analysis of latent negative transference?
<#FROWN:J32\>
Chapter 4
The Frequency of Dirty Word Usage
How frequently are dirty words used in American culture? How often does one person use them throughout the day? Although many people in broadcasting, education, law, and social science make judgments about the frequency of dirty word usage, there are no sound data on which to make those decisions. If one examines the literature on general word frequency, one finds little information on dirty word usage. Entries in standard dictionaries are standard speech. Studies of word frequency tend to be studies of standard word frequency. One can examine colloquial dictionaries or those dedicated to argot, slang, or jargon. These provide no indication of how frequently such terms are used. A few minutes on a busy street corner convinces the listener that Americans use dirty words frequently in public but how frequently? This chapter is about determining the relative frequencies of dirty words in American English using standard empirical techniques. We will look at traditional means of establishing word frequency and some of the problems with establishing how often people swear. The conclusion from the available data is that these words are used relatively frequently.
Why Word Frequency?
For many years those interested in human communication and verbal behavior have known that words are used with different frequencies in communication. The notion of variable frequency has led investigators to study how frequency differences affect a wide variety of language-related activities (reading, creativity, language learning, problem solving, or comprehension). Given that dirty word usage is a behavior that most social scientists have constantly ignored, how can we tell how often dirty words are used? What impact does differential usage have?
One of the early concerns was the role of word frequency in both face-to-face communication and through electronic media like radio, short wave, or encoded channels. It was pointed out by Zipf (1949), for example, that relatively few words were used extremely often in communication and that the higher the frequency of occurrence a word had, the more likely it was to be used in general. Shannon and Weaver (1949) incorporated 'Zipf's Law' into their grand theory of communication, using word frequency to make predictions about the probability of various sayings or messages during communication. These predictions were important to produce efficient and quick communication and to eliminate errors by using predictable words. These findings were applied in military and government communication systems following World War II. However, the grand theory building that dominated psychology in the 1950's has died out and contemporary psychologists now focus on specific issues and problems involving human communication. Interest has shifted to how word frequency influences a specific type of language processing. The frequency question remains present in almost every aspect involving information processing abilities, such as attention, problem solving, pattern recognition, learning, rehabilitation, and memorization. Below are some of the recent applications of word frequency in contemporary investigations to show why the dirty word frequency question has remained unanswered.
Word frequency has been a fairly consistent predictor of response time differences in language processing tasks. High frequency words are processed faster and more easily than words at low frequencies. The result has been called 'the word frequency effect' in whatever process is under investigation. In a lexical decision task, for example, subjects are shown a string of letters and asked to tell whether the string forms a word. One of the best predictors of response time is how frequently the word is used; the higher the frequency, the lower the reaction time.
Word frequency effects have been reported in other language and memory tasks. In an object-naming task, where subjects are asked to name objects as fast as possible, results indicate that word frequency is related to the ease of naming the object. In word legibility tasks, frequent words were found to be as legible as single letters, although infrequent words are less legible than either. In lexical access tasks, word frequency effects the time it takes to decide if a target word has occurred in a sentence. Word frequency effects have been found in many memory tests of recall, recognition, and others (see Jay, 1980a). People who study reading and learning to read know that frequent or familiar words make information processing tasks easier and faster than obscure or low frequency words.
It should be clear that in each of these studies dirty words are never used. What is known about linguistic processes has had little or nothing to do with dirty words. If all science on language stopped now, we would know little about dirty word usage or how dirty word usage relates to more normal language use.
A recent question has been whether word frequency is related to the age of acquisition of a language by children. Carroll and White (1973 a,b) found that frequent words tend to be those that are acquired early in life. Another variable related to acquisition is the number of different meanings that a word has. One point here is that frequent words have more interpretations than low frequency words. So, age of acquisition depends both on meaning and frequency aspects of word usage. It is important to know what children learn early in life when examining patients with brain damage. Lesser (1978, pp 110-112), for example, shows that word frequency is needed to study aphasic (language loss) patients. Patients may use childish or highly frequent language when recovering from brain damage. Some of these aphasics may use only dirty words when recovering because the words were learned early and well, and may be stored in parts of the brain untouched by the trauma.
There should be no doubt that word frequency has an important influence on communication and communication problems. But, have we stopped to ask the question, where did all these frequency data come from in the first place?
The Frequency Estimation Problem: Why There Are No Dirty Words
There are hundreds of millions of speakers of English using tens of thousands of words on thousands of different occasions. To accurately count actual word usage is impossible. Therefore, estimates must be used. Throughout the history of frequency research, the question of proper frequency estimation has received minimal attention, and one wonders whether any of these previous estimates were accurate. While the tasks mentioned above have been conducted with appropriate methods, those interested in taboo language estimation would find them inadequate and inaccurate to estimate taboo language frequency. Future interest in controlling taboo language frequency or offensiveness requires an alternative to traditional methods and reports.
One major problem is the use of inappropriate normative samples of word frequency. Since its publication, psychologists and others interested in language frequency have been using the Thorndike and Lorge (1944) norms to estimate word frequency. However, the count (or others like it) are inadequate for three reasons: (a) the sample is restricted only to written, not oral usage, (b) the written sample is restricted to a limited domain of reading materials (mostly children's and popular adult literature), and (c) it is outdated (language changes with time). Eriksen (1963) has demonstrated quite clearly that the Thorndike-Lorge count is inadequate to estimate oral usage and that it in fact underestimates the frequency of many colloquially used words. Two collections of written word frequency norms were published by Kucera and Francis (1967) and more recently by Carroll, Davies, and Richman (1971). These are updated and less restricted than Thorndike-Lorge, but they are still limited to written samples.
The problem with written norms is not that they are inaccurate, but that they are used incorrectly by researchers. The written norms generally are appropriate when applied to tasks involving textual material; that is, reading processes. The written norms are not appropriate for research concerning oral usage, however. Written norms do not apply directly to colloquial, conversational situations such as parent-child interaction, discourse processes, or language influenced by sociolinguistic variables (social or physical setting). The point is that sociolinguistic variables are crucial to understanding dirty word usage. It is context that controls dirty word usage and these factors must be accounted for.
The bottom line on studies of written samples is that one is rarely going to find dirty words used in the sample because they are collected from biased material, even though that material may be appropriate to design children's reading texts, for example.
Counting Oral Frequency: Almost Good Enough
Written and oral speech are two different forms of language. In addition to different rules (e.g., the future perfect does not seem to exist in oral speech) and different distribution of rules they have in common (e.g., the perfect form of the verb appears much more in written speech), written language has the benefit of more polish. Oral speech is marked by hesitations, interruptions, incomplete expressions, is more prone to imprecise or incorrect definitions - all of which can be corrected with a little proofreading in the written form.
Oral speech is also at the mercy of a number of sociolinguistic influences. Spoken language, particularly in its more colloquial form, is more sensitive to the relation between speaker and listener, the degree of social relaxation, and the topic of discussion (see Jay, 1978c). When these variables <{_>flucuatefluctuate, so does the kind of language selected and used in conversation. In light of these contextual constraints, then, the estimation of oral frequency requires looking at oral - not written - samples of language for taboo speech.
Several attempts have been made to examine the frequency of word usage in conversational English (Jay, 1980a). One widely mentioned study is French, Carter, and Koenig's (1930) collection of words and sounds from telephone conversations. This study is cited because it provides information about vowel, consonant, and word frequency data. The study has one major flaw. They omitted some 25% of the data representing utterances such as exclamations, interjections, proper names, titles, letters, numbers, and profanity. In fact profanity was 40% of the material omitted! While the study could give a relatively accurate picture of conversations, it compromised a true picture of dirty word usage. Fairbanks (1944) updated the French et al. study and compared the spoken language samples of college students with diagnosed schizophrenics. More recently, Black, Stratton, Nichols, and Chavez (1985) published a word count based on college student classroom language, as did Berger in 1968.
Several studies concentrate on children's oral language. The need for these data in reading, learning, and comprehension applications should be clear. These data serve the purpose of designing age-appropriate textbooks but do not indicate anything about how children use dirty words or how often. One of the best collections and most current counts is Spoken Words (1984) by Hall, Nagy, and Linn. These investigators report the data as a function of situation and social status of the children's family. However, when speech was recorded in classes or at the dinner table, very few dirty words were spoken. Dirty words are highly context dependent and even young children have learned not to use them most of the time at school or at the dinner table. Thus few of them appear in the corpus. While these norms provide a useful foundation for writing or evaluating children's reading texts, as intended, they give the impression that children do not produce dirty words. Consequently, there is a risk of thinking children are much more naive about matters involving sex and aggression than they are, if the conclusion is based on biased word counts. Cameron (1969) made one of the more natural attempts to collect adult language data as a function of social setting. His sample, though restricted to a few college settings, is useful to indicate how speech changes from relaxed to formal social environments. The major problem with the Cameron norms stems from an inadequate sampling procedure. He asked his 'overhearer' to record by hand "the first three words they heard during the conversation at 15 second intervals" (p 102). Such recording is subject to constraints from attention, perception, or recording bias, especially when overhearers knew he was interested in certain types of words.
<#FROWN:J33\>The most successful representations make fairly specific assumptions about the way in which the text should be interpreted (e.g., as a story about one of a small number of topics) and about the kinds of questions that might be asked about the text (e.g., questions about actor's intentions). Some work has been done to adapt the natural language understanding techniques to an information retrieval setting, but there is little near-term hope that these techniques could be used to represent large document collections and arbitrary queries [Sparck Jones and Tait, 1984].
Within the information retrieval community, a number of techniques have been developed that can represent the content of documents and information needs. These representations have a much different flavor than NLP representations. They are generally based on simple, very general, features of documents (e.g., words, citations) and represent simple relationships between features (e.g., phrases) and between documents (e.g., two documents cite the same document). The focus here is on simple, but general, representations that can be applied to most texts rather than on specialized techniques which capture more information but are applicable only in narrow contexts. Information retrieval representations also make extensive use of the statistical properties of representations features and attempt to make use of information produced by human analysis (e.g., manual indexing) when available.
Over the last decade there has been considerable interaction between the AI and information retrieval communities; AI techniques have been adapted to an IR setting and the IR focus on 'real' document collections and on thorough experimental evaluation has helped to expand the focus of AI research.
Given the availability of a number of representation techniques that capture some of the meaning of a document or information need, our basic premise is that decisions about which documents match an information need should make use of as many of the representation forms as practical. The remainder of this paper develops a theoretical framework for retrieval that allows multiple representations to be combined.
In the next section, we describe the major types of retrieval models. Section 7.3 presents the motivation for a retrieval model based on inference. In Section 7.4, we review related research on inference and network models. Sections 7.5 and 7.6 describe the basic inference network model and how it is used. Section 7.7 addresses the issue of causality in a network model. The final section discusses recent results and research directions.
7.2 Current Retrieval Models
A retrieval model fixes the details of the representations used for documents and information needs, describes how these are generated from available descriptions, and how they are compared. If the model has a clear theoretical basis we call it a formal retrieval model; if the model makes little or no appeal to an underlying theory we call it ad hoc. We use the terms theory and model here in the mathematical or logical sense in which a theory refers to a set of axioms and inference rules that allow derivation of new theorems. A model is an embodiment of the theory in which we define the set of objects about which assertions can be made and restrict the ways in which classes of objects can interact.
Four retrieval models are particularly important in IR research: the Boolean, cluster-based, probabilistic, and vector-space models. Most of the commercial and prototype IR systems currently available are based on some variation of these models, and some understanding of them is necessary for our discussion of the inference net model in the next sections.
Boolean. Boolean retrieval forms the basis of most major commercial retrieval services, but is generally believed to be difficult to use and has poor recall and precision performance since the model does not rank documents. In the Boolean model we have a finite set of representation concepts or features R = {r<sb_>1<sb/>, ..., r<sb_>k<sb/>} that can be assigned to documents. A document is simply an assignment of representation concepts and this assignment is often represented by a binary-valued vector of length k. The assignment of a representation concept r<sb_>i<sb/> to a document is represented by setting the i<sp_>th<sp/> element of the vector to true. All elements corresponding to features not assigned to a document are set to false.
An information need is described by a Boolean expression in which operands are representation concepts. Any document whose set of representation concepts represents an assignment that satisfies the Boolean expression is deemed to match the information need, all other documents fail to match the information need. This evaluation partitions the set of documents, but provides no information about the relative likelihood that documents within the same partition will match the information need.
Relevance in Boolean retrieval, then, is defined in terms of satisfiability of a first-order logic expression given a set of document representations as axioms. Several attempts have been made to extend the basic Boolean model to provide document ranking.
Cluster-based retrieval. Cluster-based retrieval is based on the Cluster Hypothesis which asserts that similar documents will match the same information needs [van Rijsbergen, 1979]. Rather than comparing representations of individual documents to the representation of the information need, we first form clusters of documents using any of several clustering algorithms and similarity measures. For each cluster, we then create an 'average' or representative document and compare this cluster representative to the information need to determine which clusters best match. We then retrieve the clusters that are most likely to match the information need rather than the individual documents. There are several ways to identify the clusters to be retrieved, particularly when using hierarchical clustering techniques that allow navigation of the cluster hierarchy.
Since many techniques are used to compare the query with the cluster representative, there is no single definition of relevance for cluster-based retrieval. Rather, relevance is partially defined by the model that forms the basis of the comparison. The similarity measures used to define clusters and the method used to create the cluster representatives also play a part in defining relevance since they determine which documents will be judged similar to a cluster representative that matches the information need.
Vector-space retrieval. In the vector-space model, we have a set of representation concepts or features R = r<sb_>1<sb/>,..., r<sb_>k<sb/>}. Documents and queries are represented as vectors of length k in which each element corresponds to a real-valued weight assigned to an element of the representation set. Several techniques have been used to compute these weights, the most common being tf.idf weights which are based on the frequency of a term in a single document (tf) and its frequency in the entire collection (idf). These tf.idf weights are discussed in more detail in [Salton and McGill, 1983].
Documents and queries are compared using any of several similarity functions, the most common function being the cosine of the angle between their representation vectors.
Since several techniques have been used to compute weights for the vector elements, the vector-space model has no single form of document or query representation, although all representations have a common form. Similarly, since several similarity functions have been used, relevance has no single definition.
The vector-space model is historically important since it forms the basis for a large body of retrieval research that can be traced back to the 1960's. The vector-space model has been criticized as an ad hoc model since there is relatively little theoretical justification for many of its variations.
Probabilistic retrieval. Probabilistic retrieval is based on the Probability Ranking Principal which asserts that the best overall retrieval effectiveness will be achieved when documents are ranked in decreasing order of probability of relevance [Robertson, 1977].
There are several different probabilistic formulations which differ mainly in the way in which they estimate the probability of relevance. Using a representative model, a document d<sb_>i<sb/> and an information need f<sb_>j<sb/> are represented as the now familiar vectors of length k in which each element is true if the corresponding representation concept is assigned to the document or query. If we let F represent the set of representations for information needs and D represent the set of document representations, then we can define an event space F x D and our task becomes one of determining which of these document-request pairs would be judged relevant, that is, estimating P(R|d<*_>i<*/>,f<sb_>j<sb/>). We then use Bayes' theorem and a set of independence assumptions about the distribution of representation concepts in the documents and queries to derive a ranking function that computes P(R|d<*_>i<*/>,f<sb_>j<sb/>) in terms of the probabilities that individual representation concepts will be assigned to relevant and non-relevant documents. Different independence assumptions lead to different forms of the model. Given estimates for these two probabilities (say, from a sample of documents judged relevant and from the entire collection), we can compute P(R|d<*_>i<*/>,f<sb_>j<sb/>).
Probabilistic models are in many ways similar to the vector-space model [Bookstein, 1982; Turtle and Croft, 1992]. Both can be considered to be generalizations of the Boolean model in that they can support partial matching using Boolean queries. Probabilistic models, however, provide a sounder theoretical base for the design of IR systems, and have significantly contributed to our understanding of some aspects of IR, such as term weighting, ranking, and relevance feedback.
7.3 Retrieval Based on Inference and Networks
Recent retrieval research has suggested that significant improvements in retrieval performance will require techniques that, in some sense, 'understand' the content of documents and queries [van Rijsbergen, 1986; Croft, 1987] and can be used to infer probable relationships between documents and queries. In this view, information retrieval is an inference or evidential reasoning process in which we estimate the probability that a user's information need, expressed as one or more queries, is met given a document as 'evidence.'
The idea that retrieval is an inference or evidential reasoning process is not new. Cooper's logical relevance [Cooper, 1971] is based on deductive relationships between representations of documents and information needs. Wilson's situational relevance [Wilson, 1973] extends this notion to incorporate inductive or uncertain inference based on the degree to which documents support information needs. The techniques required to support these kinds of inference are similar to those used in expert systems that must reason with uncertain information. A number of competing inference models have been developed for these kinds of expert systems and several of these models can be adapted to the document retrieval task.
In this paper, we describe a retrieval model based on inference networks. This model is intended to
Support the use of multiple document representation schemes. Research has shown that a given query will retrieve different documents when applied to different representations, even when the average retrieval performance achieved with each representation is the same. Katzer, for example, found little overlap in documents retrieved using seven different representations, but found that documents retrieved by multiple representations were likely to be relevant [Katzer et al., 1982]. Similar results have been obtained when comparing term- with cluster-based representations [Croft and Harper, 1979] and term- with citation-based representations [Fox et al., 1988].
Allow results from different queries and query types to be combined. Given a single natural language description of an information need, different searchers will formulate different queries to represent that need and will retrieve different documents, even when average performance is the same for each searcher [McGill et al., 1979; Katzer et al., 1982]. Again, documents retrieved by multiple searchers are more likely to be relevant. A description of an information need can be used to generate several query representations (e.g., probabilistic, Boolean), each using a different query strategy and each capturing different aspects of the information need. These different search strategies are known to retrieve different documents for the same underlying information need [Croft, 1987].
Facilitate flexible matching between the terms or concepts mentioned in queries and those assigned to documents. The poor match between the vocabulary used to express queries and the vocabulary used to represent documents appears to be a major cause of poor recall [Furnas et al., 1987]. Recall can be improved using domain knowledge to match query and representation concepts without significantly degrading precision.
The resulting formal retrieval model integrates several previous models (probabilistic, Boolean, and cluster-based) in a single theoretical framework [Turtle, 1990].
<#FROWN:J34\>
Manipulating the +/- UG distinction in terms of Coppieters' percentage figures turns out to be something of a vacuous exercise, however, since many of the percentages are based on small numbers of exemplars, thus lowering their face validity. Recall that the percentages for NNS-NS deviance are derived from varying numbers of tokens of each linguistic variable. At one extreme of NNS-NS divergences are the variables Imparfait vs. Pass Compos (39.5% divergence) and <*_>a-acute<*/> vs. de + Infinitive (34.7%), which are exemplified by 5 tokens and 2 tokens, respectively. At the other extreme is the A-over-A variable (14.4%), exemplified by 6 tokens. As numbers of tokens decrease, percentage figures become less valid indices of response patterns. It would be a simple matter, for example, to deflate or inflate the deviance for the <*_>a-acute<*/> vs. de + Infinitive category by changing (or adding, or subtracting) a single well-chosen item. Clark 1973 has demonstrated the dangers of inference from few tokens of a given category: there is no evidence that the findings will generalize beyond the items sampled and apply to the category as a whole.
A contingency in data elicitation further clouds the issue of locus of competence differences. For some 41 of the 107 items, subjects were given a choice between paired forms (e.g. Est-ce que tu {as su/savais} conduire dans la neige? 'Did you {manage/know how} to drive in the snow?'). They were then asked by the experimenter to decide: (1) whether both forms were acceptable; (2) if so, whether a difference in meaning was entailed in the differing forms; (3) if so, "what that difference consisted of in the framework of sentence interpretation" (550). These 41 items were tokens of five linguistic variables: the Imparfait/Pass Compos distinction, il/elle vs. ce, prepositions <*_>a-acute<*/> and de + Infinitive, prenominal vs. postnominal adjectives, and article use. The other elicitation procedure required subjects merely "to provide straightforward well-formedness judgments" (550) for single (unpaired) sentences. This was done for "most of the other 66 sentences" (550), representing four linguistic categories: the Object + Predicate construction, Causatives and Clitics, A-over-A constraint, and the de + Adjective construction. Reviewing the data in Table 1, above, one may discern a striking pattern. For the top five variables - just those variables tested by the first technique - NNS diverged dramatically from the NS (NNS-NS average = 29.2%) and the range of NNS-NS divergence is disparate (19.9% - 39.5%). For those variables tested under the second procedure, on the other hand, the NNS-NS divergence is relatively small (average = 17.3%), with the range clustering tightly (14.4% - 18.9%). Comparing these figures with the deviance figures for the +/- UG distinction (average for putatively - UG items = 27.2%, range = 16.9% - 39.5%; average for putatively + UG items = 17.4%, range = 14.4%-18.9%), one notes that the magnitude and heterogeneity of NNS-NS divergence are somewhat better predicted by the procedural contingency than by Coppieters' version of the +/- UG distinction.
The issue of elicitation procedure has special significance for de + Adjective structure. Coppieters registers his surprise that NNS should diverge so minimally from NS (16.9%) on such a "comparatively obscure" construction (561). It is an ad hoc move to invoke frequency as a predictor of divergence for this structure alone. Left unmentioned is the fact that minimal divergences for this structure are an embarrassment for the +/- UG distinction as a general predictor of NNS-NS differences. As it happens, the de + Adjective construction was tested under the second procedure; for it and the other structures tested in this manner, NNS-NS divergences were small. There is no anomaly to account for if one allows for the possibility that the asymmetry in NNS-NS differences is an artifact of a procedural asymmetry. Coppieters is apparently aware of the procedural asymmetry, but does not attribute differential results to it (553-54).
Though one can only speculate about whether differences in procedure are in fact causally related to differences in results, it is nevertheless easy to see how there might be a relationship. The first procedure involved a three-part, fine-grained decision task in the context of a face-to-face interview with the experimenter. In the second procedure, subjects were asked merely to give nominal statements of well-formedness. In the latter instance, it is unlikely that subtle cues could be communicated by the experimenter to the subjects. However, that likelihood necessarily increases in an elaborated discussion of form, meaning, and interpretation, as was the case in the first procedure. I offer this observation about the relative possibility of experimenter effects as a statement of objective fact about experimental procedure (see Heringer 1970, Labov 1975, Rosenthal 1966), not to dispute Coppieters' assertion that he reinforced all subjects in their judgments and gave no indication when judgments diverged from native norms (552-53). Experimenter effects can be unintentional and unnoticed; measures should be taken to reduce their likelihood.
More to the point, one must consider the nature of the data resulting from the two different elicitation procedures. Recall that the results displayed in Fig. 1 and the two left columns of Table 1 represent cumulative departures from NS norms, the norms being in the form of nominal statements of acceptability or grammaticality. The second procedure generates just such nominal statements for each item, and departures from these norms are straightforwardly tallied. The first procedure, by contrast, involves direct comparisons of two sentences. In this case, the subjects declare that the sentences are equally acceptable, or that they prefer one sentence over the other. Thus in one procedure judgments of acceptability reflect assessments against some external criterion of acceptability, while in the other procedure the judgments reflect comparisons of one sentence with another. A further difference between the procedures should also be noted. The first procedure, in eliciting three-part answers, generates complex commentaries on subtle interactions of syntax and semantics. Anyone who has ever elicited (or offered) such comments recognizes that they may contain references to imagined contexts, statements reflecting judgments based on degrees of well-formedness, expressions of ambivalence, etc. This type of data may be contrasted with the simple nominal statements ('correct or good'/'incorrect or bad') generated by the second procedure. (Moreover, what is meant by 'good' and 'bad' is not spelled out for participants, creating the possibility of intersubject variability as to the basis of judgment.) In light of multiple differences between the two procedures, one must recognize that their results are inherently dissimilar and therefore not comparable. The two types of data should not be lumped together as if they were identical.
However, Coppieters seems to do just that. The results of the first procedure, in order to be considered relevant input for the NNS-NS comparisons displayed in Fig. 1 and Table 1, have apparently been converted into nominal data. The details of this conversion operation are not spelled out. How does one establish a unique NS norm, and nominal departures from it, when the task generates a complex response? By what algorithm is a comparison of two related items, with an accompanying three-part, fine-grained discussion, transformed into an independent statement of acceptability for each of the items? These questions speak to the reproducibility of the results, as well as to their validity.
With respect to the more general question of whether there are NNS-NS competence differences, Coppieters' methodology is again problematic. Recall that the differences he attests are derived from comparisons of NS and NNS deviations from a prototypical native norm. The norms represent the majority of NS nominal judgments. For 90 of the 107 items, there was 80% or more agreement in judgment among NS. For the other 17 items, "the majority opinion of NS's" (553) must be understood as any level of agreement above 50%. Clearly, it would have been desirable to establish norms at a high level of NS agreement for all items. Even if this had been the case, one must question the use of nominal data to establish norms. Nominal judgments are insensitive to degrees of grammaticality or acceptability. Chaudron 1983 has argued that scaled judgments of grammaticality (e.g. 1 = totally unacceptable/ungrammatical; 5 = perfectly acceptable/grammatical) are more informative than nominal data, inasmuch as they permit intergroup comparisons (e.g. t-tests and ANOVAs) on the basis of means and variance from means. The psychological validity of scalar grammaticality statements is supported by Barsalou 1987 and others working in category/concept theory. The category of well-formedness is susceptible to gradedness effects in experimental performance, just like other categories such as BIRD (robins and sparrows are judged more birdlike than emus, ostriches, penguins, and apteryxes). Moreover, Chomsky 1986 has argued for relative grammaticality in syntax as a function of the number of barriers crossed in extractions. In sum, there are numerous reasons for preferring scalar responses to nominal statements in determining differences between NS and NNS.
The composition of the subject groups in the Coppieters study presents additional problems. To qualify for participation, nonnative subjects had to perform at a near-native level of speech, as determined by native-speaker friends of the subjects and by the experimenter himself. In addition, by Coppieters' reckoning, all performed orally at a level of 'superior' by ACTFL criteria. (For a description of the ACTFL oral proficiency interview, see ACTFL 1986.) However, both of these screening procedures are imprecise and nondeterministic. Informal near-native ratings cannot assure native-like performance across a variety of linguistic structures. It is quite possible to sound 'native-like' by avoiding production of complex structures which are not fully mastered. It is also possible to sound 'native-like' in restricted contexts, such as bank transactions or philosophy lectures. Indeed, for restricted and practiced contexts, there may be apparent pragmatic mastery, while for other contexts there may be manifest inadequacies. The second screening measure leaves a great deal of latitude for composing the near-native group, inasmuch as the ACTFL category of 'superior' ranges between the 3rd and 5th levels of a 5-point scale. By virtue of these screening procedures, it is quite possible that the near-native group was not particularly proficient. It is not clear that Coppieters satisfies the desideratum of Long 1990 for investigating the general question of competence differences, viz., that at least some subjects sampled be among "the very best" L2 learners.
The biographical profiles of NNS and NS subjects pose further problems. As has been documented by Ross 1979 and others, one may expect differences in grammaticality judgments - even among native speakers - if sociological and educational variables are not controlled. Coppieters' 21 NNS were "highly educated" (551) and included 19 professors or researchers, eleven of whom were professors of linguistics or literature. In contrast, his 20 NS had merely "had some education" (551). Alone, such background differences could account for a good deal of the NNS-NS deviance. That is, failure to control for background differences between the NNS and NS increases the chances that the two groups will not coincide on judgments for "subtle areas" (Long 1990:281) of the grammar, just as NS or NNS with heterogeneous backgrounds would be likely to disagree among themselves.
The risks of heterogeneity are also evident in the native language backgrounds of the NNS. Recall that Coppieters' subjects are from a variety of disparate linguistic backgrounds. In principle, this design feature would allow one to determine whether results generalize beyond speakers of a single native language. However, with only 2 subjects representing Farsi, only 5 subjects representing 'Oriental' languages, and 6 different native languages among the 14 native speakers of Germanic and Romance, one cannot assume validity of results for the individual native language groups, much less generalize across groups.
Because of these shortcomings, and others that will come to light in later sections of this paper, the Coppieters study must be viewed with some skepticism. At the very least, it should be evident that the two main questions of maturational effects in L2A - whether there are competence differences between NNS and NS, and if there are, which linguistic domains are affected - remain open. Further examination of the issues is therefore warranted. As a first step, it is important to determine whether the principal findings of Coppieters can be replicated with methodological modifications and under conditions of tighter procedural control.
<#FROWN:J35\>The other sources for the Eastern dialect give forms containing [ts] or [<*_>c-hacek<*/>] rather than [ks]. These include Lean and Mulia, the two archaic sources cited by Lehmann (1920:656), who give //lochac//, presumably [lo<*_>c-hacek<*/>ak], and all of the modern sources, e.g., Dennis and Royce (1983:14), who give lots'ak<sp_>h<sp/>. Campbell and Oltrogge (1980:219) reconstruct *lots'ak. The [ks] reported by some sources is likely an erroneous transcription, due to lack of familiarity with ejectives. Moreover, none of the sources contains any word of the form sak or sax that might be what Greenberg considers to be the second member of the form, or anything like ala-k 'glowing' or lak 'glow'. Nor do these forms appear in the notebook. The notebook contains no headword 'glow', and no entry for Jicaque under the headword 'shine', nor has inspection of all of the Jicaque entries yielded any plausible sources for these forms. The only entries under 'sun' are loksak, latsak, and loksaki. It thus appears that there is no evidence whatsoever internal to Jicaque for analyzing laksak into lak-sak. In other words, the comparison is between two words, each containing the sequence lak (if the Jicaque contains a [k] at all), in neither of which there is the slightest evidence for an analysis in which lak is a morpheme. In contrast, Greenberg fails to mention the etymology proposed by Mason (1918:19), namely, that ma<*_>unch<*/>alak is derived from ma<*_>unch<*/>al 'burn, blaze' (cf. ma<*_>unch<*/>ale 'flames'). The weakness in Mason's etymology is that no suffix -ak is known, but, there being evidence for one of the two morphemes, it is nonetheless more plausible than Greenberg's.
246 STONE<sb_>2<sb/>
LIA cites M. i<*_>unch<*/>ak and i<*_>unch<*/>ik 'knife' and asak'a 'flint'. The last is attested but is indicated by Mason (132) to be Antonia<*_>n-tilde<*/>o. Neither of the forms for 'knife' appears in the notebook, which gives M. <*_>c-hacek<*/>a:k, A. <*_>c-hacek<*/>ik<sp_>h<sp/>, <O_>formula<O/>, and <O_>formula<O/>. Nor do they appear in Mason (130), who gives M. <*_>c-hacek<*/>ak, A. <O_>formula<O/>. The form i<*_>unch<*/>ak is cited by Greenberg and Swadesh (1953:219). The forms in LIA as well as the form in Greenberg and Swadesh (1953) appear to derive from Sapir (1917:8), who gives the forms (i)<*_>unch<*/>a:k and (i)<*_>unch<*/>ik. In the footnote to this entry, Sapir cites the form as <*_>unch<*/>ak, suggesting that the forms with the initial i may be typographical errors, with i in place of the articular prefix t, the parentheses intended to segregate it from the stem.
255 THROAT
LIA lists A.p-e:nik'a. The notebook has pe:nik'a:i, which is the form given by Mason (127). The basis for segregating the /p/ is unknown.
4.2. Grammatical evidence. Chapter 5, 'Grammatical Evidence for Amerind,' cites Salinan data in eleven places, sections 6, 13, 16, 19, 45, 74, 80, 84, 88, 90, 100. Of these, the following five call for comment.
In section 19 (p. 288) LIA gives the Miguele<*_>n-tilde<*/>o first-person plural pronoun as ka. The correct form, according to Mason (32) is <O_>formula<O/>.
Section 80 (p. 311) reads in its entirety:
The following etymology represents a past-tense marker in HOKAN: Yuru-mangu<*_>i-acute<*/> iba, Coahuilteco pa-, Tequistlatec -pa, Salinan be, Pomo (hi)ba, Karok -<*_>unch<*/>ipa<*_>approximate-sign<*/>pa- (near past), and Shasta p'- (distant past, habitual past). The marker probably also occurs in Salinan iwa-<*_>unch<*/>, which when suffixed to nouns means 'that which was formerly', e.g. noq<*_>unch<*/> 'head', <O_>formula<O/> 'skull'. The last element, -<*_>unch<*/>, is a common noun formant in Chumash. The agreement in a form *ipa among Yurumangu<*_>i-acute<*/> in the extreme south, Coahuilteco in the middle, and Karok in the far north is striking.
This passage presents a number of difficulties. First, Salinan be is not a simple past tense morpheme as Greenberg glosses it. Mason (1918:35) glosses be as 'when, definite past time', as in such examples as be:-ya 'when I went'. Sapir (1920:307) suggests that be:ya is really "an indicative e:ya 'I went' subordinated by the demonstrative stem pe, pa 'the, that'." Sapir's view is supported by Mason's statement that "Pure sonant b has been found only in the case of the demonstrative article pe..." (1918:11). If Sapir is right, be is not a tense morpheme at all.
Second, the agreement among the forms is perhaps less striking than Greenberg suggests. Of the eight forms cited, only four (Karok, Pomo, Yurumangu<*_>i-acute<*/>, Salinan iwa<*_>unch<*/>) have the initial i, as we shall see, the initial i of iwa<*_>unch<*/> is probably not original. Moreover, some of these morphemes are prefixes while others are suffixes. Finally, as discussed below in 5.3, it is unclear whether the Yurumangu<*_>i-acute<*/> past tense morpheme is the infix -iba- or the suffix -bai.
Third, the discussion of iwa<*_>unch<*/> is seriously flawed. To begin with, this suffix and the examples cited are not Salinan. The suffix iwa<*_>unch<*/> is not mentioned in any source on Salinan that I have consulted, nor are the examples cited as illustrating the use of this suffix (noq<*_>unch<*/> 'head', <O_>formula<O/> 'skull'). Nor do they appear under the headings 'head' and 'skull' for Salinan in Greenberg's Hokan notebook. Indeed, Salinan has no [q] at all.
Rather, the suffix iwa<*_>unch<*/> and the examples of its use are Chumash. noq<*_>unch<*/> and <O_>formula<O/> are among the examples of the use of the suffix iwa<*_>unch<*/> in Barbare<*_>n-tilde<*/>o Chumash given by Beeler (1976:259), one of the sources listed in the notebook. The notebook lists noks under 'head' for Inese<*_>n-tilde<*/>o and Barbare<*_>n-tilde<*/>o Chumash.
Nonetheless, the issue of the cognation of Chumash iwa<*_>unch<*/> with the other forms cited still arises. Two facts militate against Greenberg's analysis. First, the initial i is very likely not original. There are two related suffixes in Chumash, the suffix iwa<*_>unch<*/> which derives nouns meaning 'dead, defunct, former', and the verbal past tense wa<*_>unch<*/>. The nominal suffix is invariant, as is the verbal suffix in Barbare<*_>n-tilde<*/>o (Beeler 1976). In Inese<*_>n-tilde<*/>, however, according to Applegate (1972:102-3), an epenthetic copy of the last vowel of the stem is inserted before the past tense marker when the stem ends in a sonorant. Although the history is not known, it seems very likely that the nominal suffix is historically derived from the verbal past tense, and that the initial i represents a frozen epenthetic vowel.
Second, there is little basis for segmenting out the final <*_>unch<*/>. If iwa<*_>unch<*/> is derived from the verbal suffix wa<*_>unch<*/>, it is unlikely that the final <*_>unch<*/> is a noun-forming suffix. Noun-forming suffix <*_>unch<*/> is not described in any source on Chumash that I have consulted, including the two most detailed grammatical descriptions, Applegate (1972) for Inese<*_>n-tilde<*/>o and Beeler (1976) for Barbare<*_>n-tilde<*/>o. There is a suffix -V<*_>unch<*/> described by both Beeler (1976:258) and Applegate (1972:213) as "resultative," and it is to this suffix that Greenberg refers. The V means that the suffix takes the form of a vowel followed by <*_>unch<*/>. According to Beeler (1976) this vowel is unpredictable, but Applegate (1972:93) gives rules for predicting the vowel. In any case, the derivation envisioned by Greenberg is not clear. If we start from a verbal past tense suffix *ipa, which in Chumash comes to be added to nouns as well as to verbs, it would seem unnecessary to add a resultative suffix, and one has to wonder why the basic past tense suffix would acquire the resultative suffix attached to it in nouns.
In section 84 (p. 311), LIA cites a suffix -<*_>unch<*/>e 'desiderative' for Antonia<*_>n-tilde<*/>o. Mason discusses such a suffix, but with final glottal stop <O_>formula<O/>, on page 49, and although he glosses it 'desiderative', he makes it clear that he is far from confident of this interpretation, a hesitation fully justified by the varied meanings of the examples adduced.
In section 88 (p. 313), LIA refers to a Salinan imperative morpheme -i-. This suffix is not described by Mason in his discussion of the verbal morphology (1918:34-54), nor in Turner (1987), so that it appears to be quite spurious, and given the lack of documentation, one despairs of tracking it down. The key turns out to be Rivet (1942:33), which presents the same equation as in LIA, minus the Karok form, which was evidently added by Greenberg. Rivet cites Sapir (1921:71), in which we find, as entry number 28, the following:
Sal. -i-, imperative suffix with third person pronominal object (e.g. m-alel-i-k ASK HIM!): Yana -'i', imperative suffix.
This constitutes the entire discussion of imperative -i- in the literature. At the very least Greenberg is to be taken to task for using as evidence a morpheme for which the evidence is so skimpy, especially when it does not appear in Mason (1918), the only published grammatical description of the language and the source of Sapir's data.
Whether Sapir's analysis should be accepted is unclear. Salinan has a number of third-person singular objective suffixes, which according to Mason (1918:46-47) take the form -o, -ko, -xo, and -k. The different suffixes are associated with different classes of verb: "... the -p prefix nearly invariably takes the suffix -o or -ko as its third personal objective form while the objective form in -k occurs exclusively with the -k prefix" (Mason 1918:39). That is, the forms in -k are associated with what Mason considered to be the intransitive verbs, later argued by Sapir (1920:307-8; 1921:69-70) to be stative. In his discussion of the imperative, Mason (1918:41) states that : "The imperative takes its third person pronominal object in -ik, never in -o or -ko."
Sapir's reasoning appears to have been that the i of this suffix could be isolated as an imperative since the suffix appears only in imperatives and since other third-person singular objective suffixes contain k. Against this we may consider the fact that imperative i occurs nowhere else and that it is not possible to analyze the third-person singular objective morpheme simply as k; even if we could extract k from ko, we would be left with the forms in o. Moreover, judging from Mason's examples of the use of -ik, namely k-<*_>a-acute<*/>mamp-ik 'take it out!', <*_>a-acute<*/>mes-ik 'shout to him!', and m-alel-ik 'ask him!', it appears that -ik occurs even with active verbs, which according to Mason's generalization do not take the objective prefixes containing k. On balance, it seems that -ik must be treated as a unit, and that Sapir's analysis is overly aggressive.
In section 90 (p. 313) LIA cites k- as the imperative prefix. Actually, according to Mason (1918:41) this prefix appears only in the plural, and not in all cases.
5. Yurumangu<*_>i-acute<*/>. Yurumangu<*_>i-acute<*/> forms occur in a total of twenty-six entries in LIA, to wit: H15, H32, H35, H44, H61, H69, H72, H87, H142, H158, H161, A9, A79, A86, A102, A104, A191, A243, A269, G19, G61, G80, G88, G90, G102.
5.1. Segmentation. Morphological analysis is often difficult even in languages for which we have unlimited data; in a case like Yurumangu<*_>i-acute<*/>, where we have only a tiny corpus containing few related forms, we must proceed with the utmost caution and must expect to remain unsure on many points.
Rivet made a valiant effort at analysis on the basis of the data available to him, using both what little language-internal evidence there was, and what was suggested to him by similar forms in Hokan languages. Many of these proposals are interesting and would be worthy of pursuit if we were convinced of the affiliation of Yurumangu<*_>i-acute<*/> with Hokan, but until such an affiliation is proved they remain the merest speculation.
The result of the dearth of Yurumangu<*_>i-acute<*/> data combined with the fecundity of Rivet's etymological imagination is that most of the morphological analyses he proposed rest on hypothetical affiliations with Hokan, and so remain undemonstrated. Rather than exercising caution and utilizing only justifiable analyses, Greenberg simply accepts Rivet's analyses and presents them as if they were clearly justified internal to Yurumangu<*_>i-acute<*/>, as we shall now see in detail.
Rivet (1942:28-29) posits a prefix a-, appearing on both verbs and nouns. Other than the fact that a considerable number of words begin with a, he cites no language-internal evidence for the existence of such a prefix. In many cases he gives no evidence of morphological complexity of any kind. Where he does give evidence, it consists of comparisons with other languages.
<#FROWN:J36\>
One
Palestine and the
Coming of the Cold War,
January 1947 - March 1947
During the winter of 1947, the future of Palestine became immersed in East-West policy considerations extending far beyond the Eastern Mediterranean. Britain's sudden retreat from empire enabled American Zionists and the U.S. State Department to exploit rising war jitters as they competed with one another in the mass media to create a new conventional wisdom on Palestine. The American Zionist Emergency Council (AZEC) insisted American involvement in Greece and Turkey should be followed by support for a Jewish state. The State Department's Near East desk argued that policy would only lead to Communist gains throughout the oil-rich region. Both sides attempted to use the American mass media, particularly the New York Times, to win converts to their cause. They competed in describing a vision of reality that would serve as the context for policymaking. The beginning of this struggle to shape conventional wisdom at the outset of the Cold War and the role the mass media played in that conflict is the subject of this chapter.
Americans reading the daily press and listening to the nation's network commentators in the opening days of 1947 had reason to believe that the world was entering an era of peace. Although a state of war still technically existed, President Harry S. Truman on December 31, 1946 declared the hostilities of World War II officially terminated. Emergency laws exercised by the Executive Branch since 1941 were ended. A presidential proclamation pledged the United States would work with other nations in building a world in which justice would replace force.
The New York Times reported that American Secretary of State James F. Byrnes and Soviet Foreign Minister Vyacheslav M. Molotov had just completed negotiations in New York that amicably resolved differences on more than forty issues. United Nations Secretary General Trygve Lie indicated in his New Year's message that a "sound foundation" had been laid and he looked to the future with "sober confidence." The chief United States delegate to the United Nations, Warren R. Austin, told a radio audience that he was convinced the United States and the Soviet Union "had come to a better understanding of each other" and that he had "great faith" that a start had been made on the "long road away from war." The nation's foremost radio commentator, Hans V. Kaltenborn of the National Broadcasting Company (NBC), caught this spirit of optimism when he told listeners, "the days ahead will confirm mankind's steady march to better things." The United States, Kaltenborn was convinced, had "seized the torch of freedom and held it aloft for others to see and follow."
The New Year's cheer proved short-lived. The British, bankrupted by the war and wracked by labor problems, abruptly announced in February a pullback from their commitments in the Eastern Mediterranean. The collapse of the British economy and the Labour government's sudden fallback from empire jolted New York Times publisher and board chairman Arthur Hays Sulzberger, who remained convinced that the "common good" of the free world required a strong Britain. The prospect of a disintegrating British empire had Times editorial writers fearing the emergence of a world in which the United States stood alone in defense of democratic freedoms everywhere under attack. Kaltenborn shared this sense of sudden foreboding. Gone was the "new age of peace" that seemed to have been dawning only weeks before. The British cabinet's decision to reduce Britain's armed forces by 340.000 men, plus the acute shortages of food, fuel and clothing throughout the United Kingdom, created a "dangerous situation," Kaltenborn believed, in the postwar world.
Britain's pullback from the Eastern Mediterranean exposed misgivings long held by members of the American media and the State Department over the Soviet Union's postwar aims. In the fifteen months that followed, the world was rhetorically split into two hostile camps, one Soviet, the other American, separated by "deep distrust and mutual suspicion." Palestine, a country with a violent history of its own, became absorbed within the East-West war of words when Britain, eager to trim a costly 100,000 man peacekeeping force, announced in late February it was turning the Palestine problem over to the United Nations. It was these events that set the stage for a mass-mediated battle for public opinion at the beginning of the Cold War.
JOURNALISM'S FIRST DRAFT ON HISTORY
The early postwar rhetoric of Kaltenborn and Sulzberger reflects a journalism of self-promotion and self-conscious public service, if not self-abnegation. Each man intended to play a bigger part in policymaking than either was permitted during the war years. Born in Milwaukee, Kaltenborn joined the Fourth Wisconsin Voluntary Infantry to fight in the Spanish-American War. In the two decades that followed he hustled stories for the German language press in Wisconsin before joining the staff of the Milwaukee Journal, and later Herbert F. Gunnison's Brooklin Daily Eagle. On April 4, 1922 Kaltenborn presented a radio address at the Army Signal Corps station in New York City, which years later he billed as "the first editorial opinion over the air." He made news when he climbed Mt. Fuji in 1927 and visited the Soviet Union two years later. In 1932 he covered the Democratic National Convention for the Columbia Broadcasting System (CBS), later interviewed Hitler and Mussolini, reported on the Spanish Civil War and the appeasement at Munich.
Kaltenborn considered himself a "contemporary historian" and saw his fifteen-minute newscasts for NBC during the war as first drafts of history. NBC nourished this image and on October 16, 1941 organized a town meeting in which Kaltenborn answered questions about the war from two hundred of San Francisco's "outstanding citizens." NBC also arranged for a newsreel camaraman to be present. Subsequently, Kaltenborn appeared weekly in newsreels shown across the country. During the first six months of the war, more than one million theatre goers asked for Kaltenborn's views. While he admitted that he "didn't know all the answers," he believed it was his responsibility to "guide public opinion.
Both Kaltenborn and Sulzberger saw the press as a major player in postwar policymaking. Kaltenborn observed that the responsibility to lead opinion belonged to the press as much as to those "who place their hands over the Bible and swear to do their duty in public office." Each made democracy possible by their public service. Sulzberger told the nation's leading editors that how the press handled its "sacred and special mission" might well "determine the destiny of the world." The war had shown too clearly, Sulzberger claimed, that only "an informed democracy" was strong enough to survive in a world where totalitarianism was on the march.
During the war the press had been an agreeable, if at times, reluctant partner in reporting the Roosevelt administration's war news and views. Part of the reason was that the press generally supported the war effort. In April, 1942, Kaltenborn endorsed the idea that total war required a cooperative press in April 1942. "We want freedom of speech and press for patriotic Americans whose one concern is to win the war," he noted. "We want silence from all others." But as the war went on, many in the press wearied in the role of messenger service for the Roosevelt administration. Radio newsmen in May, 1943 went public with a complain that government censors often left "the Japs better informed than Americans." Wire service reporters were convinced a false image of the war was being communicated through the government's determination "to hold back and play down American casualties." The American Society of Newspaper Editors, representing the nation's largest dailies, resented the growth of "pernicious propaganda" disseminated by government bureaucrats. It charged in April, 1944 that news management had confused the American people.
Arthur Krock, the Times veteran Washington correspondent, believed the war and the controversy growing out of government efforts to manage the news would alter the policymaking environment in the immediate postwar period. Krock thought the press would be less likely to accept the administration's call for a non-partisan foreign policy. He wrote that the State Department would be incapable of managing postwar foreign policy without significant participation by the Congress and the American people. Times management shared Krock's conviction. Sulzberger conceived of the Times as an "American institution" called upon to preserve the country's fragile freedoms through vigorous editorial crusading. Charles Merz, the paper's editorial page chief, determined to "stir the American people and the Congress to their responsibilities." Sunday editor Lester Markel saw that responsibility as "educating public opinion" to what diplomat correspondent James Reston described as "the changes and convulsions in the world in which America must operate."
The determination of Times management to participate in postwar policymaking was a significant departure from the days when Adolph S. Ochs seriously considered scrapping the paper's editorial page. The Ohio country boy turned grocer's clerk and druggist's apprentice told an interviewer that he could get a larger circulation by printing a newspaper with all advertisements and no news rather than a paper with all news and no advertisements. His work as an assistant in the composing room of the Knoxville (Tenn.) Tribune and as staff writer and editor of the Chattanooga Times convinced him that "news which told the exact truth so far as possible" also made good economic sense. In 1896 he became publisher and controlling owner of the New York Times, then a struggling enterprise in the era of the yellow press dominated by Joseph Pulitzer and William Randolph Hearst. Ochs' determination to cultivate an elite, monied readership resulted in a twenty fold increase in Times earnings during the first generation of his stewardship, and it led to the Times' dominance "within its own sphere of usefulness." Ochs reasoned that only one in 100 readers read all the news that Times' editors saw fit to print. But that one "would tell the other ninety-nine" and the Times would get the reputation of being the "complete newspaper."
Sulzberger saw interpretation and leading elite opinion as central to the postwar role of the Times. Sulzberger had been a member of a noteworthy Jewish family that featured philanthropists, scholars and jurists, when he married Ochs' only daughter in 1917, positioning himself as the heir apparent within the Times hierarchy. On Ochs' death in 1935, Sulzberger became publisher, president, and chairman of the board at the Times and wielded his power to pick associates who shared his commitment to a journalism of "public service." Sulzberger saw protecting the nation's freedom from totalitarian intrusions at the center of that service. He warned fellow editors that the press was like the canary miners took down into the shafts with them. "It fell over at the least sign of poisonous gas", Sulzberger pointed out, and this warning gave others a chance to escape. While the Times primarily considered Soviet foreign policy before and during the World War II in terms of that nation's security interests, by the winter of 1947 it was beginning to smell poison gas. "Communists and fellow travelers" represented a threat to the American way of life, Sulzberger told the nation's editors in April, 1947; the publisher who "knowingly employed a communist or any other type of totalitarian" or gave him "any place of influence" in news or editorial departments "threatened the United States itself." The defeat of Nazi Germany did not mean fascism was dead, Times diplomatic correspondent Herbert L. Matthews wrote, expressing this new orthodoxy. Communist authoritarianism was a "Red Fascism."
Sulzberger and the Times were seen as an important target for those within the Truman administration who were pressing for a get tough policy toward the Soviet Union. In December, 1945 Navy Secretary James V. Forrestal told Sulzberger and Brooks Atkinson, the paper's Moscow correspondent, that "the only thing /the Russians) recognize is stark force." Forrestal's case to Sulzberger was that the Russians had "no respect for the normal human weaknesses, such as justice, kindliness and affection." At a cabinet meeting three weeks later Forrestal urged President Truman to call an emergency meeting of Sulzberger and other leading newspapermen, as well Kaltenborn and the nation's leading radio commentators, impressing on them "the seriousness of the present situation."
<#FROWN:J37\>During these sessions, the PDRY was consistently reported in Sanaa as arguing that the transitional period between the announcement of unity and the holding of national elections should be extended beyond the six months stipulated in the draft constitution. PDRY leaders reportedly feared that six months was not enough time to overcome their regime's poor image in the eyes of the electorate. There was also speculation that both the YAR and PDRY leaders feared that opposition to the draft constitution was growing, so they wanted to accelerate the unification process to deny opponents time to organize.
Predictably, the issue of assigning senior posts in the new government proved difficult. Throughout the government structure, the principle of a 50-50 split of positions, and of assigning northern deputies to southern department heads and vice versa, was maintained. The bargaining sessions were long and at times rancorous as officials jockeyed endlessly to ensure themselves favorable spots in the new hierarchy. Salih and Bidh were named president and vice president, respectively, of the five-man Presidential Council provided for in the draft unity constitution; the PDRY head of state, Haydar Abu Bakr al-Attas, became prime minister of a 39-member cabinet, and PDRY prime minister Said Yasin Numan was installed as speaker of the parliament. The unity agreement said that sufficient time should be allowed for transition to permanent unity and specified a two and one-half year, rather than a six-month, transition period. The constitutional referendum called for at Aden and stipulated in the draft constitution was not mentioned.
The new state was something of a Potemkin facade. The situation was symbolized by the new flags seen around Sanaa, from which it was obvious that the green star of the YAR had simply been removed to leave the unadorned red, white, and black tricolor of the new state. The 'united' ministries tended to be divided into two camps of rival northern and southern officials who were more inclined to keep to themselves than to integrate with their new colleagues. These divisions were particularly sharp in bodies such as the Education Ministry, where northern and southern ideas differed sharply on fundamental substantive issues such as curriculum. As more than one senior official admitted, the complex problems of unifying the PDRY's socialist economy with the free market system of the YAR were at first swept under the rug with the two economies left to function largely as they had done in the past. In the field of defense, the general staffs were successfully integrated into a unified defense ministry, and a few units were moved from south to north and vice versa. Unification at the rank-and-file level, however, was left to take place over a protracted period of time, in order not to sacrifice military efficiency.
The haste with which the two sides came together had its costs in the early months of unity. The low salaries that had been paid to southern officials were at first not raised quickly enough to keep pace with the cost of living, which rose rapidly toward northern levels despite the maintenance of some consumer goods price subsidies. This problem, combined with administrative failure to meet payrolls in Aden, created severe hardship and short-term discontent in the south. Another cause of friction was a hasty attempt to roll back the land reform program under which the PDRY government had distributed agricultural land in the years following independence from Great Britain. The dislocations caused by this effort led to demonstrations in at least one southern province. At the same time, the government was slow to deliver on promises to restore real estate confiscated from businessmen who had fled north from Aden following southern independence. This failure hurt efforts to induce these businessmen to reinvest in the stagnant southern economy.
Despite these difficulties, however, the merger took place without the major disruptions some had predicted, and the small size of demonstrations mounted in Sanaa on 22 May against the union was a fair measure of the weakness of anti-unity sentiment in the country. Officials noted at that time that their strategy was to capitalize on the enthusiasm for unity that existed in 1990 and not allow the formidable practical problems of unification to cause them to miss this major opportunity. In retrospect, and notwithstanding the severe problems the Republic of Yemen has faced in its first two years, their calculation appears sound.
THE POLITICS OF UNIFICATION
There were several reasons why the leaders in Sanaa and Aden were able to overcome a legacy of mutual suspicion and fear created by 22 years of politico-military combat and to unify Yemen in less than six months. The first and foremost was the sea change in the international political climate brought about by the Soviet Union's acceptance of change in Eastern Europe in the late 1980s and the implications for the Aden regime of Moscow's new attitude toward its friends elsewhere in the world. The southern leadership doubtless felt compelled to alter radically its political course and to strike the best deal possible with Sanaa as quickly as possible.
The other reasons for the success of the unification effort are interrelated. In the 1980s, the consolidation of the authority of the YAR government in tribal areas left Salih, whose personal stake in a successful unity initiative was high, greater room for maneuver in unity negotiations than he had previously enjoyed. Salih's greater maneuverability vis-<*_>a-acute<*/>-vis the tribes grew out of the evolution of the complex Yemeni-Saudi relationship, which was an important element influencing the success of the initiative. To understand the radical transformation in the attitude of the two sides toward one another, the political and economic positions of Aden and Sanaa in November 1989 must be considered.
Aden: Fall 1989, Nowhere to Turn
By the fall of 1989, the regime in Aden had very limited options. The bloodletting of January 1986 was a repudiation of Ali Nasir Muhammad and his effort to seek an opening to the West, seen by many as the key to the PDRY's economic salvation. Also gone from the scene, however, were the most prominent of Ali Nasir Muhammad's hard-line opponents. Although they were eulogized as 'martyred' heroes, the country's new leaders showed little enthusiasm for returning the country to a radical communist course.
PDRY politics following the January events were essentially stagnant. The regime was divided into factions opposed to one another along regional as well as ideological lines, with hard-line North Yemeni members of the former NDF pitted against moderates, many of whom came from Hadhramaut Province. These splits sharply limited the regime's ability to develop a clear policy direction. The only success the new PDRY leadership could claim was the rapprochement with the YAR that produced the joint investment zone, the proliferation of other less significant agreements with Sanaa, and the gradual diminution of the status of Ali Nasir Muhammad and his exiles in the YAR.
In social terms at least, the regime had some reason for self-congratulation. Between 1967 and 1985, illiteracy was reduced from 97 to 59 percent. In the YAR, the illiteracy rate stood at 80 percent in 1985. In achieving this progress, the PDRY relied largely on native, well-motivated teachers whereas the YAR depended heavily on Egyptians. The PDRY could also point with pride to its liberal family law that gave women rights unparalleled in most Islamic countries.
The economy, by contrast, was largely dysfunctional. Land and fisheries reforms had produced declines in agricultural production, and the port of Aden, once one of the busiest in the world, played almost no role in international commerce. Industrial production had declined, and, despite 20 years of socialism, more than 50 percent of the gross national product still came from the private sector, while hard currency remittances from workers abroad accounted for half the government's annual budget.
Nowhere was the bankruptcy of the PDRY's system more apparent than in the petroleum sector. The discovery of oil in the YAR by the US-based Hunt Oil company in 1984 had led to exports of 200,000 barrels of oil a day five years later. A more or less contemporaneous Soviet discovery just across the border in the PDRY had put Aden $500 million in debt with nothing to show for it other than 60 or so mostly non-functioning wells, a pipeline suspected of leaking, an outmoded oil processing complex, and a few barrels of crude oil trucked sporadically to Aden's decrepit refinery. Ali Nasir Muhammad's opening to the West had brought several Western companies into the PDRY, but none had made a discovery.
Against this somber backdrop came the changes in Moscow's policy toward Eastern Europe. The political message that Moscow was unwilling to stand behind the regimes there was complemented by widespread reports in the latter part of 1989 that the Soviet military and economic aid that had been vital to the Aden government's survival was to be drastically reduced. Exactly what the Soviets said to the Aden leaders about aid levels, and when they said it, is not well known. There is little doubt, however, that such cuts were an imminent prospect.
The PDRY leaders' strategy for improving their bleak situation was to move cautiously in the direction of political and economic reform. In the latter part of 1989, the government organized local council elections in which candidates outside the ruling Yemen Socialist Party were allowed to run; a high percentage of non-party members were elected. The government also appeared to reach a tacit understanding with newspaper editors permitting freer journalistic expression. On the economic front, discussions began with the Soviet Union in 1989 aimed at making available to Western companies a substantial portion of the area in Shabwa Province where the Soviets had been looking for oil. The government also adopted a law intended to encourage domestic and foreign private investment and to permit investors to take a percentage of their profits out of the PDRY. Finally, wealthy Saudis of South Yemeni origin, as well as southern businessmen who had moved to the YAR, were invited to make investments in the PDRY economy.
Aden's limited economic liberalization efforts, however, proved a failure. Whatever negotiations there had been with the Soviets about the release of oil acreage seemed to be going nowhere. Soviet exploration in Shabwa largely ceased, and the date for the opening of the pipeline from Shabwa to the coast was continually postponed. Efforts to attract private investment languished as businessmen waited until the fate of the regime became clear.
The response of South Yemenis, like that of investors, made it plain that the regime's political strategy was too little too late. The spring of 1990 saw considerable ferment in the media - in contrast to the silence of the still highly 'disciplined' fourth estate in the YAR - as well as the first stirrings of political demonstrations and strikes. South Yemenis also made clear their strong pro-unity feelings. When President Salih went to Aden in November 1989, he was greeted by crowds calling for unity whose enthusiasm was obviously not staged. At the same time, some of the political parties that had been active in both the YAR and the PDRY in the early 1960s began to resurface in the PDRY. While these parties were insignificant as political forces, their appearance reflected the popular interest in genuine liberalization.
Pressure for unification from South Yemenis continued in the spring of 1990. By then, the leadership in Aden, short of other alternatives, appeared to have reached the conclusion that unification was its best option. In theory, it could have continued to resist pressures for unification, opting instead for more radical political change, for a renewed effort to open the oil fields to Western companies, and for strategms designed to avoid the need both to play second fiddle to the YAR leadership and to share their oil wealth with Sanaa. Their decision against such actions and instead to respond favorably to Salih's unity initiative almost certainly stemmed from a calculation that popular feeling in the PDRY both against them and in favor of unity was too strong to permit them to stay in power.
