We have stated above that contexts and constructs are primarily relations. From the strategic point of view we have adopted, these relations involve power. Since the organizational construct Is revealed only through the actor's experience, our method can begin only with the actor's strategic choices. It is the study of power which will enable us to analyze this construct, because power, as a fundamental stabilizing mechanism in human behavior, is at the root of the collection of relations which constitute it.(1)
Our approach to power will be to continue our study of strategy, beginning with a consideration of power from the actors' point of view.(2)
The phenomenon of power is simple and universal, but the concept of power is elusive and complex. Therefore, we will begin with simple formulation of what in some sense represents the common denominator in all manifestations of power: whatever its "type"-meaning, its sources, objectives, degree of legitimacy, or means of exercise, power, on the most general level, always implies that certain individuals or groups are able to act on other individuals or groups.(3)
As vague as it is, such a formulation has the advantages of not taking for granted a theory of the essence of power. of being applicable to any form of power, and, especially, of drawing attention to what we consider the essential point: the relational character of power. To act upon another, in fact, is to enter into a relation with him,(4) and it is within this relation that the power of person A over person B develops.
Thus power is a relation and not an attribute of the actors. It can appear, and consequently become constraining for one of the parties involved, only by being introduced into a relation(5) which confronts two or more actors dependent on one another(6) for the achievement of a common objective which modifies their personal objectives. More precisely, it can develop only through exchange among the actors involved in a given relation. To the extent that every relation between two parties presupposes exchange and reciprocal adaptation between them, power is indissolubly linked to negotiation: it is relation of exchange, therefore of negotiation, in which at least two persons are involved.
Several remarks deriving from this fact will help to make the nature of the relation more precise.
First, it is an instrumental relation. This is not to deny that its existence always entails a series of very powerful affective phenomena which profoundly condition its development.(7) Nor does it signify that all the consequences and effects of a relation of power are necessarily conscious or intentional. To take D. Wrong's example(8), it is surely not part of the intention of an overprotective and domineering mother to "feminize" her son's character. In this and other instances, purposeful action by individuals entails its portion of unforeseeable, unexpected, and "dysfunctional" consequences. Furthermore, we do not thus exclude from the analysis the forms of social domination and control which, when internalized by the various actors, give rise to the phenomena known as "deferential adjustment" or "anticipatory adjustment"(9) and which no longer require a conscious commitment of resources on the part of any of the actors.(10) To say that every relation of power is instrumental is merely to underscore the fact that, like every relation of negotiation, power is conceivable only from the standpoint of an aim which, in the context of an instrumental logic, motivates the actors' commitments of their resources.
Second, the relation of power is a nontransitive relation: while person A can readily get person B to perform action X, and B can elicit the same from C, it may nevertheless be true that A is incapable of obtaining X from C. Moreover, if power is thus inseparable from the actors engaged in a relation, it is also inseparable from the actions demanded. Each of these represents a specific stake around which a particular relation of power is erected. Thus A might be able to obtain X from B quite easily, Y might be a little more difficult to get accomplished, and Z, impossible, while, on the other hand, C might be able to obtain Z from B without difficulty.
Finally, power is a reciprocal, but unbalanced, relation. Its reciprocity derives from the fact that negotiation always involves exchange. If one of the two parties involved has no further resources to commit to the relationship, it no longer has anything to exchange; hence it cannot enter into what may properly be called a relation of power. In other words, if B can no longer bargain over his willingness to do what A is asking, there is no longer any question of a relation of power between them, because B has now ceased to be an autonomous actor with regard to A, since he has been reduced to the status of an object.(11) In addition to being reciprocal, however, the relation in question is unbalanced. If A and B have equal advantages in the relationship, so that the exchange between them is on equal terms, then there is no reason to consider that one of them has power over the other. But once the exchange becomes unbalanced in one or the other's favor, where this inequality reflects the respective situa tions of the two parties, then it is legitimate to speak of a relation of power.(12)
It then becomes permissible to redefine power as a relation of exchange, hence a reciprocal relation, but one in which the terms of ex change favor one of the parties involved. It is a relation of force from which one party can obtain more than the other, yet in which neither party is totally defenseless. To paraphrase R. Dahl's previously cited definition, it may be said that the power of A over B corresponds to A's capacity to obtain favorable terms of trade in his negotiations with B.(13)
What is the basis of power? What is its source? The answer to these questions seems obvious: the advantages, resources, and forces of each of the parties involved -- or, in short, their respective strengths -- determine the result of a relation of power. What is meant by "strength," however, needs to be made more explicit. What is actually exchanged in a relation of power? The answer is that it is not so much the forces or the strengths of the different parties involved as it is their possibilities for action. A does not enter into a relation of power with B solely in order to test his strength against B's. He has a more definite objective: to obtain from B a behavior on which his own capacity to act depends. In other words, B in some sense controls, by means of his own behavior, A's capacity to achieve his objectives. The more B can bargain over his willingness to do what A wants, and the more his available resources allow him to keep his future behavior from becoming perfectly predictable for A, then the more power he will have over A in this particular relation. Power therefore, lies in the margin of liberty available to each partner in a relation of power. In other words, the more one partner is free to refuse what the other asks of him, the more power he has. Force, wealth, prestige, authority -- all the resources, in short, which any one of the parties may claim -- play a role only to the extent that they provide a greater freedom of action.
