Interest Groups and Political Development in Turkey [Course Book ed.] 9781400853311

Citation preview

ROBERT BIANCHI

Interest Groups and Political Development in Turkey

PRINCETON UNIVERSITY PRESS PRINCETON, NEW JERSEY

Copyright © 1984 by Princeton University Press Published by Princeton University Press, 41 William Street, Princeton, New Jersey 08540 In the United Kingdom: Princeton University Press, Guildford, Surrey AU Rights Reserved Library of Congress Cataloging in Publication Data will be found on the last printed page of this book This book has been composed in Linotron Times Roman Clothbound editions of Princeton University Press books are printed on acid-free paper, and binding materials are chosen for strength and durability. Paperbacks, although satisfactory for personal collections, are not usually suitable for library rebinding. Printed in the United States of America by Princeton University Press, Princeton, New Jersey

To my father,

Edward Bianchi, and to the memory of my mother,

Julia Bianchi, with love and gratitude

Preface

the aim of this study is to use certain social science concepts to learn about politics in Turkey and then, in turn, to use evidence from Turkey and other systems to improve the concepts themselves. I have attempted to show how ideal types of pluralism and corporatism can sharpen our understanding of Turkish political change. But I also have sought to bring out the more general implications of Turkey's unusual and persistently heterogeneous associational life in order to contribute to ongoing efforts to refashion and refine conventional dichotomous distinctions between pluralism and corporatism and between "state" and "societal" corporatism. In this sense, the study reflects a widespread concern that theory and research inform one another closely in a rapidly expanding field of comparative political inquiry. In all phases of this project I have been fortunate to enjoy the encouragement of many friends and colleagues in the United States and Turkey. Leonard Binder, Philippe Schmitter, and Aristide Zolberg have provided indispensable support and guidance since my earliest days at the University of Chicago. The late Lloyd Fallers and his wife, Margaret, were two of the earliest supporters of this study, and their personal example taught me much about how to conduct intelligent and sensitive fieldwork in another cultural setting. Frank Tachau and Frederick Frey each had the kindness and stamina to read two different versions of the manuscript. Together they provided an invaluable combination of praise and challenging criticism that allowed me to tighten and clarify some of the major arguments. I would also like to thank Kemal Karpat, Dankwart Rustow, and Walter Weiker for encouraging me to develop viewpoints with which they sometimes disagreed. Several Turkish friends in Ankara and Istanbul aided my fieldwork and influenced my thinking long afterward. The contributions to this study by such scholars as Mubeccel Kiray, Ahmet Yiicekok, Yal9in Ku(JUk, Cigdem Kagitciba^i, Bahattin Aksit, and Ayse Oncii will be obvious to all who are familiar with their writings. The most important assistance came from dozens of leaders of the Turkish labor movement and business world who gave generously and enthusiastically their time and enormous insight into Turkish political life. Their uncommon patience, candor, and cordiality IN ITS SIMPLEST TERMS,

viii — Preface made this work possible and exciting. I must emphasize, however, that the responsibility for the views expressed in this volume rests exclusively with the author. My fieldwork in Turkey was supported by the Office of Education through a Fulbright-Hays Research Abroad Grant. The American Research Institute in Turkey, and especially Prentiss de Jesus, exposed me to a wide variety of foreign scholars in Turkey and provided many of us with a home away from home. The Center for Middle Eastern Studies at the University of Chicago financed part of the manuscript preparation and the Department of Political Science was particularly generous with recurrent requests for "more computer time." Both in Turkey and the United States several friends lent great personal support while struggling with their own projects. I would particularly like to thank David Barchard, John Taylor, Goksel Turk, and Dennis and Lilina Williams. Finally, I wish to acknowledge the assistance of Helen K. Bailey, who typed several versions of the manuscript; of Rudy Banovich, who prepared the illustrations; of Sandy Thatcher, who guided me through the long journey of first authorship; and of my wife, Layla, who kept assuring me that eventually this would indeed become a book.

Contents

Preface

vii

List of Tables

xi

List of Figures

xiv

1. Interest Groups, Political Participation, and Political Development PART I. The Bases of Interest Group Politics 2. Structural Differentiation and Uneven Development: Socioeconomic Bases 3. Political Culture: Attitudinal Bases 4. Public Policy Toward Associations: Formal-Legal Bases PART II. The Network of Interest Group Politics 5. The Emergence and Diffusion of Interest Groups: The Art of Association as a Dimension of Modernization 6. Strategies of Interaction and Influence 7. The Political Subculture of Turkish Interest Group Leaders 8. Turkish Interest Groups as Channels of Political Participation or Agents of Social Control? 9. Turkey in Comparative Perspective

3

35 73 108

149 202 282 338 355

Appendixes A. Regional Distribution of Three Dimensions of Associability, 1965

406

B. English Translation of Survey Questionnaire

409

Selected Bibliography

413

Index

421

List of Tables

2-1.

Economically Active Population by Major Occupational Groups, 1927-75

37

The Growth of the Urban Population and Its Distribution among Types of Settlements, 1935-75

40

Distribution and Growth Rates of Gross Domestic Product by Economic Sector, 1950-75

44

Indicators of Savings, Investment, and Capital Intensiveness, 1950-72

45

2-5.

The Diversification of Manufacturing, 1955-70

47

2-6.

The Structure of Manufacturing in State and Private Sectors, 1955-71 Concentration and Productivity in Manufacturing, 195571—Private Sector

2-2. 2-3. 2-4.

2-7. 2-8. 2-9. 2-10. 2-11.

48 50

Concentration and Productivity in Manufacturing, 1955— 71—State Sector

52

The Regional Distribution of Credits from the Turkish Banking System, 1971

60

The Relative Economic Importance of Istanbul and the East Marmara Subregion

62

Distribution of Personal Income and Landownership, 1963-73

63

5-1.

The Diversification of Associational Life, 1946-70

159

5-2.

Phasing and Slopes of Associational Growth Curves

163

5-3.

Parameters of Indexes of Modernization and Associability, 1945-70 The Changing Social Bases of Associability (Correlation Matrix)

169

The Changing Social Bases of Associability (Beta Coefficients from Multiple Regressions)

172

5-4. 5-5.

167

xii — List of Tables 5-6.

Overassociability and Underassociability by Region (Total Associations)

179

Overassociability and Underassociability by Region (Class Associations)

181

Overassociability and Underassociability by Region (Community Associations)

182

Changing Dimensions of Associability, 1946-65

188

5-10.

Correlates of Major Party Vote, 1950-73

194

5-11.

Associability and Major Party Vote, 1950-73

196

5-7. 5-8. 5-9.

6-1.

Density of Union Organization and Distribution of Union Membership by Economic Sector, 1955-71 Distribution of Membership in Rival Union Organizations by Economic Sector, 1975

231

7-1.

