Important Developments in American Social Psychology during the Past Decade (Concluded)

Leonard S. Cottrell, Jr. and Ruth Gallagher
Cornell University

In the preceding section[1] we have discussed the general trends and the specific developments in social anthropology, psychiatry and collective behavior studies as they have affected social psychology during the past decade. The first part of the present section will continue the consideration of these trends as they have affected quantitative and experimental work in social psychology. In the second part of this section we seek to point to the problems of method which must increasingly occupy the attention of social psychologists.


It is obvious that thus far this paper has not dealt with developments in the field of quantitative techniques during the past decade. The avoidance was intentional because we have held that the fundamentally important developments in social psychology during the past ten years have been in orientation rather than in technology. We wished therefore to establish our point of view before evaluating some of the developments in quantitative technique.

In the present discussion we are not raising the question of whether to quantify or not to quantify in social psychology. The problem by been discussed rather thoroughly by Lundberg and others.[2] Rather we prefer to limit ourselves to

( 303) indicating the evidences of orientational shifts in quantitative methods in social psychology.

Gordon W. Allport, in his analysis of 1600 sample articles in 14 selected psychological journals covering a period of 50 years, [3] has shown among other things that so far as a sample of articles in the general psychological journals could show: [4]

1) Interest in problems of method rose rapidly from 1918 to the end of the twenties and continued at a fairly high level.

2) The use of statistics showed the same pattern to a more pronounced degree.

A study of the articles over a period of ten years, beginning about 1916, gives rise to the distinct impression that research workers were preoccupied with refining instruments of precise measurement. The best work on reliability and validity of personality and attitude tests appeared in this period. The proliferation of so-called mental tests that followed Binet's work was followed about ten years later by a barrage of so-called personality tests. In all of this work the drive to precise mensuration forced a more rigorous definition and delimination of the variable to be measured. This tendency aided and abetted the atomistic and fragmental approach already well-developed in laboratory psychology. It was rare indeed to find within the ranks of those who operated under the symbol psychology any who protested this excessive atomization of social behavior.5 Of course the Gestalt psychologists must be excepted, but the full impact of their work on American social psychology was not felt until the early thirties.


The search for clear-cut personality variables that would lend themselves readily to statistical procedures presents research workers in social psychology with two important problems. The first is that of integration of Isolated bits of behavior into categories that are meaningful in a vocabulary seeking to describe social interactional phenomena. The second problem is relating behavior to situational contexts.

With regard to the first problem there have been two general attempts at solution, one using mathematical procedures of integration checked by less formal interpretative judgments, and the other using less formal judgments and insights [6] to arrive at meaningful categories and checking these by quantitative techniques.

Thirty-five years ago the assumptions regarding specific mental abilities and the multiplicity of tests of these abilities had forced Spearman to develop his famous "factor g" hypothesis. His hypothesis was that a group of intercorrelated tests when analyzed by his two-factor method would show a common factor g and special or unique factors for each test.[7]

When the development of measurement of personality traits other than intelligence occurred, there were some attempts to use the two-factor method to integrate clusters of isolated variables. These applications of the method did not get very far before the brilliant generalization of the method was made by L. L. Thurstone in his multiple factor theory.[8] By using the theory of determinants and vector analysis, Thurstone was able to develop a mathematical theory and procedure for reducing any matrix of intercorrelated variables to a set of clusters which would define the general factor or factors one needed to posit in order to account for all the intercorrelations.

( 305) Thus, Spearman's single-common factor theory was found to be a special case of a more general theory of multiple factors.

Most of Thurstone's recent work has been directed to the application of the theory to mental abilities. However, the method was taken over by workers interested in measuring other personality variables and even by those working on situational factors [9] in the effort to tie together a multitude of items into a smaller number of categories.

The work of the Guilfords [10] is perhaps as representative as any in the application of multiple factor procedures to personality test items. From our study of these and other efforts we get the impression that when the method is applied to behavior variables as now measured, it can do little more than establish new cluster categories that are not much more usable than the current ones. The same general observation holds with respect to its application to attitudes or opinions.[11]

It is also our distinct impression, however, that the factor method will yield astonishingly helpful results when it is applied to variables that are described and measured with reference to a specified situational field or context. Indeed, we are fortunate enough to have some tentative confirmation of this hunch in the as yet unpublished research of Kurt Lewin,

( 306) Ronald Lippitt and Ralph K. White. As a part of their elaborate studies of group behavior under various types of leadership, they have intercorrelated a number of items of behavior related to the specified situational field. Preliminary manipulation and inspection of the data seem to indicate that without doubt two major clusters will emerge which are meaningful in describing rile patterns of children in relation to adult authority symbols.[12] This suggests that results of applications of factor analysis in social psychology to sterile non-contextual descriptions of isolated behavior units will themselves prove sterile. But when behavior-in-specified-situational-fields descriptions are used the results will support the claim that the method is one of the most useful products of the past decade. This is not the appropriate place to discuss the application of the factor method to problems of measurement of intelligence. We suspect that here again if a contextual or behavior-in-situation orientation characterized analyses of intelligence, factor methods would yield more fruitful results.

A second type of attempt to solve the impasse of atomization has been the selection of apparently meaningful behavior categories on the basis of general experience and observation and the attempt to determine by quantitative procedures whether or not a given category behaves as an integrated configuration. A correlative assumption in such work is that such complexes are general trait attributes which are possessed in varying degrees by individuals.