<#FROWN:J38\>
10.2 Connections between City Growth and Economic Development
Compared with Third World averages, India did not develop rapidly between 1960 and 1981. Much of the difference can be attributed to far lower total factor productivity growth rates (chap. 7). Indeed, while poor productivity growth can account for only a portion of India's slow city growth experience after 1964 (chap. 6), it can account for a great deal of the poor economic performance overall.
Underlying India's dismal productivity growth performance are structural and intrasectoral distortions. To some degree, these could be reduced by appropriate rural policies (notably, providing insurance incentives to encourage peasant farmers to switch to higher yield but riskier crop mixes: see Singh, 1979) and industry-oriented packages (in particular, by removing location and input pricing distortions created by current policies: see Sekhar, 1983). In the presence of such distortions, traditional policies that are oriented toward capital formation will have only a modest impact, mainly because returns are low. This was the key flaw of the Mahalanobis strategy (chap. 9). Policies that favored urban manufacturing at the expense of rural production ultimately resulted in increasing input and urban food costs, as well as in sharply declining returns to investment.
Many public sector policies needlessly reduce real incomes of the poor as well. Excessive modern sector investment is one such practice. As chapter 9 shows, the Mahalanobis strategy channels investment to the most capital-intensive sectors, thereby minimizing the impact of new investment on labor demand, and hence on unskilled wages. The neglect of agriculture also results in food price hikes, again hurting the poor. Migration restrictions and incomes policies tend to diminish the earnings of most unskilled workers still further.
Furthermore, highly 'prorural' development strategies reduce city growth rates only slightly, largely because urban-rural linkages have become so strong in India. Manufacturing has always been a major user of raw materials, so an augmented supply of those raw materials encourages urban-based manufacturing growth. Because city workers are heavy consumers of foodstuffs, any improvement in food production that lowers food prices also lowers urban wages, thus stimulating city growth. In addition, because Indian agriculture has become increasingly dependent on manufactured intermediates, agricultural output growth stimulates the demand for manufacturing output, creating urban employment. As the Indian economy is relatively closed to trade, all of these urban-rural linkages tend to be stronger than in the SOE in which such linkages spill over into changes in foreign trade volumes. India's poor export performance also implies that the major outlet for increased output must be the domestic market. Thus, if rural demands lag behind, urban output will suffer. Furthermore, the growth of Indian towns and cities appears to have spurred agricultural productivity in nearby areas by generating increased demand and improving the supply of critical intermediate inputs (Dasgupta and Basu, 1985).
In summary, while Indian economic growth may well be greater in the 1990s than it was in the 1960s and 1970s, strong forces limit the extent to which it can be carried by highly unbalanced sectoral productivity advance favoring urban sectors. Development is likely to be most dramatic in the coming decade if urban and rural productivity advance are both enhanced simultaneously.
10.3 Modeling India and Other Developing Countries
Let us turn to a third set of findings coming from the previous chapters and consider what we have learned about computable general equilibrium models. Our general assessment in chapter 3 was the BMW's simulation of a recent portion of Indian economic history provided a surprisingly close fit to reality (even though our vision of that reality was often blurred by imperfect data). In addition, counterfactual simulation with BMW has offered detailed insight into the determinants of city growth. Nevertheless, some of our tales about Indian city growth could have been told without the help of an elaborate CGE model. But this statement can only be made ex post facto: the analyst cannot start by assuming which urbanization and growth forces are most important, which interactions matter most, and which assumptions are most critical.
Although BMW is a complex computer program, it is not a black box. We were surprised at the number of times the model yielded results inconsistent with our economic intuition. But unanticipated results were never accepted (nor hidden) out of deference to the program's inscrutable omniscience. Instead, virtually all results were subjected to scrutiny that enabled us to determine exactly which forces were, and which were not, driving the results. In other words. BMW is not a black box; nor should other CGE models be viewed as black boxes. The model is a tool, not a Ouija board.
Like all models, a CGE can only provide plausible answers to the questions it was designed to address. And like all models, it makes assumptions. The BMW model of India assumes that prices and quantities are, for the most part, determined by the simultaneous clearing of all markets. As such, it is designed to explain medium-term and long-run phenomena. It was not designed to assess short-run responses to shocks, for here disequilibrium rather than equilibrium assumptions seem appropriate. Similarly, BMW has been developed so that we can address questions relating to spatial movements and sectoral growth. But it cannot be expected to provide great insights into policies and shocks impinging on only a very narrow part of the economy.
One of the most valuable contributions of BMW (and something that can in principle be obtained from other CGE models) is a clear understanding of the model's errors. Just as one would examine residuals in econometric work, we devote considerable attention to those episodes and those parts of the economy in which the model did not seem to fit Indian history well. By analyzing the errors, we have been able to identify periods and sectors for which a neoclassical model does poorly and to highlight the need for further research. It turns out that the main errors all have fairly obvious explanations.
One also may ask what can be learned from BMW that could not be learned from a different framework. Relative to other CGEs, the BMW India model richly details government behavior and focuses on infrastructure, capital market fragmentation, and the rural-urban spatial division (as do Kelley and Williamson, 1984). But CGE models do not offer the only means by which to analyze urbanization. It would have been possible to develop a small econometric model of Indian city growth. Such models are obviously useful. Furthermore, they avoid 'one-point estimation' of many parameters, which is inherent in most CGE models. But such models are highly aggregative; they typically contain few explanatory variables and cannot be used to analyze urbanization with the richness offered here. Rather than designating one method as inferior to the other, it makes more sense to regard them as complementary, each with different strengths and weaknesses.
Demographic forecasting offers yet another way to analyze urbanization (see, for example, Rogers, 1984). These models typically have extensive detail with respect to demographic variables; in particular, they usually contain many regions, population groups, migration probabilities, fertility rates, and age- and sex-specific death rates. What they do not have are behavioral specifications or endogenous economic variables. But it is imporper to assert smugly the superiority of economic models: ours, after all, aggregates highly over demographic variables and treats all of them, but not migration, as exogenous.
10.4 Can We Generalize from India?
BMW has been designed to capture critical features of the Indian economy. Our assessment is that it does so rather well, enabling us to draw a large set of conclusions concerning the country's urbanization and growth process. But to what extent can these conclusions be generalized to other countries? The answer hinges on the similarity between India's demo-economic structure and that of other developing countries. Naturally, authors like their findings to be as general as possible. Alas, there are some major obstacles in the way.
A leading obstacle is the greater openness of most developing countries relative to India. In India's semiclosed economy, a sectoral demand boom can fizzle if the costs of traded goods or nontraded inputs are bid up rapidly. In an SOE, traded goods' supply curves are flatter, since traded goods can be imported at a fixed price. Thus, a booming sector is not held back so much by the presence of a stagnant, low-productivity sector in an SOE; indeed, the more unbalanced the productivity advance, the more unbalanced the output growth. But India, with its small, moribund export sector and pervasive import controls, cannot accommodate truly dramatic unbalanced growth quite so easily. In consequence, unbalanced urban-rural sectoral productivity advance, a driving force behind rapid city growth, will have a much more pronounced impact in an SOE than in India.
Two other key features of India's economic structure are its relatively large and developed rural nonagricultural sector and urban manufacturing's dependence on domestically produced raw materials rather than on imported intermediates. The latter feature ensures that rural and urban areas are far more closely integrated than, say, in many African countries. This linkage makes 'runaway city growth' difficult if not impossible and ensures reasonably balanced growth, but it also tends to put the brakes on the overall rate of economic progress. The large rural service sector essentially competes with the urban sectors for released peasant labor, but it can also release labor to agriculture and to the cities, thus reducing competition between them. If Indian agriculture continues to modernize rapidly -and one should not be excessively optimistic here -then the presence of a large rural service sector will provide something of a surplus labor pool that will continue to permit rapid growth of output (but, less happily, continue to depress unskilled wages) both in farming and in the cities.
In short, India has many unique features. Thus, the question remains: Are there any generalizable findings? The most important ones probably center on the role of government. The inegalitarian nature of public sector employment (bidding up skill differentials) and demands (consuming urban goods) is unlikely to be unique to India. So is the need to follow proagricultural policies if large welfare gains for the poor are a target. As a whole, the inefficient government policies outlined above are unlikely to be dramatically different in most other Third World economies. This comment applies most strongly to urban incomes policies and migration restrictions. Furthermore, while IMF-style policies are likely to produce much more growth in SOEs, they are unlikely to be more egalitarian than in India. The distinction between policies that generate large real GDP gains and those that increase living standards of the poor is also generally valid.
The forces driving India's city growth slowdown are to some extent also general, although they operate most strongly in a semiclosed economy. The relative unimportance of rural push factors is probably wide-spread; it was certainly true of nineteenth-century industrial revolutions as well (Williamson, 1990). Similarly, the importance of unbalanced rates of productivity advance, rather than their average level, in driving urbanization is likely to be even stronger in the SOEs, as Kelley and Williamson (1984) have shown. The importance of urban nonmanufacturing sectors and rural nonagricultural sectors in understanding growth and urbanization in India is undoubtedly matched elsewhere as well.
One of the striking findings from our counterfactual simulations was that the misallocation of investment has never been terribly important in India. In a sense, economic inefficiencies are more likely to matter at the firm level: returns to investment are, after all, fairly low everywhere. Since SOEs have more rigidly determined spheres of comparative advantage, sectoral investment choices are probably more important than in India.
In extending the model to other countries -or to India of the 1990s -several modifications need to be considered. In particular, the choice of sectors and the degree of openness must vary from country to country. So, too, must the assumed degree of factor market integration and assumptions concerning domestic restrictions on the tradeability of goods across space. Public sector behavior is also largely country specific. Finally, countries such as India that have experienced significant modernization in portions of their agricultural economies probably should have more than one agricultural sector in a CGE model.
<#FROWN:J39\>
Building Trades
The building and construction trades have made the most direct gains in using pension funds offensively. Funds from building trades union pensions formed three labor banks in Colorado, New Mexico, and Arizona. The building trades have also channeled funds into union-built construction. Over eleven projects across the country with varying arrangements and goals have pooled money from regional pension funds. One notable example is the Bricklayers and Laborers Non-Profit Housing Company, Inc. The Bricklayer Union and Laborer Union pension fund buys federally insured certificates of deposit (with small subsidies from participating unions) from the U.S. Trust Bank in Boston, which, in turn, finances construction loans for union-built, low-price housing for South Boston residents.
Other building trade investment projects operate like the Union Labor Life Insurance Company (ULLICO) 'J for Jobs' project, which invests union pension funds in an open-ended trust that invests in union-built construction. The Housing Investment Trust (HIT) has, since 1964, financed union-only housing construction projects with union pension money. In 1987, the AFL-CIO has expanded the HIT scope and established the Building Investment Trust (BIT), which will finance commercial and industrial real estate investments. In May 1988, the Bakery and Confectionery Pension Plan and the Bricklayers National Pension Fund committed $15 million to the BIT; its first project was a $5.8 million hotel complex in Taos, New Mexico. BIT's advantages to pension funds are its relatively low administrative costs and the assurance that union labor will be used in the construction and maintenance of the project.
The Sheet Metal Workers are using their national pension funds in a fashion that could be interpreted as a type of industrial policy. They seek to increase the demand for their members by closing gaps in financial markets and in the surety industry. In order to help contractors in the asbestos removal industry obtain bonds with lower premiums, the union bought a bonding company. The union pension fund already owns a stake in an asbestos removal contractor (U.S. Bureau of National Affairs 1988c). Edward Carlough, President of the Sheet Metal Workers, calls pension funds the labor movement's 'Star Wars Weapon' (U.S. Bureau of National Affairs 1988c). Approximately $500 million has been directed into real estate projects by building trades investment financing foundations (Westerbeck 1985).
The building trades have come closest to systematically challenging the appropriateness of using only the market rate of return in assessing the quality of an investment. Instead, they use a Keynesian multiplier argument and argue that pension funds, unlike other trust funds where all the income comes from financial investments, have two sources of income - employer and employee contributions and investment earnings - and the rates of return from both sources, not just from the more visible financial markets, must be considered. The multiplier argument is primary rhetorical and has not been quantified.
Most of the DOL challenges to building trade investment projects have failed because ERISA's prudent expert rule allows each investment vehicle to be evaluated within the context of the portfolio and the interests of the participants (Leibeg 1980). The unsuccessful 1981 DOL challenge of the Florida Operating Engineers low-interest mortgage program (for plan participants) showed that legal pension-fund investments must have a reasonable rate of return and must serve the needs of the participants (U.S. Bureau of National Affairs 1986a, 1986c). The Reagan administration had sent mixed signals about these union initiatives. The DOL continues to issue warnings about union social investment projects, but then President Reagan, speaking to a building trades meeting in 1982, applauded the unions' efforts in setting an example for local initiatives (Smith 1984). In fact, the building trades' activities have not threatened the traditional control of capital.
Each project has different effects in terms of reallocating capital. Some projects clearly correct a market failure, or barrier, by providing funds to projects that would have otherwise faced a credit problem. The Boston Bricklayer and Laborer project is an example of reallocating capital. Other projects can be viewed as substitutes for organizing new members. The enemy in the building trades is not always perceived as the nonunion worker or the aggressive employer; and in periods of high unemployment the nonunion contractor is not blamed as much as insufficient demand. The building trades have sidestepped organization in favor of using pension-fund monies as a primary way to increase the market share of unionized construction (U.S. Bureau of National Affairs 1985; Freeman 1985).
State and Local Governments
State and local social investing, practiced by public funds in about twelve states, is confined to mortgages or loans to small businesses. The main purpose is to increase employment in the relevant region by filling a capital gap and not by providing loans below 'market rates.' The state of Michigan's project to coordinate investments of private and public funds to provide small business loans and the Pennsylvania MILRITE project are just two examples of the many affirmative investment strategies of governments. These trends can be tracked through the BNA's Pension Reporter, Labor and Investments, and other pension-industry publications.
From a distance, these programs look like models of socially responsible investment because the government, endowed with the duty to maximize social goals, invests. Only some projects reallocate capital to worthy projects that otherwise would have faced a credit shortage. Other projects involve granting some privilege to the private sector and usually do not involve the public sector labor unions. Moreover, I have not found cases where the state and local governments have opted for owner control; they have not operated the projects they finance through pension funds. Another, more political, state and local use of pension plans is represented by the efforts of the late California State Treasurer, Jesse Unruh, who, at the end of 1984, organized the Council of Institutional Investors (CII), which now has representatives from twenty-two state and local pension funds, six multiemployer funds, and one single-employer plan administrator (Pensions and Investment Age 1989). The CII seeks to monitor and influence corporate policies that affect its state and local pension-fund investments, such as green mail, golden parachutes, and poison pills, because these practices threaten dividend income. CII was very vocal in its criticism of GM management's buyout of its largest individual shareholder, H. Ross Perot. Despite GM's assertion that a $720 million price tag on $350 million worth of shares represented a good investment, CII, as well as most of the financial press, cited H. Ross Perot's increasingly shrill criticism of GM management, for example, his public questions, such as, "Why does it take longer to design a GM car than it did to fight World War II?" as a leading motive for the buyout (Kraw 1989). CII forced a meeting with senior members of GM management. A few years later, in January 1990, two CII members, the New York and California state funds, in reportedly uncoordinated efforts wrote to GM asking for a say in the replacement of Roger Smith as GM chief. Most of the CII's tools are public embarrassment, shareholder resolutions, proxy votes, and capital boycotts. CII actions often seek to protect shareholders from entrenched management resisting a takeover. In many cases, labor would be directly opposed to the interests of the shareholders and in agreement with managements' efforts to save the plant from a hostile takeover and possible plant closing.
Under criticism from Edward V. Regan, New York State comptroller and sole trustee of the $44 billion New York State and Local Retirement Fund, Governor Marie Cuomo's Commission on 'The New York State Pension Investment Task Force' recommended that pension funds obtain guarantees from federal agencies to invest in projects to improve the state's infrastructure. Regan argued against the Commission's report and financier Felix Rohatyn's similar suggestions that the federal government was overburdened by guarantees and the first and foremost duty of the fund was to earn the highest return.
This sort of reasoning was further displayed in 1990 when several state and local pension funds threatened to divest themselves of the stock of all Pennsylvania-based companies to protest Pennsylvania's restrictions on hostile takeovers. The pension funds' position is that hostile takeovers bid up the price of the stock and any restriction on hostile bids constrains the potential profit from holding that equity. Lawmakers in Pennsylvania want to halt hostile takeovers in order to stabilize communities and the business environment. What the Pennsylvania pension plans do will reveal how sharp the horns of the dilemma actually are when it comes to defining socially responsible pension-fund strategies.
Churches, Wealthy Individuals, and Endowments
The Interfaith Center for Corporate Responsibility and the Investor Responsibility Research Council are research and advocacy institutes with strong church connections. Their publications provide advice on share-holder rights and strategies. Most of the alternative investments of university endowments have been in South African divestment and clean fund investments. In addition to specific actions and research functions, the financial industry has responded to the concern of churches, unions, endowments, and wealthy individuals who fear their investments are encouraging antisocial activities by the offering of 'clean funds' (Domini and Kinder 1984). Currently, in the U.S., eight mutual funds and three money market funds advertise holdings that are screened according to various criteria. Two Canadian funds, the Summa Fund and the Ethical Growth Fund, and one British fund, The Stewardship Unit Trust, use social screening criteria in selecting investments. Each fund seems specifically tailored to one or more groups. For instance, the PAX World fund does not contain 'sin' stocks, such as holdings in tobacco or liquor companies or defense contractors. The performance of these socially screened funds beats the market average. As a 1986 Wall Street Journal headline noted, "Investors Can Do All Right By Doing Good". So do the brokerage houses and money managers who have developed this market niche.
From the point of view of labor and churches, however, the structure of capital markets limits the effectiveness of the clean fund strategy. Multinationals and conglomerates issue debt and equity. Targeting money into a particular region, or to a particular subsidiary of a conglomerate, is not possible, because capital allocation decisions are made internally. Managers make investment decisions. Unions face the same problems a shareholders if they attempt to transform companies through the ownership of corporate equity. The advantage the labor movement has over most shareholders - its national presence - can either create enormous clout, or unresolvable conflicts of interest among unions. (How successful would the Shell boycott have been if the Oil, Chemical and Atomic Workers Union represented Shell employees?)
But, trustees of funds of many corporations and entities with trust funds are asking the same question Harbrecht (1959) and Barber and Rifkin (1978) exhorted labor unions themselves to ask. In the words of a pension attorney, "What is it that says trustees of the American Cancer Society can't say (its pension fund) won't invest in tobacco stocks" (Gerald Feder, quoted in Pensions and Investment Ages 1989).
Corporations
Corporations have discovered in the 1980s that their pension finds can be convenient sources of cash and can serve other corporate needs. A combination of legal, political, and economical factors explains why single-employer pension funds are now used as a corporate financing tool.
The interpretation, and Reagan administration's enforcement of ERISA, all but encouraged corporations to put their funds to innovative and, indeed, alternative uses. To fend off hostile takeovers, or to court 'white knights', a corporation can often obtain DOL permission to direct its employee pension fund to buy the company's stock (as long as it is not 'overpriced'). A corporation can terminate a plan that happens to have more assets than legal liabilities, pay the liabilities, and use the extra cash. The corporation can also manipulate the rate of return the fund is assumed to earn in order to alter the cash contributions required to fund the liability. Raising the assumed rate assumes the fund will earn more and, therefore, require less from the firm.
Between 1980 and 1987 about $9 billion had been recaptured by corporations in defined-benefit pension plan terminations (VanDerhei and Harrington 1989: 189).
<#FROWN:J40\>
Straight talk from Wall Street
A highly regarded stock analyst points out what's right - and wrong - with bank investor relations
By Thomas K. Brown
What makes an effective investor relations program? The answer depends to an extent on your position. I will provide the perspective of a Wall Street bank stock analyst.
We'll start with the person in charge of investor relations. This person must have a thorough understanding of the company and its financial statements, as well as access to senior management.
He or she must be able to perform the delicate task of providing information to analysts and investors without crossing the line of differential disclosure.
This person cannot be purely a cheerleader, nor a sugar-coater. The goal is to present a fair picture of current conditions.
There is no one spot in the organization from which the head of investor relations should come.
For example, at Wachovia Corp. and Mellon Bank Corp. the heads of investor relations are individuals located in the finance part of the organization. Both are excellent. Bank of America Corp. has one individual whose sole responsibility is investor relations and he is the best in the business. Fleet/Norstar Financial Group operates with a team involving the chief financial officer and two individuals with responsibilities in addition to investor relations. Yet the result is quite effective.
The form may vary, but in all these cases the end result is an effective program.
Unfortunately, in too many companies, including some of the largest banking firms, the head of investor relations is either not knowledgeable about the company and its financials or lacks timely access to senior management. These companies are known for their mishandling of important information, for differential disclosure, and for the dreaded 'surprise' announcement.
Responsive voice. As important as the quality of information disseminated is its timeliness. Institutional analysts and investors are under ever-increasing pressure to gain access to important information and draw investment conclusions.
At First Union Corp. and Wells Fargo & Co., investor relations staffs are briefed about the company's quarterly earnings just before they are released. In this way, several individuals at the company are then able to answer questions from analysts and investors once the earnings release is made public.
At other companies there is only one individual available to answer questions. This inevitably leads to long delays in receiving answers and effectively puts some analysts at a significant competitive disadvantage.
However, because there are more investors and analysts than staffers, at times priorities will have to be made as to who is called first.
Don't just return calls in the order they are received. In my opinion, the top priorities should be large shareholders and those Wall Street analysts who have been particularly visible (both positive and negative) in providing research on the company. This may seem unfair, but it will lead to quickest dissemination of information.
Good, timely disclosure. Analysts are never satisfied with the level of disclosure. So it's only natural that one of the traits that I find in the best investor relations programs is excellent disclosure of timely information.
Recent examples of breakthroughs in the disclosure of important new information include:
A reconciliation of flows in nonperforming assets provided by Valley National Corp., Bank of Boston Corp., and Shawmut National Corp.
First Tennessee National Corp. provides a breakdown of its loan portfolio by risk rating.
Wells Fargo provides excellent disclosure of the cash payments received on nonperforming loans.
The goal of increased disclosure should be to put meat on the bones of the company's required financial statements. The enhanced disclosure of asset quality trends previously described, as well as line-of-business results, are quite helpful in this regard. A better understanding of what is occurring at the company can only help to reduce investor uncertainty. Over time, this will lead to a higher relative stock price valuation.
One of the principal sources of disclosure is the annual report. I believe this is the single most important communication a company makes to its share-holders and potential investors. It should provide the reader with a thorough understanding of the company's products, markets, strategy, and recent financial performance.
Frankly, I am amazed that the investor relations departments at numerous companies do not provide any input into the company's annual report - a mistake.
There are numerous examples of companies that publish outstanding annual reports. Among them are Fleet/Norstar, First Union, Barnett Banks Inc., and Norwest Corp.
Unfortunately, some of the highest-quality companies publish some of the worst annual reports. As long as these companies continue to deliver superior results, this won't be a major problem. But if they stumble, the lack of disclosure will lead to investor uncertainty and a sharply lower relative valuation.
No differential disclosure. It is difficult to be responsive and provide timely disclosure of information without crossing the line and providing differential disclosure to one analysts or investor. The issue can only be managed. It cannot be avoided completely.
However, there are some obvious steps that can be taken to reduce differential disclosure.
For example, don't allow analysts and investors to visit the company after the quarter is over and before earnings are reported. Try as they might, senior management seldom is able to conduct these meetings without providing an important insight into the quarterly results.
In addition, when analysts and investors call after the quarter is over and before the results are announced, the investor relations department should discuss only the results from the first two months of the quarter and make this clear to the caller.
Finally, senior management must spend more time telling employees what constitutes inside information. I am surprised at the number of company executives who know they must not disclose certain information to analysts and professional investors. Yet they will freely discuss such information with friends in social settings.
Leaks of information destroy an investor relations program and raise doubts about management's credibility - and internal controls.
Educate the Street. Effective investor relations isn't simply being responsive to Wall Street's ever-changing desires. Sometimes investor relations programs need to take a stand on issues.
My personal peeve is the widespread use of the reserve-to-nonperforming loan ratio as the sole measure of reserve adequacy. Any good banker knows that this ratio is a poor indicator of reserve adequacy. Yet management let Wall Street adopt this measure over the past few years as its principal tool in reserve adequacy.
Educate the brass. An effective investor relations program must educate the company's senior management, and particularly its chief executive officer, in the workings of Wall Street, for example:
The difference between good companies and good stocks. A good or even great company does not necessarily make a good investment over the short or intermediate term. And those are the time frames of most investors, whether they say so or not.
As of mid-November in 1991, J.P. Morgan had had an outstanding year and it is widely regarded as an excellent company. Yet it has been one of the market's worst-performing bank stocks. Analysts can recognize J.P. Morgan as an outstanding company and yet not recommend its stock - and be helping their clients in the process.
Analysts' varying perspectives. Most Wall Street analysts are not good stock pickers.
Many are really investment bankers hiding out in the research department; others are just good at understanding industry trends. A good investor relations program will explain to senior management the strengths and weaknesses of the analysts who are making public comments or writing reports.
For example, when the analysts at the company's principal investment banking firm write a glowing report about the company, it's up to the investor relations department to remind senior management of the analysts' bias. Or, when an analyst who doesn't know the company well publishes a report that is somewhat off base, the investor relations department should explain this to senior management and then contact the analyst.
Don't let the CEO blow up over negative comments. When a negative report is written or negative comments are made, it is up to the investor relations department to encourage the CEO not to act irrationally by cutting the analyst off from the information flow or by preventing the banking company from doing any business with the analyst's firm.
Such steps are only understandable when careless analysis has damaged the company or if the analyst is deliberately spreading misinformation.
An effective investor relations program can significantly influence how investors view the company, which will impact how the company's earnings stream is capitalized over the long run. Too often, managements judge the effectiveness of their investor relations programs by short-term movements in their stock price - a critical mistake.<*_>square<*/>
ADA compliance is uncharted territory
With just a short time remaining before some provisions of the Americans with Disabilities Act (ADA) take effect, community bankers are generally less concerned about compliance with the law than their counterparts in large metropolitan-area banks.
Nevertheless, they have some real concerns, not the least of which is how to finance structural and other changes that may be required. "From an employment standpoint, we have a second floor without elevator access, which pretty well leaves the handicapped out," notes Gerald Hansen, vice-president and controller of Itasca (Ill.) Bank &Trust Co., with $172 million in assets. "We can make arrangements that will alleviate the problem in terms of interviews, but if someone were qualified to work in that area, we would have to consider the cost of installing an elevator in the building." The law states that "readily achievable" measures to remove physical and communications barriers must be taken, but such language has yet to be tested in the courts, and the expense of installing an elevator probably would not be found readily achievable.
But the problem is not an isolated one. Michael Derr, assistant vice-president, operations, at the $144 million-assets Bank of Glen Burnie (Md.), faces a similar situation at his bank's main office. That facility was built in the early 1950s, notes Derr, with very limited access to the second floor.
The bank has four remote branches as well, "and the bank is surveying those facilities and identifying areas that could be considered barriers," says Derr. Like most banks, accessibility to automated teller machines ranks high on the list of potential trouble spots, but access to night depositories and rest rooms can't be overlooked either, notes Derr. "We'll have to make a judgement as to what is practical," he says.
Derr and dozens of other bankers attended an ABA-sponsored symposium on ADA in late October. Speakers, including those representing various disabled groups, encouraged banks of all sizes to have a committee in place whose task is to formulate a strategy for complying with the law. Such a measure will help persuade a court of law that a bank is attempting to meet the requirements of ADA, if and when litigation occurs.
Disabled neighbors. "We're not anticipating litigation," says Paul Sciacchitano, executive vice-president and cashier of Old Point National Bank, Hampton, Va., which has $263 million in assets. "A lot of disabled people deal with us already, since we're one street away from a Veteran's Administration center," he adds.
A year ago, Sciacchitano's bank surveyed its locations and identified areas where access for the disabled could be improved. Since then, Sciacchitano has become more sensitive to the issue of accessibility for the disabled. He's observed that a local fast-food restaurant, for example, put a wheelchair ramp in a virtually inaccessible location leading to a doorway that is only about 20 inches wide. "And the ramp was placed in the middle of a parking space with no clear access to it," says SciaccitanoSciacchitano. "It was done as an afterthought and showed no real consideration."
Most disconcerting to Sciacchitano and other bankers is the fact that the only real enforcement of the law regarding facility accessibility lies with the Department of Justice. The Equal Employment Opportunity Commission (EEOC) will oversee compliance of the law's Title I provisions, which deal with human resources issues.
"Compliance is more subjective than objective," says Sciacchitano, "but the monetary penalties involved are rather objective.
<#FROWN:J41\>
Motor Carrier Deregulation and Highway Safety: An Empirical Analysis
DONALD L. ALEXANDER
I. Introduction
The passage of the Motor Carrier Reform and Modernization Act of 1980 (MCA 1980) effectively marked the end of rate and entry regulation in the interstate trucking industry, and most economists agree that this legislation has created many important economic benefits during the 1980s. For example, shipping rates fell in real terms, service expanded and improved in quality, and generally resources were used more efficiently. There is increasing concern, however, that deregulation has led firms to reduce safety expenditures in order to remain competitive, and that trucking accidents and fatalities have increased as a consequence [4; 9]. Indeed, Brock Adams (former U.S. Secretary of Transportation) claims that trucking accidents have increased since 1980 for this very reason.
The limited empirical evidence reported in the literature suggests otherwise. Moore [17], for example, shows that accident, fatality, and injury rates have fallen since 1980 despite the rapid increase in the number of truck-miles traveled [13]. His analysis, however, is based on a comparison of rates across time, and does not reveal any of the potential economic or institutional forces that may be affecting the evolution of these data. In a more systematic analysis, Traynor [22] finds that deregulation has reduced accident rates in California, although it may be difficult to extend his results to the national experience as a whole.
This paper has two major objectives. The first is to estimate an empirical model using a pooled, cross section of state data to determine whether the evidence Traynor [22] reports for California holds, in general, for all other states. The second objective is to determine empirically those factors that may be driving the results in Moore [17]. Pooling these data allow us to test for the impact of deregulation on accident, fatality, and injury rates while holding state-specific factors constant; an empirical approach that is not possible at a higher level of aggregation.
The paper is organized in the following manner. In the second section I discuss three possible ways deregulation may have affected accident rates in interstate trucking. The third section presents the empirical model and regression results, and the final section summarizes the major conclusions drawn from this investigation.
II. Motor Carrier Deregulation and Highway Safety
The current theoretical literature provides a framework for discussing the potential link between changes in deregulation and highway safety. I will use this framework to focus on three aspects of deregulation that are likely to affect highway safety in the trucking industry.
Before deregulation, rate bureaus acted as cartels and set shipping rates at supracompetitive levels. Since the Interstate Commerce Commission (ICC) restricted the entry of new trucking firms, incumbent firms were able to earn economic rents that were shared with organized labor. After deregulation, however, the ICC essentially permitted free entry and naturally there was an influx of new firms. Winston et al. [31], for example, report that the number of truckers with ICC operating authority increased from 18,045 in 1980 to 36,948 in 1986. One implication for highway safety is that trucking accidents may have increased because of the additional traffic congestion. Moreover, if the new entrants were inexperienced in handling trucks on congested highways, it is likely accidents would have increased until the new drivers gained the necessary driving experience.
The implicit assumption underlying this argument is that a driver's behavior towards safety remains unchanged in response to a change in the regulatory environment, which Peltzman [18] has questioned on theoretical grounds. Although deregulation in the trucking industry did not involve a specific change in safety regulation, Peltzman's model provides a useful framework to examine the effect of deregulation on the behavior of owner-operators.
The basic model is that drivers face a choice between driving intensity (e.g., faster speeds, longer hours) and the probability of an accident, which is affected by several economic factors: the price of an accident, income, and a regulatory parameter which we will associate with the driver's promotion of safety. In a cross-section model, it is difficult to argue that interstate drivers respond to differential 'accident prices' across each state. Moreover, it is equally difficult to argue that secular changes in income that Peltzman discusses are at work to the same extent in this analysis since we are using data for a shorter time span. Therefore, I will ignore the first and second factors and focus on the third.
We can think of deregulation shifting the driver's demand for driving intensity in several ways. The widely-held view is that deregulation has led owner-operators to reduce safety expenditures to remain competitive, which raises the probability of an accident for a given level of driving intensity. It is not quite clear, however, why owners would reduce safety expenditures when their discounted future profits depend on providing timely and safe deliveries today.
On the other hand, deregulation may have induced firms, which employ drivers for hire, to increase their safety expenditures. Suppose a trucking firm uses two complementary inputs, labor (L) and safety (S), per truck to deliver a given quantity of goods. We can think of S broadly as the resources the firm uses to maintain some level of safe operation, which may include any investment in safety training for drivers or the purchase of truck-related safety equipment. In competitive markets, the level of safety the firm provides determines the wages the firm must pay to compensate drivers for any relative risks in trucking. If, for example, the firm offers better training or installs more (or better) safety equipment, then wages would be commensurately lower. As wages fell after deregulation, this may have created an incentive for firms to hire more drivers and to provide additional safety to compensate drivers for any apparent risks in trucking. In addition, it is quite possible that firms provided the drivers with greater safety resources (i.e., equipment or training) because the new drivers were relatively inexperienced. The implication is that if firms increased safety expenditures because wages fell, then accidents would have probably fallen as well. Moreover, if the additional expenditures were used to purchase truck-related safety equipment, then fatalities and injuries would have fallen too.
And finally, deregulation may have affected driving intensity directly if drivers were induced to make more deliveries by driving longer hours and at faster speeds. Therefore, it would be important to control for differences in speed limit enforcement and vehicle inspections when attempting to explain differences in accident rates across states.
The above discussion indicates there are reasons to expect that accidents may have increased or decreased as the result of deregulation. The next section discusses the empirical model which attempts to determine the net impact of deregulation, while holding various other factors constant.
III. The Empirical Model and Regression Results
The Empirical Model
The sample consists of a pooled, cross-section of state data for 1977, 1982, and 1987. Since the MCA was passed in 1980, the sample can be partitioned into two periods: 1977 represents a regulation year, whereas 1982 and 1987 represent deregulation years. The variable descriptions and sources are discussed in the appendix.
The dependent variable is the number of accidents per truck-miles traveled per state. The numerator represents accidents that occurred in a particular state, which were reported by truck drivers engaged in interstate transportation. The denominator is an estimate of the truck-miles traveled per state, and includes all trucks that traveled more than 200 miles from their base of operation. To the extent that the denominator includes some intrastate travel, the denominator is likely to overstate the miles logged by interstate truckers. Therefore, the dependent variable is likely to understate the actual accident rate per truck-miles traveled for interstate carriers, and any effect of deregulation that we uncover would be a conservative estimate of the actual impact.
The set of independent variables include factors which have some theoretical or institutional basis for explaining the variation in accidents per truck-miles traveled across states. The first is the number of highway police officers per highway mileage (POLICE). This measure is intended to control for differences in enforcement resources used to detect speed and weight violations. More officers per highway mile should lead to more careful driving and, consequently, less accidents. Thus, we expect a negative sign for this variable.
The second, third, and fourth variables relate to traffic conditions on interstate highways which could arguably affect accident rates across states. These conditions are: the average speed of interstate traffic per state (AVESPEED); the variance of speed on interstate highways; and a density variable (CONGESTION) to proxy the level of road congestion. Lave [14] argues that it is the variance of highway speed that causes accidents and not the mean speed. The intuition is that if all vehicles were traveling at the same speed (i.e., variance is 0), then chances of an accident occurring are almost zero at any mean speed. Recently, Levy and Asch [15] challenge this view and report some evidence which shows that both the mean and variance of highway speed affect accident rates. Since this issue appears to be unresolved, I have included both measures in the model.
AVESPEED is the estimated statewide average highway speed for all vehicles. The variance is calculated as the difference between 'the speed at or below which 95 percent of the vehicles are traveling' (85TH) and the statewide average. In the regression model, one could write the expression as
<O_>formula<O/>.
Lave, however, suggests that the expression be rewritten as
<O_>formula<O/>.
The rationale is that the coefficient for AVESPEED in the second expression reflects the relative effect of both the average and variance of speed (i.e., <*_>beta<*/>-<*_>THETA<*/>). Thus, if the variance has a larger impact than the average (i.e., <*_>beta<*/><<*_>THETA<*/>), then the coefficient on AVESPEED would be negative while the coefficient on 85TH would be positive. This explains why a negative sign for AVESPEED is counter intuitive, since the coefficient measures the relative size of each effect.
CONGESTION is simply the total number of automobile registrations normalized by the estimated highway mileage for each state. This factor is intended to control for differences in road congestion across states, since greater congestion is likely to increase the probability of a collision between two vehicles. Thus, we expect a positive sign for this variable.
I included two additional variables to control for differences in weather conditions across states. The first is the average number of rain days per state (RAINDAYS). I anticipate a positive coefficient for this variable since rain is likely to impair vision and road conditions. The second is the average snowfall (SNOW) per state, and the expected sign for this variable is uncertain. On the one hand, one may argue that more snow leads to more accidents because of slippery roads. On the other hand, more snow might reduce travel if drivers wait until the roads are cleared, which makes a negative sign plausible.
Finally, I included a dummy variable (MCA80) in the model to control for differences attributable to deregulation; 1977 is a regulation year while 1982 and 1987 are deregulation years. Thus, MCA80 equals 1 for 1982 and 1987 and 0 otherwise. It is possible, however, that any empirical difference between the time periods is unrelated to the shift from regulation to deregulation, and I acknowledge this potential interpretation. Nonetheless, any difference that is uncovered in the analysis will be discussed in the context of the existing empirical literature [3; 17].
The Empirical Evidence: Rates per Truck-Miles Traveled
Table 1 reports the estimates from four regression equations, and several interesting findings emerge from these results. First, the MCA80 variable has a negative sign in each of the four equations, but is only significant (using a two-tailed test) at conventional levels in the fatality, injury, and property-damage equations. These results suggest that drivers experience the same accident rate that they did before deregulation, but that the accidents involved fewer fatalities and injuries.
The insignificance of MCA80 in the accident equation presents a puzzle; that is, how and why would deregulation affect the fatality and injury rates, but not the accident rate?
<#FROWN:J42\>
Part 1. Jurisprudence is Adjudication
a. Critique of legal positivism
Dworkin constructs his legal theory largely in response to legal positivism and utilitarianism. Legal positivism, Dworkin claims, defines law merely as a set of rules or 'social facts,' and allows judges to legislate when rules run out in hard cases. As such, Dworkin argues legal positivism contains two main defects: (1) an inaccurate description of adjudication in American legal practice, and (2) an inadequate theoretical account of law and legal obligation. Utilitarianism, Dworkin claims, enforces the preferences of the majority over the preferences of the minority to improve social welfare. Dworkin considers utilitarianism a deficient political theory because individual rights depend upon the shifting sands of political compromise and majority will. Consequently, Dworkin's legal theory establishes certain individual rights beyond the control of the majority, and links positive law to political theory and thereby to virtue and justice. Dworkin thus rejects, partially at least, both legal positivism and utilitarianism.
Dworkin starts his examination of law with a critique of Hart's version of legal positivism. Hart's positivism has three central tenets: (1) the law of a community consists of special rules identifiable by the manner in which they were adopted, (2) the set of legal rules is exhaustive of the law, and (3) a legal obligation derives only from a legal rule. Hart further divides rules into two types: primary and secondary rules. Primary rules grant rights or impose legal obligations upon members of the community. For example, the criminal law consists of primary rules. Secondary rules stipulate how primary rules are formed and validated. Hart calls a fundamental secondary rule a 'rule of recognition.' The latter is legitimate because it is accepted by the community. In the United States, the rule of recognition is the federal constitution since the legitimacy of any particular law can be traced through a complicated chain of validity back to the federal constitution.
According to Hart, judges decide cases by applying rules of law. When a case is not governed by any existing rule of law, a hard case, the judge decides the case by exercising his discretion. The new rule the judge forms becomes part of the legal order and is valid because, under Hart's system, the judge has power to create a new rule when existing rules do not provide guidance in a particular case, Judges hence "may be said to make 'choices' among possible alternatives or to exercise a 'legislative discretion.' Since most cases involve the simple application of rules, the legislative powers of the judge are limited.
Dworkin charges that Hart's theory of judicial discretion inaccurately describes what judges in the United States do when they decide a hard case. Judges use principles, in addition to rules, to decide cases. Principles differ from rules because principles are abstract, general and flexible. In support of his claim, Dworkin cites Riggs v. Palmer, denying a murderer his inheritance on the ground that "no man shall profit by his own crime," and Henningsen v Bloomfield Motors, Inc., voiding a limitation of warranty provision in a consumer contract principally on the ground that unfair bargains are unenforceable. Because the courts in Riggs and Henningsen invoked principles not rules to decide the case, Dworkin concludes, Hart's theory of judicial discretion fails to describe judicial decision making and therefore is wrong.
The origin of legal principles "lies not in a particular decision of some legislature or court, but in a sense of appropriateness developed in the profession and the public over time. Their continued power depends upon this sense of appropriateness being sustained." When a judge decides a hard case, the judge does not simply create a decision in a vacuum; rather the judge invokes the applicable principles of law and applies them to the case. The judge does not create law and apply it retroactively to the parties; the judge enforces moral and legal rights preexisting the case although not captured by any single rule of law. Principles exist independently of legal institutions enacting rules of law because they are part of the community's moral and political culture.
Dworkin draws two broad conclusions from the putative failure of Hart's theory of discretion to reflect actual judicial practice. First, he claims Hart's theory of law does not identify all laws in the society because it fails to account for the existence of principles that judges commonly use to decide cases. Second, the master rule of recognition, to the extent that it ignores rules, is not a master rule defining all laws. If the master rule were redefined to capture principles, Dworkin maintains, it would become so broad as to be meaningless. Dworkin concludes "if we treat principles as law we must reject the positivists' first tenet, that the law of a community is distinguished from other social standards by some test in the form of a master rule." This raises the possibility, Dworkin contends, that legal obligation rests on constellations of principle, as well as rules of law. More important, the critique of legal positivism provides the primary material for Dworkin to create the rights thesis.
Scholars claim Dworkin reduces positivism to a theory no one actually holds since positivists recognize restraints upon judicial discretion. Sullivan correctly notes that "judicial discretion is more tightly circumscribed than Dworkin's caricature indicates." Though the judge is free to weigh various considerations, "this does not entail that decisions resulting from this process are arbitrary, or that the judge's discretionary power is therefore completely unconstrained." Nevertheless, genuine differences differentiate Dworkin's legal theory from positivism. For positivism "law is fundamentally characterized by the notion of a rule," whereas for Dworkin it is a process of discovering the political morality implicit in positive law. Positivism distinguishes law from morality by identifying a master rule of recognition; Dworkin denies the existence of a master rule of recognition and locates law in the practice of interpretation. While some scholars argue that Dworkin's legal theory merely amends positivism, his theory nevertheless investigates the origin of law beyond the mere fact of its enactment by a legitimately constituted legal institution.
b. The rights thesis
The rights thesis corrects two flaws in the positivist's account of judicial discretion: (1) treating the judge as deputy to the appropriate legislature, and (2) claiming judges decide cases in two stages, first reviewing the law books to locate pertinent rules, and second setting aside the law books when pertinent rules are not found. Dworkin says judges are not and should not be legislators for two reasons. First, judges are not elected and therefore, under democratic theory, are not entitled to make law, and second, the judicial creation of ex post facto legislation punishes the losing party. Dworkin also denies that judges decide cases in two stages as positivism maintains. Rather judges enforce the preexisting rights of parties grounded in legal principles. According to Dworkin, adjudication should be as unoriginal as possible.
Central to the rights thesis is the distinction between arguments of policy and arguments of principle. Arguments of prinicple justify a political decision by showing that the decision respects or secures some individual or group right. Dworkin states the "argument in favor of anti-discrimination statutes, that a minority has a right to equal respect and concern, is an argument of principle." Arguments of policy justify a political decision by showing the decision advances or protects some collective goal of the community as a whole. Dworkin states the "argument in favor of a subsidy for aircraft manufacturers, that the subsidy will protect national defense, is an argument of policy." While Dworkin realizes principles and policies mix, and therefore recognizes the distinction between them is more subtle and complex than his examples suggest, he nevertheless advances the claim that a principle cannot be outweighed by every social policy.
Dworkin rejects arguments of policy as a legitimate basis to decide cases because they fail to recognize the existence of rights and require the judge to legislate. Arguments of policy do not provide a stable vehicle to secure rights since they depend upon variable factors designed to promote the social welfare. If arguments of policy determined rights, the latter would fluctuate according to whatever factor advanced the social welfare at a particular historical moment. The law and economic analysis theory illustrates the problem of basing rights on utility, since the efficient decision may deviate from prior law, and hence frustrate the expectations of the parties. Dworkin claims citizens are entitled to rely on rights and duties flowing from the law, and are entitled to request the court to enforce them. If a plaintiff is entitled to win a lawsuit, Dworkin maintains, the plaintiff always had the right to win and the defendant always had a duty to act. Economic analysis runs afoul of this conception of adjudication because economic analysis defines rights ex post facto on grounds of efficiency.
Dworkin divides rights into four categories; (1) background (2) institutional (3) abstract and (4) concrete. Background rights are rooted in political theory and are not necessarily recognized as rights by legal institutions. For example, a political theory may demand "to each according to his needs, from each according to his ability," although no legal institution in the United States yet recognizes that claim. On the other hand, an institutional right "provides a justification for a decision by some particular and specified political institution" and therefore, unlike certain background rights, has the force of law. An abstract right is "a general political aim" such as "Congress shall enact no law abridging the freedom of speech," while a concrete right gives practical content to its corresponding abstract right. For example, the right of newspapers to publish defense plans classified as secret provided the publication will not create an immediate physical danger to troops is a concrete expression of the abstract right contained in the first amendment.
The rights thesis enforces existing concrete and legal rights of an institutional type. Dworkin draws an analogy between the institution of chess and the institution of law to explain what he means by the institutional character of legal rights. He considers how a chess referee would interpret a rule of chess which provides that "the referee shall declare a game forfeit if one player 'unreasonably' annoys the other in the course of play" when one player smiles continually at his opponent to unnerve him. Since the rule does not define the term 'unreasonably,' Dworkin says, the referee must construct a theory of the game of chess to interpret the rule. The theory of the game of chess is derived from the rules constituting the game.
Dworkin first observes that the chess referee cannot interpret the rule by imposing personal convictions. For example, the chess referee may believe that individuals have a right to equal welfare without regard to intellectual abilities and rely upon this conviction to find that annoying behavior is reasonable so long as it reduces the importance of intellectual ability in deciding who will win the game because chess is a game of intellect. However, "(s)ince chess is an intellectual game, (the chess referee) must apply the forfeiture rule in such a way to protect, rather than jeopardize, the role of intellect in the context." Therefore, the discretion of the referee is fettered by the nature of the game of chess which disqualifies personal convictions of the referee contrary to the game's point.
The chess referee must determine the abstract concept of the game of chess and interpret the rule to implement that concept. The abstract concept of the game of chess is identified by analyzing its institutional rules and by posing a series of questions designed to identify the game's character. Since chess is a game of intellect, Dworkin suggests, the referee may need to construct not only the concept of chess, but also the concept of intellect itself to interpret the rule forbidding annoying behavior. 'Intellect' is the point of the game. While the referee exercises judgment to define the concrete right of the players, the exercise of judgment does not reflect the referee's personal convictions.
<#FROWN:J43\>
Conceptualizing Anti-Gay Violence
JOSEPH HARRY
This chapter attempts a conceptualization of the motivations and situations surrounding the hate crime of violence against gay males and lesbians. (The term gay will henceforth be used to refer both to gay males and to lesbians. Gay males and lesbians will be used to discriminate between the groups.) Violence is anti-gay when its victims are chosen because they are believed to be homosexual. This definition excludes common crimes committed against gay males or lesbians when the homosexuality of the victim is unknown or irrelevant to the choice of victim. Although some research has been done on the victims of anti-gay violence (Committee on the Judiciary, 1986; Harry, 1982; Miller & Humphreys, 1980), there is little knowledge about the perpetrators. In this chapter, I attempt to enlarge on this scarce data.
MOTIVATIONS FOR ANTI-GAY VIOLENCE
As Berk and his colleagues suggest in Chapter 8, the perpetrators of anti-gay violence are very largely male, in their late teens or early twenties, strangers to the victim(s), in groups, and not engaged in victimization for profit. Anti-gay violence seems to be committed during the peak years of delinquency/criminality (Hindelang, Gottfredson, & Garofalo, 1978). Anti-gay violence may thus be but one element of the general delinquency complex in which correlations are found among most kinds of illegal behaviors. If so, it may require little special explanation beyond those usually offered for delinquency and crime; that is, no special psychological propensities on the part of the offender need be assumed.
Even if, however, the typical anti-gay offender is a generic criminal disengaged from the conventional moral order, some closer examination is required to explain why or when he may engage in a particular type of offense (male pronouns are used throughout to highlight the likelihood that perpetrators are male). Whereas disengaged delinquents are free to commit a variety of illegal activities, such freedom does not mean they will engage in any one particular activity. Motivations and situational circumstances are needed to focus their attention on a particular illegal possibility. Why commit anti-gay violence versus rape or armed robbery or burglary? What is there about beating homosexuals that appeals to offenders?
I suggest that most anti-gay violence arises out of the interactions of male groups in their late adolescence or early twenties. For many persons, the period of adolescence constitutes an extended 'moral holiday' during which bonds to the adult moral order are attenuated by involvement in an adolescent subculture, the principal emphases of which are hedonism and autonomy from adult control. Such adolescents find themselves most at home not in school or in the family but in the company of same-age peers. Such company is unstructured, informal, and largely devoted to recreational pursuits, both legal (e.g., sports) and illegal (e.g., drugs). Although the social groups of the adolescent and immediately postadolescent worlds consist of both same-sex and mixed-sex groups, groups of gay-bashers seem to be almost exclusively male. In a few accounts of gay-bashing incidents, a female consort of the offenders served as appreciative audience. This acknowledged, it remains that the offenders are overwhelmingly male and usually act in groups.
One depiction of such male adolescent groups has been provided by Matza (1964, pp. 49-64) in what he calls the "situation of company." In this situation, adolescents are constantly mutually pressured to prove their commitment to the male gender role. Engaging in a variety of illegal or deviant acts is one way to prove their daring, their maleness, their adulthood. In the rather primitive eyes of the adolescent male, sexual and violent acts are the two main means through which they can prove their male commitment. For example, adolescent males have been found much more likely than females (68% versus 44%) to tell their friends about their first experience with sexual intercourse (Carns, 1973), apparently because reporting such intercourse has status value in the eyes of peers.
Although violence can also validate one's commitment to being a male, it has risks. In the legitimate forms of sports, one can lose. Also, many forms of available sports are supervised by adults and hence do not fit well with the emphases of the adolescent subculture. Most illegal forms of violence, such as fighting, offer the possibilities of losing, being injured, possibly being arrested, and having one's status considerably deflated. Hence, although it is important for the adolescent male to be able to talk a good fight, actually engaging in one is risky business.
The option of gay-bashing offers a nearly ideal solution to the status needs of the immature male. When done in groups, it offers little risk of injury. It provides immediate status rewards in the eyes of one's peers because, unlike verbal reports of sexual conquest, it provides direct and corroborated evidence of one's virility. It offers only minimal likelihood of arrest both because the offenders are rarely known to the victim and because the victim is unlikely to report the incident to the police. Gay-bashing serves to validate one's maleness in the areas of both violence and sexuality. It is a sexual, but not homosexual, act because it reaffirms one's commitment to sexuality exclusively in its heterosexual form. Occasionally, gay-bashing incidents include forcible rape, either oral or anal. Given the context of coercion, however, such technically homosexual acts seem to imply no homosexuality on the part of the offenders. The victim serves, both physically and symbolically, as a vehicle for the sexual status needs of the offenders in the course of recreational violence.
The offenders' choice of victim is made appropriate by the institution of gender. Although young males living in the situation of company and morally adrift may find anti-gay violence appealing, such behavior requires that the laws and norms of civil society be morally neutralized. In cases of gay-bashing, the offender is not simply on a moral holiday, as he may be when committing common property offenses, nor is he simply grabbing excuses out of thin air to justify seriously criminal behavior. He is resorting to an alternative set of norms based upon the institution of gender: that set of norms, imbibed mostly unconsciously from birth, that prescribes our sense of what is 'masculine' and what is 'feminine' in thought, affect, and behavior.
The gender institution often operates as a set of subterranean values justifying illegal conduct when more acceptable justifications (e.g., self-defense) cannot be found within the law. Our dominant institution of gender contributes to the view that male-female rape is justified if the victim behaved in a 'provocative' or 'unladylike' manner. It also allows the perpetuation of wife-beating. Gay-bashing seems similarly to be based on a popularly accepted belief, in this case that the only justifiable forms of sex are those between males and females. In the case of gay-bashing at least, moral neutralization is based upon "denial of the victim" (Sykes & Matza, 1957) and of her or his moral worth as a human being. By viewing the victim as worthy of punishment for having violated gender norms, the offender not only excuses himself from opprobrium but sees himself as rendering gender justice and reaffirming the natural order of gender-appropriate behavior.
The above arguments may seem to predict too much gay-bashing, just as Matza (1964, pp. 25-26) argued that cultural theories of delinquency predict too much delinquency because they imply a continuing commitment by the juvenile to delinquent behaviors. Matza's point was that, if juveniles are so committed to delinquency, they would engage in it almost on a full-time basis. Similarly, if anti-gay violence is an ideal means for the attainment of sexual status by young males and is based on such a basic institution as gender, it would seem that gay-bashing should be a daily occurrence involving significant percentages of both the homosexual and the heterosexual populations. To deal with this issue, we need some idea of the extent of anti-gay violence. Because relevant statistics are few, we divide gay-bashing incidents into three types based on the age of the victim.
First are serious physical assaults and homicides committed against adult lesbians and gay males such as those reported in the House Criminal Justice Subcommittee hearings on Anti-Gay Violence (Committee on the Judiciary, 1986). These reported assaults are clearly the most serious ones and do not include the common, random beatings of homosexuals that occur in the streets, parks, and parking lots of America. Most assaults go unreported either because the victim fears being discredited by family, the law, or employers or because the assault was less serious, although still criminal.
Second are assaults and related harassments of lesbian and gay male adolescents by their peers, such as those that gave rise to the Harvey Milk School in New York City for homosexual adolescents. The existence of such a school implies that mistreatment of homosexual adolescents is pervasive in the adolescent world.
Finally, probably far more common than either of the other forms of assault and harassment are the beatings of effeminate boys, both future homosexuals and heterosexuals (Saghir & Robins, 1973, pp. 18-23) that occur during childhood. These beatings occur because the boys do not confirm to the extremely rigid rules of the male gender role. They also reaffirm the offender's commitment to that role before his peers. Psychologically, they serve the same function as the more serious gay-bashing of adulthood. They differ from the latter in two ways, however. First, they are more accepted in conventional adult norms. Second, they do not suggest as much criminality and probable moral disengagement from the norms of civil society on the offenders' part as does adult gay-bashing. Culturally, however, the childhood and adult incidents are the same. Whether persons who engage in adult gay-bashing have also engaged in childhood 'sissy-bashing' is unknown.
If we view the above three age-based types of incidents as gay-bashings that differ only in the ages of the participants involved, the ideas offered to explain anti-gay violence may not predict too much. Gay-bashing may be endemic during childhood and decline in frequency with age while at the same time it increases in seriousness and leathality. As males approach adulthood, most become more secure in their gender roles, so that proving their gender adequacy becomes less obsessive and gender deviance in others becomes less salient. Hence the motivations for gay-bashing may decline with the advent of an adulthood that is not defined in the stark imagery of the immature male.
THE SITUATIONS OF ANTI-GAY VIOLENCE
The views of gay-bashers are clearly in agreement with those of the large majority of the population who disapprove of homosexuality (see Chapter 5). Reporting date from the General Social Survey (National Opinion Research Center, 1988) in 1987, 82% of the population found homosexuality "always wrong" or "almost always wrong." This percentage has changed little since 1973 and may have increased slightly since the 1970s. For purposes of analysis, we divide this 82% into three categories. Most strongly opposing homosexuality are a small number of activists who go out of their way to find homosexuals to assault. Such strongly motivated persons would typically go to a place where homosexuals are known to gather such as a gay ghetto (Levine, 1980) or to the environs of a gay bar. Somewhat less opposed to homosexuality would be the larger number of opportunists who are not sufficiently motivated to seek out homosexuals to victimize but will assault them as occasions arise. Such situations would typically arise in non-gay-defined settings when persons who are visibly homosexual appear. The remainder of the 82% are those who disapprove of homosexuality but not strongly enough to engage in gay-bashing. This group is theoretically important because it is by far the largest of the three and it consists of those who might normally be expected to serve as guardian citizens in cases of assault (Cohen & Felson, 1979). In the case of common crimes among heterosexual participants, such guardians serve the function of being interveners or of calling the police. In cases of anti-gay violence, however, it is doubtful that many of this large group who disapprove of homosexuality would be willing to actively assist the victim.
<#FROWN:J44\>
AMERICAN LEGAL THOUGHT AND LEGAL REFORM
A. Introduction
The Federal Rules of Civil Procedure, implemented in 1938, and the Federal Rules of Evidence, enacted in 1975, are designed, we are told, to promote the "just, speedy, and inexpensive determination of every action." They are to be "construed to secure fairness in administration, elimination of unjustifiable expense and delay, and promotion of growth and development of the law of evidence to the end that the truth may be ascertained and proceedings justly determined." The stated goals of these transsubstantive rules, then, are that the truth be determined and disputes be justly resolved.
The foundational assumptions underlying the claim that the Federal Rules of Civil Procedure and the Federal Rules of Evidence are instruments that permit the discovery of truth and the 'just' resolution of disputes are three related phenomena: first, the general 'optimistic rationalism' pervading most of Western legal and intellectual thought from the Enlightenment; second, the legal 'progressivism' of influential early to mid-20th-century American reformers, who acted as catalysts for both procedural and evidentiary reform; and third, the jurisprudential reaction to American legal realism, which coalesced after World War II into legal process or reasoned elaboration.
The Federal Rules of Procedure and the Federal Rules of Evidence were explicitly presented as means to the goals of 'truth' and 'justice' in part due to this broader Western and narrower American intellectual milieu. These goals were also channeled by a deep public and professional reverence for both Law and the Rule of Law. Finally, the legal profession was dedicated to the beauty and utility of the adversary system, the hallmark of the 'Anglo-American' system of adjudication. These explicit statements were not part of the wellspring of the Federal Rules of Evidence, the American Law Institute's 1942 Model Code of Evidence. Part of the failure of the Model Code of Evidence was due to its apparent disavowal of these goals.
The abiding belief of early 20th-century legal progressive thought was that legal reform could rationally aid in the progress of a legal system toward consensual notions of 'truth' and 'justice'. Legal realism, while having little contemporary impact on the legal profession, shattered a jurisprudential faith in legal progress toward truth and justice. The restructuring of legal progressive thought into reasoned elaboration or legal process after World War II required a fundamentally different justification for a 'rational' and progressive administration of justice. This justification, however, was unacceptable to a legal profession then essentially unaffected by legal realism. While legal academics could not longer faithfully argue that the goal of the trial was truth, nor that the administration of justice was concerned with substantive rather than procedural justice, the legal profession and the public continued to believe in both goals. Invoking the goals of truth and justice to garner public and professional support was necessary to the passage of the Federal Rules of Evidence; the structure of the Federal Rules, because it is based on the Model Code of Evidence, undermines those goals.
