'74
Bureaucracy as an Organizational System
able to eliminate the usual power dependencies but must rely instead on a rigid and authoritarian system of hierarchical social relations. Such a system can survive only because of the very deep and anachronistic dichotomy between managers and managed and the acceptance, partial at least, of their inferior social position by the subordinates. It is reproducing, forty years later, the paradox and contradictions of scientific work organization.
THE BUREAUCRATIC SYSTEM OF ORGANIZATION
The Problem
Study of the way power is allocated and an analysis of the bargaining strategies between individuals and groups within an organization are unusual starting points from which to try to reach an understanding of the function of an organization. We have taken this course because of its provocative and challenging properties in a domain that often seems paralyzed by formalism. Our effort appears to have been worth-while. For example, the distinction between organizations characterized by a stationary power equilibrium and organizations with constantly shifting systems of bargaining is apparently useful for the theory of organization. However, as we have also shown, the world of power is only one aspect of the complex relationship between the individual and the organization. The world of consensus and the world of the co-operative game are other important aspects of this basic relationship.
We shall approach our subject in another, broader way. We shall attempt to analyze these two complementary aspects of the functioning of the organizations we have studied and we shall interpret our data in the light of the theoryof bureaucracy—that theory in terms of which sociologists since Max Weber have been considering the processes of organization.
There is a paradox, however, in the long series of discussions over the theory of bureaucracy. During the last fifty years, many first-rate social scientists have thought of bureaucracy as one of the key ques- j tions of both modern sociology and modern political science. Max j Weber had furnished a very brilliant description of the "ideal type" ola bureaucratic organization, and a suggestive analysis of its historical development, that apparently paved the way^for a poskive yalue-free sociological analysis. Yet the discussion about bureaucracy is still, to a large extent, the domain of the myths and pathos of ideology.
The paradox, indeed, exists already in Weber's work. In his studies on law and on Prussian bureaucracy, it is true, Weber presents a view which is richer and more sophisticated than the rationales
iy6
Bureaucracy as an Organizational System
ist model of the famous chapter of Wirtschaft und Gesellschaft. But when he claims the superiority of modern bureaucracies that embody the "ideal type," one may wonder^ whether JiejJojsjMOlunkj^t these organizations succeed only inasmuch as thej^can impose^ ^hstanjJMjaid."^n8g^Siamount of .standardizatiQn pn'IEeir menv bers, Some of his statements show how much he was worried by such a trend.1
Among Weber's successors and contemporaries who used similar models more rigidly, contradictions are still deeper. On the one hand, most authors consider the bureaucratic organization to be the embodiment of rationality in the modern world, and, as such, to be .intrinsically superior to all other possible forms of organization. On j the other hand, many authors—often the same ones—consider it a I sort of Leviathan, preparing the enslavement of the human race. Optimism and pessimism are mixed in various ways. Whatever their proportions, there is always a double belief in the superiority of bureaucratic rationality—in the domain M.„cffipi§Bcy» and in its threateningjmp domain of human values.
This was well exemplified by Robert Michels' syllogism of the "iron law." Michels, after cruelly "disenchanting" the "charismatic" enthusiasm of the Socialist movement, pointed to the dilemma of all ! democratic social action, reformist or revolutionary: democratic social : action is possibly and. bureau:
i cratic organization is destructive or democratic values.2
This paradoxical view of bureaucracy in Western thought has paralyzed positive thinking on the problem and has favored the making of catastrophic prognostications. This has been most clearly perceptible in the great stream of revolutionary pessimism which has so pervaded Western political and social thought from Rosa Luxemburg and Trotsky to Bruno Rizzi, Simone Weil, and C. Wright Mills—to cite only the most famous names. A desperate gamble is the only hope. The picture is blackened, and the attack against what is felt to be inevitable if human beings accept their fate is exaggerated. Thus, by a pointing up of the paradox, only one alternative is left—the dialectical overstepping of revolutionary belief.
The same paradox is also perceptible in more conservative circles. For example, it may be discerned in Burnham's simplifications, and in the great number of attacks, such as those of William H. Whyte,
1See, for example Max Weber's remarks at the 1909 convention o£ the Verein fur Sozialpolitik, as quoted by J. P. Mayer, Max Weher and German Politics (London: Faber & Faber, 1943), pp. 127-28; and by Reinhard Bendix, Max Weber: An Intellectual Portrait (New York: Doubleday, i960), pp. 455-56.
% Robert Michels, Zur Soziologie des Partetwesens in der modernen Demokratie (Leipzig, 1912).
The Bureaucratic System of Organization 177
Jr., against bigness and the bureaucratization of modern life. As Alvm Gouldner quite sensibly said:
Wrapping themselves in the shrouds of nineteenth century political economy, some social scientists appear to be bent on resurrecting the dismal science. Instead of telling men how bureaucracy might be mitigated, they insist that it is inevitable.3
March and Simon have argued that Max Weber's thinking about bureaucracy corresponds to the early rationalist theory about numan behavior in organizations. This statement may be exaggerated as regards Weber himself,4 but it is quite accurate for Michels and all the revolutionary analysts of bureaucracy. We':go even farther and suggest that the modern "disjn^alsc
leHuat"sophistication^ results, from its. relying on the .same crude theory" o£., KumaS motivation as the scientific engineers. It is only because its exponents accept, at their face value, the Taylorian arguments, that they can believe that modern organizations are succeed-j ing because of their evil features. '
The first decisive progress in resolving this contradiction was made when Robert K. Merton, and after him other American sociologists like Gouldner and Selznick, began to question the perfection of the "ideal type," ancTto chscuss, m aj^sjBwand"^^^KaI way, whether tne^pp^iBolTbetween organizational efficiency and the freedom of. th^ndividual was actually possible. Beginning with the theory of unintended con sequences successive research by the sociologists of bureaucracy has suggested that the routine and oppressive aspects of bureaucracy are so many elements of what may be described as a
....."viciouscircle" j* that develops from the resistance of the human
factor to the mechanistic rationalist theory of behavior which is being imposed on it. This very resistance, paradoxically, tends to reinforce the use of the theory. Research.has demonstrated that the ideal type! of bureaucracy is far from Joeing cpmpl^ly^efficient. \
"Alvin Gouldner, "Metaphysical Pathos and the Theory of Bureaucracy," in Amitai Etzioni, Complex Organizations (New York: Holt, Rinehart, 1961), p. 82.
1 One can argue that Weber is more interested in the problem of social control than in the problem of rationality, and that he never forgets to take into account the fact that a bureaucracy must have some kind of legitimacy and thus rely ultimately ^n,,consensus^ Yet one can notice that, in his analysis of bureaucratic rationality, he emphasizes above all the predictability requirements and the standardization of behavior that provides the only way to meet them. There we can understand how his ideas can converge with those of promoters of scientific work-organization. If the latter do not want to see anything but the simplest economic motivations in human behavior, this is because such a simplified approach makes it possible for them to consider each human being as an interchangeable instrument whose response to organizational stimuli is entirely predictable.
6 March and Simon were the first to point out these logical implications of Merton's analysis; the image of the "vicious circle" is ours.
'78
Bureaucracy as an Organizational System
The Bureaucratic System of Organization
*79
This demonstration, however, has so far been more negative than positive. It does not enable us to answer Weber's question about the evolution of industrial society. It suggests that the link between the rational and successful features of a bureaucratic organization and its dysfunctional ones is much more complex than the Taylofian engineers and their contemporary critics thought. It has not, as yet, explored to any extent their interrelations and the conditions of their symbiosis. In order to avoid the paradox, one should analyze more precisely the general etiology of the dysfunctional features of bureaucracy, their limits, and the extent to which the development of modern organization is influenced by such features.
In this chapter, we should like, through a re-reading of our data, to examine the possibilities of going beyond this first stage of understanding by considering bureau^ not as dysfunctions, nor even as lecessary dysfunctions of the modern system of organization, but as... rational„parts„o£„the„„"bureaucratic systems of organization." This will enable us to answer, in completely different terms, the question Weber raised about evolution. We shall try to show that a bureaucratic system is restricted in its possible rate of development, and that evolution toward large-scale organizations js^ depends, i to a large extenfTon tne ability of men to break out of the bureaucratic "vicious circle."
Before attempting such an exploration, however, we must consider more seriously the decisive contributions made by the sociologists who have used, directly or indirectly, the theory of dysfunctions to analyze bureaucratic phenomena.
The Theory of Bureaucratic Dysfunctions: The "Human Relations" Approach to Bureaucracy
Interest in the dysfunctions of bureaucracy evolved with the discovery of the human factor and the wide development of the human relations approach in industry. This is not fortuitous. As March and Simon have shown, there is a logical link between the rationale of human relations theories, whatever their kind, and the theory of bureaucratic dysfunctions.7
If we accept the conclusions of the interactionist school or of the
" "Bureaucratic" is used here in the popular, pejorative sense.
'James March and Herbert Simon, Organizations (New York: Wiley, 1958), pp. 36-47. We will rely heavily throughout this section on the extremely penetrating analyses of these authors.