The work of Gordon W. Allport perhaps best represents this approach.[13] The weakness in this approach inheres not so much as many think in its dependence on "insight" or informal quantitative knowledge; but rather on its failure, to date, to deal

( 307) explicitly with situational factors. Thus, it deals with traits as attributes to be found in the personality quite independent of the self-other(s) context. To be sure, Allport's Ascendance Submission Test is framed in terms of behavior in situations. It is one of the few "trait" tests to show any such sophistication.[14] However, the combination of responses in many types of situation to get a total score which is presumed to indicate the degree of ascendancy or submissiveness possessed by the person reveals failure to follow far enough the view of behavior as functional to a situational context. Thus, responses which indicate ascendant behavior in some situations and submissive behavior in others will yield an intermediate score which actually predicts nothing of what behavior a person will manifest in a specified situation. The rather high validity of this test is probably due more to the stability and uniformity of the situational fields represented in the test than to the fact that it has uncovered a general trait-attribute which suffuses the behavior of the organism. From the point of view represented in this paper, measurement of traits will become more fruitful when the variable to be measured is tied to a specified type of relational context and when the thinking about traits is situational-contextual in orientation. In fact, the trend in this direction has already definitely established itself [15] and may be stated as one of the important developments in social psychology during this decade.

It is difficult to evaluate the efforts now being made to make quantitative descriptions of the situational field. As long as research workers were content to assume that an isolated measured response could be correlated meaningfully with a single artificially isolated "stimulus," the problem was relatively simple. But when a dynamic community situational field in which the interacting groups must be located, or a single

( 308) group in which persons have social psychological positions, or a marriage in which each partner represents several overt or potential rôles becomes the "stimulus" to be described, then the problem is a difficult one indeed.

The best known efforts to develop a system for quantitative description of situational fields are those of Lewin and his students.[16] Their application of the concepts of topological mathematics to psychological interpretations is most ingenious. While we are not certain that their regions, boundaries, barriers, paths, life space, et cetera, will prove the most useful tools of description and interpretation for a matured behaviorin-situational-context vocabulary, they certainly point the researcher in the direction of such an orientation and most certainly will serve as one of the major points of departure in this crucial undertaking.

Another possible lead in this direction, and one which may have possibilities of further elaboration in describing both interpersonal and intergroup situations is that of J. L. Moreno's [17] sociometric descriptions. The simple charting of the attraction-repulsion networks of members of a situational field of course has nothing of the mathematical elegance and sophistication of Lewin's work. However, the results of such chartings may be translated into topological language, and application of the method may actually be more practical for certain purposes than procedures formulated for topological studies.

The fruitful situational analyses now being made with little formal quantitative technique will prove stimulating challenges to quantitative methodologists. The descriptions of class relations in communities found in the work of John Dollard,[18] Allison and Elizabeth Davis,18 and Burleigh and Mary

( 309) Gardner,[19] and the forthcoming work of W. Lloyd Warner,[20] will undoubtedly tempt many workers to give mathematical description to such dynamic fields. These efforts are bound to test present topological and sociometric formulations and lead to other schemes of situational quantification. Similarly the behavior-in-situation analyses made by clinicians, noted previously in this paper, call for the same testing and reformulations of a mathematics of the dynamic situation. It is quite probable that the multiple factor method applied to situations rather than to "minds" and "personalities" will lead to some form of "factor loading" description of situations.[21]

Since the attitude and its measurement has been such an important focus of attention in social psychology, it is necessary to add a note on developments in this connection.  The efforts to apply the brilliant precision methods of Thurstone to the analyses of attitudes in many dynamic situational fields have led to the recognition that except in very stable contexts the method is too cumbersome. Thus, modification and short cuts have been developed. The work of R. Likert [22] is especially notable in this respect. He has been able to construct reliable and valid instruments with much shorter procedures. Even the Likert technique, however, is too complex for purposes of quick approximations in rapidly changing situational fields, particularly large-scale fields.

The demand for quick large-scale assessment of rapidly changing attitudes has led to the use by the various opinion polls and government agencies of "yes," "no" responses to single items. None of these developments, however, discards or discredits the Thurstone technique. Where time allows and precision needs demand, his technique is indicated.


It is sufficient for our purpose to give a very brief indication of the shifts in experimental work which seem to reflect the trends we have already described.

( 320)

The best indication that a major revolution has taken place is the marked change in orientation and emphasis seen in the revision of the Murphys' book, Experimental Social Psychology.[23] In the short space of six years (1931-1937) what had been a handbook of references to isolated bits of more or less meaningless experimental research became a systematic evaluation of research work with a major emphasis on integrative and contextual orientation. With this change in point of view, there came a greater tolerance and appreciation for the work of cultural anthropologists, and psychiatrists who, as we have pointed out, were leading in the newer orientation.

The productive results of the interpenetration of social-anthropological research and laboratory research are wellillustrated by the work of Muzafer Sherif.[24] Be it noted that Sherif, in addition to being a member of another culture (Turkish), studied in the fields of cultural anthropology and sociology (particularly of social classes), as well as in psychology. His application of the experimental technique to problems of the establishment of norms in defined and undefined situations demonstrates the necessary interdependence of perspective giving work in larger social contexts and the precisely defined situations of the laboratory.

With the scholarly work of the Murphys and Newcomb revised and brought into line with the dominant trends, it becomes as presumptuous as it is unnecessary for us to attempt a review of experimental work in social psychology. We need merely to state that there is widespread evidence of attempts to make the complex situation explicit in experimental analyses. This has not yet reached the point where experimenters are fully aware of the necessity for making explicit not only the laboratory context but also the impinging situational contexts which may be projected by the subject, e.g., his class, community and family situations. Moreover, it is true that as yet experimenters have not reached the point of sophistication where they realize the functioning in the experiment, its formulation, execution and interpretation, of their own positions in relevant contexts. In our opinion, however, the time is soon coming when the observer will scrutinize his own explicit and implicit situational frames in interpreting the problems he formulates and the interpretation he gives his findings. Bias

( 311) is inevitable and, what is more, if we use it explicitly and invite many biases to a problem, it may become one of the most valuable tools in our entire kit.