B. Optimistic Rationalism
William Twining describes the tenets of 'optimistic rationalism' as a congery of beliefs in truth, reason, and justice under law. Events occur independently of human observation, and past events can be truthfully reconstructed in the present, although "establishing the truth about alleged past events is typically a matter of probabilities or likelihoods falling short of competecomplete certainty." Ascertaining the truth is accomplished by listening to experts explain and interpret relevant data and through the 'common-sense' generalizations of society. In adjudicating disputes, establishing the truth must be based on relevant evidence and justice can be accomplished only if the truth is established on the basis of relevant evidence. Further, justice can be accomplished only if the method of fact finding is 'rational.' Rational decision making means making decisions based on inferences from relevant evidence. Rational decision making based on relevant evidence will thus lead the fact-finder to the truth and to 'correctness' in decision making. The search for truth, then, is at the core of a system of justice. Since, however, decisions about the truth of factual allegations occur in an imperfect, human setting, the concern for justice is not a concern for an idealized justice but a justice under [positive] law, which means that truth will not always be discovered or a correct decision rendered and further means that the goal of 'correctness' may be matched or superseded by other social goals.
The 'Anglo-American' system of adjudication - the adversary system - structures and channels these tenets of optimistic rationalism. Unlike trial by compurgation or trial by ordeal, the adversary system was perceived as a rational system for the discovery of truth and the pursuit of justice. In the adversary system, each participant, with the notable exception of the parties, plays a significant role in fulfilling the requirements of optimistic rationalism. The attorneys for the parties investigate and sift the facts pertinent to their (opposing) cases and offer and object to the introduction of evidence; the judge impartially decides disputed issues of law, including the admissibility of evidence; and the jury, given the conflicting evidence presented by both parties and instructions on the applicable law by the judge, decides the disputed issues of fact and renders a verdict for a party. This system provides checks on abuses by counsel (by the judge), by the judge (by counsel on appeal), and by the jury (through jury instructions, limiting their purview to issues of 'fact' and, in egregious cases, permitting the court to render a judgment notwithstanding the verdict or to inquire into the validity of the verdict), and so limits any departures from rationality.
C. Legal Progressivism and Procedural Reform
The story of the codification of the rules of evidence is further linked to the story of legal progressivism, for the interest in a code of evidence rules is based on the legal progressives' spirit of legal reform. In 1904-5, Wigmore's Treatise was published. This four-volume first edition was an immediate critical and commercial success. Dean Wigmore became the unchallenged authority on the law of evidence in America.
The publication of Wigmore's Treatise was "the most important event in the history of the law of evidence in this century." Wigmore's Treatise was not simply a compendium of cases and a rationalization of inconsistencies in the law of evidence but also a call for reform. If the legal system was to be a rational system for the discovery of truth, as Wigmore believed, the rules of evidence needed to be applied consistently with those goals and to be workable in practice, that is, in trials. Wigmore's ideas for reforming the law of evidence were part of the emergence of legal progressivism, or sociological jurisprudence, led by Roscoe Pound.
In 1906 Pound spoke at the annual meeting of the American Bar Association in St. Paul, Minnesota, about the reasons for public dissatisfaction with the administration of justice in American courts. Among the reasons for public dissatisfaction with the American legal system was contentious procedure, which turned litigation from a search "for truth and justice" into a game or sport "that the parties should fight out ... in their own way without interference." Decrying the sporting theory of justice, Pound cited Wigmore for the proposition that this view inaccurately depicted the adversary system. The sporting theory disfigured the administration of justice and mistakenly led even the "most conscientious" judge to believe that he was "not to search independently for truth and justice" and to assume that "errors in the admission or rejection of evidence are presumed to be prejudicial and hence demand a new trial." This gave the community "a false notion of the purpose and end of law."
Pound's call was for a true "scientific jurisprudence" based on the use of experts to make the legal system more efficient. Making judges 'scientists' would instill in judges an expertise which would create a greater efficiency in the administration of justice. It would also alter the administration of justice by creating an emphasis on substantive justice in the courts. Two years later, Pound fleshed out both these themes in a Columbia Law Review article. The science of law was a means to the end of "reason, uniformity, and certainty." A scientific jurisprudence was a search for full justice, for "solutions that go to the root of the controversies," for equal justice, and for exact justice. Law was scientific in order to eliminate "the personal equation in judicial administration, to preclude corruption and to limit the dangerous possibilities of magisterial ignorance." The scientific administration of justice, however, was not to be confused with a mechanistic jurisprudence, although a degeneration of legal science could lead to stagnation and "petrification" in the legal system.
The antidote to the problem of "petrification" was "a pragmatic, a sociological legal science." "The sociological movement in jurisprudence is a movement for pragmatism as a philosophy of law; for the adjustment of principles and doctrines to the human conditions they are to govern rather than to assumed first principles; for putting the human factor in the central place and relegating logic to its true position as an instrument."
Pound then noted that the law of procedure and evidence suffered "especially from mechanical jurisprudence." An insistence on perceiving procedure and evidence in conceptual terms led judges to view them as ends rather than means, and Pound gave examples of this error. He concluded by suggesting the enactment of "a common-sense and business-like procedure."
The advent at the beginning of the 20th century of sociological jurisprudence, also known as legal progressivism, progressive proceduralism, and progressive-pragmatism, was part of the general progressive movement and specifically part of the intellectual departure from formalism. Pound, the progenitor of sociological jurisprudence, relied, like all good progressives, on the "ideology of bureaucracy" to support his efforts at reforming the legal system. In general, "[p]rogressivism believed in the management of government by experts and advocated the expansion of the executive branch, primarily in the form of administrative regulatory agencies, at the expense of the Congress and the courts." Specifically, formalist jurisprudential theory employed a priori reasoning rather than reasoning based on actual economic and social conditions. The use of disinterested experts in adjudication would make the administration of justice more rational and just. Such reform was a gradual reform, conservative in the sense of taking the best from the American past and molding it to the present. Political and legal progressives, as their name suggests, believed in the evolution of human progress, a gradual but continued movement toward greater enlightenment about the human condition. As advocates for efficiency, expertise, and progress, progressives claimed that their movement was nonideological. All bureaucrats, including judges, if correctly trained and learned as 'scientists,' could act disinterestedly in support of progress. Finally, some legal progressives, including Pound and Wigmore, believed in moral absolutes. While society's values were often inchoate and in flux, there was some consensus about values.
Pound's ideas for legal reform gradually captured the attention of influential academics and 'elite' members of the legal profession. In a 1937 article looking back at the early proposals for legal reform, Wigmore called Pound's 1906 speech "the spark that kindled the white flame of progress." Wigmore noted that on the morning after Pound's speech was given, he met with William Draper Lewis, then of the University of Pennsylvania, and they, along with others, "resolved to do something about it in our own limited spheres." In 1936, a writer discussing the third draft of the proposed Federal Rules of Civil Procedure in the American Bar Association Journal traced the movement for reform of the rules of civil procedure to Pound's 1906 speech.
When Pound spoke to the American Bar Association, he was dean of the University of Nebraska School of Law, a "hitherto obscure Nebraska jurist." Two years later, Wigmore recruited Pound to Northwestern, and shortly after that, he was named Story Professor of Law at Harvard Law School. By 1916, Pound was dean at Harvard, and during his 20-year term he consolidated Harvard's preeminence in legal education. The preeminence of Pound at Harvard and Wigmore at Northwestern eased the transition of the legal academy from formalist to 'progressive-pragmatist' notions of jurisprudence.
<#FROWN:J45\>
Chapter 5
Rebuilding the American City
There is a potential cycle for change. It begins with the local problem of urban poverty and central-city decay, then moves to local public recognition, which generates a local response. That response is severely constrained and confounded by lack of resources and power. In the best of circumstances - and we will argue the case for this - the conflict between attempts to deal locally with the problems of poverty, on the one hand, and lack of resources, on the other, will lead to coalitions and pressures on Congress, the federal judiciary, the White House, and federal agencies. In the face of these pressures, Congress will pass better federal laws and offer more generous budgets, the executive branch will better regulate the national economy, and industry will develop a more progressive response to competition in the global economy. These changes, in turn, will lead not only to better conditions, such as stronger labor demand, more attention to education, and broad health care coverage, but will also provide the funds municipalities need to become better places in which to work and live.
Changes of this sort will not happen automatically or easily. Even when reforms begin, desirable as they would be, major changes in either markets or policy are unlikely in the short run. Neither is any set of partial reforms likely to 'solve the poverty problem.' Cognizant of these severe limitations, in this chapter we aim to be practical, to search for means by which - at the least - the serious problems of urban poverty will get written prominently into the political agenda.
It is not enough to call for a return to generous, liberal federal policy. Neither our analysis and the recommendations we make nor the excellent and more detailed proposals of others will stimulate governmental generosity. The authors of such proposals have no access to the White House basement, where they might push good legislation through Congress, to remake the country in their (and our) better image. Instead, we believe, better policy to minimize poverty will result only from new political forces, which are most likely to be rooted in the poverty of the central city. We believe, that is, that an urban political strategy is the most practical approach for attacking America's poverty problems.
The time is ripe for this plan. City governments are poor and weak, and although they would like to solve the poverty problem, they are unable. The federal government, so distant from urban poverty, is preoccupied with international economic and political affairs. But as the problems mount, city officials and community-based organizations will increase their pressures and try to form new political coalitions. As these problems threaten national productivity, new solutions will become more attractive to various national groups, such as industrial leaders who fear for their international competitive advantage. If these city-based coalitions can be formed, then inroads can be made to improve federal policies and transfer some real power to the cities, and a cycle of positive feedback can begin.
This argument will proceed, section by section, through this chapter. First, we review the history of federal-local relations in fighting poverty. We begin by pointing out that federal aid has drastically declined. Cities are short of resources and nearly powerless in the face of suburban disparities and economic pressures from big business. The situation has been made worse by the rivalries forced on cities by federal programs and their antineighborhood bias.
In the second section, we provide a selection of proposals for sensible, efficient, and efficacious federal programs to solve the urban poverty crisis. We observe the various options for public policy. The major portion of this section is devoted to a review of proposals for better federal policy. It is well for the reader to recall that the national response to global economic pressures can vary: Japan, France, Germany, Italy, and the Scandinavian countries, for example, have adopted policies considerably different from those adopted in the United States. Even in Great Britain, intercity rivalry is less destructive, because national laws and budgets provide a common base for family and urban services. In particular, countries make political choices among technical options to help guide capitalist development. The United States has chosen, partly by lack of plan, regressive policies that guide choice of technology and work arrangements in counterproductive directions. The country could, however, plan more progressively. Reforms could encourage the educating and strengthening of the workforce, from the bottom up. This would be in contrast to the current practice of dividing and further separating labor, destroying opportunity for those at the bottom.
In the third section of this chapter, we examine the potential for political support to implement governmental programs. We raise a troubling question: from where will political support come for these reforms? We briefly review four possibilities, but feel compelled to judge three of them unlikely. The fourth, which stresses the latent strength of grass-roots politics in cities, leads us to the last section, where we focus on strengthening the urban role in the quest for better policy. There we will turn briefly to the heart of the matter - how we may work collectively inside cities to gather political support to fight poverty.
We examine the possibilities for a renewed and revived municipal politics. We first observe that one way to attack the set of problems treated in this book (poverty, low productivity, social division, and urban decay) is through local, progressive experiments. Their success has been documented in several cities. Chicago, Hartford, and Cleveland are among the examples, along with the more widely discussed but smaller city experiments in Burlington, Santa Monica, and Berkeley. If these experiments were to be multiplied and extended, they could show the way to the needed reconstruction of urban America. The evidence suggests there is room for municipal maneuvering in spite of the dismal prospect of a continued negative federal policy toward global competition, and it also suggests what kinds of programs are most effective.
We are more optimistic, still. When enough local change takes place, and when more experiments arise from the economic demands and political pressure of impoverished ghetto populations of African Americans, Latinos, and recent immigrants, they will provide the stimulus for coalitions to fight for better national policies for raising productivity and improving the U.S. response to global challenges. Once there are better national policies, they will stimulate still better local reforms, and the cycle may reinforce positive change.
Federal Aid, Municipal Expectations, and Antipoverty Programs
We open this section on federal-urban relations with a brief response to conservative pronouncements on the problems of the poor and the central city.
A Note on Neoconservatism
The American city shows a pressing need for more adequate national-level policies. The core of the metropolis is failing. Central cities are falling apart physically, economically, and socially. Whole neighborhoods are decaying, the people in them are suffering, and social disorganization threatens entire cities. Far too many people are poor across the nation, not just in the cities and not only when out of work. Their numbers are not declining, even during what the indicators say are economic good times. A generation has reached adulthood in poverty, and the children of that generation are threatened with worse. The gulf between haves and have-nots in this country has never been greater, and political communication never worse.
Few can doubt that the United States needs a new approach to problems of poverty, nor can they doubt the needs of the central city. It is difficult, therefore, to accept conservatives' arguments that we should leave well enough alone. It is hard to believe their theories that the situation will get better by itself. The evidence of the 1980s casts great doubt that problems of poverty will be resolved or even seriously reduced by benefits trickling down from general prosperity.
The conservative argument has been much popularized, but it is false. Most troublesome for our work at this juncture is a tendency in much contemporary discussion to use rhetoric that at once trivializes systematic causes of poverty and magnifies the problems thought to derive from improper individual behavior. To put this bias in context, we borrow ideas from political economist Albert Hirschman, who has examined the problem of rhetoric in a broader but closely related context.
The rise of the welfare state in the twentieth century, Hirschman asserts, can be seen as the third stage in a protracted zig-zag struggle over centuries for the "development of true economic and social citizenship." The first stage was the back-and-forth struggle for civil rights of speech, thought, religion, and justice. The second stage involved the effort to win political rights by extending the vote; and the third stage was the broader struggle to expand social and economic rights, "recognizing that minimum standards of education, health, economic well-being, and security are basic to the life of a civilized person." Arguments for and against these developments of modern society have used greatly exaggerated rhetoric: progressives extoll the advantages of expanded rights, while conservatives warn of dangers. At each stage there may be progress, followed by proposals that attempt to undo the most recent gains. We are now in a period when 'reactionary rhetoric' is particularly prominent.
Rhetorical and ideological backlashes stem not simply from gloomy estimates of human capacity (as by Edmund Burke on the French revolution or Thomas Malthus on the utility of starvation for checking the growth of the English working class), or from fear by the privileged classes that derives from their being outnumbered by the common people. Support for reaction is also provided by theoretical predispositions of the social sciences, especially the myth of self-regulating economics, which allows free-market enthusiasts to denounce as strongly 'perverse' any effects from progressive interferences with the 'natural' laws of supply and demand. The argument that welfare is the cause of poverty is a prominent example of this sort of reactionary argument, neatly echoing centuries of similar reaction to various stages of progress.
The ideological onslaught of the last twenty years against redistributive policies has been widely justified in terms of national economic policy. Although the most negative and racist accompaniments of this policy have been kept usually out of sight, the agendas of those who abuse the theories of free-market economists and other archconservative social scientists have sometimes been transparent. The theories lend themselves to this abuse, as is suggested by the quantity of 'counterintuitive' reasoning to which we have been subjected. Simulation models are designed to show that "at times programs cause exactly the reverse of desired results," as would be the case, for example, if by providing good housing for the poor the City of Boston would attract impoverished migrants and therefore worsen its average housing conditions. It is claimed that "our efforts to deal with distress themselves increase distress." Conservatives argue that "we tried to provide more for the poor and produced more poor instead. We tried to remove the barriers to escape from poverty, and inadvertently built a trap."
These expectations of counterintuitive, reversed, and inadvertent consequences of progressive social policy exist more in the flawed reasoning of the right-wing critics than in reality. Although unanticipated consequences do often result from public (and private) actions, it is important to recognize that, as Hirschman points out, "there is actually nothing certain about such perverse effects." It is claimed by conservatives, to take but one example, that minimum-wage legislation dries up jobs for the poor by making labor too expensive. But there is in fact little such evidence, and it could be in theory that a higher legal floor to wages would have precisely the intended salutory influence, that is, higher minimums would have "a positive effect on labor productivity and consequently on employment." As the terms of public debate have shifted so as to frame a more conservative and less compassionate view, reformers have more and more difficulty defending in public perfectly reasonable attempts (such as the legislation of a higher minimum wage) to improve basic conditions for the poor.
<#FROWN:J46\>
International Evidence on the Historical Properties of Business Cycles
By DAVID K. BACKUS AND PATRICK J. KEHOE
We contrast properties of real quantities with those of price levels and stocks of money for ten countries over the last century. Although the magnitude of output fluctuations has varied across countries and periods, relations among real quantities have been remarkably uniform. Properties of price levels, however, exhibit striking differences between periods. Inflation rates are more persistent after World War II than before, and price-level fluctuations are typically procyclical before World War II and countercyclical afterward. Fluctuations in money are less highly correlated with output in the postwar period but are no more persistent than in earlier periods. (JEL E32, E31)
We study fluctuations in output, prices, and money in ten countries for which at least a century of annual data are available. We also examine the cyclical behavior of components of national output: private consumption expenditures, fixed investment, government purchases of goods and services, and net exports. Our objective is to document some of the salient features of business cycles. We know that in many respects these countries and time periods have been markedly different. The ten countries differ in their institutions, their monetary and fiscal policies, their industrial compositions and structures, and their average aggregate growth rates. The question is whether they share, despite these differences, similar features of business cycles.
We find a great deal of regularity in the cyclical behavior of real quantities. Although the magnitude of output fluctuations varies across countries and over time, relations among variables are remarkably stable. Investment is consistently 2-4 times as variable as output; consumption is about as variable as output; and both investment and consumption are strongly procyclical. The trade balance is generally countercyclical, exhibiting larger deficits during booms than during recessions. The exception to this regularity in quantities is government purchases, which exhibit no systematic cyclical tendency. Patterns of price-level fluctuations, however, have changed markedly. Before World War II, prices were predominantly procyclical; since then, they have been consistently countercyclical. They have also been, in most countries, substantially more persistent since World War II than in earlier periods. We also find for the post-World War II period that fluctuations in the stock of money have been less highly correlated with output. There is no general tendency across countries, however, toward greater persistence of money growth rates.
Our study is an outgrowth of business-cylce research by Robert Lucas (1977), Finn Kydland and Edward Prescott (1990), and others that, in turn, retains some of the flavor of the tradition of Arthur Burns and Wesley Mitchell (1946) at the National Bureau of Economic Research (NBER). The goal of this work is, for the most part, to summarize the properties of macroeconomic data without imposing much theoretical structure. The resulting empirical regularities can then serve as a guide to a variety of future theoretical developments. A common theme in this line of research is that the business-cycle phenomenon consists not simply of fluctuations in aggregate output, but also of common patterns of correlation between different aggregate time series. We report properties of international fluctuations in a manner that conforms with some recent work on American business cycles and thus extends this work to a much wider range of countries and time periods. Our motivation is international in another respect: our own research (Backus and Kehoe, 1987; Backus et al., 1992) concerns the dynamics of inernational trade and the relationships among business cycles in different countries. A useful by-product is additional evidence on the question of whether output fluctuations since World War II have been smaller than those prior to World War I. This question has been the subject of active debate in the United States, including papers by Christina Romer (1986, 1989), Steven Sheffrin (1988), and Nathan Blake and Robert J. Gordon (1989). Like Sheffrin's study, ours puts this debate in an international context. We include several countries not studied by Sheffrin, notably Australia, Canada, and Japan, and introduce new data for Sweden.
Our data set covers ten countries with at least a century of annual data on national output: Australia, Canada, Denmark, Germany, Italy, Japan, Norway, Sweden, the United Kingdom, and the United States. For the most part, countries with national income accounts for such a long period are also those with the highest per capita output today. Several others, including India, report partial time series, but we doubt that these series are sufficiently accurate for the study of short-term fluctuations. Estimates of national output in the ten countries vary in quality, but in some cases we think they are superior to the U.S. data.
We begin, in Section I, by describing the data. While data for earlier periods are unquestionably less reliable than modern data, in some countries they appear to be good enough to provide an accurate picture of business cycles prior to World War II. The data for several countries seem to be significantly more accurate than the Kuznets-based estimates for the United States, primarily because raw-data sources are better in these countries. In Section II, we compare output volatility before World War I (the prewar period), after World War II (the postwar period), and between the wars (the interwar period). Until recently, the presumption has been that prewar U.S. output fluctuations were two or three times larger than those of the postwar period. Romer (1989), however, suggests that at least part of this difference is the result of systematic measurement error in prewar GNP that overstates its cyclical variability. Our international data set provides additional evidence on this question.
For the ten countries, we find that interwar fluctuations in real output are uniformly larger than those of the postwar period. With the single exception of Japan, the standard deviations of output fluctuations are from two to four times larger in each of the ten countries. We find, however, no consistent pattern for the prewar-postwar comparison. In six of the ten countries studied by Sheffrin (1988), prewar fluctuations are no more than 60-percent larger than those of the postwar period. However, in the other four (Australia, Canada, Sweden, and the United States) the fluctuations are considerably larger in the prewar period. The U.S. case has been discussed extensively, and it appears that part of the excess volatility of the prewar period can be attributed to measurement error (Romer, 1989). Romer's preferred estimate of prewar volatility is only 30-percent higher than for the postwar period, but Balke and Gordon (1989) argue for a number closer to 100 percent. Sheffrin (1988) considers a similar case for Sweden and concludes that the excess volatility in the prewar era is not primarily the result of measurement error. We find, as do Michael Bergman and Lars Jonung (1989) with different methods, that about half of the excess volatility Sheffrin finds in the prewar period disappears when revised as estimates of prewar output are used. Australia and Canada have the most extreme differences between periods, with output three and two times more volatile, respectively, in the prewar period. The data for both countries are reasonably good, so the greater volatility of measure output probably indicates a change in the variability of real economic activity.
In Section III, we examine the behavior of components of the national product: consumption, gross investment, government spending, and net exports. We find that many of the properties of postwar business cycles in the United States are evident in other countries and periods. Consumption expenditures have been procyclical and have approximately the same standard deviation as output. Investment has also been uniformly procyclical and generally varies, in percentage terms, from two to four times more than output. Government spending has generally been more variable than output, but it has been countercyclical almost as often as procyclical. Net exports have been, for the most part, countercyclical. We also find that correlations between measured output movements in different countries are typically positive and more pronounced in the postwar period than in the prewar period.
In Section IV, we examine movements in price levels and money stocks. Here we find, in contrast to the regularity of real quantities, two significant changes in the cyclical behavior of prices. We find, first, that price changes in most countries have been more persistent in the postwar period than in the prewar period. This finding extends related work by Jeffrey Sachs (1980), Charles Schultze (1986), and John Taylor (1986) on the United States and work by Gordon (1983) on the United States, the United Kingdom, and Japan, to a larger set of countries. We also find, in the prewar and interwar periods, that output and price-level fluctuations are positively correlated in most of the ten countries. However, in the postwar period, price fluctuations have been consistently countercyclical. We find a slight decline in the correlation of money and output in the postwar period, but no general tendency for greater persistence of money growth rates. We conclude this section with some speculative remarks on potential explanations for the observed changes in price behavior.
I. The Data
We start with a description and evaluation of the data, emphasizing in particular the methods used to construct prewar national income accounts; sources and definitions are described in Appendix A. Although national accounts are based to a large extent on a common framework, sources of raw data differ across countries, especially in the prewar period. Countries with the best source material tend to have the most reliable estimates of national income. The United Kingdom, for example, has had an annual income tax in effect continuously since 1842, while in the United States the federal personal income tax was only made possible in 1913 by the Sixteenth Amendment to the Constitution. As a result, the United Kingdom has much better data on the income side of the national accounts for the prewar period than the United States. In other countries, the establishment of statistical bureaus to measure production and employment, frequently on an industry basis, makes production-based accounts feasible. In the United States such sources of annual data are extremely limited. For this reason, and because accounting methods have improved in the decades since Simon Kuznets's (1961) work on prewar U.S. GNP, estimates for several of the countries we study are likely to be better than the American data examined by Romer (1989) and Balke and Gordon (1989).
Problems with prewar U.S. data have been well-documented by Kuznets (1961) and Romer (1986, 1989). Kuznets and his coworkers constructed national income accounts for the United States from 1869. The cornerstone of this work, and most later work as well, is William Shaw's (1947) commodity output series: estimates of value added in manufacturing, mining, and farming. Shaw's estimates, and therefore those of national income, were severely constrained in the prewar period by the absence of comprehensive annual data sources. The most informative sources (see Shaw, 1947 part II) were periodic federal censuses, including especially the Census of Manufactures, available every ten years from 1869 to 1899 and every five years from 1899 to 1919. One source of annual data is reports on industry published by eight states. These states accounted for between 10 and 39 percent of total manufacturing in census years, and the reports typically covered only part of each state's manufacturing output (see Shaw, 1947 table II:4). The state reports were supplemented with occasional government reports and industry publications. Kuznets (1961) interpolated further between the census benchmarks of 1869, 1879, and 1889 by using a variety of industry-output indicators (see the notes to tables II:1-5 in Kuznets [1961]), since neither he nor Shaw was able to measure commodity output directly for the 1869-1889 period. Finally, both Shaw and Kuznets estimated nominal value added, which was converted to real terms at a disaggregated level using producer price indexes.
Romer (1989), however, bases her criticism of prewar U.S. data not on the fragmentary source material used to produce estimates of commodity output, but on the method Kuznets (1961) used to extrapolate from commodity output to GNP. Kuznets's problem was to estimate GNP from information on commodity output alone, since direct measures of other components were not consistently available even for census years.
<#FROWN:J47\>
AFRICAN-AMERICAN SOCIETY AND EDUCATION
Letha A. (Lee) See
The study of social inequality in the United States has properly focussed on the fate of African-Americans. Although other minorities have endured privations based on their language or religion, their identity is in their own hands to some extent: there are no physical barriers to a change of an individual's tongue or faith. This is not so with race. Gender would similarly define identity by nature, but even though women are still treated unequally across almost all racial and ethnic groups, civilized societies increasingly denounce sexual discrimination.
The African-Americans, at twenty-eight million, considerably outnumber the Native Americans (about one million), the Asians, or the mixed race Hispanic community (sixteen million). Their situation still constitutes the 'American problem' that has been identified for generations. It should be clear that the problem is only partly theirs to solve, for the society as a whole must change too, and would benefit immensely from its solution.
On the other hand, there is evidence that a minority of African-Americans have succeeded, despite the inferior opportunities available to most of their members. Of course, without their handicap of widespread victimization, these notables would likely have won achievements which were more substantial, earlier, easier, and achievements might have been recognized over a broader range of ventures.
This chapter deals briefly with four aspects of inequality in society to set the context for a discussion of education: income, housing, criminal justice, and health care. Other data on inequality (such as social class) in the United States are drawn upon where appropriate, but this chapter argues that African-Americans exist as a statistically significant sub-group within most of the other categories of disadvantage found in American society - in fact a statistically larger share of disadvantaged categories than would be expected. In short, race does not explain everything, but if you are black and in the United States, it has a pervasive inhibition on opportunities of every kind. Educational programs of schooling and teacher education are then addressed to see where intervention is most promising. Self-help programs are identified, recognizing the difficulties of securing broad public support.
Background to Inequality for African-Americans
African-Americans came to the United States as slaves in most cases. Although they sometimes came via the British colonies of the Caribbean, most came directly from Africa. Although other peoples were sometimes enslaved and the Africans mixed with various other races, slavery remained the dominant experience of their group (more or less exclusively) until the Civil War. From the time that slave trading was abolished during the 1830s (with legislation from several states reinforced by the effective blockade of slaving by the British navy) the numbers of African-Americans have grown by natural causes rather than by continued migration. Only a few thousand have left the U.S. for other nations such as Liberia or Canada. In short, for over 150 years the United States has been the only home of African-Americans, and for 125 years they have been citizens. But not equal ones.
Deprived of the vote initially, threatened by lynch mobs until the present generation, denied equal access to many public services in both government and private institutions, the African-American is still not able to enjoy equal status. This inequality remains institutionalized although no longer formalized in law. Most of the social functions of American life create separate categories for white and non-white, and black is both the largest and probably the most disadvantaged group among the latter. At the personal level, racism may not be evident, but almost any set of statistics can be broken into categories that reflect the racial exploitation. Of course, current data do not describe a society that is inevitable or desirable. Since the systems they describe are capable of being changed if there is sufficient social and political will, education has an important role in improving the deplorable conditions revealed in studies of other aspects of society: income, housing, health, and crime.
Income
Economic developments for African-Americans in the United States reflect the continuum of possibilities. Blackwell (1985) asserts that "many segments of the black community experienced major economic progress" between the Civil Rights Act of 1964 and the election of Ronald Reagan in 1980. U.S. News and World Report (1986) claimed that the majority of African-Americans are prospering and that they had doubled their proportion in the middle class. But these two articles still portray these gains as insignificant in the context of economic inequality. The Reagan period brought appalling reverses in economic equality for the majority of African-Americans. An Urban League report entitled The State of Black America (Swinton 1989) shows per capita income for African-Americans remained steadily at 51 percent of that for whites. The aggregate incomes of both groups continued to grow, but the income gap widened by about $2000. Black family income was only 42 percent that of whites and spelled disaster for large families in the urban setting.
U.S. News and World Report (1986) reported that 1.1 million black males were unemployed in 1985, compared to .5 million in 1970. Half of all black teenagers who had started a job are now unemployed. The explanation offered by the Department of Commerce (1986) is that there are no longer well-paid jobs in manufacturing where unskilled blacks would be employed. In the last fifteen years, of twenty-three million jobs created in the private sector, more than 90 percent were in the service sector.
High unemployment rates among African-Americans are related to the entry of new immigrant groups into the employment force, and to the structural shift from low-skill jobs to jobs requiring technical skills (See 1986). The internationalization of the world's labor force (partly because American businesses are opening factories offshore) points to an increasing need for American youth to have job training and general education. Indirect evidence indicates that investment in public schooling can be partly offset against added costs of long-term unemployment, or can provide a partial solution to the high rate of unemployment among blacks, thereby providing a respectable cost/benefit argument for offering training services.
Black unemployment and poverty are both high, and high relative to the figures for white Americans. The 1986 Census Bureau reports that 33.7 million people (14.4 percent of the U.S. total) are poor. Among African-Americans, the rate is 33.8 percent, for black children 51.1 percent, and for the black elderly 31.7 percent. Blacks are three times as likely as whites to find themselves in poverty. Black families headed by women are twice as likely to be poor. Swinton (1989) contends that not only did African-Americans not share in the recent economic revolution, but the black-white gap is growing. The black poverty rate doubled from 1969 to 1988 (12 percent instead of 6 percent) and unemployed tripled (1.7 million instead of 570,000).
This economic inequality in America reflects upon the marginal participation by African-Americans in the economic community. Not only are there smaller numbers but the nature of the jobs makes workers vulnerable to displacement from automation, technological changes, and shifts to off-shore operations. Unskilled and semi-skilled jobs are disappearing at the rate of 35,000 per week, nearly two million per year. This pressure on the African-American community is curiously functional, for Gans (1974) noted that any social system can ensure that its 'dirty work' is done at low wages if there are no alternatives for part of the work force.
African-Americans are losing ground, giving rise to a nation of the truly disadvantaged. Evidently America is growing into two nations, one black, one white; separate and unequal.
Housing
Blackwell (1985) argued that African-American housing should be judged by the standards created and used by the empowered Americans - the whites. By these standards, African-American housing is grossly overcrowded, substandard, and expensive. It contributes to homelessness even as it provides a limited form of housing.
This situation evolved as a succession of Republican administrations shifted the focus from construction of public housing to permitting private landlords to build or convert rental units for eligible families. These changes were followed by a series of rental subsidies (Bell 1970). To reach provisional agreements with landlords, landlords were allowed to subdivide existing apartments into exceedingly small units, resulting in overcrowding becoming commonplace (Forman 1978).
The population density in public housing was three times as high for African-Americans as for whites. In 1980, black families were larger than those of whites, but their apartments were smaller. The 1986 Census also indicated that a substantially larger number of ancillary indvidualsindividuals resided with black families.
There are not enough housing units for the African-American population, and the existing units cost too much. Today's problems arose from the urban renewal programs of the late 1950s and the 1960s, when the goal was to demolish blighted areas and construct new dwellings, office buildings, and highways. African-Americans were forced from their homes and traditional neighborhoods and the reduced number of available housing units became available for the poor in new ghettos (Gilderbloom 1989). The Reagan administration added to existing problems by massive cuts in domestic spending, including a large reduction in housing aid. Spending for low income housing fell from $32 billion in 1980 to $7 billion in 1988. This national agenda halted construction of low income housing despite residential density being at its highest levels ever, housing problems at their maximum.
Another consequence of these government housing policies is that some African-Americans have no housing at all. It is estimated that less than 2 percent of new housing guaranteed by Federal Housing Authority (FHA) mortgages is available to African-Americans. For non-government housing, the percentage open to black people is even smaller. One-third of the 23 million African-Americans now live below the poverty line (U.S. Bureau of Census 1988). It is highly probable that they live in desolation and squalor, devoting an increasing portion of their income to rent. Waiting lists for public housing swell dramatically, forcing many cities to close off new applicants.
Inequality is evident in the housing shortage, in racial exclusion, and in homelessness. The exclusion of persons from residential areas because of their race, color, creed, or ethnic attachment, despite their needs and ability to pay denies African-Americans a fundamental right. While complete freedom of selection is never achieved, compulsory or manipulated segregation is inherently wrong, damaging both for the immediate victims and for the general public. Housing segregation leads directly to segregation in other areas of life: schools, churches, hospitals, public accommodation, recreation, welfare and civic activities, and the workplace. Although segregation of schools is a violation of the orders of the Supreme Court, many schools of the north and west are segregated not by law but by racial patterns of residence.
Health
Universal health care is hotly debated in the United States. Conservative ideology suggests that those in need must fail in seeking help from their families and from the marketplace before they can depend on the government for medical assistance (Enthoven 1980, Hornbrook 1983). For poor people, seeking medical care from the marketplace drains them of hope and resources (Trevino and Moss 1983). The numbers are substantial: in 1983 the number of people living in poverty in the United States exceeded the entire population of Argentina, Australia, Canada, Sweden or Taiwan - in fact of all but twenty-three nations in the world. African-Americans represent a large number of those in poverty.
In 1969, 19.9 percent of African-Americans sixty-five years of age or more could not work because of ill health. (Only five percent of whites were in the same situation.) Low incomes for African-Americans explain many of these discrepancies, as they have done throughout the century. At the beginning of this century, white men outlived black men by 15.7 years; white women outlived black women by 16.0 years. These gaps have continued to narrow throughout the century, to become respectively 6.8 and 5.3 years. Significant differences are evident in the proportion of each race that lives beyond the age of 65: 74.8 percent of white and 58.1 percent of African males; 85.7 percent of white and 74.9 percent of African-American females (Statistics of the United States 1986).
<#FROWN:J48\>
Academic Achievement in Mathematics and Science of Students Between Ages 13 and 23: Are There Differences Among Students in the Top One Percent of Mathematical Ability?
Camilla Persson Benbow
Iowa State University
The predictive validity of the Scholastic Aptitude Test-Mathematics subtest (SAT-M) was investigated for 1,996 mathematically gifted (top 1%) 7th and 8th graders. Various academic achievement criteria were assessed over a 10-year span. Individual differences in SAT-M scores obtained in junior high school predicted accomplishments in high school and college. Among students in the top 1% of ability, those with SAT-M scores in the top quarter, in comparison with those in the bottom quarter, achieved at much higher levels through high school, college, and graduate school. Of the 37 variables studied, 34 showed significant differences favoring the high SAT-M group, which were substantial. Some gender differences emerged; these tended to be smaller than the ability group differences; they were not observed in the relationship between mathematical ability and academic achievement. The predictive validity of the SAT-M for high-ability 7th and 8th graders was supported.
"Standardized testing is much in the news. New testing programs, test results, and criticisms of standardized testing all are regular fare in the popular media today" (Haney, 1981, p. 1021). Moreover, "with the possible exception of evolution, no area in the sciences has been as filled with emotional and confusing mixtures of science, politics, and philosophy as the field of mental testing" (Carroll & Horn, 1981, p. 1012). These remarks portray quite well the status of mental testing at the beginning of the 1980s, yet they seem to be equally appropriate for describing mental testing at the beginning of the 1990s. Some might perceive this as a rather recent development. However, concern over standardized testing has been voiced ever since the introduction of the Stanford-Binet Intelligence Scale and the Army Alpha test (Cronbach, 1975; Haney, 1981).
The concerns over mental testing primarily have been threefold: test bias against certain groups (primarily women and minorities at present, but children from families of low socioeconomic status in earlier decades), the role testing might play in perpetuating social and economic injustice, and the utility of test information (Cleary, Humphreys, Kendrick, & Wesman, 1975; Cole, 1981; Gottfredson & Crouse, 1986; Haney, 1981; Jensen, 1980; Scarr, 1981). The questionable value of test information has been a particularly frequent criticism levied against college admissions tests, such as the College Board Scholastic Aptitude Test (SAT; see Linn, 1982b, for a review). This study was conceptualized to address the latter concern, namely, the predictive validity of the SAT for a special population. I assess the value of the SAT, not for high school seniors and the college admissions process, but rather for identifying highly mathematically gifted seventh and eighth graders and making predictions about their achievement over a 10-year period following their SAT-Mathematics assessment. Specifically, I asked whether the SAT-M can detect individual differences in the top 1% of the ability continuum that bear on subsequent academic achievement in mathematics and science.
The use of the SAT to identify intellectually precocious students in Grades 7 and 8 dates to 1972 when Julian Stanley launched the first talent search (Keating & Stanley, 1972). Stanley was interested in students who ranked in the top 1 % in mathematical ability. Because considerable variance in academic ability is found among students in the 99th per-centile and because Stanley was interested in differentiating among such students, out-of-level testing (i.e., using tests designed for older age groups) was required. For that reason among others (see Stanley & Benbow, 1986), Stanley chose the SAT as the instrument with which to screen highly gifted students. Since 1972 more than 1,000,000 seventh and eighth graders have been tested with the SAT, and more than 100,000 such students now take the SAT annually through various talent search programs across the United States. The distribution of scores of such students on the SAT is about the same as found for a random sample of high school students (Benbow, 1988). The scores tend to maintain their ordinal ranking over time, increasing 40 to 50 points per year (Benbow & Stanley, 1982; Brody & Benbow, 1990; Olszewski-Kubilius, 1990). Thus, from a psychometric viewpoint, the use of the SAT with seventh and eighth graders seems justified.
It has not been demonstrated, however, whether use of the SAT with young but academically competent students has utility. Is the SAT a valid tool for assessing individual differences in current development, and can this instrument be used to refine predictions of exceptional academic achievements? As Cronbach (1971) pointed out, "validation is the process of examining the accuracy of a specific prediction or inference made form a test score" (p. 471). In assessing the validity of the SAT for highly gifted 7th and 8th graders, I evaluated whether academic achievement, especially in mathematics/science, during the 10-year period after these students were identified is much higher for those students with exceptionally high SAT Mathematics subtest (SAT-M) scores (top quarter of the top 1%) than for those with comparatively 'low' SAT scores who were nonetheless in the top 1% in ability (i.e., the bottom quarter of the top 1%). I hypothesized that meaningful differences would be detected. Several studies have revealed that individuals with the most potential for high academic achievement in mathematics and science are generally considered to be those students with high ability, particularly, high mathematical ability (Davis, 1965; Green, 1989; Walberg, Strykowski, Rovai, & Hung, 1984; Werts, 1967). Moreover, Kuhn (1962) noted that an overwhelming majority of "scientific revolutions" can be ascribed to the works of mathematically brilliant persons.
Nevertheless, many researchers and educators, most notably Renzulli (1986), have argued that there is a threshold effect for ability. According to this argument, after a certain point, there is a decline in the power of ability to influence academic achievement and other variables, such as motivation and creativity, become increasingly important. The precise location of this threshold for ability has not been determined. However, it is thought to be at some point well below the top percentile for ability. If Renzulli and others of this viewpoint are correct, then there should be no statistically significant differences in mathematics/science achievement between the two high-ability groups. All students in the top 1% should achieve highly, and placement within the top 1% should not affect the results.
The reasoning in the above paragraph assumes that there is only one threshold for ability. Yet there could be a threshold effect for ability within a certain range (e.g., between the 90th and 98th percentiles) but not within the top 1%. That is, differences in ability within the 90th and 98th percentiles may not relate much to subsequent academic achievement in mathematics/science. This view is reasonable given that the possible differences in ability within a range, for example, within the 90th-98th or 80th-89th percentile ranges, are small and not reliable in comparison with the ability differences found within the top 1% when out-of-level testing is used. I do not test this possibility in this study. If, however, one is interested in scientific eminence or productivity, and a threshold effect of ability for this level of achievement, it is within the top percentile of ability that one must focus.
Although my prediction is contrary to Renzulli's position, it should be noted that there are data that support the validity of Renzulli's position. For example, students who were in the top 1% in mathematical ability in the 7th and 8th grades were studied at 23 years of age to identify those factors that affect the ways in which childhood potential or ability is translated into adult achievement (Benbow & Arjmand, 1990). As a group these students had achieved academically at a very high level but not uniformly so. When those students who were classified as high academic achievers in mathematics/science areas (i.e., those who were attending graduate school in mathematics/science or medical school; n=261) were compared with those students in the sample who were classified as low academic achievers in those areas (those who were not attending college or had withdrawn, those who graduated with mathematics/science major but with low grades; n=95), a difference in previous ability between the two groups was found (the ability difference approximated two thirds of a standard deviation on the SAT-M). The canonical correlation (from the discriminant analysis) between (a) 7th-grade/8th-grade SAT-M and (b) high school SAT-M, SAT Verbal subtest (SAT-V), and achievement group membership was .30 for male students and .29 for female students. (Too few cases had 7th-grade/8th-grade SAT-V scores to allow inclusion in the analysis.) Nonetheless, ability exhibited the weakest relationship with academic achievement in mathematics/science as compared with variables in the areas of educational opportunity, family characteristics, and attitudes. Similarly, Sanders, Benbow, and Albright (1991) found that among mathematically talented female students, previous ability on SAT-M was not a primary factor relating to choice of mathematics/science career or to educational aspirations.
Thus, the aforementioned studies indicate that among those students in the top 1%, SAT-M performance was a factor but not the major factor predicting the students' academic success. That is, a bright mind will not make its own way. The educational opportunities provided to gifted children make a difference in the children's development. In the present study, I ask the central question: Do individual differences within the top 1% in ability make a difference in the eventual display of achievement?
In sum, I examine whether use of the SAT in out-of-level testing of highly gifted students yields useful information for the prediction of academic achievement up to 10 years after assessment. That is, is it useful to diagnose level of talent within the top 1%, as is currently being done with well over 100,000 seventh- and eighth-grade students on an annual basis? More succinctly, is there a benefit to knowing where in the top 1% a student's ability lies? It has been popularly assumed that such information is not helpful. In essence, I assess the predictive validity of the SAT for use with gifted 7th and 8th graders.
Method
Subjects
Intellectually talented students were identified by the Study of Mathematically Precocious Youth (SMPY), in which the SAT was administered to intellectually able 12- and 13-year-olds in the 1970s and early 1980s (Keating & Stanley, 1972). During that 12-year period, more than 10,000 preadolescents (mostly 7th graders) participated in SMPY 'talent searches.' (Since that time more than 1 million students have taken the SAT through other talent search programs.) About 3,500 of the students in the talent searches were included in the SMPY 50-year longitudinal study. As part of this study, researchers in the SMPY are currently tracking four cohorts of students and studying their development longitudinally.
Students in Cohort 1 comprised the sample in this investigation; they were drawn from the first three talent searches of the SMPY (i.e., those conducted in 1972, 1973, and 1974). In those talent searches, 7th and 8th graders in Maryland were eligible to participate if they had scored in the upper 5% (1972) or the upper 2% (1973, 1974) nationally on any standardized mathematics achievement test. Qualified students took the SAT-M and, in 1973, the SAT-V also. These tests are designed to measure developed mathematical and verbal reasoning ability, respectively, of high school students. However, the SAT is believed to be a more potent measure of reasoning for 7th and 8th graders than for 11th and 12th graders (Minor & Benbow, 1986; Stanley & Benbow, 1986).
A score of at least 390 on the SAT-M or 370 on the SAT-V in the 7th or 8th grade was required for inclusion in Cohort 1 of the longitudinal study. These SAT criteria resulted in the selection of 2,118 of 2,582 students who, as 7th or 8th graders, scored as well as the average high school female; the criteria also provided a wide range of talent to study. SAT scores had been grade adjusted (7th-grade scores had been adjusted upward to be comparable to 8th-grade scores, with the procedure outlined in Angoff, 1971). Mean SAT scores at age 13 were as follows for male students, 556 (SD = 73) on SAT-M and 436 (SD = 85) on SAT-V, and for female students, 519 (SD = 59) on SAT-M and 462 (SD = 88) for SAT-V.
<#FROWN:J49\>
CULTURE CLASH: A MODEL IN ACTION AMONG HAWAIIAN-AMERICAN CHILDREN
Children of color may begin the schooling process having been socialized in a way which may be in conflict with the expectations of the school; when this occurs, children and teachers may fail due to the cultural incompatibility between the culture of the school and the culture of the child. In this section, efforts to remedy the clash between the culture of the home and the culture of the school among one particular American minority group - Hawaiian-Americans - will be reviewed.
The cultural incompatibility approach has been the basis of considerable work at the Center for Development of Early Education (formerly the Kamehameha Elementary Education Program, KEEP), a privately funded, multidisciplinary, educational research and development program directed at remedying academic underachievement of native Hawaiians. As with many other ethnic minorities in American schools, poor school performance among Hawaiians was at first attributed to a variety of cultural and home deficiencies. This cultural deficit model, implying a superior-inferior dichotomy, is unfounded, unhelpful, and often rightfully labeled racist. All neurologically normal children have already learned a substantial amount of relatively complex material that is specific to their culture by the time they are of school age. Employing a cultural incompatibility model as opposed to a cultural deficit model implies that all children can learn prerequisite skills for any future need, including school readiness, if given the opportunity.
Researchers at Kamehameha schools proposed that a school environment that was compatible with the child's home culture could be developed. This culturally compatible classroom might elicit from children those skills, attitudes, and behaviors that would contribute to the desired learning and help children achieve early school success.
While research findings are numerous and complex, some summary can be attempted here. To begin with, the Hawaiian socialization system teaches children to be contributing members of a family. For instance, even when adolescents work outside the home, rather than spend their hard-earned money strictly on themselves, they often contribute to the overall family resources. the family is not seen as a training ground for independence as is typical in many dominant culture families. Personal independence is not a goal; rather, interdependence is stressed. A collective orientation develops, as opposed to the individual orientation prevalent in middle-class caucasian society. This has implications for motivation and instructional strategies, as well as the reward structure in the classroom. For instance, the Hawaiian child may not be motivated by individual rewards (gold stars, grades) as much as a caucasian counterpart. Nor would a Hawaiian child desire to achieve independence from the group.
The sibling care system, whereby children from a very young age are placed in the care of older siblings, also promotes a high degree of interdependence by giving children early experience caring for younger children and carrying out many meaningful family chores. Adults tend to structure their relationships so they can relate to the sibling group as a whole, not to individuals on a one-to-one basis. As a result, children do not have as much one-to-one verbal interaction with adults. In addition, because Hawaiian children learn from peers from an early age, they are comfortable in the role of teacher as well as in the role of learner.
As a result, conditions in typical classrooms may not be sufficient to elicit and sustain appropriate learning strategies. Sibling care and interdependence may diminish the degree of authority alloted to any one adult. Peer orientation and affiliation, while frowned upon in the typical classroom, has been found to contribute to school success of Hawaiian children. Learning stations which consider this orientation facilitate learning. Reading instruction modeled after the culturally familiar 'talk-story' activity improves reading skill and comprehension. Modification of instructional practice, classroom organization, and motivation management that takes into consideration the culture of the child has been found to make a significant difference in the achievement of Hawaiian children in school.
Figure 5.1 illustrates some aspects of mainstream culture which are congruent with the culture of the school but which may be in significant conflict with the cultural knowledge and attitudes of Hawaiian-American students.
Analysis of the classroom experiences of other minority children confirms the usefulness of the cultural incompatibility hypothesis. The KEEP model, for example, has been applied among the Navajo at Rough Rock. While not directly transferable, there is every indication that culture-specific modification of the program is possible. Efforts such as KEEP should be applauded. Even if not directly transferable to other contexts, the implications and motivations behind such work can be applied, especially where there is a large population of a single minority group in the schools.
SOCIAL CLASS
We have said that differences in school achievement may be attributed to the cultural influences of race, ethnicity, and gender on learning style. Other aspects of human diversity may also be critical in determining school success. Such factors include motivation, aptitude and achievement, self-concept, peer pressure, family, health, teacher expectations, and socioeconomic status. It is to issues of socioeconomic status and its influence on school achievement that we now turn.
Most Americans believe they live in a classless, rather egalitarian society. At the least, American ideology promotes the idea that, through proper attention and diligence (and some luck, which Americans also believe in), an individual may 'rise above' his or her social class. Part of what has been called an 'American religion,' this faith in the reality of upward mobility may account for the relative lack of attention to the concept of social class in the educational and psychological literature in the United States. Certainly it accounts for the difficulty sociology professors encounter in helping young people understand the bases of class differences in this society. Nevertheless, as we all know, there are significant variations in standard of living, status of occupation, and extent of expectations of upward mobility among American citizens.
Social class has been defined in a number of ways. all of which refer to a hierarchical stratification, or 'layering,' of people in social groups, communities, and societies. Assignment to social class categories is one of a number of stratification systems that can be used to distinguish one individual or group from another in such a way as to assign 'worth.' The urge to organize people in layers almost appears to be a culture-general characteristic; indeed, it has been said that whenever more than three people are in a group there will be stratification. While many Americans would identify class membership in terms of income, it is important to understand that money alone does not determine one's social class. Rather, one's social class standing depends on a combination of prestige, power, influence, and income.
<O_>figure&caption<O/>
Traditional class markers in the United States thus include family income, prestige of one's father's occupation, prestige of one's neighborhood, the power one has to achieve one's ends in times of conflict, and the level of schooling achieved by the family head. In other cultures, markers of social class may include bloodline and status of the family name, the caste into which one was born, the degree to which one engages in physical labor, and the amount of time which one might devote to scholarly or leisurely activities of one's choosing.
For purposes of analysis, it is often helpful to divide American society into five social classes. At the top there is a very small upper class, or social elite, consisting chiefly of those who have inherited social privilege from others. Second is a larger upper middle class, whose members often are professionals, corporate managers, leading scientists, and the like. This group usually has benefited from extensive higher education, and while family history is not so important, manners, tastes, and patterns of behavior are.
The third (or middle) social class has been called the lower middle class. Members of this group are largely people employed in white-collar occupations earning middle incomes - small business owners, teachers, social workers, nurses, sales and clerical workers, bank tellers, and so forth. This is the largest of the social classes in the United States and encompasses a wide range of occupations and income. Central to the values of the lower middle class are a "desire to belong and be respectable .... [f]riendliness and openness are values and attention is paid to keeping up appearances."
Fourth in the hierarchy of social class is the working class, whose members are largely blue-collar workers (industrial wage earners), or employees in low-paid service occupations. Working-class families often have to struggle with poor job security, limited fringe benefits, longer hours of work, and more dangerous or 'dirtier' work than those in the classes above them. It is not surprising, then, that members of the working class often feel more alienated from the social mainstream.
Finally, fifth in the hierarchy is the lower class - the so-called working poor and those who belong to what has been termed the underclass - a designation that refers to people who have been in poverty for so long that they seem to be unable to take any advantage at all of mobility options and thus lie 'below' the class system. Clearly, poverty is both the chief characteristic and the chief problem of this group. Webb and Sherman point out that this simple fact needs to be underscored:
Being poor means, above all else, lacking money. This statement would be too obvious to mention were it not for the fact that most Americans see poverty in other terms. Middle-class conversations about the poor often depict them as lazy, promiscuous, and criminal. Misconceptions about the poor are so widespread that it is difficult to appreciate fully what life is like at the lowest stratum of society.
Complicating the issues of social class is the fact that in the United States there is a large overlap between lower-middle-class, working-class, and lower-class membership and membership in minority groups. African-Americans, Hispanics, Native Americans (including the Inuit and native Hawaiians) are the most economically depressed of all groups in the United States. These groups also have the highest school dropout rates. To the extent that social class status depends on income and occupation (and therefore, usually, prestige and power), women and children of all racial, ethnic, and religious groups constitute a large proportion of the lower classes. This is, in part, a consequence of the descent into poverty that characterizes the lives of women who are divorced and their children. At the present time, nearly 25 percent of all American children under 6 years old are members of households trying to exist below the poverty line.
The working poor - those people who do work but in jobs that are minimum wage or slightly above, with no benefits, and hardly any job security - must also struggle to make it in today's society. To reach a middle-class lifestyle, a family of four in 1987 needed an annual income of about $31,000, and inflation will continue to raise this figure. In many cases, to reach this level both husband and wife must work. Only 25 percent of men and women reach this level if only one partner in the marriage earns an income. Thurow states that
although the dominant pattern today is a full-time male worker and a part-time female worker, the pattern is rapidly shifting toward a way of life in which both husband and wife work full time .... As an increasing number of families have two full-time workers, the households that do not will fall farther and farther behind economically.
Brislin comments on the effect this reality has on women and children:
The people left behind in the movement through social class levels include households headed by women, and these reached a staggering 31% of households in 1985. Dependence on one income, combined with the well-known fact of lower salaries earned by women, can result in poverty. Women and children constitute 77% of people living in poverty, and 50% of these poor people live in female-headed households with no husband present.
Those of similar socioeconomic status, at whatever level, also share similar cultural knowledge, attitudes, and values.
<#FROWN:J50\>
TEACHING SOMEONE TO THINK
The most important concept in educational theory is the concept of thinking. The concept has been clearly understood for many years, thanks to the insights of philosopher-educators like John Dewey and Alfred North Whitehead, but it has yet to be grasped in the teaching/learning processes of American schools.
Thinking, says Dewey, is what gives meaning to experience. Through thinking we apprehend a connection in experience that enables us to act intentionally. Thinking is the basis of responsible action, as contrasted with capricious behavior (which accepts things as they happen to fall out) or with routinized behavior (which accepts things as they have always been). Education aims at enhancing the powers of thinking in the broad sense of the acceptance of responsibility for action.
Dewey believes that all thinking is problem-solving. All thinking is instrumental. The starting point of any process of thinking is something going on, something that just as it stands is incomplete, unsatisfactory, or unfulfilled. Its point, its meaning, lies literally in what it is going to be, in how things are going to turn out. To fill one's head with facts about what is going on, says Dewey, is not to think but to function as a registering apparatus. Thinking means considering the bearing of what is going on upon what may be, but is not yet. It is applying what is known about the structure of experience to the task of projecting a plan of action for an unknown that lies beyond experience.
The function of a good teacher, as Stanford Erickson says, is to give voice to knowledge linking the present to the past and the future. Dynamic teaching is not a matter of developing innovative techniques for telling things to students. Dynamic teaching is teaching that engages the student's thinking in the imaginative consideration of learning. The moment of instructional truth, as Erickson says, occurs "when the student grasps the meaning of an important idea." Everything else in schooling is nothing more than preparation for learning.
The fundamental aim of all education is the enhancement of thinking. The cultivation of skill obtained apart from thinking has nothing to do with education; and information detached from thoughtful action is dead (even though it may resemble knowledge) and is a mind-crushing load. Education is the attempt to mobilize the imagination of individuals in the activity of thinking. "The sole direct path to enduring improvement in the methods of instruction and learning consists," as Dewey declares, "in centering upon the conditions which exact, promote, and test thinking."
For Dewey, the requirements of a dynamic educational method are (1) that the student has a genuine situation of experience that proposes a problem in which he is interested for its own sake; (2) that this problem should develop within the learning situation as a stimulus to his thought; (3) that he should have access to information bearing on the problem, its causes and its solution; (4) that he should be responsible for coming up with alternative solutions to the problem and for exploring them in an orderly way; and (5) that he should have opportunity and occasion to test his ideas by application, to make their meaning clear, and to discern for himself their validity.
Human learning, as Dewey always insists, begins with recognition of a situation that calls for some sort of doing. The starting point of learning is a problem. Beginning with a problem posed by experience, the student must then be helped to gain command of data that will help him to deal with the difficulty or the challenge that has been presented. To think effectively one must be able to command information that bears on the problem at hand.
The assimilation of data in relation to an experienced problem or concern generates suppositions, tentative explanations, and interpretations - what Dewey calls ideas. Data are facts; the ideas that spring from them forecast possible results. Thinking is an act of inference that always involves an invasion of the unknown, a leap out beyond what is known. Thinking is an incursion into the novel and demands some measure of inventiveness.
Ideas, as Dewey uses the term in this context, are always generated out of the originality of the thinker. All ordinary thinking is creative. Ideas in Dewey's sense are anticipations of possible solutions to problems, of possible connections between action and desired consequences. Ideas are not the final end of learning; they are intermediate goals, significant only insofar as they guide and organize future experience and action. The testing of an idea, of course, involves not only assessing the adequacy of the data that support it, but also acting upon it and seeing what results.
Dewey's understanding of the relationship between education and thinking is closely akin to Whitehead's celebrated definition of education as "the acquisition of the art of the utilization of knowledge." Whitehead advances a passionate protest against the notion of education as the accumulation of inert ideas, as the transmission of scraps of information accumulated by the learned professions. Education at every stage, he says, must permit each individual student to experience the joy of discovery - must permit each student, that is, to discover that ideas can provide an understanding of the stream of events that is Life and can establish a basis for practical decision. A merely well-informed person is the most useless bore on God's earth. But an education that is merely pedantry is not only useless; it is positively harmful, for a mind loaded with the dead weight of routinized learning is unable to think.
Imagination is the capacity to think about the relevance of learning for life. The "problem of problems" in education, as Whitehead sees it, is the provision of a corps of teachers whose learning is lighted up with imagination. A school must be a place in which the adventure of thought meets the adventure of action. Education is learning for life. And what is needed for life, as Whitehead says, are ideas that provide the basis for determining through foresight what actions are appropriate. Education is "a preparation by which to qualify each immediate moment with relevant ideas and appropriate ideas."
The basic virtue of the university, in Whitehead's view, is the power of imagination. The primary reason for the existence of a university is not to be found in the knowledge that it succeeds in imparting to its students or in the scientific advances achieved through its research. The justification for a university is that it preserves "the connection between knowledge and the zest for life," by uniting the young and the old in "the imaginative consideration of learning."
The imaginative consideration of learning, as Whitehead argues, requires a dialectical combination of freedom and discipline. The rhythm of education involves a series of stages in the growth of educated imagination: (1) the awakening of interest and the general apprehension of a subject in its vaguest details; (2) the acquisition of specific knowledge and mastery of the relevant details through the pursuit of an objective method; (3) the comprehensive ordering of the subject as a whole in the light of all relevant knowledge; (4) the understanding of the general principle for purposes of creative application in novel situations. The aimless acquisition through education of precise knowledge, inert and unutilized, only paralyzes thought. The wisdom that education seeks is the habit of the active utilization of well understood principles.
