A Discussion of the Issues: (I)

Muzafer Sherif



I would like to make an over-all remark about the general methodological problems of which we should always be mindful in defining research topics. In my opinion, this consideration should be at the core of this Symposium.

Our problems in social science today are not new problems. They are as old as human beings themselves. For example, such topics as the structure of a human relationship, leader-follower relations, intergroup relations, interpersonal attitudes, communication, and so on, are as old as human history. Moreover, people did not wait for social science to be concerned about these problems before they theorized about them and discussed them. By their very nature, these are vital problems to human beings.

There have been solutions offered for them on a religious basis. In fact, all religions have their own systems which define what human relations are and what they should be. Philosophers from at least the time of Plato to the great systematizers of today have offered answers for these same problems. The distinctive claim of the social sciences is that systematization should rest on available empirical findings and should be capable of predicting new findings. As a result, some policy-makers have turned their attention to the social sciences in the hope that they may be helpful.

For example, in this shrinking world today, the over-riding problem is intergroup relations. Problems of war and peace, which directly or indirectly include most of the problems of

social science, have led some policy-makers to support social science research in the hope that we operate like engineers. If a bridge is ordered, engineers are supposed to design and build it on the basis of established laws in the physical sciences. Nowadays the social sciences are expected to perform in a similar manner. However, the policy-makers and the foundations that support social science research as if it were engineering do not know that, as yet, we have very few established principles on which to base predictions. Even the basic definitional problems have not gone beyond the stage of controversy between more or less closed

( 209) intellectual schools. If we are clever, let us keep that fact to ourselves; but let us not delude ourselves in the process.

Every social science discipline has found out by now that it is not a law unto itself. No one of these sciences is sufficient for the huge task of dealing with human problems single-handedly. Interdisciplinary research is needed as well as interdisciplinary theory. But what kind of interdisciplinary research and theory?

One interdisciplinary orientation centers under the banner of "behavioral science," a term used very loosely and meaning all things to all people. This orientation frequently over-stresses behavior as the datum of social science. The logical outcome of this emphasis, of course, would be to make psychologists of everybody. And, in fact, it is not difficult to find, for example, social scientists studying a hospital ward who have become so over-awed by psychology and psychiatry that they adopt the jargon of those disciplines. If these social scientists would stick to their problems at the social science level, they could make a greater contribution to interdisciplinary study, and, let me add, they would be in a much stronger position in the establishments where they work. For an interdisciplinary approach to become effective, it is much better to develop first-rate social scientists instead of turning them into second- or third-rate psychiatric workers or psychologists.

Some other interdisciplinary efforts seem to proceed from the assumption that just because a number of specialists from different disciplines talk about the same problems the result will be interdisciplinary. As my illustration, I shall take---with respect— one of the more ambitious theoretical undertakings of our time, Parsons and Shils' Toward a General Theory of Action. Before commenting on this work, let me hasten to say that there should be more and renewed attempts such as this on a grand scope. Hence, my evaluation does not concern the desirability of the attempt as such, but the particular contents between its two covers.

More properly, I believe, this book might have been called something like Views on Social Science and Related Topics by Some Eminent Scholars. Parsons and Shils have written not a

(210) theory, but a categorical scheme. One can search far and wide within their book for "action" and one finds not "a theory" but several very different theories by some of the most distinguished psychologists and sociologists of our day. For example, one of the authors, Sheldon, dissents from the very basic proposition that culture "as a system is on a different plane from personalities and social systems."[1] Another author, the anthropologist, Kluckhohn, disagrees with the separation of social structure and systems, on the one hand, and culture, on the other.[2] The distinguished psychologist Tolman focuses on learning in a way that may be congenial to other authors but has very little to do with what they write.; An equally eminent psychologist, Gordon Allport (by the way, a teacher of mine), treats prejudice but specifically denies the applicability of a general theory to the phenomenon "for the time being."[4] Robert Sears, cautions that "an appallingly small number of the relationships that have been discovered in social psychology can be generalized. . . . With respect to attitude measurement, for example, one might well ask whether any general principles of an antecedent-consequent nature have been found. [5] And, finally, Stouffer, an empiricist who developed outside the grandiose theorizing, warns against "an excess of optimism about the necessary usefulness of any new system of highly general orienting ideas ." [6]

What is wrong here? Two things, I think. One is that talking about the same or related problems does not necessarily build a coherent interdisciplinary theory. The other is that general theory presumes the incorporation of findings from all social science disciplines into an integrated system. We preach and preach about empirical research, with the aim of attaining reliable and valid indicators as the end product of conceptual definitions and research operations-just as Professor Lazarsfeld has been insisting. Yet some systematists take too little stock of empirical results being accumulated in different social sciences on the same problems. Theorizing is necessary. But, as I tried to stress in my paper for this Symposium, theorizing must be checked through the evidence that is accumulating at the time.

Let me illustrate one of the things that needs to be done. One of the principles that everybody preaches-in different schools of

( 211) psychology, sociology, and anthropology-is that a system affects the properties of its parts. Psychologically, this has been established time and again-by Gestalt psychologists, by Bartlett, and by many other investigators. But when we engage in interdisciplinary study, we act as though this principle were unimportant. We compartmentalize our tasks and our findings and say, "Oh, I'm a psychologist. The sociology part is being done in New York or Buffalo or somewhere else. When we are through, we will put the studies side by side." The researcher cannot operate this way. The theorist cannot operate this way. If the aspects being studied are parts of the same system, there has to be coordination from the very beginning, so that the different parts will fit together as they do in the system.

If the system affects the properties of its parts, the systematizer cannot give mere lip service to this principle; he must translate it into logical operations in terms of concepts, procedures, and measures. For example, when we study the roles of individuals in a group, we first have to place the group in its cultural setting. Then, in a definite sequence, we are prepared to study the group as a system, and only then to understand the roles of particular individuals. The point that I am trying to make here has been made for many years by many social scientists, at least in words. But when we get down to the real brass tacks of research, its implications are either forgotten or muffled somewhere. For example, in theorizing about youthful sub-cultures and misbehaviors, there has been a pronounced tendency to use the concept of "reaction-formation," which is borrowed from a highly individualistic psychological system, instead of bringing the behavior into the context constituted by the gamut of influences from the general social system and noting the similarities as well as differences among youth in different socio-cultural settings within the system.


  1. Richard C. Sheldon, in Talcott Parsons, et al., "Some Fundamental Categories of the Theory of Action: A General Statement," in Talcott Parsons and Edward A. Shils (eds.), Toward a General Theory of Action (Cambridge: Harvard University Press, 1951), p. 7, n. 9.

  2. Clyde Kluckhohn, in ibid., pp. 26-27, n. 31.
  3. Edward C. Tolman, "A Psychological Model," in Parsons and Shils, ibid., pp. 277-361.
  4. Gordon Allport, "Prejudice: A Problem in Psychological and Social Causation," in ibid., p. 384.
  5. Robert R. Sears, "Social Behavior and Personality Development," in ibid., p. 466.
  6. Samuel A. Stouffer, "An Empirical Study of Technical Problems in Analysis of Role Obligation," in ibid., p. 480.
  7. Samuel A. Stouffer, et al., The American Soldier (Princeton: Princeton University Press, 1949).



Valid HTML 4.01 Strict Valid CSS2