Lewinian school, or of both, it is as difficult to uphold Max Weber's thesis as to believe in the credo of scientific work organization. When one believes that human activities depend on the feelings and sentiments of the people involved, and on the interpersonal and group relationships that influence them, one cannot expect that imposing economic rationality on them will bring constant and predictable results. The functioning of a bureaucracy can never henceforth be totally explained by the combination of impersonality, expertness, and hierarchy of the "ideal type." If, in addition, one thinks that the most efficient leader of a group is a permissive leader, one must also believe that the best results will not be achieved by the most rational organization in the Weberian sense. They will be attained by the organization in which subordinates participate most in the decision-making process.
Sociologists, however, did not break with the Weberian model in the way that "human relations" analysts did with the classical Taylorian model. When Merton pointed the way with his two famous pioneering articles of 1936 and 1940,8 he did not contest directly, the. validity^ He simply indicated that
suci^.s£sj:ejn..,QJL .action entails .secondary consequences that run counter to.its objectives and princ he did even
this indirectly. Little by little, however, a new scheme was suggested, the rationale of which is more clear today. Merton contends that the discipline necessary for obtaining the sjandard^ required in a bureaucratic organization will bring about a displacement of goals. Bureaucrats will ..show "nt^alist'^'jtjitudes that will make them unable to adjust adequately to the problems they must solve. This will entail the development of a strong esprit de corps at a group level and create a gap between the public and the bureaucracy.
Such an analysis is implicitly based upon the following argument. The behavioral rigidity, difficulties of adjustment to the task, and conflicts with the public that exist in a bureaucracy, reinforce the nj^Jforjpritrol and regulation. Thus, finally, the unintended and dysfunctional results of the bureaucratic model tend to reinforce their hold. In human relations terms, dysfunction appears to be the consequence of the resistance of the human factor to standardized behavior that is imposed upon it mechanically. Immediately, however, naive questions emerge: Why do people stick to the mechanical model if it does not bring the desired results? And if they do stick to
s Robert K. Merton, "The Unanticipated Consequences of Purposive Social Action," American Sociological Review, I (1936), 894-904, and "Bureaucratic Structure and Personality," Social Forces, XVIII (1940), 560-68.
180 Bureaucracy as an Organizational System
it, why is the model a static one, why does it not deteriorate? After all, if the dysfunctional consequences of the model entail the use of more control and more regulation, one should gradually find more and more dysfunction, Merton does not discuss such questions, which is the reason one can say that he remains within Weber's frame of analysis. His aim at the time of writing was only to show that the "ideal type" includes a fair amount of inefficiency and to understand the reason for such a discrepancy between Weber's model and the reality.
Many American authors—e.g., Reinhard Bendix, Philip Selznick, Alvin Gouldner, Peter Blau, and Robert Dubin—have followed, in Merton's footsteps. Their empirical studies have confirmed Merton's views. Their considerations have made possible the elaboration of a much clearer and richer scheme for the understanding of what we should to call the "vicious'circle" of" bureaucracy. We will examine here only the works of Selznick- and Gouldner^ the two best-:. elaborated studies. Their theses demonstrate well both the progress realized by, and the inherent limits of, the 'Mysf^cJd^oal-SjdbflQL" Selznick's 1949 study9 deals with the TVA, at that time at the height of its prestige as a model of •.democratic organization and as the symbol of the yearnings of the New Deal. In a way, Selznick's book is similar to the Michels' study; it shows the development. pL. bureaucratic... oligarchy, .behind the _ye.il...of democratic procedure. Selznick's aims, however, are different from, if not the opposite of, those of Michels. Selznick starts with the assumption that bureau-..
cralic pressures must be taken for granted and that the .problem is to.....
understand how people manage, to .control......them partially—thus
answering the second of our questions about Merton's scheme.
Selznick's domain, however, is not the same as Merton's. Selznick uses similar reasoning about unintended results, but applies it to a different part of Weber's ideal type, expertness. He shows how the same kind of vicious circle can develop with expertness and specialization as with hierarchical control and standardization. Specialization grows because decisions have to be made on neutral technical grounds. However, specialization makes the experts more narrow-minded and caste conscious at the same time that outside economic interests.and- pressures t6:3ver|p-iwith their caste policies. These dysfunctions naturally call for more specialization, and a new vicious circle develops.10
' Philip Selznick, TVA and the Grass Roots (Berkeley: University of California Press, 1949).
" We are borrowing this scheme of analysis from March and ^imon, op. cit., pp. 4°-43-
The Bureaucratic System of Organization 181
Selznick's most original thinking, however, deals with the problem of the remlfttiffm.. "f.. J^f1..l.Bg!J^H- How does an organization limit the cumulative influence of its dysfunctions? Selznick observed two mechanisms widely used in the TVA. The first concerns the power situation: this is the mechanism: of co-optation, through which outside sectional interests and internal, narrow, expert points of view are given a .share in the decision-making .process—opponents are co-opted to policy-making bodies. The second is the diffusion of a specjalTVA ideology that secures, by its pervasive influence on people in every echelon, the necessary minimum of conformity and loyalty to the organization.11
In this new perspective, bureaucratic dysfunctions are perceived as more diffuse than in the scheme of Merton. They correspond to the rigidity of the organization that may manifest itself in the logic, of specialization as well as in thejogic of centralizatijan. Moreover, Selznick deals not only with the problem of the resistance of the human factor to the mechanistic.. model,, but also with the. problem of power and participation. This second—and, in our opinion, central—problem, however, is analyzed only as regards the possibilities of controlling the development of dysfunctions, rather than as the source of them. Thus Selznick still remains within the Web-erian scheme.
Gouldner s study is, in a way, more limited since it does not raise questions outside the hierarchical control pattern delimited by Merton. Nevertheless, it goes one step further in this direction, by showing better the inherent contradiction of the ideal type.
Gouldner makes a primary distinction between bureaucracy centered about expertness and bureaucracy centered about punition. He deals (much too rapidly for our taste) with bureaucracy founded on expertness—accepting, on inadequate evidence, that it can escape dysfunction, since the values on which it rests can be accepted by everyone, and the rules that regulate it can be elaborated with enough participation by those who must submit to them. Gouldner's actual subject is punitive bureaucracy. He sees it from three different angles: (1) as a vicious circle around problems of subordination and control; (2) as a behavioral pattern with some latent functions; and (3) as a rational response to an accidental but ineluctable event, succession.
The central bureaucratic vicious circle, in Gouldner's view, concerns the problem of close supervision. Impersonal bureaucratic rules evolve because they alleviate the tensions created by subordination
"This analysis was first presented in Selznick, "Foundations of the Theory of Organization," American Sociological Review, XIII (1948), 25-35.
182 Bureaucracy as an Organizational System
and control; but at the same time they perpetuate the very tensions that bring them into being.12 They especially reinforce the low motivation of the workers that makes close supervision necessary.13
The kterj|Jwctipn,Q£j2!£^u^dc rules is to reduce the tensions that are."3ue, in part, to the differences of values among groups, to the impossibility of elaborating norms acceptable to everyone, and to the steady decrease of friendly informal interactions.14
Gouldner's analysis, at this stage, is not completely satisfactory. At first, it is not easily perceived why values of different groups have to be different, why friendly interactions must decrease, and why acceptable norms cannot be developed. One must seek another underlying factor; and this Gouldner does with his theory of succession. In his final view, the irngerjpjjyaljtyjjEJbK^U^^ is a global organizational response to the problem of succession. This summoning of an external, accidental cause is hot very satisfactory, although the case study is an extremely perceptive one. All modern organizations must face the problems of succession, but they nevertheless have very different and disproportionate kinds of dysfunctions.
Gouldner's decisive contribution, however, does not lie in his last and somewhat confused explanatory stage. It is in his earlier and more suggestive interpretation of the role of punitive bureaucratic rules. He has shown how such rules may beusedJn an organization by supervisors and subordinates alike. Both groups take^dvantage or them for punitive purposes and as a bargaining took/inasmuch as each can suspend their application or insist upon it.15 t/Infortunately, Gouldner does not use this insight to broaden his functional analysis. His functional analysis remains partial insofar as it does not integrate the problems of power relationships and the bargaining processes between groups.
This time, nevertheless, the Weberian model is at least partially overstepped. Gouldner does not' unders^dJbjS^?^?, features as merely a mearirfpr achieving efficiency; he views them, rather, as a way to"reducetensions within an organization. ,,vv:f "
These three examples, the most striking of human relations researches on bureaucracy, demonstrate a substantial moving away from the original Weberian model. These three authors give more, and more place..taihfi.joutine and oppa^vje-aspects of bureaucracy that finally become an autonomous parau^Lsystem^orJiiHrian .Cftusa-tion. But all three—with the partial exception of Selzniclc—tencTto neglect the aspect of rationality and efficiency that is central to any
u Gouldner, Patterns of Industrial Bureaucracy, p. 177.
"Ibid., V. 178.
" Ibid., p. 240.
M Ibid., pp. 172-74.