No area of education has more to gain from attention to the rhythmic law of human development, as Whitehead says, than moral and religious education. The principle of progress in moral and religious growth is from within. The teacher of moral and religious values has an important but limited function. He can elicit an awakening of concern in moral issues by resonance from his own personality and character, and he can create an environment that is conducive to the nurture of a higher wisdom and a firmer purpose. But the ultimate motive power for learning (whether in science, in morality, or in religion) comes from the learner. It is the sense of value, wonder, and reverence, the eagerness to merge one's personality in something beyond itself. Without this sense of value, as Whitehead says, education is a great emptiness and life sinks back into the passivity of its lower types.
EDUCATION AND VALUATION
In the course of all complex human learning, questions of value will inevitably arise. Yet the method by which intelligence addresses problems of values is not fundamentally different from the method by which it addresses questions of fact. Questions of values arise in all thinking, since thinking deals with problems arising out of human concerns and seeks a course of purposive action that will bring about a desired circumstance. Unless it addresses the values-related questions, education, like science itself, is simply irrelevant.
Valuation is a phase of the process of thinking that concerns the formation of strategies for solving problems. Informed thinking about alternative courses of action leads to appraisals of which action is better or worse, more or less serviceable/feasible. Such appraisals require an experimental justification of the same sort as is involved in all scientific generalization. Valuations, in other words, are concerned with what Dewey calls "rules for the use, in and by human activity, of scientific generalizations as means for accomplishing certain desired and intended ends." Moral education depends on a theory of valuation that advances a method of reaching well-founded judgments of value to guide the intelligent conduct of human activities, both personal and social, in the solution of human problems.
An instrumentalist understanding of thinking opens new avenues for the systematic application of disciplinary scholarship in the moral domain. Of course, as Dewey himself granted, it does not immediately resolve all the questions that philosophers and other scholars might wish to raise. Instrumentalism remains debatable as an epistemological or meta-ethical theory. Yet in the context of our struggle for better answers to the problems of life, Dewey's analysis of the way we think identifies an approach to values-related questions that is by no means foreign to the academy. It is, indeed, the method by which all human understanding progresses 'from sounds to things,' from appearances to reality, from illusion to truth.
Thinking is an ongoing process. It is never finished, and its conclusions are never final. Thinking is the search for greater coherence in our expanding experience, but it is always limited by the particular context in which it functions at any given moment. Thinking is always contextual; so, too, is all thinking about ethical issues. Every ethical problem arises in a particular context of experience; and all thinking about ethical issues proceeds by stages as the mind seeks new data and perspectives for forming more coherent and comprehensive understandings. Moral thinking does not deal with ultimate or intrinsic values any more than scientific thinking deals with absolute truth. The only truth to be found in either domain is the increasing coherence of our expanding experience.
A pragmatic program of moral education might well be guided by the assumption of ethical contextualism, which views moral reasoning as always taking place within a specific context of moral decision. All ethical argumentation depends on a set of premises in which at least one ethical premise is included. Contextualism is the claim that in every context of ethical discussion we have available some moral principles that set the problem. If these principles are themselves challenged, we simply move to another context of investigation, where we shall need to identify another principle or set of principles with which to carry on the new moral inquiry.
Ethical contextualism is based on the understanding that there is no working ethical premise that cannot be inferred by a process of logical inference from other ethical premises in combination with factual information. Such a view is analogous to the philosophy of causality, which assumes that there is no cause of a phenomenon that is not itself the effect of some other cause. For ethical contextualism the logic of evaluation is coordinate with the logic of science.
<#FROWN:J51\>
INTRODUCTION
The Grammar of Argumentation
It is generally regarded as mandatory for a textbook in logic to begin with a definition of logic. With some misgivings I shall attempt to comply with that convention. However I think that more than a simple definition is required. The concept of logic has undergone numerous changes since the Greeks first took up the study in the fourth or fifth century B.C. Since then the term has taken on new meanings which it was not originally intended to carry, without, however, entirely loosing its original sense. The result is that nowadays the word "logic" does not identify a single enterprise so much as a confusing conglomeration of enterprises related primarily by historical accident. By way of introduction, I shall attempt to clarify the concept of logic, and introduce at least some of the basic vocabulary of syllogistic logic:
History of the Concept of Logic
The people who conceived and developed syllogistic logic had quite different views on what they were doing than the people who call themselves logicians today. In taking up the study of syllogistic logic - a form of logic with deep historical roots - it may therefore be advisable to begin by acquiring some sense of how the word 'logic' has changed its meaning over the centuries.
The Greek word '<translitG_>logos<translitG/>' from which the word logic is derived, means (among other things) an 'account', in the sense in which a naughty child might be asked to give an account of his actions. Thus when Plato defines knowledge in the Meno as 'true belief with a <translitG_>logos<translitG/>', he means that a belief can properly be called knowledge only if (a) it is true, and (b) the person holding the belief can give an adequate reason, or an adequate account of why he holds that belief.
Of course any account which is adequate to justify one person's beliefs should also be adequate to convince someone else to adopt the same beliefs. To act rationally or logically means to act in a way that can be understood and justified. Hence, since logic was concerned with the nature of an adequate account, the Greeks understood logic to be concerned with the nature of persuasion, refutation, and inquiry.
But what does it mean to say that an account is 'adequate?' Plato thought that one should distinguish between legitimate persuasion and refutation, such as a genuine philosopher employs when conducting an inquiry into the truth, and illegitimate persuasion and refutation, such as a sophist employs in trying to sway public opinion. Granted that people justify their beliefs and actions by telling a story or giving an account, what sort of account qualifies as a legitimate and proper justification and what sort does not? It was this question which, I believe, logicians originally set out to answer. Hence the original conception of logic seems to have been that logic was supposed to give a legitimate account of the way in which we legitimately account for our beliefs and actions.
Plato, however, offers no criteria for drawing the distinction between legitimate and illegitimate accounts, other than to point out that a legitimate account is easier to remember, since the same line of thought, if it is genuinely rational, can always be reconstructed. The only method which Plato offers for telling the difference between legitimate and illegitimate accounts is self-honest evaluation based upon a commitment to the truth. A person who is being honest with himself, and is genuinely concerned to discover the truth, will simply be able to tell the difference.
The problem with this account (as Plato was well aware) is that it gives us no recourse against someone who stubbornly persists in being irrational. In the Gorgias Socrates falls into a discussion with such a person, and the dialogue degenerates into a shouting match. Socrates accuses Callicles of dishonesty - of being in disagreement even with himself - but without being able to give a better account of logic, he has no way to make the accusation stick. Logic may begin with self-honesty and a commitment to the truth, but it cannot end there.
The earliest surviving essays which attempt a systematic study of logic are a group of lectures by Aristotle. Aristotle did not necessarily intend these lectures to be treated as a single work, nor did he even necessarily regard them all as essays on logic. In fact, they cover a wide range of topics and were written at widely different periods of his life. The Alexandrian editor, Andronicus of Rhodes, nevertheless grouped these lectures together under the title Organon, which means 'organ' or 'tool', and placed them at the beginning of Aristotle's works. The implication is that these essays were not themselves lectures on philosophy, but were rather an explanation of the tools with which philosophers work. As such they should be regarded as preliminary to the study of philosophy. The subjects treated in Aristotle's lectures can be summarized as follows:
(1) In the Categories, a discussion of the various types of terms, the classifications into which terms can be divided, and the distinctive features which terms in each category display.
(2) In the Hermeneia, a discussion of propositions, how they are put together and how they are related to each other.
(3) In the Prior Analytics, a discussion of the formal patterns which arguments may take. It is this work in which syllogistic logic first appears.
(4) In the Posterior Analytics and the Topics, a discussion of the practical application of argumentation and proof to the acquisition of knowledge.
(5) In the Sophistical Refutations, a discussion of common errors (or fallacies) made in argumentation.
Medieval logicians learned their dicipline chiefly by studying Aristotle's Organon, so it is not surprising that they considered logic to include the same subjects covered by Aristotle's Organon. The medieval logicians made some significant changes in the syllogistic logic, gave names to each of the valid argument forms, and put the system into more or less its modern form. They also wrote commentaries on the other books in Aristotle's Organon, especially the Categories.
But it was not entirely clear why logicians, concerned with the question of legitimacy in methods of persuasion and refutation, should take time to examine types of terms. Some conception of logic was needed to give a sense of unity to these diverse matters. The medieval response to this problem was to point out that all the topics in Aristotle's Organon are concerned with 'second intentions'. The word 'intention' may be taken as roughly synonymous with the word 'meaning'. Let us call anything which has an intention or meaning a 'sign'. Words, for example, are signs, since there is some object which any word 'intends', though this object need not actually exist. (The word 'unicorn', for example, has a meaning or intention, since it refers to a type of being, even though the being to which it refers does not actually exist.) However, words are not the only signs. Thoughts are also signs, since thoughts, like words, 'intend' some object, i.e. they have a meaning. (To prove this to yourself, try to imagine what it would be like to have a meaningless thought. Wouldn't that be the same as not having a thought at all?)
According to the medieval logicians, signs could be divided into two groups:
(1) those whose object is a thing or relationship that (if it exists at all) exists in nature. Examples would include 'green', 'tree', 'unicorn', etc. These were called signs of first intention.
(2) those whose object is another sign. Examples would include 'noun', 'verb', 'subject', 'predicate', etc. These were called signs of second intention.
Logic, they thought, was the science which studied signs of the second kind rather than signs of the first kind. Hence logic could be defined as the 'science of second intentions'. By this definition, of course, logic incorporates all of the linguistic sciences, including what we would now call grammar and semantics.
This definition gave a sort of unity to the list of subjects which were thought to fall within the scope of logic, but it also caused some of the subjects to be pushed into the forefront, while others took on a merely peripheral importance. For example, the discussion of types of terms became particularly important, since the purpose of the discussion was not to classify types of things, but rather to classify what we could sensibly say about things. It was, in other words, an attempt to give some order to signs themselves; it was not an attempt to give order to the objects which the signs intended. Hence the categorization of terms fit neatly within the medieval definition of logic. The discussion of sophistical reasoning, which would have been regarded as central to logic according to the Greek conception, still had its place in medieval logic but it no longer occupied a position of overweaning importance.
Thus while the Greeks thought that logic should be primarily concerned with persuasion, refutation, inquiry, and the justification of beliefs, the medieval logicians extended the notion, so that logic was thought to be concerned with the nature of signs in general.
It is important to understand that the notions of 'sign' and 'intention' are very broad. They include, not only words and language, but also thought itself. Because of this, it was easy to make the transition from thinking of logic as primarily concerned with language (especially language used to persuade), to thinking of logic as primarily concerned with thought. The philosophers of the early modern period came to regard logic as concerned with 'the laws of thought', and this frequently meant that they regarded logic as a sort of psychology. Logic was thought to study the operations of the mind, specifically, of course, the operations associated with rationality. Certain operations of the mind were thought to be more fundamental than others; so fundamental, in fact, that it would be contrary to human nature, or outside the scope of human capacity, to reason in any other way. This view of logic was held by philosophers as diverse as Leibniz, Locke, and Kant. Even Hegel, when writing on logic, seems to have in mind a thorough-going analysis of the structures of the conscious mind, though he is careful to make clear that he means any conceivable conscious mind, not simply the human mind. Hence, Hegel does not regard logic as a type of psychology but rather as related to (though not identical with) phenomenology.
During the 19th century a revolution occurred in the methods employed by logicians. George Boole was able to demonstrate that simple mathematical operations, such as multiplication and subtraction, could be made to parallel the familiar laws of logic. This meant that logicians were able to use quasi-mathematical formulas to represent or model patterns of argumentation. This new technique was incredibly powerful, and gave logicians the tools with which they could analyze more and more complex arguments. However, most logicians still considered logic to be, at root, the study of the laws of thought. They simply assumed that these laws could be expressed mathematically. Hence mathematics changed the methodology of logic, but not (at first) its essential subject matter.
However, the notion that there are such things as laws of thought, and that it is the business of logic to study them, began to break down with the advent of non-Euclidean geometry, and with the many other sudden advances in mathematical theory also made during the 19th century. It would be false to claim that the non-Euclidean geometries destroyed the concept of 'laws of thought' by proving that humans could conceptualize on a much grander scale than had previously been supposed. In fact they proved no such thing. Rather, the non-Euclidean geometries proved that even apparently absurd presuppositions could be developed into consistent (and perhaps even useful) systems. Hence discussion among mathematicians concerning what was absurd or inconceivable was supplanted by discussion concerning what could and could not be consistently systematized. Mathematicians could no longer regard mathematics as founded upon a priori truths, or as an attempt to elaborate some particular system of well understood formal relations.
<#FROWN:J52\>Speaking of the poetic writers of this period, Harold Bloom has argued that they characteristically exhibit a compulsion toward priority in intellectual and artistic creation and a fear of being regarded as derivative. Bloom terms this state of mind "the anxiety of influence." In his study with this same title, Bloom observes that "all quest-romances of the post-Enlightenment, meaning all Romanticisms whatsoever, are quests to re-beget one's own self, to become one's own Great Original." Surprisingly, in this study Bloom repeatedly draws on Kierkegaard's statements about artistic and intellectual creativity to illustrate his own points about the poet's efforts at self-creation, the poet's demand for priority and for freedom from the influence of his precursors.
Although Bloom is entirely unaware of Kierkegaard's relationship to Kant, it is not accidental that he draws on Kierkegaard's preoccupation with originality of authorship in describing this "anxiety of influence." As Kierkegaard reminds us from virtually the first to the last words of his published work, he was always a poet, and a poet writing and thinking in an era when originality - the secure possession of one's own daimon, one's genius - was the mark of greatness. Even a cursory reading of Kierkegaard's Papers tell us how much he was concerned with the singularity of his religious-literary effort. From Either/Or and Fear and Trembling onward the poetic Kierkegaard always remained concerned with his place in literary and intellectual history. Furthermore, as Christoph Schrempf reminds us in his exhaustive biography, Kierkegaard's strong sense of sacrificial destiny - his wish to offer up his life as the vehicle for a unique and redemptive idea - exhibits a turn of mind characteristic of many other 'poetic' writers of this period.
In view of this, we might suppose that Kierkegaard also suffered from "the anxiety of influence," that he, too, sought to cut the ties to his predecessors and, above all, to that predecessor on whose originality and genius his own novel creation so depended. In the course of his discussion, Bloom repeatedly quotes a remark by Kierkegaard that Bloom presents as a virtual synopsis of the poet's effort to creatively appropriate the work of his predecessor: "He who is willing to work gives birth to his own father." If the pattern of borrowing we've seen is indicative, Kant is, in a sense, Kierkegaard's intellectual 'father.' Understood in terms of Bloom's analysis, therefore, Kierkegaard's authorship becomes an impassioned effort to 're-create' Kant's philosophy in a way that makes it fully the product of Kierkegaard's own creative genius. The absence of any tradition of scholarship relating Kierkegaard to Kant and the difficulty many Kierkegaard biographers have had in tracing the lines of his descent show how well Kierkegaard was able to obscure his own intellectual paternity.
This first explanation of Kierkegaard's deliberate effort to erase the lines leading back to Kant assumes a measure of concealment and intellectual ambition. A second explanation moves in a different direction and relates Kierkegaard's handling of Kant to the central intellectual concerns of his authorship and to his ongoing employment of irony as a philosophical tool. Simply stated, it sees Kierkegaard's covert use of Kant as a subtle, necessary, and deserved trick played by Kierkegaard on his arrogant Hegelian foes.
To understand this second effort at explanation, it helps to keep in mind Louis Mackey's point that Kierkegaard faced a daunting task in taking on Hegelianism from an orthodox Christian-religious point of view. The challenge before Kierkegaard, Mackey observes, was to call into question a philosophy that regarded itself as the dialectical fulfillment of human thought and as thus able to comprehend all possible philosophical and religious positions. To accomplish this seemingly impossible task, Kierkegaard chose to avoid philosophical argumentation and to employ the method of Socratic irony, arraying its "infinite negativity" against Hegelian pretensions.
I would deepen Mackey's observations by adding that part of this ironic strategy for Kierkegaard may have involved a decision to employ Kantian philosophy in his struggle against the Hegelians: to use a thinker who had been 'transcended' (aufgehoben) and fully assimilated into the dialectic against the discipline's now reigning giant. Philosophy could thus be turned against its own methodological and substantive pretensions. To prevent this from becoming just another page in this history of philosophical debates, however, and to avoid the Hegelians' premature dismissal of his position, Kierkegaard would have had to conceal Kant's presence in his own reformulated statements of Christianity. In this way Kierkegaard could appropriate and use Kant's brilliant destruction of the tradition of rationalism while ironically exposing the hollowness of the Hegelians' claims to have mastered all preceding thought.
There are hints in Kierkegaard's writings that he was aware of the joke he was playing on the smug but philosophically less-than-well-trained Danish Hegelians. Kierkegaard knew that Martensen and his acolytes, for the very reason that they did not take the past seriously, were often ill-versed in the writings of philosophers whose work they purported to have transcended. In a remark to his brother in 1841 he caustically dismisses possible criticisms of his doctoral dissertation by what he calls "one or another half-educated Hegelian robber." Having endured Martensen's and others' lectures on Kant, and knowing how little the Hegelians really understood Kant's profound ethics and philosophical theology, Kierkegaard may well have enjoyed the one-upmanship involved in surreptitiously turning Kant against his teachers.
I have already mentioned one possible instance of this kind of playfulness on Kierkegaard's part: his handling in the Postscript of the matter of the philosophical pedigree of the idea of the 'leap.' Although Johannes Climacus, the Postscript's pseudonymous author, repeatedly confesses his debt to Lessing for this idea, even giving the title 'Attributable to Lessing' to the section where he discusses the leap, the section itself ends with mention to the fact that Johannes de silentio, author of Fear and Trembling, had previously discussed a similar idea. Climacus adds that he had read Johannes de silentio's book before encountering Lessing's essay, and he closes his discussion with the remark "Whether Johannes de silentio has had his attention called to the leap by reading Lessing, I shall not attempt to say." This comment invites the discerning reader to ask, who, if not Lessing, might be Climacus's/Kierkegaard's primary philosophical source for this idea? Since Fear and Trembling centrally addresses Kant's repudiation of the leap of faith as this was symbolized for both Kant and Kierkegaard by the episode of Genesis 22, there can be little doubt that this remark about influence points toward Kant. Almost as though he were unwilling to totally obscure his debt to Kant, in other words, Kierkegaard gives the informed reader a glimpse into the game he is playing with the Hegelians.
The Fragments offers another sign of this playfulness. On at least six occasions, the author, Johannes Climacus, raises questions of scholarly attribution and openly acknowledges the unoriginality of his ideas and his susceptibility to the charge of plagiarism. In almost all these instances, Climacus is able ironically to defend himself against such accusations because the intellectual property he is appropriating derives from the very public domain of biblical teaching. At the same time, many other elements in this volume are borrowed from Kant, possibly including at least one idea attributed by Climacus to the Bible. The project itself is stimulated, in part, by Kant's position on historical revelation in The Conflict of the Faculties, a work we might suppose relatively few Hegelians had read. Hence, concealed within Climacus's openly confessed plagiarism lies a deeper level of unacknowledged borrowing. Kierkegaard's irony and sense of humor seem to me to be at work here. Openly signalling the derivative quality of this work, he leaves it to the Hegelians to detect his employment of Kant against them, while remaining confident that, despite their vaunted philosophical erudition, the Danish Hegelians will surely fail to see or understand the presence of the philosopher they had 'gone beyond.'
Was it, then, authorial ambition or a only half-veiled playfulness and irony that led Kierkegaard to obscure his debt to Kant? In the complex world of human motives, which Kierkegaard himself so masterfully explored, these two explanations may be less opposed than they appear. Perhaps both motives were at work, a youthful author's passionate quest for originality tempered by conscientious self-disclosure and a willingness to reveal his philosophical legerdemain to those able to appreciate it.
Whatever Kierkegaard's motives, there can be no doubt that Kant played a major role in Kierkegaard's intellectual formation. From the pattern of borrowing we've seen, I am led to conclude that as early as his student years Kierkegaard was deeply intrigued by what he had read in Kant's mature works on religion. This interest shaped the course of his writing and thinking in the period immediately following. We saw that as early as 1835 Kierkegaard expressed the wish that he might find "the idea" to which he could give his life. Over the course of the next five years, reading Kant in preparation for his degree examination, Kierkegaard found this idea. Although his best teachers treated Kantianism as merely a way station en route to Hegel, Kierkegaard perceived that Kant's contributions to religious thought had not even begun to be assimilated. If Kant's ethics were taken seriously - and despite many criticisms, including Hegel's, Kant's ethical writings everywhere commanded respect - then his contentions in the Religion also had to be taken seriously. They had to be viewed neither as the work of a great philosopher in his dotage (the opinion of some philosophers) nor as a wooden and perhaps dangerous effort to translate biblical truth into philosophical terms (the orthodox view). Instead, they had to be seen as a rigorous exposition of ideas that pointed inevitably to their limits and to the requirement that they be transcended by faith. Over the next five years, during the intensely creative period of the major pseudonymous writings, Kierkegaard sought to develop a new form for the communication of these ideas, and he tried to carry Kant's beginning to the thoroughly religious or "faithful" conclusions Kant resisted.
A journal entry for 1836-1837 written at the time Kierkegaard was beginning his formal study of Kant provides an early hint of the importance Kierkegaard attached to Kant's philosophy. Kierkegaard here presents Kant, along with Goethe and Holberg, as belonging to the "royal procession" of thinkers whose work represents a fundamental breakthrough in thought and who, because of this, are resisted by their contemporaries. These thinkers, Kierkegaard suggests, play a decisive a role in conveying the "password which God whispered in Adam's ear, which one generation is supposed to deliver to the next and which shall be demanded of them on judgment day." Since Kierkegaard's own sense of personal mission was very much in formation at this time, it is reasonable to conclude that he regarded himself as a recipient, in his generation, of the "password" uttered by Kant. Remarkably, in an omission characteristic of the whole tradition of Kierkegaard scholarship on this matter, the Hongs fail to include Kierkegaard's mention of Kant in their otherwise meticulous translation of this entry.
Either/Or offers a further hint that carrying on and fulfilling Kant's initiative is an accurate depiction of Kierkegaard's project in the pseudonymous works. In Chapter 3 we saw that the ethical stage of life sketched there by Judge William is broadly Kantian in nature. Kierkegaard himself shows his sympathy for this position when he states that in every way the ethics which Judge William champions "is quite the opposite of the Hegelian." Yet Either/Or also seeks to point to the possibility of a religious stage of existence where the ethical is dialectally transcended and fulfilled. This stage is suggested in the position presented by the unnamed country pastor whose view is developed in the "Ultimatum" of the book, a view that moves toward faith via a recognition of the unavoidability of sin. The ethical must be transcended (but not eliminated), the pastor tells us, because if ethics alone were to shape our destiny, we would all succumb to moral despair.
<#FROWN:J53\>
Wittgenstein on Understanding
WARREN GOLDFARB
Wittgenstein's treatments in the Philosophical Investigations of the cognitive or intentional mental notions are evidently meant to persuade us that, in some sense, understanding, believing, remembering, thinking, and the like are not particular or definite states or processes; or (if this is to say anything different) that there are no particular states or processes that constitute the understanding, remembering, etc. Such a dark point desperately needs clarification, if it is not to deny the undeniable. For surely we may (and Wittgenstein does) speak of a state of understanding, or of thought-processes; surely when one understands - understands a word, a sentence, or the principle of a series - one is in a particular state, namely, the state of understanding the word, sentence, or principle. Now Wittgenstein sometimes puts the point more specifically, speaking of "a state of a mental apparatus" (<*_>section<*/> 149), or, with respect to understanding, of "a state which is the source of the correct use" (<*_>section<*/> 146), as that which he wishes to deny. But clearly these denials, though they sound somewhat less paradoxical, are equally in need of elaboration.
However Wittgenstein's point is clarified, though, there is an objection that will inevitably be made, an objection claimed to arise from a scientific viewpoint. The objection asks: isn't Wittgenstein here usurping the place of empirical inquiry? Is it not possible that empirical science - neurophysiology, in particular - will find specific states and processes that will fill the bill, as far as understanding, believing, remembering, etc. are concerned? And so, whether or not there are definite states or processes of understanding is something that we will discover. That there are such can certainly be entertained as a scientific, empirical hypothesis, and its consequences discussed. To preclude such an outcome now is just to claim that we shall never obtain certain sorts of results in future neurophysiology. But then Wittgenstein is simply making a bet on the future course of science or else he is engaged in a priori anti-science, denying a priori that certain projects could bring results, and hence they ought not even be investigated empirically.
For brevity I shall call this objection, that Wittgenstein is simply ruling out something that science could discover, the 'scientific objection.' I mean it to be narrowly based on envisaged possibilities of results in neuroscience and thus not to be the charge that Wittgenstein fails to leave room for the discoveries of psychology, or cognitive science, or psycholinguistics, or some as yet unknown science of the mind. The latter charge has been made frequently since the publication of the Investigations, but it is not so much an objection to a specific point of Wittgenstein's as a denial of all of Wittgenstein's philosophy of mind. Much in the Investigations is devoted to exposing the conceptual confusions that would be involved in a "science of the mind"; in any case, any such science would presuppose just the picture of mental states or processes and the whole notion of a mental apparatus that Wittgenstein is concerned to undercut. In contrast, the scientific objection as I am imagining it exploits the unarguable status as empirical science of neurophysiology and the independence from psychology of the conceptions of state and process that it employs. Consequently, the objection requires detailed attention to what Wittgenstein is saying or denying about intentional notions, and so can function to elicit clarification of Wittgenstein's dark point about states or processes.
The scientific objection is, it seems to me, a most natural one, and for that reason powerful. And it puts Wittgenstein in a poor (and stereotypical) light: as a philosopher, making ungrounded empirical claims or, worse yet, a philosopher trying a priori to deny the progress of science, like a late Scholastic refusing to grant the coherence of the Copernican conception or an Idealist 'proving' the impossibility of Einsteinian physics. Indeed, the charge that Wittgenstein is ruling out certain empirical possibilities a priori is doubly damaging against a philosopher whose major concern is to fight against a priorism, to demolish pictures of how things must be, to expose "preconceived ideas to which reality must correspond" (<*_>section<*/> 131).
A common view of Wittgenstein takes his response to such an objection to proceed by noting that a physiological state would be 'hidden'; it would not reflect central conceptual features of our notion of understanding (or believing, thinking, etc.). No such state would have conceptual links to correct responses and the other criteria of understanding; but the transparency of those links is essential to our concept of understanding. In short, the point is that no physiological state has the 'grammar' of understanding. Hence, Wittgenstein is taken to urge, understanding cannot be conceptually constituted by any such state.
Although there are passages in the Investigations that suggest this response, to my mind it fails to capture Wittgenstein's thought. The response requires a general and sharp distinction between conceptual and empirical, between criterion and result of investigation. Wittgenstein does not have the resources to make this distinction; in fact, the distinction is of a piece with the a priorism that he wishes to attack (see <*_>section<*/> 79, last paragraph). That is, the response requires a form of essentialism which I do not believe Wittgenstein accepts.
Moreover, the response leaves it open that understanding, although not 'conceptually' constituted by a neural state, might be 'empirically' constituted by one. This underreads Wittgenstein's conclusion. Thus the suggested response does too little. Yet it is not totally off the mark, for Wittgenstein does think that the identification of any state of understanding would have to bear some conceptual burden. His conclusion will be that, in some sense, "nothing is hidden." But this is meant to result from detailed considerations of the notion of understanding, not to be assumed more or less as a premise, as an unargued feature of his methodology.
The notion that Wittgenstein dismisses or ignores questions of 'empirical' identity also reinforces the view of him as simply anti-scientific. Now we do know that Wittgenstein was distrustful of the claims of science and its role in contemporary culture. He makes this quite explicit when he expresses his opposition to "the spirit of the age" in the Foreword to the Philosophical Remarks, and the theme is repeated elsewhere in his work, both early and late. Some have gone on to ascribe to him a hostility to or belittling of projects aimed at scientific explanation, of such proportions as to depict him as simply dogmatic and blinkered. Bernard Williams, for example, complains that Wittgenstein's general practice and teaching serve to "stun, rather than assist, further and more systematic explanation," and offers as illustration a story about Wittgenstein used by Georg Kreisel in 1960 as evidence for Wittgenstein's hostility to legitimate scientific explanation. The story bears repeating. Imagine a child, learning that the earth is round, asking why then people in Australia don't fall off. I suppose one natural response would be to start to explain about gravity. Wittgenstein, instead, would draw a circle with a stick figure atop it, turn it upside down, and say "Now we fall into space."
Now I myself do not find the story emblematic of an attitude against science, or of a desire to 'stun' further explanation. In fact, I think it shows Wittgenstein being highly insightful. For he is examining the source of the child's question, in the concepts with which the child is operating. Given those concepts, an appeal to gravity can do nothing but mislead: the child will take it that the antipodal people are upside down, but they have gravity shoes, or glue, or something similar, that keeps them attached to the surface of the earth; as for us, we are right side up, so the problem does not arise. What Wittgenstein's trick does is precisely to expose the conceptual confusion in the way the child is thinking of up and down (cf. <*_>section<*/> 351). Once the child sees the relativity in the notions of up and down, she may then go on to ask, "Well, why don't all these objects - us, the Australians, and the earth - go careening around independently?" or perhaps, "Then why do objects fall down rather than up?" At that point, explaining gravity might well be in order; the child is now prepared to appreciate it correctly.
The lesson, of course, is general. It is central to Wittgenstein's teaching that the conceptual underpinnings to a felt need for explanation must be scrutinized, for what it is, exactly, that wants explanation may only become clear through such an investigation. This does not make him anti-scientific. It does make him anti-scientistic, against the smug and unexamined assurance that what wants explanation is obvious, and that scientific tools are immediately applicable.
For Wittgenstein, it is characteristic of the notions that figure in philosophical problems - prominently, mental concepts and linguistic concepts like meaning - that a structure is imposed on them, without grounding in the ordinary use of these notions and without being noticed, when they are taken to be amenable to certain explanatory projects. (Much work in the Investigations, on my view, is precisely devoted to getting us to notice, to see that there is a place at the start of philosophizing, where this imposition happens.) Hence, only through clarification of what the legitimate questions are can proper sense be made of the applicability of science. A scientistic viewpoint ignores this need for clarification. As a result, for Wittgenstein scientism is just as misguidedly metaphysical as traditional, more transparently a prioristic, approaches.
I am depicting Wittgenstein as thinking that conceptual work must be done before the question of the applicability of science should be raised. Now it is also true that whatever empirical, scientific results may legitimately be foreseen or hypothesized after that work is done are of little interest to Wittgenstein. He seems to believe that what is really at issue in a philosophical question will be answered, or dissolved, by the conceptual work, and not touched by science. That is, science is simply not of use in dealing with the sorts of problems with which he is concerned. But this broad characterization can be misleading. Wittgenstein is not being dismissive; he is not urging a distinction between questions of the mind vs. questions of the heart. Nor is he saying that in doing science we are talking about different things; he would have had little patience with Eddington's 'two tables' and with Goodman's different worlds (although the latter claim may be thought controversial). Moreover, as I mentioned above, I do not believe that he wishes to rely on a sharp distinction between conceptual and empirical, that is, a Fregean divide between logic and analysis on the one hand and 'mere' psychology and physics on the other. Rather, Wittgenstein operates case-by-case. For each philosophical question we treat, we are to tease out what we are aiming for, or what we think we are aiming for, and then to come to see how our objectives will not be served by a scientific investigation; and we are to recognize how the inclination to look for science for answers elides or ignores so much as to suggest that a philosophical picture is at work.
This point is made in <*_>section<*/> 158, the one explicit appearance in the Investigations of something like the scientific objection. Here he suggests that, although the objection says that the scientific investigation may come out either way, that it is only a scientific hypothesis or conjecture that such-and-such a process or state will be found, at bottom the objector is being moved by an a priori demand that things must turn out a certain way. The claim of a modest empiricism is mere lip service.
With respect to understanding, the point might be phrased thus. Wittgenstein asks us to look in detail at the range of our practices relevant to an ascription of understanding. We find an enormous variety of considerations that can enter, a dependence on context that is impossible to describe accurately by any general rules, a lack of uniformity in mental accompaniments.
<#FROWN:J54\>
While it is not likely that Greek agriculture often produced more than limited dividends, those who were masters of the rural landscape gained an even greater advantage - they also controlled almost all the population of a polis either directly or indirectly. Accordingly, aristocrats tended to hold on to their land. It has often been argued that an ancestral plot (kleros) could not be alienated outside the family before the fifth century, but a variety of evidence proves otherwise. Hesiod's father migrated from Asia Minor to Boeotia, where he acquired a farm; the poet himself advised his auditors to honor the gods so you may buy another's holding (kleros) and not another yours. Aristotle's assertion that "to part with family estate was one of the things that were 'not done,'", does not prove the inference that it could not be done. Normally, nonetheless, formal sale or transfer of land rights was unlikely in so rurally based a world; political citizenship and economic security rested on an independent connection to land.
Various areas saw their smaller farmers of earlier times reduced to the level of bondsmen or serfs, though true agricultural slavery was uncommon in the Greek world. Usually rural domains were not large in view of the fragmented geographical nature of Greek landscapes; the largest known at Athens, as noted earlier, ran only about 50 hectares, though "in fifth- and fourth-century Athens there were landowners possessing from three to six estates in different partparts of Attica." Those who could be expected to serve as hoplites probably were masters of at least 12 hectares; free farmers after the Solonian reforms would scarcely have been able to cope with more than 4 hectares unless they drew in outside labor at critical points in the agricultural cycle.
How were the larger holdings managed? In modern times nobles relied on stewards, factors, and the like to bear day-to-day responsibilities; this was probably the case also in ancient Greece, though our evidence is very limited. Cimon's liberality to his fellow demesmen was famous, but he was absent from Athens so often on military operations that he must have had a resident aide. For Pericles we do have the comment in Plutarch's life that he arranged his paternal estate so "that it might neither through negligence be wasted or lessened, nor yet, being so full of [public] business as he was, cost him any great trouble or time with taking care of it." Thus he sold "all his yearly products and profits" in a lump and bought in the market for his household needs - obol-pinching and keeping precise records to the discontent of his family. "His manager in all this was a single servant, Evangelos by name." To secure fuller information on the careful management of agricultural resources for economic gain we must come down to Hellenistic Egypt where the financial director for Ptolemy II had an extremely astute manager, Zenon, many of whose detailed records have survived on papyri.
The picture usually drawn of upper-class concentration on landed possessions may as a whole stand, but before we accept it as representing the exclusive interest of the well-to-do further reflection is necessary on the nonagricultural aspects of the economy. For example, Demosthenes, on reaching his majority, brought suit against his guardians for the recovery of his inheritance. The estate he itemized to the jury under two headings is rather surprising: "(1) the active (energa), which included 32 or 33 slave swordmakers, bringing in 3000 drachmas a year; another 20 slaves engaged in the manufacture of furniture, 1200 drachmas annually; and 8000 drachmas on loan at 12%; (2) the inactive: raw materials on hand at his father's death nine years before, worth 15,000 drachmas, the house worth 3000, the furniture and his mother's jewelry, 8000 in cash in a strong-box at home, a maritime loan of 7000 drachmas, and 4700 on deposit in two banks and with a relation." Land proper does not appear at all; Demosthenes was rather, in Bolkestein's summary, "accustomed to make [his fortune] bear interest in many ways" as a capital-investor, living on the interest of his money.
Demosthenes' family was not, indeed, of aristocratic stock, so it may have been willing to extend the employment of its capital more widely in the thriving Athenian industry and trade of the fifth and fourth centuries. How far did the aristocrats do the same, or alternatively shun this area?
Industry may be dismissed briefly. Men of standing were not likely to sully their fingers or break their backs in the physical toil of stonemasons, smiths, potters, and other trades. Even so they did have an important, twofold role: they were the consumers who bought the wares of craftsmen, and they provided to a large degree the capital necessary for the purchase of the slaves who furnished a valuable share of the labor. The swordmakers and furniture fabricators of Demosthenes have already been noted; even more remarkable was the fact that Nicias, of aristocratic stock, owned a thousand slaves, whom he leased out to the entrepreneurs running the state silver mines of Laurium.
Commerce was another matter. Retail trade, such as that in ribbons, could be left to vendors, many of them women, but large-scale activity especially by sea required wider attention. One principal mark of the Aegean world in and after the eighth century was overseas voyaging, an this was without doubt initially in the hands of aristocrats.
Sappho's brother Charaxus, for example, carried wine to Egypt and there fell in love with a courtesan, to Sappho's disgust; Solo also engaged in foreign commerce in order to recoup his father's prodigality. Coleus of Samos, blown off course to Egypt as far as Tartessus whence he gained so much that he had to replace his stone anchors by silver ingots, and the later Sostratus of Aegina, who dedicated a statue in the Greek shrine of Etruscan Pyrgi, very probably were both men of the leading classes inasmuch as they entered Herodotus' pages.
It was, after all, aristocrats who had surplus resources that could be ventured abroad and also were leaders, able to face possible hostile resistances on foreign shores. Contrary to the views of many modern scholars, moreover, both they and the potters described in Hesiod's Works and Days sought earnestly after wealth. Already in the epics Odysseus was taunted as not looking like an athlete, that is, a man of leisure, but "one, who faring to and fro with his benched ship, is a captain of sailors who are merchantmen, one who is mindful of his freight, and has charge of a home-borne cargo, and the gains of his greed." Solon categorized the diverse ways of gaining wealth and concluded that those who are richest "have twice the eagerness that others have"; his contemporary Alcaeus quoted Aristodemus - a Spartan no less - as saying that "wealth makes the man." At first aristocratic seafaring might not have been much distinguished from piracy and coastal raids, but eventually it settled into more ordered communications.
Aristocrats were also the men most interested in the wares that could be acquired in the advance workshops of the Near East - ivory, glass, faience, perfumes, ointments, and spices (many of which had names of Semitic root) - for such luxuries were the backbone of the earliest overseas trade. Greek lands, even including Athens down to the time of Solon, fed themselves, though they did have need for foreign slaves, metals, wool, stone, and other bulk items. Hesiod drank wine of Biblis while relaxing in the heat of summer, and to a remarkable degree men and also women of the upper classes desired wools dyed in Tyrian purple and fine linen. Since textiles do not survive well in archeological contexts this item is often overlooked, but even in modern times the textile trade has been very significant; in the English colonies of North America in the eighteenth century the main import consisted of English, Irish, and German cloth and textiles.
By the later sixth century aristocrats had become more conscious of the duties and limitations of their position and largely yielded long-distance trade to professional shippers, but as they withdrew into the background their interest in this realm did not disappear. The men who scurried about the Aegean and farther afield had to have capital to outfit their ships and finance cargoes. To an extent what we cannot measure they may have done so out of their own resources, but at least occasionally they had to secure a bottomry loan at rates up to 33.33% - Demosthenes' estate, it may be remembered, included such a loan, and in one of his orations a money-lender/banker asserted that without the support of men of his type "no ship, shipper, or sailor can put to sea." And who provided the money of the banker? Undoubtedly the well-to-do of Athens; in imperial Rome as in early modern times the rich supplied funds by the back door to large-scale traders. Nor did aristocrats totally surrender the field; Andocides, who traced his ancestry back to Hermes via Odysseus and Telemachus, actually engaged in maritime commerce throughout the winter in the late fifth century and after his return from exile "continued to think and act like the businessman he had turned himself into". It is unsafe to assume that the word kerdos (profit) totally disappeared from aristocratic lips even after the developed ethos of the class frowned on undue interest in economic activities. Aristotle in his Nicomachean Ethics judiciously stresses the need for a competence but not a search for gain per se.
Through control of the land and the revenues from investments the well-to-do economically commanded the Greek communities, sometimes almost completely, though at Athens only in major degree, and used every available opportunity to enjoy an elegant, luxurious life. Modern hostility toward elites swiftly rises into view at this point in the common assertion that aristocrats in all ages spend money rather than improving the economic machinery of their world in a bourgeois fashion. Thus early modern aristocracies were reproved as being engaged in "the unjust, unhealthy, brilliant and anti-economic utilisation of any surplus produced in a given society:" An interesting study of men who participated especially in the Thirty Years' War of the seventeenth century after Christ suggests that their wealth was committed to building mansions and to acquiring adornments such as gold necklaces. Only one noted scholar, to my knowledge, finds merit in this type of expenditure. Writing about early modern English aristocrats, G.M. Trevelyan raises the question as to how else the English nobles could have expended their money save by building magnificent houses in a period when stocks, bonds and general loans were unknown and land was not easily bought - but then Trevelyan is nowadays generally dismissed as an elitist.
If a phenomenon recurs frequently in different historical societies, then there must be significant reasons for its presence; and an understanding of those reasons will be more useful than the common expression of indignation or reproof. Braudel, just quoted, also observes that luxury "scarcely changes at all" as a concept accepted both by privileged and unprivileged classes and, as he notes, both Mauss and Sombart emphasized the role of luxury in promoting demands on artists and others in early modern Europe. So too the aristocrats of ancient Greece stimulated the amazing outburst of Hellenic civilization by their patronage of the arts and crafts, as we shall see in the next chapter.
Some modern aristocracies unfortunately have been largely parasitical and have been overthrown in violent revolutions. In Greece the political, social, an economic position of the upper classes was too deeply anchored ever to be seriously threatened; even the Athenian democracy commonly entrusted its leadership to men sprung from aristocratic families of high standing. The upper classes of Athens, however, paid a heavy price both personally and economically. If they were wealthy enough to afford hoplite armor or horses, they had to be prepared to face the dangers inherent in the almost unceasing wars of the fifth and fourth centuries; an inscription of 460 or 459 lists no less than 177 men of one Athenian tribe who died in one year in Cyprus, Egypt, and elsewhere, including two generals (an Athenian tribe might have had in the order of 4000 adult male citizens, but the number who in practice might be drafted is much reduced if one takes into account the men of age and those mentally and physically incompetent).
<#FROWN:J55\>
More recent events reinforced French insistence that 'the Boche will pay.' Rehabilitating France's ten northeastern dpartments, which had served as one of the main theaters of the war, required an expensive infusion of capital. France had incurred substantial war debts to her Allies, whereas Britain's debt to the United States was offset in part by France's debt to Britain. American refusal to provide concessional reconstruction loans or to forgive these debts did much to harden the French position, rendering it inevitable that German reparations and Allied war debts would be bound up together.
A reparations bill as large as $200 billion was contemplated at Versailles. Ultimately, the assembled delegates were only able to establish a deadline for the conclusion of discussions: May 1921. Negotiations seemed to stretch on interminably. The Reparation Commission charged with settling the matter could agree only on a principle: that while France and her allies were authorized to press their claims for full damages, actual transfers would be linked to Germany's capacity to pay as gauged by the rate of growth of her exports and her success in obtaining foreign loans.
By linking reparations payments to the condition of the German economy, the Allies diminished the incentive for German policymakers to put their domestic house in order. Hyperinflation was only the most dramatic illustration. Politicians were not encouraged to implement painful programs designed to promote growth by the knowledge that the fruits of their labor would be transfered abroad. The form of the reparations bill hardened German resistance. Including pensions, as insisted on by Britain and the Commonwealth to inflate their share of the total, cast doubt on the French justification for reparations based on the cost of reconstructing devastated regions and reinforced the German belief that the dominant Allied motives were avarice and spite.
An unstable German economy had far-reaching economic and political ramifications. Anything that depressed trade in Germany depressed trade throughout Central Europe. Economic instability in Central Europe intensified fears of a Bolshevik threat from the east, reviving familiar Anglo-French conflicts over spheres of influence in Eastern Europe and undermining the spirit of cooperation developed during the war. Prospects for compromise among the Allies grew increasingly remote.
In the interim, Germany was instructed to begin transfers in kind, mainly coal but also stocks of Reichsbank gold, war matriel, public property in ceded territories and colonies, railway rolling stock, and ships. The coal was essential to a French steel industry handicapped by the destruction of French mines by retreating German armies. These 'interim payments,' justified as a way of defraying occupation costs, were formally distinct from other transfers, although they eventually came to be regarded as the first installment of reparations. Transfers completed prior to May 1921 amounted to 8 billion gold marks (marks of prewar value). This amounted to some 20 percent of German national income in 1921, although it represented only 40 percent of the interim payment specified at Versailles.
It seemed noteworthy that these sizeable interim transfers did not destabilize the German price level or the government budget. They were effected despite continued uncertainty about the size of the reparations bill and despite capital flight from territories scheduled for cession. Since a large part of the interim transfer took the form of public property such as railway rolling stock rather than private-sector production that the government had to pay for by borrowing or taxing, it was relatively easy to mobilize. But insofar as it would be necessary eventually to replace that public property, Germany was mortgaging her future, a fact that could not have reassured outside observers. The presence of Allied troops along the Rhine and the Baltic and the return of domestic political stability following the Kapp Putsch of 1920 have also been invoked to explain the ease of transfer. But troops were no guarantee of compliance, as the Allies would learn in 1923. Only with benefit of hindsight could the failure of the Kapp Putsch be seen as strengthening moderate tendencies within the military. At the time, each of these developments, rather than reassuring domestic and foreign observers, heightened concern over both economic stability and Germany's fragile political equilibrium.
More than the presence of occupation forces or the political climate, the key factor in the interim transfer was Germany's hope that a demonstration of good will would elicit Allied concessions and permit the early extinction of reparations. The Allies had not yet irrevocably committed to their excessive demands. By evincing a willingness to pay on the scale of France's reparations after 1871, Germany might encourage the victors to adopt a more conciliatory stance.
The fiscal implications of the transfer were accommodated by tax reforms guided through the Reichstag by the finance minister, Matthias Erzberger, over the strident opposition of a right wing led by Helfferich. Erzberger's tax package featured an emergency levy and transferred the income tax from the states to the Reich in return for a commitment by the central government to redistribute some of the revenues back to local authorities. The tax increase was essential for maintaining fiscal balance in the face of the interim transfer. German politicians and their constituencies tolerated higher taxes because they anticipated that the revenue would be transferred abroad for only a limited period of time. Rather than provoking capital flight and other forms of evasion, the tax increase was followed by short-term capital inflows in anticipation of possible stabilization of the mark. Since the interim transfer provoked neither capital flight nor currency depreciation, the revenue base of the new income tax was not eroded by inflation.
Following a series of preparatory conferences, the Allies assembled in London in 1921 to set Germany's payment schedule. The U.S. Congress had already indicated its unwillingness to ratify the Versailles Treaty. The American representative to the Reparation Commission was reduced to observer status, limiting his ability to support the British delegation in its opposition to the more extreme demands of France and Italy. Congress's refusal to ratify signalled the resurgence of isolationist tendencies within the United States, which bode ill for those who hoped for war debt cancellation. Given American inflexibility regarding war debts, the prospects for French, Italian, and British compromise on reparations appeared increasingly bleak.
The negotiators at London delivered a reparations bill of 123 billion gold marks, or 31 billion U.S. dollars. This staggering sum was a concession relative to the Reparation Commission's initial recommendation of 225 billion gold marks. Denominating the debt in gold insured that inflation and exchange rate depreciation could not be used to erode its value. Germany was to begin service immediately on 50 billion of the 132 billion total, on which 5 percent interest and 1 percent amortization amounted to 3 billion gold marks (roughly 7 1/2 percent of national income). In addition, she was charged 1 billion marks annually for occupation costs and in settlement of prewar debts (bringing the total to perhaps 10 percent of national income). Payment of the second tranche of 82 billion gold marks was deferred pending an adequate increase in Germany's capacity to pay. These contingencies heightened the uncertainty surrounding the date at which the reparations burden would finally be extinguished. All that was certain was that Germany would be obligated to make substantial transfers over a period of decades.
No issue in twentieth-century economic and political history has been more hotly contested than the realism of this bill. Contemporaries gauged the burden by comparing it to the reparations paid Germany by France following the Franco-Prussian war. France had paid a total of 5 billion francs, roughly one-quarter of French national income in 1872. In comparison, Germany's immediate burden of 50 billion gold marks represented 125 percent of national income in 1921. Including the deferred payments (known as C Bonds) raised the ratio to the 330 percent. At 10 percent of national income, the first year's payments under the London Schedule were very large by prewar standards.
Defenders of the London Schedule observed that Britain had transferred abroad fully 8 percent of national income through foreign lending in 1911-13. This proved, they argued, that the balance-of-payments adjustment mechanism was capable of absorbing a transfer on the requisite scale. But at least some British investment abroad had returned to London as foreign deposits and some in the form of export demands. Together these mechanisms minimized the impact on British industry and on the balance of payments. It was unlikely that either mechanism would operate as powerfully to recycle German reparations.
The politics of the two transfers were even less comparable. Britain had not sacrificed domestic wealth in the amount of the transfer. The British had invested abroad voluntarily with the option of devoting those resources to future consumption. No necessary impact on British living standards resulted. The problem for Germany was how to mobilize for transfer 10 percent of national income and to reduce both present and future consumption without provoking domestic political unrest.
Transforming 10 percent of national income into foreign currency required an external surplus equivalent to 80 percent of 1921-22 exports. One can imagine that strict controls modelled on wartime practice might have succeeded in reducing German imports by 80 percent. But radically curtailing imports was inconsistent with the maintenance of exports given the economy's reliance on inputs from abroad such as copper, cotton, and wool, a dependence that had been heightened by war-time losses of territory and stockpiles. Expanding exports by 80 percent required a further increase in imported inputs, multiplying the gross increase in exports necessary to effect the transfer. And even these calculations left aside the implications of massive import compression for domestic living standards.
Even had Germany somehow been able to provide this astonishing increase in exports, the Allies would have been unwilling to accept it. The problem was not that the incremental exports were so large relative to the British, French, and U.S. economies. The projected transfer amounted, on an annual basis, to perhaps 1 per-cent of their combined national incomes. But German exports would be heavily concentrated in the products of industries already characterized by intense international competition, notably iron, steel, textiles, and coal. The same difficulties would be posed for Allied industries if Germany instead flooded markets with exports. Representatives of these industries were unlikely to accede graciously to a sudden expansion of German exports. Even while complaining that Germany's effort to meet its reparations obligation was inadequate, the Allies raised their import barriers. Keynes, in The Economic Consequences of the Peace, insisted that proponents of reparations specify "in what specific commodities they intend this payment to be made, and in what markets the goods are to be sold." Thomas Lamont of the U.S. delegation to Versailles brought this same point to the attention of the negotiators. The American economist Frank Taussig echoed the warning.
That 1920-21 was a period of recession aggravated both problems: those of Germany's ability to export and the Allies' willingness to import. The Allies would have been happy to accept additional in-kind transfers had they taken the form of raw materials (British reservations about coal notwithstanding). But the German economy could provide these only to a limited extent. Transfers of raw materials disrupted Germany's capacity to export manufactures. Proposals to import German labor for the work of reconstruction were rejected as immoral and politically unpalatable in light of unemployment among demobilized Frenchmen, Belgians, and Italians.
Hence the theoretical question of what change in prices would be needed to clear international markets in the presence of reparations (known as the 'transfer problem') was ultimately beside the point. Keynes's conclusion was that to generate a trade surplus on the order of 80 percent of initial exports, a very considerable decline in the relative price of German goods would be needed to switch foreign demands toward German exports and German demands away from imports. He raised the possibility that, if demands were sufficiently inelastic, a decline in German export prices might reduce the value of German exports at the same time it raised their volume, rendering the transfer impossible at any price.
<#FROWN:J56\>
Such an assessment accurately condensed the realities of Eire's military situation - as well as its political position at the start of the war. Others, however, noted that buoyant morale in the Irish services outweighed mere physical constraints, that it was guerrilla tactics that carried the day in the War of Independence.
After hostilities broke out, the total strength of Ireland's army would reach only a little over half of the authorized levels. Nineteen thousand men - or two divisions - would be in uniform within days of the Emergency's beginning; this compared to the 136 divisions the Germans fielded in the Battle of France. Nearly every Irish unit was understrength. Of the eight authorized rifle battalions, none was yet organized. The cyclist squadrons - known as the 'Piddling Panzers' - would have posed precious little threat to real panzers. The army had at its disposal two 'serviceable' tanks and 21 armored vehicles, most of the latter of which were already in 1939 antiques - 1920, and earlier, Rolls-Royces. The air force, a branch of the army, was equally toothless, with only 24 craft, of which 10 might be called modern. There was no navy other than a small coast guard unit. Prior to May 1940, no strike force capable of resisting any invader existed. Few commands had reserves. Armaments, ammunition, and vehicles were extremely scarce.
But so quickly did the land war turn into Phony War that the army's General Headquarters announced on September 21 that "the present Emergency does not constitute a war situation and it would not be justified in maintaining its Establishment and that the strength should be reduced." By Christmas, most of the army personnel who could manage the journey had gone home for the holiday. Only an IRA raid on the central army munitions magazine in Dublin's Phoenix Park, in which the raiders escaped with large quantities of arms and ammunition, caused headquarters to order troops back to their duty stations.
During these early months of the war, Irish soldiers weren't even sure of receiving a uniform. That was perhaps as well. With some irony, and undisguised distaste on the other side of the Irish Sea, the soldier of the Twenty-six Counties looked uncannily like his counterpart in the Wehrmacht - most notably so in the same 'iron scuttle' steel helmet both forces wore. It was likely this unfortunate symbolism rather than the uniform's scratchy uncomfortableness that caused the Army Department to scrap it in 1940. In its place came a new uniform little distinguishable from the British pattern, soup-plate helmet and all (the latter being the same style helmet worn by American GIs until replaced in 1942 with the familiar rounded, nearly brimless model).
Not only were the soldiers dressed poorly; they were housed poorly as well. Most of the barracks remained as leftovers from the British regime, with little new military shelter built since. Many derelict country houses had been fixed up to provide minimal standards, and farmboys turned soldiers suddenly found themselves living in once grand but now sadly dilapidated mansions.
Privates earned 14 shillings a week - about one dollar in contemporary terms, but then with strikingly greater purchasing power, of course. The income wasn't entirely discretionary, though. A forced haircut deduction of two pence was taken out of each pay packet, as was six pence for laundry and another tuppence for 'social welfare.' In theory the Irish soldier was fed better than his civilian compatriots - supposedly a daily three-quarters of a pound of 'best home-fed beef,' a quarter pound of fresh vegetables, a 'liberal' quantity of butter, cheese, jam, eggs, sausages, bacon, etc. In fact, his diet consisted of the usual monotonous regime of many armies: oatmeal, brown stew, jam rolls, bread and butter, and tea.
To meet the costs of their new defense requirements, the government forced the Irish taxpayer to pay taxes higher than any ever known. The first increase predated the outbreak of war: in the spring of 1939, the new pounds5.5 million defense appropriation meant jumping the income-tax rate by a half shilling to a shilling on the pound - 5 percent. Along with this income-tax increase came new surtaxes, as well as additional taxes on the richest ratepayers. Two months after Germany attacked Poland, the income tax went up another shilling on the pound, together with higher increments in estate duties and new levies on beer and whiskey. Through the next two years, tax increases on income could rise until the Irish citizen was paying on average 37.5 percent of his income to the government. Though the Twenty-six Counties remained at peace, their government was assessing tax levies as onerous as those of most of the countries at war.
<*_>three-bullets<*/>
If the island was not yet threatened by a Wehrmacht held in its traces, it was threatened by the ruthless and bomb-prone activities of the IRA. The Irish Republican Army's raison d'<*_>e-circ<*/>tre was to end British control of Ulster by whatever means necessary, however appalling or murderous. The terrorist organization's position was, simply stated, that "England's difficulty is Ireland's opportunity." To achieve its goal, the organization maintained a complete, albeit underground, government, a constitution, and some 7,500 mostly youthful members - plus perhaps 15,000 more or less dedicated supporters (the figures are from Time magazine).
What popular support the IRA received in the south during the Emergency was given almost entirely in token of the perceived injustice of the island's partitioned status. As to the organization's relationship with the Dublin government, its policy provided that no terrorist activities would be carried out against Eire, as long as the IRA was free to carry out from southern bases operations against Ulster and that province's British targets. De Valera refused such a concession, understanding full well the danger of British retaliation against Eire.
Some months before the outbreak of war, the IRA had undertaken to traumatize the British people into demanding that their government leave Ulster. The shock was carried to Britain itself in the form of a series of terror bombings. In January 1939, young Irishmen recruited in Britain set off explosions in what were, with war approaching, the kingdom's most vulnerable sites: factories, power stations, and telephone exchanges. At the time, even a few well-placed blows against British defense facilities were enormously crippling to the catch-up effort to match Germany's industry. The campaign reached its moral nadir back home with a bombing of the Irish country hotel where Prime Minister Neville Chamberlain's son Francis was spending a hunting holiday, the fortunately ineffectual assault an apparent attempt to sour the personally cordial relationship between Chamberlain and de Valera. Britain endured over a hundred more explosions in July alone, with blast sites including Piccadilly Circus and Madame Tussaud's waxworks. Harried police waded through crowds arresting anyone with a brogue. The IRA's outrages culminated in an act that finally crystallized public opinion and marshaled concrete action against the outlaw organization.
On August 25, a package-laden bicyclist made his way through the crowded streets of Coventry, an ancient and, by 1939, heavily industrialized Midlands city. The rider left his parcel - a pre-fused bomb - at a caf in crowded Broadgate, in the center of the city. To hide the device's origins, its makers had assembled it in one place and brought it carefully to another, where the last man in the deadly chain put it in his cycle's carrier basket. Delayed by traffic and worried that the bomb would blow him up along with the innocent bystanders who were its intended victims, the anxious IRA terrorist hurriedly threw his bicycle against the wall on the caf and left. The explosion a few moments later blew off the front of the building, along with the windows of the neighboring shops. Ankle-deep debris settled over a wide area. Five people lay dead, including an eighty-one-year-old man and a small boy. Seventy more were injured.
Recognizing the threat this and the earlier outrages represented to an Anglo-Irish accord, the authorities reacted by searching every Irish home in Coventry, jailing hundreds of activists, and ending with the apprehension of three members of the city's IRA unit. Two others were later arrested for complicity in the terror attack. The dragnet resulted in an immediate and sharp decline in IRA terror activities in Britain for the rest of the war. But among Eire's citizenry who deplored the IRA's methods, so deep was the vein of antipathy for Britain that when two of the accused were hanged in Birmingham in February 1940, almost the entire country mourned them, with flags dropping to half staff, theaters closed, and masses offered for the repose of the executed men's souls.
Hitler understandably regarded people who could commit such acts against Britain as his natural allies. In fact, Germany had been trying to cement a relationship with the terrorist organization since at least 1937. The military intelligence agency, the Abwehr, directed by Admiral Wilhelm Canaris, initiated planning in November and December of 1939 to send its agents into Ireland by submarine to establish contacts with the IRA, with German agents instructed to tell prospective recruits among the Irish that Germany strongly desired a united Ireland and that the best course for the IRA would be to join efforts with the Reich in destroying 'England,' the sooner both their goals being fulfilled.
But Hempel warned his superiors in November 1939 that Germany had best not rely too heavily on playing the IRA card. On the fourteenth, he wrote to Berlin that "the I.R.A. is hardly strong enough for action with promise of success or involving appreciable damage to England and is also probably lacking in a leader of any stature." He pointedly cautioned the Foreign Ministry that open cooperation with the IRA would very likely lead the more moderate sections of the Irish republic into blaming the organization for making the country's national interests dependent on Germany, which "in view of the widespread aversion to present-day Germany, especially for religious reasons, could rob the I.R.A. of all chances of future success." Hempel also noted, again, that such a course would give Britain an excuse to intervene militarily in solving its own outstanding problems with Eire.