At this point, it might be useful to digress a bit in order to confront Schelling's argument, apparently in contradiction with the above, which can be found in a remarkable book cited previously.(14) In rich, stimulating, and frequently amusing analyses, the author offers numerous examples of situations -- or rather of games and game structures(15) -- in which the "winner" is not the one who manages to maintain his margin for maneuver by keeping his future behavior unpredictable, but rather the one who does the opposite by tying his hands and reducing his maneuverability to zero, thus making his future behavior mpletely predictable. His is the power of the weak, of the player without alternatives, who, having "burned all his bridges," finds himself with his " back to the wall." Shelling gives the example of the striking railwaymen who succeed in halting a "strike breaking' train by chaining themselves to the rails.
In our opinion, Shelling somewhat overestimates the generality and universality of the situations he studies. His games have special structural conditions and dynamics.(16) A closer examination reveals that they are games in which each of the adversaries/partners has a more or less urgent but always imperative need to arrange and settle things with the other, and, moreover, in which this other cannot be chosen but has been part of the game from the outset. The fundamental logic of such games is that of the bilaterial monopoly, a structure which already limits drastically the margin for maneuver of each partner/adversary. In the author's example, chaining oneself to the rails makes sense only because there are rails, structures which drastically limit the other's freedom of action: the train's engineer can choose only between surrender and massacre.
Such structural conditions are found, though rarely, in reality.(17) It should be noted, however, that in the actual cases familiar to us, the fact of making one's future behavior perfectly predictable has significance and consequences which extend beyond the confines of the game itself. To do so always amounts to changing the nature of the game, or to displacing the stakes and the zones of uncertainty, in order to take advantage of circumstances to force one's adversary onto less favorable terrain or into a position where he must yield. In the case of the railway strikers, the stakes are not the stoppage of the train or even the continuation of the strike. Human lives are at stake. With these new stakes, the strikers as a group (hence indirectly the men chained to the rails) take control of a new zone of uncertainty, involving the future behavior of their group and allied groups in the event that the train runs over the strikers. In short, by chaining themselves to the rails, the strikers place the train's engineer, and particularly the company which chartered the strikebreaking train, in the position of having to choose between the relatively predictable consequences of stopping the train and the less predictable, indeed incalculable, consequences (riots, mass movements, etc.) of the death of the strikers. This explains why it is probable that the train will stop.
If, however, the conditions of the expanded game introduced by the strikers are changed -- for example, by the supposition of a lack of solidarity among the railway workers or of a society in which the value of human lives, or simply of the strikers' lives, is nil -- a different "solution" to the conflct may be anticipated. The consequences of a fatal accident being negligible, the zone of uncertainty controlled by the chained strikers will not be relevant. The train will continue on its way. The weak will remain weak.
The power of an individual or group, or social actor, is thus a function of the size of the zone of uncertainty that the unpredictability of the actor's conduct enables him to control vis-à-vis his partners. We have already explained that not every zone of uncertainty is important. It must be relevant to the problem at hand and in relation to the parties involved. It must be a zone of uncertainty whose existence and control condition the capacity of these parties to act. The strategy of each of the partners/ adversaries will then naturally take the form of manipulation of the predictability of his own and others' behavior, either directly or indirectly, by means of judicious modification of the structural conditions and "rules" regulating his interactions with others.(18) In other words, each actor will seek to widen his own margin of liberty as much as he can in order to maintain as broad as possible a range of behavior, while trying at the same time to limit his partner/adversary's range and to shackle him with such constraints as to make his behavior foreseeable.
An example will help to review and clarify the argument. We have deliberately chosen one whose simple, if not simplistic, character will aid in pinning down our ideas, which we will then be able to broaden as the analysis proceeds.
Mr. Dupont, a prominent and well-to-do resident of a small provincial city, asks Mr. Durand, a craftsman with a small business, to make some repairs on his house. Thus, by this request, a relation of power established between them. The price which Dupont will ultimately agree to pay will be a function of the balance of power which prevails.19 If Durand is the only one in town capable of undertaking the necessary repairs, if he has all the work he needs, and if, for various reasons, Dupont cannot take his business out of town, the zone of uncertainty which Durand controls by his behavior alone is maximal: his client has, in fact, no choice. The balance of power is clearly tilted in his favor. Dupont, however, is not completely empty-handed. He can, in fact, refuse to have the repairs done at all if Durand' conditions seem exorbitant. He can even tip the balance in his own favor if he is able to call on competing craftsmen. This is true, at least, unless these competitors can agree among themselves and thus eliminate Dupont's freedom to choose, which would put us back, mutatis mutandis, in the original situation.
