Actors and Systems: The Politics of Collective Action

Michel Crozier and Erhard Friedberg Translated by Arthur Goldhammer

The University of Chicago Press Chicago and London

Originally published as L'acteur et le système, ©1977 by Editions Seuil

The University of Chicago Press. Chicago 60637
The University of Chicago Press, Ltd., London
©1980 by The University of Chicago
All rights reserved. Published 1980
Printed in the United States of America
87 86 85 84 83 82 81 80 54321
Library of Congress Cataloging in Publication Data
Crozier, Michel.
Actors and systems.
Translation of L' acteur et le système.
Bibliography: p.
Includes index.

  1. Organizational behavior. 2. Organization.
    I. Friedberg, Erhard, joint author. II. Title.
    HD58.7.C7613 302.3'5 80-13803
    ISBN 0-226-12183-6

MICHEL CROZIER iS the director of the Centre de Sociologie de Organisations de Paris and the former president of the French Association of Sociology. He is coauthor of The Crisis of Democracy and author of The Stalled Society, as well as the author of The Bureaucratic Phenomenon and The World of the Office Worker, both published by the University of Chicago Press. ERHARD FRIEDBERG, a fellow of the International Institute of Management in Berlin from 1973 to 1976 is assistant director of the Centre de Sociologie de Organisations de Paris.


This book was written for a French audience. Given the usual transatlantic misunderstandings, the English-language edition requires a special Introduction. Our intellectual trajectory is very different from that of our American colleagues. We started from the American organization theories of the fifties and in the course of trying to apply them in the interpretation of French realities, came to realize that we needed new theories. The limitations of the old paradigms were apparent in their inability to account for conditions in France, but the significance of this failure was even more profound. We came to believe that it revealed the fundamental inadequacy not only of the original models of organizations and systems theory but also of that of their contemporary American successors.

Despite its past glories, French sociology had subsided into a major lull from the twenties to the fifties. It began to flourish again in the fifties and caught up with the American-dominated international quantitative mode only in the mid sixties. ==In the field of organizational study there were two starting points: the Merton-Gouldner-Selznick sociological model, and the Michigan psycho-sociological model==. But French researchers working from these models faced very specific cultural constraints. Trade unions, for example, were much more resistant to sociological surveys whose claims to neutrality remained to be demonstrated. More important was the tremendous gap within French management between theory -- which tended more and more to be that of the American business schools -- and actual practice. Under these circumstances the American paradigms could not be successfully applied. As usual in new scientific fields, misunderstanding and failure were strong incentives to break new ground. The elaboration of new paradigms led once again to major differences between European and American approaches. The issue, however, is not only one of European versus American sociology but of alternative approaches to the study of organization. This book presents one such approach developed in the course of twenty years of field research in France.

During most of this period the dominant Anglo-Saxon mode has been theoretical regarding the aims, quantitative regarding the methods.(1) Sociologists, particularly Americans, had become confident enough to believe that they could build theory or present theoretical propositions of universal value by comparing quantitative results derived from statistically significant variance of crucial indicators within samples of organizations. These theories, even when they were more sophisticated than the very crude first organizational laws proposed in the late sixties(2) could not be applied to the French case. Viewed from France, even the dominant structural-contingency theories which appeared later were unduly deterministic.(3) To account for French empirical evidence and to obtain significant data, one had to develop a very different research strategy, one that led gradually to the elaboration of new paradigms and to the questioning of the American paradigms for the American context.

We believe that the social sciences have not progressed sufficiently in this field to make it possible to use such sophisticated instruments as our American colleagues use despite the poor quality of the very fuzzy indicators they are able to define. In our view research strategy should focus on organizations as distinct autonomous phenomena, but without treating them as quantitative units. We also believe that ethnomethodologists, even when they use the same empirical evidence as we do, do not provide a satisfactory alternative. They tend to extrapolate from the description of primary face-to-face encounters to generalized propositions about global society. Not only do they ignore the goals and results of human action, but they bypass the complex interrelations at the organizational or at the organized-systems level -- market or institution -- which do influence and give meaning to individual interaction.(4) For us, organizations and organized systems have enough autonomy to make it impossible to understand their functioning either from a deterministic standpoint or as an extrapolation from primary group interaction. ==Our research strategy therefore consists, first of all, in using individual behavorial data. These indicators are the only reliable evidence for understanding not only the cultural or psychological forces but the specific "human construct" which our interviewees confront daily, namely, a single organization or a broader organized system.== Instead of considering organizations as neutral transmission belts of human wishes, we assume that such transmission belts are more important than their motors. Our focal point is what Chandler has very aptly called the management of the throughput between the technological input and the economic results.(5) This may seem to bring us back to an earlier period when case studies were basic research strategy. Although we do appreciate and revere such works of the fifties as those of Gouldner, Selznick, and Whyte, we nevertheless have quite different concepts and methods.

The new paradigms we shall present and subsequently test on specific cases in this book can better be understood by examining ==our basic concepts: collective action, games, uncertainty, and power.==

The concept of collective action is not original, of course, especially for organization theory. Nevertheless, its full implications have not been explored in the sociological literature.(6) It seems crucial to return to them in the present circumstances for one basic reason: when one begins from the collective-action standpoint, human behavior appears as neither a logical or necessary outcome of social interaction nor a result of the structures of the problems to be solved. The means and devices by which collective action is brought about can therefore no longer be considered as some thing obvious or natural (in the sense of "the nature of things"). It be comes a problem for research and explanation.(7)

Organizations are not "natural" beasts whose existence is a given of nature. They are "human constructs" developed half consciously, half unconsciously, by man to solve problems of collective action, and above all the most basic of these -- cooperation for the production of some collective good by relatively autonomous social actors pursuing diverse and always, in a certain sense, conflicting interests. The most urgent question posed by the existence of organizations, therefore, is not what kind of formal and structural regularities exist in organizations. A basic paradox of collective action, the counter-intuitive effect, may be presently considered, in that respect, the first useful paradigm of organization and system theory.

The counter-intuitive effect arises from the logical structure of the problem. If one looks further, however, it is also, and may be principally, an organizational or systemic effect. When results of collective action are the opposite of what individual actors wished, it is never merely because of the properties of the problem. It is always also the result of the human structuration of the field of action, that is, because of the characteristics of the organization or of the system in which the actors interact and to which they are committed.

Moreover, the real advantage of organizations and organized systems may lie in their ability to offer a more reliable and more useful way to structure the human field of participants, actors, and clients than that offered in a nonorganized field. In other words, organizations control the counter-intuitive effects inherent in problems. But when organizations regulate these first-order distortions, they produce second-order distortions of their own, which are those we usually discover first in practice.

For the creators of the concept of collective action, the counter intuitive paradox was only logical. We believe it has a sociological, that is, a "human construct" dimension -- not because management problems would tend to be strongly influenced by "natural sociological''(8) determinants, but because very often it has become a social or organizational paradox. We therefore use the concept of collective action as yet another device to help people -- and many an organizational sociologist as well -- understand that organizations are extraordinarily binding human constructs, but that they are also contingent. In them there never was and never will be either one best way or the best contingent solution. Men build organizations to solve problems otherwise intractable. Before measuring the instrument or even trying to improve it, one should consider the actual problem to be solved.

But sociology has yet another contribution to offer. Organizations provide a crucial area in which the link between material results and human wishes may best be studied.

Let us recall, for the sake of clarity, the dilemma faced by two criminals arrested for the same crime but with no material evidence against them other than what might be elicited by the one's informing against the other. In this situation, each of the two prisoners has only two strategies for extricating himself: to deny the facts, or to incriminate the other. If both deny the facts, the police have no proof for their charges; all they can do is get the defendants convicted of lesser charges and sentenced to a year in prison. If one of the defendants agrees to turn state's evidence by in forming on the other, the informer goes free, while the other is sentenced to twenty years in prison.

Aware of these consequences, each prisoner knows that the success of his own strategy will depend upon the strategy adopted by his accomplice. But since they are held in different jails, they have no opportunity of communicating and cooperating with each other. The logical structure of the problem is such that if they act "rationally," that is, serving their individual interests, each will betray his accomplice and both will find themselves back in prison for ten years. It should be emphasized that no judgment on "human nature" is implied here; it is merely assumed that each defendant will try to win and will place his own interests above those of his accomplice. Knowing the result makes no difference; both are caught in an "infernal logic" that leads them inevitably to defeat and is a consequence of the structure of the problem.

Only one thing can free people from the dilemma: the capacity to trust one another. Although this presents an ethical dimension, ethics here, as is often the case, is a result of a "social construct." The unwritten laws of the underworld are a good example. The law of silence may be subsumed, as it is in many romantic movies and thrillers, as an ethical principle; it is first of all a human construct achieved by training and sanction. What seems simply affective retaliation has become an organizing principle which is manipulated by the leaders of the underworld and which will be the cornerstone of whatever organization may be developed around these activities.