Because of the IRA's potential to damage Eire's neutrality policy more than any other group in Irish politics, the organization's British atrocities had in June 1939 given de Valera a good excuse to outlaw it. The move enabled him to marshal the resources of the state in chasing down its members and generally branding it a menace to Eire's survival in the Emergency. But by the end of the year, the IRA openly declared its sympathies lay in a GermanyGerman victory in the war, evidently on the amazing deduction that such an outcome would, by Britain's defeat, mean the end of partition. The perhaps more likely possibility that Hitler would occupy Ireland - the whole of the island - apparently didn't occur to the IRA leaders. Though Nazi Germany's most scorching depravities and betrayals lay in the future, its many double crosses up to this point should certainly have put the IRA off any hope that it or the nation for which it purported to fight would be treated with respect by Adolf Hitler.
On December 23, 1939, the organization carried off one of its grandest coups - though one of its last - against the Dublin government. When the IRA stole more than a million rounds of ammunition and cases of guns from the Phoenix Park arsenal, it looked as though the phantom army might turn itself into a real army and attempt a coup d'tat, or even try to start another civil war. D<*_>a-acute<*/>il member James M. Dillon warned of the raid: "I believe the ultimate end of the activities of these gentlemen [the IRA] must be assassination. God knows how many of us may be victims!"
The reaction of the government was to arrest every member of the IRA who could be rounded up, sending 5,000 Special Police armed with rifles to seal the frontier with Northern Ireland and hunt down the clandestine terrorists. To further the search-and-destroy operation's success, the government rushed through Parliament a bill suspending the constitutional guarantee against holding suspects for more than forty-eight hours without evidence.
<#FROWN:J57\>Moreover, in recent years, most of those in pursuit of a truly 'British' past have been in thrall to a series of remarkable articles written by J.G.A. Pocock. In the Journal of Modern History in 1975, and in the American Historical Review for 1982, Pocock argued that British history could only be understood as "the interaction of several peoples and several histories." By this, it is important to note, he meant not only the relations that existed over time among England, Wales, Scotland, and Ireland but also the broader connections between these four countries and North America and the rest of Britain's 'white' empire, including Pocock's own native New Zealand. Predictably, though, it has been his insistence on the need for study of the four component parts of the United Kingdom - and not his geographically more wide-ranging manifesto - that has generated the greatest interest among British historians.
Some of the results of this new scholarly fashion have been entirely benevolent. Our collective consciousness has been raised, and we are now much less likely than we were even ten years ago to describe exclusively English events and trends as though they were necessarily synonymous with British developments. We have come to understand with more precision than before that Great Britain is a composite structure forged, as France and Spain were forged, out of different cultures and kingdoms. And by examining how these entities effected each other in the past, we have been able to approach familiar historical events in a different and revealing light. Sir John Seeley remarked as long ago as 1895 that the interaction between Scotland, Ireland, and England was so extensive in the 1640s that he wondered whether the civil war "had really its origins in the necessity of revising their mutual relations." But it is only in the past few years that this insight has been pursued to the full.
These are substantial gains. But while acknowledging them as such, we also need to be aware of the problems and limitations inherent in this approach to the British past. To begin with, some of its practitioners are undoubtedly swayed by current political preoccupations, and this can lead to a certain amount of special pleading. Especially since the 1960s, both the Welsh and the Scottish nationalist movements have increased in size and self-consciousness (as simultaneously has support for an independent Basque country and Catalonia in Spain and for separate Breton and Occitanian nations in France). In addition, one of the consequences of Margaret Thatcher's long premiership, which saw a savage reduction in Tory electoral support in Scotland, and a less dramatic but still significant fall in Tory support in Wales, has been the reemergence of a right-wing Little Englandism. (The Labour party, for reasons that will become clearer later in this essay, remains emphatically British in its electoral base and ideology.) Put crudely, the current political situation has encouraged some English scholars to view the Welsh and the Scots as the Other in a more deliberate fashion than before, and vice versa. If we add to this the fact that Protestant Britons have traditionally viewed the predominantly Catholic Irish as the Other, and have been so viewed in return, it is easy to see why the appeal of a Four Nations view of the United Kingdom can seem so overwhelming quite independent of its scholarly value. Such an approach can reduce Britishness to the interaction of four organic and invariably distinct nations (or three if Ireland is left out of the story). As such, it can sit comfortably not only with Welsh, Scottish, and Irish nationalism but also with a newly assertive English nationalism.
The breakup of Britain, or at the very least the emergence of a federal Britain existing as part of a federal Europe, may well be desirable goals for the 1990s. I am not concerned here with vindicating unionism or to argue for its continuation in the future. But I would argue that the Four Nations approach, if pushed too hard or too exclusively, is an incomplete and anachronistic way to view the British past and, also, a potentially parochial one.
It conceals, if we are not careful, the fact that the four parts of the United Kingdom have been connected in markedly different ways and with sharply varying degrees of success. Most conspicuously, Ireland, as a whole, was only part of the Union between 1800 and 1920. It has always been divided from the British mainland by the sea and since the sixteenth century has been severed from it even more brutally by its strictly limited response to the Protestant Reformation. There is considerable evidence that at grass-roots level the Welsh, the Scottish, and the English saw (and often still see) the Irish as alien in a way that they did not regard each other as alien. None of this means that we should ignore Ireland's many and important political, cultural, and economic links with Britain. But we should recognize that, mainly for religious reasons, the bulk of its population was never swept into a British identity to the degree that proved possible among the Welsh, the Scots, and the English. We also need to recognize that, until the late nineteenth century, at least, the majority of people in all of these countries were never simply and invariably possessed by an overwhelming sense of their own distinctive identity as Englishmen, as Scotsmen, as Welshmen, or even as Irishmen. As in the rest of Europe, intense local and regional loyalties were always there to complicate and compromise.
Even in the early 1800s, for example, and despite the enormous impact of Sir Walter Scott's heroic evocation of the lochs and glens of the North, some Lowland Scots still automatically referred to their Highland neighbors as savages or as aborigines. They regarded them, as they had traditionally done, as impoverished and violent, as members of a different and inferior race, rather than as fellow Scots. Conversely, whereas the word 'sassenach' is now one of the kinder epithets used by all Scots to refer to the English, before 1800 the Gaelic sasunnach (meaning a Saxon) was commonly employed by Highland Scots to refer to Lowlanders in general as well as to the English. Quite logically in ethnic terms, Highlanders could view both Lowland Scots and the English as foreigners. By the same token, the inhabitants of northern England had (and still have) far more in common with their Lowland Scottish neighbors than with the inhabitants of southern England. They read the same books, ate the same kind of food cooked in similar ways, frequently intermarried, and shared similar literacy levels. Much the same could be said of men and women living in Herefordshire and Shropshire with regard to their Welsh neighbors. Here, again, people living close to the border, whether on the Welsh or on the English side, could have more in common with each other than with the rest of their respective countrymen. As Hugh Kearney has demonstrated, with a scrupulous honesty that threatens at times to undermine his own arguments, imposing a strict three- or four-nation model onto these intricate and myriad regional alignments is difficult and distorting. In practice, men and women often had double, triple, or even quadruple loyalties, mentally locating themselves, according to the circumstances, in a village, in a particular landscape, in a region, and in one or even two countries. It was quite possible for an individual to see himself as being, at one and the same time, a citizen of Edinburgh, a Lowlander, a Scot, and a Briton.
The invention of a British national identity after 1700 did not obliterate these other, older loyalties. True, both before and after that date, London was always ready to employ military force, parliamentary legislation and various kinds of indoctrination to limit the autonomy of a few, particularly dangerous regions - the 'pacification' of the Scottish Highlands after 1746 would be an obvious example. But Britishness was never just imposed from the center, nor can it be understood solely or even mainly as the result of an English cultural or economic colonization of the so-called Celtic fringe. The extent of such anglicization has, to begin with, often been exaggerated. Scotland always preserved its own religious, educational, and legal structures and its own sophisticated network of printing presses and cultural centers, while even in the 1880s, some 350 years after the Act of Union between Wales and England, three-quarters of all Welshmen still spoke their own language out of choice. More broadly, though, we need to stop thinking in terms of Britishness as the result of an integration and homogenization of disparate cultures. Of course, a degree of integration did occur, mainly by way of the advance of communications, the proliferation of print, the operation of free trade throughout the island, and a high level of geographical mobility. But what most enabled Great Britain to emerge as an artificial nation, and to be superimposed onto older alignments and loyalties, was a series of massive wars between 1689 and 1815 that allowed its diverse inhabitants to focus on what they had in common, rather than on what divided them, and that forged an overseas empire from which all parts of Britain could secure real as well as psychic profits.
It is this vital and external dimension of British development that is most likely to be obscured by too narrow a concentration on the Four Nations model. The interaction of Wales, Scotland, Ireland, and England is an important and fascinating theme and is a particularly pertinent one at the end of the twentieth century. But in the eighteenth and nineteenth centuries, Britishness was forged in a much wider context. Britons defined themselves in terms of their common Protestantism as contrasted with the Catholicism of Continental Europe. They defined themselves against France throughout a succession of major wars with that power. And they defined themselves against the global empire won by way of these wars. They defined themselves, in short, not just through an internal and domestic dialogue but in conscious opposition to the Other beyond their shores.
II
The absolute centrality of Protestantism to the British experience in the 1700s and long after is so obvious that it has often been passed over. Historians, always reluctant to be seen to be addressing the obvious, have preferred to concentrate on the more subtle divisions that existed within the Protestant community itself, on the tensions between Anglicans and nonconformists in England and Wales, between Presbyterians and Episcopalians in Scotland, and between the older forms of Dissent and newer versions such as Methodism. These internal rivalries were abundant and serious. But they should not obscure what remained the towering feature in the religious landscape, the gulf between Protestant and Catholic.
Even after the beginning of large-scale Irish immigration, the Catholic community on the British mainland was a small one, and its members were usually able to socialize with their Protestant neighbors, own land, earn a living, and even attend mass openly. Yet in terms of prejudice, none of this mattered very much. Irrespective of their real strength and of how they were treated as individuals, Catholics as a category remained in popular mythology an omnipresent menace. Every November 5 until 1859, worshipers at virtually all Protestant places of worship in England and Wales would be reminded that it had been a Catholic who had tried to blow up James I and Parliament back in 1605. In England, Wales, and Scotland, almanacs, sermons, and popular histories made the point, year after year, that it had been a French Catholic Queen, Henrietta Maria, together with her interfering priests, who had led Charles I astray and the whole island into war; that the would-be tyrant James II had been Catholic, just as those responsible for the Saint Bartholomew's massacre in 1572, or the Irish 'massacres' of 1641, or the Great Fire of London in 1666 had been Catholic also. "While Britain continues to be a nation," wrote a Scottish pamphleteer at the end of the eighteenth century, "she ought never to forget."
<#FROWN:J58\>It could be argued from Hutton's own evidence that many of the problems facing Charles II, especially in his early years, were the result of the legacy of the civil war and of the difficulty at the Restoration of finding a satisfactory settlement that could accommodate the interests of former Royalists, Parliamentarians and ex-Cromwellians, Anglicans and Dissenters alike. Certainly the religious tensions generated by the issue of Dissent and the unsatisfactory Church settlement perpetually bedeviled the government in its dealings with all three realms throughout the reign. However, Hutton sees the year 1673 as marking a crucial turning point in Restoration politics, with the duke of York's public profession of his Roman Catholicism after the passage of the Test Act and his marriage to a Catholic princess ushering in a new period of political instability, centering around the problem of the Catholic succession. The succession issue was to lead to the clash between Whigs and Tories after the revelations of the Popish Plot in 1678, but whereas Hutton is prepared to accept the existence of a "Whig party" (p. 422), he is adamant that "there was no 'Exclusion Crisis' at all" (p. 357), because, he says, the situation never became critical for the government. This argument is somewhat strained, since by Hutton's own account the government faced a series of acute difficulties during the Exclusion period that virtually crippled its ability to govern properly: there were problems in forming and keeping together an effective ministerial team; it was impossible to get anything done through Parliament; Charles was unable to conduct any meaningful foreign policy; and the government faced the opposition of an important section of the nation.
Although we learn a lot from this book about what Charles and his various ministers actually did, we do not get a very clear sense of the nature or essence of royal power at this time. The position of the monarchy is said to have been fundamentally strong, and even to have gained in strength in the last years of Charles's life, but the basis of this strength is never properly analyzed. Rather than placing sole emphasis on the inherent powers of the Crown, I would suggest there was also a crucial ideological dimension to the strength of the monarchy in the first half of the 1680s. After the Exclusion scare, Charles made every effort to cultivate public opinion - through propaganda, public pronouncements, and even royal entries and processions. He portrayed himself as a king committed to the rule of law and the defense of the Church of England against the subversive threat of the Whigs and Nonconformists, thereby attempting to recapture the soft Anglican middle ground that had become alienated from the Crown during the 1670s. The monarchy was stronger in 1684 than in 1679 because more people supported what it was doing; that support was achieved partly as a result of a successful public relations exercise but also partly as a result of a shift in policy by the king himself, with Charles at last fulfilling the role of a Cavalier-Anglican monarch that so many of his subjects wanted him to play. Hutton has all the material at his disposal to discuss such questions; it is disappointing that he decided not to do so in a direct matter. Instead he seems to have seen his task as getting the facts straight for his readers so that they could be in a position to draw their own conclusions from the evidence he has chosen to present.
Two of the other books under review look at individuals whose lives and influence spanned the crucial period between the civil war and the 1680s. Conal Condren focuses on George Lawson - long thought to be significant for his criticism of Hobbes and influence on Locke - and on his tract Politica, which was published first in 1660 and again in 1689. Condren's book operates on a number of levels: in addition to being a work of political philosophy and of the history of ideas, it is also a study in political linguistics, and it has much to say on the use of metaphor and rhetoric and on Lawson's manipulation and subversion of existing political vocabularies. As a result the argument is complex and not easy to summarize. Lawson, a clergyman who worked most of his life in Shropshire, was a supporter of Parliament who not only found it possible to accommodate himself to the changing political and religious regimes of the 1640s and 1650s but could accept the Restoration of monarchy in 1660 and work within the reestablished national Church as well. Although the Politica was in gestation for some time before its publication, Condren shows that its context belongs very much to early 1660; therefore it should not (as has been suggested) be seen as a defense of the Commonwealth or of Cromwell but, rather, as a settlement tract that sought to diffuse the divisiveness of constitutional and religious differences on the eve of the Restoration. Thus Lawson's ideal governmental form allowed for any variation of rule by king, Lords, or Commons, or even a republican elite. Likewise any church form might do in a pinch, so long as it did not lock itself into the rhetoric of divine origin. Condren sees Lawson's arguments as in many ways reflecting Clubmen rhetoric, with loyalty to the community and the country emerging as major themes in his work. As many will know, Lawson made a distinction between real and personal sovereignty, with the former being invested with the people as a community, and the latter being the attribute of a particular governmental form. His precise views on the limits of people's subjection, however, are not easy to unravel (necessarily so, perhaps, given that Lawson's stress was on settlement and subjection), so Condren prefers to offer two competing readings of Lawson's position on resistance (the 'cobweb' and the 'seesaw' hypotheses), which he feels circumscribe the limits of plausibility. Nevertheless, Condren maintains that Lawson essentially had a theory of dissolution rather than resistance per se.
Different aspects of this book will appeal to different types of scholars in varying ways. There is a fascinating section on Lawson's interpretation of the civil war, which he saw as having primarily religious causes - although Condren warns us that the attempt to separate discrete factors (religious, constitutional, economic) is really misguided, given the way contemporaries conceptualized their world. He is skeptical of the impact that Lawson is often thought to have had on Locke. Rather it is John Humfrey, that reluctant Nonconformist, who emerges as the main vehicle for the transmission of Lawsonian ideas between the Restoration and the Glorious Revolution. There is also a good discussion of the importance of the Politica in the "Allegiance Controversy," which reminds us how relevant the rhetoric of interregnum religious and political discourse still was to the world of 1689.
Toward the end of his study, Condren offers an intriguing, though all too brief, criticism of the way other scholars have deployed the label 'radical' and its siblings 'moderate' and 'conservative.' None of these, he argues, are particularly helpful to our understanding of either Lawson or his text, since they are anachronistic terms that apply to groups in nineteenth-century politics in a way that would not be acceptable to the world of the seventeenth. This important point has been made many times before, though there is no harm in saying it again, so long as it does not blind us to the fact that the late seventeenth century did have its own concept of 'radical,' even if it was very different from that which came into existence after the French Revolution. A theory of government that placed ultimate sovereignty in the people was, so far as the Tories of the Exclusion period were concerned, a 'radical' theory. As one author alleged, the Whigs did their best to infuse into people's minds the belief "that Power is radically and revokably in them." Thus, according to Whig theory, "they [the People] do not absolutely part with this their so natural Right, but commit onely the Administration of such Power as is radically in them to others. But they retain to themselves much of this Right, as upon the Male-administration of the Power so delegated, they may revoke the Delegation, and take all the Power into their own hands again." Many of the more extreme Whigs, Locke among them, would qualify as espousing a 'radical' theory in this sense of the term; whether Lawson would no doubt partly depends on whether one has a cobweb or seesaw reading of the Politica.
Not all are going to accept the conclusions of this book. Julian Franklin, a leading scholar of Lawson and Locke who comes in for serious criticism from Condren, has already replied in print by identifying what he sees to be a number of errors of interpretation. More worrying, I fear, is the fact that most people will have terrible trouble just understanding what Condren is trying to argue. There is a certain irony in the fact that a student of linguistics should find it so difficult to articulate his views in a clear and comprehensible manner, and on one occasion even be forced to give up words altogether and start writing in mathematical equations (p. 108). This is too bad, since I fear many will be put off by the style and therefore miss the many stimulating and provocative ideas that Condren's book undoubtedly contains.
Jonathan Scott, by contrast, has produced a lively and highly readable account of that great seventeenth-century 'radical,' Algernon Sidney. (The term 'radical' still seems appropriate in this case, even after reading Condren.) Although at times Scott's prose styles is rather rough and even colloquial (contractions and split infinitives abound), he writes with pace, verve, and a degree of wit that ensures his reader remains perpetually enthralled with his argument. The present book, which is the first in a two-part study, concentrates on the years of Sidney's life up until 1677 and contains an analysis of his other major political treatise, the Court Maxims written in 1665. The sequel (which was not available at the time of writing this review) is to be entitled Algernon Sidney and the Restoration Crisis, 1677-1683, and will contain a detailed analysis of the more famous Discourses. Not that Scott adheres to the strict chronological limits that the titles of his respective volumes suggest. The main purpose of the first study is to examine what Sidney believed and how he came to hold such beliefs, and this is achieved through a combination of biographical narrative, analysis of his intellectual influences, and exposition of his ideas, which takes us back and forth across the whole of Sidney's lifetime and has much to say about the Discourses as well. This is an admirable book: rich, clever and provocative in its revisionism, and to judge from the glimpses we are offered here, the subsequent volume promises to dust off the old cobwebs and set the seesaw rocking through a powerfully argued reconceptualization of the Exclusion period.
The main thrust of Scott's argument, which explains why a separate volume on the period up to 1677 is necessary, is that Sidney was not an Exclusionist Whig whose ideas were formed in response to the so-called Exclusion Crisis; rather, his attitudes were critically shaped by the age of the English Republic (1649-53 and 1659) and his struggles during the period 1635-77 more generally. Scott shows that the traditional view of Sidney as the great popular and patriot philosopher, who was a reformer and a moderate rather than a radical, has got the man seriously wrong. Not only was he "one of the most passionate and bellicose rebels of his age" (p.3), but he was also far from being "the perfect Englishman," since he spent half of his life outside England, was deeply influenced by Continental ideas and his internationalist perspective, and was engaged among foreign princes and republics in a variety of acts of treason against his own country.
<#FROWN:J59\>
Walter Benjamin would ultimately admonish the modern intellectual's ambiguous politics, exemplified by this kind of neutralization of race, gender, and class allegiances. Western intellectuals, he observed, did not see themselves as "members of certain professions" but as representatives of a "certain characterological type," a type located somewhere between the classes. Advocating a more activist role for the intellectual, Benjamin called "for the transformation of the forms and instruments of production in the way desired by a progressive intelligentsia - that is, one interested in freeing the means of production and serving the class struggle."
Piper's realignment of the artist/spectator relationship rests on her desire to work beyond such a characterological suspension of the artist's connection to the rest of humanity. This challenge to the forms and instruments of art production resulted both in a withdrawal from the precious art object and a travesty of modernist conceits: the 'apolitical' intellectual, the 'classless' dandy, and the 'objective' fl<*_>a-circ<*/>neur who wanders the streets of Paris observing the heroism of modern life are for Piper the subject of parody and even derision. Take, for example, her Mythic Being. Masquerading in dark glasses, Afro, and pencil mustache, the Mythic Being was Piper's male alter ego. Neither dandy nor fl<*_>a-circ<*/>neur - yet strangely reminiscent of both - the Mythic Being represented himself as tough black street kid who engaged in charged and sometimes hostile encounters with strangers. Piper sent him into white middle- and upper-middle class social contexts - theaters, gallery receptions, museum exhibitions - in order to observe racist patterns of avoidance and aggression. After moving to Cambridge, Massachusetts, in 1974 to begin graduate study in philosophy at Harvard University, Piper involved her Mythic Being in such actions as 'cruising' the streets to experience male sexuality, wandering the Cambridge commons, and venting class antagonisms by mugging a male accomplice in public view after a provocative conversation.
If Piper's 'whiteness' allowed her to escape from the cruelest side of racism, the racial and class specificity of her male alter ego left her particularly vulnerable. Gaining a sense of her "own marginality as a nonwhite (but not obviously black) member of society," Piper realized that she was now even more threatening to white society at large. As Homi Bhabha observes, "the black presence ruins the representative narrative of Western personhood: its past tethered to treacherous stereotypes of primitivism and degeneracy will not produce a history of civil progress, a space for the Socius; its present, dismembered and dislocated, will not contain the image of identity that is questioned in the dialectic of mind/body and resolved in the epistemology of 'appearance and reality.' The white man's eyes break up the black man's body and in that act of epistemic violence its own frame of reference is transgressed, its field of vision disturbed."
Ultimately, Mythic Being personifies the mythic black man - the presence who ruins the representative narrative of white America. He is the being for whom miscegenation laws were invented, codes that pretended to 'protect' white women "but left black women the victims of rape by white men and simultaneously granted to these same men the power to terrorize black men as a potential threat to the virtue of white womanhood." The tragic irony of this social equation is that black men's 'power,' though threatening to white society, is most often fictive and allusive; since male power within patriarchy is relative, poorer men, frequently men of color, are most often denied the material and social rewards of their participation in patriarchy. With her Mythic Being, Piper conflated issues of race and gender in order to question problematic feminist constructions that pit white women against white men in a struggle for power. Any understanding of patriarchal relations which examines the power of men over women outside of broader racial and economic issues is, of course, problematic. Arguments that feature patriarchy as the primary determinant of women's oppression fail to see "the inapplicability of such a concept in analyzing the complex of relations obtaining in the Black communities both historically and at present." Such discourses of patriarchal relations, for example, most often ignore the struggle that men and women of color must fight together against the white ruling class. Mythic Being - a black woman in the guise of a black male youth - metaphorically represents this alliance, reminding us that in Western society racism effortlessly crosses the boundaries of gender.
Piper's desire to speak over the oppressive master texts of modernism results in the blasting of another aspect of the artist's mythology - the specter of anonymity. Like Frantz Fanon's existentialist evocation of the 'I' that restores the presence of the marginalized (and his own presence within dominant narratives that would silence a black man's voice), Piper at times speaks through personal or autobiographical narratives. The Big Four-Oh (1988), a self-portrait video installation, juxtaposes a display of materials soaked in her bodily fluids, an open journal, a suit of armor, forty hard balls, and a repeat videotape of Piper with her back to the camera dancing nonstop to soul music. The work, which represents the artist at the time of her fortieth birthday, affirms her complexity and strength, each of its parts serving as a metaphor for an aspect of her existence. This installation would seem to continue the project of the earlier Political Self Portraits (1980) in which Piper superimposed images of her face and body over detailed autobiographical accounts written from the three perspectives that define her marginalization - class, race, and gender. A chronology of the artist's life written by Piper and published in the catalog for her recent retrospective exhibition includes the following note: "This chronology was created solely by Adrian Piper and is presented as part of her artistic work." Rather than narcissistic evocations of the self or acts of romanticized self-expression, autobiography serves a crucial political function in Piper's <*_>oe-ligature<*/>uvre. As part of a broader structure for dismantling oppressive systems, such narratives acknowledge the extent to which marginalized peoples are spoken at and for, the degree to which people of color, women, gay men, and lesbians have been defined and judged by the narrow standards of a dominant culture governed by white heterosexual males. Indeed, who constructs the master narratives of culture? Who are the patrons of academia, the publishers, the financiers of industrial society? Who writes history?
Marginalized peoples, of course, are generally excluded from defining their own role in the narratives of history. Women of color, for example, have been alternately categorized as exotic or 'pathological' or both - universalizing conditions that deny difference as they create stereotypes of passivity or abnormality: "But what does the sameness of the exotic women represent? Female heroism, humor, carnivalesque gesture, triumph, movement ... 'trans-cultural, trans-historical, trans-social' - exotic. Here exoticism marks a universality which systematically negates the very raison d'<*_>e-circ<*/>tre of women's different experiences, strategies and actions." Like the mythologized dandy, these ciphers of universality, exoticism, or sameness mask deeper political motives. As such, autobiographical writing can serve an important therapeutic function for marginalized peoples. The fear and uncenteredness associated with psychic and physical oppression can often be overcome or helped by reconnecting with the personal narratives of the past. Remembering can be part of a cycle of reunion, observes Bell Hooks, "a joining of fragments, 'the bits and pieces of my heart' that the [autobiographical] narrative made whole again." Such speech will always be difficult for the dominant culture to accept. Representational marginalization exempts the exotic 'others' - whether they be women, gays, blacks, or even artists - from serious consideration by the ruling class. To allow such peoples to speak in their own voices is to risk hearing their oppositional speech - discourses that demand rather than passively accept, that scream rather than whisper.
Although Piper's project is directed at a cultural community that is mostly privileged, she reconstructs the ideological role of the artist in a way that directs her audience to join her in the struggle against racism and sexism. Indeed, she is always polite in her address, always conscious of the psychological threshold of complacent viewers who would rather look at 'art' than confront the painful reality of their own racism. "I can't bear the thought of violating the norms of etiquette," Piper has said. "Such norms help me to grease my way ... through a hostile white world." The inherent difficulty of disseminating upsetting information makes this politeness a matter of packaging for Piper; like a good advertising executive, she understands that style is often as important as content in reaching an audience. It is through this "power of passive provocation" that Piper hopes to transform her audience psychologically, by presenting them with "an immediate and unavoidable concrete reality that cuts through the defensive rationalizations by which we insulate ourselves against the facts of our political responsibility. I want viewers of my work to come away from it with the understanding that their reactions to racism are ultimately political choices over which they have control - whether or not they like the work or credit it for this understanding."
To this end, Piper embarked on a series of audience performances in which predominantly white, art world groups were engaged in various consciousness-raising activities. The ground-breaking Funk Lessons, for example, began as a question in Piper's mind: Why are white people indifferent, even hostile, to soul music? "I found that response [to funk music] so often," Piper observes. "It seems to me there was a gap between the purported attitude of openness and receptivity to popular culture that is usually espoused by the art world, according to which anything is adequate subject matter for appropriation and reuse within the context of high culture. And what actually seemed to be the case is that in fact only some things can have that function, and in particular black working-class culture cannot have that function." Having incorporated funk music into earlier politically oriented performances, Piper found that white audiences misunderstood her motives or attacked her use of this music as "cheapening" the serious political content of the performance. Such resistance, Piper suggests, is rooted in several areas: "One problem has to do with the overt sexuality of that music - it talks about fucking, it talks about making love, it talks about bodies, and it's very hard to assimilate that in a way that's not threatening to white upper-middle-class culture. Another problem is that it requires a very highly structured use of one's body in order to respond to it."
Piper decided to 'educate' white audiences about the significant (if not always acknowledged) role played by this music in both dominant and marginal culture. Funk Lessons was structured in an academic format as a participatory performance with Piper as instructor and audience members as students. Piper distributed a bibliography and photocopied handouts that listed some of the 'characteristics' of funk dance and music. She proceeded to discuss certain mainstream presumptions about funk music in an attempt to free her audience from discomforting misconceptions. She then led the group in body isolation exercises, discussed the structure of the music, and practiced dance movements with musical accompaniment. (An extraordinary video version of Funk Lessons centers on a particularly successful performance at the University of California at Berkeley in 1983. Piper augments film of the Berkeley evening with voice-over and on-screen commentary and archival footage of influential soul performers [e.g., James Brown, Little Richard, and Aretha Franklin] and the white entertainers [Elvis Presley, Mick Jagger, The Talking Heads] who capitalized on the black funk idiom. The video yields a level of meaning somewhat independent of but not unrelated to the performance - that of the mass-media suppression of working-class black culture and its comfortable, and profitable, appropriation by white artists.) Ultimately, Piper did not want to scare or intimidate her audience; rather she constructed a "comfortable and safe" format that encouraged people to explore their apprehensions about the music and their ability to soul dance. Individual audience reactions to Funk Lessons varied from antagonism and resistance to jubilation. Successful performances culminated in a celebratory dance party; failed ones fizzled out in a morass of confusion and resentment.
<#FROWN:J60\>Lady Elisabeth's disappointment in him was most vocal by the middle of that year. Talbot was undoubtedly affected, at least to some degree, by the general state of affairs in Britain. In 1820, the 'Swing Riots' swept through Wiltshire and other areas of the country. They were the inevitable outcome of the intolerable economic pressures that increasing industrialization was inflicting on agrarian workers. As a result of his compassionate and surprisingly expert management of Lacock, Talbot succeeded in avoiding any local violence. He soon found that in trying to alleviate the condition of the poor, he was out of step with the local farmers. Reform was in the air. Henry Talbot expressed his opinion that representation should be equitable; he stood for Parliament in 1831 as a reformer, but was defeated on this first try. Perhaps his entry into politics was less anomalous than it might at first appear. Talbot's ancestors (and many of his current relatives) had been deeply involved in politics. His cousin Kit, Christopher Rica Mansel Talbot, had entered Parliament in 1830. Lady Elisabeth considered her son's talents wasted in the game of politics, but she would have encouraged him to at least do something, anything, to break out of his ill humor. The passage of the Reform Bill in 1832 increased his constituency, and Henry Talbot was finally elected as a Member from Chippenham. He was to serve only one term.
Talbot's short political career proved unsatisfying but perhaps also convinced him that a life in science was what he wanted. A sure sign that Talbot was getting a grip on his doldrums came with his March 1831 election as a Fellow of the Royal Society. Proposed by his relatives Charles Lemon and William Fox-Strangways, Talbot's election was supported by the testimony of Michael Faraday, Davies Gilbert; George Peacock, Thomas Philipps, and William Whewell (Herschel would have been happy to sign for Talbot as well save for his continuing protest against the Society). Perhaps even this short experience in politics was for the best; Talbot's point of view on the relationship between science and society had been expanded. In 1833, he wrote to the botanist William (later Sir William) Jackson Hooker that "the difficulty which you complain of, of getting any bookseller to publish a scientific work on Botany is not confined to that science." Talbot suggested that there was a societal benefit to be gleaned from support:
Abroad, not only is paper & printing cheaper, but assistance is rendered by the Governments. In my opinion public libraries ought to be established in all our principal Towns at the national expense. A considerable sum should be voted annually for the encouragement of science, which should be in part expended in patronizing literary undertakings of merit. From 20 to 50 copies of such works should be purchased by government & distributed to these provincial libraries, which small at first, would soon become important.
Another sign that Henry Talbot was emerging from melancholy was his December 1832 marriage to Constance Mundy, of Markeaton. The youngest daughter of the MP for Derbyshire, she came from a family possessed of more than the usual artistic interests and was a reasonably accomplished amateur sketcher herself. Constance's cheerful countenance was an important foil to Talbot's occasionally brooding behavior. She was highly supportive of Henry's efforts and proved to be a good moderator for the well-motivated ambitions of Lady Elisabeth. However successful in this, Constance never became the intellectual partner to Henry that Maggie was to John Herschel. The evidence for this is distressingly clear. Since Henry spent much time away from home, his correspondence to and from Constance is voluminous. It is perhaps understandable that she carried on no discourse about her husband's highly-specialized researches. But it is extraordinary that Constance, as an amateur artist herself, never once made any serious assessment of any of Henry's photographs. Constance rarely commented on his subjects and virtually never made suggestions for possible photographs. When Constance was vacationing separately in Wales in 1835, at a time when Henry's interest in his new discovery was at its first peak, she wrote that "I wish you could have taken the outline of the castle & fine elms behind just as I saw them - but I think you told me you could not produce the desired effect by any light except that of the sun." But even this breezy comment was an exception. Her very occasional references to photography after it became public in 1839 indicate that she made some attempts to sensitize paper and make some prints; however, they betray a very superficial understanding of how the process worked. Perhaps Constance Talbot's artistic training exerted some influence on her husband during his picture making, but one is disappointed in efforts to trace hints of such an influence in their correspondence. Constance's understanding of her husband's work, including his photography, was conversational, and her role was to be minimal. It would be an overstatement to call her a photographer.
Completing his article on colored flames in 1826, Talbot had explained to Herschel, that "as I am not acquainted with Dr Brewster perhaps you will have the kindness to transmit it." The relationship between Talbot and Sir David Brewster obviously flourished, for, judging by the surviving correspondence, Brewster and Talbot were in regular contact at least by 1833 and were good friends by 1836. They shared interests in light and perhaps some personality traits as well. Brewster; as influential as he was, struck many as an abrasive personality. He had many admirers but few friends. Talbot was reclusive by nature but the two men hit it off very well. A letter from Constance to Lady Elisabeth written on the occasion of Sir David's visit to Lacock Abbey in August 1836, is particularly revealing of Talbot' character:
You are perfectly right in supposing Sir D. B. to pass his time pleasantly here. He wants nothing beyond the pleasure of conversing with Henry discussing their respective discoveries & various subjects connected with science. I am quite amazed to find that scarcely a momentary pause occurs in their discourse. Henry seems to possess new life - & I feel certain that were he to mix more frequently with his own friends we should never see him droop in the way which now so continually annoys us. I am inclined to think that many of his ailments are nervous - for he certainly does not look ill. I hear from Sir David that he distinguished himself at the Meeting in a conversation on the Improvement of the Telescope. ... When I see the effect produced in Henry by Sir D.B's society I feel most acutely how dull must our ordinary way of life be to a mind like his! - and yet he shuts himself up from choice.
Henry Talbot would never achieve the same levels of fame that Herschel (unwillingly!) reached. His scholarly contributions were generally less influential; much of this stemmed from the fact that he rarely took on the more general questions of the organization and role of science that Herschel approached. Unlike Herschel's minutely detailed and lengthy treatises, Talbot's journal publications tended to be very short. Frequently they reflected scientific concerns common to the day and were just as often as suggestive as they were declarative. Even so, Talbot made two fundamental contributions to science before photography was even announced. These stemmed from a growing awareness in Talbot's time of the interrelationship between forms of energy and matter. Talbot made one very practical contribution that is still in use today. In July 1834, he read a paper to the Royal Society summarizing his experiments on light: "I have lately made this branch of optics a subject of inquiry, and I have found it so rich in beautiful results as entirely to surpass my expectations." In this, Talbot revealed his discovery of the polarizing microscope; his method of placing one polarizer close to the eyepiece and another below the stage is still considered highly effective. It provides an increase in contrast between the subject and the background and is an important tool in the analysis of internal structure. Talbot applied it immediately to studying the structure of crystals.
Whereas the polarizing microscope was a discrete invention of Talbot's, his work in spectral analysis ( the analyzing of the physical makeup of substances through optical means) was more in the nature of fundamental contributions to the beginnings of an important branch of science. As Talbot was to remember in later years:
About the year 1824 or 1825, Dr. Wollaston gave one of his evening parties, to which men of science and amateurs were invited and it was the custom to exhibit scientific novelties, and to make them the subject of conversation.
On the evening in question I brought as my contribution to the meeting some very thin films of glass (such as are shown in glass-houses to visitors by a workman, who blows a portion of melted glass into a large balloon of extreme tenacity, and afterwards crushes the glass to shivers.) Such a film of glass I brought to Dr. Wollaston and his friends, and after showing that in the well-lighted apartment it displayed a uniform appearance without any markings, I removed it into another room, in which I had prepared a spirit lamp, the wick of which had been impregnated with common salt. When viewed by this light, the film of glass appeared covered with broad nearly parallel bands, which were almost black, and might be rudely compared to the skin of a zebra.
Talbot's optical 'zebra skin' in fact provided some of the earliest evidence of the presence of the lines of sodium. These bands had been discovered by Josef von Fraunhofer (the same man who was the cause of Herschel's and Talbot's first meeting). Henry Talbot's first non-mathematical journal article, his 1826 'Experiments on Coloured Flames,' was a foundation stone for spectral analysis. The following year, Herschel incorporated Talbot's findings into his treatise on light. Two types of objects can be analyzed this way. In absorbent subjects (where the effect of incident light is analyzed), important pioneering work was done by Sir David Brewster and others. In the treatment of self-luminous subjects (such as the sun), Talbot and Herschel were the pioneers. Many of their subsequent publications returned to this subject of study.
Henry Talbot, through his substantial family connections and through his own efforts, now exercised a certain amount of political influence. An excellent demonstration of this was his effective support for the establishment of a national botanic garden at Kew. Long known as 'the Royal Vegetable Patch', Kew had declined by 1838 to a most precarious position and was threatened with closure. Talbot's longtime botanical friend, Sir William Jackson Hooker (recently knighted for his botanical contributions), led the fight to establish a national collection. Agreeing that "it would be a pity indeed if Kew Gardens were to be sacrificed to a pitiful & false economy," Talbot persuaded his influential relatives to take up the cause. He met with the Chancellor of the Exchequer, determined that petitions to Parliament from learned societies would have great weight, and proposed to the Council of the Linnean Society that they present a petition to the House of Commons. His idea was accepted unanimously. By May, Talbot could report to Hooker that the Chancellor of the Exchequer "spoke confidently to me last night that something satisfactory would be done about Kew Garden. ..." Although he was leaving town and could be of no further assistance, "I trust that it is put into a right train, & will issue favourably." This support was perhaps Talbot's greatest and most effective contribution to the politics of science. Kew was saved and revitalized that year; in 1841, Hooker became its new director.
When considering the dynamics of the beginnings of photography, to paint Talbot as being junior to Herschel, as is often done, is terribly misleading. No one was like Herschel and only a few even approached his status.
<#FROWN:J61\>Instead of focusing just on how her sympathy feels, she remains interested in understanding what provokes it, what responses it provokes in others, how it reflects childhood experience. Somewhere along the way, however, she finds that she takes great pleasure in contemplating (if rather coolly) the fact of her feeling sympathy when she does; there is something exactly right about it, she finds, that makes it worth appreciating apart from its mere psychological significance or effect on others: like a dancer's gesture or a poet's turn of phrase, it is, somehow, satisfying to behold. Or perhaps the response itself is nothing special; it becomes interesting only through beholding it as Bullough's aesthete beholds the fog at sea: she neither anxiously interrogates it for clues about her psychological health or her moral worth, nor is she so detached that she is indifferent to it altogether. Instead, she has learned how to contemplate, with appreciation and from the 'proper psychical distance,' the mere fact of it. This, too, counts as an aesthetic meta-response.
No doubt, there is something vaguely troubling about routinely aestheticizing one's feelings. But even more troubling is the habit of aestheticizing feelings that play a central role in our moral lives.
IV
Feagin's failure to note the aesthetic character of some of our meta-responses underlies a further important difference between our views. She thinks such responses are a good thing, from a moral point of view, while I am not convinced they are. I am skeptical in part because the aesthetic and moral points of view often conflict; and when they do, the aesthetic often triumphs, if in unexpected and subtle ways.
The extent and obviousness of the conflict vary, as do the ways we resolve it. Formalist purists go far enough when they proclaim, with Clive Bell, that while content may or may not be harmful to aesthetic experience, always it is irrelevant. Formalists of this sort pay attention only to the look or sound or structure of things; moral concerns, like content, are beside the point. But one can go further. Consider the Futurists, whose sensibility echoes in confused and disturbing ways through contemporary culture. One Futurist goal seems to be to reject morality altogether by aestheticizing (rather than ignoring) just those aspects of a situation which would ordinarily be most appalling:
We will glorify war ... militarism, patriotism, the destructive gesture of anarchist, the beautiful Ideas which kill .... Art can only be violence, cruelty, and injustice.
Of course, most of us are even less likely to be Futurists than formalist purists. If we would feel uneasy ignoring moral content for the sake of aesthetic satisfaction, we would be outraged at the glorification of immorality, in life or art. But compromises, especially confusing ones, can be struck.
One sort of compromise involves a kind of response that, unlike the formalist's, attends to morally compelling content, but not for the Futurist purpose of rejecting morality. This seems a step in the right direction; paying attention to a work's morally compelling content seems to be an improvement, from a moral point of view, over ignoring it. But, as we have seen, Danto worries that when we confront a work of art, we respond aesthetically even to morally compelling content. And this is wrong, for Danto, because representations of human misery demand moral rather than aesthetic response. Indeed, one might even argue that aestheticizing what we should respond to morally may be worse than ignoring it as the formalist does.
I do not doubt that the sort of response Danto describes sometimes occurs. But just as Bell's formalism fails as a descriptive theory by ignoring audiences' widespread concern with content, Danto's view fails to recognize how fictional portrayals of human misfortune may engage most people's moral responses. Danto seems to imply that even if one writes a play for the sake of making a moral difference, the project is doomed from the start; injustice, as soon as it becomes the subject of the play, is put "at a distance which is exactly wrong from a moral perspective." However, if my analysis is correct, we may, after all, have our moral cake and the aesthetic pleasure of eating it, too: We respond to the injustice morally rather than aesthetically, but then diminish the moral significance of this first-order moral response by aestheticizing it.
Something like this may be going on with Alice Walker's character. That the sight of the black woman is for her a weepy miracle suggests not the austere ecstasy that some formalists talk about, but a pleasure that is born, however inappropriately, of moral sentiments, in particular those of a white liberal woman confronting a poor black woman. Here, though, the character knows it is wrong to treat black people as art, as a means to those weepy but enjoyable sentiments. So the character falls somewhat short of a full-blown aesthetic meta-response: her guilt keeps her from fully enjoying her sensations, let alone from contemplating how fine they are.
V
So what exactly is wrong with aesthetic meta-responses to tragedy? To answer this question, we must look more closely still at what counts as such a meta-response, and we must talk about the moral significance of sympathy itself. We must also consider whether such responses, even if inappropriate in 'real life', are unobjectionable in our experience of fiction. The best way into this labyrinth is through a more thorough look at Feagin's view.
That Feagin doesn't consider the possibility of aesthetic meta-responses may make it easier for her to sing their moral praises. Given the tension between the aesthetic and the moral points of view, it may be easier to exonerate morally, if not celebrate, the pleasures we take in tragedy if we do not understand these pleasures to be aesthetic ones. Indeed, Feagin avoids the traditional notion of the aesthetic altogether. When she talks about aesthetic response and aesthetic pleasure, she is simply using 'aesthetic' as a synonym for 'artistic'; she is not using the language of aesthetic-response theorists. At one point she even implies that differentiating between direct and meta-responses to art may allow us to avoid talk about aesthetic response altogether. Nevertheless, some of what Feagin says about these meta-responses seems to indicate that she has aesthetic ones in mind. This, we shall see, makes it easier to question their moral credentials.
Why does Feagin find meta-response to tragedy praiseworthy? First, she claims that the sympathy we enjoy when we respond to tragedy also underlies our capacity for moral action. But does this establish its special moral status? Here I think we must inquire further into the grounds of our pleasure. One might say, "I am glad I can sympathize; sympathy helps me to be a better person, to care about others, to do what needs to be done. Helping others is important, and sympathy is part of this, if not a means to this end." Insofar as we sincerely give this sort of explanation, our meta-response is a moral rather than an aesthetic one. Furthermore, it seems more natural here to talk about valuing or appreciating one's capacity to be moved by the suffering of others than about taking pleasure in or enjoying the feeling of being moved. The latter sort of language, which suggests that we savor a feeling in abstraction from a larger moral context, as we would the look of a painting or the sound of a symphony, seems to be a paradigmatic case of the sort of quasi-perceptual aesthetic meta-response I discuss above.
To be fair, I should note that Feagin does once use the word 'satisfaction.' But elsewhere her emphasis is different. Consider how she explains the moral difference between taking pleasure in one's reaction to fictional suffering and taking pleasure in one's reaction to the real thing. When we sympathize with merely fictional suffering, Feagin points out, our enjoyment of our reaction does not have the price of other's pain:
In real life, the importance of human compassion is easily overshadowed by the pain of human suffering. It is not possible in real life to respond to the importance of human sympathy as a distinct phenomenon, since that sympathy depends on, one might even say 'feeds on,' human misery. It is not, in life, an unequivocal good. In art, however, one experiences real sympathy without there having been real suffering, and this is why it is appropriate to feel pleasure at our sympathetic responses to a work of art, whereas it is not appropriate to feel pleasure at our sympathetic responses in reality. There the sympathy comes at too great a cost.
This may be true enough. But here, again, the enjoyment of our sympathy seems to be an end in itself, far removed from its moral context - especially any connection to the practical action sympathy often motivates in the real world. Indeed, given that Feagin is talking about the moral significance of our meta-response, it is odd that she doesn't say more here about the connection between sympathy and action, if only to strengthen her case by pointing out that we need not worry about rescuing a fictional victim. On the other hand, if a concern with acting, when we can and as we should, does not somehow figure into our meta-response, how morally significant can it be?
Another sign that Feagin has in mind an aesthetic meta-response is her admission that this pleasurable meta-response would be morally inappropriate in a real situation. But unless she means an aesthetic meta-response, there would be no moral impropriety; after all, if our meta-response reflects only a concern with doing the right thing, there would be nothing wrong in saying, for example, "I'm glad I could sympathize with her; it made her feel less alone," or even "I'm glad that after all I've been through I can still feel for people." We need not regard another's pain as a means to a self-congratulatory, narcissistic end; rather, through our feelings we may come to understand and ease another's pain and, perhaps, become better people. But this means just that not all meta-responses to sympathy are aesthetic ones, and that the moral (as opposed to the aesthetic) significance of sympathy does not rest in its being valued as an end in itself.
Determining the moral status of sympathy is, of course, an old and thorny problem. J.S. Mill complained that some of his contemporaries regarded the sympathies as "necessary evils, required for keeping men's actions benevolent and compassionate." Such a view certainly is too harsh. But surely the moral significance of sympathy emerges as part of a larger moral picture, including a commitment to understanding those with whom one sympathizes and to acting when appropriate. Without such connections, the feeling of being at one with humankind seems, from a moral point of view, neither here nor there. If a world of purely rational Kantian agents, completely dutiful and completely unfeeling, is a grim prospect, a world of sensitive souls so involved in sympathy that they are distracted from or uninterested in understanding and acting on the situation at hand may be even worse.
Indeed, in the nineteenth century, obsession with sympathy drew criticism on just these grounds. Historian Walter Houghton puts it this way: the cult of benevolence, typified by Dickens and Eliot, "took a new direction in the nineteenth century when the misery of the industrial workers became sufficiently apparent to demand redress - and all the more so because it constituted a threat to social order. If one solution proposed [by Carlyle and Arnold] was a more earnest sense of social duty, another lay in quickening the moral sensibility to an acute sympathy for suffering humanity." Such a sensibility, however, may care more for sentiment than for understanding and eliminating the conditions that give rise to it. Houghton sees this tendency culminating in the aestheticism of Walter Pater, for whom morality was "all sympathy," contemplation the end of life, and the job of the artist to disengage our thoughts from life and fix them, "with appropriate emotions, 'on the great and universal passions of men.'"
<#FROWN:J62\>
Nevinson's Elegy: Paths of Glory
CHARLES E. DOHERTY
Christopher Richard Wynne Nevinson (1889-1946) was among the first British artists to witness the horrors of the First World War. A volunteer ambulance driver on the western front, Nevinson observed the suffering and carnage resulting from the new trench warfare in the winter of 1914-15. He returned to the front in 1917 as a member of the British government's official war-artists' program. For the program Nevinson produced a controversial painting, Paths of Glory (fig. 1), that was banned from his one-person show at the Leicester Galleries, London, in March 1918. The exhibition took place at a time during the war when morale seemed at its lowest, and the government supposedly censored the painting because it portrayed dead British soldiers. Since other contemporaneous paintings and photographs of Allied and enemy war dead were shown in London galleries and reproduced in the British press, the reason for the censorship of Nevinson's Paths of Glory demands further investigation.
Nevinson was the son of two eminent figures in British society, the Manchester Guardian war correspondent Henry Woodd Nevison (1856-1914) and the suffragist writer Margaret Wynne Nevinson (1860?-1932). The progressive Nevinson household became involved in Futurist activities during F.T. Marinetti's prewar visits to London. The younger Nevinson joined forces with Marinetti to write the only English Futurist manifesto, 'Vital English Art,' published in the Observer on June 7, 1914. Although stopping short of the warmongering rhetoric of earlier Italian manifestos, the English proclamation, like its Italian counterparts, urged that society be cleansed of her ills, her tired platitudes, and long-worshiped traditions and extolled the virtues of sport, adventure , and the heroic instinct of discovery.
A quest for adventure may have drawn Nevinson to the Friend's Ambulance Service, a group of Quaker volunteers who risked their lives tending the injured and dying in Flanders. Following his arrival in Dunkirk on November 13, 1914, he observed death on a daily basis; on one occasion his ambulance was blown up. His personal experience of the grisly results of aerial bombardment and automatic weaponry proved advantageous in producing early, innovative, modern images of the war.
When the fatigued Nevinson returned to London on January 30, 1915, suffering from rheumatic fever, one newspaper reporter stated that the Futurist artist was a victim of neurasthenia, or shell shock. Nevinson denied the accusation, calling the war "a violent incentive to Futurism" and proclaiming, in a quotation borrowed from Italian Futurist manifestos: "There is no beauty except in strife, no master-piece without aggressiveness." H. W. Nevinson's journal entry of October 25, 1914, however, indicates that his son was interested, even prior to visiting the front, in distancing himself from the Futurist movement.
Following a period of convalescence, Nevinson enlisted in the Royal Army Medical Corps and served as an orderly in a London military hospital. He was invalided out in late 1915, due to recurring rheumatic fever. During 1915 and 1916, he produced Cubofuturist portrayals of fighting men, exploding shells, and destroyed landscape. One early image, Returning to the Trenches (fig. 2), of 1914?-15, captures the buoyant mood of marching French soldiers, equipped with kits and bayonets. Their feet barely touch the ground as they move in synchronized formation toward the front.
This oil on canvas, along with Nevinson's other representations of marching men, is sometimes associated with Vorticism, the short-lived British abstract-art movement. Although a woodcut version of it, titled On the Way to the Trenches, was reproduced in the second issue of Wyndham Lewis's Vorticist publication Blast, the painted version, in its emphasis on simultaneity, lines of force, and fractured planes of colour, derives its abstract qualities more from Futurism than Vorticism.
Expressing none of the action and vitality of war, his La Patrie (fig. 3) of 1916 anticipates the grim portrayal of war dead in Paths of Glory. Both are among the most moving British paintings of the First World War. Set in the poorly lit wooden shed that served as a casualty clearing station in the Dunkirk railyard, La Patrie depicts row upon row of casualties lying on stretchers on a straw-covered floor. The clearing station was known as 'the Shambles,' and Nevinson later described the blood, stench, typhoid, and agony in which six workers attempted to rescue more than three thousand maimed and dying men.
Despite its portrayal of grimacing faces, bloody bandages, and squalid conditions resembling those of a temporary morgue, La Patrie presented no problem for the British censoring authority. Not only was its exhibition permitted twice in 1916 (in June and September-October), it was praised by critics, reproduced in newspapers, and included in a 1917 volume of Nevinson's war images. One reviewer hailed the artist for his non-flag-waving, non-drum-beating depiction of war, while the influential P. G. Konody stated that "C. R. W. Nevinson stands alone, in England, as the painter of modern war." Thus, an image portraying the pain and suffering of Allied soldiers, which not only implied the possibility of death but questioned the invincibility of the British forces, was sanctioned by the British government in 1917.
Under the direction of the journalist and politician Charles F. G. Masterman and, later, the novelist John Buchan, the government hired artists to produce eyewitness accounts of the war after June 1916. These artists retained ownership of their work, while the Department of Information (known as the Ministry of Information after February 1918) had right of first refusal to purchase and use the art to illustrate propaganda literature. Nevinson, who departed for France and Belgium as an official war artist on July 5, 1917, was the first of the young avant-garde artists to be hired for the innovative program. He toured the British line until mid-August, returned to London, and produced more than seventy-five paintings, drawings, and prints over the following seven months.
The controversy surrounding Paths of Glory began on November 24, 1917, when Masterman asked Alfred Yockney, the former editor of the London Art Journal, who served as an advisor within the Department of Information, to show him two of Nevinson's paintings. One of these Masterman called 'Dead Men.' Five days later, the department's censor, Major A. N. Lee, voiced an objection to the one portraying dead British soldiers. Although it would not yield "military information to the enemy," he wrote to Yockney, its "subject matter raises a point of policy." Given a probable concern for public morale, he deemed it necessary to consult the War Office for their opinion, and refused to permit the work to be released.
The elder Nevinson accompanied his son to the government offices on December 4 for a meeting with Buchan, the director of information. They wished to learn more about Lee's reason for suppressing the work. They learned that two Nevinson paintings, Paths of Glory and an oil titled A Group of Soldiers (fig. 4), were considered hindrances to the war effort. The latter, a painting of four Tommies standing at ease, was suppressed because of its purportedly unflattering representation of British soldiers, who were thought to resemble mannequins or ventiloquists' dummies. On this painting Lee's decision was later overruled by Masterman, who claimed that paintings should only be censored "from a military point of view," rather than an aesthetic one. A Group of Soldiers was approved for exhibition and reproduction on December 13, 1917, and for inclusion in Nevinson's one-person show the following March, but Paths of Glory was not.
The artist did not fare as well with Lee regarding this painting. Although no details of the Nevinsons' meeting with Buchan survive, most likely they were told what Yockney was reporting to Masterman, that "representations of the dead have an ill-effect at home" and that all such paintings were "now rigidly suppressed." The painting was purchased by the government in January 1918, perhaps as a means of acquiring full control over its use and reproduction.
The brown-and-green oil on canvas depicts two dead British soldiers sprawled on a hillside littered with rifles, helmets, and barbed wire. The soldiers appear, compositionally, to recede from the lower left to the upper right of the canvas. Unlike Returning to the Trenches, in which military might is suggested by an infinitely long line of marching men, Paths of Glory portrays the human cost of war, with a recessive line implying an untold number of dead beyond the limits of the canvas. A mood of morbidity and death is heightened through use of an eerie, unnatural light that casts aqua-colored shadows across the dead soldiers and the debris that surrounds them. The small zone of background sky is transected by a skeletal web of posts and barbed wire, and appears only at the very top of the composition, conveying a sense of restriction, rather than liberation.
Scenes of inglorious death in the trenches resembling that in Nevinson's painting were commonplace along the front line that stretched from the North Sea coast of Belgium to Switzerland. Burial in a prepared grave was often an impossibility: bodies were frequently left where they fell, sometimes used as shields for trench reinforcement, as gun supports, or as guideposts. The living became accustomed to the sight of trenches "rotthen with dead," who looked "like ghastly dolls, grotesque and undignified," as the poet and soldier Siegfried Sassoon described them. Nevertheless, the British government preferred that the public at home be shielded from knowledge of such images.
Paths of Glory was painted in November and December 1917, during the blackest period of Nevinson's war years, when the government was calling up discharged soldiers and previously reflected men to replenish the drastically depleted ranks at the front. Among those under review were men suffering from neurasthenia and shell shock. Entries in H. W. Nevinson's journals of the winter of 1917-18 provide proof that his son was depressed, extremely fearful of returning to the front, and a patient of the eminent British neurologist Sir Henry Head. The younger Nevinson's depression during these dark days was exacerbated by his desire to succeed in the war-artists' program, which he viewed as his only means of avoiding conscription. He had failed in an earlier attempt to emigrate from Great Britain after Parliament began debating the issue of reexamining discharged soldiers. Therefore, the idea that the Futurist sympathizer and former member of the rebel avant-garde deliberately chose to paint an alarming or potentially controversial work of art for political or personal reasons during these traumatic and difficult months is not consonant with his profound wish to remain in the good graces of government officials at the time. His other official war art, in a more representational and stylistically conservative mode than his 1914-16 innovations, also bears this out.
Nevionson's shift toward a more representational style may have been independent of his commitment to the government-sponsored art program. After the trauma and illness following his service at the front, he appears to have distanced himself from the Futurist rhetoric of 1914-15, regarded by some as having contributed to the war's madness. Moreover, this shift is in keeping with a similar move on the Continent toward more legible and orderly stylistic tendencies during the mid and late teens.
A crowd of dignitaries and luminaries from the military and art worlds, as well as invited members of society, gathered at the Leicester Galleries on March 1 to see Nevinson's exhibition and to hear an opening speech by the newly appointed minister of information, Lord Beaverbrook. Their surprise at and intrigue with one painting in a semidisguised state was recorded in the press: despite its proscription Nevinson had hung Paths of Glory and placed a piece of brown paper diagonally across it, covering the dead soldiers. Upon the paper he had written in bold letters "CENSORED."
The press photographed and reproduced the painting in its altered state (fig. 5). One reviewer concluded: "Probably no picture in the Nevinson show excites more interest and speculation than the one which is partly obscured by the 'censored' label." The issue was raised in the London Mail's column, 'Things We Want to Know,' which asked, "What is hidden by the patch?"
<#FROWN:J63\>Nonetheless, both ruler and ruled always benefit in some way - intentionally, accidentally, or indirectly - because they share some common task or purpose (Pol 1254a27-28).
That ruling and being ruled are according to nature does not mean that either is easy. What is according to nature appears to be divine insofar as it appears to be in the best state possible; but it is not "sent by the gods," or the same as fortune, because it requires effort on our part (NE 1099b9-24). Indeed, Aristotle observes, "in general, it is difficult to live together and be partners in any human activity" (Pol 1263a15-16). This observation seems to move Aristotle's notion of the household toward Arendt's interpretation - that the household is a place of toil yielding no real satisfaction. According to Aristotle, however, things brought into being through effort - nature's or man's - are the greatest and noblest of all things (NE 1099b22-24). They thus yield much pleasure, for "actions in accordance with virtue are by nature always pleasant" (1099b13-14). Furthermore, the difficulty of living together decreases to the extent that the parties recognize their common aim, a life as complete and self-sufficient as possible (Pol 1280b33-35, 1260b13, 1254a27-28).
THE AIM OF HOUSEHOLD RULE: VIRTUOUS INDIVIDUALS
In that the best household's aim is to instill unqualified moral virtue through some sort of rule, its aim appears to be indistinguishable from that of the best regime. Moreover, the aims of the best household and the best regime are alike in that they both seek to acknowledge the distinctiveness of individual human beings; according to Aristotle, diversity more than sameness gives rise to unity (Pol 1261a29-30, 22-24). Both the household and the city should promote similarity in the sense of virtue, but neither should promote homogeneity (1263b31-32). "Habits" deriving from household activities and "laws" from the regime can together make the city "one and common through education" (1263b36-40) without sacrificing diversity. Nonetheless, as noted earlier, household activities are better suited to individualized instruction and thus to acknowledgment of individuality than is public education. Cities, then, should rely more on households than on laws and public, institutions to maintain diversified excellence. The question is, what should household rule instill to achieve this diversity?