To analyze a relation of power: then, two series of questions must be answered. First, what resources are available to each partner, and what are the advantages by which, in a given situation, each might enlarge his margin of liberty? Second, what criteria define the relevance of these resources and determine their degree of mobilizability; or, to put it another way, what are the stakes involved in the relation, and what structural constraints does it exhibit?
The first of these questions involves us with every sort of resource (individual, cultural, economic, and social) which an actor may have at hand by virtue of his over-all social situation. These define the social, spatial, and temporal framework in which his strategy is set. Such an inventory allows inequalities among the actors to be introduced explicitly into the analysis. These inequalities depend on the relative positions of the actors within a structured social field.
Knowledge of an actor's social situation allows us to understand how he might diversify his investment, or, in other words, how he might be come involved in several power relationships at once.20 The multiplicity of an actor's commitments offers him a considerable advantage from two points of view. On the one hand, such multiplicity protects him against the risks and losses inherent in relations of power insofar as he can thereby distribute his investments and thus avoid "putting all his eggs in one basket." On the other hand. he can strengthen his ability to play offensive games. By his involvement in several relations of power. an actor can parlay the resources accruing to him from other commitments and invest heavily in a specific relation in order to reinforce his position. We will illustrate this statement by expanding our earlier example.
Dupont, as a prominent member of the community, is on the best of terms with the local tax collector. He is therefore in a position to intervene judiciously at the right moment in order to smoothe Durand' periodic negotiations with the authorities over the size of his tax bill. In addition, Dupont's profession is that of estate administrator. As a result, he con trols a large amount of business which Durand might well be interested in obtaining, and for which his particular skills might not in all cases be as indispensable as they are in relation to the repairs on Dupont's house. As thus redefined, the situation is rife with important advantages for Dupont which allow him to "open" the game up and put the relation on a new footing, even if a priori the balance of power was not in his favor.
The social situation of the actors also makes it possible to understand how each actor can make use of the time factor in relations of power. 1S even conceivable that the temporal dimension is essential for a relation of power to develop, since it is a source, and sometimes the only one, of diversification of investments. In fact, an actor can accept loss in the short term only if this loss seems temporary, and he can hope subsequently to win. Above all, time is a dimension of an actor's margin for maneuver. As a result, the capacity to adopt a more distant temporal horizon in a power relationship becomes an important advantage.21 To return to our exam ple, if Durand urgently needs the money from the repairs on Dupont's house, and if Dupont is not in a similar hurry to have the repairs done either because the problem is a minor one or because he has other places to live--then Durand's negotiating margin, "cornered"* in the short term, will be weakened with respect to Dupont, who can "wait and see.
Taking the respective resources of the actors in a power relationship into account thus complicates the initial design considerably. It becomes clear that, faced with the same power relationship, different actors do not share similar alternatives or temporal horizons. In short, their capacity to proportion their commitment or adjust their investment, thereby limiting the risks to themselves which inevitably accompany relations of power, is not the same.(22) By dint of their social situation, actors have different "strategic capacities."(23)
By situating the actors in a structured social field, and by showing how their strategic capacities are limited by their position therein, this type of analysis allows us to give more operational definitions of such notions as "social power," "strength," and "social ascendancy." These can now be defined in terms of a player's capacity (24) to extend the area over which power is exercised so as to move the relation onto terrain where the balance of power becomes favorable to himself. Such an analysis also allows social inequalities among the actors to be integrated into the discussion. Clearly, it is these very inequalities that are responsible for the fact that, from the outset, certain actors will be more likely to occupy dominant positions than others. To return once more to our example, in spite of his weakness in one particular relation, the likelihood is great that Dupont will win in the end: his more numerous involvements allow him to expand the game until the balance of power becomes favorable to himself.
While it is probable that this will be the actual course that events will follow, it is not necessarily so. While economic and social inequalities among actors are fundamental factors in understanding the development of given power relationship, only rarely are these inequalities reflected mechanically and without modification. The special structural constraints of a definite situation can attenuate or cancel out such inequalities. This is because it is insufficient to look only at the actors' resources. These must be mobilizable in the specific relation, and they must be relevant to the objectives of the partners. In the example, if the tax authorities refuse to tolerate any interference in the assessments on tradesmen, and if the sort of work which Dupont, as estate administrator, controls does not fall into Durand's province, then Dupont's particular resources lose much of their relevance and are no longer mobilizable in the relation. Thanks to these specific structural constraints, which at least at a given moment apply to both actors, the balance of power may shift toward the "weaker" partner, Durand.(25)
Now that we have examined power relations from the point of view of the actors involved, it is useful to change our perspective and investigate the structural constraints characteristic of a given negotiating situation. Through structural analysis of the constraints imposed on all the actors in a given relation of power we will be able to answer the second question raised above, namely, what resources can an actor actually mobilize in a relation of power. and how relevant are they?
This is the level at which the structural characteristics of an organization come into play. They establish and limit the range of power relationships among the members of an organization and thus define the conditions for negotiation. They are constraints imposed on all participants.