The free-rider problem can be analyzed in an analogous way. As trade unions everywhere well know, organization is the only way to force free riders to pay for their benefits.(9) More generally, organization is a way of dealing with the otherwise insoluble problem of aggregating interests whose contradictory pursuit would otherwise frustrate all participants. The trade union is, for example, built on such unnatural social behavior as trust, loyalty, accountability, responsibility, and so forth, all qualities whose social character is suggested by the sacred aura they still had in the early Middle Ages and by the very strong moral -- that is, nonrational -- connotation they retain now.

How is it done? Not by using these emotional moral feelings directly. These are indispensable props that make it possible for people to interact and cooperate, but they do not dictate behavior. They are merely the consequences of organizational life and, conversely, the condition for its development. They certainly do not integrate the conflicting strategies of the actors involved. This is accomplished through the creation and use of structured games, which must be considered as the real discovery of man as an organizational animal.

To engage in collective action, people pursuing contradictory interests have either to be coerced (or manipulated ideologically or affectively) or to bargain. Bargaining, first of all, is itself unnatural, and people do not engage in it without protective devices. Its possibility hinges upon the existence of those unnatural "moral" feelings we have just mentioned, and even then, it always means facing up to power relationships and to the constraining consequences of one's own course of action. Bargaining, moreover, is threatening, as its dynamics can operate in such a way that all parties involved may suffer.(10)

A good game is the real organizational trick. Suppose the field is structured and the problem redefined so that people can still pursue their own interests, while their winning and losing will not alter the collective good -- will, in fact, improve it. We then avoid any kind of prisoner's dilemma; cooperation is achieved without suppressing people's freedom, that is, their right to pursue antagonistic interests.

One may object that reality is not so simple. Fair enough; this is why organizations, when considered as a human construct, appear such difficult, fragile, and almost impossible endeavors. Considering the context not only of bounded rationality, but also of bounded legitimacy (in complete normative integration of members) and of bounded interdependence (incomplete functional integration of members), it is a miracle that organizations exist at all. Indeed, one could say that they do exist not so much because, but in spite of, the strategies of their members. The solution to this apparent paradox -- and also a key to an understanding of an organization -- lies in the analysis of the different power games which indirectly structure the strategies of the actors involved and constrain their freedom of choice.

When thinking about games, however, people have been too obsessed by the problems of the rules. Rules are necessary to maintain the game, to make it possible for people to be sure that no cheating will jeopardize their stakes, but they do not command behavior.(11) People are channeled, oriented, and led not by the rules but by the structure of the game, which determines a range of possible strategies and results. The game remains open, and several strategies remain possible, among which the players must and will indeed choose. Moreover, a player can even -- if his resources allow it -- accept initial losses in the hope of a later reversal of the game which will be profitable to him. In any case, the constraint here is always of an indirect nature; it is the result of the simple fact that so long as a player wishes to continue to play, and in order for his participation in the game to be profitable to him, he must adopt some strategy whereby he will contribute to the continuation of the game and thus, nolens volens, to the production of the collective good.

What about coercion or ideological or affective manipulation? We do not deny that they exist and constitute a means to enforce the rules and to command behavior, but we rather analyze them as another kind of game, a very poor one indeed, which is at best a zero-sum game and which may indeed have negative results.

Now let us examine the paradox with which we began. Organization is a human construct with which people hope to solve the problems that remain intractable so long as one is limited to the singular logic of collective action. With the help of such a construct, people do not get rid of counter-intuitive effects, but replace one kind of negative counter-intuitive effect with another, more positive, kind.

Two questions then arise. First, what kind of a link exists between the logical and the organizational structure? Second, what about the new counter-intuitive effects?

The first question brings us back to the results of organizational analysis. ==We propose uncertainty as the key concept of the logical structure of the problem to be dealt with and power as the key concept for the human construct that will have to deal with it==.

There are no social problems without uncertainty, none which do not, in other words, leave the actors with a substantial margin of liberty in the choice and implementation of practical solutions. Problems without uncertainty are not really human problems. They can be dealt with by automatic or cybernetic (in the sense of self-regulating) devices.(12)

Uncertainty, however, is the basic resource in any kind of bargaining.(13)

Here we can find the basic link between the two structures. People are unequal before the uncertainties of the problem. Those who can cope with them because of their situation, their resources, or their competence have an advantage.

The logical structure of the problems as embodied by individual or collective actors thus unavoidably becomes a structured field. This means that real solutions will develop along three possible lines: 1. No bargain can be struck; the logical structure cannot be dealt with by the concerned players as they can presently organize; the problem will be ignored. 2. Bargaining will develop empirically along lines corresponding to the structure of the problem and to the personal and social strengths and capacities of the players. 3. An organization already in charge will handle uncertainty according to its own capacities; it will, that is, develop a game that will be its contingent solution.

The third eventuality is, of course, the crucial one for the discussion of formal organizations. Organizational games are built around logical, "objective" uncertainties, those of the techniques, of the markets, of the different constraints which are, at least for the short-run, stable, unshakable givens.(14) Those who get the upper hand in the game are those who control most of the crucial uncertainties. But the way the game is structured and the stakes determined will reduce the possible gains and losses to acceptable proportions. Artificial uncertainties such as authority distribution, information channels, and legal constraints, can, on the other hand, counterbalance the "objective" ones. Indeed, the creation of such "artificial" uncertainties will be a precondition for the establishment of a game. In any case, this means that to be processed, problems will have to be redefined either to fit prevailing games or at least to allow for the creation of some artificial uncertainties without which no bargain can be struck. Finally, most formal organizations cannot be understood on the basis of one game that explains them entirely. There are many games and layers of games. Part 1 and 2 of this book will deal with formal organization games. Part 3 will deal with organized systems which correspond to the second eventuality.(15)

One last point: what is uncertainty in the logic of the "objective" problem is also power from the point of view of the actors and for the organization. Organization as a human construct deals basically with power.(16) It is organizing, regularizing, taming power that makes it possible for men to cooperate. Organizing, however, is a difficult endeavor: if it goes too far in trying to suppress empirical "natural" power, it will risk raising artificial power to an unbearable degree; if it does not try hard enough, a growing maze of privileges will develop.(17)

This leads us to the problem of the counter-intuitive effects of the organizational construct. They are the immediate consequence of the course taken. Each organizational solution, because it deals directly with the most threatening element of human interaction, power, raises problems of communication and commitment. Because power is involved, threatening information will not be processed. Bureaucratic dysfunctions will develop. Vicious circles of low commitment, poor communications, and inadequate performances will develop. Efforts to shake them will bring back privileges and will initiate other vicious circles. New games will emerge, games of privilege and cheating. We will then have a real organization, the human construct of the practitioner, not the one with which the theoretician is familiar.

From this standpoint organizations and organized systems can be considered as, at the same time, the necessary means and the basic operational constraints of collective action.

Now we come to the second question: What about the new counter-intuitive effect? Organizations and organized systems generally appear to be the basic operational constraints of collective action. From a short-run perspective, at least, collective action takes place only through already existing organizations, or at least through stable organized systems, the underlying games of which mediate between the means and the ends of collective action. Substitutions among these organized structures are possible, but on the whole there is not much possibility of selection. We are in a sellers' market: many problems and few solutions, which in turn implies that, in the short run at least, means have more importance than ends. Ends are chosen to a surprising degree because of the availability of means. Very often means command over ends.

Of course, new organizations can be built; new systems to regulate cooperation and conflict can be achieved; even new markets can be developed. But it takes time. Achieving flexibility through all-purpose or very malleable organizations can be hoped for, and there is certainly a good deal of progress in that direction. But we generally remain within the domain where constraints prevail.

In the past, the counter-intuitive dimension of organizational life was obscured behind religious, moral, and patriotic taboos. People played games, of course, but they would not admit this. The gradual disappearance of the aura of "natural law" which hung over this order of things, which we now begin to think of as a construct, has led to an intellectual as well as a moral crisis. While the contingent, that is, man made, nature of existing arrangements becomes more visible and thus more difficult to legitimize, our intellectual tools -- and above all the still dominant mode of reasoning from ends to means -- become more in adequate to deal with our difficulties.(18)

==Behind this crisis and the revival of half-baked utopias such as the transparent society, the self-management society, the relational society, as well as such scientific naiveties as cybernetic models,(19) there lies a real organizational crisis, for conditions of organizational life are in the process of radical transformation==. The basic game patterns we had been accustomed to and which were, to some extent, self-perpetuating, have been upset. New patterns do not emerge easily. Our capacity to handle problems decreases, while the pressure for solving the ever increasing number of them rises considerably. The main reason for this new course can be traced back to the impact of the very success of economic development and the rapid increase in human interactions and communications which that development brought about, on the basic bargaining relationship of the individual with an organization or with any established authority. In this respect, one can never emphasize enough the fundamental changes that take place when the existence of real alternatives allows the individual to take risks, for as soon as the freedom of choice of the individual has passed a certain threshold, the structure of the games changes. Once the individual can choose effectively, once he has alternatives, he can and will bargain differently. Not only can he compare overall alternatives, but he can bargain on a day-by-day basis. Government by fragmentation, isolation, and control of possible alternatives will give way. To give but one example: once the parish priest can really opt out of his institution without being sanctioned for his deviant behavior, as he traditionally would have been, the ecclesiastical authority of the bishop lost much of its power, and the traditional game between the two changed. Value changes may eventually follow these more basic changes. But in this case, as in many others one could think of, traditional authority disintegrates not so much because of value changes as because of the changing structure of the conventional game.