According to Aristotle, instilling moderation and judgment makes human beings virtuous without eradicating any distinctiveness other than a lack of virtue. The man and the woman of the household may exercise both moderation and judgment as well as "show who they really and inexchangeably are" by selecting and remaining with each other, managing the household, and caring for their children. Likewise, children and servants may also acquire and demonstrate moderation, judgment or understanding, and distinctiveness by the ways they conduct themselves and respond to the heads of the household. Indeed, the extent to which members of the household practice moderation and judgment is itself expressive of distinctiveness.
TEACHING MODERATION
All household members must learn to be moderate toward things and each other. The various forms of household rule can teach members moderation by revealing to them the natural ends of their natural desires (Pol 1257b19-34). For example, household management (rule over the material conditions of a household) teaches that specific things must fulfill specific needs and desires: food satiates hunger, a bed satisfies the need for sleep; money itself cannot satisfy such needs. Thus, household management teaches human beings to check their desire for money - itself an unnatural, because unfulfillable, desire. The various household relationships also teach moderation in various ways. Forming a household entails the exercise of moderation in that it requires limiting oneself to one out of many sexual partners and companions. Parenthood teaches both the parents and the children moderation. Since children's reasoning powers are not developed, parents must find the mean between arguments and force which is effective for teaching their children (Pol 1260a13-14, b6-7, 1332b10-11; NE 1179b23-29). It is because children are potentially reasoning and reasonable beings - or "free persons" - that one ought to rule them in "kingly fashion" (Pol 1259a39-b1, 1253b4, 1285b32). And children, who are not inclined to be moderate, must learn to be so if they are to live well (NE 1179b24-34). Finally, as the next chapter shows, ruling slaves teaches both the masters and the slaves moderation.
Aristotle's characterization of the ideal household as requiring the exercise of moderation contrasts with the general contemporary liberal view according to which what goes on in the household is entirely a matter for the (undefined) discretion of household members. Indeed, activities are private according to Aristotle only when the actors heed the limits established by nature.
The moderation learned in the household not only helps to sustain the household but facilitates all human engagement. Moderation is both the result of and fosters seeing what is required for living together. It is thus neither a strictly private nor a strictly public virtue, and so it - not courage - might be said to be in Aristotle's eyes the political virtue par excellence.
TEACHING JUDGMENT
In addition to moderation, the good household teaches judgment (Pol 1253a15-18). Forming a household requires judgment in that it requires choosing a good partner. Raising children involves judgment as something to be taught. Ruling servants involves judgment in trying to compensate for the servants' lack of it. What is pertinent to this inquiry, however, are the ways judgment required by the household differs from that required by the regime. One significant difference involves natural affection; another, the end each aims to realize.
According to Aristotle, the end of the city is justice, which all take to be "some sort of equality" - that is, equal things for equal persons (Pol 1282b14-21). But this definition encompasses both natural justice, the fundamental principle of which is proportionality or desert, and conventional justice, the fundamental principle of which is arithmetical equality (NE 1134a26-28, b18-19). The regime that is "by nature" - realized natural justice - is best (NE 1135a5). But since realizing natural justice in a regime presupposes many deserving human beings and the ability to detect them - that is, requires fortune and virtue to achieve (Pol 1331b21-22, 1277a1-5) - cities should aim first to realize conventional justice.
Should the household also then seek conventional or ordinary justice? In two places, Aristotle says that it should not. "Political justice seems to consist in equality and parity," "but there does not seem to be any justice between a son and his father, or a servant and his master - any more than one can speak of justice between my foot and me, or may hand, and so on for each of my limbs. For a son is, as it were, a part of his father" (MM 1194b23, 5-15). As he explains in the Nicomachean Ethics, "there can be no injustice in the unqualified sense toward what is one's own, and a chattel or a child until it reaches a certain age ... is, as it were, a part of oneself, and no one decides to harm himself. Hence there can be no injustice toward them, and therefore nothing unjust or just in the political sense. ... what is just in households ... is different from what is politically just" (1134b10-17).
By proceeding immediately to discuss natural justice, Aristotle suggests that it characterizes the household. The household appears to be even a paragon of natural justice in that inequalities within it are evident and determine who rules and who is ruled. And, as Arlene W. Saxonhouse explains, "the family, because its differences in eid<*_>unch<*/> are observable, demonstrates a unity in diversity which perhaps becomes impossible in political life. In the polis obvious differences in eid<*_>unch<*/> are absent. ... The family with its definition of differences ... attains a certainty in nature not available to the city." Or, at least, not available to most cities. In other words, it appears that the household, being a model of natural justice, is a kind of model for the best regime. Aristotle would apparently like the natural superiority holding together the (best) household to hold together the (best) city. Indeed, he may insist on the preservation of households (in all regimes) because they have the potential to exemplify perfect unity or justice and by their examples point the city toward a higher justice.
Aiming to realize natural, not conventional, justice, the good household ruler does not treat all members equally or give each a turn at ruling; rather, it is incumbent on this ruler to detect the virtues of each member and treat him or her accordingly, giving guidance or instruction when needed and freedom to make choices when deserved. The household is a compound of "unlike persons" - man, woman, servants, and children - who, moreover, have multiple functions or obligations - as husband and father, wife and household manager, son or daughter and future citizen (Pol 1277a5-8, 1253a4-14). There are thus not only manly virtues, womanly virtues, servile virtues (1277b20-23), and presumably even youthful virtues but also virtues attached to being a husband, father, wife, and child. A household thrives when each member performs his or her function, or upholds his or her obligations, in accordance with the virtues proper to doing so (NE 1098a14-15).
The variety of virtues indicates the variety of judgment in the household. Most notably, the judgment of those ruling differs from that of those being ruled, as becomes clear when we take into account the deliberative capacities of each kind of member and Aristotle's distinctions among intellectual virtues in the Nicomachean Ethics. One acquires prudence by repeatedly putting into effect good judgments about at least one's own affairs, if not the affairs of others (NE 1141b12-21, 29-1142a10). Lacking experience, the young cannot have prudence (1142a15-16). Lacking good judgment, or the ability to detect through deliberation what action to perform, and how and when to perform it, the slavish, who lack the ability to deliberate, cannot have prudence either (NE 1143a29-31, Pol 1260a12). The nonslavish adults of the household, however, having both experience and the ability to deliberate (Pol 1260a10-13), may have prudence. In fact, household management requires that they do (NE 1141b31-32). Nonetheless, the prudence of the man and the woman apparently differ. Although it is the responsibility of both to manage the household, the man should acquire possessions and the woman should oversee their use and consumption (Pol 1277b24-25). It follows that the man should acquire the household servants (Pol 1255b37-39), since they are animate possessions (1253b32), and that the woman should command them, since their function is to assist in the use of other possessions (1252b32-33, 1254a2). Moreover, Aristotle indicates in several ways that the man, at least more than the woman, should guide their children; for example, "the man rules the child" (Pol 1260a10). In addition, Aristotle assigns marital rule to both the husband and the wife; that is, spouses rule each other (Pol 1253b9-10, 1259a39-b1, 4-10). Since the man and the woman each rule over others, at least in part for the good of those others (Pol 1278b32-1279a8), each has complete moral virtue, which Aristotle calls justice and prudence (NE 1130a2-14, 1145a1-2; Pol 1260a17-18, 1277b25-26). But because each rules over different persons, they again exercise prudence differently (Pol 1260a10-12, 20-24, 1277b20-23).
In contrast to the judgment of the free adult members of the household, the judgment of children and servants is lacking. Children have only the potential for judgment and prudence; servants can only follow judgment and comply with prudence (Pol 1260a12-14, 1254b22-23).
Variety of judgment appears naturally in the household; even more, in the good household, those who rule acknowledge it. Good household rulers do not command their spouse, children, and servants in the same way (NE 1134b15-16). By way of presenting the household, then, Aristotle suggests that private judgment differs from the judgment required by most regimes in that it acknowledges differences in kinds of, and aptitude for, virtue among human beings. Moreover, in trying to promote the virtues peculiar to each member, household rulers promote individuality.
In addition to promoting individuality, private differs from public judgment in not having law to aid it (Pol 1282b1-6). Both political and household rulers must employ "knowledge and choice" (Pol 1332a31-32) to bring about, respectively, the city's and the household's excellence.
<#FROWN:J64\>
Complementation, re-presentation, reworking of strategic elements - all of these are present in the finale. The subset center C is insistently thrust into attention by the wrong-footing opening, and at crucial articulative points the movement also reworks the Neapolitan. In fact, the finale shows remarkable new interpretative roles for C, which connect it closely in pitch or function to F-natural. The second subject of the finale, for example, has a strong Neapolitan element. In the exposition, the second subject (mm. 70-78) is in the key of B minor, so the Neapolitan pitch for that key is C. This means that C, the subset pitch of the first movement and the wrong-footing opening of the finale, now assumes a new role as the Neapolitan of B minor (Ex. 18).
Ingenuity is demonstrated on many fronts: the second subject in E minor - a recapitulatory feature - comes back in the development (m. 216), not in the recapitulation. This placement indicates not so much an exchange of function between development and recapitulation, but the interpolation of a recapitulation element in the development. In a movement where surprise and ingenuity are of the essence, this is one further instance of a technical virtuosity, with Beethoven reordering the elements of sonata and rondo design in an individual solution of dazzling skill. In the E-minor appearance of the second subject, the Neapolitan pitch is F-natural. The second subject is extended by a strongly marked four-measure phrase, dynamically highlighting F major and its dominant C. Since C major is also the Neapolitan of B minor, one could say this passage encapsulates both uses of the Neapolitan, as well as the Neapolitan and its dominant (Ex. 19).
<O_>figures&captions<O/>
The most stunning use of the Neapolitan, though, comes in the coda, where it is reintegrated into E minor. Like the hammered-out dominant minor ninth in the first movement, this Neapolitan is the dramatic and structural hinge of the finale. It reworks precisely elements from the scherzo - strong dynamics and sudden dynamic change, wide registral span - but more powerfully. The extreme range of the fortissimo Neapolitan is cut off by a shocking measure's silence, then the V of E minor comes in softly. Silence, as well as the pitch and harmonic elements from the first movement, is another element in this powerful return (Ex. 20). This searing Neapolitan figure and its resolution in E minor fulfills all the functions of the Neapolitans from the beginning of the work - except one. The coda has one last card to play. In the last page of the score, one final reference to the Neapolitan appears as part of an ascending chromatic line in the first violin from E to E'. The movement finishes with three pairs of perfect cadences: the first has F-sharp - G in the upper line, so raising the F-natural to its diatonic F-sharp and resolving it to G; the second has D-sharp - E in the upper line, reversing the E - D-sharp of the opening i-V<sp_>6<sp/> at the beginning of the work; the third pair of cadences falls a fifth, B - E, reversing the space-opening gesture of the first movement. The circle is now closed.
<O_>figure&caption<O/>
Drama, conflict, design - these features characterize the works of Beethoven's middle period that fall within the description of 'heroic'. To say that Mozart's elegant gracefulness or Haydn's witty equipage have vanished in the face of a more powerful musical conception sounds like a truism, but truisms are rarely the only truth. The description 'heroic' sits oddly on the shoulders of the Fourth and Pastoral Symphonies, the Piano Sonata in F, op. 54, or the C-major Razoumovsky Quartet, op. 59, no. 3, works smaller in frame , more relaxed and graceful in style than their bigger-boned neighbors. But drama and conflict, the characterizing features of the heroic concept, are central to the Eroica and Fifth Symphonies, the E minor Razoumovsky Quartet, the Waldstein and Appassionata Piano Sonatas.
The Appassionata Sonata
The nature, or rather the status, of conflict is treated differently in the Appassionata Sonata than in the other middle-period works just cited. In the Eroica and the E-minor Razoumovsky Quartet, first-subject elements are lightened and partially transformed in the finale. This reworking in turn affects the formal design of the finale where the underlying framework - respectively variation and sonata rondo - is shaped into a highly individual rescoring of the form. The Waldstein Sonata releases its first-movement intensity in the finale by a wonderfully expansive rondo. Only in the Apassionata is conflict unresolved. Conflict in the first movement is taken up in the finale and replayed in a correspondingly intense mode, but while the first-movement tonal opposition of contrasted keys is resolved, the discourse of conflict is not.
At the beginning of the first movement, two distinct elements are presented: First, the linearization of the tonic-triad F minor is followed directly by the flat supertonic G-flat major, forming the harmonic juncture i-<sp_>b<sp/>II. The second element is the modern C - D-natural - C, which is a variant of the movement's prime mordent, C - D-flat - C. Even more closely than the Eroica or the E-minor Razoumovsky Quartet, the finale of the Appassionata is a direct reworking of first-movement elements in a movement of parallel intensity, rather than a reinterpretative reworking in a lighter vein - an intentional matching of mode and material that underscores the cyclic nature of the work. Not only will first-movement elements, particularly the Neapolitan and the prime mordent, return prominently in the finale, but there will also be close parallels of formal placement and pacing between the two movements.
The developments of the two movements vividly illustrate these parallels. In both instances, conflict is central to the dramatic action of the development, with the Neapolitan its most powerful agent for engendering tension and deflecting tonal direction. In the first-movement development, the Neapolitan pitch, G-flat, appears in the context of B-flat minor at m. 115. By a rising bass motion, V-VI, the Neapolitan is then tonally reinforced in G-flat major. Immediately, it is respelled enharmonically as F-sharp, and for the first time in the movement, it rises to G in the key of C. The cumulative energy of this intensification erupts into sweeping diminished sevenths and climaxes onto hammered fortissimo D-flats - C, the latter segment of the movement's prime mordent, now magnified at its structural dominant. (Compare this emphatic segment with its bleached-bones version at the end of the recapitulation).Prolonged through the sweeping line of diminished sevenths and the structural dominant. G resolves to F only at the recapitulation (Ex. 21))
The central action of the development may be seen, therefore, as the conflict between the chromatic <sp_>b<sp/>II and the diatonic supertonic. If this description is reminiscent of the first movement of the Razoumovsky, op. 59, no. 2, what differentiates the first movement of the Appassionata is the way Beethoven locks the chromatic and diatonic supertonics in conflict, rather than allowing the diatonic supertonic to form a point of relaxation as part of the dominant harmony of F minor, so that conflict between the two supertonics is polarized, but not resolved.
<O_>figure&caption<O/>
In the finale development, the Neapolitan G-flat is again at the center of conflict against G-natural. Just as in the first-movement development, G-flat appears in the context of B-flat minor, but here replays the work's other prime element, the mordent, on the Neapolitan pitch in a sharply condensed focus of the two main structural components of the outer movements. (See Exx.Ex. 22, 23.)
<O_>figures&captions<O/>
Striking parallels between the outer movements of the Appassionata demonstrated by matching formal placement of salient harmonic features and the contextual reworking of the Neapolitan. Far beyond general similarities of mood and expressive delineation, these precise parallels make viable the description of the work as cyclic form in which the finale provides a complete variant reworking of the first movement. In variation form, the variation - normatively, in classical style - exhibits the same key planning, phrase structure, harmonic progressions, and large-scale formal sections as the theme, while elaborating texture and changing register, articulation, and dynamics. The variant concept also opens up a larger dimension of formal integration for the sonata as a whole; just as prime first-movement elements are reworked (in the manner of development and variant) in the finale, so the principal elements of the slow movement are elaborated (but without development) in a set of variations. While individual variations in the slow movement are self-contained, the movement overall is not, but the end of the slow movement leads into the finale in a way comparable to the slow-movement link to the finale of the F-major Razoumovsky Quartet (F-minor, op. 57, and F-major, op. 59, no. 1; the closeness of compositional period will make such parallels of compositional procedure understandable). Yet the two transitions are different in the way the two movements are connected: the slow movement of the quartet dissolves out in a swirl of elaborate figuration that fines down onto the trill initiating the finale. At the end of the slow movement of the Appassionata, the opening theme returns, but its resolution is blurred by two hazy diminished sevenths. These same chords, hammered out thirteen times fortissimo at the beginning of the finale, shatter the serene repose and consonance of the variation movement. Dynamic rupture, though, is anchored by pitch continuity. D-flat, the pitch and key center of the slow movement, is carried through to the diminished sevenths and falls to C on the turning sixteenth-note figure which heralds the thematic statement of the finale. This juncture, D-flat - C, forming a hinge between the two movements, takes up, at a larger level, the mordent intersection D-flat - C, which is the structural hinge of the first movement.
Just as with development sections, so the recapitulation and finale coda both re-present and compress strategic elements from the first movement. The coda presents the final clash of the supertonic and wrenches it back into the constraints of F minor.
Some of the techniques discussed here with reference to important works in Beethoven's middle period were not in themselves necessarily new, nor had they originated with him. For example, the favored harmonic shift of unmediated I-ii (although not <sp_>flat<sp/>II) which Beethoven used virtually as his own style characteristic for strongly defined first-movement openings (in the Appasssionata, and the String Quartets, op. 59, no. 2, and op. 95, as examples) had been used by Haydn at the beginning of the first movement and also at the beginning of the second movement - scherzando - of his C-major String Quartet, op. 33, no. 3. Expanded dimensions and contrapuntal enrichment of development can be found in the first movement of Mozart's C-Major String Quintet, K. 515, in the first movements and finales of his G-Minor Symphony, K.550, and Jupiter Symphony, K. 551. What is new in Beethoven's middle period is the consistently expanded formal framework as distinct from individual instances; the versatility and daring with which he rearranges the modules of formal entities, yet retains their underlying coherence and sense of internal logic; and the use of strategic pitches as axial centers around which a movement is built and from which it diverges by contextual reinterpretation into new and surprising key areas to form, frequently, the structural hinge of the movement.
In the middle-period works, Beethoven differentiated the four movements of the quartet and symphony by sharply defined characterization, yet, conversely, drew closer relationships between first movement and finale, binding them by similar or related mood, and even more, by the reinterpretation of prime first-movement elements in the finale. This compositional procedure produces a kind of cyclic form., but one conceptually different from Schubert's Wanderer Fantasy or from later nineteenth-century works lie Liszt's B-minor Sonata. Beethoven's expansion of dimensions, of dynamic and tempo ranges, has as its context the retention of formal design and structural principles. The strong individualization of movements is set against the concern for the organic form of the work overall. Accordingly, large-scale integration of the outer movements depends for its effect on the internal formal autonomy of each movement and its specific delineation of character and material - in the relationship of self-standing parts to a larger whole.
<#FROWN:J65\>
5
Conclusions
SUE BRIDEHEAD: THE CASE FOR A FEMINIST READING
In his depiction of Sue, Hardy shows remarkable sensitivity to feminist issues. The novel's tragedy turns on marriage, and it is a double tragedy. This view is augmented by looking into the historical context of women's issues.
In discussing Jude in the context of Hardy's fiction as a whole, Patricia Ingham (1989) observes that over the course of his novel writing Hardy's treatment of women increasingly diverges from the traditional misogynist stereotype which had been 'scientifically' justified by Herbert Spencer in his popular The Study of Sociology (1873).
Owing to the far-reaching influence of Darwin, Spencer's discourse was scientific rather than moral. Darwinism, with On the Origin of Species (1859), but more particularly with The Descent of Man (1871), gave momentum to biological determinism as it related to the female nature and role. By appealing to so-called objective laws, and adopting a tone of neutral and dispassionate observation, science sought to establish the biological link between physiology, psychology, and sociology, and to effect the ratification of the status quo - that is, to confirm the old stereotypes, and to reaffirm the disabilities of being a woman. This spurious science replaced the Bible as the underpinning of the double standard of sexual morality. Social critics used concepts of evolution to show that sexual difference was the result of adaptation to the conditions necessary for social survival. Woman's position in society was seen as the natural result of processes designed to strengthen her essential function - maternity. Spencer claimed that women had less power than men for abstract thinking because their vital energies went toward nurturing offspring. As a result, women lagged behind men in the evolutionary process, having smaller brains as well as weaker physiques. As the weaker sex, Spencer argued, women had learned to disguise their feelings, to please and persuade, and to delight in submission. Because Victorian women were not only dependent but ready to cultivate and display that dependence, a husband was their only goal. Although in the Victorian period women were often revered as being morally superior - more devout and devoted to caring for others - some writers were of the opinion that woman's reproductive capacity gave her a far more menacing nature. Havelock Ellis, in Man and Woman (1894), stated that since menstruation is disgusting, women are ashamed of it, and shame makes them deceitful. This tendency toward dishonesty is, he asserted, reinforced by the duties of maternity, and much of the education of the young, which is entrusted to women, consists of skillful lying.
Childbearing, one of the few acceptable activities for women in Victorian society, dominated women's lives. In 1900 a quarter of all married women in England were pregnant. Most deliveries took place at home, where the experience could be nothing but a struggle with poverty, pain, and death. Stillbirths, miscarriages, attempts at abortion, uncaring doctors, and incompetent midwives caused women to fear pregnancy. The infant death rate in 1900 was 163 per thousand, compared to 9.4 per thousand in 1985. In all, 145,000 infants died in 1900. The medical profession did not explain about or provide contraception or abortion. The condom and vaginal sponge were unreliable and in any case unavailable to most people because birth control was relatively costly. Among the working classes there was a flourishing trade in abortion-inducing pills.
Employment for women outside the home was effectively limited to 'women's work' - work that required nimble fingers or no great physical strength, for example, dressmaking, schoolteaching, bar keeping, assisting in a shop, doing office work, or working in domestic service. The telephone, typewriter, and bicycle widened career possibilities for women in the 1890s, but professional opportunities did not exist for women until after the First World War. Therefore the ideal woman was supported by her husband and had no independent legal existence.
The movement to change this state of affairs came to the fore in the last two decades of the nineteenth century, a period during which organized labor was pressing for social and political emancipation. Throughout Europe mass socialist and working-class parties were organizing and demanding fundamental changes. A group of middle-class feminists sponsored debates and campaigned for legislation giving women access to the professions, secondary and higher education, the right to own property, and the right to vote. They crusaded against adult prostitution, the flourishing trade in child prostitution, and the protection of 'innocence' that made sexuality furtive and dismal.
By the 1890s the woman question was being widely debated in newspapers, journals, and novels. J.S. Mill had argued in The Subjection of Women (1869) that the so-called disabilities of women were maintained in order to make women servants of men who feared the competition of women in the working place, and who could not tolerate living with them as equals. The most popular writer on the woman question, Grant Allen, portrayed marriage as a degrading form of slavery. In his novel, The Woman Who Did (1895), the heroine deliberately has a child with her lover whom she refuses to marry, regarding herself as a moral pioneer doomed to martyrdom.
But the whole weight of social orthodoxy brought to bear on maintaining the stereotype was deeply ingrained in the majority of women as well as men, and the women's movement was not able to reach a consensus on most of the feminist issues raised in the 1880s and 1890s. Some writers argued that promiscuity was the path to self-fulfillment; others asserted that such freedom could only come from celibacy. Another point of contention was childbearing. Was it a woman's most sacred calling? Or was it rather an aspect of her degradation? Could fulfillment come from working in the man's world? Or was the man's world a trap for women?
In spite of much debate on the fundamental place of women in society, many women maintained that there was no escape from their established role. In June 1889 more than one hundred well-known women signed their names to the 'Appeal against Female Suffrage,' which was printed in a leading journal. The appeal stated that women's direct participation in politics "is made impossible either by the disabilities of sex, or by strong formations of custom and habit resting ultimately upon physical difference, against which it is useless to contend." Among those who endorsed the appeal were Beatrice Webb, Mrs. Humphry Ward, Eliza Lynn Linton, Mrs. Matthew Arnold, and Mrs. Leslie Stephen; a supplementary list of two thousand names was added two months later.
A woman might decide to escape the life marked out for her by "the inexorable laws of nature" - that is, the controlling and channelling of her sexuality into marriage - by refusing the sexual dimension of a relationship. According to Penny Boumelha, Sue Bridehead elected this option. Boumelha contends that Sue's situation is confused and confusing because Sue is not sure whether she wants love without sex or sex but not marriage. Yet one must not see her as lacking sexual feeling, Boumelha argues; rather, Sue's actions should be seen as her response to the dilemma of how to have love without "the penalty."
According to Boumelha, whether Sue denies her sexuality or risks pregnancy, she is reduced. Boumelha says that the tragedy is not brought on by her frigidity but by motherhood: "It is motherhood - her own humiliation by the respectable wives who hound her and Jude from their work, Little Father Time's taunting by his schoolmates - that convinces her that 'the world and its ways have a certain worth,' and so begins her collapse into 'enslavement to forms'" (Boumelha, 148). Hardy, Boumelha observes, is alone among writers of stature in drawing attention to motherhood's role in confining women within the nuclear family. Sue's sexuality destroys her, whereas Arabella's, by contrast, helps her survive. Both reject their husbands, take up with other men, sublimate their sexuality into religiosity, and eventually return to their husbands. Yet a crucial point that emerges from the ironic paralleling of Arabella's life with Sue's is that Arabella never plays a maternal role. Whereas Arabella is identified with sexuality and fecundity - she barters her sexuality for security, seducing Jude by flinging a pig's penis at him and pretending to hatch an egg between her breasts - Sue assumes the role of mother - to her own children as well as to Arabella's son. Thus, according to Boumelha, Hardy understands that a woman's freedom depends on remaining free of the maternal role.
Ingham (1989) argues that Hardy gradually developed his sensitivity to the woman question over the course of 25 years of novel writing. She traces the emergence of metaphors in which workingmen suffering from self-devaluation are compared to women, demonstrating that Hardy ever more insistently subverts the social ideal that a woman's self-fulfillment is rooted in self-denial. By the time Hardy wrote Jude, the workingman and woman are "two in one," twins who suffer a similar oppression. This affinity in oppression, Ingham says, is highlighted by the emphasis on Jude's and Sue's similarities. Their being cousins on Jude's mother's side, children of a family doomed by a hereditary curse, points up a sameness that is continuously stressed: by the rhyming circumstance of each taking refuge in the other's room, by Sue's actually appearing in Jude's clothes as a kind of double, and most emphatically by Phillotson's view of them as the lovers in Shelley's 'The Revolt of Islam' - transcendent beings, martyred in the cause against tyranny.
Jude comes to see himself and Sue as martyred pioneers. They are yanked "back into pre-determined forms of marriage," Boumelha notes, and this is the tragedy (Boumelha, 150). Conventional notions of sanctity and free will are exploded by the novel: neither the home nor the love relationship is a protected zone, and individual acts and intentions cannot reform society.
Boumelha claims that Hardy understood the crucial importance to women of socialized child care, that it was his expressed reason for supporting female suffrage. In his 1906 letter to the Fawcett Society he states that he hopes suffrage would tend "to break up the present pernicious conventions in respect to manners, customs, religion, illegitimacy, the stereotyped household (that it must be the unit of society), the father of a woman's child (that it is anybody's business but the woman's own), except in cases of disease or insanity" (Letters 3:238). The view that unless the institution of marriage is radically changed, women will continue to be enslaved, is also expressed in the novel. Phillotson tells Gillingham, "And yet, I don't see why the woman and the children should not be the unit without the man" (4:4). When faced with the possibility that Little Father Time may not be his child, Jude makes a statement that gives Hardy's position a sharper focus: "The beggarly question of parentage - what is it, after all? What does it matter, when you come to think of it, whether a child is yours by blood or not? All the little ones of our time are collectively the children of us adults of the time, and entitled to our general care" (5:3).
I do not see Jude as a novel primarily about marriage, nor do I think of it as "the Sue story" (Boumelha, 138), as Hardy called it in a letter to Florence Henniker. Yet, Boumelha's and Ingham's arguments offer persuasive interpretations that illuminate Sue's life.
Sue's behavior is confused and confusing, which is to say her behavior indicates her self-control. Marrying Jude would invite oppression, yet loving him without marriage would invite the penalties reserved for sinners. Ambivalent, she appears to be coquettish, half inviting his advances, yet sidestepping them. "[E]picene tenderness," "boyish as a Ganymedes" (2:4) are what the love-sick man sees in her external behavior, but he has no clue about what is at issue. Sue wants to be loved, but she cannot bear to lose her freedom.
SHAME
Hardy's alternation of scenes - one on a comic plane, the other tragic; one lofty, the other low; one affirming, the other repudiating - is the novelist's method for grasping the ambiguous real world.
<#FROWN:J66\>New York, however, was the magnet that drew artists of all sorts. Space was still cheap in lower Manhattan. America's far-flung universities had little interest in recruiting experimentalists-in-residence, and the one serious exception to this rule, Black Mountain College, folded in the fifties, sending much of its teaching staff and student body to New York as well. Bohemian Manhattan was an intimate, small-scale scene: a band of outsiders easily recognizable by their dress and demeanor. Groups that later would seem diametrically opposed or at least very different - for example, the Beats and the 'New York School' of poetry - rubbed elbows amiably and frequented the same bars and jazz clubs. Being few in number, they were obliged to stick together; in Eisenhower's blandly conformist America, all weirdos were brothers until the opposite was proven. In addition, artists shared an exhilaration born of their recent liberation from Europe. The old American colonial complex - a sense of being on the periphery of things, still strong among the modernists of the 1920s - had been swept away by the triumph of abstract expressionism, by William Carlos Williams's appropriation of American speech as a basis for new poetry, and, of course, by jazz, the American art form par excellence.
Needless to say, not all artists frequented bars and jazz joints, but a remarkable number did. The abstract expressionists - Jackson Pollock, Willem de Kooning, Franz Kline, and others - were hard drinkers and inveterate hangers-out. One of them, Larry Rivers, was also an accomplished jazz saxophonist and served as a point of intersection between the worlds of painting and jazz. The Beats also spent a lot of time in night spots. For the New York School (Kenneth Koch, John Ashbery, and others), a key connection with jazz was Frank O'Hara, whose best-known poem, a poignantly oblique homage to Billie Holiday, is entitled "The Day Lady Died." Judith Malina's and Julian Beck's Living Theatre, with its mixture of raw psychodrama and dreamy pacifism, sponsored poetry readings at its headquarters on 14th Street and featured Jackie McLean and hard-bop pianist Freddie Redd in its production of Jack Gelber's play The Connection.
Almost anyone's account of the era includes both this heady mixture of scenes and the centrality of jazz as an artistic model and jazz clubs as meeting places. Ron Sukenick's Down and In: Life in the Underground, a combined study and memoir covering the period from 1945 through the eighties, describes the clientele and atmosphere at the Five Spot in the late fifties: "If the painting seemed more consciously American after 1950, the uniquely native American art form, jazz, became, through the fifties, more central than ever for underground artist of all kinds. It came together at the Five Spot, a bar on Cooper Square where the brothers Iggie and Joe Termini hosted a basically flophouse clientele until the artist started coming in during the mid-fifties. Painters like Grace Hartigan, Al Leslie; David Smith, de Kooning became habitus. Larry Rivers, the painter, played jazz there, poets read poetry to jazz, and avant-garde film makers even showed their films to jazz. Writers like Kerouac, Frank O'Hara, and Kenneth Koch moved in, and finally the great jazzmen themselves came down to play - Charlie Mingus, Sonny Rollins, Cecil Taylor, Thelonious Monk, Ornette Coleman."
In response to a question about which jazz musicians down-town artists and intellectuals were friendliest with, the painter Emilio Cruz (who also writes poetry and plays jazz drums) gave this description of the scene: "In the case of Jackie McLean, I would think that if he was around the Living Theatre crowd, he was friendly with Judith Malina and Julian Beck and that crowd, and a number of interesting people in that company. Also Cubby Selby; we all called him Cubby. I never knew his real name was Hubert Selby, Jr., until his book [Last Exit to Brooklyn] came out. Paul Blackburn, the poet, who was a good friend of mine, knew a lot of those people. Bob Thompson the painter, Allen Ginsberg at times, Bob Kaufman the poet, some of the Black Mountain poets. Rollins lived downtown and knew a number of artists, though he was very solitary.
"Others who were there not just because it was cool but because they had a deep interest in the music were [Amiri] Baraka and Larry Rivers. [Rivers] was a friend of Zoot Sims, Stan Getz I know he considers himself as serious a saxophone player as he does a painter. During that period I lived on Jeffferson Street, down by the river. Pepper Adams lived in that building underneath me. In fact, he was the only one in the building who had heat. He had a gas heater, so in the winter when he was on the road I was a lot colder than when he was there. Donald Byrd used to come by a lot because they had that band together at that time. Jefferson Street was south of East Broadway, maybe southeast, not that far from where the old Fulton Fish Market was. That's all changed now. I don't even know whether that street exists anymore. Ed Blackwell used to come by that building too.
"That was a period, in my life, when a lot of things were integrated. Jazz was integrated within artists' lives. A lot of people lived in lofts, and oftentimes the musicians might not have lived in lofts themselves, but they would come over there (at least on the Lower East Side) to play, to rehearse a band, so there were a lot of connections because of that. That connection would then extend what a person's capacity was, so that one person might learn more about music, another person might learn more about painting or poetry, and I knew a number of musicians who were interested in learning about all of it. Like I'm friends with Grachan Moncur III. I was his drummer last year at his workshop in New Jersey. He used to come visit my studio all the time. There was a lot of openness between various people. Herbie Lewis, the bass player, he lived around the corner. He was a neighbor of mine, so we spent a lot of time together. Miles ... I know numbers of people from the beat generation who were friends with Miles, like Allen Ginsberg, Jack Kerouac, Robert Frank the photographer. So there was an integration of lots of the arts. The beat generation spent much of their lives in clubs.
"Now one of the things that it's very important to understand is that there weren't thousands of people involved in the arts in those days, so when you would walk down the street in New York City, you would look at somebody and you would recognize them instantly as an artist, and you would immediately find that you had some kind of rapport, and there's another thing. I can't speak for right now, but there was youth. Youth reaches out. There's one thing that was consistent in all the artists in New York City at that time. They were reaching out for new things, new ways of expressing themselves. They were attempting to discover new values. Those values were not necessarily ones that were supported by the society at large, so there was this that they had in common. Another thing was that there was an attempt to break down racism, and politically there was a sense of hope that America could arrive at a higher moral state. So racism was something that in that world was frowned upon."
What did jazzmean to those experimental artists who took it most seriously? Part of the answer, as Sukenick indicates, has to do with its American qualities, also underlined by Hettie Jones, author of Big Star Fallin' Mama, a study of black female vocalists from Bessie Smith to Aretha Franklin: "I think jazz was the music they [downtown bohemian types] felt closest to, the way someone feels close to music that's part of the zeitgeist. It was American. There's that whole idea that abstract expressionist art was the first truly American art movement; and those people saw themselves as an avant-garde in what they were trying to do. It was a shared feeling that they were all part of a changing American art scene."
Jazz was also influential in its improvisational freedom and structural openness, as Allen Ginsberg indicates in his description of jazz's relationship with his poem 'Howl': "In the dedication of 'Howl' I said 'spontaneous bop prosody.' And the ideal, for Kerouac, and for John Clellon Holmes and for me also, was the legend of Lester Young playing through something like sixty-nine to seventy choruses of 'Lady Be Good,' you know, mounting and mounting and building and building more and more intelligence into the improvisation as chorus after chorus went on ... riding on chorus after chorus and building and building so it was a sort of ecstatic orgasmic expostulation of music. So there was the idea of chorus after chorus building to a climax, which was the notion of part one of 'Howl,' with each verse being like a little saxophone obbligato or a little saxophone chorus, as though what I was doing was combining the long line of Christopher Smart, the eighteenth-century poet, with notions of the repeated jazz or blues chorus, till it comes to a climax, probably in the verse 'ah, Carl, while you are not safe I am not safe.' And then there's a sort of a coda from then on."
Ginsberg also claims jazz as an important model for his work and that of his contemporaries: "The whole point of modern poetry, dance, improvisation, performance, prose even, music, was the element of improvisation and spontaneity and open form, or even a fixed form improvisation on that form, like say you have a blues chorus and you have spontaneous improvisations, so in 'Howl' or 'Kaddish' or any of the poems that have a listeny style, 'who did this, who did that, who did this,' you start out striking a note, 'who,' and then you improvise, and that's the basic form of the list poem or, in anaphora, when you return to the margins in the same phrase, 'Or ever the golden bowl be broken or the silver cord be loosed or the pitcher be broken at the fountain,' as in the Bible or as in some of Walt Whitman's catalogues or in Christopher Smart's 'Rejoice in the Lamb' poem or the surrealist example of Andr Breton's free union, 'my wife with the platypus's egg, my wife with the eyes of this, my wife with that and that ...'
"It [jazz] was a model for the dadaists and it was a model for the surrealists and it was a model for Kerouac and a model for me and a model for almost everybody, in the sense that it was partly a model and partly a parallel experiment in free form. The development of poetics, as well as jazz and painting, seems to be chronologically parallel, which is to say you have fixed form, which then evolves toward more free form where you get let loose from this specific repeated rhythm and improvise the rhythms even, where you don't have a fixed rhythm, as in bebop the drum became more of a soloist in it too. So you find that in painting, the early de Koonings have a motif or a theme, the woman or something like that, but it gets more and more open, less dependent on the theme, and in poetry, where you have less and less dependence on the original motifs and more and more John Ashberyesque improvisational free form flowing without even a subject matter, though I always kept a subject matter like the old funky blues myself. It was partly a parallel development within each discipline: painting, poetry, music. There were innovators who opened up the thing after Einstein, so to speak - you know, relative measure, as Williams said - which is in a sense something that happened with bebop: not the fixed measure but a relative measure.
<#FROWN:J67\>
One indication of poetry's place in the current construction of African American critical theory is revealing. In The Signifying Monkey, Henry Louis Gates, Jr., proposes a compelling way of tracing the lines of continuity among major African American texts. With consummate skill, he teaches us to read this literature through the paradigms of vernacular culture as a way of chronicling a common African American struggle for authority and voice. Taking off from Roland Barthes, Gates defines Hurston's Their Eyes Were Watching God as the first instance in our written tradition of a "speakerly" text - that is, a text "oriented toward imitating one of the numerous forms of oral narration to be found in classical Afro-American vernacular literature" (181). Gates's characteristic reinvention of literary theory, grounded in the language and rituals of African American culture, is insightful, yet this particular moment in his revisionary strategy raises an important concern. By slighting poetry in his newly constructed canon, Gates neglects such "speakerly" nineteenth-century texts as Paul Laurence Dunbar's seminal poem 'An Ante-bellum Sermon' (1895). Elsewhere, Gates has written instructively about Dunbar and African American poetry. His early essay 'Dis and Dat: Dialect and the Descent' remains crucial to our understanding of the conventions of 'dialect' and vernacular poetry. My point here is simply that his exclusion of poetry from The Signifying Monkey, given the ensuing authority of that work, reenacts what has become a familiar pattern in African American literary history. The effect of such omissions in our most influential readings of African American literature may well serve to lessen the perspective of poetry in current revisions of black critical theory and the African American canon.
In his essay 'Performing Blackness,' Kimberly Benston surmises that it may be the "relative susceptibility" of African American poetry "to discrete analyses [that] resists theoretical efforts that move toward totalization, toward recuperation and ideological closure" (184). He confronts the risks involved in theorizing about the lines of continuity between disparate poetic works only to underscore the necessity of defining the shared themes, literary forms, and rhetorical strategies that constitute a distinctive poetic tradition. In particular, he entreats us to recognize that seemingly discrete, fragmented expression - enacted as a mode of performance - characterizes African American discourse. Black utterance, like other counterdiscourses, masks its own continuity as a way of veiling its challenges to the legitimacy of dominant political institutions and cultural traditions. In theorizing about African American poetry, critics might be guided by what Richard Terdiman describes, in a different context, as the "capacity [of counterdiscourses] to situate: to relativize the authority and stability of a dominant system of utterances which cannot even countenance their existence," as a way of enacting a new realm of subjectivity (15-16). When black writers turn to African American vernacular performance, they call into question the authority of the literary conventions and racial ideologies of the dominant society. Critics should neglect neither the insistence of African American poets on challenging the authority of dominant American institutions nor the individual black poet's struggle for recuperation and wholeness. Their individual texts make up an ongoing tradition of black poetic subjectivity.
While the term performance can be applied variously to a range of cultural and literary phenomena, I use it to designate verbal performance viewed as a cultural event. In revising the boundaries of what has traditionally constituted the folklore text, recent scholars, especially Roger D. Abrahams and Richard Bauman, have reconceptualized the distinctions between textual representations and what Robert Georges calls "complex communicative events" (313). Such scholars often associate the term vernacular with the modern concept of folklore as an intricate interaction between performer and audience that relies on linguistic, paralinguistic, kinesic, and thoroughly contextual codes and conventions. This implied notion of performance has become a model for how to discuss formal literary texts. In written texts that draw on the aesthetics of vernacular performance, the relations of orality and literacy are continuous. The tensions between repetition and improvisation that operate in a verbal performance are translated into competing structures of creation and recollection for literary artists and their audiences.
Construing performance in this way, I argue that Paul Laurence Dunbar's early poem 'An Ante-bellum Sermon' is an instructive example of an African American "preacherly text." The poetic heritage to which this poem belongs stretches back at least to the eighteenth- and nineteenth-century sermons of African American religious leaders, even as it looks forward to the preacherly performances of James Weldon Johnson, Langston Hughes, Robert Hayden, and Gwendolyn Brooks. In scaling what Hughes calls the "racial mountain," Dunbar's vernacular performance illustrates the solid ground of African American poetic traditions.
I
For black poets of the nineteenth and twentieth centuries, the challenge has been to establish the grounds of their literate and poetic authority without sacrificing the distinctiveness of their experiences in the New World. These poets have often relieved the disruptions resulting from different aspects of themselves by locating their work in a continuum of African American expressive culture. In their search for poetic voice, they have struggled to define their experiences as African Americans through tropes of indigenous and communal subjectivity. Many of them, including Dunbar, Hughes, Hayden, and Brooks, have made vernacular sermonic performance of their heritage a primary site of cultural authority and artistic creativity. The narratives, rhetorical strategies, and rituals of performance of the vernacular sermon have helped shape the recurring aesthetic and ideological tendencies of African American poetry. The relations of preachers and their congregations have provided these poets with a model for what Barbara Bowen has called "untroubled voice": the performance of perfect continuity between artist and audience. By performing these sermons, black poets have mended the divisions in their artistic voices and the contradictory expectations of their various audiences, thereby recovering for themselves and their communities the privileges of vernacular eloquence.
To explicate the cultural status of the vernacular sermon, I begin with the historians of the American slave - including John Blassingame, Eugene Genovese, Lawrence Levine, and Al Raboteau - who have drawn on 'folklore,' interviews, autobiographies, and various other kinds of texts to suggest that slaves were able to survive their oppression by evolving sustaining forms of cultural expression. The early black sermonic performance was one of the rituals that defined for slaves and free African Americans their participation in a unique religious fellowship. According to the available evidence, slave preachers were esteemed as exemplary figures in the black community, whether they preached in church, in the quarters, or in the fields, whether they preached out of the Bible or from the depths of their hearts. While many slaves were required to listen to white preachers, some of whom were well educated and well trained, they clearly preferred to hear black preachers deliver messages from God. As Nancy Williams, and ex-slave from Virginia, put it: "Dat ole white preachin' wasn't nothin'. Ole white preachers used to talk wid dey tongues widdout sayin' nothin' but Jesus told us slaves to talk wid our hearts" (Yetman 13).
The slave preachers did not necessarily derive their cultural authority from their ability to read or their facility in speaking standard English. Often the important issue was whether they could perform sermons that moved their congregations toward freedom. One former slave, Clara Young, told her interviewer that it did not matter that her favorite preacher, Mathew Ewing, was illiterate; what she cared about most in judging his sermons was his style and his message. "He never learned no readin' and writin'," she reminisced, "but he sure knowed his Bible and would hold his hand out and make like he was readin' and preach de purtiest preachin' you ever heard" (Yetman 335). She must have been impressed with his wit in reenacting the gestures of white preachers who read Scripture to their congregations, but not as much as she was stirred by his eloquence. Judging from his vast experience with slave congregations, the white minister Charles C. Jones could say that they were natural judges of a "good sermon," even though such a sermon might corrupt Christian theology and the English language (14-15). While recognizing that the slaves preferred to hear black preachers, he was not willing to admit that these preachers were evolving their own style of address. Jones championed the use of religion among slaves as a means of social control without ever realizing that slaves would use their religion as a way of fighting oppression.
Many black congregations expected their preachers to be vigorous, dramatic, and instructive in language, theme, and gesture. The textual accuracy of their preachers' readings of the Bible may not have been a consideration, but the dynamics of performance were essential. Even if the preachers had to create the biblical references in their sermons, their congregations demanded that they breathe life into those creation. In her Letters from New York, Lydia Maria Child, the well-known abolitionist, describes an especially dramatic sermon that Rev. Julia Pell, a black itinerant preacher, delivered to a Methodist congregation in 1841. Although the regular minister of this church felt compelled to apologize to the congregation for Pell's misquotations of Scripture, her performance must have been remarkable. Even the skeptical Lydia Child admitted that this "dusky priestess of eloquence" made her shout and cry with religious fervor (67). Child offers a rare early description of the rhythmic pacing of a black sermon, noting that Pell "began with great moderation, gradually rising in her tones, until she arrived at the shouting pitch ... [that] she sustained for an incredible time, without taking breath, and with a huskiness of effort." This rhythm of performance, and the physical effort it required, undoubtedly reinforced the effect of the sermon's thematic climax and encouraged the enthusiastic response of the congregation. The section of this climax that Child records can be transcribed as lines of poetry:
Silence in Heaven!
The Lord said to Gabriel,
bid all the angels keep silence.
Go up into the third heavens,
and tell the archangels to hush their golden harps.
Let the sea stop its roaring,
and the earth be still.
What's the matter now?
Why, man has sinned,
and who shall save him?
Let there be silence,
while God makes search for a Messiah.
Go down to the earth;
make haste, Gabriel,
and inquire if any there are worthy;
make haste, Gabriel;
and Gabriel returned and said,
No, not one.
But don't be discouraged.
Don't be discouraged, fellow sinners.
God arose in his majesty,
and he pointed to his own right hand,
and said to Gabriel,
Behold the Lion of the tribe of Judah;
he alone is worthy.
He shall redeem my people. (65)
While the members of the church congregation could not have found this conversation in their Bibles, they would have understood that the invented narrative testified to the possibility of redemption. Although Child cannot record the sound of the sermon, one can discern a pattern of rising and falling intonations and perceive the shrillness of Pell's dynamic "shouting pitch." As the black preachers tell us at the beginning of their sermons, they preach as God's instruments. Here Pell literally creates music with her voice. She provides the congregation with proof that she is one of "God's trombones."
Pell could neither read nor write, Child reports, yet this preacher's 'illiteracy' does not limit her talents as an artist. Without 'book learning,' she nonetheless reads and performs the cultural conventions of the black vernacular sermon. While societies define and use literacy differently, "literacy is always connected with power" (Pattison viii). The sermon quoted above demonstrates that Pell is literate in the rituals of African American culture, and it measures the extent to which she exploits her power as a prophet. Unable to read the Bible, Pell still knows what sermons her congregation needs to hear. Such knowledge is communal. It is shared by preacher and audience.
Knowing the tacit rules of performance, many of the blacks in Pell's audience would have recognized her skill in drawing on the conventions of delivering a vernacular sermon: the conspicuous patterns of repetition; the Old Testament imagery and diction; the allusions to the spirituals; the preacher's freedom to assume the identity of a biblical character; her dwelling on the inadequacies of language in the face of divine revelation; her own version of a biblical character's conversion, narrated through the devices of vernacular storytelling.
<#FROWN:J68\>These people are observers who contribute to the composite picture of Gary Gilmore, but they also help Mailer achieve the broad social panorama he admires in writers as different as Tolstoy and Dreiser. Indeed, Mailer has chided himself for doing so little with the secondary characters in his previous novels, a 'flaw' he hoped to correct in The Executioner's Song. Here Mailer develops virtually every 'minor' character and permits each to speak in something like his or her own voice, however much the several idioms blend into the flat, colloquial style for which the book is famous. Mailer's defense of his unadorned prose might apply to the minor characters themselves: "one's style is only a tool to use on a dig." Like the style by which we know them, the secondary characters are supposed to contribute to the book's larger formal ends.
One such end is to 'examine' the American reality exposed by the strange saga of Gary Gilmore. Joan Didion sees Mailer as capturing two crucial features of western America. The first is "that emptiness at the center of the Western experience, a nihilism antithetical not only to literature but to most other forms of human endeavor." The second is an inability to direct our own lives, a failing so pervasive that all the characters seem to share in "a fatalistic drift, a tension, an overwhelming and passive rush toward the inevitable events that will end in Gary Gilmore's death." I believe that Didion's insights are exaggerated, but they do point up suggestive connections between Gilmore and the people who surround him. Bessie Gilmore, Brenda Nicol, Vern Damico, Kathryne Baker and her daughters Nicole and April - all are 'trapped' in their futile efforts to find a life worth living. Indeed, almost every woman in the book first marries at fifteen or sixteen and eventually marries at least three or four times, and the men seem equally caught up in the fatalistic drift Didion notices. Didion does not do justice to the admirable stability of people like Brenda Nicol and Vern Damico, but the wasted lives of those around Gilmore suggest that his own fate is only an exaggerated instance of that moral emptiness Didion hears in the book's western voices.
In this respect as in others, Nicole Baker is the second most important character. Mailer has called her "a bona fide American heroine," but most readers will think she is rather the quintessential American victim. Promiscuous at eleven, institutionalized at thirteen, married at fourteen and again at fifteen, Nicole suffers three broken marriages before she is twenty. "Sex had never been new to Nicole," we are told (143), and it is more than plausible when she runs off with an older man because "she didn't care where she was going" (117-18). Yet Nicole has virtues to match her troubling irresponsibility. As Gilmore sees, she is fearless and fiercely loyal. These are the very qualities that Gilmore counts on when he manipulates her toward a suicide pact. In his many letters from jail, he pleads with Nicole not to make love with other men (350), to give up sex altogether (403-04), and to join him on the other side in death (472). At the end of book 1, he leads her toward a double suicide attempt that epitomizes both his romanticism and his selfishness, even as it climaxes Mailer's portrait of Nicole as an endearing victim. Later Nicole will be denied the 'clean' resolution of death, will emerge from yet another institution to tell Larry Schiller (and Mailer) the story of her love for Gary Gilmore, and will finally drift off to Oregon to new lovers if not a new life. Nicole's story is a familiar one among her family and friends: years of acute aimlessness followed by an utterly hopeless commitment. Surely it is no accident that Nicole comes to love Gilmore most fiercely when he is cut off from her forever. For the Nicoles of the world (and perhaps this means for all of us), there is no consummation except in an imagined future.
The stories of Nicole and the other witnesses point to one of Mailer's most crucial decisions in structuring book 1. Rather than trace Gilmore's grim history from reform school through his term in Marion, Illinois, Mailer chooses to focus on Gilmore's last months in Provo in 1976. The reasons for this no doubt include Mailer's desire to achieve greater dramatic unity and to emphasize Gilmore's 'mystery' instead of the familiar stages of American crime and punishment. But another important reason is to allow Mailer to flesh out the human context in which Gilmore plays his final role or sings his final song, as the title would have it. This context is dominated by the same hateful 'habits' that take more spectacular forms in Gilmore. Yet the human resources displayed in book 1 should not be dismissed quite so easily as Didion's formulation would suggest. Here we get example after example of human folly, western style, but also many instances of what Mailer calls "American virtue," the American's dogged determination to do his or her best in the worst of circumstances. The range of such portraits is really quite extraordinary, from Gilmore's mother, Bessie, to Brenda Nicol, to the Damicos, to the irrepressible Nicole. One of the earliest reviewers called The Executioner's Song "a remarkably compassionate work," and the truth in this judgment should remind us that, like Mailer's portrait of Gilmore, book 1 is structured to highlight the human frailties as well as the abominations of American life.
It might seem that book 2 offers a less sympathetic, more satirical history of Gilmore's last months. The very title of part 1, "In the Reign of Good King Boaz," signals a new kind of irony. Here lawyers and the press are omnipresent and one eighty-two-page section, 'Exclusive Rights,' is devoted to virtually nothing but Larry Schiller's and David Susskind's efforts to corner the Gilmore market, so to speak, by securing exclusive rights to his story. Packs of reporters are everywhere, confirming Mailer's worst fears about press. The many lawyers introduced are often distinguished by one bizarre detail or another, as when Earl Dorius, Utah's assistant attorney general, is excited at the prospect of an execution and proceeds to work himself into a near breakdown to ensure that the state of Utah gets its execution on 17 January 1977 (500), or when Dennis Boaz, Gilmore's second lawyer, supports his client's desire to be executed until it occurs to him that Gary would prefer to live if he could have connubial visits from Nicole (590-91), perhaps in Mexico (611)! Gilmore's final lawyers, Bob Moody and Ron Stanger, are a good deal less eccentric, but they too partake in the grim legal struggle in which the state of Utah pursues its pound of flesh, and the ACLU and other liberal groups fight stubbornly to save a man who does not want to be saved. The ironies here are obvious and may even seem undramatic. In the film version of The Executioner's Song (1982), scenarist Mailer and director Schiller chose to leave out most of the materials of book 2, as if they were less relevant than the more 'immediate' events of book 1.
My own view is that book 2 is at least as interesting as book 1, a remarkable feat when one considers that the protagonist is all but unavailable and the heroine is locked up throughout. Once again Mailer gets great mileage from his so-called minor figures, a few of whom (e.g., Boaz, Schiller, Barry Farrell) are among his most memorable characters. Of real interest for their own sake, they also provide perspective on Gilmore. For example, Gary's brother Mikal is at first reluctant to allow his brother to die and participates in legal actions to prevent it. When he finally talks with Gary, however, Mikal is won over by his brother's seriousness and depth of feeling. As they part, Gary first kisses Mikal, then utters perhaps the most haunting words in this very long book: "See you in the darkness" (840). A cellmate of Gilmore's named Gibbs also effectively testifies on Gary's behalf. A police informer, Gibbs refers to Gilmore as the most courageous convict he has ever been (759). And Gilmore's relatives, especially Vern Damico and Toni Gurney, find themselves moving ever closer to Gilmore as he approaches death. Toni's relationship with Gilmore is especially moving. She first visits him the day before he is to be executed and is overwhelmed by his gentle affection (874-75). Later that day, after her own birthday party, she returns to the party Gilmore has been permitted at the prison and again experiences Gary's new warmth (884-86). Toni is sufficiently moved to try to attend Gary's execution (929). This sequence blends with many other small but affecting moments to verify the change in Gilmore that is sensed by many people during his final weeks.
Mailer uses Barry Farrell and Larry Schiller to temper the more sentimental implications of book 2, but ultimately these veteran journalists also testify to Gilmore's surprising depth. The title of book 2, "Eastern Voices," seems to refer to all those safely established in the social system, whether in the East or the West: lawyers, reporters, producers, assistant attorney generals, and so on. Farrell and Schiller are such voices. Each brings a heavy load of urban skepticism to the Gilmore assignment, hating Salt Lake City, as Farrell does, and believing there is no 'center' to this story, nothing of real human resonance (577). When both men come to see Gilmore in a very different light, Mailer is able to bring his book to a genuine climax.
Farrell is at first confident that nothing sets Gilmore apart but his willingness to die. If Gilmore is not executed, Farrell suggests, he will become indistinguishable from the hundreds of others condemned to die but never executed (611). As he works with Gilmore's responses to hundreds of questions, however, Farrell notices that Gilmore "was now setting out to present the particular view of himself he wanted people to keep" (711). Later Farrell responds profoundly to Gilmore's tapes: "Barry was crying and laughing and felt half triumphant that the man could talk with such clarity" (804). Farrell still believes that Gilmore "had a total contempt for life" (805), but this makes it all the more impressive when Gilmore responds so "humanely" to the massive attention of his last months (805). Farrell is stunned at Gilmore's apparent complexity. In the transcripts Farrell spots "twenty-seven poses," twenty-seven different Gilmores ("racist Gary and Country-and-Western Gary, artist manqu Gary, macho Gary," 806). Farrell begins to pursue the single Gary who presumably stands behind these multiple poses, but he is "seized with depression at how few were the answers" to his inquiry (811). There is an "evil genius" in Gilmore's planning Nicole's suicide, but much else in Gilmore's life suggests sheer ignorance (812); Gilmore's relations with Bessie, his mother, seem a potential key, but the answers to many related questions provide no "hope of a breakthrough" (827; see 844). Continuing to ponder Gilmore's transcripts just before the execution, Farrell turns to yet another possible solution to the Gilmore mystery: Gilmore's fascination with small children. But this 'answer' is also unsatisfactory: "It was too insubstantial. In fact, it was sheer speculation .... beware of understanding the man too quickly!" (855). Beware indeed. Farrell's final comment on Gilmore takes us back to the passage from Andr Gide ("Please do not understand me too quickly") that Mailer first used as his epigraph to The Deer Park (1955). Farrell's conclusion should caution us against reductive readings, psychological efforts to pluck out Gilmore's mystery. Indeed, Gilmore's complexity should impress us as much as it does Farrell, whose prolonged efforts to understand Gilmore are akin to Mailer's.
Larry Schiller's role is in part like Farrell's. Schiller also looks for the human side to Gilmore, the "sympathetic character" buried inside the cold-blooded killer (629), for Schiller cannot imagine making a successful book or film unless he first makes this discovery.
<#FROWN:J69\>
1.1 PEX and PEXlib
PEX is the 3D extension to X. It adds over 200 X protocol requests for defining and displaying 3D pictures. PEX provides all the common features found in most modern 3D graphics systems, but provides them in a way that's seamlessly integrated with X.
The PEX graphics model is very unlike the 2D graphics model of basic X. With PEX you can define objects in convenient coordinate systems, rotate and move them with modeling transforms, view them from different angles, and create lit and shaded scenes.
1.1.1 The PEX Protocol
As most X programmers know, the underlying mechanism of X is the X protocol. While most of us think of X as consisting of Xlib, which generates X protocol, and the X server, which interprets X protocol, the essence of X is really the protocol itself. It is what defines X and allows it to be interoperable.
PEX is an extension to the X protocol. Like the X protocol, PEX protocol travels between the client - the application program - and the X server (see Figure 1-1). The server contains the PEX server extension, which receives and interprets the PEX protocol and executes the PEX requests. As always in X, the client and server can be on the same machine or on different machines.
<O_>figure&caption<O/>
Although PEX is defined as a protocol, it specifies a highly capable graphics system. This system and how you use it is what this book is all about.
1.1.2 PEXlib
Just as applications use Xlib to create and send X protocol, so, too, they use PEXlib to create and send PEX protocol. Xlib and PEXlib act as the application's agent, formatting and sending requests to carry out the application's will. You do not need to know PEX protocol to use PEXlib. You need only to understand the PEX capabilities and PEXlib functions.
Figure 1-1 shows the relationships between an application, Xlib, PEXlib, and the X server. The application sits above Xlib and PEXlib. It calls Xlib for the usual window management tasks, such as making server connections and creating windows, and calls PEXlib to draw 3D images. The X and PEX requests travel to the server intermixed on the same communication channel. The request dispatcher in the server routes the X requests to the server's core, and sends the PEX requests to the PEX extension. If other extensions are being used, then requests for them, too, are intermixed with the X and PEX requests; the server's dispatcher routes them to the correct portion of the server.
In our descriptions of PEX and PEXlib, we'll refer mostly to PEX, because that's really the system that's providing the 3D functionality. We'll use PEXlib only when we mean specifically the PEXlib interface to PEX.
1.1.3 What's In the Name? Only History
PEX is an acronym for 'PHIGS Extension to X.' This implies that the purpose of PEX is to support PHIGS, a popular 3D graphics standard (1.4). But while PEX does indeed support PHIGS and has had much of its definition taken from PHIGS, PEX goes beyond PHIGS in its functionality. The name, at this point, is merely historical, indicating where PEX had its origins, but not where it is today.
