The Prediction of Cultural Change: A Problem Illustrated in Studies by F. Stuart Chapin and A. L. Kroeber
Floyd H. Allport and Dale A. Hartman
I. INTRODUCTORY METHODOLOGICAL STATEMENT
One of the most interesting developments of the recent anthropological school in sociology has been the consideration of culture as an independent object of scientific investigation. To the anthropologist the culture of a human tribe or nation seems to be an intrinsically interesting phenomenon. In both its material and immaterial aspects it has been studied as more or less independent of the biological nature of man, and even of his innate psychological equipment. Material inventions, fashions, and institutions appear to some to follow an order of development or recurrence so definite as to give rise to the belief in natural laws on the purely cultural or superorganic level. Closely following this conclusion comes the corollary that prediction of the rate, levels of development, and rhythmic changes in culture is a possibility. The curve of culture phenomena being known, it may be possible to predict changes in invention, fashion, and institutions with some degree of reliability.
It is the purpose of this paper to examine the general merits of this cultural argument, and in particular to analyze the claims which have been made for cultural prediction in the two recent and interesting investigations by Professors Chapin and Kroeber, cited in footnote 1. Though differing in their conceptions of cultural units and types of index, these two researches are alike in their hypothesis of an orderly and predictable sequence in culture. In its essentials this represents an attempt to import into the somewhat intangible field of culture, material and immaterial, the concepts, techniques, and principles employed in the natural sciences. It will make our task easier if we begin not with a direct analysis of the
(308) researches cited, but with a formulation of the various possible methodological approaches to cultural phenomena.
There are four principal ways in which objects of common experience (including cultural objects) may be approached. These ways of approach represent not so much the ultimate character of the objects as our attitudes toward them. Indeed, the foundation of any method of investigation, it seems to the analysts, lies in the way in which the investigator looks at his world. The four ways of approaching experience here referred to are as follows: (a) the natural-science approach, (b) the genetic-explanatory approach, (c) the telic approach, and (d) the stylistic approach. Our first task is the description of these various approaches and the consideration of their logical interrelations, limitations, and possibilities.
a) THE NATURAL-SCIENCE APPROACH
By this we refer only to a specific kind of attitude and procedure which seems to have been present in the making of significant' scientific discoveries. That is, the attitude of looking at, or into, phenomena in such a way as to see what is there and to discover invariable sequences between one identifiable happening and another. Such sequences as are found always to occur under uniform conditions are known as natural-science generalizations, or "laws."
The examination of material in order to discover its composition and laws involves two corollaries as requisites of method. The first is that of an analytical attitude, with the reduction of the material into simpler levels; the second is the criterion of explicit denotation. We shall describe these in order.
If one follows strictly the approach just described, the environmental objects one confronts will be found to have more than one level of complexity. A behavior psychologist, for example, looks at the human organism at first as a whole. He is interested in what people do and say, in other words, in behavior at the integrated, "human" level. If he looks
(309) more closely, however, with the purpose of discovering certain generalizations as a basis of understanding or prediction, he begins to see the parts of which the organism is composed, or at least to think in terms of these parts. He begins to interpret behavior through the generalizations which can be given him by the neurologist and the general physiologist. The physiologist, in his turn, describes the. action of nerve and muscle fibers and then analyzes the cells of which they are composed, either actually or conceptually, into their organic, and finally into their inorganic, components. By the aid of generalizations in the fields of organic chemistry and physics the nerve impulse and the muscle con-traction are interpreted in the simpler and more general terms of chemical dissociation, electrical polarization, and the like. The physical chemist, in his turn, peers into such phenomena as electromagnetism and "ether conduction," seeking to identify a still more elementary plane upon which even broader generalizations can be discovered.
Our second corollary, the criterion of explicit denotation, is here offered as a condition which must be fulfilled by all objects which are to be approached by the natural-science method. In order to understand this criterion let us recall a distinction made by psychologists between explicit and implicit behavior. Explicit responses consist, for our purpose, of skeletal movements which are capable of manipulating or modifying things in our environment. Implicit responses, on the other hand, consist of abridged skeletal movements, verbal reactions, and postures which we substitute in our thinking process for explicit contacts with objects. The phenomena studied by the natural-science method are characteristically things toward which one can make some sort of explicit reaction. They are stimuli for our responses of denoting, manipulating, measuring, weighing and other discriminatory and graded reactions. There occurs the possibility of some explicit response to a natural object as a beginning of every natural-science investigation. Such investigations never begin from purely implicit responses. Something, in other words, always obtrudes itself upon our experience and presents to us a problem for study. Natural-science material is thus more than that which we can see, hear, taste, touch, or smell. There is also the possibility of altering and refining our impressions from it. This means, in the last analysis, that we can obtain from the situation a kinaesthetic (or motor) experience arising from our manipulation of the object of study. Connected with explicit denotation is the important factor of verification of our experience by others—a check which guards against hallucination in the sense fields already named and which is made possible only by our capacity for explicitly
(310) denoting the phenomena concerned. It is our thesis that entities which can be "encountered" only in a metaphysical sense or by definition, for example, "immaterial culture," do not satisfy the criterion of explicit denotation, and therefore cannot be approached by analytic reduction to simpler levels and discovery of laws at these various levels.
The significance of the natural-science approach as a possible avenue of prediction deserves special comment. All statements of natural laws are essentially statements not of effective or causal agencies, but of probabilities that, given the same conditions, certain explicitly identifiable changes in the phenomena will follow. Laws, are, in other words, merely statements of the likelihood that an event will happen. They are, there-fore, the bases of prediction. Laws are relative in degree. Probably no law as stated by natural scientists is infallible. Those laws which express the higher degrees of probability that an event will occur have the higher predictive value. Predictability, being based upon laws of varying re-liability, is therefore also a relative matter. On the whole the laws formulated in such sciences as physics, chemistry, and biology express a fairly high degree of probability and have therefore relatively high predictive value.
b) THE APPROACH OF GENETIC EXPLANATION
The second approach, which is often confused with that of natural-science procedure, is concerned not with experimental analysis and observation, but with tracing the history or genetic development of the phenomenon in question. Take, for example, the mineralogical findings in a given region. The natural-science approach would take a sample of the material and analyze it into its constituents, discovering laws of crystal formation, molecular combination, cohesion, etc., at the various levels. It would make no difference from the standpoint of this approach what had been the history of the specimen or how long a time had been required for its formation. The geneticist would look backward in time, attempting to explain by the use of laws at the various levels the course of events by which the particular mineralogical compositions of the region had come into existence. A large part of historical theory follows this method. Another example, from the field of human behavior, will show that much which passes under the name of psychology belongs to this approach rather than to the natural-science method. From the natural-science view in the study of habits, for example, we should seek
(311) for laws as mere descriptions of processes which are always found in the changes of neural resistances involved in the learning process. From the genetic approach one would be interested not in the laws of neurological changes involved in all habit formation, but in what particular habits had been formed. The interest would be, not in discovering the laws of neurological change, but in applying these laws intellectually to a study of the individual's past so as to account, through a knowledge of environmental stimuli, for the formation of particular traits. We thus see that the entire case-history method in the study of personality, as well as the more academic accounts of the modification of native responses through learning, belong in the field of genetic explanation rather than that of natural-science investigation. The process is essentially one of explanation. It gives us a body of knowledge which is satisfying not -because it represents a discovery of laws, but because it seems to combine all known laws into a coherent explanation of the origin and the development of the phenomenon.
Another fact is especially significant here. In the natural-science approach we find continual necessity for caution against accepting scientific laws as statements of efficient causes. In the approach of genetic explanation, however, this caution is forgotten and laws are assumed to involve an inner condition or a necessity in the universe, by virtue of which sequences are compelled to occur in a certain way. Since, however, we are forsaking explicit contact with nature in making this assumption, we must remember that the only authority for our interpretation in the genetic approach is its satisfying character, and not its infallible one-to-one correspondence with data that may be explicitly encountered. It therefore follows that we must expect far less reliability in prediction than from the natural-science approach. Furthermore, both historical explanations and predictions must be altered whenever further discoveries reveal more generally valid laws in the natural-science sense.
Turning now to the main question, that of prediction, there can be claimed for the genetic approach in comparison with the natural-science method only an inferior degree of reliability. Suppose that, as explanatory geologists, we are facing a cut of soft material recently made by a running stream. We might predict that this sharp cliff will eventually be eroded to a low bank, and estimate the approximate length of time required for this erosion. The estimate will hold good only providing that a vast array of attendant conditions remain unchanged, and is based upon a knowledge of "how natural laws work." But it may be that through some unforeseen chance some other natural law than the one
(312) counted on will be "brought into operation." The continual shifting of conditions, even in a field so stable as that of geology, makes the alteration of phenomena, from the human standpoint at least, an almost fortuitous matter.
If one gives up the goal of tracing the history and foretelling the future of such a complex pattern of phenomena, and takes the natural-science approach, prediction becomes fairly certain. The natural scientist in the narrow sense, would not attempt to predict the fate of a cliff or river valley, but would say, with respect to any specimen of the region, that when erosion or other changes were taking place one would always find certain definite molar or molecular processes at work capable of being formulated as precise laws. We have thus quite a reliable prediction with regard to the composition of the substance, as found upon analyses, and the behavior of its parts, but a very limited and unreliable prediction of the future pattern of the whole phenomenon which we have tried to explain by the genetic approach. We may summarize this matter by saying that prediction from the standpoint of the past of a given phenomenon, while not impossible, is relatively unreliable, because of (a) the impossibility of counting upon a repetition of the same conditions in the future, and (b) the impossibility of predicting the unusual without penetrating, through the natural-science approach, to the more elementary levels.