Biographical Characteristics of Survey Respondents

286

7-2.

Indicators of Traditionalism

290

7-3.

Indicators of Religiosity

292

7-4.

Indicators of Authoritarianism

293

7-5.

Indicators of Nationalism

295

7-6.

Indicators of Paternalism

298

7-7.

Indicators of Articulation

300

7-8.

Indicators of Bargaining

303

7-9.

Indicators of Dependence

307

7-10.

Indicators of Efficacy

308

7-11.

Mean Scores on Attitudinal Indexes

310

7-12.

Intercorrelations between Attitudinal Indexes

312

7-13.

Factor 1 ("Traditionalism")

314

7-14.

Factor 2 ("Activism")

315

7-15.

Factor 3 ("Accommodativeness")

316

7-16.

Mean Factor Scores for "Traditionalism," "Activism,"

6-2.

229

and "Accommodativeness"

318

7-17.

Indicators of Corporatism and Pluralism

319

7-18.

Intercorrelations between Corporatist and Pluralist Items

326

7-19.

Ideological Factors

329

7-20.

Mean Factor Scores for Ideological Factors

330

List of Tables — xiii 7-21. 7-22.

Intercorrelations between Ideological and Attitudinal Variables (Individual Items)

333

Intercorrelations between Ideological and Attitudinal Variables (Indexes and Factors)

335

List of Figures

1-1. 5-1. 5-2. 5-3. 5-4. 5-5. 5-6. 5-7. 5-8. 5-9. 5-10. 5-11. 5-12. 5-13. 7-1. 7-2. 7-3. 7-4.

A Framework for the Analysis of Group Development Associational Growth and Socioeconomic Change, 194570 Growth Trends in Association Formation, 1946-68 Annual Number of Community Associations, 1946-68 Annual Number of Class Associations, 1946-68 Annual Percent of Increase in Number of Community Associations, 1947-68 Annual Percent of Increase in Number of Class Associations, 1947-68 Correlates of Associability, 1945-70 (Total Associations) Correlates of Associability, 1945-70 (Community Associations) Correlates of Associability, 1945-70 (Class Associations) Associability and Development, 1965 Regional Distribution of Three Dimensions of Associability, 1965 Associability and Justice Party Vote, 1969 Modernization, Associability, and Major Party Vote Relationships between Biographical Characteristics, Attitudes, and Ideology Intercorrelations between Biographical and Attitudinal Variables Three Views of Interest Group Politics in Contemporary Turkey Intercorrelations between Biographical, Attitudinal, and Ideological Variables

29 157 158 161 161 165 165 171 175 176 178 189 200 201 283 313 327 337

Interest Groups and Political Development in Turkey

CHAPTER 1

Interest Groups, Political Participation, and Political Development

SINCE THE END of World War II, the rapid proliferation of a wide variety of interest groups along with their increasing interaction with one another, with political party organizations, and with various governmental institutions has added an important new dimension to the politics of contemporary Turkey. The rapid emergence and diffusion of this network of groups representing specialized interests suggest that the Turks have been particularly precocious in developing "the art of association" while implementing broad social and economic change within the context of liberal democracy. This confirms the global hypothesis of most students of associational life since de Tocqueville and Durkheim that increasing group organization and activity are the result of the growing division of social labor and the expansion of formal political equality.1 But one of the most interesting aspects of the development of associability in modern Turkey is that it has so often displayed sharp discontinuities with the general level of socioeconomic development and structural differentiation. In different historical periods, social and economic sectors, and geographical regions, association formation commonly has either lagged behind or surged ahead of other aspects of social change. In other words, one of the most distinctive and striking characteristics of the art of association in Turkey is that historically it has been an independent and highly politicized dimension of modernization.

The highly politicized nature of Turkish associational life is observable not only in the close interconnections between government, party, and group leaders but also in the very different ways in which their relationships have been structured in order to advance or inhibit the representation of different socioeconomic and political interests. The complex and variable nature of interest group politics in Turkey is largely the result of the division of the Turkish associational universe into two different and often overlapping networks—the pluralist network of private voluntary associations and the corporatist network of semiofficial compulsory associations. In certain 1 Alexis de Tocqueville, Democracy in America, 2 vols. (New York: Vintage Books, 1954), 2:114-118; Emile Durkheim, The Division of Labor in Society (New York: The Free Press, 1964), pp. 1-32.

4 — Political Development historical periods and in certain social and economic sectors, pluralist interest groups have been predominant. These have tended to broaden the scope and increase the intensity of interest conflict, to serve as important channels for expanding political participation, and to contribute to the reshaping and invigoration of party-electoral politics. But in other periods and sectors corporatist interest groups have been predominant. These have tended to suppress interest conflict, to constrict political participation, and to serve as instruments of social control. One of the major objectives of this study is to describe the emergence of both corporatist and pluralist associations as competing means for structuring interest representation and to examine the tensions in Turkish political development that have been produced by historical shifts in the relative importance of these two types of groups. Particular emphasis will be given to analyzing the background and evaluating the consequences of associational policies pursued by recent governments in their attempts to implement central economic planning since the 1960 revolution. These policies have aimed at corporatizing the most important areas of Turkish associational life in order to limit political participation and reduce demands for economic redistribution without abandoning the formal framework of liberal democracy. Regardless of the stated or implicit intentions of these policies, however, they have contributed to a heightened perception of interest conflict between and within several important social and economic sectors and to a rising militancy among several groups in both corporatist and pluralist associational networks. This means that the principal, though unintended, consequence of recent attempts at corporatization has been the creation of simultaneous crises of political participation and distribution that seriously threaten Turkey's attempt to combine ambitious new plans for rapid economic development with the continuation of a democratic regime. Much of what will be presented in the following chapters will describe and attempt to explain two of the most striking aspects of the development of interest group politics in modern Turkey: first, the common discontinuities between the degree of socioeconomic development and collective action through various associations; and, second, the rising militancy of several long-established and well-organized interest groups, which has contributed to more intense social conflict and to a major realignment of the increasingly fragmented and polarized party sytem. Both of these important outcomes conflict with some key hypotheses and assumptions concerning the development of interest groups that have become prevalent among the major theorists of political development and group politics.2 2

This is not to suggest that one can find in the literature on political development or in