This development forces upon organizations an enormous and hardly manageable increase in complexity in two ways. Those who can choose among several games are much less predictable and therefore cannot be led as easily. Moreover, when a choice exists, the structure of the problem itself becomes more complex. The communication explosion, which can be viewed as a consequence of freedom and complexity and which feeds back on both, has also had its own special impact. It helps change the balance of the bargaining, since communication is a basic resource of the power play. But it also invariably changes the problem itself; and by forcing the counter-intuitive effects into the open, it is at the root of the intellectual and moral paradox which we have presented and which brings another form of pressure upon the traditional order of things.

All these converging pressures and the vicious circles that tend to disintegrate old forms of organizational games can thus be viewed as the results of an irrevocable trend toward more complex and more open kinds of collective action. This should be considered with hope, or at least with equanimity. Man will be forced to invent new games and more flexible organizations and organized systems. He is gradually liberating himself from the constraint of the means, of which he had been a prisoner. The accompanying moral crisis, the exaggerated hopes and fears we are suffering, may be simply episodes of false consciousness that were to be expected.(20)

This may be true in the long run, but one ought to beware in the short run of the sheer limits of the problem structure and the extreme difficulties one meets in developing these exceptional human constructs: new games, new organizational forms, more sophisticated organized systems. Here, we feel, we converge very much with the seminal work of Argyris and Schon on double-loop learning which, for us,(21) revives the best tradition of American pragmatism. The salience of the problem does not make it easier to tackle. We will discuss in the last part of this book the problem of change(22) and the contribution the sociology of organizations could make and should make to it. Two preliminary remarks seem pertinent here.

Present circumstances increase the relevance of organizational analysis and at the same time make it more difficult. The disappearance of ritual forms and traditional smoke screens has eliminated a great many of the old taboos. Authority, especially, has been exposed in a ruthless way; it has become quite easy to get at it, now that it has lost all its earlier paraphernalia. But a last taboo remains: power, a taboo which indeed is most stifling for the analysis of organized action.

Some people, including many an American sociologist, may be puzzled by this proposition, since they implicitly equate power with established authority. But this is exactly what has to be questioned. ==Power relations and the problems they raise are not merely the result of superimposed authority structures, be they organizational or social, that is, the expression of patterns of social domination. They are a basic and ever-present ingredient in any kind of social relation which can always be analyzed as an embryo of collective action involving power and influence bargaining.(23)== There is no getting around it: any attempt at defining away the phenomenon by labelling it differently (as, say, influence) and by subordinating it (that is, making it a function of some superordinate mode of domination)(24) does not change ==this basic fact: power and its hidden counterparts, manipulation and blackmail, are unavoidable components of any collective endeavor==. As such, they raise concrete problems that have to be handled. However critical we must be of the traditional and present hierarchical structures built to deal with these problems, getting rid of the former will not be enough to bring about a solution for the latter. To solve these, men will have to find new ways, develop new games, to handle power. And the first way in which social science might make a real contribution is to break the taboo of power by dealing with it in a more detached and scientific way, and thus to train people to face the problem by themselves.

This brings us to our second remark concerning the methodology of organizational research. When considering the problem as it has been outlined, especially in view of the strength of the power taboo among politically progressive people as well as among conservatives, we believe that the basic strategy of organizational research should revolve around organizational analysis understood in an empirical way.(25)

It seems obvious that we should first of all try to understand present power arrangements as the key to these contingent constructs we call organizations. Though this should be simple enough, it seems to present some difficulties. When one realizes that few organizational papers published in American journals have been written by people who have had personal contact with the organization studied, and that few of the re maining articles have dealt with any kind of power issue, one can under stand why American sociology has lost so much ground and why a change of paradigm in Kuhn's sense is strongly needed.(26)

In a very provocative recent book Roberts, Hulin, and Rousseau(27) discuss the "sense of discouragement" with which one confronts the present state of knowledge in the field of organizational theory. "The continued emphasis on narrower and narrower views of responses made by individuals in organizations, concentration on individual differences to the exclusion of environmental effects or concentration on organizational variables to the exclusion of individual differences or societal variables will generate more precise knowledge about increasingly trivial matters."(28)

In this book we seek to establish a high priority for empirical analysis over any kind of hypothetico-deductive theory. This does not mean discarding theory but using it to enable us to get at the facts and question them in a more sophisticated manner. This means that we are trying to elaborate a new paradigm which should help overcome the present complacency and discouragement so prevalent in the American sociological milieu.

This of course is only a first attempt and cannot be a refined intellectual system. But in the present state of the art, we believe that explorations are more useful than formal coherent theories.(29)

Introduction Footnotes

  1. There had been, however, a timely warning against this trend, especially in B. G. Glaser and A. L. Strauss, The Discovery of Grounded Theory (London: Weidenfeld & Nicolson, 1968).

  2. As in P. M. Blau and R. Schoenherr, The Structure of Organizations (New York: Basic Books, 1971). The vitality of Blau's flimsy model, even after the sharp and pertinent rebuke of Chris Argyris (1972), is attested to by the recurrent publication of laborously sophisticated articles in the Administrative Science Quarterly. See for example Robert Dewan and Gerald Hage, "Size, Technology and Structural Differentiation: Towards Theoretical Synthesis," Administrative Science Quarterly, March 1978, and Janice Beyer and Harrison M. Trice. "A Reexamination of Relations between Size and Various Components of Organizational Complexity," Administrative Science Quarterly, March 1979.

  3. As shown by Paul Lawrence and Jay Lorsch, Organizations and Environment (Cambridge, Mass.: Harvard University Press, 1967), and John Child, "Strategies of Control and Organizational Behavior," Administrative Science Quarterly, March 1973.

  4. Not only H. Garfinkel, Studies in Ethnomethodology (Englewood Cliffs, N.J.: Prentis Hall, 1967), and A. V. Cicourel, Method and Measurement in Sociology (New York: Free Press, 1964), but also E. Goffman, Asylums (Chicago: Aldine, 1961), and Howard Becker, ed., The Other Side: Perspectives in Deviance (New York: Free Press, 1964).

  5. Alfred D. Chandler, Jr., The Visible Hand: The Managerial Revolution in American Business (Cambridge, Mass.: Belknap Press, 1977).

  6. As a matter of fact, it is only during the last years that the basic paradigm of Mancur Olson (The Logic of Collective Action, Cambridge, Mass.: Harvard University Press, 1965) has begun to be used seriously by young American sociologists. See for example the seminal article by Mark Granovetter, "Threshold Models of Collective Behavior," American Journal of Sociology, vol. 83, no. 6 (1978), and the excellent case study by Gerald Maxwell and Ruth Ames. "Experiments on the Provisions of Public Goods: Resources, Interest, Group Size, and the Free Rider Problem," American Journal of Sociology, vol. 84, no. 6 (1979). In both cases it is characteristic, however, that the references are only to economics and that the reasoning does not account for the organizational or even social construct.

  7. Here our debt to Herbert Simon is great. We feel much closer to some American political scientists and economists (such as Herbert Simon, James G. March, Tom Schelling, and Graham Allison) than to most of our fellow organizational sociologists.

  8. The two words should of course be antinomic, but we think that in many of our present trade discussions they are almost equivalent: God-given facts man has to cope with. The same remark is made by Andrew Van de Ven in his extremely pertinent critique of Howard Aldrich's book, Organizations and Environments (Englewood Cliffs, N.J.: Prentice Hall, 1979).

  9. See the whole discussion of public goods in modern economics and especially G. Hardin, "The Tragedy of the Commons," Science, December 1968, p. 1243, and Gerald Maxwell and Ruth Ames, "Experiments on the Provision of Public Goods," American Journal of Sociology 84, no. 6 (1979):1335.

  10. Collective bargaining in many Western countries is a good current example. It can very well be understood as a mild prisoner's dilemma whose German or Swedish solution of complicity seems to be giving better results than working-class purity (and bourgeois egoism) of Italian or English partners.

  11. Cheating may be understood as another free-rider problem, a but as a matter of fact, it may be and usually is organized in another kind of game.

  12. This seems to us the basic fallacy of the cybernetic analogies that seem to gain ground in social science and in systems theory; it is based on the assumption that human systems can be compared to systems controlled and governed by self-regulating services.