The Bureaucratic System of Organization
183
kind of organization. Moreover, they do-not try to • analyze the~ possible interdependence between rationality and dysfunction. This makes it impossible for them to question the dynamic part of the Weberian model, its analysis of the unrelenting evolution toward large-scale bureaucratic organization. Their contribution is inadequate for a discussion of whether or not the resistance of the human factor to the rationalistic push toward more and more efficiency can be successful, or whether or not it can have indirect consequences on organizational rationality itself. Even Gouldner's analysis remains static. At best, it can be described as viewing bureaucratic evolution as a cyclicaLpxocess with alternate bureaucratization and debureau-cratization phases according to the accidents of succession.18
Finally, as March and Simon have pointed out, all these contributions, while setting the stage for further progress, are limited by the human relations theories on human behavior upon which they rest. To go one step further, one must accept the fact--that members of an organization are not governed solely by affective, motivations buf operate as autonomous actors, each one with his own personal strategy*~thls means giving a central place to the power problems we have analyzed in the last chapter. To do this, we shall first examine theproblems.of government raised by the functioning of an organization. \Power relationships^can be considered as the operational consequences of these.
The Problem of the Achievement of Conformity
Let us accept the main weakness of the sociological theories of bureaucracy as their difficulty in mte^ra.tin^_change and „dey_elop-ment and their reliance on exterior factors to explain the emergence of bureaucratic patterns. Then, in order to go beyond their process-centered analysis, we must raise, in political science terms, the more general question of an organization's government and its minimum requirements.
Planned co-operative action is possible only if one can rely on a greafdeal of r^guIarity^Fbehavior on the part of all the participants. In other w6rds,~arvy~orgahTzation members a
variable but always substantial amount of conformity. Members will conform partlyjypluntajjly, partly because of coercion. Whatever the
14 In their recent book, Formal Organizations (San Francisco: Chandler, 1961), Peter Blau and W. Richard Scott present a more sophisticated point of view on change, which they tend to see as a much more natural and complex process. This excellent contribution, which I was able to read only after my book had gone to the printer, seems to converge on many points with mine.
184 Bureaucracy as an Organizational System
- proportions, the adJsxsi^ be the
central problem of an organization's government.
When we review more analytically the means used to influence people to conform, we can observe very striking changes in modern times. Only two jpenriiries ago, conformity within an organization was obtained!T&ougFTvery narsh and direct means, with a great deal of open coercion. No regular army could function without the painful and unavoidable experience of drilling that we have all but forgotten today. Supervision by overseers in the eighteenth-century mills was almost equivalent to brutal coercion, and, after all, the galleys were one of the great organizations of their time. Religious or religious-like ideologies partly helped people to internalize organizational aims, but these ideologies themselves were of a fanatic type and terror-ridden. Moreover, conformity com obtained in a specialized and limited way; rather, it required a lifelong total commitment. As a consequence, people were molded in a way that would deeply mark th^eir„,:wM life. Whether they were employees of the Fugger House, members of the Jesuit order, or Prussian grenadiers, they had to devote themselves completely to the organization. Leaving it was equivalent to treason; no large-scale organization could be efficient without such drastic conditions.17
Comparisons with modern organizations may seem overdrawn. Yet they must be made, to bring some perspective into the traditional debate on the standardization and the threatening enslavement of modern man. Modern organizations, in contrast to their predecessors, use a much more, liberal set of pressures. They deal witfujpeople who, through theireducation,have already internalizeoTa^number of basic conformities and a general ability to conform easily to., an organization's way. Then, too, there has been a great Heal of progress in the field of training, and no one feels obliged any longer to make people spend months trying to master the exact observance of petty details. Most important of all, human ...behavior is n^ stood.jndtherefore ^m Because of this, a modern
organization does not need the same amount of conformity to get as good results as did earlier organizations. The modern organization can tolerate more deviance, restrict its requirements to a more specialized field, and demand only temporary commitments. For all these reasons, it can and does rely more on indirect and intellectual means to obtain conformity: communication structure and work
17 Sociologists have so far neglected die very important and in some ways crucial body of knowledge to be gained by examining the records of the first large-scale commercial organizations, integrated armies, and religious orders. The theory of organizations could well be enriched by a renewal of such studies.
The Bureaucratic System of Organization 185
flow, the technical setting of jobs, economic incentives/and also, perhaps, rational calculus of a higher sort. The punitive aspect of the conformity achievement process has declined. Direct coercion is still in reserve as a last resort, but it is very rarely used, and people apparently no longer have to see it operate often to retain it in their calculations,
The contrast is further demonstrated by comparing the quasi-monarchal type that seemingly was necessary for maintaining large-scale organizations at the beginning of the capitalist era with the relatively easy-going and tolerant corporation of the affluent society.18 But even during the last thirty years, there has been a not insignificant change throughout the Western world. Compare, for example, the discipline and conformity of dress and behavior imposed on the sales force of a large department store today with the standards of thirty years ago—a great deal more was required then, and in a much harsher way.
Such a reminder makes it possible to understand that, contrary to some easy generalizations, certain "bureaucratic" traits were more likely to appear in earlier organizations than in those of today. Displacement of goals is a case in point. People trained to a rigid conformity, entailing consequences for their whole personalities during their whole lives, are much more likely to indulge in "goal displacement" and "ritualism" than people who are only temporarily specialized—even, as is most often the case, if this specialization is a far narrower one.
Rigidity, however, is not engendered only by pressures coming from above. Conformity is not a one-sided process. Subordinates will bargain with their own conformity and use it as a tool with which to bind management. This is just another aspect of the fight for control. Subordinates tacitly agree to play the management game, but they try to turn it to their own advantage and to prevent management from interfering with their independence. When this double pressure is stabilized and leaves very little freedom for adjusting difficulties, then an organization has become deeply rigid. This was the case with the earlier ritualistic clerk who made a point of following his instructions to the letter and ignored the reality with which he had to deal, not only because of his "trained incapacity," but because he needed protection against too harsh treatment in case of error. His "bureaucratic" behavior was not the consequence of the weakness of his own human )nature. It resulted from the way conformity and rationality had to be observed at that time, when there were so few possibilities of intellectually understanding and
" Notwithstanding William H. Whyte's and many others' criticism. For a full discussion of their stand see below, pp. 297-98.
i86 Bureaucracy as an Organizational System
foreseeing complex situations, and thus of devising in advance flexible programs to adjust to them.19
There are comparable modern patterns—e.g., in the Clerical Agency. Supervisors prefer taking routine, inadequate decisions to facing hostile relationships and possible risks of failure. This modern / bureaucratic behavior originates in more complex organizational processes; but these processes, too, are the outcome of bargaining among various groups, and between the organization and these groups, about the way to impose conformity and rationality.
jEvigryj)rganization, how^ex, must continually adjust to some kind of/change] It must b^]|exibley To achieve this vague but primary end/it must rely on individual and group ingenuity and cannot discourage it too much.20 The organization must consider this goal when devising ways tOm irflpose^conformity, since it is counter to the other primary goal, 'predictability^ What will be the outcome of ; these two conflicting aims? Its environment, its goals, the kind of fluctuating reality to which an organization must adjust, will be the most important factors to be taken into consideration in this per- , spective. But market uncertainty, as such, is not an omnipotent deterrent to rigidity. Extreme conditions of uncertainty will tend to result in more conformity and rigidity, since trying to adjust to com-pletely unpredictable situations will not be rewarding enough. Too little uncertainty, on the other hand, will make it feasible to pre-scribe-in great detail all possible forms of behavior, thus achieving a ;, high degree of rigidity. There will be a tendency to escape from reality at the two extremes, when reality is too difficult to cope with and when it is no longer a challenge. I
It is clear, in any event, that in any kind of organization there is a constant pressure to escape from reality. This tendency corresponds , to what popular sentiment calls "bureaucratic tendencies." Centralization is one of the ways to achieve it; completely impersonal rules are another. Both permit escape from otherwise necessary adjustments. .
But in a normal case, evasion cannot go too far. There is a constant feedback of information that permits and even obliges the organization to take account of its errors and to correct them. We shall describe as a "bureaucratic system of organization" any system
10 Even not so long ago, at the beginning of mass production, the product was imposed on the consumer without any possible adjustment. Motorists had to like' their Fords black until it was deemed possible to predict well enough the probable variation of tastes.
20 In the traditional framework of the nineteenth century, it is the entrepreneur ■ who monopolizes the organization's capacity of adjustment. But such a system was feasible only because production units remained small. *
The Bureaucratic System of Organization 187
of organization where the feedback process, error-information-correction, does not function well, and where consequently there cannot be any quick readjustment of the programs of action in view of the errors committed. In other words, dijbw^amsrMie^rgan^at^n-is an organization that cannot correctits he}igviqr h% learning from its errors^ Bureaucraticpatterns of action, such as the""imp^rs^bnaKtyrof the rules and the centralization of decision-making, have been so stabilized that they have become part of the organization's self-reinforcing equilibria. Finally, when one rule prevents adequate dealing with one case, its failure will not generate pressure to abandon the rule, but, on the contrary, will engender pressure to make it more complete, more precise, and more binding.
The Basic Elements of a Bureaucratic Vicious Circle
Let us now try to build a modejj^sjich^ self-reinforcing "bureau-cra_tic_sys.tein'.'-.by using the characteristic traits that emerge from our two case studies. In doing so, we shall, of course, be limited by the specifically French nature of our data. But we shall at least have a starting point, from which we can see whether other models can be imagined and whether there are some traits, some underlying pat-terns,-common to all possible "bureaucratic systems."