Political Development — 5 Most students of political development and most group theorists appear to agree on two global hypotheses concerning the emergence and development of group politics. The first is that the emergence of a network of specialized agents of interest representation is a basically spontaneous consequence of the growing division of labor in society and the expansion of citizenship. The number and variety of associations and the degree of group organization are generally assumed to be closely related to, if not basically determined by, the level of socioeconomic development. The second hypothesis is that as group organization becomes more highly structured and complex, interest groups tend to become politically less militant. Longestablished and well-organized groups are assumed to contribute to a more exclusionary and less conflictive political process whose outcomes favor established elites and reinforce existing social and economic inequalities. While theorists may approve or disapprove of the consequences, there seems to be remarkably wide agreement that as interest group politics becomes more organized and more structured it tends to limit political participation, narrow the scope of interest conflict, and reduce its intensity. The discussion of interest groups in Turkey will be preceded by a critical and comparative examination of some of the leading theorists of political development and group politics. This will identify and make more explicit the assumptions concerning the nature and consequences of interest group politics in societies at different levels of economic and political development. In particular it will describe different views of the ways in which interest groups structure political participation and thereby contribute to both the creation and possible resolution of different problems or "crises" in political development. Theorists of political development and group politics generally have not attempted to elaborate frameworks for the comparative analysis of group development that would explain variations in group organization and activity in different societies at similar levels of socioeconomic development or in individual societies over time. Instead they have provided us with what amounts to three different sets of middle-range hypotheses about how interest groups tend to shape and structure political participation in societies at low, intermediate, and advanced levels of economic development and social differentiation. The juxtaposition of these three synchronous views into a unilinear theory of group development has tended to serve as a poor the literature on group politics an explicit "middle-range, developmental theory of interest group politics." In fact, while most studies of political development are quite concerned with the general issue of political participation, they indicate little about interest groups as specific channels of participation. Similarly, most studies of interest groups in particular political systems tend to be "snapshot" analyses that provide little insight into the origins and development of groups or about their changing impact on political participation in general.

6 — Political Development

substitute for a comparative framework that would account for alternative developmental patterns. Collectively, most developmental and group theorists have provided the following unilinear theory of group development: In the early phases of modernization, the absence of well-organized groups and the unstructured nature of group interaction contribute to potentially excessive levels of interest conflict and political participation. In the intermediate phases of modernization, the emergence of a wide variety of newly organized groups provides a specialized set of channels for managing interest conflict and for expanding participation more gradually and moderately than mass party organizations. And in the advanced phases of modernization, a small number of highly organized interest groups acquire effective control over much of the decision-making process and tend to severely constrict meaningful participation and to suppress interest conflict. These three views of group politics are prevalent not only among. Western social scientists but also among some of the most important participants in contemporary group politics in Turkey. Descriptive generalizations that several social scientists have advanced to explain the nature of group politics in societies at different levels of development are very similar to the conflicting ideological orientations expressed by Turkish associational leaders when they describe their own system of interest politics during a specific historical period. Thus, a better understanding of prevalent political theories can also help to provide a better understanding of the perceptions and values of many key figures in Turkey's current associational life.

THREE VIEWS OF INTEREST GROUP POLITICS IN DEVELOPING SOCIETIES

General treatments of the development of interest group politics can be classified according to the different aspects of political development that seem, implicitly or explicitly, to be the most salient to the authors in question. In attempting such a classification we can benefit from James Coleman's conceptualization of political development as an open-ended process involving a "continuous interaction among the processes of structural differentiation, the imperatives of equality, and the integrative, responsive, and adaptive capacity of a political system" (emphasis added). Coleman argues that differentiation, equality, and capacity are distinct and interdependent dimensions of the "development syndrome" and that variations in political development can be understood in terms of different attempts to prevent or resolve certain political "crises" that can result from the "inherent contradictions" among these three dimensions.3 In a 3

James Coleman, "The Development Syndrome: Differentiation-Equality-Capacity," in

Political Development — 7

similar tripartite conceptualization, Schmitter distinguishes between theorists who identify political development primarily with "the centralization or concentration of authority," with "the dispersion or distribution of participation," or with "an 'integration' or 'balance' " of these interrelated but antagonistic processes.4 By approaching the work of developmental and group theorists from this perspective, three theories can be identified: (1) "gap" theory, which focuses on capacity and the concentration of authority; (2) "integration-equilibrium" theory, which focuses on differentiation and the balance between capacity and equality; and (3) "critiques of 'democratic elitism,' " which focus on equality and the dispersion of participation. By distinguishing the clear differences in emphasis that various theorists place on the three dimensions of development, we can observe corresponding differences in their views of the ways interest groups structure political participation and thereby contribute to the creation or potential resolution of various crises in political development.5 Gap Theory The leading representative of gap theory is Samuel Huntington, for whom the key factor distinguishing societies is not the type of regime but the degree of political stability.6 Stability, Huntington argues, can be conceptualized as the ratio between political participation and political institutionalization. He defines political development as the creation of "civic society," in which institutionalization exceeds participation, and he defines political decay as the emergence of "praetorian society," in which participation exceeds institutionalization. The central feature of modernization for Huntington is that rapid social change results in the increasing consciousness and political organization of new groups at a much faster rate Crises and Sequences in Political Development, ed. Leonard Binder et al. (Princeton: Princeton University Press, 1971), pp. 73-100. 4 Philippe Schmitter, Interest Conflict and Political Change in Brazil (Stanford: Stanford University Press, 1971), pp. 4-6. 3 Of course, none of these general political orientations is definitively represented by the work of any single author, and no author can be described accurately as focusing on a single aspect of development to the exclusion of others. In this sense, any attempt to classify the work of a large number of writers on such broad and complex issues as political development and interest groups is necessarily simplified to the identification of common emphases on predominant themes. Most of the authors treated here have produced several notable works, which often vary in their relative emphasis on different dimensions of development as the theorist himself changes his vie ws. Any selection of works can only provide a critical overview of what are generally regarded as some of the most influential recent contributions to developmental and group theories. 6 Samuel Huntington, Political Order in Changing Societies (New Haven: Yale University Press, 1968), pp. 1-92.

8 — Political Development than the development of strong political institutions and that the net outcome is often violence, instability, and a "corrupt society" incapable of self-government or economic growth. The dominant theme in Huntington's presentation is what he calls "the dialectical rather than complementary relationship between polity and society" which is exemplified by a series of gaps—between mass and elite, aspiration and capacity, rates of participation and rates of institutionalization, newly mobilized social forces and the state. His preoccupation with such gaps is, in turn, based on an image of an emergent mass society. Modernization tends to produce "alienation, anomie and normlessness" and to increase the consciousness and aggressiveness against the state of the disaffected.7 The fundamental problem of modernization for Huntington, then, is its tendency to encourage the unlimited and excessive expression of unaggregated demands upon weak government structures. To combat those aspects of modernization that encourage the emergence of "corrupt" mass society and to turn the tide from political decay to political development, Huntington proposes two interrelated strategic responses for political elites—an increase in the degree of political institutionalization and the control and possible reduction of political participation. Although he phrases the first response in general terms of increasing the strength and adaptability of all political institutions, Huntington is clearly most concerned with increasing the power of the state, and when he speaks of its "adaptivenss," he appears to employ a kind of vulgar functionalism in which mere persistence is equated with adaptability. Having elaborated at length his view of the "dialectical" nature of the relationship between state and society, Huntington seemingly reverses himself and speaks of a complementary relationship. What is good for the state is regarded as good for society as a whole. "The primary problem [of modernization]," he declares, "is not liberty but the creation of a legitimate public order." Huntington's understanding of legitimacy, however, appears to equate the public interest with the interest of the state. He maintains, for example, that "the legitimacy of government actions can be sought in the extent to which they reflect the interests of the governmental institutions," and that "governmental institutions derive their legitimacy and authority not to the extent to which they represent the interests of the people or any other group, but to the extent to which they have distinct interests of their own apart from all other groups." 8 In addition to a general increase in the capacity of the state to satisfy demands, Huntington also sees an immediate need to reduce the level of 7 8

Ibid., p. 37. Ibid., pp. 7, 27.