As illustrations of the trend in the direction of more explicit situational orientation, we may cite in addition to Sherif [25] the work of Lewin [26] and his associates, that of H. A. Murray [27] and his associates, of Lois B. Murphy,[28] D. M. McGregor,[29] Hadley Cantril,[30] and R. R. Sears. [31] These names by no means exhaust the list nor do they represent an adequate sample.

It is also important to note here the growing tendency for experimenters with animal subjects to become more aware of total situational factors (even including the experimenter's personality) and to make them explicit in their interpretations. Moreover, there is a greater tendency to give attention to setting up situational fields roughly congruent with human situations and to point to possible implications for the social psychology of humans. Fifteen years ago Hamilton[32] made use of

( 312) animal experimental work In psychopathology and suggested more extensive use of it, but it was several years before his suggestions bore any fruit.[33]

This impressionistic report of trends in experimental social psychology may be concluded with the prediction that during the next ten years experiments will show a rapid shift to the orientation we have outlined in this paper. There will be less patience with fragmental work which merely allows the use of ingenious instrumentation and more demand that experiments be set up to answer questions raised at critical points in a dynamic field theory of social behavior. Experimenters will think more in terms of situations and less in terms of isolated segments of behavior. They will be more conscious of their own positions in the experimental situation and the inevitable interpenetration of experimenter and subject. And finally, the problems they formulate will be viewed Increasingly from the point of view of their relevance to clarifying our understanding of behavior in interpersonal as well as large-scale social situations.


The foregoing sections of this article were written in the summer of 1940 at the request of the Sociological Research Association. Originally our intention was to outline the important developments in the field of social psychology which occurred during the decade of the thirties. A s we have seen, however, it was impossible to talk about developments in social psychology as though they cropped up full-blown; rather we had to show these developments as trends, relevant both to the particular field of social psychology and to the general field of

( 313) social sciences. We found ourselves engaged in the never-ending task of trying to reconstruct history. Instead of listing events in a few pages, we attempted to give them direction and movement, and tried at least to make the context of these movements as broad as the limitations of time and our particular perspective would permit.

Necessarily personal selection figured large in the process. Lacking any rigorous basis for assigning relevance or importance to the trends we observed, we constructed one from the elements of our own bias. First of all we weighed the trends in terms of their contribution to social psychology at the conceptual or theoretical level, a fairly arbitrary choice; and secondly we set up an arbitrary standard of what we considered to be the requirements of an adequate social psychological theory. We accepted the assumption that all knowledge is relevant to particular perspectives and went ahead to describe what we saw from our perspective as fully and honestly as possible. Recent evidence[34] gives some support to the validity of the point of view we presented.

The present section was written in the summer of 1941, just one year later. It is not our intention to prolong this article by citing works published within the last year that would substantiate our position. Such excellent article as G. P. Murdock's "Anthropology and Human Relations [35]   and A. Kardiner's "Psychoanalysis and Psychology" [36] fall readily into the picture as continuations of the developments we have noted. And sufficient others of the same temper have appeared within the last year to suggest the scientific utility of the frame of reference we have outlined. There is a point however beyond which trend-finding degenerates into an "aha we told you so" attitude, and this is particularly inexcusable when the trend-finder focuses his specialized observation on a span as short as one year. We can take advantage of the year intervening between this section and the previous ones much more profitably, we think.


In reviewing the trends we built a certain picture of the status of social psychology up to 1940. If we take this as a given, it affords us a fairly good springboard into the future of social psychology. From comments and reactions to that picture, including our own, we have taken a new perspective toward it--one which affords some insight into the possible direction of future social psychological developments. In one sense this new perspective represents a shift from our former "criterion of relevance and importance" to a new criterion; it takes us from the realm of social psychological theory to that of method.[37] Instead of using the "requirements of an adequate social psychological theory" as our yardstick, we propose to measure social psychology against the canons of operational science.[38]

As a matter of fact our choice of a conceptual standard for the foregoing sections was perhaps not so arbitrary as we made it sound. It was as much a commentary upon the role of social psychology in the thirties as it was upon our own interests and bias. If we were to characterize this rôle, we might say that social psychologists conducted a kind of clearing house for the theoretical output of other social scientists. They battened on the research efforts in other fields, but offered little in the way of research return from their own field. Recall how much of what turned up as grist for the social psychologist's mill came from anthropology and clinical psychiatry, political science and even economics. Recall too the lack of speculation about method, the unwillingness to dogmatize about scientific provinces, in short the whole brokerage aspect of social psychology in this era.

If our emphasis on theory was prompted by the nature of social psychology in the thirties, it would not be straining fact too far to say that our present choice of emphasis is

( 315) prompted by a critical need which these developments have finally produced. Sooner or later social psychologists should be brought to book to define their subject matter and their research goals within it, and unless the forties are to be nothing but a reiteration of the thirties, the coming decade should show an increasing preoccupation with this problem.

The important point established by inter-disciplinary cooperation was this: Each of the social sciences furnishes just another perspective on the varied phenomena of human behavior; each isolates for study only a particular aspect of the common primary datum, social action, and no one of them can thrive in even partial independence of the others.[39] This is a conclusion worthy of a decade of effort. But further crossfertilization of the kind which established this point is no longer needed. To consolidate this gain, social scientists should begin to ask themselves some methodological questions. What justifies these separate approaches to the same data? How do we define the subject matter of a particular social science, and its rôle in relation to the other sciences which study social action? It is difficult to see how anything but a rigorous concern for method in each of the social sciences will answer these questions satisfactorily.