First, the organization makes the development of continuous relations of power possible. Power. we have seen, does not exist in itself. It can be exercised only in a relation in which two actors are willing to engage, or are forced to engage, for the purpose of accomplishing a given task. Thus, temporarily, at least, they become part of an organized structure. Power and organization are thus inextricably interwoven. Social actors can attain their objectives only by exercising power. At the same time, however, they can exercise power over one another only in connection with the pursuit of collective objectives whose constraints very directly condition their negotiations.26
Second, the structures and rules governing the official functioning of an organization determine where relations of power may develop. These structures and rules, by defining some sectors in which action is more predictable than in others and by setting up fairly easily mastered procedures, create and circumscribe organizational zones of uncertainty. Individuals or groups will attempt to gain control of these zones in pursuing their own strategies, and as a result power relations will grow up around them. The power and capacity for action of an individual or group within an organization depend, in the final analysis, on the control which can be exercised over a source of uncertainty affecting the organization's capac ity to attain its objectives, as well as on the importance and relevance of this source of uncertainty to the other members. Thus the more crucial the zone of uncertainty controlled by an individual or group, the more power he or it will command.(27)
Finally, the organization regularizes the development of relations of power. Its formal hierarchy and internal regulations place constraints on what individuals and groups within it may do, which fundamentally influence their strategic orientation and content. This imposes a certain minimum predictability on all behavior within the organization in two ways. First, the organization affects each member's capacity to play by determining what resources may be employed in power relationships. Second, it influences the members' willingness to make use of these resources in their strategic choices, by setting the stakes which each member can hope to win by committing resources to a particular power relation.
Not all the resources available to an actor are equally relevant or mobilizable within a given organization. The organization's objectives and the types of activities they determine encourage the use of some resources and rule out others. The ability to play the violin is of no use in a machine shop. If you are, on the other hand, one of the few people familiar with the mysteries of an extremely complex machine on which the smooth running of this same shop depends, your negotiating position will be enhanced and your power increased as a result. Furthermore, the organization establishes channels of communication among its members and so defines the opportunities for access to information which the actors may need for carrying out their respective strategies. The organization also invests certain of its members with a legitimate authority over others, or. in other words, with special punitive or recompensatory powers. The members thus favored obtain advantages which may give them a greater weight in negotiations. In our example, it is not a matter of indifference whether the person who knows the machine in question inside and out is the chief official in the shop or an ordinary workman without formal prerogatives. The strategy and hence the behavior of this person, as well as of all the others In the shop, will be fundamentally affected by the circumstance.
To possess advantages, however, is not enough. The members of the organization must also be willing to use these advantages in particular relations of power. Insofar as the organization is always just one area of strategic investment among others for its members, this commitment is in no way automatic, as we have previously emphasized. They will agree to mobilize their resources and take the inherent risks only if the organization offers sufficiently relevant stakes with respect to their advantages and objectives, and stakes that are of sufficient importance to justify their mobilization.28 At this level, again, the organization has a role to play. For example, by the way it organizes promotion internally, by the number and importance of the zones of uncertainty which are left to be read between the lines of the regulations, by the prerequisites established for this or that post and the prerogatives and advantages attached to it, the organization eliminates certain individuals or groups from competition for a given source of power by defining their respective potentials for gain. Not all the organizational zones of uncertainty are stakes for all members of the organization. Competition is organized around intermediate stakes or objectives.(29) Take, for instance, the maintenance function in a production shop. To the extent that the satisfactory performance of the shop depends on maintaining the machines in good working order, the person responsible for this maintenance derives power from his role and obtains advantages which he would like to hold on to. As a result, this function can become a stake for anyone who has the technical qualifications to exercise it. Such a person might attempt to use this position for his own strategic purposes by mobilizing whatever resources he can. Suppose, however. that a rule is put in force prohibiting access to this job to all bachelors, and that all the potential candidates happen to be single. The job will cease to represent a stake for them. The outcome is determined in advance; there is nothing to win. They will therefore invest their re sources elsewhere.
The concrete relations of power which form within an organization, it follows, are never merely carbon copies of the balance of power and modes of domination found in the ambient social structure, nor do they simply reflect the relations of production and the derivative technical and social division of labor. This does not mean, as S. Clegg seems to believe, that the structural inequalities which characterize the action po tential of the various "players" within an organization are neglected.(30) Negotiations among these players are, of course, overdetermined by these inequalities, and the outcome of such negotiations is by itself incapable of abolishing them. To use Clegg's example, borrowed from chess: in the organizational game, the queen (the general manager, roughly speak ing) usually occupies a position of considerable and often insuperable privilege as compared to a simple pawn (the worker), since a certain number of nonnegotiable cultural and legal rules bestow a greater possibility for action on one player than on the other. Similarly, the queen (the manager) clearly cannot act with an altogether free hand: just as the unalterable rules limit the queen's freedom to maneuver, a business man ager's choice of policies is constrained by the logic and rationality of the dominant mode of production and exchange.