Political Development — 9 demands made upon the state through the moderation and restriction of political participation. Huntington argues that in highly developed or "institutionalized" political systems the important task of regulating participation is performed by parties and interest groups. While parties and interest groups are described as ensuring "broad, organized, and structured" participation, they are seen primarily as mechanisms to slow down and control the entry of newly mobilized social forces into politics and to make them less threatening and disruptive by changing their attitudes and behavior. Huntington describes party and group organizations, not as conduits for channeling political demands, but primarily as "filters" that shield the established political leadership from unlimited access by nonelites and that socialize nonelite leaders to preexisting political values and practices. In a sense, the top positions of leadership are the inner core of the political system; the less powerful positions, the peripheral organizations, and the semi-political organizations are the filters through which individuals desiring access to the core must pass. . . . These [intermediate] institutions impose political socialization as the price of political participation. In a praetorian society groups become mobilized into politics without becoming socialized by [preexisting patterns of] politics. The distinguishing characteristic of a highly institutionalized polity, in contrast, is the price it places on power (emphasis added).9 The most striking problems in Huntington's presentation are, first, his narrow and undifferentiated conceptualization of political participation as the expression of frustration and social discontent and, second, his understanding of political institutionalization as involving the concentration of power in the state and the restriction of nonelite access to authoritative decision makers. Because he identifies political participation with the expression of discontent arising from the dislocations of modernization, Huntington presents newly emergent interest group and party organizations as a restriction rather than an expansion of participation. It is this understanding of political participation that leads him to view it as being inversely rather than directly related to political institutionalization. Huntington's concern with the disruptive aspects of social change leads him to rely on de Tocqueville as another theorist of mass society, but his preoccupation with political order causes a complete misreading of de Tocqueville's notion of participation. Huntington asserts the basic similarity of de Tocqueville's injunction that the "art of association" must grow as equality increases with his own injunction that institutionalization 9

Ibid., pp. 22,83.

10 — Political Development

must increase as the level of participation rises.10 For de Tocqueville, however, the art of association is only secondarily a matter of demandfiltering socialization. It has nothing whatever to do with an increase in the power of the state. On the contrary, he saw the development of intermediate social structures as mechanisms that would allow common men to overcome the powerlessness of fragmentation by engaging in politics collectively and to protect themselves from what he regarded as a dangerous and potentially irresistible tendency toward the centralization of power in the state. De Tocqueville prescribes structures for the promotion and not the restriction of political participation that would protect both the individual and the state. But Huntington stands him on his head by identifying the art of association with increases in central authority and decreases in participation.11 In identifying the development of more and better organized interest groups with a possible constriction of political participation, Huntington's gap theory is similar to Gabriel Almond's functionalist scheme, in which both interest groups and parties are seen as demand-reducing mechanisms.12 For Almond, the primary task of group and party organizations is to ensure "boundary maintenance" between polity and society in order to prevent what might be called a "performance gap" characterized by the overperformance of "input functions" and the underperformance of "output functions." The demand-reducing role of interest groups is much less obvious in Almond's discussion than in Huntington's because, at least in his earlier writings, Almond focuses attention mainly on the input functions, especially "interest articulation" and "interest aggregation." This implies that development is to be understood more in terms of increases in political participation than in terms of greater output and capacity. But this focus is curiously deceptive since Almond's overriding concern with boundary maintenance (which in plain English amounts to shielding the bureaucracy from political demands) leads him to a subtle redefinition of the function of interest groups in developed political systems. The chief implication of this redefinition is the conceptual alienation of interest groups from interest articulation and of interest articulation from political development. At first glance it seems that for Almond political development is identified primarily with the emergence of more differentiated structures, which acquire increasingly autonomous responsibility for the performance of specific political functions. While realizing that in reality all structures are 10

Ibid., pp. 4-5. de Tocqueville, 2:114-119. 12 Gabriel Almond, "A Functional Approach to Comparative Politics," in The Politics of the Developing Areas, ed. Gabriel Almond and James Coleman (Princeton: Princeton University Press, 1960), pp. 3-64. 11

Political Development — 11

multifunctional, his image of a highly developed political system is one in which a clear division of labor exists, so that interest groups articulate demands and parties aggregate them, and legislatures make policies and independent bureaucracies impartially implement them. Yet Almond appears to argue that the hallmark of a developed political system is not really its sophisticated division of labor but its ability to achieve and preserve boundary maintenance, not the multiplicity of structures performing different functions but rather the common contribution of all structures to the single overriding task of reducing the level of demands and protecting the autonomy of authoritative decision makers. Though the division of labor associated with development is generally seen as improving and increasing the performance of all "universal" functions, interest articulation is not to be augmented but rather modified so that it serves or appears to become the function of interest aggregation. Although Almond considers all functions to be "universal," he does not appear to consider all functions to be equally important. This in itself is not surprising and would not justify a rejection of functionalist analysis in general. It is certainly possible to imagine a hierarchy of functions in which some are "more vital" than others or in which some are merely desirable but still expendable if they impede the performance of other functions higher up in the hierarchy. Almond himself has retained a functionalist approach while explicitly shifting his general emphasis from input functions in his earlier writings to output functions and governmental capabilities in his later work.13 But a major problem with all of Almond's functionalist reincarnations is that each leaves the hierarchical nature of the functions implicit and forces us to reconstruct it anew for ourselves. The greatest source of confusion in this regard is his repeated assertion that "among the input functions, interest articulation is of crucial importance." 14 This strongly implies that Almond considers interest articulation to be a necessary and perhaps vital function. It would seem reasonable to conclude from this and from his earlier encouragement of research on interest group politics in developing systems that political development and increasing interest articulation go hand in hand. Yet, on the basis of Almond's own discussion, it is just as reasonable to conclude the opposite. Like Huntington, Almond identifies higher levels of interest articulation with increasing modernization, but not necessarily with political development. For Almond, interest articulation and interest groups are important, not in their own right, but because, in his view, they pose the greatest potential threat to the performance of the more crucial output functions 13 Gabriel Almond and G. Bingham Powell, Comparative Politics: A Developmental Approach (Boston: Little, Brown and Co., 1966), chaps. 5, 7. 14 Almond, p. 33.