C) THE TELIC APPROACH
Our third approach to experience is not concerned so much with the, phenomenon in itself as with the significance of the phenomenon to the observer. This significance, moreover, does not consist merely in an interest in explaining the phenomenon to our satisfaction, but in its use for some other purpose in our life-adjustment. For what are the objects about us useful? Conversely, which objects are useful in fulfilling certain of our needs? These questions express our attitudes. In the telic approach, it is the implicit rather than the explicit significance of objects which is important. A knife and a fork have meaning not from the stand-point of an analysis of their substance into its components, with laws at various levels (the natural-science approach), nor yet from the stand-point of explaining the human activities and events through which they were manufactured (genetic approach), but from the standpoint of what one can do with them; in other words, what they imply in human life. For this reason it will be seen that the problem of reduction to simpler levels is wholly irrelevant. Our behavior is upon a single level, if indeed the notion of levels can be applied at all. There is, however, a certain
(313) similarity of this approach to that of genetic explanation. Both fulfill some sort of purpose and both are prone to employ natural-science laws as effective causes in the task of prediction. In the telic field one predicts no less than in the natural-science and genetic fields, but the prediction in this case is not made for the purpose of confirming some hypothesis or establishing a law; it is made rather for the purpose of getting some practical basis of decision or action in human affairs. There is no interest in the law or its validity in itself, but only in the use of the law in helping us to choose a course of action which will be most likely to yield us profitable results. It goes without saying, therefore, that the telic approach must rest for its predictive ability upon the formulations made by the natural scientists and can only follow these formulations more or less blindly without itself contributing to the problem of predicting either similarity or change.
From another standpoint, however, there is a peculiar type of pre-diction inherent in the telic approach. One's purpose may for the time being be regarded as a causal agent in itself. Thus Mussolini on the day before taking over the local governments (podestàs) of Italy might have predicted, because it was his will, that by the following evening he would be the dictator of citizens in their community relationships through local governmental machinery. That there is some value in this type of pre-diction no one will deny. But its limitations are equally clear. An untold number of occurrences may thwart our purpose. It will be noted also that the prediction of what will take place in the future can only be couched in telic (not explicit) terms. It might have been possible fifty years ago, for example, to predict that people in 1930 would travel through the air. This prediction would be based mainly upon the telic approach. For example, we have the desire to be able to fly and we know that modern human ingenuity and determination have overcome an increasing number of obstacles in the past. Our prediction thus is entirely in terms of fulfilling our purpose. It has no relation, however, to the explicit manner in which the purpose may be fulfilled. Thus, while we might have predicted with a certain assurance that we would fly, no one fifty years ago could have predicted just what the modern flying machines would be like. The prediction of an invention in anything more than a bare telic sense is practically impossible. The explicit prediction would be the invention itself and hence its own fulfilment.
It will be noted that as we have passed through three types of approach, we have been getting nearer to that definition of experience connoted by the term culture. The natural-science approach, which is mainly explicit
(314) and divorces the phenomenon from significance in human life other than an understanding of itself, can have little meaning in the cultural field. It will not do to argue, as some might, that science itself is a part of culture; for this is to shift one's ground so as to destroy the meaning of the natural-science approach, which is a psychological and not an objective cultural phenomenon. A scientific treatise embodying natural-science laws as principles of explanation and utility has, of course, large cultural implications. But the actual taking of an attitude to approach a phenomenon explicitly and analytically, while it may be fostered by custom and practice in various places, does not depend upon a telic or a cultural attitude toward experience. Many persons seem to have a natural-science interest in and for itself.
It is the thesis of the present writers that the telic approach is necessary in order to speak in the same universe of discourse with the cultural anthropologists and sociologists. Without implicit denotation and the significance of culture objects or patterns for human life there would be no such thing as culture. But this consideration is equal to an admission that, as one leaves the natural-science approach and enters the field of cultural phenomena, prediction becomes more and more limited and unreliable. One loses that constancy of surrounding conditions which renders the prediction of repetition of similar occurrences a reliable one, and qualitative prediction, or the foretelling of unexpected configurations, becomes extremely difficult (as indeed it is from any approach).
AN ILLUSTRATION OF APPROACHES a, b, AND c
Before proceeding to our fourth approach let us consider a simple illustration which will help to fix in mind the relationships discussed up to this point—the development of the automobile. An automobile may be studied purely as a natural-science object. The physicist, for example, would first study its larger parts under such formulations as the laws of motion and mechanics. He could then descend to the more elementary levels and study the molecular phenomena of cohesion, the laws of gaseous expansion, or the subatomic phenomena of electrical conduction. In this approach it will be observed that the automobile as such disappears. Any other specimens of the same materials in the physicist's laboratory would do as well. On the other hand, the knowledge of the natural-science laws at these various levels is an indispensable aid in inventing automobiles as telic, or cultural, objects. It will be further observed that, whereas the elementary natural-science laws are statements having a high reliability for prediction of the events with which they deal, these events have nothing to do with the automobile as
(315) a final product of human ingenuity. They do not enable one to foretell how human beings might adopt and combine these laws in automobile construction. The particular combination of parts making up a modern automobile was in other words, unpredictable before it was actually achieved.
Taking the same automobile from the genetic-explanatory approach, we may study historically the modifications of human behavior which have been involved in producing it. In such a study we should deal with prepotent or instinctive responses as modified by the development of more effective methods of transportation. The automobile becomes an extension of our locomotor behavior. We would deal also with conditioning of various manual responses to different types of materials in construction, and the mechanisms of implicit response used in thinking. We should also find the stimulation of one person by another through language; and one person working with the problem where another left off. There would be expressed, perhaps, habits of a more complex order, such as desire for aesthetic qualities. Our natural-science approach could here be applied to the human organism, and we could descend again to the biochemical processes of the organisms which are involved in learning, thinking, and constructing. Thus at any point, in dealing with the auto-mobile or the human being himself, we may take either the natural-science or the genetic approach. We should be compelled, however, to bear in mind the limitations of the latter approach in the matter of pre-diction. That is to say, we could never hope to predict accurately just what kind of object would be constructed or what lines its future development would take.
Starting again from the automobile, at any point in its development we may adopt the telic approach. Here we proceed not in the backward direction (genetic explanation) but in the forward manner, putting our-selves in place of the man who is trying to invent or perfect the object. As soon as we think in terms of a human purpose to be fulfilled and reflect on how we are to accomplish it, we take an attitude very different from that of the natural scientist or the geneticist. Invention is to be understood only through the telic approach. Without such an approach there would be no invention. This fact marks the difference between invention and scientific discovery. Our sole basis of prediction upon the telic ground is our desire and determination to perfect the invention and our confidence in our ability to do so. Since we have entered the field of objects culturally conceived, we must lose much of the support of prediction of repetition or change which is possible within the natural-science and to some extent within the genetic approach.
d) THE STYLISTIC APPROACH
We have, finally, a fourth approach to experience. It is possible to derive significance from such an object as an automobile without taking either the natural-science or the genetic approach, and still without having as our main interest a utilitarian motive. We may look at the object as something which is aesthetically satisfying, as something which harmonizes with our habits in perceptions of form, proportion, and balance. Our approach to the phenomenon is thus stylistic. We may examine the size or proportions of the hood of the automobile, the color, the form of the decorative devices, etc. Such considerations are treated by the anthropologist as a part of culture. They are also related to the telic approach, since our enjoyment of proportion and balance are probably de-pendent upon the empathic responses of keeping our balance and the like. These phenomena, moreover, may be studied by the method of genetic explanation, since we may try to understand the origin and the development of aesthetic attitudes in human beings in terms of natural-science laws.
It is, however, possible to observe these stylistic phenomena in and for themselves. And this is essentially the approach taken by Professor Kroeber in the study to be considered. His problem is that of measuring and determining the laws (if any exist) of progressive changes in the pattern of stylistic units. For material he has taken the dimension of women's garments as shown in the files of fashion magazines. The central question of Professor Kroeber's research concerns the possibility of pre-diction from this linear stylistic approach. Let us assume with Professor Kroeber that we can detach the stylistically conceived units from all other human purposes, and from all approach by way of natural-science levels —in short, that we can regard them upon a plane of their own, which he calls the "superorganic." We have to remember here that prediction means simply reliance upon a fixed trend of events learned from past experience. If, therefore, fixed trends can be established in a given stylistic measurement over a long-enough period of time, we might rely upon a continuance of that trend, or upon a reoccurrence if it is cyclical. The proof of this lies entirely in the evidence produced. It should be mentioned, however, that there are some drastic handicaps. Since we have isolated the stylistic measurement from all contact with the rest of nature, it is impossible to use the natural-science laws which are serviceable both to the inventor and to the geneticist. The case must rest entirely upon its own merits. Removing the stylistic sphere from the natural-science levels renders impossible that universal type of prediction based upon
(317) discoveries at more elementary levels. Each case must therefore be dealt with by itself. Thus the establishment of a trend or a cycle in the pro-portions of an automobile-hood design would not prove a similar trend or cycle in the proportions of other decorative features, to say nothing of proving trends in stylistic units derived from other culture objects. If the measurement of a great many of such stylistic units should establish similar trends almost without exception, we might predict with fair probability that the next stylistic unit measured would show the same trend. But the proof of this must lie entirely in the data themselves. We cannot assume it from the existence of universal laws derived from the explicit and analytical methods of the natural-science approach; for that approach is in a different world from the superorganic reality of stylistic units.
In summary, we may say that any object may be viewed from the stand-point of (a) natural science, (b) genetic explanation, (c) telesis, and (d) style. Only in the last three approaches, however, can it be under-stood as a culture object. It must be remembered that the preceding analysis is a priori, and wholly tentative. From a purely logical stand-point, however, we are led to wonder whether we may really attain in cultural phenomena an order of prediction similar to that in the natural-science approach. We are justified in at least questioning the assumption that the grounds of scientific prediction must exist in the world of cultural phenomena. Let us pass on, however, to our analysis of the investigations of Professor Chapin and Professor Kroeber.