  13. This mode of reasoning was first tried by March and Simon in a logical, nonsociological way. Crozier used it first in an article published in French in 1962 (De la bureaucratie comme système d'organisation. Archives Européennes de Sociologie, 1, no. 2: 18-52). Crozier developed it substantially as a sociological concept in The Bureaucratic Phenomenon. Many Anglo-Saxon sociologists have since used it: D. Hickson and Ch. Perrow in the late 60s; more recently P. Abell (1975) and G. Benveniste (1977).

  14. Ultimately, all constraints, even technical ones, are amenable to be considered for the change and can also long run as a human construct.

  15. In this part we will a join a growing and more promising trend in American sociology: networks, interorganizational systems. See especially J. K. Benson, "The Interorganizational Network as a Political Economy," Administrative Science Quarterly, June 1975; William Evan, Organizational Theory (1976); and Herman Turk, Organizations in Modern Life: Cities and other Large Networks (San Francisco: Jossey-Bass, 1977).

  16. American organizational sociologists never seemed to understand this point. American sociology in any case did not face up to the problem of power, except for the recent Marxian treatment, which was more emotional than reasoned. Europeans have been more open to this new paradigm. See for example David Hickson et al., "Grounds for Comparative Organization Theory: Quicksand or Hard-Core?"' in C. J. Lammers and David Hickson, eds., Organizations Alike and Unlike: Towards a Comparative Sociology of Organizations (London: Routledge and Kegan Paul, 1979); Mauk Moulder, The Daily Power Game (The Hague: Martinus Nijhoff, 1977); and Peter Abell, ed., Organizations as Bargaining and Influence Systems (New York: Halsted Press, 1975).

  17. This too sketchy presentation could give the impression that we pretend this is a black-and-white or zero-sum game proposition, which it is not. The capacity of people to play more sophisticated games of conflict and cooperation must be added as a necessary third dimension. Developing this capacity is the only alternative to the increase of "artificial" bureaucratic power. We deal with those problems specifically in part 1 of the book.

  18. Here again we recognize the influence of Herbert Simon and James G. March. whose concept of bounded rationality helped us immeasurably. We discuss the problems of rationality in part 4, "'Decision Making.'

  19. Their common characteristic being the absence of any thinking about power and the problem of power as an unavoidable ingredient to any human interaction.

  20. This has been discussed at length in the European section of a Trilateral Commission report by Michael Crozier, Samuel Huntington, and Joji Watanuki, published in English as The Crisis of Democracy: Report on the Governability of Democracies to the Trilateral Commission (New York: New York University Press, 1975).

  21. Chris Argyris and Donald A. Schon, Organizational Learning: A Theory of Action Perspective (Reading, Mass.: Addison-Wesley, 1978).

  22. See part 5, **Change."

  23. Among many others, this has been shown very forcefully by the work of Ronald Laing. This has at last been recognized by Michel Foucault, Discipline and Punish: The Birth of the Prison, trans. Alan Sheridan (New York: Pantheon, 1978).

  24. As for example Stuart Clegg, Power, Rule, and Domination (London: Rutledge and Kegan Paul, 1975).

  25. This point is developed in the Appendix.

  26. European sociology seems much more lively by comparison. Works of young En glishmen like Peter Abell, of Swedes like Eric Rhenman and Richard Normann, of Germans like Niklas Luhmann, not to mention one's own countrymen, testify to Europe's presently greater creativity.

  27. Karlene Roberts, Charles Hulin, and Denise Rousseau, Developing an Interdisciplinary Science of Organizations (San Francisco: Jossey Bass, 1978).

  28. ibid., p. 136.

  29. This is what Roberts, Hulin, and Rousseau call a Dionysian strategy as against an Apollonian one.


The Theory and Practice of the Research Method (pp. 259 - 272)

Strategic analysis and systems analysis are not merely abstract theories. They are above all modes of research. This has been the general thrust of this book, and in treating a good many examples we think we have given a concrete demonstration of the character of these types of analysis. It may be useful to provide a more complete description of the research procedure as it is actually practiced. We propose to do so in this appendix. We will also give a brief theoretical justification of the principal assumptions on which it is based.

The Investigator and His Area of Research

 Inductive Logic

It will be useful to return for a moment to the point of departure for strategic analysis. This form of analysis rejects all determinism of context, environment, "objective structures" of the problems, etc; as we have shown, it does so by relativizing such determinisms. In their stead, it raises a question which belongs to another analytic level: what constraints are placed on the capacity of an organization and each of its members to act, develop, and change by the conditions, modalities, and constructs of the games through which they have managed to achieve cooperation?(1) This question has no general answer. The constraints it envisages are always linked to concrete and specific problems and solutions, situations, and actors.

To understand and explain the origin and nature of these constraints, we have proposed the strategic method, which cannot study a field of action in the abstract, on the basis of some a priori rationality. Instead, we must adopt a restricted phenomenological viewpoint in order to reconstruct the intrinsic logic and rationality of the relations and interactions which operate within the group; this is the only procedure which makes it possible to discover the weight of the constraints on the actors and on the organization they comprise. In each given instance, the contingent nature and rules of the games which structure the relations among actors, and consequently condition their strategies, are sought by the investigator. Once this has been done, it then becomes possible to go back to the modes of regulation which assure the integrated interdependence of these games within a system of action.

In approaching his area of study with such a project in mind, the investigator does not, of course, begin at absolute zero. Apart from his problematic, of which we have just given a summary sketch, and his mode of argument, which arms him with an analytical instrument, built around such concepts as strategy, capacity, zone of uncertainty, the game and its rules, and suited to the treatment of this problematic, he is in possession of certain preliminaries from which he constructs in a more or less formal way an initial framework for his study.

For one thing, he has "experience." He will himself have previously studied this field or others that are more or less similar. He will be familiar with research done by others in related areas. Consequently, his attention will naturally be drawn to certain situations which "experience," in some sense, will have equipped him to presume will be critical zones or areas of conflict. By giving priority to the analysis of these situations, it will be possible to make more rapid progress in understanding the special features of the field. He may know of certain particular structural features and "objective" constraints characteristic of his field of study; in any event, in his first concern will naturally be to obtain such knowledge if he does not already have it.(2)

Once this initial orientation is achieved, however, the job has just begun. The investigator may be aware that every system of action is constituted through power relations among actors seeking to control zones of uncertainty relevant to the problems to be solved. But this general formulation reveals neither the particular and contingent dynamic of such confrontations, nor the specific configuration of the actors concerned, nor the extent of their strategic fields, nor the limits of the system whose existence must be demonstrated and whose workings must be analyzed. More concretely, if he knows, for example, that every formal rule constitutes an artificial source of uncertainty which may be used by the actors in their strategies, this will suggest leads to follow up. Only after investigating them, however, will he be able to specify the relevance of such and such a category of actors, or measure their general importance.

The problem is the same with knowledge of the particular structural features of the field considered, and with the "objective" constraints that characterize it. Though analysis of such factors is indispensable, it can do no more than indicate a series of limits circumscribing the participants' strategic fields and channeling the possibilities of action -- excluding some, creating others. In themselves they are insufficient to enable the investigator to answer the only interesting question: namely, which of the potential alternatives is actually chosen by the participants? By the same token, the investigator will not yet be in a position to understand how and why one strategy is pursued rather than another, or what the significance of each strategy is.

The answers to these questions can be found only by clinical (and necessarily contingent) analysis of the effective relations which grow up among actors in the specific field under study. At this point, the investigator may proceed to study the games and modes of regulation characteristic of the particular system of action.

Strategic analysis is designed to study phenomena which are irreducibly contingent. Once this is acknowledged, it is clear that there is no choice but to adopt a hypothetico-inductive method in order to define, in successive stages, an object of study, through observation, comparison, and interpretation of the manifold processes of interaction and exchange which compromise the life of the system under analysis. The method uses the actual experience of the participants in order to formulate and verify increasingly general hypotheses concerning the characteristics of the ensemble.

The Path of the Research toward Its Goal

To illustrate how the investigator's thought progresses toward its goal, it is perhaps best to take an example from the field of organizations. We have deliberately chosen a crude one.

An investigator walks into an office where lower-echelon employees are performing extremely monotonous and routine tasks. Contrary to his expectation, however, all, or almost all. find their work interesting, maintain that they are able to take a fair amount of initiative, and seem to be quite satisfied. The investigator's common sense, outside experience, tastes, and knowledge of the formal prerogatives of these employees would lead him to say to himself that they were dulled by their work to the point where they were no longer even capable of recognizing its real nature, and consequently he might be tempted to discount their testimony. As a sociologist, however, he must say: "These people find work interesting which apparently is not. If they say it is, there must be a reason. What is it?" Following this lead, he has to begin searching for possible explanations. This raises problems of comprehension, for he is outside the game. He will first compare this testimony with that of other groups in similar situations, or with what the same group has to say about other problems. Next, he will use these comparisons to formulate hypotheses concerning the strategy on which the statement of the employees is based, and the game in relation to which such a strategy might appear rational. In attaching a significance to their expressed "interest" in their work, he will be led to discover certain key elements of the situation, e.g., the nature and characteristics of the game within which members of this group develop their strategies. It will be useful to review and comment rapidly on the principal steps in this process.