We shall begin with the basic elements of the bureaucratic "vicious circle" as we observed them in the daily operations of the Clerical Agency and of the Industrial Monopoly. This will give us a static model comparable, in many respects, to the models of Merton and Gouldner. But this first model's main achievement will be to enable us to discuss the system's reactions to change and its own pattern of change. This last step will make it possible to present a general assessment of the self-reinforcing equilibrium.
Four basic elements seem to be necessary for the stability of the vicious circle we have observed: the extent of the development of impersonal rules; the centralization of decisions; strata isolation and concomitant group pre^u3rej^.._.the..in<Iividual; and the development of parallql power relationships around remaining areas of uncertainty. We shall analyze them successively.
the extent of the development of impersonal rules
Impersonal rules delimit, in great detail, all the functions of every individual within the organization. They prescribe the behavior to be followed in all possible events. Equally impersonal rules determine
188 Bureaucracy as an Organizational System
who shall be chosen for each job and the career patterns that can be followed. In our two cases, as well as in most sectors of French public administration outside the higher executive class, two basic rules dominate the field. The first rule is that open competitive examinations {coricours) govern promotion from one main category to another. The second rule is that seniority determines job allocation, transfer, and promotions within each main category. Candidates' personalities and past work achievements must be ignored when these examinations are evaluated, as well as when the seniority rules are applied. Judgment must be made only on their most abstract abilities or their most impersonal characteristics.
This system, of course, always affords some loopholes, but viewed from a certain distance it looks extremely tight. In the Clerical Agency, the work behavior of the people occupying all routine and lower supervisory jobs is minutely prescribed. All operations to be performed, the way to proceed, and even their sequential succession, are specified. As regards promotions, exceptions were made to the rule of competitive examinations as determining career patterns at the time of the Liberation, when many people were "integrated" into the next higher category during the great reshuffle of the organization. Since then, no exceptions to this rule have been possible. The seniority rule does not apply absolutely to the supervisory jobs, but exceptions are made only at the end of an individual's career— i.e., after twenty years of seniority promotions. Rules are even more strictly enforced at the Industrial Monopoly, where for many years there have been only a very few promotions21 from one to another of the six categories, and where seniority prevails even for the executive jobs.
As a consequence of the combination of these two sets of rules about job specification and job allocations, nothing seems to be left of the arbitrary whim and individual initiative of an organization member. The daily behavior of everyone, as well as his chances of having to perform a different routine later, can be predicted exactly. In such a system, as we have established, hierarchical dependence relationships tend to disappear or at least to decline considerably. If no difference can be introduced in the treatment given to subordinates, either in the present definition of the job or in the fulfilment of their career expectations, hierarchical superiors cannot keep power over them. Superiors' roles will be limited to controlling the application of rules. As a counterpart, subordinates also have at their disposal no possibility of pressure, no bargaining power over
21 And only from production worker to lower supervisor after many years as temporary supervisor.
The Bureaucratic System of Organization 189
supervisors, inasmuch as their own behavior is entirely set by rules.
Every menibex--of the organization, therefore, is protected both from hisjupe.riars..and from his subordinates. He is, on the one hand,
totally deprived, of ^.initiative andf completely controlled, ..by......rules,
imposed on him from the outside. On the otherTiand, he is completely free rrom^jeflsonal jntej|ejr£rjce. by any other individual— as independent, in a sense, as if he were a non-salaried worker. Such a system of human relations devaluates superior-subordinate relations. Data have shown, both in the Clerical Agency and in the Industrial Monopoly, that such rapports have lost their affective significance for the supervisor as well as for the subordinates, and that they exist only on a conventional basis, with little emotional commitment from either side.
Practically, of course, the system can never be so tight as it can theoretically. There is always some possibility of play within the framework delimited by the rules, and therefore dependence relations and bargaining are never completely suppressed. The curious practice of the greve du zele—striking by slowing down the work flow and paralyzing the functioning of the organization just by observing, to the letter, all the required prescriptions—has been repeatedly used in many sectors of French public administration, precisely as a way of expressing the fact that rules cannot take care of everything and that management must rely on workers' support and must therefore bargain for it.
the centralization of decisions
The power of decision-making within a bureaucratic system of organization's located exactly at idie points where the stability of the intern^
functjond_g6als-biE &eoj^mkztion. This trait can be considered as strongly complementary to the first one. If one wants to keep strictly to the climate of impersonality, it is essential that all decisions that have not been eliminated by the system of rules be made at a level where those who make them are protected against personal pressures from those who are affected by them.
Therefore, the power to make decisions and to interpret and complete the rules, as well as the power to change the rules or to institute new ones, will tend to grow farther and farther away from the field where those rules will be carried out. If the pressure for impersonality is strong, such a tendency toward centralization cannot be resisted. As a consequence, constant priority will be given to internal political problems—the struggle to get rid of favoritism and discrimination
19°                       Bureaucracy as an Organizational System |
and to maintain the. equilibrium between the different parts and |
subsections of the system over the problems of adjusting to the j
environment. The pressure of the latter suggests locating the power 1
of decision-making where all the relevant facts will be best known i
at first hand. This is not possible if one gives precedence to internal ,|
considerations. The power of initiating decisions, given at one arbi- ■ ."i
trary level, cannot but introduce differences among individuals at |
this level, and, as a secondary consequence, introduce dependence 1
relationships above and below. Such dangerous power can be placed ,j|
only very carefully, and due account must be taken of all groups' J
and subgroups' particular claims. Centralization is thus the second 1
means of eliminating discretionary personal power within an organ- I
ization. The price the organization has to pay for it is still greater |
rigidity. People who make decisions cannot have direct firsthand . ]
knowledge of the problems they are called upon to solve. On the |
other hand, the field officers who know these problems can never -j
have the power necessary to adjust, to experiment, and to innovate. |
The Clerical Agency was the best example of such centralization. |
We have dwelt at length upon the reciprocal relationships between :
lower and higher supervisorsrThese embodied perfectly the dilemma 1
of a bureaucratic organization and the choice it is bound to make. J
But the same pressures operate and the same solutions are taken j
above that level, in the Department to which the Clerical Agency |
belongs. The same is true, as noted, in the Industrial Monopoly, ;|
where the director and the assistant director have to decide on all j
shop problems, about which they are not in a position to know first- i
-hand. By and large, patterns of this sort, if not always so easily per- ; , j
ceptible, exist throughout all French public administration. 1
■if
the isolation of the different strata and the concomitant ji
group pressure on the individual j
The suppression of most possibilities of discretion on the part of -i superiors and of most possibilities of bargaining interference on the part of subordinates, through the system of impersonality and centralization, has another important consequence. Each hierarchical cate- ; gory, each stratum, will be completely isolated from all other strata, above and below. A bureaucratic organization, therefore, is composed of a series of superimposed strata that do not communicate very much with each other. Barriers between strata are such that there is very little room for the development of cliques cutting across several I categories.
This strata isolation is associated with a pressure of the peer^roup
The Bureaucratic System of Organization 19 r
on the individual that is much stronger than usual. Where hierarchical pressure dwindles and where there can be no cliques that unite people of different strata, the peer group—i.e., the group of equal members of the same stratum—becomes the only force that stands between the individual and the organization. Since there must always be complete equality among the members of one stratum, the only possible discrimination is that of seniority; competition and bargaining occur between strata, and the pressure of the peer group on the individual is inescapable, at least in all matters that are subject to bargaining. Deviant impulses will be severely sanctioned, and the discipline imposed by the peer group will be one of the main forces, apart from the rules, which regulate behavior. Since supervisors may not interfere, and since there cannot be much sanction by comparison of individual work results, individuals must conform only to impersonal rules and to the group norms that support, interpret, and complete these. The importance of the peer group was marked in our two case studies, and especially in the Industrial Monopoly, by the remarkable concordance of answers among members of the same group for all relevant matters, and also by the discrepancy between private opinion, which could be deviant, and publicly expressed opinion, which had to follow the official line.22
This pressure of the peer group is one of the most relevant factors for understanding the bureaucrats' esprit de corps and ritualism, .The displacement of goals that is basic to them could not take place if it were not enforced by the peer group as a way of protecting itself against other groups and against the organization. Task impersonality and petty regulation are, in fact, very well developed in many private large-scale modern organizations, without important effects on ritualism. The same forces have a direct impact in a bureaucratic system of organization because the isolation of each stratum allows it to control its own domain and to ignore the organization's wider goals. We should further like to argue that, in order to get the best bargain for its own members, ^
gafHaToBjective is an end in itself. The members' ritualism provides good means to achieve such an end. It enables the group to assert its own differences and uniqueness, to pretend its own functions are the most crucial for the success of the whole organization. Then, finally, it helps develop and reinforce group solidarity among the group's own members.