Accepting this point, however, in no way helps to account for peculiarities in the functioning of an organization: the same logic of action, as can be observed every day, may be adapted to a wide variety of situations and embodied in quite different policies. As far as organizational analyses are concerned, it is this diversity which interests the in vestigator and which he has to attempt to understand and explain. This is possible only if the development of relations of power within an organization is related to the relatively autonomous structural constraints as sociated with that organization. For it is within the constraints discussed above, and around its formal hierarchy and official rules, that the organization produces its own sources of power. about which a little must now be said.
As a first approximation, four broad sources of power can be distinguished, each corresponding to different sources of uncertainty particularly relevant for organizations. First, there are sources of power which derive from special skills and from functional specialization.(31) Second, there are those connected with the relations between an organization and its environment, or, more precisely, its several environments. Third, there are sources of powerengendered by control of communication and information. Finally, there are sources of power in the existence of general organizational rules.
Before developing these points, however, the limits of any typological classification should be stressed. Clearly, its only purpose is to clarify the ideas involved and to illustrate concretely a mode of argument. As with all resources which actors employ in their strategies, organizational sources of uncertainty are not objective and unequivocal facts.32 While rooted, certainly, in the exigencies of certain technologies or production processes, as well as in the characteristics and particularities of the formal structure of a given organization--all of which may summarily be referred to as the "objective" factors of a situation they are simultaneously integrating elements in the human system which underlies the organization, a response to its essential problems, and, as such, human constructs. They should therefore be analyzed as artifacts, as so many props which the actors in the organization "put together" out of their material and cultural resources to support and sustain their exchanges.
It is therefore imperative to avoid arguments of the following type: such and such a group is in possession of such and such a structural, "objective" source of uncertainty; hence it has a certain power: hence it exhibits such and such a behavior or strategy. The nature of the subject precludes any simple determinism. Here, as elsewhere, a source of uncertainty exists and becomes significant for organizational processes only when it is understood and invested in by actors pursuing their own strategies.(33) The "objective" existence of a source of uncertainty tells us nothing about the willingness or even the capacity of actors to use the opportunity it represents. This explains how, as in the case of the Industrial Monopoly treated above, a single source of uncertainty of an apparently technical nature could become an important source of power profoundly influential in the operation of the organization, or else could remain a relatively minor and unexploited circumstance in another situation, even though the same technology was involved in both cases.
Let us now say a few words about the first important source of power. In fact, we have already had a good deal to say about this factor, as it was the most apparent of the four. Recall that we are speaking of the source of power which has to do with the possession of a special skill or functional specialization for which no ready substitute is available. The expert is the only person with the competence, knowledge, and experience of the situation needed for solving certain problems crucial to the organization. He therefore occupies an advantageous position in negotiation both with the organization and with his colleagues. When the smooth operation of an activity depends on his services, he can negotiate for advantages and privileges in return for them. The mechanism is well known, and there are few exceptions. Take, for instance, the privileged situation of maintenance services in most shops, or the advantages in the hands of the technocrats in France. Or consider the evolution of the power structure in large business or industrial groups where power has passed from the hands of bourgeois families to managers capable of controlling the great uncertainties within these still poorly integrated amalgams.
What is properly called "expertise" is relatively limited. Few persons in a society as complex as ours are really the only ones capable of solving a problem within a given setting. Yet a great many people enjoy a de facto monopoly because it would be too difficult or costly to replace them, in view of their having succeeded, generally by way of organizing as a group, in making or keeping esoteric and inaccessible what special experience and knowledge they possess. In the strict sense, every person within an organization possesses a minimum "expertise" of which advantage can be taken in negotiations, in the sense that it is difficult to find a replace ment (because of the costs of advertising, training, etc.).
The second major source of power found in the organization has to do with the uncertainties which arise in connection with the organization's relations to its environment. It is rather similar to the previous source since control of the environment may be considered a form of "expertise." No organization can exist without establishing relations with its environment or environments.(34) The organization depends on its environments in two ways: first. as the source of the human and material resources necessary to its functioning (supplies, personnel, etc.); and, second, as the outlet for its product, be it a material good or something intangible. Consequently, an organization's relevant environments. the sectors of society with which it thus enters into relation, always represent a source of potential disturbance of its internal functioning and hence a major and unavoidable source of uncertainty. Individuals and groups capable, by means of their manifold connections with one or more sectors of the environment, of at least partially controlling this zone of uncertainty, and of turning it to the organization?s use and profit, will enjoy a considerable power within that organization as a result. This sort of power has been called boundary spanning, 35 since it is possessed by an actor participating in several related systems of action. As a result, he can play the indispensable role of intermediary and interpreter between different and even contradictory logics of action. The sales representative, with his wealth of outside relations, as well as the worker who is an official of his union and who may be instrumental in triggering a strike, are two examples, among others.
Power does not, however, derive solely from the "objective" factors inherent in the technique, the job, and the manifold problems created by relations with the outside world.36 It comes, too, from the active use to which the actors put their positions within the organization. The organization creates power merely by the way in which it organizes communication and information flow among its subunits and members.