In order to be in a position to observe, compare, analyze, and interpret observed behavior, the investigator first has to be skeptical of the appearance of this behavior as perceived and described by the actors themselves. He must obtain a certain critical perspective on sense experience and common-sense categories (the actors' as well as his own). The observed facts must be stripped of their "evident" character, which is the way in which the participants ordinarily regard them.

This is a well-known epistemological requirement. Bachelard,(3) among others, has devoted some luminous pages to the question; it is no doubt useless to treat it again here. We have neither the desire nor the capability to engage in the traditional endless and intractable discussion as to the status and role of the social scientist in society; it should be noted that this often revolves around a requirement which it is clear is never completely fulfilled and which, moreover, does not depend solely on an "intellectual attitude" of the sociologist. Indeed, the whole problem of the investigator's autonomy -- his institutional and financial as well as his personal and intellectual, if not cultural, independence -- is involved.

Initially, the investigator is in the position of an outsider generally skeptical about the nature of the reality he observes. But this position offers no criteria for judgment, no normative base, which would allow him to "evaluate" what he sees. He must avoid "ethnocentrism." There is no one best way, and, a fortiori, no rationality exterior to the field that he can take for granted as evident. His problem is not to evaluate or even criticize the observed practices, but to understand them. The basic heuristic assumption which guides the whole approach is this: no matter how anomalous, contradictory, or senseless the observed phenomena may appear at first sight, they all make sense and correspond to some rationality from the moment they come into existence. It is precisely the search for this "deep sense," often hidden behind the apparent "sense" or "nonsense," which is the essential part of the investigator's work. To accomplish his task, he must find the particular constraints in relation to which apparently irrational conduct is irrational no longer.

This means that he must abandon the position of detached observer standing outside the field of inquiry in order to take a "detour" through the "inside" of the actors' own situations. It is only by reconstructing from within the inherent logic of the situation as perceived and experienced by the actors themselves that he will reach a position where it becomes possible to discover the implicit factors in relation to which apparently aberrant conduct takes on sense and meaning. In a manner analogous to that which he saw in the above example, the "objective" structure of the field can be elucidated by identifying its key elements, not all of which will necessarily be evident at the outset. This is accomplished by explaining the "subjective" significance of the behavior of the actors.

Thus the research method involves two complementary but contrasting orien tations, and constant interchange between them. Initially, the investigator must gain some perspective to preserve his autonomy and his unjaundiced view. But he then has to move right to the heart of the matter, as it were, to "put himself in the place" of the various actors in order to reconstruct the logic of the diverse situa tions they face. Finally, having taken this "plunge," he will in one way or another have to recover his own exteriority and compare the many contingent rationalities or strategies he may have observed; then, little by little, he can gain access to the implicit characteristics and rules of the game which structure the field in which he is interested.

What really interests the investigator are the relations of power among the actors and the implicit rules that govern their interactions. In order to get at this underlying reality, he must be able to separate from the mass of available obser vational data the information useful to his argument.

The criteria that govern this sifting of data are elaborated by the investigator himself, who uses what might be called the method of anomalies. Its principle is simple. In a more or less formalized manner (depending on the phase of the re search), the investigator will use the available descriptive data relative to his field in order to formulate a series of hypotheses as to what ought to be observed if everything went "normally," i.e., in a way consistent with the logic and "ration ality" used in elaborating the hypotheses. By then comparing these predictions with what actually occurs in practice, he will discover a whole series of "anomalies" or processes and conducts which do not seem to obey the rational "norms" embodied in his hypotheses.

These anomalies are invaluable. They indicate the areas where the hypotheses are in contradiction with the actual behavior of the actors. It is thus possible to pinpoint those places where a rationality different from the one known to the investigator is guiding the relations among individuals and groups. These anomalies thus indicate parts of the argument which need to be developed, and this in turn leads to the discovery of further characteristics of the organization.

The Importance of the Actors' "Experience"

It is apparent from the foregoing that the logic of the strategic analytic method accords a primary importance to the experience of the participantsb. This means that interviews must be regarded as a prime source of information.

The usefulness of the interview depends on its ability to provide the investigator with the insight into the actor's inner experience which he is seeking. Interviews are a technique for rapidly acquiring a significant amount of concrete information regarding the actors' experience and the implicit factors associated with the field under study. Using this information, the investigator may ask how each actor con fronts his situation and its inherent constraints, what objectives he sets for himself, and how he perceives his potential for attaining these objectives within the given structure. In other words, he seeks to know what resources the actor pos sesses, what his margin of liberty is, and in what way, under what conditions, and within what limits he can make use of them.

This means that the tension between the two poles identified above -- the exterior and the interior -- is reproduced in each individual interview. This is characteristic of the method. The investigator who wants to explore largely un familiar terrain must display a particularly open attitude in order to demonstrate to his subject that he is an interested and understanding partner whose role is limited to asking open questions which allow the other person to talk of his work, describe his situation as he sees it, etc. By definition, the interviewees are always "right," since it is they who experience the situation, not the investigator.

For all these reasons, the investigator's attitude toward his subject must be respectful and open; but obviously this does not man that it should be passive, as a completely nondirective approach would require.(4) The interviews have a precise goal: they must bring out the opportunities and capacities to act that are charac teristic of the various actors in the specific field under study. The investigator's problem in other words, is to lead his subject to reveal the sources of his action and to explain the characteristics of the strategic field in which he must act as he sees them. Each interview, then, is a strategic situation, in which the researcher's problem is to use the "spontaneous" dynamic of the session in order to achieve his own objectives.

It will be said that the testimony gathered in this way by the sociologist will not in general reflect "objective" reality, but rather the way the actor perceives and experiences it from his point of view: it is inevitably "subjective." This is not a valid objection to strategic analysis. The essential point of this method is precisely to surpass the artificial dichotomy between "objective" and "subjective" reality. The strategic analyst regards the subjectivity of others, their manner of choosing strategies according to their perception of the constraints binding them, as an important element quite as "objective" as the technical or economic con straints-in defining the situation in relation to which each individual elaborates his particular strategy and conduct.(5)

It is perhaps worth noting that from our viewpoint this recourse to the "experience" of the actors is much more than a kind of symbolic tip of the hat to "the importance of the human factor." By the same token it is much more than a perfunctory complement to knowledge of the formal characteristics of the field. It is the primary condition for a genuine understanding of the field in question. "Objective" knowledge can be constructed only through analysis of "subjective" experience, and this is the only procedure that allows the investigator to discover and set out in detail the effective significance, importance, and range of some supposedly objective constraint which was not at all obvious at the outset.

The problem, therefore, lies not in the subjective character of the testimony, but rather in the use to which strategic analysis puts the data thus gathered. In effect, we are treating the information obtained from interviews as a sort of primary indicator, ultimately the only one, of the subjective choices made by the actors among the opportunities offered by their respective situations. The actor's tes timony is taken as an expression of his strategy. It is not self-evident why this should be justified. The point is worth considering at greater length. The Problem of Materials Implicit in what appears at first as a choice of method is, as we have indicated, a different conceptualization of attitudes, a conceptualization which touches on both the links between attitudes and behavior and the processes by which these links are formed. From the standpoint of strategic analysis, indi viduals develop attitudes not as a function of the past (their socialization, their past experience), but as a function of the future, of the present and future oppor tunities that they see in the games they are playing and in relation to which they orient their strategies. For the sake of simplicity, it might be said that here atti tudes are explained not so much by the individual's past insofar as it conditions present behavior, but rather by the present behavior itself. Through it, we come to understand the nature and rules of the games which direct it, and whose charac teristics are at the heart of the organizational phenomenon. This is what the investigator tries to grasp in examining individual attitudes.

Before we turn to setting out in detail the premises and consequences associated with such a reversal of perspective, and to giving some illustrations, we must describe very succinctly how social psychology has used attitudes, what evolution the notion has undergone, and what theoretical premises are implicit in this evolution.(6)

Attitudes in Social Psychology

When it made its appearance in the vocabulary of German experimental psychology toward the end of the last century, the notion of attitude was assimilated to a neuropsychic state which preceded and facilitated an action. In this first restrictive usage, it designated primarily a mental process, a sort of a "mental adjustment,"(7) a provoked and therefore conscious "attention" of individuals in an experimental situation which facilitated and accelerated their re sponse to a stimulus given them.(8)

The discussion of the concept in social psychology and its introduction into situations of a less experimental nature greatly extended and broadened its range of application and essentially modified its status. Through various stages about which there is no need to go into detail here, the term attitudes came to designate relatively permanent normative orientations of individuals with respect to certain privileged social objects, selected according to the central interests and specific needs of the research: unemployment, women's suffrage, religion, political par ties, work, the company, etc.(9)

Discernible behind this extension of the concept's usage is a significant shift in the burden of research, on both methodological and substantive levels. On the methodological level, first, the technique of observation and measurement changed. Attitudes were no longer observed directly, but rather indirectly through a series of more or less complex measurements of the degree of satisfactions felt by individuals with regard to the various dimensions of such and such an object of the social reality under study.(10) As a result, attitude became the inferred basis, constructed after the fact, for the judgments and opinions collected from the subjects. It was the stable element, the psychological structure underlying the opinions and giving them sense and coherence. Opinion--which came to mean instantaneous evaluation of a social object depended in some way on a more stable evaluation of this same object, which was the attitude.