12 — Political Development (especially "rule application") and to the strengthening of governmental capabilities (especially "extraction," "regulation," and "distribution"). Among the input functions, interest articulation is of crucial importance since it occurs at the boundary of the political system. . . . The structure and style of interest articulation define the pattern of boundary maintenance between polity and society, and within the political system affect the boundaries between the various parts of the political system—parties, legislatures, bureaucracies and courts.15 The concept of boundary maintenance makes no sense without recognizing the implicit hierarchical nature of the functions, which may require some lower-order functions to be restricted or even suspended for the sake of higher-order ones. In Almond's changing functional hierarchies the most consistent theme has been the primacy of the output functions and the expendability of interest articulation. Interest articulation may be "of crucial importance," but it is also the most expendable function of all. Both its importance and its expendability appear to stem from its potential for disrupting and impeding the performance of the higher-order output functions. If development is basically a matter of improving boundary maintenance, and if this in turn requires decreases in interest articulation, then how can political development be identified with the emergence of interest groups? Almond's answer lies in his distinction between two types of interest groups—those that contribute to bad boundary maintenance and those that contribute to good boundary maintenance. The former type includes what he terms "institutional, non-associational and anomic interest groups," whose style of interest articulation is characterized as "sporadic, latent, diffuse, particular and affective." The latter type he terms "associational interest groups," whose style of interest articulation is characterized as "constant, manifest, specific, general and instrumental."16 What is distinctive about "associational" interest groups is that unlike all other types they play a "regulatory role . . . in processing raw claims . . . and directing them in an orderly way and in an aggregable form through the party system, legislature and bureaucracy'' (emphasis added). n In other words, associational groups transmit demands in such a manner that even a good functionalist would be tempted to describe their function not as demand articulation at all but as demand aggregation or, more accurately, demand reduction. Interest groups other than the associational variety, however, are not seen as processors of demands but as mere 15

Ibid. Ibid., pp. 33-38. 17 Ibid., pp. 35-36. 16

Political Development — 13

creators and conductors of demands, that is, as interest articulators par excellence. To Almond's mind that is exactly the problem with them. They inject into the political system "the 'raw,' unaggregated demands of specific interests, or . . . diffuse, uncompromising, or revolutionary and reactionary tendencies," so that "special interests and ideological tendencies penetrate the bureaucracy and undermine its neutral, instrumental character" (emphasis added).18 Despite his use of the terminology of general systems theory, Almond's framework does not resemble a systematic model or even a simpler cybernetics model so much as the notion of a communications net or flow chart in which a series of differentiated and interrelated structures (interest groups, parties, and legislatures) sequentially collect, process, and transform demands in a gradual, steplike demand-reducing process. He implies that this process should be kept as separate as possible from the crucial policy-implementing function of an ideally independent bureaucracy. Thus, if Almond appears to regard rule application as the most important function and interest articulation as the most expendable function, he also treats interest aggregation as the most ubiquitous function. He particularly emphasizes the notion that "interest aggregation can occur at many points in the political system."19 Indeed, this function seems to be the primary task of all types of nongovernmental structures in Almond's vision of the developed political system. If Almond tends to describe interest aggregation as a ubiquitous function without a specific structure, he tends even more to describe interest groups as marginal structures without a specific function. Interest groups seem to be excluded from the effects of the prevailing increase in the division of labor and denied any distinctive or unique function, since development requires their transformation from demand producers and transmitters into demand processors and filterers. As other structures supposedly become more unifunctional, interest groups actually become more multifunctional, since they articulate less and aggregate or filter more. As other structures supposedly become more autonomous, interest groups may become less so if the functions of others are considered more important. In fact, there is more confusion about the supposed function of interest groups than about the function of any other structure in Almond's model. "Articulating," "aggregating," "filtering," and "screening" clearly are not the same, yet all are included in the functional repertoire of "boundary-maintaining,'' "associational" interest groups in developed political systems. To summarize, then, whereas Huntington understands political devel18 19

Ibid., p. 38. Almond and Powell, p. 99.

14 — Political Development opment as requiring a closure of the supposedly widening gap between participation and institutionalization, Almond sees development as dependent on preventing the appearance of gaps between the performance of input functions and the performance of output functions. For Huntington the gap is to be closed by promoting institutionalization more rapidly than participation, and for Almond the gaps are to be avoided by improving boundary maintenance through the promotion of interest aggregation and the reduction of interest articulation. Both Huntington and Almond tend to see the major problem of political development as the excessive and unmediated expression of demands on weak governmental structures. Both argue that such demand overloads can be prevented or remedied by increasing the capacity of governmental institutions to satisfy demands and by promoting intermediate structures to control and limit the expression of demands. Accordingly, both maintain that the principal task of interest groups in developing political systems is not to create and transmit demands but to reduce their volume and alter their content. Thus, while both authors hold that interest groups provide an important contribution to political development, that contribution is seen in terms of their limitation rather than expansion of political participation and interest articulation. Integration-Equilibrium Theory Whereas gap theorists emphasize the disruptive and destabilizing aspects of modernization, integration-equilibrium theorists see modernization as containing the potential remedies for many of its own ills. Whereas the former see rapid social change mainly as a threat to public order, the latter are more impressed by its overall contributions to greater social and economic adaptiveness. For both types of theorists, political development involves an ongoing tension between antagonistic tendencies toward the concentration of authority and the dispersion of participation. But whereas gap theorists view these tendencies as dialectical and contradictory, integration-equilibrium theorists view them as interdependent and reconcilable. The former are more concerned that development have a clear theme and a regular rhythm, the latter that it have a complex yet pleasing harmony. The former are annoyed at the distractions of polyphony; the latter are intrigued by the depth of well-balanced counterpoint. The former are jarred by dissonant chords; the latter believe that they can be resolved by consonant ones. Integration-equilibrium theorists are no less concerned with political stability than gap theorists. They see it as attainable not by encouraging one developmental tendency to the detriment of the other, not simply by increasing the capacity of the state to close a series of supposed gaps, but