There is a great difference of attitude, at the same level of task prescription, between an employee controlled by the kind of bureaucratic system of organization we are elaborating here and an em-
13 Sec chapter 4, p. 104.
19 2 Bureaucracy as an Organizational System
ployee of another, less bureaucratic system of organization. Ritualism is an asset for the former, since the possibilities for betterment depend on the success of his group and on his status within his group. It is a liability for the latter employee, whose chances depend on his ability to join other groups and to show that he is able to understand larger goals.23
the development of parallel power relationships
We devoted space in the preceding chapter to the problem of parallel power relationships; we shall summarize our conclusions here. Since it is impossible, whatever the effort, to eliminate all sources of uncertainty "within an organization by multiplying impersonal rules and developing centralization, a few areas of uncertainty will remain. Around these areas, parallel power relationships wilLdevelop, with the concomitant phenomena of dependence and conflict. Individuals or groups who control a source of uncertainty, in a system of action where nearly everything is predictable, have at their disposal a significant amount of power over those whose situations are affected by this uncertainty. Moreover, their strategic position is all the stronger because sources of uncertainty,are.very few. Paradoxically, in a bureaucratic system of organization, parallel power increases in direct ratio to its rarity. In many French public agencies, for example, there are cases of low-ranking employees whose opinions may be decisive in important affairs, simply because of their fortuitous occupation of a strategic position in an otherwise over-regulated system. There are similar examples of groups that maintain exorbitant privileges in the face of over-all egalitarian custom.
Parallel power relationships can develop within the hierarchical line. However, as a rule—and in the Industrial Monopoly especially —they develop outside it. This has as a consequence a complete reorganization of the organization's human relations.
Strata of experts are..often..privileged-..gra view, inasmuch as their task cannot be prescribed and regulated in a detailed way. Decentralization of the type described by Selznick will be the consequence of their successful pressure. Such decentralization is not so inconsistent with administrative centralization as it may seem to be. It is within the general framework of the impersonal and centralized "bureaucratic system," and because of the rigidity of such a framework, that experts' privileges can be maintained and developed.
23 This is often, of course, only partially true, since in many instances the peer group in a non-bureaucratic organization can also control its members very strictly.
The Bureaucratic System of Organization 193
We have learned, in studying the Monopoly, that the position of the experts is much stronger in an organization where everything is controlled and regulated—i.e., as long as their own task cannot be rationalized. Paradoxically, the more narrowly the organization is regulated, the greater the independence of the experts.
the vicious circle
The chjficulties, the poor work, and the fiustrations that are the con-secnienc^joj^^ tend to develop new pressure*
inTavor of the climate of impersonality and of centralization that has produced them. In other words, the "bureaucratic system of organization" is primarily characterized by the existence of a series of relatively stable vicious ^
i2S§Iifc..The schemes suggested by Merton and Gouldner are good examples of such vicious circles. But it is possible to elaborate new ones and to integrate them within a still broader scheme.
We have already discussed the vicious circle of the displacement of goals. In our view, it could be explained not only by the rigidity of the human personality, which maintains itself within the mold to which it has been submitted, but by the isolation of competitive strata, which use the displacement of goals to assert their influence against one another. The dysfunctional consequences of displacement of goals—i.e., difficulties with the customers, poor communication with the environment, and unsatisfactory adjustment to it, difficulties in achieving the task, a lower productivity, etc.—cannot and will not lead to greater flexibility within the system. The only weapon that can be used by the people who must make decisions is a greater elaboration of the rules and further centralization. Also, individuals and groups who directly face these difficulties and poorer results at the field level do not apply pressure to obtain more autonomy. On the contrary, they attempt to use the dysfunctions to reinforce their position vis-a-vis the public and within the organization. Their struggle against centralization is not directed toward helping the organization to adapt better to the challenge of the environment, but rather toward safeguarding and developing the kind of rigidity that is protecting them.
The vicious circle of control and supervision that has been analyzed by Gouldner can also be extended. He argues that the proliferation of impersonal bureaucratic rules reduces the tensions created by too close supervision, while, at the same time, the frustrations and the poor performances that develop in an impersonal bureaucratic world reinforce the need for close supervision. One can go further
194 Bureaucracy as ati Organizational System
and argue that inasmuch as the bargaining power of the supervisor dwindles—the natural tendency in all bureaucratic systems of organization—the vicious circle goes further the foreman-worker level and tends to include the whole set pf^hierarch within^ the organization,.
The Clerical Agency provided the perfect example of such a generalized vicious circle of close supervision, impersonal rules, and centralization. The frustrations of the different groups, which cannot discuss the decisions that will affect them and must submit to the close supervision of their activities, build up so much that higher-ups do not feel solid enough to face the problem, and the whole process of decision-making tends to move one rung higher. If people who make decisions do not have to confront those who will be affected by these decisions, tensions are reduced; but frustrations go on, and so does the pressure for centralization. Of course, efforts to change the whole system, to open it up are possible; but such attempts would run counter to the general fear of dependence relationships that is a contingent cultural trait of great relevance for the understanding of the development of bureaucratic systems of organization. This fear, in turn, is fed upon and reinforced by the frustrations emerging from the parallel power relationships that are likely to arise in such regulated organizations. The existence of those privileged relationships is the indirect consequence, as we have seen, of the development of impersonality and centralization; it tends to generate a very powerful secondary drive for more centralization and impersonality.
By and large, the common underlying pattern of all the vicious circles that characterize bureaucratic systems is this: the rigidity of task definition, task arrangements, and the human relations network results in a lack of communication with the environment and a lack of communication among the groups. The resulting difficulties, instead of imposing a readjustment of the model, are utilized by individuals and groups for improving their position in the power struggle within the organization. Thus a new pressure is generated for impersonality and centralization, the only solution to the problem of personal privileges.
Such a scheme of interpretation is no longer founded on the passive reaction of the human factor, offering resistance to certain kinds of interference and manipulation. It is based on the recognition of the active tendency of the human agent to take advantage, in any circumstances, of all available means to further his own privileges.
The Bureaucratic System of Organization
'95
The Problem of Change in a Bureaucratic System of Organization
Our model of a bureaucratic system of organization as one unable to correct its own errors, whose dysfunctions, therefore, are part of its self-reinforcing equilibrium, meets certain criticisms made earlier of the human relations theories of bureaucracy. It is general and systematic; it emphasizes rational patterns instead of the logic of sentiments; and, finally, it provides us with a good operational definition. But it still falls short of making it possible to understand the development and the limits of bureaucratic phenomena, which it still presents to us in a static and descriptive way.
To go one step further, we must consider the problem of change for such a bureaucratic system. Any organization, whatever its functions, goals, and environment, has to face change from within and from without. And if it is first necessary to study the steady state of day-to-day operations to discover regular patterns of action, then the way the organization reacts to change and tries to control it raises questions basic to the understanding of the real meaning of the regular routines, of which functional analysis could make only the internal mechanism understood.
The first, and quite obvious, point to make here concerns rigidity. A system of organization whose main characteristic isjts„rigidity will not.adjust easily to.change and will tend to resist change as much as -P-Qssible. CjQnsUnt-.trarisiQrrrjiatiQns' affect a modern organization. They concern the services it provides the customers and the public with whom it has to deal, the techniques of performance, and even the attitudes and capacities of the personnel it employs. Adjustment to these transformations can be gradual and more or less constant, if the agents of the organization who are at the level where the necessity of these changes is more obvious can introduce the wanted innovations or obtain such innovations from the competent authorities. But, as we have already pointed out, a bureaucratic organization does not allow for such initiative at the lower echelons;24 decisions must be madej^herejower isjp_cated, i.e., on top. However, this concentration and the concomitant strata isolation make it impossible to expect a permanent adjustment policy from the higher echelons. For one thing, they will be spared any advance warnings and predictions, because of the failures of the communication system. Then, when 21 Especially as regards human problems.
196 Bureaucracy as an Organizational System
they do learn of it, they have difficulties making decisions, because of the weight of the impersonal rules that may be affected by change. One may validly argue that, as a result, a bureaucratic system will resist change as long as it can; it will move only when serious dysfunctions develop and no other alternatives remain.
However, while a bureaucratic system of organization resists change longer than a non-bureaucratic system, it must nevertheless resort to change quite often in our adaptive modern society. Rigidity can obtain only within certain limits, and dysfunctions will reinforce the bureaucratic vicious circle only within a certain margin. Resist-Ji?c§.,j5.,£hange, therefore, is only one part of the picture. Another important part—one that there is a tendericj„t.Q_£QIgei:—is the very peculiar way a bureaucratic organization adjusts to change.
From the above analysis, it emerges that change in a bureaucratic organization must come from the top_,down and must be uniyersal-islic, i.e., encompass the whole organization en hloc. Change will not come gradually on a piecemeal basis. It will warFuntil a serious question pertaining to an important dysfunction can be raised. Then it will be argued about and decided upon at the higher level and applied to the whole organization, even to the areas where dysfunctions are not seriously felt. Only in this way can the impersonality system be safeguarded. One may even contend that very often change will lead to further centralization, by providing a way to get rid of local privileges that have developed around the rules.