1.1.4 How Do I Know if I Have PEX?
You can determine whether a server has a PEX extension by invoking the xdpyinfo command and searching its output for the string 'X3D-PEX'.
% xdpyinfo | grep X3D-PEX
The R5 sample sever from the M.I.T. X Consortium includes a PEX extension (1.5).
1.2 PEXlib as a Graphics Library
Putting aside PEXlib's role as an X extension, you can look at it as simply another 3D graphics library, one that meshes smoothly with X. Graphics libraries sit between the application and the display device, manipulating the device in response to application commands. Graphics libraries can be low-level or high-level (see Figure 1-2). Low-level libraries typically provide commands for manipulating pixels, device registers, and other features of the display hardware. High-level libraries deal in abstractions like geometric objects and color. They hide low-level and hardware-dependent details.
<O_>figure&caption<O/>
1.2.1 PEXlib Is a High-level Library
PEXlib is a high-level graphics library. It allows a programmer to describe a graphic image in terms of familiar objects and attributes, without having to deal with the details of producing that image in a window. PEXlib lets a user say, Draw a wide red line from this point to that point, without requiring her to figure out which pixels to turn on and how to make them look red. All the details of producing the picture are handled by the PEX server. The user merely specifies the geometry, the location, and some appearance attributes for the objects.
This convenience extends to objects more complex than lines and attributes more subtle than color. Say an application wants a picture of a room full of furniture. PEX lets the user describe where the walls of the room are, the shape, location and color of the furniture, what types of lights are on and their locations, and where the viewer is. It then figures out how the room looks to that viewer - which parts are bright or dim, how the aim of the lights varies the color across the walls and upholstery - and sets the pixels on the screen to the values required by the picture (see Figure 1-3). PEX frees the programmer to compose a picture and not worry about the details of producing it.
PEX provides a set of familiar graphics objects called primitives, each with attributes that control its location, orientation, color, and appearance. You can define simple models or complex scenes with these primitives. Multiple views of the same objects can be displayed. You can even animate a scene by setting it in motion.
<O_>figure&caption<O/>
1.2.2 Displaying a Picture
PEX gives you two ways to draw a picture: 1) Pass all the primitives in the scene to PEXlib one by one and have PEX display them immediately. 2) Store the primitives in a graphics database, then tell PEX to display the database. In the first method, commonly called immediate mode, you must send all the primitives to PEX each time you want to change or redisplay the picture; once PEX draws the primitives, it forgets about them. In the second method, called structure mode, you need not re-send the primitives for each redisplay, but merely tell PEX to redisplay the stored data-base. You can selectively edit the database between redisplays, changing only those parts of the picture that should be different in each scene. The primitives and their attributes are stored in the database in containers called structures.
There are advantages and disadvantages to both these methods. Most applications are suited to one or the other. Immediate mode is best for applications that must continuously re-specify the complete images, such as animated simulation programs that display different geometry with each frame. Structure mode is best for programs such as computer-aided design programs that create a graphics model and edit portions of it frequently or view it in different ways.
It's also possible to mix both immediate mode and structure mode. You can store the static or infrequently changed parts of your model in structures, and draw the other parts using immediate mode.
1.2.3 Limitations of PEX
PEX has some limitations. It does not do ray tracing or radiosity. It doesn't compute shadows or follow light as it bounces from one object to another. It does not yet provide texture mapping, so you can't tell it to accurately display your marble table. And motion blur and realistic fog would be tough to do with PEX.
PEX implementations can extend PEX functionality, so some PEX implementations may indeed calculate shadows or perform texture mapping; but they'll do it in an implementation-dependent way that may not be the way it's done in other implementations.
PEX implementations may have other limits. Most restrict the number of light sources you can use (although often to a high number), some don't do depth cueing, and some provide only flat or Gouraud shading. Most PEX implementations do provide the majority of the PEX functionality, however, and offer a powerful, useful, and portable set of graphics capabilities, even though they don't support all PEX features. PEXlib provides functions for determining which features a PEX server extension does support.
1.3 Overview of PEXlib
Table 1-1 shows you what's in PEX. It lists the major PEX features and tells you where in this book we've put their primary description. Examples of using the features are throughout the book. See the Index to find the location of specific examples.
<O_>caption&table<O/>
1.4 The Relationship between PEX and PHIGS
PEX owes much of its origin to PHIGS. PHIGS [14] is a 3D graphics standard sanctioned by the International Organisation for Standardisation (ISO). It's been available since 1988, and is used by many graphics applications. A primary goal of PEX is to support PHIGS in the X environment, although that is not its only goal.
To an application, PHIGS serves much the same purpose as PEXlib, sitting between the application and the server, as shown in Figure 1-4. In the X environment, PHIGS sends PEX protocol to the X server to carry out the PHIGS functions.
With the birth of PEXlib, PHIGS implementations can themselves use PEXlib to format and send the PEX requests.
<O_>figure&caption<O/>
Most of the PEX requests were designed to support the identical functionality in PHIGS. Consequently, PEXlib and PHIGS use the same graphics primitives and attributes, and share the same model for defining a scene and rendering it. But PEXlib goes beyond PHIGS, providing immediate-mode rendering and better integration with X.
Because the PHIGS standard was defined before X was popular, it has no knowledge of X, specifically, the role of X in controlling the display and input devices and the aim of X to be 'policy free.' PHIGS defines its own ways to control the screen and the available input devices. It defines a central graphics database rather than a distributed one, and makes several of the decisions that a policy-free interface would leave to the application. Therefore it's not well suited as an extension to Xlib. PHIGS can be implemented so that it's well-integrated within X, but parts of it would have to be redefined to serve as an extension to Xlib - a nearly impossible task now that PHIGS is an approved standard and in wide use.
PEX and PEXlib are the results of a fresh start on a 3D interface for X. While accommodating and building on PHIGS, they begin with the assumption that 3D graphics should be an integral part of the window system. PEXlib uses the existing X mechanisms: the communication channel, protocol requests and replies, the X event queue, and error events. It avoids specifying policy, and instead provides the functions needed by the application to execute its own policy. By providing an immediate-mode capability, PEX meets the needs of more graphics applications than PHIGS.
1.5 Where to Get PEX and PEXlib
PEX server extensions are available from most major workstation vendors, as well as the M.I.T. X Consortium. PEXlib, too, is available from these same sources.
The PEX and PEXlib products provided by workstation vendors are usually highly optimized for the vendor's computer and display devices. These implementations typically provide the best possible performance for those devices. Some vendors require you to buy their graphics accelerator hardware to get reasonable or high performance, however.
The X Consortium's PEX implementation is known as the sample implementation, or simply, the PEX-SI. As of late 1992, the PEX-SI does not support some high-level PEX features, such as shading and hidden line and hidden surface removal. It also does not provide the highest possible performance on many display devices. Nevertheless, it's a reliable and useful implementation, and provides a good basis for learning PEX and a good starting point for building more complete implementations.
There is only a C interface for PEXlib. We have not heard of an effort to define a FORTRAN binding or any other interface.
<#FROWN:J70\>
Enhancing CRP values
By Ted Hawn and Mike Getman
Cooperative, interagency assistance to CRP contract-holders in Montana has resulted in some model acreages from the standpoint of soil erosion control and wildlife habitat development
A model Conservation Reserve Program (CRP) contract. That's how one might describe Lee Berry's land in central Montana. Berry is a landowner with an eye to the future. His 445 acres, enrolled in the CRP in 1988, now provide optimal values for soil erosion control and wildlife habitat.
Seeding the recommended mix of three wheatgrasses, alfalfa, and clover, Berry also planted nearly 24,000 feet of tree rows - more than 4,400 trees. At a county tour held during the fall of 1989, county commissioners, conservation district supervisors, and other local agency personnel saw first-hand what it takes to successfully establish CRP cover. At a stop on the Berry farm, tour delegates were duly impressed when several walked into a seven-foot-high clover and grass field resembling a willow thicket and disappeared from view. With assistance from the U.S. Fish and Wildlife Service (FWS), three ponds were constructed and food plots were added. The diversity of habitat has resulted in phenomenal changes in the wild-life species present and their populations.
Opportunities aplenty
The 1985 farm bill's CRP has provided excellent opportunities for the management of natural resources on private land. Federal and state agencies have taken this opportunity to complement and maximize the benefits on CRP acres. Most agencies have tailored their programs to benefit the resource for which they are responsible. The Private Lands Program was created by the FWS to develop wetland habitats in conjunction with CRP with emphasis on improving waterfowl production. Improved nesting habitat on private land should aid continental duck populations, which are at record lows. Grass and legume plantings on highly erodible land provide excellent erosion control and the necessary cover for waterfowl to nest on millions of acres throughout the region.
The scarcity of water is a limiting factor for waterfowl production in Montana. The Private Lands Program emphasizes water development, and projects typically consist of restoring wetlands, repairing dams, or constructing new impoundments. Interested landowners also can receive funding to include certain grass and forb species in their seeding mixtures. This assistance, in conjunction with cost-sharing funds from the Agricultural Stabilization and Conservation Service, generally covers most of the cost of obtaining seed for these plantings.
Data collected on first-year water projects in CRP fields confirmed that ducks readily occupy newly created habitat. Waterfowl pair counts showed an average of 2.2 breeding pairs per wetland acre. This is similar to the pair counts during the Prairie Pothole region's highest production levels in the 1950s and 1960s. Production most likely will increase in the future because progeny return to the same area to nest.
Field observations also have shown increased benefits to resident wildlife. An estimated 300 pheasants wintered in a 50-acre CRP field where the diverse seed mixture provided optimal cover and food needs. An ongoing radio telemetry study has shown that white-tailed deer and mule deer have preference for and concentrate in CRP fields. Elk also have been seen using these fields, especially where sweetclover or alfalfa are planted.
More than 150 mule deer wintered in 1988-1989 on Fred Lahr's CRP fields just south of Denton, Montana, in an area where few deer were seen on annual census routes. Biologists for the Montana Department of Fish, Wildlife and Parks attributed the increased wildlife populations to the excellent diversity of wheatgrasses and alfalfa on Lahr's CRP acres. Soil Conservation Service conservationists have worked closely with Lahr on his 700 CRP acres. They assisted in designing four separate shelter-belt and habitat plantings. In addition, FWS constructed three dams to provide water for wildlife.
Residual vegetation is a key factor in providing suitable nesting cover for waterfowl, upland game birds, and other ground nesting birds. For high quality nesting cover, seed mixtures include rhizomatous grasses that provide more contiguous nesting cover than does the scattered cover associated with bunchgrasses. Grass species with fibrous stalks and leaves that are resistant to flattening from winter winds and snow provide higher quality nesting cover. Forbs, such as alfalfa and clover, provide diversity for nesting cover and are preferred forage for mule deer, white-tailed deer, and elk.
The CRP acres on Pat Burton's Double B Farms also exemplify what can come of extra effort to further program goals. Burton's 1,100 CRP acres are situated in the fragile semiarid area of the rugged Missouri River Breaks and adjoining the Charles M. Russell National Wildlife Refuge. Drought and grasshoppers plagued his 1987-1988 plantings, causing the seeding on about 620 acres to fail. Another factor also affected seeding establishment. Elk from the adjoining refuge sought out tender seedlings and 150 to 200 head spent part of the winter on the fields.
Working through the Montana Department of Fish, Wildlife and Parks' Upland Game Enhancement Program, Burton seeded a mixture of three wheatgrasses, clover, and alfalfa. Native upland gamebirds, such as sharp-tailed and sage grouse, should benefit from the improved nesting and protective cover. Also, a number of shallow water developments will be constructed using Private Lands Program funds.
Willing landowners
Landowner attitudes toward this multi-agency approach to enhancing resource management generally have been positive. Landowners favor programs that provide an annual income, improve the equity in their land, and allow them a decisive role in management. Some landowners were hesitant at the outset. But after several projects were successfully completed and the benefits from these programs became obvious, the interest increased to the extent that it now exceeds the financial and manpower capabilities of the several agencies involved in the project.
Maximum accomplishments occur when personnel from agencies with programs to assist landowners coordinate and cooperate with each other. Landowners with specific objectives can be directed to the agency that best serves his or her needs. In many situations, a landowner's objectives can be met and funded cooperatively among agencies in a way that multiple resource benefits can occur.
Everyone gains through these cooperative ventures. Decisions are made at the local level, and the land is treated according to its suitability. Soil erosion is controlled, wildlife habitat is developed, and a healthy working relationship is established between landowners and the government agencies involved.
The pressure is on all people involved in natural resource management to be creative and flexible in managing natural resources. Through these cooperative working relationships and mutually beneficial programs, all of society wins.<*_>square<*/>
Citizens in Mason County, Illinois, have found that by working together they can deal effectively with agricultural-related threats to local drinking water supplies
Local resource planning for water quality improvement
By Dale A. Boyd
THE scene was a local coffee shop in Mason County, Illinois. Farmers were discussing a research study that showed high levels of nitrates and traces of pesticides in the local aquifer that supplies drinking water to most of the county. Nitrates exceeded the drinking water standard in 70 percent of the samples from 10 monitoring wells and in 39 percent of the samples from 14 generally deeper domestic wells. Trace levels of metribuzin; atrazine; and/or simazine, propachlor, and trifluralin were detected in groundwater samples from 10 to 13 monitoring wells located downgradient of corn and soybean fields (3). No trace levels were found in the domestic wells.
The farmers were concerned for two reasons. First, the aquifer was their supply of drinking water. Second, they were worried that if the problem persisted they would be forced to stop using chemicals vital to their farming operations.
The result of that early morning discussion was the Illinois River Sands Water Quality Project, which has been a reality in Mason County, Illinois, since the spring of 1990. The project came about because of a unique grassroots effort organized by local individuals and farm organizations to address existing and future water quality problems. It is an excellent example of local organizers working successfully with state and federal agencies to solve problems within a community.
Nationwide, 74 hydrologic-unit-area water quality projects have received funding. The Mason County project is one of two in Illinois. The activities of a group of concerned citizens and local organizational representatives known as the Water Issues Resource Planning Committee were important factors in the project's selection. Resource planning is a process that encourages individuals in a community to come together to identify and discuss mutual problems and needs and, with the help of federal and state technical advisors, develop a plan of action to address local resource concerns
The backdrop
Mason County soils include large areas of sands, the remnants of a prehistoric river bed. Water in the aquifer is held in pore spaces in a sand layer that ranges from 60 to 200 feet thick, depending upon the location in the county. The water table is high. In most places it is only 3 to 12 feet below the surface. The available, abundant supply of water has resulted in 35,000 acres of specialty crops grown annually (1) and more than 100,000 acres of irrigated fields. Combined with shallow domestic wells, conditions are such that the drinking water supply is quite susceptible to contamination by farm chemicals. According to Hallberg (2), while nonfertilizer uses, such as septic systems or manure disposal, may cause nitrate contamination in specific circumstances, regional increases in nitrate levels in shallow groundwater directly reflect increased use of nitrogen fertilizers, especially in sensitive areas, such as western Mason County.
Some families with small children have had to haul water because of high nitrate concentrations. Excessive nitrates can reduce an infant's ability to carry oxygen in its blood, causing a condition known as methemoglobenemia or 'blue baby' syndrome.
Concerned farmers approached the Soil Conservation Service (SCS) district conservationist in March 1989 to ask what the agency could do to help. SCS representatives took the case to the Mason County Soil and Water Conservation District Board, which formed the Water Issues Resource Planning Committee. That committee's first meeting was held May 24, 1989. Efforts were made to have diverse interests in the community represented. Members included individuals from local and state governmental agencies, local farm organizations, and agribusinesses. This initiative led to the area qualifying for the U.S. Department of Agriculture's water quality hydrologic-unit-area designation.
Committee members realized that, while they understood the nature of the problems in their community, they did not have the technical expertise or the financial resources to arrive at a solution by themselves. They again asked the SCS district conservationist for help. As part of the resource planning process, a technical advisory committee was formed and met for the first time in June 1989. This subcommittee of the Water Issues Resource Planning Committee included technical experts from the University of Illinois; members of local farm organizations; and representatives of federal, state, and county governmental agencies. It provides technical assistance and guidance to the local resource planning committee, which makes decisions based on technical advisory committee findings and reports.
The resource planning committee then met and discussed problems and concerns about the local aquifer. Members decided that their overall objective should be to maintain and enhance the quality of water from the aquifer. They also decided to limit the project area to the western two-thirds of the county, where more than 600 irrigation systems irrigate about 100,000 acres.
The result of this work was a resource plan. The purpose of a resource plan is to document a resource problem in an area and outline alternative solutions that can be used in addressing the problem. In this case, the resource plan was used as the basis to apply for USDA water quality hydrologic-unit-area funds.
The technical advisory committee recommended a number of practices, many related to agricultural production, that could reduce the amount of chemicals reaching the aquifer. After the planning committee chose practices from the alternatives presented, they asked SCS, the Agricultural Stabilization and Conservation Service (ASCS), and the Cooperative Extension Service (CES) to assist in developing a joint plan of work to maintain and improve the aquifer.
<#FROWN:J71\>
Submarine Maneuver Control
By William P. Gruner and Henry E. Payne III
Submarine warfare is no longer limited to launching weapons against enemy ships, submarines, and land targets and evading enemy antisubmarine (ASW) attacks. It includes collecting intelligence information on activities near foreign shores and at sea and tracking, intercepting, and escaping from foreign submarines. To meet these challenges, U.S. submarines must be able to maneuver rapidly while engaged in their missions.
The public generally is not aware of the close encounters our submarines experience. Table 1 shows a few examples.
The need for great maneuverability in these deadly games of observe, tag, and evade is evident. To make an effective and safe maneuver in the proximity of other submarines, ships, and mines, a diving officer needs two things - precise control of his own submarine and accurate knowledge of the positions and movements of other vessels and objects. Our existing attack and ASW submarines are hampered by three major shortcomings.
Although capable of speeds in excess of 30 knots, they are unable to make small-radius turns at high speeds safely, because they become unstable.
The normally used four-man, manually operated control system lacks the rapid-response capability for directional control of an unstable submarine.
The captain of a submerged submarine has little knowledge of the precise locations of other submarines and ships in his vicinity. Consequently, he does not know how best to maneuver.
The Stability Problem. A submarine at rest has static stability. When it starts to pitch or roll, moments are generated by buoyancy and weight forces to restore it to the rest position. When propelled, it has dynamic stability if it can be made to follow a predictable path. However, pitch-yaw hydrodynamic coupling (the dreaded snap roll) causes severe stability problems when maneuvering at high speeds. Dynamic stability is a prerequisite for directional control, and directional control is a prerequisite for maneuverability. High maneuverability cannot now be achieved because directional control cannot be maintained when large rudder angles are applied at speeds in excess of 15-20 knots. Once control is lost, the submarine may be lost if the diving officer cannot regain control.
Recognition of this instability problem did not come until very powerful propulsion plants and better stream-lining made high speeds possible. The problem was high-lighted in 1954 when the experimental high-speed, battery-powered submarine USS Albacore (AGSS-569), designed with a sleek body-of-revolution hull, became operational. With a top speed greater than 30 knots, she was equipped with specially designed control surfaces and a fully automated control system that could be operated with from one to four men, with or without selected automation. Instability soon became evident. One commentator noted, "If in a melee situation, a modern high-speed sub pilot tries to turn too sharply at too high a speed, he might find himself in a snap roll, hanging from his seat belt and with a loss of several hundred feet in depth at a markedly slowed speed."
<O_>table&caption<O/>
An early lack of concern about the stability problem probably was because, until the end of World War II, very few of the world's submarines had submerged speeds in excess of ten knots. Manual control of bow and stern planes and helm by separate operators directed by a diving officer was considered adequate at normal speeds of two to six knots. If 'the bubble was lost' while increasing depth, an alert diving officer could soon regain control by backing and blowing tanks.
Later, when nuclear-powered submarines of the Skipjack (SSN-585), Sturgeon (SSN-637), and Los Angeles (SSN-688) classes - all characterized by body-of-revolution hull designs and large sails - became operational, they, too, had stability problems similar to the Albacore's. Their instability is caused by external variable forces acting on their hulls - a result of pressures exerted by the swirling masses of water that are continuously generated by the hull as it turns and drives its way through the ocean. These masses of water, termed vortices, rotate inward toward the center of the hull. The higher the submarine's speed, the more energy is imparted to them, and the greater the pressures acting on the hull.
Figure 1 is a computer simulation of vortices generated by a submarine during the mid-stage of an evasive maneuver at 24 knots with a 30 rudder angle. Note the displacements of the twin vortices as the ship progresses through the water and the effect of the sail on their location. Under the influence of the forces generated, the stern will squat down and the bow pitch up. Meanwhile, the forces will cause the submarine to sink and slow, and it will go into a snap roll unless speed is reduced by backing and the rudder put amidships.
Sample patterns of wind tunnel airflow tests, using a scale model of the Skipjack hull, are shown in Figure 2. The first pattern shows the airflow to be quite smooth when the submarine is on straight course. However, when the model is placed at significant roll and yaw angles as in the second pattern, the strong influence of the sail pulls the upper vortex out of its place along the hull and into the wake of the sail. Comparisons of such patterns at different roll angles, yaw angles, and speeds reveal dramatic variation in the oscillating flow and pressure patterns.
Full-scale submarines experience similar flow conditions when large rudder angles are applied at high speeds. Resulting unbalanced pressures on the sail, for example, can produce forces of millions of pounds, with corresponding moments great enough to rotate a submarine about its roll axis. This results from a submarine's relatively small righting moment (because of its small meta-centric height) about the roll axis.
The Directional Control Problem. When a roll begins under these conditions, a coupling of motions about the roll, pitch, and yaw axes occurs to cause a submarine to veer suddenly and radically in unexpected directions and to change depth and speed. Manual control under such conditions becomes highly difficult, if not impossible.
In discussing directional control of the Albacore, an early commanding officer stated, "It was evident that the Albacore performed significantly better without automated controls programmed by a single operator than she did with a standard four-man team of diving officer, helmsman, bow planesman, and stern planesman." He likened the submarine control problem to that of high-speed aircraft in which "designers turn to higher performance control machines to offset human limitations in sensing and reacting, a lack of uniformity of performance, and limited adaptability to performing multiple requirements simultaneously." He believed future submarines would be "flown" with automatic controls with pilot override.
Limited human capabilities - the human brain and muscular control systems operate too slowly and have other limitations - caused airplane designers to turn to higher-performance-control machines. Pilots are unable to keep pace with rapidly changing situations while faced with multiple tasks during aerial combat, low-level flying, and landing. Further, human systems, if under stress, are likely to err, function inconsistently, and freeze. The repeatability and reliability of modern computer-aided systems have made them the only choice for high-speed airplane maneuvers. For similar reasons, they are the only choice for controlling high-speed submarine maneuvers.
Despite the Albacore's favorable experience with the automatic one-man control system in the mid-1950s, the system was looked upon with suspicion. Although other work was performed to improve the controllability of submarines at high speeds, the basic four-man diving team used in low-speed submarines was applied to nuclear-powered submarines, with limited provision for a one-man control system. To minimize the risk of loss of control, maneuverability is now limited by operational procedures that constrain rudder angles to a very few degrees at speeds above about 15 knots. As another safety measure, the rudder control system prevents the application of large rudder angles at high speeds.
The Approach to Greater Submarine Control
For our submarines to achieve greater maneuverability, they must be modified through the application of current technical knowledge in control systems, hydrodynamics, human engineering, and physics. Much of this knowledge can be borrowed from aircraft and missile engineers, who were faced with stability and control problems as soon as Orville and Wilbur Wright demonstrated a capability for powered flight in 1903.
The well-known author and aeronautical engineer Nevillle Shute described the state of airplane control during World War I, when most airplanes flew at speeds of 60-130 knots. "We knew that a clumsily executed turn might have the effect of putting an aeroplane into a spinning nose-dive (a Parke's Dive, some of us called it, because Lieutenant Parke was one of the very few people who had come out of it alive). In general, a spin, once started, continued to the ground, the machine hitting very violently. And that literally, was all we knew about it."
As airplane speeds increased, it became obvious that manual flight control was no longer acceptable. It was recognized that, "human pilots were incapable of adequately compensating for the rapid changes encountered during tactical engagements, adverse weather conditions, ground-skimming flight, and other maneuvers. To make flight feasible, pilots were provided with assistance in the form of automatic flight-control systems. Such systems now are incorporated in all modern high-performance aircraft, including C-141 and C-5 transports, B-1 and B-2 bombers, and F-4, A-6, F-14, F-15, and F-16 fighters and fighter-bombers.
"Individual systems vary from aircraft to aircraft, but in general provide pitch, roll, and yaw control augmentation, autopilot modes of operation, and altitude hold. The pilot normally enters maneuvering orders by means of stick or control column and rudder pedals. Major elements of the flight-control system include control-force transducers, pressure and temperature sensors, directional gyros, accelerometers, a central air-data computer, and a flight-control computer. The latter interprets transducer input from pilot actions and other data sources, processes data, and sends commands to the various servo-actuators to assure that proper control surface and device responses are made to pilot orders within safe flight limits. Control system reliability is achieved by rigid parts selection; inspection and test at component, subassembly, and system levels; and redundancy in electronic, hydraulic, and power supply elements."
While scientists and engineers were developing systems to control supersonic aircraft, others were developing control systems for intercontinental ballistic missiles, to enable them to deliver payloads accurately and reliably to targets thousands of miles distant. At that time, submarine designers seemed content to modify two-dimensional control systems for submarines operating in a three-dimensional ocean. Since these systems do not provide the degree of maneuverability required, a new control capability must be developed - one similar to aircraft and missile control-system technology.
All bodies in motion - including submarines and air-planes - are subject to the same inviolable laws of physics. These laws govern relationships between mass, force, torque, inertia, and acceleration. One important law states that rotational acceleration about an axis is proportional to the moment applied and inversely proportional to the inertia characteristic (moment of inertia) of the body about that axis. The high energy vortices noted previously and the large sail are the main causes of the upsetting moments that cause U.S. submarines to lose control in high-speed turns. Therefore, special attention must be paid to reducing sail size and its resultant interaction with the hull vortices.
Concept for an Automatic Submarine-Control System
Because of the complexity of submarine dynamic motions, the rapidity with which upsetting moments are generated, the speed with which control forces must be applied, and the inability of humans to exercise manual control, the new control system must employ computer technology. Figure 3 presents a concept for an automatic submarine-control system composed of three major subsystems: an automatic maneuver sub-system, a pressure-sensing subsystem, and an automatic attitude-control sub-system.
The Automatic Maneuver Subsystem. This subsystem performs two major functions. First, it provides the man/machine interface by which the diving officer enters maneuver instructions and receives information. To initiate a maneuver, the diving officer specifies the required maneuver within the framework of an earth-oriented, north-referenced, three-dimensional orthogonal coordinate system. Maneuver instructions may call for a simple turn and a change in depth or for more complex maneuvers to avoid a collision or to reach a distant weapon launch position or a position offset from a moving-target track.
<#FROWN:J72\>
Retrospective - 1992
Internal operations and relations received primary attention. The pace of labor relations activities provided a breathing spell allowing improvements to be made to the tools for dispatch, for allocations and for labor relations information retrieval. Longer range assessment forecasts were introduced at the request of the Industry, and a complete rewrite of the Tonnage Reporting and Assessment Procedures Manual was begun to provide clearer and more explicit reporting instructions for the Industry.
Labor Relations
Labor relations in the West Coast longshore industry is governed by the Joint Port Labor Relations Committees (JPLRC's), the Coast Labor Relations Committee (CLRC), the Coast Steering Committee (CSC) and the Area Sub-Steering Committees. The CSC and Sub-Steering Committees are made up of PMA member company executives, and the JPLRC's and CLRC are composed of Union representatives and the PMA member company representatives.
The meetings and activities of these Committees handle the day to day administration of the Coast contracts, and their agenda determine the direction and scope of much of PMA labor relations staff activities. The bulk of labor relations activity this past year took place at the level of the JPLRC's.
The Joint Port Labor Relations Committees process routine grievances and administer the agreement provisions locally within the various ports. These committees, composed of local Union and Employer representatives, also maintain registration lists and rosters of casual labor pools, coordinate the travel of manpower to ports with sufficient work opportunity, and facilitate the participation of the work force in the General Safety Training, Skilled Training and Drug/Alcohol Free Work Place programs. Their efforts, with the cooperation of the Union membership, have improved the dispatch of manpower, have resulted in fewer instance of labor shortages on peak work days, and have produced a reduction in the number of on-the-job injuries.
The Coast Labor Relations Committee developed a procedure for complying with the Americans' with Disabilities Act. Additionally, the CLRC handled routine matters such as processing referrals from JPLRC's seeking registration approvals and periodic updating of Low Work Opportunity Port lists and negotiated mileage allowance formulas. The Employers and the Union held preliminary discussions on the topics of computerization and new technology in the longshoring industry and the negotiation of a separate labor agreement covering rail container transfer facilities. The preliminary discussions did not result in follow-up meetings with the Union, and no new understandings were reached on either topic.
Labor policy issues concerning several member companies occupied the attention of the Coast Steering Committee. One issue involved a company's negotiating a separate agreement with the Union in violation of policy, and the agreement was eventually nullified. Two issues concerned payroll practices that created an underpayment of man-hour contributions. Both practices were corrected, and the proper man-hour contributions were collected. The Committee continued its review of new technology issues based on periodic reports and on-site visits by its Clerks' Sub-Committee.
The relatively slow pace of labor relations activity provided PMA staff with the opportunity to improve and develop new work place tools. A five-volume index of alphabetical listings of contract interpretations, arbitrators' decisions and historical references covering the period from 1934 to the present was placed into an electronic database. Labor relations personnel in each of the Area offices can now research these data by subject, date, or arbitrator's name quickly and easily.
The PMA Allocator in each of the four Areas is now using personal computer systems to perform many daily allocation functions. (The PMA Allocator accepts manpower orders from the Employers, prioritizes the orders and transmits the information to the appropriate dispatch hall. This process traditionally includes a voluminous amount of manual record keeping and data retrieval.) These systems are in varying degrees of implementation at year-end, and the goal is for computerized allocation to be fully operational soon.
Several staff members participated in labor relations and arbitration work-shops, computer user seminars, and safety and training conferences, and the four Area Managers visited the ports of Rotterdam, Felixstowe, Tilbury and Bremen/Bremerhaven in October to observe first-hand the technology and terminal operations at these ports. Institutional and structural differences in labor relations and operational procedures were explained by our hosts at ECT/Sea-Land Delta Terminal, the Port of Felixstowe, the Port of Tilbury and Bremer Lagerhaus-Gesellschaft.
Training
A record number of employees, 7,437, completed ILWU-PMA training programs. This represents a 40% increase over the previous year, and 85% of the training was Safety training. A significant number of direct employer representatives were among the trainees, and in several programs, ILWU members and company representatives attended classes together. A list of training programs showing the number of participants by program is shown on page 38.
General Safety Training
The General Safety Training (GST), described in detail in the 1991 Annual Report, targets the entire workforce over a three year period and addresses all legislated training requirements. More than 5,000 individuals have already participated in the GST program. The program underwent planned revisions designed to fine tune the curriculum, to introduce new material and to incorporate new regulatory requirements. Three successful custom videos were added, 'Personal Protective Equipment,' 'Vehicle Safety' and 'Local Area Hazards.' The popularity of this program is in large part due to the videotaped segments depicting the longshore work force engaged in their job functions in West Coast ports.
Alcohol/Drug Free Workplace Training For Supervisors and Employees
In February, the Alcohol and Drug Free Workplace (ADFWP) Training for Supervisors commenced, and more than one thousand employees have completed the training. This program is a logical extension of the nationally recognized Alcohol and Drug Recovery Program (ADRP),and it targets both labor and management supervisory personnel.
Drug and Alcohol Free Work Place Training is also incorporated into the General Safety Training.
Winch Training
A Winch Training Program was begun for members of Locals 13 and 46, and this marked the culmination of negotiations with the Maritime Administration on the use of its USNS Curtiss (T-AVB4) for ship's gear training. The Curtiss, a Ready Reserve Fleet vessel homeported in Port Hueneme, is ideally suited for 'yard and stay,' 'swinging boom' and 'jumbo' hands-on training. Rooms on board provide space for classroom instruction and rigging demonstrations. More than one hundred trainees have completed the new program since September.
Video Production
Videotaped presentations of training materials are rapidly supplanting other traditional teaching aids. Video delivers high quality, standardized, scripted training to the work force. Several new videos were added to our library this year. In addition to the three previously described in General Safety Training and the one in Alcohol/Drug Free Workplace Training, each Area contributed ideas for videos which are in various stages of completion. The Oregon Area completed work on 'Lumber Handling,' a ten minute description of handling packaged lumber. 'Marine Fork lift' is in progress in Washington, and 'Autos' and 'Lashing' are being planned in California.
Accident Prevention
The revisions to the Man-hours and Injury Incidence Rate Reports were completed this year. The new reports provide member companies with both quarterly and year to date injury/illness statistics so that they may more readily assess the progress of their individual safety programs.
The PMA Coast Accident Prevention Awards Program continued to generate considerable interest this year. Emphasis on individual achievement has resulted in increased participation as well as deserved recognition. Awards information and other statistical data for 1992 is contained on page 39.
Several staff changes occurred during the year. Larry Hudson was promoted to Area Supervisor, Oregon Area, replacing Dale Larson upon his retirement. Joe Boettcher joined the Oregon staff as an Assistant. Tony Peredo was promoted to Area Supervisor in the Washington Area, and Fred Gordon joined the staff as his Assistant.
Legal
As in the past, third-party litigation consumed the major portion of PMA legal effort this past year. NLRB charges filed by casuals against the 'permissive rule' are pending in the State of Washington and in Southern California. The Golden consent decree and its progeny required considerable attention in the federal courts in Los Angeles in litigation involving unsuccessful applicants for registration or transfer who attempted to use the decree for their own purposes. Although the Golden decree has generated considerable litigation in its ten years of life, it also has provided a generally consistent and stabilizing forum for the adjudication of issues concerning waterfront employment. PMA counsel and ILWU counsel have worked together to provide an effective defense for the parties to the collective bargaining agreement.
While a federal court action filed by several casuals in Oregon claiming so-called 'contingency list' status was pending, the Coast Arbitrator rendered a decision approving the dissolution of the list in Portland and permitting registration based upon hours of experience in accordance with the collective bargaining agreement. The federal district court in Oregon dismissed the lawsuit in view of the Coast Arbitrator's decision.
After much administrative planning and oversight by the NLRB, the re-registration of longshoremen at Local 18 in Sacramento finally is underway. This should conclude the lengthy and costly litigation and the processing of NLRB ordered deregistration and re-registration.
In what has become known as the 'Fish Case,' a Coast Arbitrator's decision was received this fall involving an issue crucial to the future membership of steamship agents in the PMA. The Coast Arbitrator remanded the case to the Area Arbitrator because of the absence of facts showing that the husbanding agent had control of cargo operations for which it was held responsible. We await anxiously the final outcome of this matter.
For the first time in over a decade, PMA became involved in litigation against a member company. Following a refusal to pay interest penalties on delinquent tonnage assessments, Long Beach Container Terminal, Inc., filed a declaratory relief action in the California Superior Court against PMA and various agents and foreign steamship operators. PMA and other defendants have answered with cross-complaints to require LBCT to pay the interest penalties. Discovery is proceeding.
Early in the year, the NLRB sought and obtained a temporary restraining order and injunction from the federal district court in Los Angeles, barring the ILWU from interfering with PMA members doing business with the Southern Pacific ICTF in the Ports of Los Angeles and Long Beach. The PMA operations quickly returned to normal after a one-day shutdown, with the exception of the refusal by Local 63 Marine Clerks to perform work voluntarily during the meal hour in a manner consistent with conduct prior to the dispute. After a favorable area arbitrator's award finding Local 63 in violation of the contract, the Coast Committee declared the contract remedies exhausted. PMA, in a rarely necessary proceeding, sought and received confirmation of the award from the federal district court. Local 63 continued to violate the award, and PMA moved for an order to show cause why the union should not be held in contempt. The motion was denied. PMA will continue to monitor compliance with the award for future enforcement. The ILWU has appealed the district court confirmation of award to the Ninth Circuit.
PMA engages the services of local law firms in each of the four port areas on the Coast. These firms continue to serve the organization extremely well. Notwithstanding occasional exceptions, the relationship between PMA counsel and ILWU counsel remains professional and courteous and serves as an aid to the resolution of legal problems in the industry.
Finance and Administration
Building on the successful completion of the coastwise payroll system in 1990, the Information Services (I.S.) Department began the redesign and reprogramming of the Pay Guarantee Plan payroll system to make it consistent in all four Areas. Design is completed, and the new system should be operational in the second quarter. Since the payroll system has been accepting automated input, more than 90% of the payroll hours processed each week are transmitted to PMA electronically.
I.S. staff continued their involvement with several projects related to stream-lining the dispatch process. Significant progress has been made in Los Angeles/Long Beach where the UTR board in the Wilmington Longshore Dispatch Hall was the first to be modernized. In addition, telephone check-in for registered longshoremen should be operational in early 1993.
<#FROWN:J73\>
SUBSONIC AIRCRAFT PROPULSION/AIRFRAME INTEGRATION
The technology of integrating propulsion systems and airframes involves the ability to assess and control the development of wave drag, induced drag, and profile drag. Advances in CFD over the past decade have contributed greatly to this technology. It is anticipated that ongoing CFD developments will lead to even further refinements.
Two areas remain in which technology improvements are needed. One is the development of wind tunnel test techniques and powered propulsion simulators to better represent installed power effects of the forthcoming generations of very high bypass ratio engines in wind tunnel testing. The other is the need to predict the installed characteristics of thrust reversers, both computationally and with wind tunnel testing techniques. These are areas in which NASA can make important contributions.
AERODYNAMIC CRUISE PERFORMANCE
Although the fundamental physical principles of subsonic and supersonic airflow around aircraft are the same, design approaches to minimizing drag are greatly affected by the cruise speed. This section of the report discusses cruise performance in the two speed ranges separately.
Subsonic Aircraft Cruise Performance
Long-haul subsonic transports are now, and will be for the foreseeable future, the major product of the civilian aviation industry and infrastructure. As noted in Chapter 2, from 1975 to 1995, aerodynamic efficiency will have increased by approximately 10 percent, and if the current rate of improvement is maintained, another 5-10 percent is projected by the year 2020. However, ordinary development or evolution alone will not keep the United States at the forefront in the world market. Although continued evolutionary advances in methods and processes (experimental, theoretical, and computational) are needed to provide continued improvement of aerodynamic design technologies, demonstrated innovative technologies are necessary in the longer term to provide opportunities for significant improvements in performance.
Laminar Flow Control
The flow on most of the surfaces of an aircraft is turbulent. Laminar flow control (LFC), hybrid laminar flow control, and natural laminar flow are promising sources of skin-friction drag reduction on aerodynamic surfaces. Laminar flow nacelles are also being studied by NASA. Laminar/turbulent transition of the airflow next to the aircraft surface is delayed through a combination of pressure gradient tailoring of the wing and control such as suction through the skin. If full-chord laminar flow can be maintained in this fashion, fuel savings of up to 25 percent could be realized.
Transition is extremely sensitive to freestream conditions (e.g., freestream turbulence and acoustics) and surface roughness (e.g., rain and ice crystals, insect debris, surface finish, and fasteners); lack of confidence in these issues has hindered the use of this concept on vehicles. Also, of perhaps greater significance have been the questions of fabrication cost and operational cost and maintainability.
Engineering and optimization tools have outpaced the state of the art in transition prediction theory. Thus, the design of LFC, hybrid laminar flow control, and natural laminar flow systems depends on empirical bases to determine transition. This method is also limited because it cannot account for the effects of surface roughness and freestream disturbances.
Knowledge of transition - so very important to the success of LFC techniques - is, in general, limited to the simplest of geometries. Efforts to better understand the transition flow physics are under way to provide valuable guidance for the surface roughness and freestream disturbance problems.
Only a limited number of flight tests have been flown since the original and successful X-21 program of the 1960s; these are the JetStar (NASA/Langley) and Boeing 757 (NASA/Boeing). In both cases, extensive laminar flow was successfully achieved on the upper surface of the swept wing through the use of suction. Very low suction levels were required, with power penalties of the order of 1 percent. Studies with engine noise indicated no effect. The use of a Krueger nose flap eliminated a potential buildup of insect debris on the leading edge.
The remaining challenges to the implementation of laminar flow technology in large subsonic transport designs include validation of the technology in actual airline service operating environments and exploration of the technical issues associated with making laminar flow operate effectively on the inboard portion of the wings of very large aircraft. Recognizing the challenges, during 1990 NASA and the industry developed a cooperative research plan; however, these efforts have been delayed by overall program constraints. Meanwhile, the Europeans have rapidly advanced their laminar flow efforts. Airbus plans for laminar flow technology validation include extensive large-scale testing, targeting technology validation as early as 1993.
Turbulent Drag Reduction
The most promising technique demonstrated thus far has been passive control by riblets, tiny streamwise grooves on the aircraft surface. This device is useful for surfaces on which laminar flow is very difficult to achieve (e.g., the fuselage). The approach was used successfully on the U.S. entry in an America's Cup Race and then flight-tested on a portion of a business jet, achieving a reduction in local skin-friction drag of 8 percent.
The state of the art in turbulence predictions depends on empirical correlations and models, usually developed for one set of flow conditions or a very simplified model. Here also, efforts at understanding the basic physics of turbulent flow are under way. Prediction and control have been hindered by the lack of reliable, efficient models of turbulence for complex geometries.
Advanced Supercritical Airfoils
Advanced supercritical airfoils, which reduce the shock strength on transonic airfoils, have contributed to drag reduction and have been used on all commercial transport aircraft developed since 1975. Further modifications with reduced moments and weaker shock waves are under study by NASA for use with LFC systems.
Improved understanding of shock/boundary-layer interactions has led to new opportunities to greatly improve airfoil design concepts and procedures.
Wing Design
The improvement of theoretical analysis tools and CFD, coupled with a better understanding of flow physics, has enabled the design of more aerodynamically efficient wings with greater thickness and reduced sweep. This allows a wing weight reduction or higher aspect ratio. Substantial improvements in cruise Mach number and critical Mach number have also been realized. New opportunities exist to significantly improve the design optimization procedures for wings that incorporate laminar flow systems along with advanced high-lift systems.
Winglets
Winglets, or wingtip extensions, which first appeared on business jets, are now used on various versions of commercial transports (e.g., the Boeing 747-400 and MD-11). These effectively increase the aspect ratio of the wing. Advancements in understanding of the 'nonlinear' effects of wing-wake-deflection and roll-up have created opportunities to improve the design optimization procedures for winglets and other wingtip devices for drag reduction. In each of these technology topics, significant opportunities have developed to advance the state of the art; however, constraints in the aeronautics program have limited NASA's ability to support the needed advancements in experimental or computational capabilities and in ground and flight validation of these technologies.
<O_>figure&caption<O/>
Status of Subsonic Technology
As noted earlier, the past decade has seen large improvements in wing design technology, the successful demonstration of laminar flow control on a Boeing 757, the successful demonstration of riblets on a business jet, the successful demonstration of natural laminar flow on a business jet, and the use of supercritical airfoils and winglets on transports in everyday commercial service.
Advances in computer and quiet-tunnel/instrumentation capabilities have allowed details of the fundamental flow physics of transition and turbulence to be studied. Computational time for these efforts, however, is too long for extensive use in the design process at present.
NASA should play a leading, but not exclusive, role in the development of enabling technologies. There should be a cooperative effort among NASA, industry, and academia to research complicated flow physics with the goal of predicting, modeling, and controlling such flows. Research should combine theory, careful experiments and CFD; duplication of the results by another technique or in another facility, as recommended by the U.S. Transition Study Group, is also desirable. NASA is an appropriate organization to encourage flight testing of enabling technologies and the development of advanced diagnostic instrumentation for nonintrusive testing. NASA is also an appropriate organization to provide high Reynolds number, quiet testing opportunities, as well as the most current computational facilities and techniques.
To implement the above concepts, a better understanding of flow physics is required. Intensified efforts to develop useful, accurate engineering models that can be used for design are particularly necessary. This work, although seemingly basic in nature, must be actively pursued and must include companion theoretical, computational, and experimental efforts conducted under careful, well-documented conditions. A better understanding of flow physics will also afford the opportunity for effective use of flow control. Issues to be addressed should include implementation, reliability, and maintenance of the mechanical systems, as well as implications of the loss of the system in flight. Continuing advances in CFD, as well as high Reynolds number, low-disturbance experimental capabilities, are also needed in support of the evaluation and design of these concepts.
NASA needs the following resources to accomplish the foregoing:
high Reynolds number facilities (simulate flight),
low-disturbance freestream facilities (simulate flight),
full-scale Reynolds number flight research capability,
nonintrusive instrumentation,
faster and bigger computers for flow physics (model development),
more efficient CFD, faster algorithms and grid setups,
companion theory/computation/experiment efforts (validation and guidance), and
NASA/industry/university cooperation on appproximately equal levels. NASA's role in the university training of future engineers through research funding must never be overlooked.
The Europeans are using much of the technology described above - particularly LFC - even though by U.S. standards the technology is often untested or unproven. On an overall technical basis, current European offerings are equal to those of the United States. However, because Europeans are incorporating the technology faster than the United States is, their rate of improvement is significantly greater. In particular, within the BRITE EURAM consortium involving government, industry, and academia in Europe, a highly organized effort has been developed to advance the state of the art for laminar flow engineering design capabilities. Their efforts combine the best talent and facilities in all of Europe to develop and validate transition prediction tools for integration with industry engineering design methods. (BRITE EURAM efforts address all other aeronautical disciplines as well.) Although progress is being made in the United States on understanding boundary-layer transition physics, no similar design-tool-focused effort is being funded here. The problem in this country is not lack of opportunities, but rather the lack of priorities on resources.
Supersonic Aircraft Cruise Performance
U.S. expertise in supersonic aircraft performance stems from the earlier Supersonic Transport (SST) program, the ongoing Phase I High Speed Research Program, and various high-speed fighters. In cruise, the recognized promising technologies for the HSCT are the same as for the subsonic case: hybrid laminar flow control by suction and pressure gradient. Unlike the subsonic speed regime, this technology has not been demonstrated in flight. The benefits of lower drag (as described in subsonics) and thermal requirements associated with laminar flow can be realized in this flight range as well. In addition, technology advances in the fundamental understanding of high Reynolds number effects on leading-edge vortex formation and the ability to eliminate unacceptable characteristics, such as low speed pitch-up, could allow the utilization of highly swept, high-performance wing planforms. This would provide substantial increases in cruise lift-to-drag ratios over the currently favored planform concepts.
Laminar Flow Control
At present, our knowledge of high-speed transition is even less developed than our knowledge of subsonic flows; designs depend on existing theories that are not compatible with design and optimization tools (as described in subsonics). In the supersonic range, however, the effects of freestream conditions (e.g., freestream turbulence and noise) and surface roughness may be more severe than in subsonic flows. Efforts are under way to understand and predict the transition behavior in supersonic flight. The exciting progress that has been made in subsonic laminar flow technology required nearly four decades of concentrated, if not continuous, effort. Although some of the subsonic lessons learned may apply to supersonic laminar flow challenges, it would be overly optimistic to expect supersonic laminar flow technology to mature after a few years of effort.
<#FROWN:J74\>
Laser/Vision - 100% Inspection Systems
The growth of automation requires that most fastener manufacturers implement methods for providing defect-free material to their customers
The demands for stricter quality control have shifted from a high level AQL (Acceptable Quality Level) to three parts per million or zero defects product levels.
The implementation of sophisticated statistical process control has helped to insure quality of manufacture, but does not address product mix, and foreign material contamination as a result of plating, heat treating, packaging, and other bulk processes.
A growing trend for high-speed sorting and 100% gaging is emerging within industries because of the high levels of automation employed in manufacturing today. The human factor will always be present regardless of the SPC or quality control levels adopted for a particular manufacturer. The same customer that was talking AQL yesterday is demanding zero defects today at a reduced price.
The non-contact laser gaging technology available today has been in place for 20 yrs. Only the appliance, automotive, and aerospace industries have sought to minimize downtime by inspecting key attributes of material used in automatic assembly processes. A task originally performed as incoming inspection by the end user is now demanded of the supplier.
Solving the quality control problems for today's manufacturer requires systems designed with the flexibility for large product ranges, simplicity of setup, and the ability to work in harsh production environments.
Perhaps the most over-looked area regardless of the system chosen is the maintenance and level of personnel to perform that maintenance. The laser sorters utilize a single card to perform all the data processing relating to the inspection tasks. An operator is capable of correcting most all<sic!> problems with this type of inspection equipment.
Once the decision to purchase a system for 100% inspection is made, a part packaging system integrated with the inspection element will eliminate any potential mixing due to the human factor or bulk packaging. Several technologies are offered to assist in solving specific problems of the various industries.
Our Vision Inspection System is an automatic, non-contact, stand-alone inspection and measuring system. The standard version with the appropriate options is capable of simultaneously making both two-dimensional gaging measurements and surface flaw analysis of each field of view (FOV) as well as real-time image enhancements. This system is an easy-to-use, menu-driven system. The menu system is a flexible job setup tool with screen prompts to assure correct entering of parameters. No programming experience is required to learn the systems menu. The menu allows the operator to run the system on a turnkey level. Also, if new tolerance or inspection data are required, they can be inserted through the menu.
The typical system configuration includes an inspection module with the specific application programmed, optics, and camera(s), light source, part detector and associated cabling, mounts, brackets, etc. The camera is aligned to view one specific inspection point in the material handling system. During the actual inspection process, the camera captures an image of each part as it passes this inspection point.
This system utilizes an IBM PC compatible 486 processing platform for maximum throughput. A keypad is used by the operator to select and change inspection parameters as required. The inspection tools used to interpret the image and provide orientation information are adjustable in size and position to provide the flexibility to detect many configurations without changing software.
During the setup procedure, the part detector is aligned to monitor the areas adjacent to the inspection point. During the inspection process, the detector continuously monitors the material handling system. As each part reaches the inspection point the detector sends the 'Part Detect' signal to the vision inspection module. Upon receipt of this signal, the vision inspection module actuates the image capture/analysis process. A strobe light source is used, as is necessary with a moving part, to illuminate the part. A visual image of the part is then captured by the camera.
<O_>figure&caption<O/>
A small rectangular element in the camera houses an array of tiny electronic devices called photodiodes or pixels. Their function is to detect the amount of light reflected from the part to the camera. The size of the photodiode array can range from a matrix of 256 pixels to 512 pixels (vertical columns and horizontal rows). Each of the pixels continuously transmits an analog signal to the vision inspection module which assigns a value of either 0 or 1 to the binary image acquired.
The vision inspection module then transfers the binary values to the digital memory. The module can now inject the binary data into a variety of complex mathematical routines. These can perform the following analytical operations on the captured image: linear measurement, flaw analysis, and real-time image enhancement. The information obtained by these operations is then used for accept/reject decisions. These decisions can be sent out via RS-232, through discrete I/O.
The 'Magic Black Box' that inspects every dimension on every part does not exist, quality control personnel must identify and target specific areas for inspection and sorting equipment to be implemented in the most cost effective manner. The bulk handling for plating, heat treat, and packaging will remain the major areas of part contamination due to high volumes and materials handling methods.
Safe and Fast
In the 60's scientists used a laser and a passive corner reflector to measure the distance from the earth to the moon within six inches. Since then the technology transfer program helped put the laser to work for non-aerospace uses. In the past five years the development of low-power, personnel-safe gas lasers contributed substantially to their applications on production and assembly lines. More specifically, gas lasers have been adapted successfully to other tasks including automatic dimensional and surface inspection in quality control.
Lasers are fast, reliable, and non-contact. Properly applied, the automatic laser system can completely eliminate manual inspection and reduce test times by as much as 90%. Inspection speeds can exceed 600 ppm. Studies supporting automation show that human inspectors can miss as many as 20% of defective parts when operator fatigue sets in. In contrast, the use of automatic laser inspection eliminates most of these shortcomings.
Operating Principles
The beam configurations generated by various existing laser-based systems include static spot, circular spot, linear scan, and multiple spot. Because the geometry of the laser beam determines the inspection system capability and its ultimate use, laser systems are classified according to the type of their beam.
In operation, the transmitted laser beam interacts in various ways with the target being inspected. Through this interaction the beam at the receiver input will be encoded by various target parameters such as its dimensions, surface quality, etc. The receiver extracts or decodes this information and uses the resulting data for making decisions in inspection. In particular, consider the several ways that an incident laser beam can be encoded by its target and produce different detected signals at the receiver. It can be any of the following:
Sequentially interrupted - the detected signal is digital. It is either on/off or a go/no-go type. Application is in simple length sorting.
Reflected - this results in an analog detected signal whose amplitude and duration are used in the analysis of the inspected part. This approach is used basically in areas of surface flaw detection.
Shadowed - this produces a digital signal which is used for more accurate dimensional gaging, for example, where the sequentially interrupted method becomes inappropriate.
Transmitted - when detected, this is either a digital or an analog signal and is basically used in applications where the material to be inspected is clear or transparent. Using transmitted signals, parts can be inspected for surface defects, missing operations as well as dimensional accuracy. For example, threaded fasteners can be searched for surface problems, missing internal or external threads, damaged threads, or even mixed sizes of threads. The laser beam is scanned in a vertical plane so that it 'walks' down the threads. At the same time, the horizontal beam determines the location of the part under inspection. The inspection of the peaks and valleys in the thread creates a pulsing return signal that is picked up by the receiver/detector. The signal contains information on the quality and size of the thread.
Manual inspection methods using mechanical gaging still hold their own in accuracy and resolution capabilities. Thus, even though a laser-based inspection system can gage dimensions with an 0.0004" accuracy, this figure can be exceeded by traditional methods. But for industrial applications where this lower gaging tolerance is adequate, laser systems win hands down by virtue of their highest speed. Today, laser inspection systems are used in industries making fasteners, bearings, automotive and aircraft components, glass and pharmaceuticals. Laser inspection functions range from simple dimensional gaging to providing feedback signals for automatic process control.
System Components
The basic non-contact laser inspection system consists of several basic components: a laser used as a target illuminating source (the word laser is an acronym of Light Amplification by Simulated Emission of Radiation); a receiver to sense the optical radiation that carries the required dimensional information; and processing circuitry to extract this information in preparation for further use. This radiation can be reflected, scattered, or shadowed by, or transmitted through the target. Other parts include lenses, mirror, prisms, polarizers, choppers, and filters.
In practice, the gas laser provides high beam brightness, good beam collimation, small beam diameter, and long beam life. These laser qualities are used for very high speed, tight tolerance, and non-contact inspection on a wide range of parts and materials.
Parts made of metals, plastics, glass, rubber, or wood can be inspected by a laser without tool changes usually required with mechanical methods.
Functions Performed
Almost all laser system suppliers provide equipment to perform some combination or all of the following functions:
Gaging - most laser systems are used for gaging length, width, height, diameter of parts, and combinations of these at high speeds. A benefit of 100% part size qualification is uninterrupted production.
Operation Voids - the laser system will inspect parts for missing operations such as threads, slots, and holes. Or it will locate specific indicators to help in aligning parts for assembly operations. A laser system is used in the assembly of an automotive thermostat to first locate a scribe mark on one of the parts. Then using the mark as a reference, to position the part for a subsequent staking operation. The laser is used to define the exact position. This operation runs about 120 parts/hr. The laser 2 ft away positions the scribe mark within 0.001".
Surface Flaw Detection - Laser scanning capability is used to identify surface trend indicators. While tighter tolerances are possible with mechanical inspections, the laser method is much faster and more reliable and it works equally well for metallic or non-metallic surfaces.
Process Control - This function combines gaging, surface detection, and identification to provide an on-line output that can be used as a feedback signal to an automatic control system to provide corrective action. A typical application would be in weld seam inspection to generate axial information to guide the weld track and flux flow.
What of the Future?
The future of laser-based inspection seems bright. In addition to the already mentioned advantages of speed and reliability, these systemsystems are also predicable in terms of pricing. For example, the cost of a laser system can be determined accurately for planning purpose. Thus, hard figures are available to management for comparison to alternate-method costs.
For more information contact the author or Circle 259.
Developments in Self-Locking Fasteners
Some Background
Developed almost 50 yrs ago, the idea of engineering a nylon locking element to solve specific fastening problems, Figure 1, is still going strong today. This is reflected not only in growing domestic use, but in a wide range of applications throughout the world. Global acceptance of the patented process that provides a non-metallic prevailing torque element in both male and female threads has been further strengthened by the introduction and use of new technology that upgrades the performance of the self-locking process.
<O_>figure&caption<O/>
For example, a new patented nylon powder dispensing system that produces a patcxh-type locking element with more consistent performance and better reusability features.
<#FROWN:J75\>
The active double-star architecture has received considerable research attention. As shown in Figure 2.1, the architecture calls for feeder fiber to connect the central office to the RDU, which might serve around 1000 subscribers. By equipping the RDU with switching capability, the RDU electronics direct individually selected channels to each home from the many signals multiplexed on the feeder cable. Multiplexing the signals between the central office and RDU reduces costs relative to the switched-star alternative in which individual channels are switched only at the central office for transmission over separate fiber cables all the way to the home.
The first proposals for an IBN assumed an active double-star architecture due in part to its resemblance to the current DLC system. The telephone industry continues to cite this architecture as a likely approach [Shumate, 1989], and much of the ongoing effort to develop prototype IBN components assume this architecture, although other architectural alternatives are now receiving more serious consideration.
Passive optical networks (PONs) are another class of networks now receiving a large amount of research attention. PON architectures have single fibers emanating from the central office, and these fibers fan out via passive optical splitters similar to a tree-and-branch topology. In this way, PON networks achieve a high degree of shared plant throughout the network. The lowest branch in the distribution tree connects to the individual homes. Figure 2.1 shows an example of a passive double-star architecture in which the RDU now houses an optical power splitter.
The optical power budget determines the number of successive splitting nodes in a PON network. Larger bandwidth signals have smaller power budgets, so broadband services cannot be split as often as narrowband services. This leads to the same problem faced by hybrid networks to provide a mix of narrowband and broadband services. Again, WDM or coherent transmission techniques would have to be employed in the future to increase the bandwidth of the system.
An advantage of PON networks is that they provide a transparent path between the central office and the subscriber. This provides a degree of flexibility because the architecture is dependent upon the power budget and not the transmission format. In addition, the use of passive devices reduces the number of active components throughout the network, thereby decreasing the number of potential faults.
There have been few proposals calling for a bus topology in the subscriber loop. One problem is accessing the information on the bus. If the signal is tapped by bending the fiber and detecting the light that escapes - called a nonintrusive tap - then the system provides very little flexibility for evolution to new services because the taps are optimized to one service bandwidth. Another method is to receive and retransmit the information bus at every node. This doubles the number of optical transmitters and receivers in the field, raising more concerns about the environmental protection, reliability, and maintenance cost of the remote nodes. However, one could consider the passive double-star as a bus topology with a passive splitter as the only node. Configuring the network as a double-star rather than a bus simplifies maintenance by reducing the number of nodes. Also, there is less excess signal loss from a single 1:n splitter than from a succession of taps on a bus, thus allowing more subscribers to be served within the available power budget.
In summary, an all-fiber, active double-star network is still under consideration as a viable option for the subscriber loop. However, the number of competing alternatives is growing. The development of PON and hybrid architectures have served as milestones to shift attention away from the active double-star to a richer variety of options. PON networks take more advantage of the properties of optical transmission, whereas network hybrids offer an interim solution before more fiber can be placed in the network.