On the substantive level, this implied a movement toward an increasingly abstract object of research, from opinions to attitudes and finally to the value systems to which individuals were supposed to subscribe. Attitude was the bridge between observable individual conduct and the value structure which was sup posed to orient this conduct but which was unobservable. Values were sought through attitudes.

This explains why the study of attitudes came progressively to focus not so much on the attitudes themselves but rather on the processes by which they are formed and transformed.(12) From this standpoint, the problem for research was one of identifying social situations, structures, and roles in which individuals undergo significant experiences of social learning, and of defining the social ob jects concerning which they acquire attitudes. Consistencies in the constitution and acquisition of attitudes were sought through study of the mechanisms by which these attitudes were formed and transformed, and attitude itself was con ceptualized as reflecting the motivations which oriented and explained behavior.(13)

With a little perspective, it becomes clear that this use of the notion of attitude is based on certain theoretical premises which are not always very explicit. It will be useful at this point to bring them out clearly.

First, regardless of how the concept is extended, attitude is always considered an individual attribute. It refers to what is individual as opposed to what is "situa tional." In a given situation, several types of behavior are possible. Attitude makes it possible to understand the individual's choice of one of them. It thus tells us something about him.(14) It implicates the features of his personality and even his system of values.(15)

It is of course true that the concept of the individual varies and can be more or less complex. Initially, the individual was defined wholly in terms of the attitudes he adopted through meaningful experiences of social learning. Later, an attempt was made to incorporate the individual's analytical capacities into the argument; but an irrational component remained, that was irreducible to an instrumental calculation: this was the "attitude," which was based on the individual's prior experience.

Regardless of which point of view as adopted, attitude was defined with refer ence to the individual, his personality, his system of values, and its formation. Here we encounter a second more or less implicit theoretical assumption: when attitudes are used to explain behavior, they come to be based, intentionally or not, on a theory of social learning in the matrix of society. In other words, one im plicitly refers to a theory that incorporates the following chain of deductions: first comes society with a certain social structure; then the personality is formed within this structured universe through meaningful social experiences which lead to the adoption of certain values and personality traits; these in turn induce certain types of opinions and behavior.(16)

Explanations for the acquisition of values through meaningful social experi ences vary. For the behaviorists, as is well known, this is a process of in strumental learning in which attitudes form and are reinforced, little by little, in response to punishments and rewards coming from the environment. Other ap proaches, such as the theory of cognitive dissonance, deduce learning and trans formation of attitudes from a principle of internal coherence.(17) Still others use methods inspired by psychoanalysis to look for the origin of attitudes in the needs, anxieties, inhibitions, motivations, and defense mechanisms of individuals, i.e., in the functions the attitudes fulfill in relation to the subject's experience and subconscious.(18)

This explains the retrospective character of most analyses. From such a vintage as we have been describing, attitudes can only be understood in relation to the subject's past experience, of which they are in some sense the reflection. This also explains the interest in and justification for studying attitudes in themselves. Once acquired and established, they are relatively stable, hence measurable. They have a life and logic of their own, which is precisely what needs to be brought to light in order to understand the subjects' reactions. This, finally, accounts for the predictive use which has often been made of attitudes. Since they may be regarded as relatively stable, if we know what they are we can predict the reactions and behavior of subjects in different circumstances, and even direct them by controll ing their learning experiences.

An Example

Attitudes have been studied in industrial sociology and in the sociology of work. Very simply summarized, the personality of the individual on the job is analyzed in these disciplines as a set of attitudes which govern his responses to organizational requirements. In order to understand these responses, and eventually to change them so as to make improvements in the operation of the organization, the attitudes in question must be measured and explained.

Though strategic analysis belongs to this current of thought, it has gradually evolved another use of the concept. We would like to illustrate this with the following example, taken from a study made between 1956 and 1958 in six Parisian insurance companies.(19) A questionnaire was used to obtain information concern ing individual perceptions, feelings, and opinions relative to various aspects of work experience. Consistent with the "classical" outlook of the sociology of work, one of the aims of this study was to measure and understand how individual attitudes toward work were structured. The study produced disconcerting results which led to a change in our point of view. We would like to call attention to these results before we go on to discuss their implications in greater detail.

Two major dimensions were incorporated: first, the manner in which individuals evaluated the interest of their work, and, second, the satisfaction they felt with regard to their situation. At the top of the hierarchy, the policy writers exhibited interest in their work with the greatest frequency, but at the same time they were the most likely to declare themselves dissatisfied with their situations. Conversely, the individuals at the bottom of the hierarchical ladder were most likely both to find their work uninteresting and to say they were satisfied with their situations.

This pattern in the results was disconcerting: it was difficult to accept the fact that satisfaction with the situation is more pronounced at the bottom than at the top of the hierarchy, especially since these results contradicted the "objective" analysis, which found very low salaries but relatively tranquil working conditions at the bottom of the scale and higher salaries but poorer working conditions at the top.

To move our interpretation forward, it was necessary to change our view of attitudes. No longer were these seen as mere reactions to a passively evaluated reality, but rather as signs of the strategic perception of each actor. At the lower level, the employees felt too threatened to complain about their situation. But there were fewer risks in expressing dissatisfaction with their work. This even enabled them to move the implicit negotiation with the company onto ground where they felt more secure; they were able to assert their superiority to their station and thus demand better working conditions as a reward for their loyalty. On the other hand, a high-level position in the hierarchy conferred certain advan tages which reduced the risks involved in criticizing one's situation. At this level, the function was accompanied by a certain undeniable prestige which was an advantage in the implicit negotiation, insofar as the importance of the work could be used to justify an improvement in the situation.

To sum up, the policy writer complained about the conditions to which he was subjected by his company in the name of the importance and dignity of his job, while the secretary tended to use her loyalty to the company as justification for asking for more interesting work.

Attitude in the Strategic Procedure: Premises and Consequences

Although the foregoing example is limited and incomplete,(20) it is a good illustration of the change of perspective in the treatment of attitudes. For strategic analysis, these are no longer reflections of a reaction or a summation of past experience, but rather correspond to the strategic orientations adopted by the social actors, taking into account their possibilities and resources, as well as the constraints that bind them. They therefore reflect the actors' choice of an orienta tion for their action in the face of the risks and opportunities offered by the games in which they participate. In simple terms, for strategic analysis the attitudes of the actors do not depend on the past but rather on the future, as the actors, with their present resources and expectations, see it; more precisely, attitudes depend on opportunities discovered in the games in play, to which behavior must be adjusted. In short, it is opportunity that makes the thief, not his past history.(21)

Thus we may say that the analysis of attitudes in this perspective is prospective rather than retrospective in character. Certain opportunities emerge from the actors' situations and games. Depending on their resources and capacities, some of these will be taken advantage of, and corresponding strategies will be adopted. The actors will adjust their behavior to these choices. According to a schema which extends and develops the approach to social learning in terms of a cognitive dissonance that requires an effort to restore coherence,(22) these choices will in turn represent a strong pressure on the orientation of the actors' attitudes. Attitudes thus express a kind of choice:(23) they are indicative of the effective use which actors can and wish to make of the available resources in view of the opportunities offered by the game.

Thus the retrospective aspect of attitude is supplemented by a prospective dimension. The attitudes of the actors express a strategic orientation that comes out of an evaluation of their potential for action, a sort of anticipatory computation of wins and losses, and in which the present situation is analyzed in terms of advantages and resources which may be mobilized in the future. To return to our previous example, it is clear that neither the policy writers' nor the secretaries' attitudes can be explained by a simple accounting of profits and losses. Another kind of calculation is involved. Both policy writers and secretaries seek to de termine the strong and weak points of their situation, together with potential means for taking offensive or defensive action.

Of course, the past experience of the actors, their values and attitudes in the social-psychological sense, do not disappear in such an analysis: they largely condition the way the actors will perceive their game opportunities and structure their capacities so as to take advantage of these opportunities. As we showed above, 24 an individual will adopt a particular strategy not only because he sees an opportunity, but also because he has the necessary capacities-material, affective, cognitive, relationalfor accepting its risks and difficulties. Similarly, if a group adopts a certain game, it does so not only because of its goals and resources, but also because of the constraints which limits its margin for maneuver and range of options. These constraints come from the mechanisms of integration (including values) and constructs which the group has no choice but to rely on for its very existence as coherent entity. The past experience of individuals, their * socializa tion." and the resulting values therefore do not disappear from the analysis; they are merely relativized. Here, these factors become nothing more than the ele ments which structure the capacities of individuals and groups, thereby indirectly conditioning individual strategies and collective games.