II. CHAPIN'S INVESTIGATION OF THE NATURE OF CULTURAL CHANGE A. SUMMARY OF CHAPIN'S RESEARCH
The author deals with the origin, the development, and the modification of human culture, in a treatise, descriptive in character, which contains formulations concerning societal patterns, institutions, and the back-ground of inventions. With the aid of such documents as legislative records, university catalogues, minutes, municipal records, and patent-office registers, the author seeks to plot temporal or spatial successions in the development or spread of inventions, and thereby establish, if possible, laws describing the growth of cultural phenomena. For instance (chap. viii), curricular changes in a college and poor-relief legislation go through three phases, according to the author's qualitative analysis. There is first the formal and inelastic enforcement of existing mores which begin to come into conflict with requirements of a changing order.
This leads to the second phase, which consists of a mass of special and unco-ordinated legislation based upon trial and error. There finally evolves the third stage, one of consolidation, in which the problem is solved for the time being and successful group practices are achieved. In time, however, general social conditions again change, and the cycle begins anew. This three-stage "societal reaction pattern" is evidently put forth as a tentative hypothesis, with the implication that it may be found to exist as a general pattern throughout all phases of cultural change. The cycle is regarded as dependent upon a learning process, having its analogue in habit formation in an animal organism.
Invention is viewed as the process of putting together certain previously existing elements of culture, involving known scientific laws, for the production of a new complex or cultural integration to satisfy some human purpose. It therefore depends upon the previous development of the necessary parts. In other words, all inventions have a cultural threshold which may be analyzed into the state of readiness of each of the parts to be brought into the complex. This principle is well illustrated by the dates of development of different parts of the automobile. The same analysis applies to institutional inventions, such as the commission plan of government, in which the various parts were developed at different times and places previous to their final integration.
For our purpose the most important part of Chapin's research is contained in chapter xii, "The Growth of Social Institutions by the Accumulation of Culture Traits."Here the treatment is quantitative and definite, and tentative laws with predictive possibilities are stated. In general, the plan is to trace the development of a given invention, noting year by year the number of improvements or subsidiary inventions added until it reaches a point of stability where no further additions or changes are made. Thus, S. Simon Kuznet's data upon the number of patents issued in successive four-year periods between 1855 and 1923 in connection with the plow sulky are plotted in a "frequency cycle curve."It resembles the curve of normal or probable distribution, starting at a low point and increasing in number of inventions up to the year 1884 when the number begins to decline gradually, coming to a minimum in 1923. Chapin prefers, in the case of such data, to plot the cumulative frequency of the subsidiary inventions. "When we have done this," he says, "we have a curve showing changes in the rate of growth of the fundamental invention during its life cycle"(italics by analyst). He conceives of the fundamental invention growing slowly at first, then taking a spurt and ascending rapidly until it begins to taper off and reach a
(319) point of stability. A purely generalized diagram of the types of curves revealed is shown in Figure 1.
Passing now to the field of social institutions, Chapin analyzes Upson's data on the growth of the varied activities and functions of the government of the city of Detroit since its organization in 1824. After taking account of various criticisms and taking necessary precautions, hinging upon the identification and definition of the units employed, the author concludes that "in any event, the structure of the institution grew." The growth curve, when plotted, is similar in the first and the middle portion
to that which represents inventions in connection with the plow sulky. It does not show an analogous decrease at the upper end, however, which may or may not indicate that the institution has not yet reached the corresponding period of development. The functions of administrative units of the state government of Minnesota from the period 1858 to 1923 provided further illustrations. It was found that if subgroups, or administrative units of government, were studied separately, each taken by itself would follow the first phase of the growth curve representing all items added together.
Turning his attention to the growth of the commission form of government, the author adopts a radically different method of treating temporal change. Instead of plotting the accretions of subsidiary inventions, he derives a curve based upon the number of towns or cities adopting the
(320) commission form of government each year between 1901 and 1923; in other words, upon the rapidity with which, year by year, the idea was adopted. A similar study was made of the number of cities adopting the city-manager plan. The frequency cycle and growth curves of both these inventions, with respect to their spread by number of cities adopting them per year, were strikingly similar to those of the purely mechanical invention, the plow sulky.
Chapin is careful to point out that we are dealing in one case with the number of supplemental inventions, and, in the other, with the geographic spread of the final pattern of the invention. These latter data, he says, are "population data." One difficulty lies in the question whether or not the differences of size among the cities invalidate the procedure of adding them together as units. The author goes into a theoretical comparison of his work with that of the biologist Raymond Pearl. Pearl has described and predicted by means of growth curves such varied populations as the body of a male white rat (conceived as the multiplication of body cells), a pumpkin, a tadpole's tail undergoing regeneration, a collection of yeast cells, drosophylae in a bottle, and the populations of France, Sweden, and the United States. In all these varied forms Pearl found the same sort of growth curve to be present. Chapin compares the growth of the governments of Detroit and Minnesota with that of an integrated structure like the cells which go to make up the body weight of the white rat or the pumpkin. In other words, he places the institution of city government, by implication, in the realm of organisms.
A map of the United States indicates the spread of the commission plan from Galveston over the entire country from 1900 by intervals of six years. The year of greatest numerical gain in adopting the plan (1913) is also the year of greatest geographical spread of the institution.
Our three graphic methods of description, the frequency cycle curve, the growth curve, and the plot of diffusion, merely express in quantitative form the conclusion to which political scientists have come into respect to these new political institutions, namely, that they have both passed the crest of their popularity and are on the wane.
The author's tables also show a number of cities abandoning the two plans year by year; and he observes that the experience of abandonment goes through a frequency cycle.
In drawing conclusions from the data just described, Chapin seeks to discover whether such fragmentary results conform to any principle. With commendable caution he refuses to claim a basis of prediction in the curves derived from the meager data so far collected. Moreover:
Our quantitative analysis has not disposed of certain difficult problems connected with the homogeneity of the units involved. These problems can only be solved by many studies similar to the present one. If independent verification of our results comes from such future studies, then the trends and laws of growth suggested by this study may be found to be universal principles.
For purpose of "concise description of the phenomena" the author considers the possibility of fitting the curves of his data to mathematical formulas.
Chapin explains the dynamic aspect of the growth of an invention or an institution as follows:
To the original fundamental invention .... the attention of many persons with original minds is attracted. The result is that many efforts are made to improve upon the original pattern. These efforts bring about an increasing flood of supplemental inventions and modifications in the fundamental pat-tern. But since the fundamental invention was designed to serve a specific function, it is clear that in time the original pattern will by increment and modification attain a point of relatively perfect adaptation in meeting the situation for which it was designed. Thereafter the number of increments or supple-mental inventions will diminish.
Recalling his previous description of the societal reaction pattern, Chapin points out that the cumulative growth curve of the institution or invention is closely similar in principle. The period of slow growth at the beginning of the invention corresponds to the first phase of the societal re-action pattern in which the existing mores were enforced and few innovations attempted. The middle period of the reaction pattern, with rapid increase in special legislation, corresponds to the period of rapid rise in number of inventions in the cumulative growth curve. Finally, the period of general laws and consolidated and simplified structure of the reaction pattern corresponds to the period of diminished growth in the cumulative curve of inventions. Both these cycles begin again when social conditions change and new requirements produce a strain upon the existing mores or the existing level of the invention. Thus the "growth of institutions is not steady, but occurs in waves or pulsations."
The author has an interest in the apparent cyclical occurrence of cultural phenomena which are broader in character than the special examples hitherto cited. He describes alleged cycles of social change belonging to a number of different orders of time and distribution. Cycles of the first order relate to material culture. They may be illustrated by the business cycle, the displacement of one machine in industry by another,
(322) or the invention of a process of manufacturing. They include also larger phenomena, such as the rise and the fall of the slave system in Rome, or the rise of industrial capitalism in Europe. Cycles of the second order relate to non-material culture. They are illustrated by the rise and the fall of religious sects, the adoption of a commission form of city government, the patriarchal family, or the monarchical form of government. Cycles of the third order relate to larger cultural composites such as national cultures of civilizations. Within the span of the larger cultures of the third order, there occur numerous overlapping cycles of the first and the second order. These separate minor cycles may be spread over the various phases of the greater cycle or may be entirely completed within a single phase of the greater cycle.
The author suggests certain a priori assumptions in order to pass to a working hypothesis.
(1) It may be postulated that every cultural form (whether it be a manufacturing machine, a tool, a paleolithic implement, the commission form of government, a Family Welfare Association, a cultural nation, a fraternal order, a social club, hobble skirts, and so on) has its own law of change, with the qualification that for certain classes of social phenomena there may be somebasic and common law of change. (2) The law of each cultural form is probably cyclical and may be periodic. (3) It is possible to discover and express quantitatively (perhaps in mathematical formula) the law of its life cycle or periodic function. (4) When the cycles or periods of a number (majority?) of cultural forms of the first and second order are synchronous, we have the era of maturity of the cultural nation or group in which these cultural traits are found.
In his speculations upon a law of cyclical recurrence which may be characteristic of cultural forms as such, Chapin's work is similar to that of Kroeber, who likewise prefaces his factual research by allusions to a wide variety of cycles of a more general sort. Both authors would probably recognize upon analysis certain fundamental differences between these various types of cycle, but neither has taken special pains to discover the nature of these differences. It may be said that throughout his study Chapin has advanced his formulations not as established laws, but merely as tentative statements which may prove serviceable in guiding future research.
B. DISCUSSION OF CHAPIN'S INVESTIGATION
I. CULTURAL UNITS IN RELATION TO METHODOLOGICAL CONSIDERATIONS
We may ask whether the cultural data used by Chapin satisfy the criteria for a natural-science approach. It will be observed at once that
(323) material inventions and institutions are very different, in that the former can be approached as natural-science objects while institutions cannot. But even material inventions, when approached from the natural-science standpoint, lose their significance as culture objects. A natural-science view of these objects would reveal only the laws at the different levels which are operating in the perfected and working invention. It would not show the significance of the reciprocal action of the various parts for human uses, nor would it explain genetically how the given complex of mechanical features came to be assembled. One could not, in the natural-science view, react to the invention telically, or understand it as an object of culture. There would be, as we have seen, no satisfactory ground for prediction as to what the next modification or the final form of the invention might be. Such prediction, or prediction of new natural-science laws discovered and applied to the invention, would be equivalent to the discovery itself.