Because of the necessarily long delays, because of_the_amplitu.de of the™s755pe it must attain, arid because of the resistance it must overcome, changeinbufeaucfaric organizations is a deeply felt crisis. The essential rhythm prevalent in sucTiorganizatioriws, therefore, an alternation of long periods of stability with very short periods of crisis and change. Most analyses of the bureaucratic phenomenon refer only to the periods of routine, and this is the image that emerges from our description of the bureaucratic vicious circles. But it is a partial image. Crisis is a distinctive and neee&aty.jdement of th^Jbjureaucra,tic system. It provides the only means of making the necessary adjustments, and it therefore plays a role in enabling the organization to develop and, indirectly, for centralization and impersonality to grow.
Crises are important in another way. They exemplify other patterns of action, other types of group relationships—temporary, but of decisive importance. During, crises, individual initiative prevails and people eventually come to depend on some strategic individual's arbitrary whim. Forgotten, strained dependence relationships reappear. Personal authority at times supersedes the rules. Such excep-
\
^ The Bureaucratic System of Organization 197
tipns, made possible by the loopholes of the routine setup, will be tqlerated because they can last only as long as the problem to be solved remains unsolved. They are the short periods of a war of movement necessary to permit more rational realignment of the war of positions. Their role, however, remains important. First, they perpetuate the fear of direct authority and arbitrariness among the members of the organization. Second, they continually create a new demand for authoritarian reformer J_gures_in the midst of the bureaucratic routine. Such figures are just the opposite of Merton's "bureaucratic personalities," but they are also characteristic and play an indirect but powerful role in shaping the values and counter-values of the organization.
Crises, finally, can develop at different levels. In the Clerical Agency and in the Industrial Monopoly, we analyzed minor but recurrent crises that played a significant role in the strategy of the different groups. In the Industrial Monopoly, possible plant reorganizations were an important element in the strategy of the different groups. In the Clerical Agency, the periodic overwork crises that stemmed from the discrepancy between the routine, stable allocation of work and the periodic traffic jams caused by seasonal influences provided an example. These crises perpetuated older authority patterns and increased a concomitant deep-seated distrust of face-to-face relationships. However, other, more remote crises still had weight'— the crisis over seniority in the Monopoly; the introduction of the accounting machines and the personnel reorganization in the Clerical Agency after the Liberation. Crises can come from within or they can be accidents imposed from without. Wars and social and political crises that upset the customary power equilibrium provide very good opportunities for effecting changes that have long been overdue.
More sophisticated and older bureaucracies have made special efforts to deal with the problem of change on a more rational basis. The older departments of French public administration, for example, have tried to domesticate and control the difficult and dangerous roles of change agents in order to regulate, and eventually eliminate, crisis. To achieve this, they have tended to create separate castes of higher civil servants. These are isolated from the bulk of the Civil Service by their recruitment, training, and career expectations, and are therefore protected from possible pressures from within. These Grands Corps provide, whenever needed, the personalities capable of imposing the necessary reforms on the administrative units that need them. At the same time they maintain as much as possible the rules of these units and minimize the authoritarian
198 Bureaucracy as an Organizational System
aspects of their own role by their impartiality and the prestige they enjoy because of their elite situation.25 However, such a way of dealing with the problem of change is cumbersome; and it may be argued that it provides a decisive element of instability at the higher levels of the political system.26 ;]
Such patterns, of course, are peculiar to French culture. But it is significant that cultural differences are most striking in the problem of dealing with change. We shall discuss the problem more thoroughly when we try to analyze a bureaucratic system from a cultural point of view. For the moment, it is sufficient to conclude that a bureaucratic system of organization is not only a system that does not correct its behavior in view of its errors; it is also too rigid to adjust without crisis to the transformations that the accelerated evolution of industrial society makes more and more imperative.
The Bureaucratic Personality
J Traditionally, the bureaucratic personality has been viewed as developing around ritualistic behavior. This especially is implied in Merton's call for a study oFthe1 relationship between bureaucratic structure and personality.27 In Merton's categories, ritualism is char-acterizexLhy ;.e,mp.hjs]^
Qt.rejectinglhe.m§X:alI^ds.-28 It is thus opposed to the three other possibilities of the famous paradigm—-conformity, retreatism, and rebellion—as if bureaucracy were the embodiment of one special category of social action, that of giving prominence to means over ends.
Such a view is highly suggestive and stimulating. However, it remains too one-sided, and, as we have pointed out, it overlooks the complexity of individual and group strategies in modern organization. Merton stresses only die impact of specialization and training;^ he cites the famous formula,^'People may 'Beliinfitted'^y'hieing "fitTo an unfit fitness." For the bureaucrat, "Training has become an incapacity." 29 This interpretation, of course, still remains partly true
s° On the role of the Grands Corps and their gradual transformation from a function of stabilization and unification of administrative action to a function of reform, little has been written as yet. See, on the Inspections des Finances, Francois Pietri, Le Financier (Paris: Hachette, 1931), a good literary account; and Philippe Lalumiere, L'inspection des Finances (Paris: P.U.F., 1959), a less good presentation of more recent facts.
20 Cf. below, pp. 255-56.
"Robert K. Merton, "Bureaucratic Structure and Personality," Social Forces, XVIII (1940), 560-68, republished in Social Theory and Social Structure (rev. ed.; Glencoe, 111.: The Free Press, 1957), p. zo6.
'~sIbid., p. 140.
w Ibid., p. 198.
^ The Bureaucratic System of Organization 199
for any large-scale organization. But it is much more suited to old-line bureaucracies, which relied on gross, inadequate drilling of their agents, and imposed lifelong commitments on them, without any prospect of change. In modern organizations, even bureaucratic ones, even the lowest group is able to devise its own strategy, and change, even if it is resisted stubbornly, remains a basic element in the game to be played. We have already argued that, in this perspective, ritugl-f ismis^o£.lhe-sin^ is also a very userul in-';
strjLiment in the struggle for Power and control an j' in the protection; of a group's,..area (faction. Ritualism in such a context, if one con-.' siders the actor's frame of reference and not the whole organization, must be viewed as conformity to what is expected from him and no longer as over-conformity. It is the rational response and not a "professional deformation."
At the same time, however, as it becomes less clear-cut and more diffuse, ritualism appears to form a lesser part of the general picture. Other possible alternatives of behavior emerge within a bureaucratic system of organization, according to the constraints of a wider variety of roles than is usually thought. In our analysis, for instance, numerous examples of retreatism and rebellion appeared; and, even though it is partly mythical, the role of innovator retains a great deal of importance. Finally, outside the paradigm, an important category of social action remains that of submissiveness to and identification with the power of other groups.
Retreatism at first seems to be a very central category of behavior in a bureaucratic system of organization. One may even argue that it is as decisive as ritualism. Confronted with an over-demanding situation, with no expectation of significant reward, individuals will choose to reduce their involvement and to commit themselves as little as possible to the organization. The pattern of impersonality and centralization brings great pressure in this direction. On the one hand, it deprives people of the possibility of personally influencing decision-making, and thus precludes any hope of recognition. On the other hand, it does not demand anything but formal compliance from individuals. People are not invited to participate, and, if they retreat, they risk little punishment. However, such a response is not so free as we may think, because of the decisive influence of the peer group, whose powers of coercion are much greater than in a less rigid organization. Our case studies show that groups allow full retreatism only inasmuch as they are weak and discouraged. Successful groups prohibit retreatism completely. Intermediary groups are characterized by a mixture of different attitudes and behavior.
The purest example of retreatism is provided by the lower supervisors in the Industrial Monopoly. Their involvement is extremely
200 Bureaucracy as an Organizational System
slight, they rarely hold strong opinions, and their answers are often not even coherent. They do not take pride in any part of their work. One would imagine that they would emphasize those bureaucratic functions that they can accomplish without any interference from the maintenance men, but they do not—at least, most of them do not. As a group, they are not ritualists. They might have been in earlier times, according to legend, when they still had some power. But they ceased to be so long ago, except for a handful of old-timers who have kept a number of ritualistic patterns within the general framework of retreatism. A few of them, however, as we have noted, try to fight back. Those "rebels" are much more dissastisfied than their colleagues, but they are the only ones really involved in the affairs of the Monopoly.
The supervisors of the Clerical Agency present some of the same traits, but in a very different mixture. They certainly reduce their commitment to the Agency, but they still expect promotions and know they are likely to be transferred to another agency. They are, therefore, generally quite involved in the affairs or the over-all department to which they belong, and they often act at least partially as rebels within the Agency. Trade unionism is, for many of them, a good expression of their hostility. As regards their work, they fight to preserve the status quo and to keep problems and entanglements to a minimum. From this point of view, they are, above all, retreat-ists. They have a few very "bureaucratic" duties to perform, but they do not emphasize them in a ritualistic way. A very small group of zealous over-conformist individuals are found on top, among those who can expect a higher promotion. But their emphasis is not on ritual but on the specific achievements measured within the Agency —productivity and quality per head, i.e., that which is considered a goal by management, Theirs is submissive, more than ritualistic behavior.