For the strategic method, attitudes indicate not the characteristic or "personalities" of individuals as determinants of their behavior, but rather the subjective relationship that individuals establish between their situation and the game in the light of their resources and capacities, or, in short, the strategies they have adopted or will adopt. It is possible, therefore, to take attitudes as signs which reveal these strategies and consequently the characteristics and rules of the games that define the range of possible strategies and keep the system of action operating.

Attitudes thus provide access to "inside" knowledge of the way in which mem bers of a system of action intend to use the margin of liberty, hence of power, available to them within a given framework of games. Similarly, they indicate what resources and opportunities are distributed by the system among its members. Attitudes will apply to individuals, then, but in this view what is of explanatory value is their situation within a game to be discovered. In other words, at the risk of giving a somewhat caricatural description, it would be possible to characterize the reversal of perspective sketched above in the following way. In social psy chology, the explanatory value of attitudes lies in the fact that they are supposed to reveal permanent individual dispositions to act, and hence intrinsic individual values as these define and differentiate a given personality. In strategic analysis, on the other hand, the explanatory value of attitudes lies in the state of a system of action which they reveal. The characteristics and mode of regulation associated with this system state structure the games in which the members must play, which is the basis of the explanation of their behavior.

It follows that strategic analysis is not interested in observing attitudes outside of a specific situation and its associated structure. Attitudes have no other structure. The attitude no longer has any value in itself; it ceases to be an "in itself" (en-soi). There is no longer any point to collecting attitudes for the pur poses of abstract measurement; they are only a tool for inferring the strategic orientations of the actors.

If it were possible to have direct access to knowledge of games, which could be described as ranges of options aimed at certain individual capacities; and if, moreover, it were possible to identify and measure these capacities; it would then be possible to dispense with the notion of attitude. That such a procedure iS difficult if not impossible underscores the interest in an analysis of attitudes. We may think of them as tools permitting rapid access to the essential point: the various choices effectively made by the members of a system of action from a range of options.

Attitudes, then, are convenient, yet imperfect, research devices or tools which strategic analysis uses to observe the subjective aspect of a game which remains to be discovered. They are useful in a heuristic procedure whose purpose is to facilitate the observation and comprehension of the way the members of a system of action put together their situational potential in order to exploit the opportuni ties offered by the game. This is the primary goal of the research. The important point is to understand the interdependence of the subjective combinations of the potential in various situations. To put it another way, one wants to understand the interdependence of the various strategies in a system. This will make it pos sible to understand the structure and regulation of the system of action under study

Interpretation: From Feelings to Games,

From Games to Structures

The investigator's first order of business is to familiarize himself with the terrain and to make a quick inspection tour to determine its objective characteristics.25 He will then have to devote fairly extensive amount of time to gathering information by interviewing actors belonging to the various categories which he identifies with the aid of his formal, and inevitably rough, knowledge of the field.

From these interviews, the investigator will collect a variety of data providing quite concrete information as to the perceptions and feelings of individuals and groups regarding their respective situations. In particular, his data will include: detailed information regarding perception and behavior, constraints and ensuing difficulties; information on relationships among the actors, the importance placed on them, associated expectations, nature of frequent conflicts and of solutions generally attempted; information concerning the actors' evaluations of their ac tivities, situations, relations, areas of satisfaction and/or dissatisfaction, hopes and disappointments; finally, the actors' estimates of their own and others' action potential in view of the foregoing.

Whether these data are qualitative in nature, deriving from relatively un structured interviews, or the result of a statistical treatment of quantified answers to a formal questionnaire, in the strategic method the problems of interpretation are always the same. The descriptive data regarding opinions, perceptions, senti ments, and attitudes must be used to reconstruct the power structure, together with the games and their rules, which regulates the interaction and conduct of the actors.

To this end, the investigator's first priority will be to establish what different strategies are involved. This is the object of the intensive analysis that he will make of the explicit and implicit26 contents of the interviews or statistical studies.27 Inevitably, there will be at least indirect and implicit indications of the power element involved in each actor's relations with the others, and of their respective evaluations of the chance for profit and loss. The data collected will be studied as evidence of the actors' perceptions, feelings, and attitudes, their com ments on certains problems, their explanations and rationalizations for certain behavior, and their satisfactions and dissatisfactions. This is true even though these data always reflect certain "official"* aspects of their situation. Neverthe less, the investigator gains access to the strategic perceptions each actor makes of his situation, and to the adjustment he makes to a more or less conscious system or relations with the properties revealed by the research.

By a continual effort to reconcile the convergent and/or discordant data derived from the interviews, the investigator will attempt to discover the internal logic that implicitly structures the set of perceptions, feelings, and attitudes of each category of actors. The goal is to identify and detail the various strategies involved.28

If quantified data are available and the field of investigation is suitable,29 the investigator may use a supplementary technique to "test" directly the validity of the strategies he thinks he has found. He can communicate the raw results of the investigation to the parties involved, without any commentary, but after mak ing certain hypotheses concerning their probable reactions. For instance, when the results were presented to the workers in the Industrial Monopoly, the re actions of the maintenance workers were extremely revealing as to their particular strategic situation. The members of this group denied and challenged all the re sults which did not correspond to the majority strategy of aggressive domination. There was one noteworthy exception: the union officials from this group made comments which clearly were worded so as to minimize the sharpness of the prior attack. This is easily comprehensible: it was important that the power of the maintenance workers and the accompanying conflicts remain hidden. Too much publicity would have threatened the status quo, which gave these individuals the upper hand in the union as well as the shop. so they tried to confuse the issue.

Once the strategies have been identified, it remains to explain them. Our fun damental hypothesis, according to which the strategies are rational, has been amply justified by its heuristic usefulness.30 If a strategy appears to be irrational, a closer look must be taken at the situation(s) which enter into the definition of rationalities. Technical, juridical, social, and economic constraints have to be considered, along with the *rules of the game" which reflect the relations of power which structure the organization's system of action. Comparison of the strategies identified with the "objective" constraints will show that the latter are only partially adequate to explain the former. As long as the power structure has not been brought into the picture, a zone of "rationally" inexplicable strategic behavior remains. The nature and "rules of the games" which provide regulation of the underlying system of action must also be taken into account.

In other words, to explain the strategies pursued by the various actors, the investigator will seek to make explicit the relationship, or, better, the implicit mediation which exists between a collection of constraints inherent in the field being studied (technical and economic constraints, official regulations, etc.), which define the critical zones of uncertainty; and a set of strategies which can be reconstructed from the responses collected during the investigation, reflecting the strategic perception of their respective possibilities by the various actors, in the light of the over-all constraints.

To reconstruct this mediation, the investigator will formulate increasingly gen eral hypotheses concerning the characteristics and rules of the implicit games within which several strategies can be rational simultaneously. He will begin with a simple hypothesis regarding the implicit game between two groups. In order to verify this hypothesis, he will first ask what consequences the characteristics and rules of this first game should have on the strategies of a third group; then he will check to see if the responses given by this group are compatible with his analysis. Depending on the outcome, he may have to reformulate the initial hypothesis, or broaden it, before submitting it to another test against attitudes and feelings, and so on. Using such an iterative procedure-going from feelings to strategies, strat egies to games, and then back to feelings, etc.-the investigator will be able to discover, from the experience of the various actors (individuals or groups), the power structure in the field under study and the games that condition the behavior of the actors.

To make this somewhat abstract schema concrete, it may be useful to apply it to a simplified concrete example, taken from the Industrial Monopoly. When his investigation is complete, the researcher will have amassed information con cerning the attitudes of production and maintenance workers. From this the undercurrent of conflict between these two groups will emerge clearly, along with the caustic comments each group is bound to make about the other. To explain this undercurrent, the first step is to try to reconstruct the strategies of the two groups. If a crude way, the strategy of the maintenance workers might be charac terized as one of attempting to assert a peremptory, if not aggressive, superiority over the production workers. The production workers' attitudes, on the other hand, indicate a strategy of official submission coupled with what amounts to indirect resistance. The next step is to ask what type of game such strategies might correspond to. An obvious hypothesis would be that the two groups confront one another, with highly unequal resources, over a stake of great importance: the extent and control of the arbitrary margin of power possessed by the maintenance workers because of their dominance over a crucial source of uncertainty: the breakdown of machines. Thanks to this dominance, they are the real "bosses" of the shop. If this hypothesis is correct, it will affect relations between production workers and shop foremen. When one examines the attitudes of these two groups toward each other, one finds, contrary to what might be expected in relations between superiors and subordinates, that there is a virtual absence of emotional content. In a similar way, if one analyzes the attitudes of the maintenance workers and shop foremen toward each other, one can support and complete the initial hypothesis, and so on. We have done no more than note the high points of the procedure the investigator will follow. It will lead him to discover the particular importance of the breakdown of machines, which is far from obvious a priori, but specific to the system of action within the Industrial Monopoly. He will gain an understanding of the maintenance workers' consistent strategy of defending their expert role by hiding the machine plans and maintenance bulletins, and he will grasp the significance of these strategies for the underlying system of action.