Political Development — 15 by maintaining a number of delicate balances—between liberty and order, between social differentiation and subsequent reintegration, and between the various segments of society in relation to one another. For them the chief issue of political participation is not demand reduction but conflict management and resolution. Participation is seen more clearly as a force for moderating conflict as well as increasing it, for achieving greater support for the regime as well as imposing greater demands, and for reintegration under a new consensus if enough social groups are given a stake in the system. This view of the integrative and stabilizing potential of participation in general (and of interest groups in particular) was expressed forcefully in de Tocqueville's early estimate of the likely consequences of expanding associability in nineteenth-century America. Freedom of association in political matters is not so dangerous to public tranquillity as is supposed, and . . . possibly, after having agitated society for some time, it may strengthen the state in the end. . . . By engaging [civil associations] more and more in the pursuit of objects which cannot be attained without public tranquillity, [governments] deter them from revolutions. . . . Thus it is by the enjoyment of a dangerous freedom that the Americans learn the art of rendering the dangers of freedom less formidable.20 It has become popular to identify both de Tocqueville and Durkheim with the notions of "mass society" and "crisis of authority" because of their common concern with the need for social reintegration during periods of rapid change. Yet, unlike the proponents of gap theory, who frequently cite them as supporting authorities, they are less troubled by a cataclysmic vision of the revolt of the masses than by the threat to individual liberty from a too powerful state. For de Tocqueville and Durkheim the danger of social atomization is not so much the "overavailability" of the masses as their basic powerlessness and defenselessness vis-a-vis centralizing government. Although both recommend the encouragement of secondary associations as stabilizing buffers and intermediary structures between the state and the individual, both are quite emphatic in identifying associational development with the expansion of participation rather than with the concentration of authority. As Durkheim says: A society composed of an infinite number of unorganized individuals, that a hypertrophied State is forced to oppress and contain, constitutes a veritable sociological monstrosity. For collective activity is always too complex to be able to be expressed through the single and unique organ 20

de Tocqueville, 2:126.

16 — Political Development

of the State. . . . A nation can be maintained only if, between the State and the individual, there is intercalated a whole series of secondary groups near enough to the individuals to attract them strongly in their sphere of action and drag them . . . into the general torrent of social life.21 Despite their common emphasis on balance and on the long-term contributions of participation through secondary groups to social harmony and adaptiveness, integration-equilibrium theorists tend to divide over the question of whether balance and conflict resolution can be achieved spontaneously through the operation of self-correcting forces or whether it must be ensured by a coordinating political center. For example, Madison, Simmel, Coser, and Truman all believe that under certain circumstances conflict can be structured so that it tends to be self-limiting by virtue of its very ubiquity and multiplicity and tends to bind contending groups in a more or less stable balance of power whose equilibrium will be maintained by a kind of "invisible hand." 22 Basic to these theorists is the notion that political participation through secondary associations can moderate conflict if they can substitute multiple and crosscutting cleavages that involve only peripheral parts of members' loyalties for the supposedly more intractable and all-encompassing cleavages that divide primary groups. All agree that "the violence of faction," as Madison terms it, can be eased not by removing its causes, since this could only be achieved at the expense of liberty, but by controlling its effects, and that this in turn is best achieved by multiplying the causes of faction and relying on their mutual checks and balances or countervailing power as natural guarantees of both stability and equity. These general assumptions about self-integrating and self-balancing tendencies lead to two further assumptions concerning the concept of the public interest and the nature of the state. Since noble consequences are seen as emerging naturally from selfish motives, it is difficult for such theorists to conceive of a public or general interest that may be different from and in conflict with the interests of the individuals and groups that constitute society. They tend either to dismiss the concept of the public interest entirely by denying any possible distinction between " i s " and "ought," or to equate the public interest with whatever emerges from the "group struggle" as a reflection of the "prevailing values" of the com21

Durkheim, p. 28. Alexander Hamilton, James Madison, and John Jay, The Federalist Papers (New York: The New American Library, 1961); Georg Simmel, Conflict and the Web of Group Affiliations (New York: The Free Press, 1955); Lewis Coser, The Functions of Social Conflict (New York: The Free Press, 1956); David Truman, 7"Ae Governmental Process (New York: Alfred A. Knopf, 1953). 22

Political Development — 17

munity.23 This lack of normative standards for evaluating public policy is accompanied by an inclination to deny or minimize the possibility that the state can act disinterestedly on behalf of the entire society. Government coordination and arbitration often are seen as both unnecessary and undesirable. In sharp contrast to the proponents of gap theory, they often regard the state as merely one of several groups constituting society, qualitatively no different from the others and pursuing its own institutional goals, which are inherently no more legitimate or disinterested than those of any competing interest. In contrast to such notions of homeostatic tendencies, Durkheim, Eisenstadt, and Smelser describe a direct and crucial role for the state in resolving conflict and avoiding what they regard as the real dangers of social breakdown. Each points to the paradoxical effects of the greater division of social labor as involving both improved adaptability and greater fragility. They note that the greater efficiency of newly differentiated structures is accompanied by their greater total interdependence, which increases the potential impact on all structures of a disruption by any one of them. This paradoxical nature of the division of labor means that although complex societies may be more capable of dealing with conflict, they are also more susceptible to its immediate effects. Thus, they argue, balance cannot be assumed to be the spontaneous result of self-regulating group interaction, but must be guaranteed by a central structure with unique responsibility for reintegrating the separate parts of society and coordinating them in the general interest of the whole.24 Durkheim, for example, sees the division of labor as leading to the gradual replacement of mechanical solidarity (social solidarity based on similarities) by organic solidarity (solidarity based on interdependent differences). Although he tends to focus on occupational associations as mechanisms for regulating the interaction of individuals and groups whose interests are becoming progressively more diversified during periods of growing social heterogeneity, he sees this role as belonging primarily to the state, which must also remain conscious of the need for solidarity based on similarities. [In] societies where organic solidarity is preponderant . . . social elements . . . are coordinated and subordinated to one another around the same central organ which exercises a moderating action over the rest of 23

Richard Flathman, 7"Ae Public Interest (New York: John Wiley and Sons, 1966), pp. 3-86. 24 Shmuel Eisenstadt, Modernization: Protest and Change (Englewood Cliffs, N.J.: Prentice-Hall, 1966); Neil Smelser, "Mechanisms of Change and Adjustment to Change," in Industrialization and Society, ed. Bert Hoselitz and Wilbur Moore (New York: UNESCOMouton, 1966), pp. 33-47.