To perform adequately the task or function assigned to him, an actor needs information that is in the possession of other individuals who occupy different positions within the work process. If, for a variety of reasons, he cannot short-circuit or forgo their cooperation, such individuals will have powerover him, inasmuch as the manner of their transmission of information (with longer or shorter delays, withholding or "doctoring" of pertinent data, more or less, etc.)(37) will have a profound impact on the recipient's capacities for action. No regulation can do any good in this respect. The recipient can defend himself in such a situation only if he in turn possesses information (or controls another source of uncertainty) affecting his counterparts' capacity to play. The same process of extortion and counterextortion, of negotiation and bargaining, will then develop around the control and transmission of information relevant to each in volved party.(38) A good example is provided by the strategy of lower-level supervisory personnel in the clerical agency analyzed by Michel Cozier(39) To make his decisions, their superior in this organization needs information concerning detailed work situations which these lower-level personnel are responsible for furnishing. As a result, they enjoy a power over their superior which they use to influence his decisions, by biasing the information so as to favor their own interests. This state of affairs is the natural and systematic consequence of their position within the organization, which puts these personnel in competition with one another and leaves them no other means of influencing their superior's decision. on which their own capacity to maintain a suitable working environment within their respective sections depends. They cannot push this procedure too far, however, as it is the superior who ultimately makes the decisions. If they were to provide information too obviously erroneous or biased, he might be prompted to investigate the actual underlying situation for himself, thus sapping the source of their power. These considerations may not always be explicitly present in the minds of the lower level supervisors. Nevertheless, it is easy to imagine that their real strategy will be quite subtle, quite unlike the conventional image according to which information moves up or down as dictated by the rational criteria embodied in the rules and structure of the organization.
The fourth source of power which we have singled out is the utilization of organizational rules. We take it up last because, to a greater extent than its predecessors, it is a construct, and can be understood as a managerial response to the problem posed by the existence of the three other sources of power. We saw previously, in the discussion of power deriving from the control of information, that managerial authority is able to adapt the circuits of information necessary for internal cooperation for its own purposes. An analogous problem arises here. In principle, the rules are supposed to suppress sources of uncertainty. What is paradoxical is that not only are these rules unsuccessful in completely doing away with such sources of uncertainty, but they also create new uncertainties which can be capitalized on by the very person whom the rules sought to constrain and whose behavior they were supposed to regularize.
The negotiation and bargaining which follows in the wake of a rule's application provides the best example. It is generally thought that a rule is a device by which a superior obliges his subordinates to behave as prescribed. By making this prescription precise, the rule reduces the subordinates' margin of liberty, hence increasing the superior's power Another analysis is possible, however, which shows that the rule's rationalizing effect is not one-way. If it limits the subordinates' margin of liberty, it does as much for the margin of choice of the superior. For example, the latter can no longer exercise his power of sanction apart from certain explicit circumstances. Similarly, the rule becomes a protective device behind which the subordinates can take refuge from their superior. By knowing how to apply the rule, they can disarm the superior so far as they are concerned. Since it is usually the case that the smooth functioning of an operation requires more than the rule prescribes, and since the superior is himself judged on the results of the operation, he is placed in a position of weakness, having no way of compelling his subordinates to do more than the rule prescribes.
How can the superior regain control of the situation? Usually, he will have not just one but several rules at his disposal. By allowing his subordinates to flout some of these, he gives himself a way to put pressure on them, as he can always threaten to end this indulgence and return to a strict regime, and can thereby extract a special effort when necessary. Yet he is aware that he cannot go too far, for then the subordinates would take him at his word and apply the strict letter of the rules, turning the tables on him by taking shelter behind them.(40)
Thus, while reducing one uncertainty as to the subordinates' behavior, the rule creates another uncertainty, since the extent of the use they will make of the rule as a protection against the superior's arbitrariness is indeterminate. Hence the power potential in the rule is not so much in its imposition of definite prescriptions as in the possibilities for pressure and negotiation which it creates. The superior's power is the power to create rules which he can use to obtain the desired behavior from his sub ordinates.
When an organization is studied from the point of view of the relations of power by means of which the actors continually use the available zones of uncertainty to negotiate over the implementation of their respective strategies, what emerges is a second power structure, parallel to the one codified and legitimated in the official organization chart. Bringing this parallel structure into the discussion allows the real extent of the official authority conferred by the formal organization to be ascertained, and the real margin for maneuver of the various actors to be measured. In short, such a study can shed light on the "anomalies" and " disparities" which never fail to exist between an organization's official façade and its actual functioning. This parallel power structure, which completes, modifies, and even nullifies the formal plan, is in fact the real chart of the organization. The strategies of all parties are shaped and guided by their relationship to this shadow organizational structure.
Consideration of the actor's experience is indispensable to the understanding of relations which make no sense apart from that experience. On the other hand, the meaning of that experience derives from aspects of power inseparable from the organizational context.