The concept of institutions as an entity to be studied empirically further violates a criterion of the natural-science method. We cannot measure that to which we cannot react, and we are not able to react to an institution explicitly. An enumeration of practices of individuals in city government which by definition comprise parts of an "institution" is really a type of measurement very different from that familiar in natural science. Such enumeration does not consist in counting similar units in the magnitude of something which we encounter. It consists, rather, in counting implicit and differing elements which seem to belong together in that they work together to serve some ulterior human purpose. The result is the measurement (if it can be called measurement) of something established rather by logical definition than by explicit denotation. It seems, therefore, that the notion of establishing laws dealing with such subjective elements of culture as city governments is quite foreign to the discovery of laws by the natural-science approach, where the attitude is one of explicit analysis rather than telic adaptation of groups of elements to human needs. This fact, of course, does not prove that such cultural laws do not exist. It only means that our confidence in them should not be based upon the prestige and past successes of natural-science methods.
It is to be noted, however, that when Chapin comes to his factual investigation he makes no attempt to predict from his curves the nature of institutions or inventions (i.e., he does not invent in advance), but merely suggests the prediction of certain quantitative features such as the enumeration of subsidiary, inventions. He does not attempt to tell what these subsidiary inventions will be, only how many there will be and the ratios of their numbers in the succeeding years.
II. ANALYSIS OF VARIOUS TYPES OF ALLEGED CYCLES IN CHAPIN'S AND KROEBER'S GENERAL HYPOTHESES
The more general type of alleged cycle is that of a civilization or national culture. One of the main problems is that of definition or identification of the phenomena. When we speak of the rise, the culmination, and the fall of the Roman Empire, what is it that is really rising, culminating, and declining? Suppose that we say it is the power of Roman consuls or emperor; that is, the number of people subject to such power and the extent to which they allow the command of the ruler to dominate their lives through institutional channels. In this case we are not measuring the power of any explicitly denotable continuing thing, but only the number of people who year by year are subservient to a man who is addressed as imperator, and the degree of their subservience in successive years. The imperator himself, of course, may change. Historians and sociologists have read into the situation a continuing entity known as the "Roman Empire" proceeding in a cycle form, through rise, culmination, and fall, from nonexistence to nonexistence. Translated, therefore, into explicit language, it will be seen that we are measuring the number and intensity of attitudes in response to symbolic stimuli rather than an institution per se. Though we have dealt here with only one phase of the Empire, namely, allegiance to the emperor, it would be possible to pursue the same method with respect to any element of the "pattern" sufficiently explicit to be measured. We should thus have countings of pluralistic behaviors rather than "measurements" of the "height of a civilization."
When we deal with the pattern of culture rather than with the political dynasty we see the subjective character of the culture object from a new standpoint. We may ask, for example, what we mean by the rise, the culmination, and the decline of the Elizabethan drama or the Hellenic culture. A number of questions arise here which demand empirical investigation. Is there a pattern of traits making up such a period which shows a certain cohesion gradually diminishing as we go outward toward the extremes of the curve, that is, toward the period just preceding and the period just following the era in question? Some objective type of definition and access to data which can be identified and tabulated is obviously required. It may be found, for example, that there is no particular modal point at which all the phenomena of the age of Pericles tend to culminate. They may instead spread themselves over a longer period of Greek history, and yet be identified in our minds with the age of Pericles through association with two or three prominent figures who
(325) attract our attention. The whole matter needs to be investigated in explicit terms before we postulate a law of universal culture cycles.
There is also a large valuational element present in the identification of culture eras, which makes quantitative treatment difficult. We are especially interested in the writings of Edmund Spenser and William Shakespeare because of their outstanding merit. Around them, at least around Shakespeare, we have, therefore, built up the concept of a literature and drama with its rise and decline. If we took instead some obscure writer halfway between Shakespeare and Spenser, we might perhaps find that he' also was the center of a certain pattern, the cohesion of whose elements diminished as we proceeded forward or backward in time. In any event we are dealing in these great civilizational complexes with a vast array of factors whose correlations we only assume and which lend themselves to actual investigation much less readily than a single mechanical invention, a governmental charter, or the width of the skirt. It would seem wise, therefore, to omit speculation upon cycles of this character, which involves elements that are so subjective in nature, as a preface to the study of phenomena affording some measure of explicit denotation. By omitting these references we shall tend to keep our approach inductive rather than deductive.
III. OF WHAT KIND OF DATA ARE CHAPIN 'S CURVES CONSTRUCTED?
Perhaps the chief difficulty in the interpretation arises from regarding a mechanical invention or an institution as a unitary thing. It is true that it is a single thing from the standpoint of name and juxtaposition of parts. That is, all the parts are said to be in the plow sulky and all the functions of city government adhere in a group of officers working in proximity. More strictly conceived, however, we have from the telic standpoint not one thing but a number of things.
In the case of the automobile, for example, the bare purpose of transportation was achieved in the first crude productions. As time went on other needs were felt. Danger to human life and increased traffic demanded improvement in safety devices such as brakes, steering gear, and horns. Thus was added to the purpose of transportation the further purposes of safety for driver and others. The purpose of thrift entered the situation and devices were created to increase economy in the consumption of fuel, oil, and electric current. The demand for physical comfort while driving led to improvements of the body and springs. Eventually there emerged the purpose of aesthetic expression giving rise to improvement of design, shape, and ornament. Finally, there is the purpose of dis-
(326) -play for raising social status. It is to be noted that these various purposes did not play rôles of equivalent importance, and that they appeared in the guise of subsidiary inventions during the course of time.
From this standpoint we would change Chapin's formulation. For the concept of a single basic invention growing to the highest degree of perfection we would substitute the concept of a finite number of purposes gradually being realized in connection with a complex culture object. An automobile from a purely telic standpoint is not a single thing, but a number of things which, however, can all be propelled, directed, and stopped at the same time and at the will of a single driver. Moreover, it is by the telic approach that the automobile can be understood as an auto-mobile. Hence it seems preferable to accept the telic view rather than the natural-science view, for from the latter we have merely a pile of steel, wood, and leather.
We believe that the multiple-purpose view that we have been developing would help us to interpret the upper portion of the author's cumulative growth curve. Since the number of purposes which can be satisfied in connection with any one telic object is limited, we find a gradual diminution of these purposes as they are satisfied one by one. Finally, none is left which is possible of satisfaction in connection with the integration of that object. In other words, the pattern of elements in the object would have to be changed radically in order to admit of the satisfaction of any further purposes. An example of such change would be the transition from automobile to aeroplane. At this point we would naturally expect the cumulative curve to flatten out.
These considerations throw important light upon the problem of pre-diction and show why prediction of the content of inventions is such a doubtful matter. We can never tell beforehand just how many values or purposes will be brought to bear upon an invention, or what its possibilities for human satisfaction may be. One stage of the invention complex may reveal the potential satisfaction of an entirely different set of purposes from the preceding stage. An example of this is the variety of uses to which radio may now be put and the devices both mechanical and institutional for securing the maximum enjoyment of these satisfactions. The reason for the appearance of prediction which accompanied the construc-
(327) -tion of the Liberty motor lies perhaps in the fact that all the purposes for which it was to be used were known and stated in advance. It was built, in other words, according to strict specifications. It was also built in a period when the perfection of motors was probably reaching the culmination of the curve.
IV. THE NATURE OF UNITS IN CHAPIN'S CURVES
The addition of successive subsidiary inventions involved in the growth of mechanical inventions or institutions raises one of the most debatable questions in Chapin's study. In referring to his analysis of the growth of a political institution, the author himself says: "We have added together such varied functions and administrative units as garbage collection, zoo, band concerts, wire inspector, forestry, ferries, continuation schools, and so- on." While conscious of the questionable nature of such additions, Chapin attempts to show that the procedure is justifiable. First, the number of increments added together is relatively large (549 in the case of the plow sulky) and in the aggregate the errors will cancel one another; second, the total is broken up (in the case of the city government of Detroit) into subgroups of relatively homogeneous units which follow the same law of growth.
It is also contended that the functions of new administrative units are woven into the texture of the old structure; for example, "Electrical inspectors help diminish fire hazard and indirectly affect the functioning of the fire department." Thus, "The concept of homogeneous units, equal and interchangeable, is seen to be over-mechanical and unsuited to the analysis of such a psycho-cultural organism as a political institution."
It is inferred further that the difficulty growing out of the heterogeneity of the units added together in the growth curves is partly more apparent than real. These units are to be thought of as extensions and elaborations of functions already potential in the expanding structure. The institutional situation is compared with growth in a living organism, since the latter is composed of different types of cells growing at different rates, but nevertheless expands and elaborates as an organism in accordance with the growth curve.
Chapin's contention concerning the large number of cases and the possible cancellation of errors does not meet the essential difficulty. The real problem lies in the impossibility of adding together heterogeneous units. One may count successive additions to the manufactured plow sulky of theless expands and elaborates as an organism in accordance with the blade, the character of the seat, etc., but one can hardly add these units
(328) in a quantitative sense unless one can find a common denominator to which they can all be reduced. Chapin seems to imply that the common denominator is the fact that they somehow participate in the entity known as the plow sulky. It would be very difficult, however, to point out what the essence of the plow sulky is, aside from the integration of its various mechanical parts. Likewise, if one abstracted from a city government the functions of mayor, aldermen, and all the bureaus, one might ask what would be left as a common denominator or as the essence of the government itself in which each of the bureaus participated.