The directors of the Industrial Monopoly are also strongly retreat-ist, but they present a different combination. The "dignified elder statesman" we have described may be considered a ritualist. His emphasis constantly bears on the forms and on the decorum of his role. He tends to ignore the productive goals of the organization. When talking with the interviewers, one of them used to repeat this, for
ihim, characteristic joke: "We are here to write reports and process papers . . . [the service given to the public] is only a by-prod-
;uct."so But this ritualism appears much more as a way to preserve
"This joke, to be sure, is ambiguous; but we can ignore the reversal of roles that makes the organization responsible to focus only on the way this director seems to be fascinated by bureaucratic practices. The rest of the interview, as a matter of fact, shows what exaggerated importance he himself gives to written documents and to formalism. . ».
The Bureaucratic System of Organization 2.01
a fagade and to hide from oneself and others the futility of one's role, than as an occupational deformity. And only a minority of directors adopt such behavior. At the other extreme are the directors who exaggerate their helplessness; these are pure retreatists who try{ to preserve their self-respect as individuals by separating themselves \ from their job and deprecating it. They simply refuse to be involved J and choose to concentrate on possible achievements outside their job. Most directors present a combination of both patterns of behavior. This contradiction, as one can discern from the incoherent comments of a few individuals, is quite characteristic of the group as a whole. But one can understand it only if one refers to the role of innovator played by two or three of them.
Innovation, Merton's supplementary category, is not absent from a bureaucratic system. Curiously, the innovator seems to be the polar figure of the whole system and innovation the most envied achievement, the one for which people are most ready to fight. The innova- i tor in a bureaucratic system, however, shows certain special charac- ij. teristics. He is a legislator, a Solon type, rather than a discoverer. He £ is someone who will once again put everyone in his own place, who will reorder the world in a better way, rather than someone who will launch new patterns, new ways, of doing things.
People try to prepare for this role. The changing strategy of the members of the Monopoly's engineers' corps, as their age and seniority increase, can be viewed as dominated to quite an extent by their hopes and realistic expectations as regards this glorious role.
The role of innovator can interest, in fact, only a small minority of civil servants, since innovation appears only on the top and people must compete for it. But its importance reaches far beyond the immediate group it affects directly. In the Monopoly, success is more a myth than a reality; but its mere possibility arouses fears and resistances which would otherwise slowly disappear. Change and a fresh meting out of justice will return some day, like a long-forgotten specter, and one had better have one's defense prepared.
In other contexts, however, where the innovation-legislation function is given more importance, the whole strategy of the system depends on it and we have a completely different setup. One typical example is provided by the role of the prefect in French provincial administration.31 Among French bureaucrats, the prefect is probably the one whose innovation-legislation function is most developed. He has discretion and is supposed to use it. The staff around him show a characteristic pattern of adjustment. Intermediate officials are extremely cautious, timid, and submissive. They pretend that they
"The prefect is a sort of appointed governor in charge of all administrative affairs for an area of a size intermediate between a county and a state.
202 Bureaucracy as an Organizational System
have no responsibilities whatever in making decisions. They feel that they are merely anonymous helpers of the great innovator figure with whom they identify. Petty officials, on the other hand, appear to be extremely ritualistic and, at the same time, rebellious. They are attached to the status quo and resent possible innovations as so many violations of the order which they must impose on the public. They feel themselves to be betrayed, and their position undermined, by the prefect's initiative. Their strategy is a strategy of opposition and rebellion; they try to impose their ritualism on the prefect and to obtain some compensation for his trespassing on their jobs. We see here the importance of the gap created by centralization. Petty officials cannot make the necessary adjustments. Power to innovate is reserved for superior figures with prestige. As a result, petty officials behave as extremely jealous ritualists for all practical purposes, and try to use to the utmost advantage the parcel of power involved in the rites imposed on them. At the same time, they question the whole system and pose as rebels. Sentiments are more complex than one would expect from reading Merton. There is a sort of paradox in this respect: when a petty official obtains promotion to a middle rank that permits him to escape the chicanery of petty regulations, he foregoes this theoretical rebellion and becomes humble and submissive.
I should like to argue that the freedom and discretion of the innovator figure require the strict ritualism of the petty officials and the submissiveness of the middle officials if they are to develop fully. Petty officials are given the unpleasant role of enforcing the rules without the slightest leeway; they must refrain from anything personal for fear of arbitrariness. This is a protection for the public and, although indirectly, for the official as well. But the dysfunctions thus created make it necessary to resort to an innovator. This figure can make the wanted adjustments, and his prestige will be all the greater, since he is entrusted with the only discretionary power within the system. Intermediate officials will act as transmission links. They are too involved in the discussion of arbitrary exceptions to the rules and in the resetting of the rules to adopt the ritualist and rebellious attitudes of their subordinates, but too helpless themselves to assume responsibilities. Power of discretion, finally, is so exceptional that it must be far removed from face-to-face relationships and surrounded by an aura of awe and submissiveness.
|   We now confront new images of the bureaucrat, the self-satisfied jglorious innovator, the submissive assistant, and the rebellious and Iritualist subaltern. This set of roles corresponds to a regular hierarchical arrangement within a very active and powerful agency. Significantly, there is little retreatism in such an arrangement. *R_etreatism
The Bureaucratic System of Organization 203
seems associated with an increase of rigidity and the disappearance of the innovator function. Whenever change and power are present, we find more and greater involvement and deeper feeling.
Finally, the last roles still presenting a problem among those we have studied are those of the maintenance man and of the technical engineer in the Industrial Monopoly. The people who assume these roles are very much committed to their jobs. Still, they áre precisely the opposite of the innovator. In a way, they are staunch conservatives—they guard the status quo to protect their own privileges and are natural enemies of the potential innovators. But maintenance men and technical engineers are not ritualists by any ordinary standards. Their privileges do not stem from the rules but, on the contrary, from the impossibility of introducing a rule. They are ritualists only as regards the system as a whole, giving priority to the technical and human relations arrangements over the functional goals. For the rest, they seem more practical, more instrumental, more goal-minded than anyone else. They exemplify still another dimension of the bureaucratic personality: practical conservatism. The other pole in this respect might be bureaucratic idealism—reliance on the power of the rules to make change. This last attitude is much more diffuse and it is often associated with retreatism. Lucidity and the emphasis on intellectual understanding are the usual compensation of the re-treatist idealist. They constitute another very important dimension of the bureaucratic personality. It was especially prominent among directors and supervisors of the Industrial Monopoly and among higher and lower supervisors of the Clerical Agency. This polarity is a good counterpart of the submissiveness-rebellion couple we have seen surrounding power and innovation. A bureaucratic system will revolve around one of these axes, inclining completely in the direction of retreatism-conservatism in the sectors and periods of routine, and admitting a large dose of power, and therefore of submissiveness-rebellion, in the sectors and periods of crisis.
The Individual's Stake in Bureaucracy
We have now described the typical processes of a bureaucratic system of organization and analyzed its ways of adjusting to change. In \ so doing, we have shown how its gross inadequacies and "dysfunc-j tionality" follow from its rigidity and from its tendency to escape! from reality. At the same time, we have suggested that these traits i are not only unintended consequences but also necessary elements of a system whose rational aim is to obtain a minimum of conformity from the members of the organization. This bureaucratic phenome-
204 Bureaucracy as an Organizational System
non has thus appeared to be a function of the equilibrium between the kind of social control used to maintain the organization as an ongoing system and the reactions of the human group that is submitted to it. Both depend on the state of the larger society's cultural norms and on the technical possibilities, of diminishing the uncertainty of social action. ^
But it can be argued—and repeatedly has been—that, whatever the requirements of organization, individuals suffer from the kind of social control always imposed by social action, and that co-operative forms of action are possible and would appear more rewarding than our present power-ridden forms. We should ask, therefore, why men still consistently choose conflicting types of games instead of more co-operative ones. Why is it that they make it necessary to resort to bureaucratic rigidity, imposing on themselves the roles of "ritualists", or "retreatists"? We have seen the problem from the organization's point of view; we must now consider it from that of the individual. We must ascertain the individual's stakes in bureaucracy.
Co-operation means participation in decision-making. The possibility of more co-operative forms of action depends on the attitude, ofc-pmplft toward.auch„^ Human relations theoreticians,
especially those of the Lewinian school, have long assumed that people want to participate, and are ready to participate, under all and any conditions.32 They have always argued as if the reasons for lack of participation were to be found only in the contingencies of organized activities and in the misgivings and apprehensions of manage-: ments—if only superiors could be converted to more permissive leadership, subordinates would be glad to participate. We certainly do not defend the contrary view: progress can always be made in the area of participation. However, the limits are narrower than one usually thinks they are, and it is necessary, if one wants to determine them to inquire into the actors' motivations.