18 — Political Development

the organism. . . . There is, above all, an organ upon which we are tending to depend more and more; this is the State . . . [which] is entrusted with the duty of reminding us of the sentiment of common solidarity.25 In the same vein Eisenstadt and Smelser emphasize the pivotal role of the state as a balancing-coordinating force, a reintegrative center managing the social disturbances and protests that can result from discontinuities between differentiation and integration on the basis of interdependence. Smelser's treatment of the problem is an especially interesting contrast to gap theory. He argues that in attempting to control the "contrapuntal interplay between differentiation (which is divisive of established society) and integration (which unites differentiated structures on a new basis)," the state must not create imbalances in political development. The symbols of solidarity based on similarities must not be used in such a way as to impede the emergence of a more sophisticated solidarity based on interdependent differences, and the response to political protest must not have the effect of preventing the long-term integrative potential of political participation. Like Huntington, Smelser notes that governments can effectively use the symbols of nationalism, the public interest, and state autonomy to enhance their own legitimacy, procure otherwise unobtainable sacrifices, and deal with disruptive protest. But, in contrast to gap theorists, he adds a key warning that is reminiscent of de Tocqueville: However, these political leaders should not take their claim to legitimacy too literally. They should not rely on their nationalist commitment as being strong enough to enable them to ignore or smother grievances completely. They should "play politics," in the usual sense, with aggrieved groups, thus giving these groups access to responsible political agencies, and thereby reducing the conditions that favor counter-claims to legitimacy. One key to political stability seems to be, therefore, the practice offlexiblepolitics behind thefaqade of an inflexible commitment to a national mission (emphasis added).26 Critiques of ' 'Democratic Elitism'' A third view of interest group politics, which focuses on the concept of equality, criticizes interest groups in Western Europe and the United States as restricting effective political participation, reinforcing existing social and economic inequalities, and contributing to what has been called dem25 26

Durkheim, pp. 181,227. Smelser, p. 47.

Political Development — 19 ocratic elitism. These writers share the integration-equilibrium theorists' belief that expanding participation in general is a key aspect of political development with positive long-term effects for stability and adaptiveness. They have a much less sanguine view, however, regarding interest groups as a particular dimension of participation, since they see these groups as excluding and alienating so many people and interests from the policymaking process that neither integration nor balance is encouraged. They agree with gap theorists that interest groups serve as conduits for the particularistic demands of specific sectors that tend to blur the distinction between public and private power to the detriment of the public interest. But they argue that this is true only for the interests of a narrow and privileged set of elites who are able to manipulate interest groups successfully and not for the interests of the majority who cannot. Consequently, although the remedies proposed by these critics vary widely, they generally include a call for more equal access to interest group politics and for increased self-government in daily life. In general, gap theorists see interest groups as conduits for the demands of too many and would prefer them to be more efficient gatekeepers for all. Critics of democratic elitism, on the other hand, see interest groups as conduits for the demands of too few and gatekeepers for too many, and would prefer them to be more accessible conduits for all. Stein Rokkan's work on changing patterns of political participation in Europe is typical of most long-term developmental analyses in identifying increased participation primarily with the extension of suffrage and the rise of mass parties.27 He sees the expansion of party organizations as the main mechanism for political mobilization of the lower classes soon after the granting of suffrage and as the means through which they gain greater access and representation in decision making. But Rokkan's treatment of political participation differs from that of most developmental theorists in two important respects: he does not regard increased party-electoral participation as necessarily leading to the expansion of political equality, and he indicates that newer forms of associational participation can, in fact, reinforce old inequalities and promote new ones. Rokkan's chief concern is to explain persistent inequalities in partyelectoral participation by examining comparatively the various elite decisions and strategies that led to the extension of suffrage and the differential rates of political mobilization with which various social strata responded to their newly gained rights. But he also describes newer inequalities that have emerged in modern Europe because of the crystallization of two 27 Stein Rokkan, "Mass Suffrage, Secret Voting, and Political Participation," in Political Sociology, ed. Lewis Coser (New York: Harper and Row, 1966), pp. 101-131.

20 — Political Development

distinct channels of participation—the party-electoral and the associational—and the gradual predominance of the latter over the former. He argues that while elites were attempting to draw greater numbers into the political system by making citizenship universal, increasing social differentiation and growing official bureaucracy encouraged the emergence of a network of interest groups, which helped render effective decision making a more covert process restricted to bargaining among well-organized groups, politicians, and bureaucrats. Thus, Rokkan suggests that efforts to achieve full citizenship may ultimately have resulted in recent declines in both the rate and significance of party-electoral participation relative to the more constricted channel of associational participation. Rokkan's conceptualization of political participation contrasts sharply with that of gap theorists, who view participation as the undifferentiated and excessive expression of demands and frustrations. For Rokkan, participation is a highly differentiated concept, including the expression of demands and supports at several levels such as voting, media exposure, party affiliation, active party involvement, interest group affiliation, candidacy, and office holding. The high and sustained participation rates assumed by gap theorists and integration-equilibrium theorists are a matter for empirical investigation for Rokkan, who calls our attention to potential changes in the degree and type of participation of various social strata over time. In particular, he notes that at different times certain social groups may shift their focus of participation from party-electoral channels to associational channels and that other groups may be poorly mobilized in one of the channels or even excluded from both. Reinhard Bendix deals specifically with this problem in his discussion of the European transformation from the medieval estate societies of the eighteenth century to the modern welfare nation-states of the twentieth century.28 For Bendix the most important political aspects of this transformation are "the simultaneous trends toward equality and a nationwide, governmental authority." Like integration-equilibrium theorists, he sees the trends of expanding participation and nation building as usually complementary, not always contradictory. But his treatment of the extension of the universal rights of citizenship distinctively centers on the paradox that the attempt to grant equality itself laid the basis for the creation of new inequalities. This was so, Bendix argues, because citizenship was extended with reference to two conflicting ideals of representation, which had contradictory effects for the ideal of equality: the "plebiscitarian ideal," in which 28 Reinhard Bendix, Nation-Building and Citizenship (New York: John Wiley and Sons, 1964), pp. 33-144.

Political Development — 21 each individual citizen stands in direct relation to the state and exercises equal rights without intermediate representatives, versus the "functional ideal," in which estates, corporations, or associations enjoy exclusive rights that are in effect unequal privileges. Bendix sees the major problem of political participation since the nineteenth century as an ongoing tension between these conflicting styles of representation, involving "ever newer and ever partial compromises" in pursuit of what de Tocqueville called "the phantom of equality." Bendix suggests that the initial triumph of plebiscitarianism in modern Europe generally confirmed de Tocqueville's expectation that governments could deter men from revolution and strengthen themselves in the end "by engaging them more and more in the pursuit of objects which cannot be attained without public tranquillity." He argues that originally the rights of citizenship were extended in Europe under the plebiscitarian ideal by central governments that sought to break the inherited privileges of feudalism while integrating the lower classes into expanded political communities under the banner of nationalism rather than socialism. He sees the late eighteenth century as a major break with the medieval tradition of liberty as corporate privilege, because thereafter equality gradually expanded at the expense of functional notions of the Standestaat and its hierarchically ordered constituent groups. He credits the growth of plebiscitarianism with the emergence of a new consensus concerning the authority of national governments and with the decline of socialism by diverting lower-class protest from challenging the social order toward demanding the full rights of citizenship and a larger share of national wealth. Like Rokkan, Bendix notes that European governments were mindful of the dangers of plebiscitarianism as well as its advantages. For decades elementary education and the franchise [were] debated in terms of whether an increase in literacy or voting rights among the people would work as an antidote to revolutionary propaganda or as a dangerous incentive to insubordination. . . . Leaders of the established elites became increasingly torn between their fears of the consequences of rapid extensions of the suffrage to the lower classes and their fascination with the possibilities of strengthening the powers of the nationstate through the mobilization of the working class in its service.29 Because of this ambivalence, Bendix says, governments devised a number of transitional compromise strategies for controlling the onrush of mass democracy, which often included some initial reliance on traditional estate privileges and delayed the adoption of universal adult suffrage. But unlike 29

Ibid., p. 94.