It would seem that the common denominator really exists only in the verbal habit of speaking of all of the parts as making up the complex known as the "plow sulky" and all of the bureaus as making up what is known as "city government." Moreover, by the continued use of these parts or bureaus certain human purposes can be satisfied at the same time. It is impossible to state a telic concept of this type in quantitative terms. The chief difficulty, therefore, is that the notion of increments of growth appears to be a figurative concept which cannot be explicitly denoted. It is not the discrepancy in size which prevents the additive relation among these increments so much as the impossibility of reducing them to a homogeneous character. If one were "measuring" a complex group of apples and pears, it would not do to use apples and pears as units; nor could we say that the variations were such that in the long run large. apples would correct for small pears and vice versa. One might, however, reduce them to a common denominator in terms of weight and add the units of weight contributed by each piece of fruit regardless of its species.
This is what Pearl did in plotting his growth curves of body weights. He ignored all differences between cells except their weight, since this was a function which could be determined in quantitative terms. Following this reasoning, Chapin has ignored all distinctive features of the subsidiary inventions, basing his conception of their totality on the consideration that they are all parts of, or partake of the nature of, the invention or the institution. It is not possible, however, to measure this "par-taking" factor by itself as one would measure the weight of cells, because both the institution and the mechanical invention, telically understood, are incapable of explicit denotation. Recourse is thus had to definition. The units are defined as part of a given whole, but the whole itself cannot be explicitly encountered or measured.
It has been pointed out that growing inventions are really multiple rather than single objects. The reason why the units are not homogeneous lies partly in the fact that they are satisfactions of distinct and qualita-
(329) -tively different purposes, which since they are subjective cannot be added. It would have seemed more logical, therefore, had Chapin employed column diagrams rather than continuous curved particularly curves in cumulative form. Column diagrams would show, in discontinuous form, the number of purposes satisfied at each of the dates given on the horizontal axis. It is true that without a continuous curve it would be less easy to illustrate the growth of a continuous thing—the institution or invention —but the existence of this single thing in any explicit or measurable sense is just what we have been questioning.
In order to reduce units of invention to a common denominator, we might employ the economist's device of reducing them to some set of exchange values. Thus the cost of plowing in successive years might be taken as a fair index of the satisfaction of all purposes except the aesthetic involved in the development and use of the plow sulky. For example, we might calculate the total cost of plowing per acre in the successive years of plow-improvement. We could infer that, other things being equal, the lower this cost, the more fully each purpose (i.e., each aspect of successful plowing) was being satisfied by the use of the plow as perfected in any particular year. A curve constructed from such data would describe directly the
(330) integration of such purposes in the behavior of the individuals using a plow.
Chapin might object to such treatment as purely economic and contend that it leaves out of account the institutions or the cultural problem as such. On the other hand, we might reply that it is a truer measurement of the growth of the cultural object in its telic significance than is the counting of its parts. And it is by means of the telic approach, as we have seen, that culture can be more clearly understood. Material culture, in other words, is as fully implicit as explicit, and a curve representing use and satisfactions would seem to be at least as significant as the adding-up of mechanical details shown in patent-office records.
V. INTERPRETATION OF CHAPIN'S CULTURAL GROWTH CURVES
Two general lines of interpretation of his curves are employed by the author. First, inventions, at least institutional inventions, represent a "process of organic growth and not of mechanical increase." That is, the growth of culture objects, by implication, is considered to be analogous to the growth of organisms studied by Pearl. Second, the "process of cultural change gives body and content to the claim that the cultural process is a learning process." This statement is made directly only in regard to the cycle of the societal-reaction pattern; but the author considers this cycle to be of the same general type as other movements rep-resented by his cumulative growth curves.
In discussing the organic analogy let us consider, first, the manner in which organisms and inventions grow. In organisms all the component organs, with such minor exceptions as the sex gonads, are laid down in their characteristic form and function at birth. Growth subsequent to birth seems to consist mainly of further cell division within the respective organs. This analogy, if strictly applied, would mean that in the plow sulky or the city government all the component structures were present and functioning when the original invention appeared, although on a smaller scale or with fewer personnel. It would have been a closer analogy had Chapin compared the growth of the culture product with the development of an embryo in the uterus, provided embryonic development follows Pearl's curve of growth. The essential difference between organisms and inventions lies in the fact that organs oh the body develop according to known laws, discoverable at the various levels in the natural-science approach, whereas the invention of the parts of an institution or tool depend upon human conditioning, and upon so complex a set of circumstances
(331) that their occurrence is sporadic and very difficult to predict. All natural events take place in accordance with law, but we cannot predict what laws will be "called into play" at particular moments in the process of future human adaptation. Then, too, the parts of an invention do not have the same telic relation to the entire invention which organs have to the organism as a whole. The unity of the invention does not consist of the integration of its parts through the purpose of the invention itself, but through the purpose of a different being, man; where-as the reciprocal action of parts in an organism seems, with respect to purpose or function, to be directed solely from the standpoint of the organism itself.
But the organic analogy may be criticized on more general grounds. Either Chapin means that institutions and inventions are on the same natural-science level as biological organisms, that is, they are organisms among other organisms; or he means that they are of a different order of existence, that is, they are not organisms but somehow exhibit and are explained by the laws of organic phenomena.
Let us take the latter assumption first. If Chapin is dealing with the approach of natural science, we must suppose that the level with which he works can be resolved by analysis, like all natural-science levels, into units of a more general and simpler order. Thus we would have the invention, the institution, or the societal-reaction pattern standing as a super-organic level with formulations and laws of its own, which by natural-science analysis could be broken up into its constituent parts, human beings or their common forms of behavior. If this is the case, we then have a right to inquire what are the formulations and laws stated for the level of the superorganic itself. And these formulations, to be in keeping with the standard of uniqueness shown at each of the other natural-science levels, must be drawn in novel terms and not borrowed from the biological, the chemical, or the physical realm. What we find, however, is that the author has borrowed the terminology of the biologist or the behavior psychologist, and has stated his superorganic in terms of "organisms" and their laws of growth. Clearly then, if the superorganic fails to have a describable content of its own, we are not entitled both to describe it in terms of a simpler level and yet to insist that it comprises an independent class of phenomena within the natural-science approach.
Suppose we now adopt the alternative hypothesis, namely, that institutions are biological organisms. We are then confronted by the anomaly that their parts (individuals) could live, in a biological sense, when the institution as such had disappeared; that is, when the individuals, no longer functioning together for certain purposes, were forced into isolation. This independent viability of the parts of the "social" or "cultural organism" is in marked contrast to our knowledge of all other organisms of a highly integrated type. Parts of the human body, or of vertebrates, quickly die and disintegrate upon removal from the organism, unless they are preserved in media in which the essential materials upon which they normally depend in the individual's body have been provided. The only organisms in which the parts have true independent viability are those of a very low order, such as sponges and coelenterate colonies. These could scarcely be used as analogies or prototypes of "societal organisms"; for they lack the necessary integration and reciprocal action of parts. Thus it is fair to raise the question whether more confusion than light has not been thrown upon the curves of cultural change by referring to them as processes of organic growth.
The author's second interpretation of his curves involves the claim that the cultural-change process is a learning process. With this thesis we agree. But the learning is by individuals and not societal, as Chapin implies. The curve derived from the study of societal change during adaptation to a crisis is not a learning curve, though the learning of contemporary and succeeding individuals has probably helped to produce it. It is rather the accumulated resultant of these individual learning's as reflected by changes in the invention over a long period of time. If we were to assume that the curve represented the learning of the group, we would attribute to a supposed complex level of phenomena (societal pattern) the characteristics and laws of its units at a simpler level (human beings). This, as shown above, would amount to tautology.
Another objection to the "group' learning" hypothesis is that the curve does not represent the inventive activity of a single person, but of a number of persons of succeeding generations; and learning has no meaning as a psychological process when spread over a succession of separate individuals. To this objection Chapin replies by the ingenious argument that within a biological organism there occurs, over a period of time, a
(333) change of the materials (units) of which the organism is composed; yet the organism as a whole continues the process of learning in accordance with the usual curve of such processes.
But upon closer examination the analogy will be found unsatisfactory. Stated more explicitly, the argument is as follows: Group learning may be a continuous process although the individual learning is in separate persons who may disappear and be replaced from time to time. That this view is not unreasonable is shown by the fact that an individual's learning also is a continuous process, although the elementary or ultimate learning is carried on by the separate neurones which disappear and are replaced by others. When thus formulated as an exact analogy the comparison is seen to be untrue to fact, since it requires that single nerve cells learn. Learning, as understood by psychologists, does not apply to the activity of a single nerve fiber, but to the functional linkage of two or more neural elements so as to combine a stimulation process with a response. It is a neural-arc, not a nerve-trunk, modification—a phenomenon of the entire organism (i.e., afferent in connection with efferent portions), and not of separate cells. Nerve fibers conduct impulses, hut human beings learn.
To be true to neurological fact the analogy would have to be restated thus: Individual learning consists of the joining-together of nerve cells (which do not individually learn) into patterns of stimulation and response which constitute the true learning and which remain constant, though the separate nerve cells may change. In the same manner the group learns, and this learning persists, although the individuals, who do not learn, and who function together in such a way as to comprise the pattern of group learning, may disappear and be replaced by others. But this statement, though probably correct in its psychological reference, is untrue regarding the cultural level. In order that an individual may retain his habits over a long period it is not necessary for the separate nerve cells, as they are renewed, to learn these habits; it is enough that they are there. But in order that culture shall be continuous it is necessary that, in each replacement of human personnel, the individuals shall learn the current cultural habits (institutions and use of inventions). Otherwise culture will not develop or even persist, but will disappear. Group learning, therefore, is either individual learning, or else it is a metaphysical, abstraction. Whenever we try to interpret group or cultural phenomena in organic terms the old tautology reappears.