For social psychologists this presents itself as a critical issue. They can continue the brokerage function which they have assumed within the social sciences and sidestep the troublesome problem of method if they like. This is an alternative not open to the other fields. But, if social psychologists choose this course, they must be content with an increasingly second-hand contact with data. If they choose the other alternative, they will address themselves directly to the problem of method, taking advantage of their cross-roads position between institutional and clinical studies of behavior to establish an operationally valid relationship with actual data.

The challenge of this latter course is such that we doubt if social psychologists will ignore it. In the spectrum of the social sciences, there is a dark spot between institutional studies of behavior and the "impertinently realistic" studies of the clinics. In this area rests the key to the meaning of culture and society to individuals, which no amount of further theoretical refinement alone will reveal. This is the area which institutional studies shoot over and clinical studies shoot under, and the one which only a methodologically mature social psychology can actually hit.

Much of the groundwork for meeting this challenge has already been laid, and it is readily accessible if we are willing to go to the philosophers of modern science for it. Bentley's

( 316) statement that "The newer science becomes able to express itself frankly on the level of its own skills"; and again that "A generation ago physics and chemistry were differentiated In terms of fact: today in terms of objective and technique.[40] represent a position won by assiduous thinking in conjunction with experiment. They contain suggestions which social psychologists in their present situation cannot afford to ignore. According to this view there are no separate scientific subject matters apart from methodological convenience. The subject matter of social psychology, no less than that of any other science, is given by operations which social psychologists perform. It is the nature of the operations, not the "nature of the data," which determines the kind of knowledge which eventually goes under the heading of social psychology [41] .

Up to the present day the kind of operations which have built a significant body of social psychological theory have resisted formulation, but they were nonetheless operations performed in a context. The theory of the social origins of mind and self, the theory of the situational determination of behavior, were arrived at by men like Cooley and Mead, who did something within the social process and recorded their reactions to the doing. Unfortunately, the operations that went into the building of this theory are lost to us. At the time when Cooley and Baldwin and others were theorizing about social experience, there was no operational movement on-foot, no awareness that successful communication of theory depended upon our ability to describe the operations which were behind it. The lack of explicit operational statement in the theory we have inherited is a hindrance to us even up to the present time, for so far 'as operational consciousness goes, social psychology is still in the position of Mr. Bentley's "physics a generation ago."

The ABC's of operational thinking, that operations must be made both exact and explicit in order that their accompanying conceptual scheme be communicable, are not yet the ABC's

( 317) of social psychology. No social psychologist has deliberately set himself down to do what the physicist Bridgeman did--to observe himself as scientist at work and to record accurately the relation between experiment and concept found in his own behavior. The difficulty in subjecting interpersonal processes to experiment is partly responsible for this, but a greater share of responsibility lies with the fact that it is possible to communicate social psychological insights in the inexact hit or miss language of everyday social life.

The physicist can safely look upon the atom or the electron as creatures of his own scientific construction.He does not have to worry overmuch about lay misinterpretations of his concepts or competition with other non-professional terms. To communicate his meanings he knows that he need only fall back on the specialized operations which produced them, he does not take for granted enough similarity of experience on the part of his audience to avoid or omit operational proof. Not so with social psychology. Here we are dealing with the stuff of everyday experience in the words of everyday experience. The concepts which we held to be important in social psychology--the self, the other, rôle, situation, are subject to myriad lay connotations,--distortions of one form or another of the specialized use to which social psychologists would put them.

The danger here Is not that students will cling unreasonably to their own usage of these concepts in the face of more rigorous definition. The danger lies with the social psychologist himself who is misled by the commonplaceness of his subject to take too much of his audience's experience for granted. In modern science anything taken for granted without being established operationally is a one-way ticket to confusion, and in social psychology, where it is so easy to assume similar experiential backgrounds, exactness is practically unknown. Too often successful communication in this science depends on such unknown quantities as the "personality" of the teacher, or on his ability to stir the imaginations of certain students who "have the knack" for social psychological investigation. However commendable this procedure may be in the arts, it has distinct limitations for science, where one should be driven by the difficulties of precise communication to be as exact and rigorous in exposition as he is in analysis.

One of the first steps toward making social psychology an operational science should be to subject this art of teaching social psychology to analysis. Investigate for instance the backgrounds of successful academic social psychologists, (or the unsuccessful ones, too, for that matter). What types of experience

( 318) predispose one to theorize about interpersonal relations in the first place? What pattern of rôles derived from interpersonal and intergroup relations is effective in the process of communicating this theory to students? Perforce the answers to these questions will throw light on the students who "have the knack" for social psychology, but it would not be time wasted to study the reactions of different students to this type of education. It would be important to know whether the verbal ability which usually wins high grades has anything to do with the ability to use insight in solving interpersonal problems. It would be even more important to investigate the change in life pattern occasioned by social psychological education. Is there, for instance, a transference between students and teacher that parallels in any way the transference between analyst and patient?[42]

To answer these questions in operational terms is to do more than describe a particular form of communication. It is an inquiry into the nature of social knowledge when such knowledge is geared to the demands of a specific and limited social act. It is the chance par excellence for the social psychologist both to reinject living operations into the theory which he expounds, and to overcome for all time the disadvantage he has felt as a non-clinical investigator of behavioral relations. If he Is willing to look upon his classroom as a clinic, he is immediately up against a horde of subtle and intricate social processes,--rôle-changes, insight, resistances, which in range at least far surpass the data of clinical psychiatry. That his sample is limited to students with their stable age and socio-economic characteristics, and to himself, need not worry him too much. If he can make clear the processes of interpersonal and symbolic group relations in their behavior, he has already generalized beyond his sample. For process, unlike configuration, is always beyond the limitations of a single sample group.