The theses developed in this chapter are the fruit of a series of efforts to extend the original notions of Michel Crozier. A preliminary exposition can be found in "Power and Organization," Archives européennes de sociologie 5, no. 1 (1964):52-64, and in The Stalled Society (New York: Viking, 1973). In addition, we have made use of ideas from the talks given by Jean-Pierre Worms in the course of a series of seminars at the Centre de Sociologie de Organisations (Paris).
This is what the American political scientist Robert A. Dahl meant in defining power as the "capacity of person A to make person B do something that he would not have done without A's intervention." See R. A. Dahl, "The Concept of Power." Behavioral Sciences 2 (1957):201-15; and "Power." Encyclopedia of the Social Sciences, vol. 12 (New York, 1968), pp. 405-15. The attractive simplicity of such a definition should not, however, hide its obvious gaps, and particularly the following: the impossibility with such a definition to distinguish between intentional power and influence without the knowledge of one of the parties, the lack of recognition of the specific dependence of A's power on the action requested, and, finally, the very clear bias toward a viewpoint predicated on the "holding" of power which is still regarded as an attribute of the actors.
The proposition could also be reversed to say that to enter into a relation with someone is in a more or less explicit way, depending on the case -- to put a relation of power into effect.
We will return to this point. For the moment, it should be noted that it is of course not a question of an abstract relation, but rather of a relation in situation, therefore one which is contingent on the actors and on the structure in which they act.
This aspect of "dependence," indeed of interdependence, which is part and parcel of every relation of power. has rightly been stressed by R. E. Emerson, "Power Dependence Relations," American Sociological Review 27 (1962):31-41.
The capacity to experience relations of power one of the central aspects of an individual's relational capacity. We will return to this point in chap. 6, which is devoted to the relations between organization and culture.
See D. Wrong, "Some Problems in Defining Social Power, American Journal of Sociology 73 (1968):673-81.
That is, one of the actors adjusts his behavior in advance to the perceived or merely anticipated desires of the other.
These effects of domination are also what Bachrach and Baratz are working toward with their concept of "non-decisions." Contributing to the community power debate, a controversy which grew up in the United States at the beginning of the sixties, notably involving the books of Robert Dahl, Who Governs? (New Haven: Yale University Press, 1961), and C. Wright Mills, The Power Elite (New York: Oxford University Press, 1956), these two authors argue that the power structure of a collectivity cannot be seriously studied merely by analyzing the decisions actually taken. Account must also be taken of "non decisions," since the power of certain groups manifests itself not in what they are able to accomplish, but rather in what they are able to block -- i.e., in their capacity to produce or obtain "non-decisions" on certain themes or in certain areas of public action. See P. Bach rach and M. S. Baratz, "Two Faces of Power," American Political Science Review 56 (1962):947-52, and "Decisions and Non-Decisions: An Analytical Framework," American Political Science Review 57 (1963):632-42. As interesting as this distinction is, its contribution to a new comprehension of the phenomenon of power seems to us limited by its noticeable tendency -- which underlies, moreover, the whole controversy around community -- to regard power as something "held," essentially an attribute of the actors. Paradoxically, such a point of view prevents the problem of the "structural overdetermination" of the exercise of power from being posed in a meaningful way -- i.e., the problem of power as relation. It is precisely this overdetermination that the authors mean to emphasize with the concept of "non-decision."
To exist thus comes down to entering a field of power since cannot exist except by bargaining with others over my willingness to do what they want of me or by not responding to the "expectations" they have of me. Thus access to sources of power -- i.e., to possible alternatives for behavior -- and the actual utilization of such possibilities become preconditions not only of all relationships to others, but also of all processes of personalization, of access to identity.
The study of powers relation makes it necessary to go back to the situational and structural overdeterminations which alone are capable of explaining the development of observed relations of bower. Power conceptualized as relation thus becomes an instrument of research which makes possible the exploration and analysis of the respective situations of the actors, as well as the structural rules which govern their transactions. We will return below to the structural contingency of relations of power: for the methodological aspects of our procedure, we refer the reader to the methodological appendix at the end of the work.
This point is emphasized and forcefully argued by T. C. Schelling when he pleads for abandoning the assumption of symmetry in the theory of games. See his The Strategy of Conflict (Cambridge, Mass.: Harvard University Press, 1960), esp. Appendix B, pp. 267-90 (1973 ed., Oxford University Paperback).
Ibid.
It is a matter of attempting to extend the theory of games to "mixed" (conflict/ negotiation) games of nonzero sum.
By reintroducing into his analyses strategies which include and make use of chance -- i.e., strategies in which promises and threats made are attenuated by their conditional character -- Shelling makes his argument more subtle by once again giving a central place to the manipulation, by each of the partners/adversaries, of their own margin of liberty as well as that of the other. See Schelling, Strategy, esp. pt. 3, "Strategy with a Random Ingredient,' pp. 173-203.
True, they are predominant in the problems on which Schelling particularly works. In fact, he is interested in the dynamics of international relations in a context of a balance of a terror between two superpowers.