VI. SUGGESTED CHANCES OF HYPOTHESIS FOR FURTHER RESEARCH
Professor Chapin has opened up a new and fascinating field. His cultural growth curves, if confirmed through later investigations, offer an interesting problem for further study and interpretation. The analysts, however, would seek for working hypotheses along a somewhat different line. The difficulties which arise in trying to assign cultural phenomena to some level of the natural-science series have been pointed out. Because of the telic factor in all culture and the impossibility of denoting this factor explicitly, culture objects cannot be made the objects of a natural science approach without violating criteria which are definitive of that approach itself. The desire to predict with something like the certainty of natural-science generalizations is of course a worthy one. But it seems better to think of cultural phenomena as "indices" of human achievement in the use of natural laws for progressive adaptation rather than as objects through which one endeavors to derive laws of universal predictive value.
Cultural objects may be regarded as tools which individuals, working singly or together, construct or perfect, and which they teach individuals of succeeding generations how to use. The cultural process is the learning processes of individuals as they invent, adopt, and learn the use of cultural objects. Culture, therefore, is not an organism, or indeed anything which can be handled explicitly for the discovery of natural laws regarding its component parts or their behavior. It is simply the ways and methods which human beings use to adjust themselves to their surroundings. The forward viewpoint in this adjustment process is, as stated in Part I, the purposive, or telic, approach. The backward view is that of genetic explanation, an approach which uses the natural-science laws as conceptual explanations, helping us to understand the way cultural phenomena have come about, but which does not attempt to establish such laws. Taking this point of view, the cultural process becomes a complicated series of modifications of human behavior, involving the construction and the use of tools for an increasing number of purposes and the stimulation of one individual by another or by objects or records made by another. Principles of general and social psychology may here be of help.
In our hypotheses for further research we would recognize two general groups of factors: first, the pluralistic aspect of the mass of inventors and' users of the object or institution; second, those special manufacturers or officials having a vested interest in its progress. Let us take these in order.
So far as the initial portion of the cumulative growth curve is concerned, we would accept Chapin's own working hypothesis, namely, that the interest of many is attracted by some new possibility opened by a scientific discovery or some basal invention. Thus, an enormous set of possibilities was opened up by the invention of the electric motor and the internal-combustion engine. At the beginning improvements are slow and based on fortuitous adaptations without special realization of the "principles" involved. The sudden rise in the curve may be due to the fact that people are "getting the idea." The case may be similar with institutions. People probably "got the idea" that just as a few functions had been efficiently administered by mayor and council so other functions might be handled through municipal departments and bureaus. This sudden realization of a principle is a factor which might have been given more emphasis by the author. Other factors which may help to explain the sudden rise of the curve belong in the field of social psychology. Suggestion, social facilitation, and rivalry must be added to the usual economic motive and to the professional curiosity of the inventor. When new ideas about a mechanical invention or a political institution are heralded they become a part of the lively news of the day. Impetus is given to the efforts of individuals, not only by the crowd, but also by the face-to-face interstimulation between inventors and other interested persons. It is not outside the realm of possibility to devise techniques for studying processes of this sort.
Next we note that the special interests of those most concerned are chiefly in the direction of increasing, rather than decreasing, the number of structures and functions to be added to the invention. High officials may de-sire to extend their power by increasing the number of bureaus and personnel. Certain countertendencies may, of course, be present in the motives of some of the lower officials. In the mechanical field, the vested-interest factor involves the desire to increase business. Competition induces industrial plants and commercial organizations to employ experts and to add new features which may serve new purposes. Automobile salesmen know well the value of particular and exclusive "talking points."Sometimes rather irrelevant interests of the prospective user are brought into play to induce him to purchase, as when dash-clocks and cigar-lighters are added to automobiles. These factors of motivation are perhaps difficult to meas-
(336) -ure or estimate in their contribution to the rise of the curve, but patient investigation might in time yield a promising technique.
The declining growth in the upper segment of the cumulative curve is less obvious in its suggestion of working hypotheses. The fact that there is a limit to the number of purposes that can be fulfilled in connection with a single inventional complex has been suggested. But we have to ex-plain why this limit is gradually rather than suddenly reached. A partial explanation might be found in the diminishing returns to manufacturers of money spent upon invention and promotion beyond a certain point. The factors of social facilitation and rivalry also decline as the limit of inventive possibilities is approached. Finally, there is the distracting influence of new mechanical or institutional possibilities, suggesting inventions in other directions. Some of these factors may perhaps be susceptible of investigation by the study of individuals actually engaged in manufacture and invention.
Because it points the way to inductive and explicit studies of this character, Professor Chapin's Cultural Change is a valuable contribution. Our present criticism is directed not against the author's technique or findings, but deals only with the question whether the attempt to measure and interpret cultural growth as one would handle the explicit objects of natural science will, in practice, justify the methodological assumptions upon which it is based. But before summarizing our conclusions on this point, let us examine the second investigation we have chosen with respect to the measurement and the prediction of cultural change.
III. KROEBER'S INVESTIGATION OF CULTURE CYCLES
In his study entitled "On the Principle of Order in Civilization as Exemplified by Change of Fashion" Professor Kroeber has selected for measurement certain style traits appearing in women's full evening toilette for the period 1844-1919. The actual data were taken from style-journal illustrations. Measurements were taken of the following eight traits. (1) total length of figure from the center of the mouth to the tip of the toe; (2) distance from the mouth to the bottom of the skirt; (3) distance from the mouth to the minimum diameter across the waist; (4) length of décolletage, measured from the mouth to the middle of the corsage edge in front; (5) diameter of skirt at its hem or base; (6) maximum diameter of the skirt at any point above the base; (7) minimum diameter in the region of the waist; (8) width of décolletage across the shoulders. All these measurements were recorded in millimeters.
As a general rule the first ten figures suitable for measurement in the style journal were taken for each year. Occasionally, however, it was necessary to take a few from the last months of the previous year. The author notes that there was scarcely a year for which ten illustrations could be found in each of which all eight measurements were recordable. In some cases estimates had to be made: "The majority of the eight features observed are represented, year by year, by less then ten measurements, sometimes only by four or five."The series of data are fullest for measurement items 1, 2, and 5. The author does not print his raw measurement data in full, but gives those of three years (1859, 1886, and 1910) as samples, the first of which is reproduced in our Table I. The absolute
|* Lines 1—8 refer to measured traits as defined above, p.336; line 9 refers to distance from mouth to middle of maximum diameter of skirt (not given in Kroeber's table for 1859); line 10 to distance from diameter to waist line or of corsage (all in refersminimumat waistpoint millimeters).Figures in parentheses are estimates.|
numbers in this table, representing millimeters, were then converted throughout into percentage ratios to the length of the entire figure. That is, each item in lines 2–10 inclusive was converted into a percentage of the corresponding item in the same column in line 1. The percentages for each measurement were averaged for each year. The figures used on charts, tabulations, and in discussion are these "year-percentage averages." These are exhibited in our Table II for the selected ten-year period 1850–59 from the entire series 1841 1919. Table II gives data for the period named on six measurements in terms of year-percentage averages. No measures of reliability for the averages are furnished, nor can the reliability be estimated by the reader since the number of cases on which the individual averages are based does not appear, except for the three samples of raw data for the years 1859, 1886, and 1910. The author has plotted his data for measurement items 2, 3, 4, 5, 7, and 8 and reproduced them in the article cited.
Kroeber considers the curve of trait 5 (width of skirts) as his best case Upon the basis of these data, and of general impressions of the width of the skirt before 1844, he believes that "they do seem to suggest a certain regularity of curve, a tendency toward recurrence."He finds here "a swing of about sixty years in each direction; a period for the whole wave or cycle of a hundred or a hundred and twenty years, in this particular trait."The following prediction is made:
By 1912 the tide has once more turned-no doubt to continue now for an-other two or three score years unless the periodicity of the rhythm is accelerated by some unknown new cause or is totally broken off by an alteration of fundamental fashion, such as the substitution of trousers for skirts.
|Length of Waist||Décolletage||Width Skirt||Width Waist||Width Shoulders|
|Average (1844-1919) .||97.7||26.4||13.8||65.3||9.4||15.6|
Kroeber reaches the conclusion that "the rhythmic period for skirt length is only a third that for width; about thirty-five years as against a century." It is pointed out that the possible length of a dress is automatically cut off when it reaches the ground, but that it can shorten in-definitely.
This brings it about that when skirt length attains its maximum, it remains apparently stationary for a time, whereas at its minimum it reaches a climax and quickly descends again. It might be said that fashion clearly tries, and is prevented only by physical impossibility, to draw the bottom of the dress several inches into the ground.
It is predicted:
The child whose braids hang down her back may be reasonably sure that in the years when her daughters are being born she will wear longer dresses than her mother now goes about in; and that her skirts promise to be wider each successive decade until she is a grandmother.
On the more important of the other traits there is
some indication that the position of the waist line may completely alter, also following a "normal" curve, in a seventy-year period; and a possibility that the width of shoulder exposure varies in the same manner, but with the longest rhythm of all, since the continuity of tendency in one direction for seventy years establishes a periodicity of about a century and a half, if the change in this feature of dress follows a symmetrically recurrent plan.
Kroeber considers that he has given "reasonable evidence of an under-lying pulsation in the width of civilized women's skirts, which is symmetrical and extends in its up and down beat over a full century; and of an analogous rhythm in skirt length, but with a period of only about a third the duration" and possibilities of orderliness in some of the other traits, as we have mentioned in the paragraph above. He states that "there is something in these phenomena, for all their reputed arbitrariness, that resembles what we call law: a scheme, an order on a scale not without a certain grandeur." While he does not hold that the fashion of a future date can be written now, he at the same time maintains that
it does seem that some forecast can be made for any one basic element whose history has been sufficiently investigated; and that, when the event arrives, if the anticipation proved to have been more or less erroneous, the source of the aberration may be clear, and the disturbingly injected forces stand revealed as subject to an order of their own.
Professor Kroeber draws conclusions of still broader significance. He states that "the fact of regularity in social change is the primary inference from our phenomena." The amplitude of the periodicities is of al-
(340) most equal importance since their magnitude dwarfs the influence any individual can possibly have exerted in any alteration of costume. Rejecting the gifted individual or innovator as the source of these trends, he offers as the other alternative the operation of "superindividual principles which determine the course of social events."