Today's Cable Television Networks
Cable operators install a tree-and-branch architecture using coaxial cable to provide distributed video services, as shown in Figure 2.4. The headend receives video signals from local studios, over-the-air broadcasts, or microwave and satellite sources and then combines and retransmits these signals over the trunk cable of the network.
<O_>figure&caption<O/>
Portions of the signal on the trunk are split off every few hundred meters to feeder cable, where in turn the signal is split to drop cable to directly serve the household. Because these systems use the standard NTSC AM-VSB transmission format, a television set used for over-the-air reception can be connected simply to the cable system through a simple frequency converter box. This makes for low entry and exit costs to cable services.
The high transmission and splitting losses that arise from a coaxial cable tree-and-branch network require the placement of amplifiers throughout the network to boost the signal. Coaxial cable introduces transmission losses of about 1 dB every 30 meters (optical fiber introduces about 0.1 dB every kilometer). This loss and the amplifier gain determine the amplifier spacing. Given a gain of about 20 dB, trunk amplifiers are typically situated about every 600 meters. Signal quality also declines as the number of amplifiers in cascade increases, because noise and distortions accumulate with each amplifier. To maintain signal quality, no more than 30-40 amplifiers are generally placed in cascade. In addition, the series connection of amplifiers limits the reliability of cable service. If one amplifier in the cascade fails, then all subsequent network branches downstream of the failure lose their signal, resulting in a loss of signal over a potentially large area.
The tree-and-branch approach is well suited to the characteristics of distributed video services. Each subscriber receives the same group of video channels, which favors an architecture with a high degree of shared network resources, such as the tree-and-branch topology. The bandwidth of the amplifiers limits the number of channels carried by the system. Improvements in technology have increased amplifier bandwidth from around 200 MHz (30 NTSC channels) to 550 MHz (80 NTSC channels) today. The bandwidth of the coaxial cable is about 1 GHz (150 NTSC channels) and represents the upper bound of capacity on typical cable systems.
Fiber Backbones
The cable industry is also exploring architectures that would incorporate fiber into their networks [Chiddix and Pangrac, 1988]. In assessing fiber alternatives, cable operators enjoy the advantage of already having a broadband connection to the subscriber. Fiber installed in any portion of the cable network can immediately improve broadband services. Consequently, cable industry proposals are exclusively hybrid networks and do not include plans to build something like an all-fiber IBN in the near future.
A common approach for adding fiber into cable television networks is the fiber backbone strategy. The idea is to replace the primary trunk lines of a current coaxial cable system with fiber backbones that extend to within a few hundred feet of the subscriber - the remaining distance to be connected using the existing coaxial cable. The fiber backbone could entirely replace the trunk line, thereby eliminating all the trunk amplifiers, or replace a significant portion of the trunk line, thereby shortening the number of trunk amplifiers in cascade to a small number.
Deployment of fiber backbones reduces the number of trunk amplifiers in cascade between the headend and subscriber. Using fewer amplifiers in cascade improves network reliability and picture quality, increases system capacity, and provideprovides greater network flexibility for offering new broadband services. Network reliability is an important attribute, as frequent loss of television signal resulting from the long amplifier cascades continues to be a common consumer complaint about current cable television services. The increase in system capacity made possible by fiber backbones could enable cable systems to carry HDTV signals over the cable network without having to reduce the number of offered channels.
2.4 MARKET STRUCTURE FOR THE SUBSCRIBER LOOP
The previous sections described a framework characterizing the mechanisms of network evolution, starting with the demand for network-based services. Network planners assess the technology alternatives and select the approach best suited to meet current and future demands. The cost characteristics of the chosen technology indicate the appropriate market structure - the most efficient form of industrial organization of firms for production - for the network-based services. Typically, network-based services exhibit economies of scale in production. However, in a multiproduct environment, natural monopoly requires the dual presence of economies of scale and scope in the production of services.
Public telecommunications networks provide the communications infrastructure vital to the activities of everyday life, and the need to establish policies favoring the development of this infrastructure is therefore clearly linked to the notions of economies of scale and scope. That is, presumably the justification for a public communications network, and regulation of the network in general, stems from strong economies of scale and scope arising in the use of such a resource. Telephone companies propose that IBNs will serve as the future communications infrastructure for society by providing a diverse set of services based upon the strong economies of scope inherent in the services provided by fiber-based networks.
2.4.1 The Importance of Economies of Scope
Without any regulatory barriers, a network operator will offer a new service if that service can be offered at a profit for an incremental cost lower than that of competitors. Economies of scope between new and existing services would lower the incremental cost of the new service. As pointed out earlier, the strength of economies of scope between narrowband and broadband subscriber loop services should ultimately determine the success of IBNs. Whether an IBN will achieve economies of scope depends upon the transmission characteristics of the network services and the technological alternatives to transport these services separately.
Figure 2.5 qualitatively illustrates this point by plotting the hypothetical cost functions for three transmission technologies versus three levels of network-based services. Assume that the hypothetical cost functions represent the costs of building a network using that transmission technology. The cumulative levels of service are narrowband, distributed video, and switched video. The shaded portion of the graph shows the area in which cumulative economies of scope are present; its boundary represents the combined cost of separate networks to provide each service category. Any cost curve passing through the shaded region exhibits economies of scope between the service categories, because the cost of an integrated network is less than the cost of separate networks.
<O_>figure&caption<O/>
The three technology curves depicted in Figure 2.5 are presented in part to motivate discussion, but also because they are indicative of the actual cost characteristics of the representative technologies [Johnson and Reek, 1990]. The curve for copper technology illustrates that this technology provides the lowest cost alternative for narrowband services, but as the capacity requirements of the services increase to broadband levels, copper wire pairs quickly becomes uneconomic versus alternative technologies and provides no economies of scope for these services. The curve for hybrid networks shows that these systems provide economies of scope between narrowband and distributive video, but economies are lost in upgrades to switched video status. The curve for an all-fiber network suggests that these systems offer no economies of scope between narrowband services and distributed video services (that is, a fiber network carrying these services is more expensive than separate copper and coaxial cable networks), but if switched services are incorporated into the network, then the fiber alternative captures more economies of scope than the other alternatives.
Although somewhat contrived, these curves do succeed in conveying the complexity of the network evolution problem. Different technologies are most efficient for transporting different combinations of services. No simple solution exists that will dominate over the entire range of services. Network planners must plan a network that can deliver an efficient mix of services given a highly uncertain environment regarding future technology developments, market opportunities, and government regulations.
One aspect of the deployment of IBNs that is not considered in the analysis is any potential economies of scope between businesses and residential applications. Businesses may have a much higher demand for broadband applications such as high speed date, electronic publishing, or video imaging services. Strong demand for broadband services by business users located in residential areas could provide sufficient economies of scope to facilitate the introduction of fiber into these areas.
<#FROWN:J76\>
Chapter IX
The Insulating Refractory Product Line
APPLICATIONS AND APPLICATION CRITERIA
Thermally insulating refractories function by providing stagnant or 'dead' gas space, which is to say they contain large volume fractions of voids. Since it is impossible to build closed-cell structures into high-void-volume ceramics, these materials are all 'open': susceptible to permeation and saturation by hot process liquids and to chemical attack by aggressive gases. It follows that they are not willingly exposed directly to liquids of any kind, nor to condensible vapors, nor to gases of more than minor chemical reactivity.
Ergo, corrosion resistance is downgraded as an index of merit. The prime criterion for material selection is refractoriness and dimensional stability sufficient for the assignment. The first property needed for insulating refractory qualification is accordingly the service temperature limit, which is related to composition, sintering temperature, and void volume.
The composition classes of Table VIII.1 turn out to be more than ample. Virtually all refractories in the insulating product line are oxidic, for the obvious reason that nonoxides are on the whole innately efficient conductors of heat. Relief from the necessity of resisting corrosion by liquids eliminates some of the more expensive oxidic types and their composites.
Insulating refractories are grouped at the bottom of Table VIII.2, under Porosity -now expanded to mean void volume fraction. But Table VIII.2 does not address the microstructures of insulating materials nor the manufacturing methods employed specifically to create them. Some further subclassifications are needed. To put those in perspective, let us first examine where and how insulating refractories are used.
Duplex Linings; Steady-State Usage
There are two reasons for interpolating an insulating layer between a hot working lining and the 'outside.' These are: (a) to cool the back face, e.g., to preserve the mechanical integrity of an enclosing metal shell or for reasons of safety outside a wall or roof; and (b) to reduce the heat flux J through the lining and hence improve process fuel economy. Both motives may apply at the same time, though the second usually predominates.
In the simple case of a plane wall at steady state, where a hot-face temperature T<sb_>h<sb/> is fixed by a given operation, then:
<O_>formula<O/>.
This equation is set up for a working lining 'w' of mean thermal conductivity <*_>unch<*/><sb_>w<sb/> and thickness Z<sb_>w<sb/>, an insulating lining 'i' of (low) mean thermal conductivity <*_>unch<*/><sb_>i<sb/> and thickness Z<sb_>i<sb/>, and metal shell 's' of (high) mean thermal conductivity <*_>unch<*/><sb_>s<sb/> and thickness Z<sb_>s<sb/>. T<sb_>h<sb/> is the (fixed) temperature of the working hot face; T<sb_>i<sb/> is that of the interface between linings 'w' and 'i'; T<sb_>b<sb/> is the refractory back face temperature or that of the interface between lining 'i' and shell 's'; and T<sb_>o<sb/> is that of the outside of the shell. J<sb_>o<sb/> is the heat flux to the 'outside,' existing by virtue of water-cooling or forced or convective air-cooling of the shell.
The equation is solvable, given all k's, once T<sb_>o<sb/> or J<sb_>o<sb/> is fixed. This may be done by applying practical criteria to J<sb_>o<sb/> with a known relation between this T<sb_>o<sb/>, or by arbitrarily limiting T<sb_>o<sb/> itself (e.g., to <*_>approximate-sign<*/>100<*_>degree<*/>C if water cooled, to 40<*_>degree<*/><*_>unch<*/> if exposed to human contact, or to some safe limit for mechanical integrity of the metal). Ordinarily the shell thickness Z<sb_>s<sb/> will be fixed by structural design considerations; and Z<sb_>w<sb/> may be limited by considerations of working lining corrosion or its end-of-life minimum allowable thickness, for example. If a metal shell is absent, delete the third equality and replace T<sb_>b<sb/> by T<sb_>o<sb/>. If an insulating backup refractory is absent, delete the second equality and replace T<sb_>i<sb/> by T<sb_>b<sb/>. If both are absent, delete both of those equalities and replace T<sb_>i<sb/>by T<sb_>o<sb/>. The equation is thus versatile, and comparisons may be made with it among various lining options.
Computer programs for solving this steady-state type of equation are in common use. Depending on their sophistication, provisions may be made for: (a) inserting each k as f(T) instead of estimating a mean; (b) inserting additional functions for interface conductivity or temperature drop, where applicable; and (c) inserting appropriate relations between J<sb_>o<sb/> and T<sb_>o<sb/> for various modes of outside cooling. Comparable equations are suited to cylindrical or spherical geometry.
For hand calculations, reasonable estimates can be made for the independent parameters. Perhaps the only one not self-evident is the connection between J<sb_>o<sb/> and T<sb_>o<sb/> for air cooling. A rough empirical curve for unimpeded convective cooling of vertical exterior surfaces by ambient air at about 77<*_>degree<*/>F or 25<*_>degree<*/>C, used for estimating purposes, is approximated by:
<O_>formula<O/> (J<sb_>o<sb/> in Btu/hr ft<sp_>2<sp/> and T<sb_>o<sb/> in <*_>degree<*/>F);
or, in convenient metric units:
<O_>formula<O/> (J<sb_>o<sb/> in kJ/h<*_>dot<*/>m<sp_>2<sp/> and T<sb_>o<sb/> in <*_>degree<*/>C).
This rough guide applies to a refractory cold face up to some 600<*_>degree<*/>F or 300<*_>degree<*/>C. It should do about as well for the outside of a steel shell.
For a conservative example of what thermal insulation can do, suppose the hot zone of a tunnel kiln, firing pottery, averages 1000<*_>degree<*/>C at the hot face. Assume its working refractory sidewalls and roof are 9" or 22.86 cm thick, exposed to the air outside, constructed of super-duty firebrick whose mean thermal conductivity is 9.5 Btu <*_>dot<*/> in./ft<sp_>2<sp/>hr<*_>degree<*/>F or 490. kJ <*_>dot<*/> cm/m<sp_>2<sp/>hr <*_>degree<*/>C. By entering these quantities into the above equations and solving simultaneously:
J =490 (1000-T<sb_>o<sb/>)/22.86 and <O_>formula<O/>,
one obtains T<sb_>o<sb/> =236<*_>degree<*/>C and the heat loss J =16,380 kJ/m<sp_>2<sp/>hr. Now add only about 2" or 5 cm of lightweight insulation to the outside, using a mean thermal conductivity of about 0.6 Btu <*_>dot<*/> in./ft<sp_>2<sp/> hr<*_degree<*/>F or 30 kJ<*_>dot<*/>cm/m<sp_>2<sp/>hr<*_>degree<*/>C. What will be the new T<sb_>o<sb/> and heat loss? One solves:
<O_>formula<O/>
simultaneous with the above air-cooling equation for J<sb_>o<sb/>, obtaining T<sb_>i<sb/> =804<*_>degree<*/>C, T<sb_>o<sb/> =105<*_>degree<*/>C and the heat loss J =4,190 kJ/m<sp_>2<sp/>hr. The saving in lost heat at steady state is (16,380-4,190)/16,380 or very close to 75%. If the kiln hot zone dimensions are 80 ft. by 10 ft. wide by 12 ft. high (24.4 x 3. x 3.7 meters), the total heat-loss area is about 250 m<sp_>2<sp/> and the saving in lost heat is about 3 million kJ/hour or 73 million kJ per day, or 69 million Btu per day. That is worth about $120,000 a year in 1990 U.S. dollars.
In general, interpolating an insulating refractory layer or increasing its effectiveness: (a) increases T<sb_>i<sb/> and decreases J at a fixed value of T<sb_>o<sb/>; or (b) Increases T<sb_>i<sb/> and decreases Z<sb_>w<sb/> and T<sb_>o<sb/> at a fixed value of J. These effects on T<sb_>i<sb/>, which is the cold-face temperature of the working lining, make that lining increasingly vulnerable to corrosion. Of the two effects on T<sb_>i<sb/> recited here, the first is much the more pronounced. Where corrosion is already economically limiting, in fact, the use of an insulating backup lining may be contra-indicated. Examples are in the O<sb_>2<sb/>-blown steelmaking furnaces and the lower parts of the ironmaking blast furnace.
For the same basic reason, insulation is never placed on the outside of a metal shell to decrease J<sb_>o<sb/> unless the shell temperature is already quite low, and never without proving analytically that the metal will not be heated above its safe limit. These same equations serve also in that proof.
Thermal Conductivity, Void Volume, and Bulk Density
Armed with the above means of computing and a list of qualified insulating refractory materials and their properties, the system designer engages iteratively in an approach to both material selection and the determination of Z<sb_>w<sb/> and Z<sb_>i<sb/>. The thermal conductivity of the insulating layer becomes both a qualifying and a design property; but it is not much used for classification. Suffice it for now that values of k<sb_>i<sb/> can be had from close to those of working refractories to nearly two orders of magnitude lower. This k depends on the void volume fraction, f<sb_>v<sb/>.
The void volume fraction is related to measurable quantities from which two corresponding densities are obtained at room temperature. Every defined solid has a theroretical density, <*_>rho<*/><sb_>th<sb/>, obtainable by x-ray diffraction for crystalline species or by weight and volume measurements for bulk glasses. The latter method is also used for multi-phase materials after their maximum consolidation by, e.g., fusion casting or hot pressing, etc. Sectioning and counting up of pore areas under the microscope can be used to improve this measure theoretical density.
Likewise, each insulating material comprised in part of such a defined solid and in part of void space has a bulk density, <*_>rho<*/><sb_>b<sb/>, obtained by measuring its weight and its 'bulk volume.' Measurement of the bulk volume is not always easy. In concept if not in actual practice, this is obtainable by gently shrink-wrapping or enveloping a weighed quantity of material in an infinitely thin plastic film and then measuring the volume by water displacement in a graduated cylinder or pycnometer.
Though there are standard ASTM procedures for obtaining both of these densities, the above conceptual definition avoids confusion. It is important for present purposes that the 'bulk volume' shall contain all of the void volume within a specimen, including the interstitial packing space in the case of unconsolidated granular materials.
With these two named densities in hand and denoting the void volume fraction by f<sb_>v<sb/>, this quantity is:
<O_>formula<O/>.
But in the classification of insulating refractories, this parameter f<sb_>v<sb/> is rarely stated. Instead, the bulk density is used for classification. Although industry practices vary, it is uncommon in the U.S. for the theoretical density to be reported, or even for the phase composition of the material to be disclosed. On the other hand, the precise computation of k from f<sb_>v<sb/> is all but futile anyway; k is measured empirically vs T and is reported for each commercial and research product. We shall deal with the measurement of k along with its numerical cataloguing in Chapter XI.
Working Configurations; Cyclic Usage
Dramatic improvements in processing economy have also been made by the use of insulating refractories in the working configuration -that is, directly exposed to the process environment. A first requisite is that this environment must consist only of 'clean' (i.e., relatively dust-free) gases. A second is that, since high-void-volume refractories are to a greater or lesser degree friable, the usage must entail correspondingly low mechanical wear. Numerous applications meet these two requisites, however. Some that exemplify cyclic process operation include the walls and roofs of batch driers, ovens, and heat-treatment furnaces, batch or periodic kilns, insulating lids or 'hot tops' for metal casting, the upper portions of a limited number of metallurgical furnaces, some kiln furniture, and much hot-gas ducting. Air, moist air, and clean oxidizing or neutral combustion products are generally acceptable. Redox cycling of the atmosphere disqualifies some insulating refractories, and others may not tolerate the production of soot. Chemicals vaporized from the charge are usually somewhat harmful. These have to be evaluated individually depending on chemical identity, partial pressure or transfer rate, and specific rate of attack on the selected refractory. Dust, likewise, need not always be nil but calls for careful evaluation.
What is the peculiar virtue of low-density refractories in these cyclic situations?
It is intuitive after Chapter IV that the analysis of thermal transients entails use of the thermal diffusivity, <O_>formula<O/>, where c = specific heat in kJ/kg<*_>degree<*/>C or J/g<*_>degree<*/>C and <*_>rho<*/><sb_>b<sb/> = bulk density as defined in the preceding section. Since <*_>rho<*/><sb_>b<sb/> in high-void-volume refractories range from low to very low, the product c<*_>rho<*/><sb_>b<sb/> runs correspondingly low while k does likewise, relative to these properties of dense refractories. Computerized mathematical methods of performance analysis using <*_>delta<*/> are widely practiced, but can not be demonstrated here. Instead, some approximations and manual calculations will be used to illustrate the economic benefits of insulating working linings when the temperature is cycled.
Consider a periodic shuttle kiln, again firing pottery at 1000<*_>degree<*/>C. Each charge of ware plus kiln furniture consumes 20 million kJ in firing, and an additional 20 million kJ goes up the stack if it is not recovered. The entire cycle occupies 22 hours, leaving two hours per day for charging and discharging. The cycle consists of 12 hours heat-up plus 4 hours steady-state at 1000<*_>degree<*/>C, plus 6 hours of slow cooling (burners off).
<#FROWN:J77\>
Colorfastness to Fulling
Developed in 1954 by AATCC Committee RR22; revised 1957, 1972; editorially revised 1981, 1984, 1986; reaffirmed 1958, 1975, 1978, 1983, 1989; editorially revised and reaffirmed 1988.
1. Purpose and Scope
1.1 This test method is intended for evaluating the colorfastness of dyed wool fabric and yarn to mill fulling.
2. Principle
2.1 Dyed wool test specimens, in contact with the desired undyed textiles of choice and steel balls, are enclosed in an undyed test cloth bag and fulled in a soap solution in a metal container in a Launder-Ometer by procedures varying in temperature of the bath and time.
3. Terminology
3.1colorfastness, n. - the resistance of a material to change in any of its color characteristics, to transfer of its colorant(s) to the adjacent materials, or both, as a result of the exposure of the material to any environment that might be encountered during the processing, testing, storage or use of the material.
3.2 fulling, n. - a textile finishing process in which cloth is subjected to moisture, heat, friction and pressure.
4. Safety Precautions
NOTE: These safety precautions are for information purposes only. The precautions are ancillary to the testing procedures and are not intended to be all inclusive. It is the user's responsibility to use safe and proper techniques in handling materials in this test method. Manufacturers MUST be consulted for specific details such as material safety data sheets and other manufacturer's recommendations. All OSHA standards and rules must also be consulted and followed.
4.1. Good laboratory practices should be followed. Wear safety glasses in all laboratory areas and a single use dust respirator while handling powder dyes.
4.2 All chemicals should be handled with care.
4.3 The dyes listed in this method belong to the following chemical classes:
C.I. 62106: (C.I. Acid Blue 78), anthraquinone
C.I. 42645: (C.I. Acid Blue 15), tri-phenylmethane
C.I. 43830: (C.I. Mordant Blue 1), triphenylmethane
5. Apparatus and Materials
5.1 Launder-Ometer (see 14.1)
5.2 Stainless steel specimen containers 8.9x20.3 cm (3.5x8.0 in.) (see 14.1)
5.3 Stainless stell balls, 6.3 mm (1.25 in.) (see 14.1)
5.4 Soap, neutral granular (Fed. Spec. P-S-566 or ASTM D 496) (see 14.2)
5.5 Sodium carbonate, anhydrous, technical
5.6 Phenolphthalein indicator solution
5.7 Standard dyeings (see 14.3)
5.8 Worsted test cloth with 12 effect floats (see 14.4)
5.9 AATCC Chromatic Transference Scale (see 14.1)
5.10 Gray Scale for Color Change (see 14.2)
6. Test Specimens
6.1 Form a bag by folding together a 2-gram dyed wool test specimen and a 2-gram cutting of the test cloth [approx. 7.6x10.2 cm (3.0x4.0 in.)] with the dyed piece innermost, and stitching two of the open edges. Insert ten 6.3 mm (0.25 in.) stainless steel balls and stitch the remaining open edges. To test dyed yarn, braid the yarn with undyed wool, cotton or other yarn of interest [5.1 cm (2 in.)] and tie the braid at the ends. Enclose the braid and ten 6.3 mm (0.25 in.) stainless steel balls in a test cloth bag formed as above.
7. Test Conditions
<O_>table<O/>
8. Procedure
8.1 Prepare a test solution containing 37.5 grams soap and 15 grams sodium carbonate per liter.
8.2 Place each test bag in a 20.3 cm (8 in.) stainless steel container. Add 100 stainless steel balls and 8 mL of the test solution, and secure the covers. Place the tubes in the Launder-Ometer which has been heated to the desired temperature, and run for the required time (see table above). Remove the tubes, release the covers, empty the contents into a colander, and rinse the test bags in several changes of water at 43C (110F) (until the rinse water is neutral to phenolphthalein). Open the bags, remove the balls and the specimens, and dry the specimens and the test cloth.
9. Evaluation
9.1 Evaluate the specimens by comparing them with the appropriate Standard dyeing which has been subjected to the same fulling procedure. These Standard dyeings represent the minimum fastness requirements for their respective classes for alteration in color and staining of accompanying undyed textiles.
10. Alternate Evaluation Method for Color Change
10.1 Define the effect on the color of the test specimens by reference to the Gray Scale for Color Change.
Class 5 - negligible or no change as shown in Gray Scale Step 5.
Class 4 - a change in color equivalent to Gray Scale Step 4.
Class 3 - a change in color equivalent to Gray Scale Step 3.
Class 2 - a change in color equivalent to Gray Scale Step 2.
Class 1 - a change in color equivalent to Gray Scale Step 1.
11. Alternate Evaluation Method for Staining
11.1 Staining can be evaluated by means of the AATCC Chromatic Transference Scale or the Gray Scale for Staining (see 14.5).
Class 5 - negligible or no staining.
Class 4 - staining equivalent to Row 4 on the AATCC Scale or Step 4 on the Gray Scale for Staining.
Class 3 - staining equivalent to Row 3 on the AATCC Scale or Step 3 on the Gray Scale for Staining.
Class 2 - staining equivalent to Row 2 on the AATCC Scale or Step 2 on the Gray Scale for Staining.
Class 1 - staining equivalent to Row 1 on the AATCC Scale or Step 1 on the Gray Scale for Staining.
12. Report
12.1 Report the class determined for color change or staining.
12.2 Report evaluation method used (Sections 9, 10 or 11).
12.3 For staining report whether Gray Scale for Staining or AATCC Chromatic Transference Scale was used.
13. Precision and Bias
13.1 Precision and bias have not been established for this test method.
Colorfastness to Bleaching with Chlorine
Developed in 1927 by AATCC Committee RA34; revised 1942, 1947, 1955, 1956, 1957, 1962, 1972; editorially revised 1974, 1983, 1984, 1986, 1991; reaffirmed 1975, 1979; editorially revised and reaffirmed 1985, 1989.
1. Purpose and Scope
1.1 This test method is applicable to cotton and linen textiles and mixtures thereof whether dyed, printed or otherwise colored, which may be subjected to solutions containing up to 0.3 per cent available chlorine.
2. Principle
2.1 A test specimen and cuttings of the appropriate control dyeings are washed in hypochlorite solution under controlled conditions.
3. Terminology
3.1 colorfastness, n. - the resistance of a material to change in any of its color characteristics, to transfer of its colorant(s) to adjacent materials, or both, as a result of the exposure of the material to any environment that might be encountered during the processing, testing, storage or use of the material.
4. Safety Precautions
NOTE: These safety precautions are for information purposes only. The precautions are ancillary to the testing procedures and are not intended to be all inclusive. It is the user's responsibility to use safe and proper techniques in handling materials in this test method. Manufacturers MUST be consulted for specific details such as material safety data sheets and other manufacturer's recommendations. All OSHA standards and rules must also be consulted and followed.
4.1 Good laboratory practices should be followed. Wear safety glasses in all laboratory areas.
4.2 All chemicals should be handled with care.
4.3 Use chemical goggles or face shield, impervious gloves and an impervious apron during preparation of sodium hydroxide solutions. Use in an adequately ventilated laboratory hood.
4.4 An eyewash/safety shower should be located nearby and a self-contained breathing apparatus should be readily available for emergency use.
5. Apparatus and Materials
5.1 Launder-Omeer (see 13.7).
5.2 Flat iron
5.3 Stainless Steel Cylinder, 500 mL 7.5x12.5 cm (3.0x5.0 in.) (see 13.7).
5.4 AATCC Chromatic Transference Scale (see 13.7).
5.5 Gray Scale for Color Change and Gray Scale for Staining (see 13.7).
5.6 Distilled water, pH 6.8-7.2.
5.7 Sodium hypochlorite (4-6 per cent available chlorine, pH 9.8-12.8) (see 13.2).
5.8 Sodium hydroxide
5.9 Sodium carbonate
5.10 Sodium bicarbonate
5.11 Sodium bisulfite
5.12 Sodium arsenite
6. Test Specimens
6.1 One specimen should be used for each of the types of test to be applied (four types are described below) and one specimen should be reserved for comparison with the tested specimens.
6.2 Specimens of not less than 2 or more than 6 grams should be taken. If less than 2 grams are available, make up with dyed controls or boiled-out unbleached cotton.
7. Tests
<O_>table<O/>
8. Procedure
8.1 Solutions of sodium hypochlorite (NaOCl) containing 0.01, 0.1, 0.2, 0.3 per cent available chlorine are made up and adjusted to a pH of 11.0<*_>unch<*/> 0.2 by means of the proper buffers (see 13.3).
8.2 The test specimen is thoroughly wet out at 25-30C (78-85F) in distilled water, or in event that it has been given a water repellent finish (see 13.4), it is wet out in a 0.5 per cent neutral soap solution at 25-30C (75-85F). The surplus liquor is removed permitting the specimen to retain approximately its dry weight of liquor and in this condition it is then placed in a one-pint glass jar containing 50 times its dry weight of the sodium hypochlorite solution adjusted to the required temperature.
8.3 The tightly capped jar is immediately placed in the Launder-Ometer and the machine run until one full hour has elapsed from the time that the specimen was placed in the jar. The Launder-Ometer is maintained at a temperature of 27 <*_>unch<*/> 3C (80 <*_>unch<*/> 5F) during the test.
8.4 The specimen is removed from the jar and rinsed thoroughly in cold running tap water for 5 minutes, squeezing or agitating at intervals. The surplus water is removed by any convenient means after which the specimen is placed in 50 times its dry weight of a 1.0 per cent sodium bisulfite solution of 27 <*_>unch<*/> 3C (80 <*_>unch<*/> 5F) for 10 minutes with occasional agitation. It is then removed and again rinsed in cold, running tap water for 5 minutes, with occasional squeezing or agitation. After removal of the surplus water by any convenient means, it is pressed dry between white cotton with an iron having a temperature not above 152C (305F).
8.5 If the test is used for arbitration of a dispute, then the test specimen should weigh 4 grams and no tolerances in temperature or pH are permissible.
9. Controls (see 13.5)
9.1 Bleached muslin or light weight cotton cloth covered with -
No.1 4 per cent Vat Violet BN 10 per cent Paste (CI 68700)
No.2 4 per cent Vat Brilliant Violet RK 10 per cent Paste (CI 63365) or their equivalents in any strengths of these dyes.
10. Evaluation Method for Color Change
10.1 The effect on the color of test specimens, by each of the four tests with different concentrations of available chlorine, is evaluated and the colorfastness classified by comparison with the Gray Scale for Color Change (see 13.7)
Class 5 - negligible or no change as shown in Gray Scale Step 5.
Class 4 - a change in color equivalent to Gray Scale Step 4.
Class 3 - a change in color equivalent to Gray Scale Step 3.
Class 2 - a change in color equivalent to Gray Scale Step 2.
Class 1 - a change in color equivalent to Gray Scale Step 1.
10.2 As stated in Section 13.6, it is prudent to include Control No. 2 (see Section 9.1) in the tests being made as a means to establish that the tests have been made satisfactorily.
10.3 When subjected to the four tests and the effect on the color evaluated by comparison with the Gray Scale; Control No.2 should merit the following classification:
<O_>table<O/>
11. Evaluation Method for Staining
11.1 In practical process bleaching with chlorine, no visible staining is permitted; however, if desired in testing, staining may be evaluated and classified by using multifiber test cloth (see 13.7) attached to the best specimens and classifying the staining by comparison with the AATCC Chromatic Transference Scale or the Gray Scale for Staining. The means used should be indicated when reporting the test results (see 13.8).
Class 5 - negligible or no staining.
Class 4 - staining equivalent to Row 4 on the AATCC Chart or Step 4 on the Gray Scale for Staining.
Class 3 - staining equivalent to Row 3 on the AATCC Chart or Step 3 on the Gray Scale for Staining.
<#FROWN:J78\>
The new detector was a Hammamatsu 928R 'extended red sensitivity' photomultiplier (PMT), pin-for-pin compatible with the previous detector, except for a change in bias voltage. Thus, the new detector was able to make use of the transimpedance amplifier built into the base of the pre-existing detector mount. To accommodate the single fibers used in the temperature monitor system, two ST bulkhead connectors were mounted near the input end of the PMT, angled so that the cones of light emitted by the fibers fell on the first dynode of the PMT. Between the PMT and the ST mounts was an optical path selector consisting of a stepper-motor driven shutter that alternately blocks the path of light coming from the sensor fiber or the 'reference fiber.' In the UV monitor, the position of this selector was verified by two photoelectric sensors; these had to be replaced with magnetic reed switches, since the new PMT proved quite sensitive to the wavelength of operation of the photoelectric sensors.
The grating in the monochromator was replaced with a 930 grooves/mm holographic focusing grating blazed for operation at 750 nm. This grating has an effective dispersion of 8 nm/mm at the output focus of the monochromator. The monochromator itself was actually operated (without an output slit) as a spectrograph, with two single optical fibers held at the output focus by a special mounting system (Figure 2). The fibers were potted with TraCon F-230 epoxy in 27 gauge hypodermic tubing inside 22 gauge tubing and polished, then inserted into a 0.5" inch section of 0.5" diameter aluminum rod having two holes parallel to the rod axis. The rod was held in a 1.375" long section of .625" O.D. tubing that slid into another tube directly attached to the monochromator body. Because of this arrangement, light of two different wavelengths is coupled into the two fibers; the spacing between the holes determined the wavelength separation between the light coupled into the two fibers. A slight adjustment in wavelength could also be accomplished by rotating the cylindrical fiber mount. No difficulty was encountered in precisely tuning the output to the two wavelengths of importance to the temperature monitoring measurement (828 and 856 nm  see Figure 3). Between the monochromator output hole and the fibers was another shutter assembly acting as a wavelength selector that could block either one, or both (to provide dark current readings) of the fibers; this output shutter assembly differs from the one used in the UV monitor.
As a final improvement, a general reduction in dark currents was achieved by making extensive use of baffling material in the interior of the optics module. The custom analog circuit board (incorporating the stepper motor driver, PMT power supply, and programmable preamplifier) developed for the UV absorbance monitor was used without modification.
A commercial 2x2 fused fiber coupler was used to couple light at both wavelengths to the temperature sensor fiber, and to the reference fiber (see Figure 1). The coupler was chosen because of its low temperature sensitivity. To further isolate the coupler from warm-up effects, etc., it was mounted in a separate enclosure attached to the outside of the main optical module housing. During early experiments, considerable 'drift' and sensitivity to fiber bending were observed. These effects appear to have been due to the presence in the sensor fiber and coupler of light carried by 'leaky modes.' In addition, the fiber lengths used in the apparatus are so short that the equilibrium distribution of energy between true guided modes, each of which has different bending loss sensitivity, is never reached under normal transmission conditions. This problem was essentially eliminated by the use of mode stripper/mode mixers spliced to the coupler. The stripper/mixers were composed of three sections: 1 m of 50 <*_>mu<*/>m (core/125<*_>mu<*/>m (cladding) step-index fiber, followed by 1 m of 50/120 graded index fiber, followed by 1 m of 50/125 step index fiber, all wound inside a 3" diameter holder.
<O_>figures&captions<O/>
3.2 Fiber
Since the temperature measurement algorithm employed by the system is based on changes in the intensity of optical absorbance peaks (i.e., since the system must work in a high-loss region of the fiber transmission curve), it is important to maximize the amount of light transmitted by the fiber. To this end, a large-core multimode fiber is preferred over the more common single mode rare-earth doped fibers. The fiber used in this system was a 50<*_>mu<*/>m core/125 <*_>mu<*/>m cladding graded index silica structure with a numerical aperture of 0.24, and an effective index difference <*_>DELTA<*/>n=0.02. The fiber was designed to give optimum performance for spatially averaged temperature measurements over a 12 meter path. Based on a conservative system operating margin of 20 dB total loss, and on the known loss behavior of Nd<sp_>3+<sp/> doped fibers, this means that a Nd<sp_>3+<sp/> concentration of approximately 50 ppm was necessary to achieve the desired performance. This concentration was obtained in a specially prepared MCVD preform by using a carefully controlled heated-ampoule system to deliver Nd<sp_>3+<sp/> during deposition of the core region of the preform, and resulted in losses of 1.19 dB/meter at 828 nm and 1.55 dB/meter at 856 nm. The host glass in the core was silica, containing 17% (wt) Al<sb_>2<sb/>O<sb_>3<sb/> to raise the index of refraction. After collapse, the preform was etched in hot hydrofluoric acid until the desired aspect ratio was achieved. For use in the temperature monitoring system, the fiber was drawn with a standard coating, and a 12 meter length was terminated with ST style connectors.
3.3 Electronic module
The only modification required in the pre-existing (UV monitor) electronics package was the replacement of the high-voltage lamp power supply with a low-voltage, high current supply. Power supplies to drive the stepper motor controller and DC-to-DC converter (PMT power supply) in the optics module, together with hardwired safety interlocks to prevent operation of the system of the failure of cooling gas or opening of the lid of the optics module or of the electronics module, were all left as originally designed. The remote display unit of the UV monitor system was used in the temperature monitoring system to display the calculated average temperature. The main component of the electronics module is a Pro-Log System 2 Model 10 single-board CPU, complete with MS-DOS operating system that was re-programmed for the temperature monitoring application.
3.4 Software
The software used in the distributed average temperature monitoring system is directly descended from software developed for the UV absorbance monitor. At any given time the program can be in any of the fiber different modes: RUN, STANDBY, IDLE, AUTOZERO, or TEST. Switching between modes is controlled by operator interaction or by software detection of certain error conditions. The software for each mode is responsible for determining any condition which should cause that mode to exist. Upon exit from any mode, control is returned to a dispatch routine, which determines what mode to activate next. Figure 4 shows a diagram of the overall software architecture used in the distributed temperature monitor.
<O_>figure&caption<O/>
On startup the main program enters an initialization routine which initializes the display, the analog board, and the shutters, turns on the lamp and high voltage to the photomultiplier tube, and waits a specified time for the system to warm up. It also reads some values from a parameter file that can be adjusted under program control.
When the initialization is done, the dispatch routine is entered. This routine reads the digital status lines and the operator switches and decides which mode to activate. When the selected mode is terminated the dispatch routine starts again and selects a new mode based on the same input or on newly generator error codes.
During normal operation the system uses RUN more to continuously make measurements, calculate the temperature, and output the results. STANDBY mode leaves the monitor in an inactive, but ready state. IDLE mode is similar to STANDBY mode, except the high voltage and the lamp are turned off. AUTOZERO mode is used to automatically compute the calibration constant C for the temperature calculation. This is done each time the system turned on, after the normal warm-up period. Recalculation of C will also be necessary when switching the monitor to a new fiber since the coupling efficiencies for the fibers to the optics will be changed. The 'AUTOZERO' calculation is done while holding the fiber at a known temperature. TEST mode is an interactive diagnostic routine which allows the operator to exercise all subsystems of the monitor. The user communicates with TEST mode via a portable terminal to move the shutters, read the photomultiplier, do an A/D-D/A check, calculate the temperature, do an auto zero calculation, change the calibration temperature used in the AUTOZERO routine, or perform other operations.
Embedded in the code for all five modes are routines which continuously check the monitor for various error conditions. Each error is assigned a unique error code which is displayed as a hex digit on the lower display. If unexpectedly high or low PMT readings are encountered, if the detector shutter is out of position, or if the calculated temperature is out of the expected range, only the error code is displayed. If the air flow to the optics module is interrupted, or if the lamp and PMT supply voltages are turned off, the system displays error codes and enters IDLE mode. If the cover of the module is opened, the system goes into STANDBY mode.
3.4.1 Data acquisition
In order to do a temperature calculation, six photocurrents must be measured: the fiber current for both wavelengths, the bypass current for both wavelengths, and the dark current for both the fiber and the bypass. The data acquisition routines must position the shutters to select the appropriate light path and read the output of the photomultiplier tube. A layered software structure has been designed to accomplish these tasks. The lowest layer is the hardware specific routine which reads the analog to digital converter. The highest level routine is the routine which coordinates the measurement of all six different photocurrents.
This highest level routine oversees the collection of all the data for both the temperature determination and the autozero function. Each call to this routine results in all six photocurrents being measured once and the proper value for each being assigned to its associated global variable, where it can be accessed by the calculational routines. To do this, the routine first initializes the integrating amplifier timer, then calls the shutter positioning routine in the proper sequence for each of the six photocurrents, waits for the shutters to come to rest, checks for shutter positioning errors, and finally calls the next lower acquisition routine, to get a value for the photocurrent.
The second highest level data acquisition routine handles the processing functions that are necessary to improve signal quality. Five measurements are made for each photocurrent, sorted by the routine, and the median three are summed for the final result, which is then returned to the highest level routine.
Each of these five measurements are made by a call to the next lower level routine, which uses a TTL I/O board to put an integrating amplifier through one complete charging cycle. The integrating amplifier works by using the output voltage of the photomultiplier base to charge a capacitor for an accurately known time interval, through a fixed resistor and an operational amplifier. The voltage developed across the capacitor is further amplified by a second amplifier stage and presented to the analog to digital converter, where it is read by the lowest level data acquisition routine. To cycle the integrating amplifier and make a measurement, the computer first issues a TTL signal to the JFET switch on the optics module circuit board which, in turn, connects a small resistor across the charging capacitor, causing it to discharge. This switch is then opened and an accurately timed signal is applied to another JFET switch which allows the capacitor to begin charging. At the end of the time interval, this second JFET switch is opened and the voltage is read by the lowest-level routine.
<#FROWN:J79\>
Seven
Nuclear Safety
Scientific, industrial, and political leaders involved with the peaceful development of nuclear energy understood from the very beginning that safety was essential to the success of the new technology. While sound radiation protection standards were not fully developed at the outset, the importance of isolating workers, the public, and the environment from radioactive materials was recognized. Further, it was clear that the peaceful benefits of nuclear power should be shared among nations, and that technical assistance with the development of civilian nuclear power programs might be used as an incentive to accept international safeguards, thus discouraging the spread of nuclear weapons. It was also known that nuclear energy would have to be economically competitive for any large-scale deployment to occur. However, the early developers of nuclear energy were aware that they could not predict all the problems the budding technology eventually would face.
One of those problems has turned out to be public concern over nuclear energy - a technology that scientists see as safe, nonpolluting, and capable of enriching the lives of people in many areas of the world.
President Eisenhower's 'Atoms for Peace' speech on December 8, 1953, opened the door to domestic and international development of nuclear energy. By that time, a large amount of data was already available from programs on nuclear safety sponsored by the handful of nations with active development programs. The United Nations sponsored its first conference on the "Peaceful Uses of Atomic Energy" in Geneva in 1955. No less than 200 (out of a total of 1132) papers on nuclear safety and protection of the environment, and a comparable number of papers on medical uses of radiation, were presented at the conference.
<O_>graph&caption<O/>
Radiation
Any concern over nuclear safety is based on the potential for release of radiation or radioactive materials. The design and operation of nuclear power plants is intended to prevent harmful releases under all circumstances, from normal operation to highly improbable accidents. Nevertheless, radiation is an inevitable by-product of nuclear power as well as a natural component of our environment. The fraction of our annual radiation exposure that comes from the generation of electricity by nuclear power is approximately 0.1%, which is insignificant relative to even the variations in the background of natural radiation.
Radiation and its effects on man continue to be the subjects of study by researchers and review by prestigious scientific panels. The annual average exposure from natural radiation for a typical American was estimated to be about 100 millirem per year until 1987, when it was revised upward to 360 millirem because of a new appreciation for the contribution from naturally occurring radon gas. Of the 360 millirem, 18% comes from man-made radiation, primarily from voluntary exposure associated with medical procedures. Consumer products such as smoke detectors and luminous watch dials contribute about 3% of the exposure. Fallout from atmospheric testing of nuclear weapons accounts for less than 0.3%. The natural radioactivity in our own bodies contributes approximately the same amount that we receive from medical x rays. Cosmic radiation, streaming in from outer space, averages 8% of the total.
Individual radiation doses are unevenly distributed, depending on where we live, types of houses and work buildings, types of medical treatment, frequency of high-altitude flights, and other factors. The 0.1% of the average exposure due to the nuclear fuel cycle is received almost entirely by workers in the industry; the rest of the population gets effectively nothing from the generation of nuclear power.
Most data on the adverse health effects of radiation have come from studies of the survivors of the atomic bombs dropped on Hiroshima and Nagasaki. Individuals receiving short-term exposure to radiation on the order of 100 rem (100,000 millirem) are subject to acute illness. Acute exposures of 500 rem are sufficient to kill a substantial fraction of individuals within weeks. But there is controversy in how to extrapolate back by a factor of 1000 or more to the very low dose rates associated with normal living, medical treatment, and occupational exposure. One theory is that the effect is linear, allowing a simple extrapolation to be done. A second theory is that there is a threshold below which there is effectively no damage, or the damage is such that it can be repaired by the body. A small minority holds the view that low radiation doses actually cause proportionately more damage. Regardless of which theory is correct, the radiation emitted from operation of a nuclear plant is far too insignificant to have any impact on public safety.
It might be expected that because there is such a large variation in state-to-state natural radiation exposure, the importance of low radiation doses could be inferred from state-to-state variations in the cancer rate. But the states with the highest radiation exposures happen to have the lowest cancer rates. The risk of cancer from normal radiation exposure is simply insignificant compared to such other causes as smoking, industrial pollution, and life-style.
Reactor Safety
In the United States, the government sponsored aggressive reactor safety experiments prior to the start of commercialization activities. Approximately 900 square miles of isolated desert land in Idaho were selected to be the site for the National Reactor Testing Station. One reason for selection of this remote location was so that any accidental release of radioactive materials from tests to study reactor accidents would not affect the public. Safety testing continued well after the commercial nuclear power era began, often with international partners. Other Western nations recognized the imperative of safe operation of nuclear power plants and conducted independent reactor safety experiments. The safety philosophy that evolved in the former Soviet Union and Eastern European nations was apparently less conservative, but there is a current initiative to 'upgrade' some Soviet-built reactors to Western operating and design standards. Nevertheless, some Soviet-built PWRs such as those in Finland and Hungary have excellent records for operating efficiency.
The early development period in the United States was characterized by a broad-based AEC program that included uranium exploration, enrichment, power reactors, space propulsion, waste management, biology, medicine, and peaceful applications of nuclear explosions. The parallel naval reactors program focused on production of nuclear propulsion systems. Development of safety technology was a key focus of the U.S. Atomic Energy Commission's (AEC's) civilian and naval reactor programs. Radiation-effects biology programs were carried out; fundamental and applied nuclear data were generated; a number of special safety-experiment reactors were built and operated; and several series of major reactor safety experiments were carried out. Fundamental radiation standards for all radiation application were developed.
From the very beginning, reactor designers understood that a nuclear power plant could not explode like a nuclear bomb - the hazard was from the 'ashes' of the chain reaction, i.e., the fission products. Reactor safety was based on containment of any credible release of radioactive material. The concept of 'defense in depth' became standard practice to meet the general nuclear safety objective, which is to protect individuals, society, and the environment by establishing and maintaining in nuclear power plants an effective defense against radiological hazard. First and foremost, defense in depth emphasizes the importance of preventing accidents. However, in the event of an accident, defense in depth emphasizes mitigation of damage and doing everything possible to minimize injuries to workers and the public. This philosophy has continued to evolve over the years into internationally accepted basic safety principles.
<O_>figure&caption<O/>
In 1957 the first bounding estimates of the consequences of a large nuclear accident, based on an arbitrarily large release of fission products, were published by the AEC in a landmark report called WASH-740. It was not until 1975, when WASH-1400 (better known as the Reactor Safety Study, or Rasmussen report, after MIT Professor Norman Rasmussen who headed the study) was issued, that a truly quantitative estimate was available for the probability of a major reactor accident in a U.S. nuclear plant. Over the years, the Reactor Saftey Study has been challenged by a number of critics, and new safety issues requiring investigation have arisen, but nothing has happened to change the basic conclusions of the report.
In any type of risk analysis, consequences become more severe as more and more improbable circumstances and events are postulated. For most reactor accidents, WASH-1400 showed that the most probable result would be property damage to the plant, but no deaths or acute illnesses due to radiation exposure of workers or the public.
The worst type of reactor accident is the so-called reactor 'meltdown,' which might be expected to happen once in 20,000 years of reactor operation. The study showed that there would be no detectable deaths in 98 out of 100 reactor meltdowns, over 100 detectable deaths in only 1 out of 500 meltdowns, and 3500 detectable fatalities in only 1 out of 100,000 meltdowns. In addition to deaths immediately attributable to the accident, the cloud of radioactive material released in some cases would expose a large population to small doses of radiation. Using the conservative linear extrapolation model for radiological effects, some small fraction of the population would be expected to develop cancer. The average impact would be 400 fatalities over several decades. For the worst meltdown accident considered, there would be an estimate 45,000 additional cancer deaths in an affected population of ten million people. This corresponds to an increase in the probability of an individual's dying from cancer by about 0.5%, which is significantly less than the state-to-state variation in the normal cancer mortality rate. If there were a reactor meltdown every five days, the effect would be comparable to the estimated 30,000 deaths in the U.S. each year as a result of pollution from burning coal
Antinuclear activists often talk only of the most severe postulated accident, the 1 case in 100,000 meltdown accidents, leaving the impression that all major reactor accidents would result in disaster. Clearly, with fewer than 500 nuclear power plants operating worldwide, there are not going to be 100,000 reactor meltdown accidents, or even 100. The equivalent cannot be said of other energy technologies. There have been incidents such as dam failures and excessive pollution that have caused a large number of fatalities. Perhaps the best known incident happened in 1952, when pollution from coal burning under unusual atmospheric conditions in London caused 3500 fatalities within a few days.
People are subject to all types of risks, which can best be understood in relative terms. For example, a person would have a 20,000 times greater chance of being killed by lightning than by the largest reactor accident described above. If all electricity in the U.S. were generated by nuclear power, it would represent the same risk as a regular smoker indulging in an extra cigarette once every 15 years, as an overweight person putting on an extra 0.012 ounce, or increasing the highway speed limit from 55 to 55.006 miles per hour. For any reactor accident, the probability of occurence is much less than other man-made or natural accidents with similar consequences. The relative risk of nuclear power, to either workers or the public, is simply quite small.
The type of analysis that went into WASH-1400 has continued to be refined. The U.S, Nuclear Regulatory Commission (NRC) now encourages all nuclear utilities to conduct probabilistic risk assessments (PRAs) that are plant-specific. The largest benefit from these expensive and lengthy studies has been the ability to identify the systems or operations that pose the most risk for bringing about an accident. Money for modifications and maintenance can then be effectively directed, resulting in real safety improvements. As a result, today's plants, even older ones, pose less risk than the plant used in the Reactor Safety Study.
Reactor Accidents
There have been two major accidents in nuclear electrical generating stations plus a major accident at a large plutonium-production reactor. There have also been a number of lesser accidents with no off-site consequences, a substantial number of accident 'precursors' - events that could lead to an accident if corrective action were not taken, and a few serious accidents at government-owned nuclear facilities.
<#FROWN:J80\>Floodlighting had its merits for early warning, but the rotating beam had two enormous advantages: there is a greater concentration of energy, and plan position indicators (PPIs) could be employed. The advantage of PPI display had been appreciated in both Germany and Britain since 1935; but it was only practical at much higher frequencies with a rotating beam system. The time-base of the PPI radiates from the centre of the screen in synchronization with the beam. The target appears as a spot of light in the direction of the echo at a distance from the centre that represents the range of the echo, as in Fig. 11.5. When the immediate priority of an early-warning system was satisfied, efforts in ground radar turned to CHL (chain home low), GCI, and gunlaying systems.
Narrow beams required antenna arrays large in relation to wavelength. At the CH frequencies (and also those used by the German system Freya), vertical antenna arrays, 70 m and more in height, were required to obtain the necessary discrimination - in particular to avoid the swamping of the picture by ground returns. Something quite different was required for ship borne radar, to say nothing of aircraft equipment. An October 1935 statement of the operational requirement for naval radar specified microwave equipment, not only to reduce the size of the ship's antenna, but to avoid the ship's radar becoming a homing beacon for enemy bombers. The technology to home on to microwave transmissions came only a few years later.
<O_>figure&caption<O/>
There was nothing new about VHF: the original Hertz oscillator probably transmitted on a wavelength of about 4 metres, had he the means to measure it. In the 1920s it had been thought that ordinary thermionic valves could not operate on centimetric wavelengths, because the transit time of the electrons from cathode to anode was too long. The real, unrecognized, problem was the stability of the transit time. Several exotic microwave transmitting devices appeared, some of them producing a fraction of a watt. In 1936 the 16 cm obstacle detector on the Normandie produced 10 watts. As many kilowatts were sought. One important development was the magnetron, a term used since 1921 to describe a vacuum electron tube (usually a cylindrical diode) in which plate current is controlled by magnetic fields. In the early 1930s Philips of Holland produced a magnetron which gave 80 watts on 13 cm continuous wave. A Philips magnetron was delivered to the German navy in 1933.
The difficulties due to the inability to produce power on centimetric wavelengths, and therefore to discriminate with small antenna arrays, are clearly seen in early British aircraft interception (AI) radar operating on 1.5 metres. The transmitting array was in the nose, two quarter-wavelength antennas on a wing acted as elevation antennas, and a half-wavelength dipole with director on each wing provided the azimuth array. The polar diagram is illustrated in Fig. 11.6, the display in Fig. 11.7. The signal received from the elevation antennas (after amplification) is fed to either side of the vertical display; and, similarly, the signals from the azimuth arrays to either side of the horizontal display. Thus the orientation of the target is indicated by the imbalance of the blip, about 20<*_>degree<*/> below and 30-40<*_>degree<*/> to the left in the illustration. There was no range scale, nor any need for one, because the ground return swamped the screen at all ranges greater than the aircraft height, which it sharply defined.
AI MK IV described above (and installed in RAF Beaufighters in 1940) is interesting also because it created a human requirement in terms of spatial visualization and quick thinking unparalledunparalleled in the history of navigation. As the armament was fixed forward, the object of the navigation was to get on the target's tail. If the target's range is allowed to exceed aircraft height, it is lost. The target's course can be inferred indirectly from the motion of the 'blip' down the time-base and the change in its shape. Assume that the Beaufighter is heading south. If the target is heading east, a sharp turn to the left is required at mximum power before the target gets lost in the ground return. If the target is heading west, the only hope is to throttle right back, drop the undercarriage to increase drag, and turn right, losing the blip temporarily, and hoping to come out of the turn behind the target and not in front of it. There are only a few seconds to decide, and, when the correct decision is taken, the enemy unkindly changes course! Only a few had the knack of it.
<O_>figure&caption<O/>
The desperate need for powerful microwave transmitters was met by the invention of the multi-cavity magnetron at Birmingham University by Henry A. Boot and John T. Randall at the beginning of 1940. The university had been given a contract by the Admiralty to research microwave transmission, and work had concentrated on developing a high-power version of the klystron invented at Stanford University, California. Randall and Boot were assigned the less promising task of trying to do something with the magnetron. Apparently they shared the research philosophy of Watson Watt and Wilkins, for in 1976 they were to write: "Fortunately we did not have the time to survey all the published papers on magnetrons or we would have been completely confused by the multiplicity of theories of operation."
The story is related that Randall went back to the original experiments of Hertz and, in his mind, extended the Hertz resonant ring into a cylinder with a slit in it, as in Fig. 11.8. He then saw how this could be developed into the six-cavity figure illustrated. Early in 1940 the laboratory model produced over 1 kw pulse power at a wavelength of about 10 cm. The first production model generated 10 kw, which was soon increased to 100 kw. At the end of the war there was megawatt power at that wavelength, and wavelengths down to 3 cm could be transmitted at less power.
<O_>figure&caption<O/>
That was the beginning of microwave radar as we now know it. The Royal Navy seized on the magnetron (and on related work on the klystron to provide a local oscillator for centimetric receivers) to produce the first fully operational centimetric radar in the world. From its formation early in 1941, the Admiralty Signals Establishment made major contributions to coastal and naval radar. Perhaps the improvement in AI is as dramatic an illustration as any of the potency of the multi-cavity magnetron. Ground returns could be put where they belonged, and any of the three displays shown in Fig. 11.9 were at the AI designer's disposal. In Fig. 11.9(a) the display shows the target as the eye sees it, in correct azimuth and elevation, but with no indication of range, since a 2D display cannot show three independent variables. This can be supported by a second display in which range is shown against either azimuth or elevation, as illustrated in Figs. 11.9(b) and 11.9(c).
<O_>figure&caption<O/>
A scarcely less dramatic outcome of the magnetron was terrain-mapping radar, in the form of H<sb_>2<sb/>S and later derivatives. In the early years of the war metric naval radar and ASV (aircraft to surface vessel radar) displayed coastlines, but no metric radar could effectively distinguish the features of terrain. Ten-centimetre H<sb_>2<sb/>S enabled the bomber to identify cities and some topographical features of enemy territory by night or through cloud. Later 3 cm equipment gave even greater detail. Accurate altitude above terrain for precision bombing was also provided.
Shortly after the achievement of Boot and Randall all resistance to Hitler on the continent of Europe ceased. In the opinion of many, not least the US ambassador to Britain, the defeat of Britain by Germany was imminent. Certainly Britain lacked the production resources its situation needed. The USA was not to become (by courtesy of Japan) an ally for another year and a half. Churchill chose this moment to do a curious thing. On 8 July 1940 President Roosevelt received a letter from the British ambassador to Washington offering to disclose all Britain's technical secrets to the American government without reciprocal undertakings from the US; but there was an implied quid pro quo. The letter concluded: "We for our part are probably more anxious to be permitted to employ the full resources of the radio industry in your country with a view to obtaining the greater power possible for the emission of ultra short waves than anything else."
The timing was perfect. Eleven days earlier Roosevelt had created the National Defence Research Council (NDRC), and one of its first acts was to establish a microwave committee. A British mission led by Tizard (Sir Henry at that date) departed on 30 August for Washington bearing, inter alia, the multi-cavity magnetron, described in 1946 by one enthusiastic American scientist as "the most valuable cargo ever brought to our shores".
The word 'magnificent' appears infrequently in this book; but the American response was magnificent. In October the Radiation Laboratory was founded at the Massachusetts Institute of Technology (MIT), and eminent American scientists flocked to it. The Laboratory's given priorities, agreed between the Tizard mission and the NDRC, were in order:
1. microwave AI;
2. precision gunlaying radar; and
3. a fixing system requiring no response from the ship or aircraft (see Chapter 12).
In the same month the first contracts were being settled with Bell Laboratories, General Electric, Sperry, Bendix, RCA, and Westinghouse.
In December 1940, for the first time ever, radar was fitted to an American aircraft. It was a 1.5 m pre-magnetron British ASV Mk II (see below), fitted to a US Navy PBY. Only three months later an American-designed and built 10 cm radar employing magnetron technology was flying in a US B18. At that precise date (10 March 1941) the staff of the Radiation Laboratory had already reached 140. In effect, American aircraft overflew pre-magnetron radar. When the US finally became a belligerent in December 1941, its leadership in the field was assured.
No small group of people ever win a major war; but sometimes quite small groups can prevent it being lost. It might be said that the group associated with ASV (aircraft to surface vessel) were in that category. ASV worked on similar principles to AI Mk IV except that only a horizontal display was required and the sea return, unlike the ground return, did not swamp the picture. The equipment did not require the special aptitude and skill required by Mk IV AI, but did require intense concentration. The early unsuccessfulness of the Mk I ASV in anti-U-boat operations has been variously attributed to the equipment itself, poor training or unsuitability of aircrew, and defects of aircraft and weaponry. In the quarter ending with February 1941 96 Allied ships were sunk by U-boats without loss. If that rate had continued, Britain would have quickly lost the war by attrition.
<O_>figure&caption<O/>
Figure 11.10 (based on Figure 6.1 of Bowen 1987) shows the effect of the introduction of ASV Mk II, which had the same 7 kW pulse output as ASV Mk I, but was better engineered and had a longer pulse and a lower PRF. Presumably improvements in aircrew, aircraft, and weaponry also played their part. The rise in Allied shipping losses in early 1942 seen in the figure has been attributed in part to the rich pickings for U-boats in American waters immediately after the USA entered the war, but was also partly due to the German development of ASV detectors and other U-boat counter-measures. The signal received at the target is of course several orders stronger than the echo signal received back at the radar transmitter. The range of receivers designed to give warning of radar surveillance may therefore exceed the range of the radar itself. German equipment to warn submarines of ASV and night bombers of AI became a technology in its own right.
After the introduction in March 1943 of microwave ASV Mk III, an H<sb_>2<sb/>S derivative, the U-boat became an ineffective weapon and the coffin of its crew.