We shall return to this problem of what the social psychologist could treat operationally, for we have by no means exhausted the possibilities of his rôle in relation to data. It is about time though that we turned our attention to the how of

(319) social psychological operations. Can we state what behaviors in social psychology are commensurate with the scale readings and the setting-up of critical experiments in physics? In brief, what is an operation in social psychology?

We have already indicated in a previous section of this article our theoretical position, that social behavior Is explicable in terms of social situations, and that social situations are made up of the interplay between rôles of expectation and rôles of response. As we have defined socialization, it is the end product of innumerable rôles of this sort, acts and counteracts which finally become internalized and foreshortened by the use of verbal and motor gesture. All of us act against an incorporated structure of expected acts from others in our environment. Indeed, in the lightning-quick process between the time of perceiving a social situation and acting in it, we make an incipient trial thrust against our world of incorporated "other" rôles. Incipiently we respond to this trial act in the role of the other, and our final overt act is modified, or perhaps simply "approved" in this process.

Operationally this means that we are all capable of acting from a multitude of rôle perspectives in any given situation. However urgently we may cling to our "self" rôle in overt behavior, we have the content of many diverse "other" rôles in our organized reaction system. Any of these may be vitalized into overt behavioral expression whenever the demands upon our habitual self rôle are relaxed. Ultimately it is the mechanism of anxiety (expectation of punishment or uncertainty) which hedges in our self rôles and keeps us toeing the line of expected behavior. But there are many ways of relaxing the vigilance of the "other" rôles which induce anxiety.    The depressant effect of alcohol, acting at the physiological level to deaden most anticipatory responses, appears at the social level in the form of greater spontaneity- -more freedom in the choice of rôles and more abandon in the playing of them. The same thing can happen at the purely social level. Witness the behavior of many staid persons playing with children when they feel themselves free of adult scrutiny; or the sometimes horrifying content of our own dreams when our selves, apart from the demands of daily social life, do pretty much as they please.

What then do these observations imply about the way a social psychologist operates. Ideally, a social psychologist is a person who has a professional license to travel, anxiety-free, from one perspective to another in the social process, and whose chief duty it is to take notes on the way. The operation which makes his travel possible is the familiar one of taking the rôle of the other, but there are important differences in

(320) his rôle taking and that of the ordinary mortal. Whether he is scrutinizing his own responses to slow down the act-counteract process of his own inner forum, or whether he is abandoning his self rôle to identify with the feeling and views of another, he has a further purpose in view than that of winning friends and influencing people. His aim is to understand and make explicit the processes involved, and his desire is to be able to communicate his insight to others with the same aim. Whatever he achieves in the way of better adjustment in his own interpersonal world by this insight is at best a check, not the goal of his efforts.

This much, then, our theory allows us to conjecture about social psychological operations. They consist of deliberate rôle assumptions within the social process, played against the backdrop of professional interest and to an audience whose objectivity rests upon the relative detachment of scientific inquiry. The subject matter of social psychology flows logically from this statement of operations--it is simply the understanding (through behavioral participation) of what behaviors are appropriate to what situations, in order that we may predict backward or forward, from behavior to situation or situation to behavior. This much puts us several steps ahead of "physics a generation ago." We know at least that our data are related to the operations we perform as social psychologists, and roughly what those operations are. But, we are still a long way behind modern physics in precise statement of them.

For instance, we know next to nothing about the necessary mechanics of identification, the kind of operations which place us accurately in the psychological shoes of another. Rough ly we know that duplicating another's physical movements, his habitual gestures and posture, is a help in making his circumstances seem plausible, just as the reverse is true. But how far is it necessary or possible to carry out this operation? Must we, in order to understand another's anxiety, suffer it with him? Or may we consider it as private as his toothache and rely upon observable behavioral signs to announce its presence and intensity. After all, most of us get along quite passably in the social world by making verbal gestures to ourselves in the rôle of another. Usually whatever neuromuscular activity is associated with these gestures is rudimentary and incipient, as when we find ourselves executing a mannerism of

(321) a friend when we are thinking about him. The problem is to develop adequate checks upon the rôle taking process so that we may know when we have actually attained to another's definition of a situation. Only empirical investigation will permit us to say to what extent we may trust our symbolic short-cut interpretations.

Ability to predict an individual's behavior from an operational survey of his incorporated rôle population will of course be the final check upon our rôle-taking. But, prediction of social events is again not an activity exclusive to the social psychologist--any more than is rôle-taking. Every social act, however unreflective, carries with it certain predictions about the behavior of others whom it implicates. A girl pauses before a door, predicting by that act that her escort will open it for her; or a child stops crying when his mother approaches, predicting that she will tend to his wants. As a matter of fact, we are so used to this predictive aspect of our own behavior that we rarely bring it to conscious attention. Ordinarily it slips silently by In the general course of social adjustment until it is brought up short before the court of consciousness, usually for some misdemeanor. When our behavior predicts inaccurately about the behavior of another, we are left holding a bag of inappropriate neuromuscular adjustments which we must somehow get rid of, and consciousness is very often called in to help. The social psychologist as human can afford to operate as blindly as this within the on-going social process, but as scientist, he cannot. He must consciously build conceptual frameworks around his predictions to retrieve them from the stream of undifferentiated process.