It is a partial view indeed which expects people to be always eager for participation. People are very ambivalent toward participation. It is difficult to have a clear and welPargued position in an area which is only beginning to be explored. We shall merely present a few remarks that are suggested by these new developments. On the-pne hand, people would like very much to participate in order to control their own environment. On the other hand, they fear that if and when they participate, their own behavior will be controlled by their £Pjarticipants. It is far easier to preserve one's independence and integrity if one does not participate in decision-making. By refusing to be involved in policy determination, one remains much
m See, for example, Erich Fromm, Escape from Freedom (New York: Rinehart, 1940- *     ' ■
The Bureaucratic System of Organization 205
more free from outside pressure. Recent research has shown that members ojan organization are not always enthusiastic when invited
to parJiapate.....more.,, mJts,.^acjripm'pg. The relative ambivalence of
subordinates' attitudes toward participation emerges from the results of scientific controls on theconsequences of ■experiences m_ Human. relatpjjgjKuiangJ? as well as from the results of the ambitious program of decentralization tried in a large insurance company by the Survey Research Center of the Unversity of Michigan.34 Chris Argyris, among other researchers, has shown that, within the framework of the usual psychological contract into which individuals have entered with their organizations, every change that implies greater participation is viewed unfavorably.35 Arnold Tannenbaum, in a series of studies on the problems of control within different kindifof organizations, voluntary as well as business, has found most challenging contradictions to the usual hypotheses.38
If one accepts participation, one is bound to co-operate, i.e., to bear one's coparticipants' pressure whether they are one's equals or one's superiors. People therefore rarely agree to participate without some substantial counterpart. They try to negotiate about their par- / ticipation, and giyejt^nlyifthey feel that they will be adequately rewarded.. This is true even at the lowest level," ^Kere','particip^fion means only being committed to one's own job's goals; it is very markedly so when participation concerns leadership problems of allocating resources and delineating duties. Retreatism can be a very rational form of behavior whenever the individual concerned has good reason to believe that the rewards he is offered are not commensurate with his efforts, and feels that there is a good chance that he will be manipulated. The will to participate, finally, depends to a large extent on the degree of trust and openness in interp^
** See, for example, E. A. Fleishman, E. F. Harris, and H. E. Buret, Leadership and Supervision in Industry: An Evaluation of a Supervisory Training Program (Columbus: Ohio State University, 1955); Floyd Mann, "Studying and Creating Change: A Means to Understanding Social Organization," Human Relations in the Industrial Setting (New York: Harper, 1957); Rensis Likert, New Patterns of Management (New York, McGraw-Hill, 1961).
"The famous experiment at Prudential has shown that the "democratic" program of increasing the employees' possibilities of participation has entailed serious problems of interpersonal relations at the group level. See Nancy Morse and Everett Reimer, Experimental Change of a Major Organizational Variable," Journal of Abnormal and Social Psychology, LII (1955), 120-29.
* Unpublished data we have gathered in a systematic comparison of forty groups in six insurance companies show concomitantly that people working in self-administered groups are afflicted by interpersonal difficulties.
M Arnold Tannenbaum and B. S. Georgopoulos, "The Distribution of Control in Formal Organizations," Social Forces, XXXVI (1957), 44-50; Arnold Tannenbaum and Robert L. Kahn, in Dorwin Cartwright (ed.), Participation in Local Unions (Evanston: Row, Peterson, 1958).
2o6
Bureaucracy as an Organizational System
characteristic of the cultural norms to which people adhere.37 In a society where proving one's own independence is considered to be a value in itself, retreatism is the most satisfactory mode of adjustment as long as the proposed participation does not adequately insure full rights of control. ^
Another basic cultural datum, the norms of attitudes toward authority relationships, is still of much importance. The possibilities of participation must, in actuality, remain limited. Participation is likely to be viewed as only partially satisfactory, since some constraint must be imposed from the top down to meet the remaining uncertainty within and without. If the inescapable authoritv relationships are not accepted easily, pressure for centralization will be strong, and a certain lypje^^rigidity, comparable lo~trie vicious circles we have analyzed in our two case studies, will tend to develop within all organizations. But other possible types of rigidity can develop. For example, if authority relationships are accepted submissively, with passive resistance as the outcome, one can hypothesize vicious circles of successive controls.38
In any case, the system's general equilibrium relies on the bargains which individuals can make. These bargains depend on individual expectations and values, on the one side, and on the requirements of organized activities, on the other. These requirements are themselves determined by the technical means employed by man to control his universe and by the same cultural data that shape individual reactions.
" j  Within this general framework, bureaucratic systems of organization such as the ones we have studied offer individuals a fairly good
/combination of indejgndence, and jecurity. The modern observer is especially struck by their dysfunctional aspect: he emphasizes the heavy price that people must pay for such results. But one should not forget that, if one takes due account of people's values and expectations and of the boundaries of social action, the bargain is usually quite a good one.
Rules protect the people who submit to them. Within the area delimited by the rules, they are free to make their own contribution according to their arbitrary whim. They can participate or take refuge in retreatist behavior, commit themselves to the organization's aims, or reserve their moral forces for some personal endeavor. Of course, they have very little chance of making a success of their activities within the organization. They will not obtain distinction among
"It depends also of course on their position in the hierarchical order and on the model of social relationships between social strata characteristic of their culture. " This is the case in Russian and Soviet bureaucracy. See below, pp. 227-31 •
The Bureaucratic System of Organization 207
their colleagues, but, on the other hand, they do not have to fear failure or face hostile reactions from competitors.
We shall go even further and argue that a bureaucratic system of organization always relies on a certain amount of compulsory participation which appears to be, under the present conditions, more gratifying for the individual than the voluntary participation for which— as is, perhaps, too readily believed—he is fighting. We do think that, even in those cases in which we have diagnosed retreatism, ritualism, and rebellion, there is much commitment and participation that should not be neglected. Their importance can be appreciated when one compares the attitudes and performances of members of such "bureaucratic" organizations in Western industrial societies, with those of any kind of modern organizations in underdeveloped societies, where people are unable to give the necessary attention and commitment to their task, whatever their apparent willingness.39
Compulsory participation of that sort implies no avowed responsibility.40 It is clandestine, and people do not feel that they may become involved. Individuals who participate in this way remain completely free vis-á-vis the organization. Such an arrangement makes it possible to solve contradictions that would otherwise remain insoluble. On the one hand, members of a modern organization must participate, and they know that the organization cannot get along without their participation. On the other hand, they know that the organization cannot reciprocate in the way that would be requisite if they were to commit themselves to it seriously and thus renounce part of their freedom. If the organization imposes on them, through its official system] of rules, a kind of compulsory participation without responsibility, they will be able to achieve two contradictory aims at the same time. They can give meaning to their work by participating in the common enterprise, and they can safeguard their independence in a situation where an avowed responsible commitment would mean a risk of alienation. Furthermore, such retreatism puts the organization in an unfavorable bargaining position; it obliges the organization to remain on the soliciting side, to beg for support.
The latent fa^ understood in
this perspective; as(.primarily a protective function. It provides the
" As we have already noted, the grěve du zěle offers a good means for subordinates to remind management that they cannot dispense with the former's intelligent participation.
"We do not distinguish here between participation in decision-making and participation in the application of decisions. For we are interested in what is common to both kinds of participation, i.e., the feeling of responsibility and commitment the actors may have, whatever their place in the hierarchy.
the
peo-
208 Bureaucracy as an Organizational System
individual with the minimuin.of security necessary.to him for deal-ing with his fellows in the pursuit of the neces^.ary co-operative activities. .This.,.kirtd. of security is valuable, especially in a world where the individual feels he is extremely vulnerable. One can anticipate, perhaps, that, as our industrialized societies become more flexible and can jjive all^their members due pjotectior^ against failure, lure^bflthei p1cv6lected'status of trie bureaucrat will mmifusn^nd pie will be progressively less ready to pjaythe price_of rigidity.
Finally, the bureaucratic model we have analyzed seemingly has still another limited but important function. It may be viewed as a very good way of maintaining some of the individualist values of a pre-industrial world within our modern societies. One may argue that its development in France is associated with the resistance of older ways of life which was brought to a special perfection before the Industrial Revolution. Resistance to participation, and preference for centralized authority and the stability and rigidity of a bureaucratic system of organization, by preserving for each member a minimum of autonomy and individual discretion, proceed from the same values which peasants, craftsmen, and noblemen embodied in the delicate balance of human relations that characterized the art de vivre of traditional France. In a certain way, a bureaucratic system of organization provides a combination of .the values 6Fa traditional ascriptive. society ImcTthose. of .a. modern achievement-oriented society.41 People can compete for any position; no formal barriers prevent them from doing so. But competition has been institutionalized and separated from the daily life of the work environment; and its formalism has, at least partially, the same prptectiye^yalue. as the older ascriptive rules.
Such arrangements, however, present substantial drawbacks for the individual. TJbe,Jiuj_e^^ People are protected, but at the price of being partially cut off from reality. They have security and are protected from the sanction of facts; but they have no way of taking the measure of their own endeavor. This engenders a secondary kind of anxiety and explains the paramount importance of human relations within a bureaucratic system. What people gain in security they lose in realism. They must rely on human relations sanctions instead of on~trie~usual achievement sanctions. Theirs is a world of petty bickering and the endless battles of the war of position. People escape the lower-middle-class status panic; instead, they develop the skimpy outlook of the petty power struggles, of a tight social system.
" For a new discussion of the Parsonian distinction between ascriptive values and achievement values and its relevance for national characteristics, see S. M. Lipset, "Democracy and the Social System," in The First New Nation (New York: Basic Books, 1963). * ■
BUREAUCRACY AS A CULTURAL PHENOMENON
THE FRENCH CASE