These alternatives, in the opinion of the analysts, do not exhaust the possibilities. One might employ a psychological explanation. Underlying these stylistic phenomena we might, for example, point to emulation as a motive within the individuals who are fashioning, advertising, and wearing the various gowns. If this factor were shown to be strong, we might have a possible explanation for trends and even cycles. It is assumed for the moment that the existence of these fashions could not stay constant, for to do so would prevent outdoing offer people. If a trend in fashions had started in a given direction, it would be more likely to continue than to return, since to go back would be to repeat a fashion quite recently current. This would be contrary to the nature of the motive of emulation. Somewhere, however, there is a limit to such a trend. It may, perhaps, be physical, such as the ground in the case of the skirt; or it may be a conflict with some moral habit or notion of decency; or it may, perhaps, be mere inconvenience, as in the case of a skirt too wide. Here the trend would stop, turn, and begin its long swing in the other direction. Traced over a period of years the general effect would be cyclical.
Underlying such long trends as the decrease in the length of the skirt, we might find such factors as changes in the economic and social position of women, their greater participation in sports, and changes in the mode of dancing. Economic factors, such as the relative cost of certain materials, are possibly present. Explanations of this nature would avoid acceptance of the interpretation in terms of gifted personalities, which Kroeber dislikes, and at the same time make it unnecessary to postulate the operation of superindividual principles which determine the course of human events.
It is difficult to accept Kroeber's conclusion that "the fact of regularity of social change is the primary inference from these phenomena."Granting for the moment that regularity in certain fashion changes has been demonstrated, it must be remembered that the phenomena have been noted within narrow limits of time and source. Only certain features or traits of evening gowns have been dealt with. The data offer far too narrow a basis for the inference that regularity in social change is a fact. The
(341) phenomena dealt with are not to be considered as representative of social change in general, but rather as a group of selected cases.
The author himself states this criticism at the beginning of his study. He tells us that he observed that some fashion traits "don't change enough, others too much." Data concerning such traits he rejected along with cases where sudden and radical change was followed by a new era of comparative stability. It is apparent that the material for the study was subjected to a highly selective process. While this in no way invalidates the results found in the separate traits investigated, it does decidedly lower their value as the basis for an inference as to the regularity of social change in general. Nor can we even go so far as to accept Kroeber's dictum that if any principle could be found in the narrow limits of his study of evening gowns, "it would apply a fortiori to the more changeable kinds of clothing." If a general law of social change is to be formulated, the materials investigated should be determined by some random method and should be sufficiently large in number to afford adequate measures of reliability.
The analysts cannot help but feel that the whole notion of determinism in relation to a natural law is employed by Kroeber in a sense foreign to that of natural-science prediction. A confusion has arisen here between the philosophical viewpoint, or what might be called "rational" or "intelligent determination," and the natural-science approach of empirical observation and experiment. Determinism of any sort is, strictly speaking, not a natural-science concept, When the scientist observes a concomitance between two variables he does not maintain that one determines the other, but merely that they are present together in his observation. He may from the appearance of one of them predict the occurrence or behavior of the other. By observing one of the variables whose concomitance with the other is expressed as a scientific law, he may adjust himself telically (i.e., in practical life) to the whole situation. In a sense, then, the sequence is for him practically determined. He can, moreover, appreciate it intelligently; and it seems to make his world more coherent and dependable.
A subtle illusion of determinism thus results from shifting our sphere of discourse from the practical or telic to the philosophical mode, without realizing that in making such a change a critical inspection of our concepts is necessary. The curve of temperature in the different seasons coincides fairly well with the other criteria by which we identify winter, sum-
(342) -mer, spring, and fall. To the man, therefore, interested only in giving an intelligent verbal expression it is sufficient to say that temperature is determined by the season of the year. In his treatment of cultural change, Kroeber, in fact, goes farther in that he implies that the height of one part of the curve is determined by the part just preceding, since the whole curve is to be understood as obeying some necessary law. Upon this view winter temperature determines the warmer temperatures of the spring to follow. It is, of course, only our prediction or understanding of what is to come next, and our corresponding practical adjustment, that is determined by the so-called law of the curve together with our present location upon it. We have no evidence that either seasonal or cultural phenomena are in themselves, and, apart from our interpretation, determined by one another in this way.
Notions of cultural determinism such as Kroeber's must be taken in the sense of philosophy of human interpretation rather than facts revealed by a natural-science approach. As a system of metaphysics or a religion they may have value in certain phases of life. They should not, however, be rep-resented as emerging inductively from data of an empirical sort which are capable of analysis and explicit denotation.
In the foregoing discussion we have granted temporarily the reliability of the data tabulated and plotted by Kroeber, and of the trends and cycles thus produced. Upon inspection of the raw data, however, it becomes apparent that little assurance can be attached to the year-percentage averages upon which he bases his conclusions. The reliability of his average is of necessity low, since in no instance is it based on more than ten cases, and in some instances upon so few as four. This lack of reliability in the year-percentage average is a vital defect. In many instances very high variability appears in the data within a given year. Consider, for example, the figures on the length of the waist for the year 1859 (items in line 3 of Table I). There are nine actual measurements and one estimate, which, for convenience, we shall consider as ten cases in all. Reducing the raw data to ratios based on the length of the entire figure (i.e., to percentages of items in line 1 of Table 1) we find a total range of 2.41. This range within the individual measurements for the year 1859 is greater than that within the yearly averages which the author assigns for the years 1859-64. The range within the year 1886 in the width of the skirt (which is the trait Kroeber considers the best) is greater than the range of year-percentage averages between the years 1870 and 1908.- There is a greater range in the length of the skirt in 1886 than there is in the averages from 1881 to 1912. Considering the small number of cases and the wide variability within a
given year, we question whether the reliability of the averages, and consequently of the plotted curves, is adequate.
IV. SUMMARY AND CONCLUSIONS
The analysts have suggested as useful for the approach to culture a number of methodological concepts, which also represent approaches to the phenomena of experience in general. The first is that which we have called, in a narrow sense, the natural-science approach. It requires material toward which one can react in an explicit manner. By reactions toward the material either by manipulating it under control or by changing the viewpoint in the process of observation, the scientist is able to derive generalizations or laws. These laws refer not only to the action of the object which is initially studied, but also to the analysis of this object into simpler components and to the discovery of generalizations which hold at various levels or stages of analysis. The second approach is that of genetic explanation. The attitude here taken is to trace the history of the phenomenon in question, showing phenomena which are its usual antecedents, or so-called "causes," and attempting to explain through natural-science laws seen in perspective how the event or the object came into existence. The third approach, which may be called the telic view, is one in which we do not attempt to analyze, to discover laws, or to explain the object before us, but merely attempt to use it for other purposes, and to make such adjustments in it or in ourselves with relation to it as to serve more purposes and to Serve them more effectively. The fourth approach is like the telic in that it is not interested in the object for itself, but only in human feelings or purposes toward it. It is the stylistic view. From this aspect we are interested only in measuring or predicting the formal or incidental changes (usually of an aesthetic character) in the man-made objects which are the materials of investigation. We have seen that culture, as understood by ethnologists, cannot be viewed or investigated from the first of these approaches, the natural-science view; but retains its identity as culture only in the latter three. Culture is fundamentally telic in_ character. When we approach an object with a view to manipulating it for the sake merely of analysis, it is no longer a cultural object.
Implied in the writings of Professors Chapin and Kroeber is the assumption that, since culture is a part of the natural world, its development must follow certain definite laws. In discovering these laws we can proceed as we would in dealing with the materials of the natural sciences. The problem, they assume, is one merely of devising accurate units of measurement and obtaining adequate samples of, the data. The fact that prediction in social science lags so far behind prediction in "physical science"is due, in Chapin's opinion, merely to the fact that, whereas in physical sciences we have reached the level of "projected invention" through symbolic trial and error (use of principles), we have not yet passed in social science beyond the level of "empirical invention" through overt trial and error. Kroeber thinks that there is a "principle of order" governing cultural phenomena, which we may find inductively by a study of cultural changes, and which both describes such change and furnishes us a basis for prediction. Acting upon these hypotheses, the authors have set out to select their units and their data, and to measure cultural changes as a natural scientist measures the changes in objects within his realm of observation.
Against such experiments, when carried out with consideration for statistical accuracy, no valid objection could be raised. In the last analysis
(345) their success and vindication must rest upon the findings themselves. It is, therefore, not the empirical procedure which we are here criticizing, but only the hypotheses upon which the authors have proceeded. The problem is probably not so simple as they imagine. The opinion that the only difference between our ability to measure and predict culture and our ability to measure and predict other phenomena is one of degree and is due to backwardness in our techniques is an assumption which leaves out of account some fundamental methodological problems. We have shown, for example, that one cannot explicitly handle, or react to, cultural phenomena in such a manner as to reveal the way they behave independently of human beings; and one cannot analyze them into parts and describe the laws of action of those parts. As soon as such an attitude is taken the object loses its cultural significance and we find we are investigating a different phenomenon, and one which tells us nothing about the behavior of the "cultural" phenomenon. This is, perhaps, the same as saying that in the natural-science approach we deal with the object and its changes, so far as possible, as independent of ourselves. It is external; our behavior toward it is explicit. In the cultural approach, on the other hand, we understand or define the object in terms of our own response toward it. Its meaning is carried in our own implicit responses. It is, therefore, largely internal or implicit in character. Experiences of this sort must be regarded as upon a plane of their own; one cannot treat them by the methods of explicit denotation and analysis. They are uni-level rather than multi-level phenomena, and their level, indeed, is not one with which one might begin a multi-level analysis.