Any framework which we impose, like the divisions of time and space, is of necessity arbitrary. However, without some such framework to capture units of the process, it is almost impossible to communicate operations with precision, let alone with any economy. The concept of "situation" in social psychology is adapted to all of these requirements, if by situation we understand all those rôles which provide the field for a given subject's behavior. It sets the boundaries for our predictions and yet encompasses behaviors of varying duration and scope. As a concept related to operational practice, it should permit of generalization as well as prediction. It is the knife with which we cut social process into manipulable sections. As yet it is not a very sharp knife for the reason that it has had very little operational honing.

As social psychologists learn to emphasize the operations they perform in analyzing behavior, the concept of the

( 322) situation will grow quite naturally in specific meaning and usefulness. For after all the term simply describes the unique grouping of operations performed in the act of understanding a particular rôle. Eventually it should be possible to use the concept as rigorously as this: Given an external situation consisting of the following rôle expectations and a subject equipped with the following rôle patterns we may predict that the rôlé of the subject in this situation will be When we have attained this much precision in our situational analyses, it should not be long before we can generalize about types of situations, which produce types of behavior, and this is what we are ultimately looking for.[43] When the subject matter of social psychology is in this form, it should be really negotiable among the social sciences; and social psychology may begin to fulfill the rôle Sapir so hopefully predicted for it.

We have already considered one neglected source of data upon which the social psychologist could begin his operational analyses: his own teaching. But we did not mean to imply that the classroom set the empirical horizons for his investigations, only that here was rich ground unexplored, and as such, not encumbered with the debris of older analytical techniques. (Unless one would count as hazard the endless attitude studies to which social psychology classes have patiently submitted.) Case studies of individuals made with an eye to the operations performed in gathering and analyzing them should continue to be an important source of data in social psychology.[44] But we suspect that the old pattern of getting life history materials, checking a person's "own story" against whatever outside information we can get on his social milieu, will probably give way before the promising suggestions of recent psychiatric developments.

One very hopeful development along these lines is the psychodrama as it has been used by Moreno. He has demonstrated that it Is not only a good clinical procedure, but has

( 323) fine possibilities as an exploratory and experimental tool as well. In these latter uses it seems that the psychodrama may be tailor-made for social psychological purposes. As a research technique it would permit us to study life histories In process. It can help us both to explore the problem of type situations and to submit our findings to constant experimental check.[45] Nor is its use limited to situations involving interpersonal relations. The work Moreno has done suggests that intra-personal processes, the relations between different rôles within the personality of a single individual, can be projected psychodramatically in a way which no other method allows. Dreams, contemplative thought, perhaps the whole inner forum, can be converted into overt action, where there are greater possibilities for operational check. In the same way rôles that are appropriate to group interaction can be cast within the dimensions of experimental psychodrama, where one can observe the meaning of group allegiances and alignments in individual experiment.

The advantage of experimental psychodrama from an operational point of view is fairly obvious. The social psychologist who sets experiments in social behavior within the psychodramatic framework cannot ignore the operations which he performs. He is himself a part of the psychodramatic experimental situation, and his actions, made overt in his director's rôle, can be observed by himself and others. This fact makes at least this much of the operations he performs subject to repetition and check.

One of the chief disadvantages of psychodramatic work as it has been done to date is that the more subtle aspects of the experimenter's activity have remained buried away from outside observational check. The preliminary rôle-taking he does in order to set up a critical experimental situation, the way he uses his past experience with or knowledge about a particular subject when analyzing that subject's behavior on the stage, have remained hidden as parts of his covert behavioral processes. If social psychologists are to turn the experimental psychodrama to any better operational use than the clinical investigator, they must devise techniques that will bring these covert processes out into the open. One technique that suggests itself is to put the director himself on the stage in a situation he wants to analyze, letting him soliloquize as analyst while he participates in the interaction.


The modern physicist holds no more advantageous position in relation to his data than the position such a procedure would create for the social psychologist. Behavior, Knowledge and Fact, that holy trinity of operational science, may under these circumstances become related for the social psychologist as intimately as they are for the physicist. If such is the case, we might venture a prediction about the kinds of review a trend-finder would write in 1950. His criterion will be neither a theoretical one nor a methodological one. It will be both. The main theme of his work will point to the growing convergence of concept and operation in social psychology, and it will underscore the advances in precision and insight made possible by this convergence. We hope that he may underscore yet another kind of progress, made possible by this first: namely, a clear definition of the rôle of social psychology in relation to the other social sciences. We suspect that out of a decade in which social psychologists have patiently tilled their own garden will come the sort of product which other social scientists can use. Eventually, we feel, social psychology will be the biochemistry of the social sciences, and somehow we hope that this role will become clear enough in the forties to make the deadline for the 1950 review.