This is the real object of Schelling's analyses. He is interested in the means and strategies (promises, threats, extortions) which allow an actor to use the structural con straints of a situation for moving the game/negotiation/conflict which opposes him to the other party onto a terrain favorable to his own interest. He shows that tying one's hands is actually a possible strategy-without, however, sufficiently specifying the structural con ditions which circumscribe the validity and effectiveness of such a strategy.
Clearly, such transactions are established within the constraints imposed by a market for such services, whose existence we certainly do not want to question. Our empirical reasoning deals with the margin of liberty left by the market, and leads to more general considerations concerning the very development of such markets.
The example discussed at length by Schelling is particularly enlightening in this respect.
It can even become an actor's principal advantage. Indeed, in a situation in which he has absolutely nothing else, he will have nothing left to lose: time itself leaves him in different. Against another actor "in a hurry," this indifference to time may become the advantage which allows him to regain power though he begins from a situation of weakness.
The less one can measure one's commitments, the less prepared one is to commit oneself: this is the origin of the well-known protective strategies which are expressed through "withdrawal" or "apathetic" behavior.
Similarly, the same resource does not increase the actors' "strategic capacities" in the same way. As in other respects, in this connection cumulative processes exist which allow certain actors to make use of resources which would be of no value to others. "Only the rich get loans," as Talleyrand said. What is more, the same "objective" resource will be noted and actually pressed into service by certain actors, while others will not be able to use it.
A capacity susceptible, this time, to empirical analysis.
Note that the completely egalitarian market of economic theory is only a limiting case-largely theoretical, moreover-in which the transaction involves only the exchange of measurable goods, making no allowance for the possibility that the unequal resources and action potentials of the various actors may be used to "rig the market." For such a game to be set up and perpetuated, it is easy to see that very complex constraints must be imposed. The market is a construct.
This does not mean that an actor can establish power relations only inside an organization; the contrary is true, as we have amply demonstrated. It means simply that, whatever the degree of formalization of an "organized group," the existence of relations of bower is synonymous with the existence of a minimum of organization" of the relationships between men.
This is what Hickson et al. hoped to formalize in their strategic contingency model of intraorganizational power. in which the power of a subunit in a given organization depends on (1) its capacity to cope with a source of uncertainty for the organization; (2) the degrees of substitutability of this capacity; and (3) its degree of centrality to the organization in question. See D. J. Hickson et al., "A Strategic Continuity Theory of Intra-organizational Power," Administrative Science Quarterly 16 (1971):216-29. Such a formulation is certainly useful, especially for describing and measuring the distribution of power within an organiza tion at time t. But this is also where its limitations lie. By treating such sources of uncertainty as givens or as "objective" resources, such an approach neglects to investigate the negotiating conditions which define a subunit's potential for using the power it obtains from its ability to cope a with a source of uncertainty, and scarcely permits an explanation of the functioning of an organization in terms of the dynamics of the underlying system of power
In other words, their "participation" in the organization will depend on the stakes they perceive.
The stake can thus be redefined as a means which several actors need at the same time for the purposes of their own strategies and which therefore there will be competition to control.
See S. R. Clegg, Power, Rule, and Domination (London: Rutledge & Kegan Paul,1975).
What is generally referred to in sociological theory as "expertise."
Of course, this distinction has an analytic value only. In reality, types of power and types of zones of uncertainty are usually mixed together, sometimes inextricably inter twined.
This explains why, in such a perspective, qualitative analysis will always precede quantification. This is why the approach of David Hickson and his colleagues is open to such serious criticism. See Hickson, "A Strategic Contingency Theory."
We will return to this problem in greater depth (pt. 2, chaps. 4 and 5).
This concept was put forward by Henri Jamous in his study of the decision-making processes leading to the reform of medical education. See Contribution à une sociologie de la décision: la réforme des études medicales et des structures hospitalières (Paris: Copedith, 1968). See below, chap. 12, pp. 195 ff.
This objectivity is actually only superficial, for all tasks, all techniques are also human constructs.
Or merely by transmitting passively and without any elaboration on their part all the data in their possession. The recipient of such information, incapable himself of distinguish ing what is important from what is not, and equally incapable, very often, of perceiving the real significance of information which comes in in such disarray, will ultimately be as para lyzed as if the information had deliberately been withheld.
Only from such a viewpoint is it possible to understand the strategy of management, or, more generally, of the superior officials in any organization, who use the information available to them in order to secure an additional margin of maneuver for themselves by mixing secrecy and candor in a calculated way.
See Crozier, The Bureaucratic Phenomenon, pp. 13-55.
In a very fine analysis, H. Popitz has shown the importance and the positive functions fulfilled by ignorance of legal rules, infractions of which, he shows, are not punished. See H. Popitz, Uber die Präventivwirkung des Nichtwissens (Tubingen: Mohr, 1968). See also the analyses of A. W. Gouldner devoted to the functions of bureaucratic rules, for example, Patterns of Industrial Bureaucracy.