The consequences of this fact are important both for prediction and for the possibility of social laws. When one asks, for example, what sort of phenomena the cultural sociologists are measuring, we must conclude that their measurements are more largely of implicit than of explicit phenomena. What Chapin is really predicting in his cultural growth curves is the number of purpose-fulfilling objects which may be found in a given year as parts of a complex cultural object to which we assign a certain name. He is, in other words, predicting our own counting behavior at some stated future time when we shall take an attitude of enumerating these telically conceived parts. He is not predicting the growth of a specific object as one would predict the growth of an organism. The more highly perfected automobile does not grow out of the primitive auto-mobile, except by a ludicrous stretch of the metaphor. It is not the same automobile whose parts we are counting after a period of inventive development; it is a different and complex object which we now denote by
(346) the name of "automobile," just as we formerly denoted its crude predecessor by that name. Or to put it in another way, we are here predicting the number of details in the pattern of implicit, or thought, reactions which will be exhibited in the behavior of a future manufacturer as he sets up a factory to produce the objects which we shall call by the old name, "automobile." The "culture object," therefore, whose "growth" Professor Chapin is measuring, is more within ourselves (in our naming and thinking reactions) than it is outside. The measuring behavior which he uses in plotting its growth is also implicit rather than explicit. It does not consist of running an eye or a finger along a meter stick laid against some explicit object (e.g., as in measuring the stature of an organism) ; but in counting the things we conceive to be parts of a "culture object" which is an object not by virtue of its own continued existence, but because things of its class arouse in us the same name continuously. Certainly, therefore, the difficulty inherent in measuring cultural change by the counting method, with resulting inequalities in the things counted, is not merely one of degree, to be overcome by the refining of our technique; but it is inherent in the very nature of the problem. It is the inevitable result of trying to apply to the telic, or cultural, approach the natural-science concept of measurement.
The discrepancy between the telic and the natural-science approach is most marked with regard to the universality of social laws and the resulting capacity for prediction. Let us suppose that the methods of Chapin and Kroeber were carefully employed in studying thousands of different inventions and institutions, and that in these measurements the same cultural growth curves and periodic cyclical changes were demonstrated without exception. Since scientific laws are merely statements of probability, we should then feel fairly confident that the next cultural phenomenon would conform to the same law. In other words, we could conclude that the author's hypothesis was verified with extreme scientific precision. Let us suppose, however (and this supposition seems more in ac-cord with the facts) , that in the measurement of a certain cultural complex the growth curve of its parts is found to be distorted, or that there is
(347) no change at all; or let us suppose that the stylistic measurements of certain objects, or of certain dimensions of those objects, fail to conform to the law of periodic cyclical change. Or again, let us merely focus our attention on the fact that periodic cyclical changes, even if their existence be proved, vary greatly in span between different elements. Now, under all of these conditions there emerge special limits to the universality of cultural laws. It seems that many cultural phenomena are obeying laws of their own, thus limiting the very notion of scientific generalization. A plurality of laws of cultural change, with apparently no explanation beyond these laws themselves, is an unsatisfactory condition. We know nothing of cultural change as such, but only the laws of change of this or that invention or organization. There are no universal principles which explain the discrepancies between cultural elements; and there is no way of predicting whether a certain element, not yet measured, will differ in growth from the hypothetical pattern, or what the direction of that difference will be. Clearly this condition calls for further investigation. We must reduce these differing phenomena to their essential components and discover laws at simpler levels, which will account for their varying appearance and will give us a wider basis of prediction.
Natural scientists are continually faced by the same problem. And this dilemma is in fact one of the most familiar and important incentives for the advancement of research. Certain unpredictable changes of appearance or position of the stars call for deeper astronomical study. The unexpected behavior of bacteria in a culture, violating the laws usually holding for such organisms, leads to a minuter analysis. It was found, some years ago, that epidemics of puerperal fever in a hospital usually attended unsanitary conditions. This relationship, however, was not a universal law, since the incidence of this disease under unsanitary conditions was not 100 per cent. Searching for some generalization to explain both the usual incurring of the disease and the cases which escaped from it, it was necessary to find a technique for breaking the present factors (human tissues, dirt, etc.) into their components, and revealing the laws of action of those parts at their own level. As soon as the microscope revealed the bacillus and its behavior in the human organism that particular problem was solved. It was easy to explain through a knowledge of the conditions of microbe communication, not only the law of connection between puerperal fever and unsanitary conditions, but its exceptions. Moreover, such exceptions could be, in some measure, successfully predicted.
A more recent example is to be found in attempts to explain the exceptions to the law of gravitation which occur in certain electromagnetic phe-
(348) nomena. Gravitation and electromagnetism have long been regarded as two separate and ultimate forces. The genius of an Einstein, however, was not content to leave them so; and he has been seeking, perhaps successfully, to conceive them under a more nearly general law which will enable us to predict and to account for exceptions to either principle. When the objects of nature can be approached in an explicit manner it is thus possible to analyze them into simpler components and arrive, through a study of those components, at wider and more valid generalizations which enable us both to understand and, to some extent, to predict the unusual within our experience.
Returning now to the problem of the exceptions to cultural laws, we meet with a sudden impasse. It is impossible to deal with objects explicitly and analytically and have them retain at the same time their cultural significance. And without explicit behavior and analysis, no wider under-lying generalizations can be discovered. If the plow sulky is analyzed into its components, giving more general laws at simpler levels, we can never come back to it as the plow sulky again. The natural laws derived are the same as would be gained by an analysis of any piece of wood or iron in the chemist's laboratory. Such laws tell us only about the action of wood or metal; they reveal nothing concerning plow sulkies. Hence we are unable to explain irregularities or exceptions to the norm of the growth of inventions which the history of this article might reveal, or to pass below these differences to some more general law through which they might be understood. We can arrive at fundamental laws enabling us to predict different kinds of change within physical materials; but we can find no such elementary principles through which we can predict variations of cultural change. The methodology which is at the very heart of the natural-science approach is entirely foreign and irrelevant to the field of cultural phenomena. We must realize, therefore, that the gap between our capacity to generalize and predict regarding culture and our capacity for generalization and prediction in the "physical" sciences is not a matter merely of the degree to which we have refined our technique. It is a fundamental difference of kind. This fact, to be sure, does not reduce the possibility that there may exist laws upon a societal or cultural plane; nor does it argue against our capacity to discover such laws upon their own level. But it does suggest that deeper penetration to secure wider generalizations, that the explanation of any variations in cultural development, that the prediction of the unusual, and that the organization of inductive principles into a coherent body of knowledge are at the present time unlikely within the field of cultural change.
The failure to grasp the fundamental difference between natural-science
(349) and cultural material has led students of culture to advance certain hypotheses of doubtful value. Chapin, assuming cultural phenomena to be-long in the world of natural science, looks about for natural-science principles through which to interpret them. Since no explicit analysis of his phenomena is possible, he resorts to describing them in terms of analogy. He borrows laws from a natural-science level, the biological, and interprets his cultural changes in such terms as organic growth and the learning process. Against this procedure we have found two fundamental objections: first, when the analogy is stated in exact form it requires a distortion of the biological facts; second, it obscures the requirement for uniqueness of formulation at different natural-science levels. It either assumes that culture is an organism, a comparison which cannot be maintained, or that we have a superorganic world which behaves in terms of the laws of organisms. This conclusion amounts to tautology.
Kroeber has attempted to meet the difficulty in another way. He boldly asserts that we should cut loose from our moorings to the levels of natural science, and seek our interpretations of the superorganic in terms of that level itself.  He sets up superorganic forces, such as the sweep of fashion or the determinism of culture-cycle curves, as explanations of his phenomena. He speculates upon the operation of some extra-human power which alters the length or the width of women's skirts, and which even attempts to push the hem of the garment into the ground. We thus arrive at a mystical notion of determinism which is quite as hopeless for directing future research as the organic analogies of Chapin.
Such unworkable hypotheses not only lead to far-fetched interpretations, but detract attention from more promising investigations within the , cultural field. The analysts believe that the approach best fitted to give us knowledge concerning the cultural changes under discussion is that of genetic explanation. Remembering that the phenomena which distinguish objects as cultural are implicit rather than explicit, we shall expect to find them in the behavior of the human beings who invent and perfect such objects. Though not to be found by natural-science research into culture objects, a conceptual linkage of cultural phenomena with the realm where natural laws are revealed may still be achieved through the study of the culture-producing behavior of human beings. In this way we should use psychological and biological laws not as descriptions of cultural change upon its own abstract level, but as' principles in the historic explanation of cultural development. The changes in culture from this view are iden-
(350) -tical with the history of human thinking and learning; and these in turn are conceptually dependent upon known natural-science laws working amid a complexity of conditions. This statement of the problem opens a field of investigation, in terms of human behavior, which would have been ignored by postulating cultural phenomena as a level within the natural-science series. We have attempted above to suggest a few hypotheses useful in such a study. It is to be regretted that the approach of genetic explanation has been so little stressed by Chapin and has been deliberately rejected by Kroeber.
Notwithstanding their methodological difficulties, these two investigators have made a genuine contribution in that they have opened the cultural problem for scientific investigation. The student of human behavior, in particular, should be grateful for their attempt. Certainly the average laboratory psychologist or case-worker would not have thought of surveying these wide fields of cultural traits or of attempting to describe them in clear and simple terms. And until such a description was given, and tentative laws formulated, the psychologist could not have seen the implications of his principles and methods for the study of human nature. The historic trend of cultural behavior, and its sweep in geographical areas, are discoveries which cannot fail to stimulate fresh thought and imagination in human research. Only commendation can be given these investigators for their boldness in asking of nature the questions: Are there laws of cultural change? And is the prediction of culture a possibility? Their techniques also are of much interest, and although not fully relevant to the purpose intended, are illuminating for purposes perhaps still more important. The sole object of this article is to question certain assumptions regarding the nature of culture and the meaning of cultural investigations. In using their culture-measuring techniques, how-ever ingenious and revealing such techniques may be, are these investigators really doing what they think they are doing? This challenge is merely a suggestion that their methodological postulates be examined afresh; it is in no way a detraction from the contribution they have made.