  1. See Sociometry, Vol. 4, 1941, pp. 107-139.
  2. George A. Lundberg, "Quantitative Methods In Social Psychology," Am. Sociological Review, Vol. I, 1936, pp. 38-54, with Discussion by Willard Waller, pp. 511-60. For further development and generalization by Lundberg see his Foundations of Sociology, New York, 1939. See also Herbert Blumer, "The Problem of Concept In Social Psychology," Am. Jour. of Sociology, Vol. 45, 1940, pp. 707-719.
  3. G. W. Allport, "The Psychologist's Frame of Reference" (The Presidential Address to the 47th Annual Meeting of the Amer. Psychological Association, September, 1939), Psychological Bulletin, Vol. 37 (I), January, 1940, pp. 1-28.
  4. A number of weaknesses in the sample exist but cannot be discussed here.
  5. Much credit must go to Ellsworth Faris, sociologist, and Gordon W. Allport, psychologist, for constantly emphasizing the fallacies and pitfalls of the atomizing tendencies in social psychology. Cf. Faris' Nature of Human Nature, New York, 1937 for a collection of his articles written during the 1920's and '30's; and Allport's article "The Study of the Undivided Personality," Jour. of Ab. and Soc. Psychology, Vol. 18, 1923-1924, pp. 132-141.
  6. Such insights may of course be the results of the informal correlational processes inherent in any common sense observation of relationships.
  7. See Spearman, C. E., The Abilities of Man, Hew York, 1927, for a summary statement of the development of the theory and method as applied to mental abilities.
  8. The first publication of his theory was made by Thurstone in his article, "Multiple Factor Analysis," Psychological Review, Vol. 38, September, 1931, pp. 406-427. Since that time he has published a series of reports culminating in his two books, Vectors of Mind, Chicago, 1335, and Primary Mental Abilities, Chicago, 1938. For a summary statement of the status of multiple factor theory see L. L. Thurstone, "Current Issues in Factor Analysis," Psychological Bulletin, Vol. 37, No. 44, April, 1940.
  9. See H. F. Gosnell and Margaret J. Schmidt, "Factorial Analysis of the Relation of the Press to Voting in Chicago," Jour. of Social Psychology, Vol. 7, pp. 375-385; and E. W. Burgess and L. S. Cottrell, Jr., Predicting Success or Failure in Marriage, New York, 1939, pp. 317-322; and J. B. Mailer, "Vital Indices and Their Relation to Psychological and Social Factors," Human Biology, Vol. 5, No. 1, February, 1933, for illustrations.
  10. J. P. Guilford and Ruth B. Guilford, "An Analysis of the Factors in a Typical Test of Introversion-Extroversion," Jour. of Abnormal and Social Psychology, Vol. 28, 1934, pp. 377-399. Same authors, "Personality Factors S.E. and M.," Journal of Psychology, Vol. 2, 1936, pp. 109-127. Same authors, "Personality Factors D., R.T., and A., Journal of Abnormal and Social Psychology, Vol. 34, 1939, pp. 21-36. Same authors, "Personality Factors N. and G.D., Journal of Abnormal and Social Psychology, Vol. 34, 1939, pp. 239248. See also, J. C. Flanagan, "Factor Analysis in the Study of Personality," Stanford University Press. 1935.
    See also, T. V. Moore, "The Empirical Determination of Certain Syndromes Underlying Praecox and Manic Depressive Psychoses," Am. Jour, of Psychiatry, Vol. 9, pp. 719-738.
  11. See H. B. Carlson, "Attitudes of Undergraduate Students," Jour. of Social Psychology, Vol. 5, 1934, pp. 202-213. See also, L. D. Whisler, "Multiple-Factor Analysis of Generalized Attitudes," Jour, of Social Psychology, Vol. 5, 1934, pp. 283-297.
  12. Conversation with Ralph K. White regarding the studies of autocratic, democratic and laissez-faire group situations. See R. Lippitt, "Field Theory and Experiment Atmospheres," Am. Jour. of Sociology, Vol. 45, 1139, pp. 26-49.
  13. See Gordon W. Allport, "The Study of the Undivided Personality," Jour. of Abnormal and Social Psychology, Vol. 19, 1924-25, pp. 132-141. Same author, "A Test for Ascendance-Submission," Jour, of Abnormal and Social Psychology, Vol. 23, 1928, pp. 118-136. Same author, "What Is a Trait of Personality?" Jour, of Abnormal and Social Psychology, Vol. 25, 1931, pp. 368-372.
    Philip E. Vernon and Gordon W. Allport, "A Test for Personal Values," Jour. of Abnormal and Social Psychology, Vol. 26, 1932, pp. 231-248.
    Gordon W. Allport, Personality, a Psychological Interpretation, New York, 1937.
  14. A. C. Rosander, "An Attitude Scale Based Upon Behavior Situations," Jour. of Social Psychology, Vol. 8, 1937, pp. 3-16, is another such test in the field of attitudes.
  15. See for example Lois B. Murphy, Social Behavior and Child Personality, New York, 1939.
    For suggestions of how personality characteristics (traits) are functional to positions in situations in the larger social situation, see John Dollard, Caste and Class in a Southern Town, Yale Univ. Press, 1938. See also Allison Davis and John Dollard, Children of Bondage, Washington, D.C., 1940. The field for contextually minded measurers is wide open for measurement of "traits" and attitudes of members of groups In defined positions In intergroup situational fields.
  16. See K. Lewin, Principles of Topological Psychology, New York, 1936. See also J. F. Brown, Psychology and the Social Order, New York, 1936. See also K. Lewin, "Field Theory and Experiment in Social Psychology: Concepts and Methods," Am. Jour. of Sociology, Vol. 44, 1939, pp. 868-896.
  17. See J. L. Moreno, Who Shall Survive? Washington, D.C., 1934, and numerous articles by him and others in the journal Sociometry published by Dr. Moreno.
    See also George A. Lundberg and Margaret Lawsing, "The Sociography of Some Community Relations," Am. Sociological Rev., Vol. 2, 1937. pp. 318-335.
  18. See John Dollard, Caste and Class in a Southern Town, Yale Univ. Press, 1937; Allison Davis and John Dollard, Children of Bondage, Washington, D.C., 1940.
  19. See Allison and Elizabeth Davis and Burleigh and Mary Gardner, Deep South to be published by the Univ. of Chicago Press.
  20. Research in "Yankee City" to be published in several volumes beginning shortly.