The key to the method of strategic analysis is thus to use the data gathered in interviews in order to define the strategies pursued by the actors and ultimately to have access to the games which correspond to these strategies. These games indicate the structure of the power relations which link the various actors. It is this structure which must be clarified if the over-all regulation of the system of action is to be understood.

Notes to the Appendix

  1. In the broadest and, therefore, most neutral sense of the term, which in no ways rules out (we remind the reader, lest there be any misunderstanding) situations of domination. Of course, the games that prevail among actors will depend on whether cooperation is obtained by force or voluntarily, whether it results from domination by one set of actors rather than another, or, alternatively, from a relatively egalitarian situation. In any case, it is legitimate to speak of cooperation.

  2. He has several ways of doing so: organization chart, internal regulations, technology, etc., in the case of an organization; the, in some sense, "objective" logical structure of the problem, the formal prerogatives, and the material resources of the principal actors involved in the case of a more diffuse system of action.

  3. See especially Gaston Bachelard, La formation de l'esprit scientifique (Paris: Vrin, 1970).

  4. Without entering into a discussion of the most general problems raised by nondirective techniques (conditions of their possibility, potential abuses, etc.), we will say that insofar as the procedure attacks a relatively formalized problem, a totally nondirective attitude would be out of place, if not basically aberrant.

  5. More generally, shouldn't all the elements of "social reality" -- as "objective" and "predetermined" as they may appear at first sight -- be regarded as social constructs main tained by social action, i.e., the product of conflicts and power relations between individuals or groups -- in short, between social actors. It is the sociologist's research strategy, i.e., his choice of terrain and his level of analysis, which reintroduces the dividing line between that part of social "reality" that he may consider as a given precondition and that which he must explore and explain as a social construct.

  6. Such a review must necessarily be brief. For a presentation and discussion of all these questions, we refer the interested reader to G. Lindzey, ed., The Handbook of Social Psychology (Reading, Mass.: Addison-Wesley, 1954); S. E. Asch, Social Psychology (En glewood Cliffs, N.J.: Prentice-Hall, 1952); A. Levy, Psychologie sociale (Paris: Dunod, 1965); G. Summers, Attitude Measurement (Chicago: Rand McNally, 1970); M. Jahoda and N. Warren, eds., Attitudes (Harmondsworth: Penguin, 1966); and S. Moscovici, "L'at titude: théories et recherches autour d'un concept et d'un phénomène," Bulletin du CERP 2 (1062): 177-191. In addition, we would like to thank W. Ackermann for the aid and advice he gave us in the writing of this section.

  7. The German word Einstellung means, first of all, adjustment in the sense that a machine must be "adjusted" for certain operations.

  8. This is, in particular, the sense given the word by N. Lange in relating his experiments on the different reaction times of individuals, depending on whether or not they were consciously prepared to make a desired gesture after having received the agreed signal. See especially G. Allport, "'The Historical Background of Modern Social Psychology," in Lindzey, Handbook of Social Psychology.

  9. This usage of attitudes is well known in the sociology of elections, psychosociology, industrial sociology, and even commercial market research, etc.

  10. Such as satisfaction with regard to various aspects of work, the situation, hierarchical relations, etc., which are dealt with in the studies of industrial psychosociology or the sociology of work.

  11. Two utilizations of the study of attitudes must therefore be distinguished. The first focuses on attitudes as way of predicting behavior, by establishing an implicit equivalence between attitude and behavior (such and such an attitude, such and such behavior). The second focuses on what attitudes reveal of regularities and modes of structuration within individual categorizations of the social world, i.e., the universes of professional, occupa tional, or national values. We have previously emphasized (see chap. 6) the deterministic reasoning that usually underlies such a utilization of attitudes.

  12. This interest in the processes of learning and transforming attitudes reflects the accent placed on the "socialization" processes by functionalist sociology.

  13. This is the perspective adopted in countless sociological studies of attitudes. In pas sing, we point out the implicit normative argument behind this research effort. Its latest avatars are met in the "motivational" approach to organizations, to which we alluded previously. There we can see the temptation to use knowledge of attitudes to give direction to training processes by manipulating their experiential and situational content. In short, the learning of "correct attitudes" can be insured by taking measures to control the places of learning.

  14. The claim, of course, is that a hold is also obtained on collective attitudes. But actually these are always derived from aggregate individual attitudes. What may be collective is the conditions under which these attitudes are acquired, as these may be relatively similar from one individual to the next. But the attitudes themselves remain individual.

  15. In this respect, the conclusions of a psychometric study by the Swedish psychologist U. Remitz are quite explicit. According to him, only ten percent of the variation in attitudes toward work can be attributed to external social factors, while factors internal to the firm (in this case a bank) could account for at most an additional ten percent. The remaining four fifths of the variation reflect, in his view, the existence of a sui generis factor, the individual disposition to satisfaction, which might be isolable and measurable in the same way as intelligence. See U. Remitz, Professional Satisfaction Among Swedish Bank Employees (Copenhagen, 1960).

  16. This makes more readily comprehensible the accent placed on the formation and transformation of attitudes in the above-mentioned research, as well as the convergence already noted between this research orientation and that of functionalist sociology.

  17. See Festinger, A Theory of Cognitive Dissonance.

  18. See, among others, Theodor W. Adorno et al., The Authoritarian Personality (New York: Harper, 1950); D. Katz and E. Stotland, "A Preliminary Statement of a Theory of Attitude Structure and Change," in S. Koch, ed., Psychology: A Study of Science, vol. 3 (New York: McGraw Hill, 1959).

  19. For further detail concerning this example, which we can summarize only briefly here, we refer the reader to M. Crozier, Le Monde des employés de bureau (Paris: Editions du Seuil, 1965), especially chap. 5, pp. 80-105.

  20. Other examples might be cited. For instance, the "antiurban attitude" which was prevalent among engineers in the State Public Works Ministry at the time of the reform of the Ministry of Equipment. As we have shown elsewhere, this "antiurban and technocratic attitude" can only be understood as the expression of the dominant strategy pursued by these middle level functionaries. Its full significance emerges when it is seen in the light of the power struggle which pitted them against their superiors, the engineers of the Ponts et chaussés, over control of the regional offices of the old Ministry of Public Works. See Thoenig and Friedberg, "Politique urbane et stratégies corporatives," Sociologie du travail 4 (1969).

  21. Except insofar as his past history, in part, conditions the "thief's" present and future opportunities, as well as the actor's capacity to act as thief.

  22. Festinger, A Theory of Cognitive Dissonance.

  23. Attitude reflection becomes in this perspective an extreme case which may be interpreted, in turn, as a sign of resignation or impotence.

  24. See chap. 6, which is devoted to links between mode of organization and culture.

  25. What we generally referred to as the "hard facts." This appellation seems to us to be an abuse of language which perpetuates a misconception that must be dispelled. In fact, these hard facts frequently turn out to be much softer and less resistant to change than the so-called soft "human" factors that are the object of strategic analysis. To consider only the area of organizations, for instance, to change an organization chart is a relatively simple matter; organizational change is often reduced to a mere redefinition of this kind. By contrast, to change the ongoing games within the organization is infinitely more difficult.

  26. Quite as important as what is said explicitly is what appears only between the lines or what is, perhaps, even passed over by certain interviewees. For to avoid mentioning a problem or conflict which is commented on at length by others may be as revealing a way of making one's opinion and point of view known as speaking volumes about it.

  27. Obviously, many kinds of more sophisticated procedures may be applied in this area, provided they are securely founded, i.e., that the relevance and significance of what is being measured is known. Indirectly, this points to the basic importance of the qualitative, exploratory investigation in such a methodology. Not merely a formal familiarization with the research terrain, this will always be one of the high points, if not the key step, in the whole research enterprise. It is actually an investigation in its own right, and its execution may be fairly time-consuming, but the results should make it possible to formulate a first, relatively precise diagnosis of the specific properties and regulations of the system of action under study.

  28. Obviously, with these strategies, the investigator forms an idea of the actors' fields and their real categories, which will usually be little in line with the official analysis or common sense. A typical example in this respect is the complicity between prefects and notables analyzed above, strikingly different prefects and notables analyzed above, strikingly different from the rhetorical antagonism prevailing between these two categories of actors.

  29. Which will essentially be the case in organizational studies.

  30. That is to say they have a "rationality" which must be uncovered. It should be noted that here the basic assumption of strategic analysis -- that individuals make rational and strategic use of their zone of liberty -- is divulged in its true colors: it is a heuristic procedure which makes it possible to structure and comprehend a social field and its various components, rational and irrational, conscious and unconscious, voluntary and involuntary.

Associated notes and readings



Pages that link to this page