  21. See L. L. Thurstone, "Isolation of Blocs in a Legislative Body by the Voting Records of Its Members," Jour, of Social Psychology, Vol. 3, pp. 425-433.
  22. R. Likert, "A Technique for the Measurement of Attitudes," Archives of Psychology, No. 140, June, 1932. See also Murphy, G. and Likert, R., Public Opinion and the Individual, Yew York, 1938.
  23. The first edition, Experimental Social Psychology by Gardner Murphy and Lois B. Murphy was published In 1931 (New York). The revised edition (same authors and T. N. Newcomb) appeared In 1937.
  24. Psychology of Social Norms, New York, 1936.
  25. Op. cit.
  26. Op. cit.; see especially the forthcoming work of Kurt Lewis, Ronald Lippitt and Ralph K. White on autocratic, democratic and laissez-faire group situations.
  27. Explorations In Personality, op. cit. An interesting feature of the work reported in this book Is the effort to develop categories of Interpersonal relationships rather than rely on trait-attribute categories. It seems to us that the theorizing is handicapped somewhat by leaning too heavily on the Freudian theory of personality organization. The work of Sullivan, Moreno and Horsey would add a good deal to the Insights of the Harvard Clinic group.
  28. Op. cit.
  29. "Determinants of Prediction of Social Events," Jour of Abnormal and Social Psychology, Vol. 33, 1938, pp. I79-204.
  30. "The Prediction of Social Events," Jour. of Abnormal and Social Psychology, Vol. 33, 1938, pp. 364-389. These experimental studies of prediction which pay attention to the degrees of clarity and ambiguity of the situational field and the correlated degrees of projection by subjects of private "definitions of the situation" are very significant for the study of any act-in-situation, since every social act Involves a "prediction."
  31. "Experimental Studies of Projection: I. Attribution of Traits," Jour. of Social Psychology, Vol. 7, 1936, pp. 151-163; 11. "Ideas of Reference," same journal, Vol. 8, 1937, pp. 389-400.
  32. Hamilton, G. N., An Introduction to Objective Psychopathology, St. Louis, 1925. Same author, "Comparative Psychology and Psychopathology," Am. Jour. of Psychology, Vol. 39, 1927.
  33. Some representative work is found in: Hall, Calvin, "Neurotic White Rats," Jour, of Abnormal and Social Psychology, Vol. 28, 1933, pp. 1-5. Liddell, H. S., "The Experimental Neurosis and the Problem of Mental Disorder," Am. Jour, of Psychiatry, Vol. 94, 1948, pp. 1035-1041. Murchison, C., "The Experimental Measurement of a Social Hierarchy" In Gallus Domesticus. See Jour, of Social Psychology, Vol. 6, 1935, Pp. 3-30; 172-181; and Vol. 7, 1936, pp. 3-18, for three articles on this work. There are several others referred to in these articles. Maier, N. R. F., "Experimentally Produced Neurotic Behavior in the Rat," paper at the December, 1938 meeting of the A.A.A.S.
  34. Some of the positions we maintained In our review have subsequently been corroborated by others, particularly Ernest Beaglehole, "The Implications of H. S. Sullivans' Interpersonal Theory for Social Psychology," Psychiatry, Vol. 4, No. I, February, 1941, and E. B. Reuter, "The Status of Social Psychology," Am. Jour. Soc., Vol. XLVI, No. 3, November, 1940. Reuter's article was written in the spring and Beaglehole's in the summer of 1940, but neither was available to us at the time of writing our review.
  35. Sociometry, Vol. 4, No. 2, May, 1941, Pp. 140-149.
  36. Philos, of Science, Vol. 8, No. 2, April, 1941.
  37. Note that our treatment of experimental and quantitative developments was made from the point of view of orientational shifts, and not concerned with the problem of method, as such.
  38. Credit for this shift in perspective is due chiefly to Talcott Parsons whose excellent critique of this paper as it was originally presented before the S.R.A., defined the role of social psychology in terms of method. His paper, (mimeographed, Cambridge, Mass.) held some invaluable suggestions as to the larger meaning of the trends we noted in a general "action schema" for studying human behavior.
  39. Talcott Parsons: op. cit. See also H. S. Fries, 'Physics, a Vicious Abstraction," Philos. of Science, Vol. 6, 1939, pp. 301-308.
  40. A. F. Bentley, "Skin and Philosophy," Philos. of Science, Vol. 8, No. I.
  41. A bibliography on this position in modern science would run to many titles. We suggest the following as a basic minimum: G. A. Lundberg, Foundations of Sociology, New York, 1939. W. M. Malisoff, "Emergence Without Mystery," Philos. of Science, Vol. 6, 1939, pp. 17-34. L. A. White, "Science is Sciencing," Philos. of Science, Vol. 5, 1938, pp. 360-388. P. W. Bridgeman, "Operational Analysis," Philos. of Science, Vol. 5, 1938, pp. 114-131.
    See also footnote on page 109, Sociometry, Vol. 4, No. 2.
  42. Observation of student reactions at Cornell seems to indicate that there Is. Many students find it difficult to maintain the "just another course" attitude toward social psychology. There is too much of an all or none acceptance or rejection, both of the material presented and of the teacher himself, to Indicate a psychologically neutral subject matter. The growing tendency to describe psychiatric treatment as process in interpersonal relations will sooner or later provide a picture which we can use for comparison on this problem.
  43. For a fuller discussion of this point, see L. S. Cottrell, Jr., "The Case Study Method In Prediction," a paper delivered before the Sociological Research Association, September, 1941 (mimeographed, Ithaca, N.Y.). In this paper a combination of sources and methods are recommended for arriving at operationally valid type-situations. They include The Cross Cultural Survey at Yale, the experimental psychodrama, factor analysis of words used In everyday speech to describe situations, Investigation of extent case studies, refinement of case study techniques, etc.
  44. Ibid.
  45. J. L. Moreno, "A Frame of Reference for Testing the Social Investigator," Sociometry, Vol. 3, 1940, pp. 317-327.

Valid HTML 4.01 Strict Valid CSS2