22 — Political Development

Rokkan, who seeks the source of persistent inequalities in political participation in the various historical compromises required of plebiscitarianism and in the different rates at which various social strata continue to make use of it, Bendix sees the roots of modern inequality in two fundamental problems of plebiscitarianism itself. The first of these problems concerns some unanticipated consequences of the growth of plebiscitarianism and the second concerns the fact that the right of association is not necessarily the same as the art of association, a distinction that de Tocqueville did not draw clearly enough. According to Bendix, plebiscitarianism, by virtue of its very success, led to a series of basic political transformations that encouraged a revival of functional representation in a new form and of inequalities based on new privileges. Plebiscitarianism affected the general nature of politics by changing it from a struggle over the distribution of sovereign powers into a struggle over the distribution of the national surplus. The content of demands changed from a concern with civil and political rights to a search for greater social and economic rights. And, finally, the concept of the citizen was redefined from the notion of an individual entitled to equal protection under the law to the notion of an individual entitled to public assistance. Bendix believes that each of these changes tended to highlight the advantages of collective versus individual action in dealing with largescale, modern government, so that functional representation continued to be relevant through newly emergent interest groups, which were formed both as causes and consequences of the proliferation of government.30 Under such conditions, Bendix concludes, the chief problem of participation becomes not the equality of rights but the unequal opportunities to take advantage of them. In societies marked by formal, legal equality alone, the political consequences may be merely that all citizens are equally powerless. The granting of civil and political freedoms is an instance of "enabling legislation" conferring merely the potential of power. Translating that potential into real power depends on the right of association. Where this basic civil right was withheld or restricted, the abstract principle of equality of individuals was often the direct cause of greatly accentuated inequalities. For example, the notion of the individual work contract as 30 "The simultaneous development of a nationwide authority and the plebiscitarian tendencies in the political realm are accompanied by the development of functionally defined, organized interests. The efforts of public officials to obtain support, information and guidance from the relevant 'publics' are matched point for point by the efforts of organized interests to influence government actions so as to benefit their members or clients. . . . [Public administrators] look for support of discretionary judgments [and] find such support in the opinions and expert advice which organized interests are only too willing to provide" (ibid., p. 136).

Political Development — 23

an agreement between legal equals upheld the worker's right to make contracts under the myth of his economic equality with the employer while rendering that right meaningless by denying him the right to combine with other workers. Yet even when the right of association was extended to correct such injustices, plebiscitarianism divided the population in a new way—into the organized and the unorganized—since the ability to develop the art of association depended in turn on such unequally distributed attributes as wealth, education, and political experience. Thus, the plebiscitarian impulse to guarantee equality before the law in fact created new inequalities with each advance it made, at first because in its incomplete form it denied or restricted the right of association without which all other civil rights remained dormant, and later because of the inability of many to develop the art of association even after the formal right was granted. At both points the extension of new rights benefited primarily the advantaged, who had the social and economic independence to use them, while the main burden of economic change fell upon the lower class. Consequently, where the extension of citizenship was accompanied by the growth of capitalism, it provided the foundation of legal equality upon which a structure of social and economic inequality was built or strengthened. The equality of citizenship and the inequalities of social class developed together and plebiscitarianism served as "the architect of legitimate social inequality" reflected by differences in mastering the art of association.31 Bendix concludes that in Western Europe the right of association gradually became detached from the plebiscitarian ideal of universal rights and was incorporated into the functional ideal of group-specific privileges. Originally intended as a remedy for the inequalities of partial plebiscitarianism, the right of association gradually promoted the inequalities of growing functionalism, thus indicating the continuing tension between these two basic styles of participation.32 What Bendix contributes here is a subtle but important revision of de Tocqueville's famous "law of associability." Writing in the context of the breakdown of feudal society, de Tocqueville warns that as material conditions become increasingly more 31

Ibid., p. 77. In so arguing, Bendix tends to read medieval political patterns into modern conditions by saying that "organizations based on common economic interest perpetuate or re-establish corporate principles analogous to those of the medieval period [because] . . . in effect legal opportunities have turned into privileges." Medieval privileges based on inherited status and limited by corresponding duties are replaced by modern privileges based on disparities in the ability to organize collective action with no clear social responsibility. In its pursuit of "the phantom of equality," plebiscitarianism earlier helps to extinguish the former variety of privilege, but then unwittingly serves as the midwife of the latter variety (ibid., pp. 8384, 86). 32

24 — Political Development

equal all men must learn the art of associating together lest they all become equally powerless and lose their freedom before an expanding state. Writing in the context of advancing industrial capitalism, Bendix warns that as material conditions become increasingly more unequal some men have a greater need than others to learn the art of association, but their ability to do so is generally inversely related to their need. Consequently, their freedom often becomes meaningless as universal rights are transformed into special privileges. Bendix's discussion of the tension between plebiscitarian and functional representation as well as his conclusion that modern associations tend to serve as the architect of legitimate social inequality is similar to the criticisms of interest group politics found in the writings of several American group theorists. Recent group theorists in the United States commonly have altered Bentley's original portrayal of group interaction as a rather formless "process" that underwent such continual fluctuation that definition of the political group as a stable unit was itself problematic. Instead, Bentley's successors have tended to emphasize the patterned and highly structured nature of group politics and to explore its links with formal governmental institutions, stable socioeconomic categories, and enduring aspects of culture.33 The closest American counterparts to Bendix's structural analysis of competing channels of representation appear in the ethical critiques of interest group politics made by such antipluralist writers as Schattschneider, McConnell, Lowi, and Bachrach.34 But they also go beyond Bendix by condemning this tendency as a violation of the open, participatory, and processual democratic ideals to which both they and their pluralist adversaries are committed. For these American critics the emergence of interest groups involves, not simply the development of an alternative and more specialized set of channels of participation, but a clear constriction of meaningful participation in general. In their view, the more limited nature 33 Even such leading representatives of pluralist theory as Truman and Dahl, who share Bentley's view of group politics as being basically open and ever changing, attempt to specify certain constraints on group activity and identify certain points through which it must flow. For Truman this involves particular attention to how group "access" can vary according to the structure of government institutions and internal group organizations, and for Dahl it involves stressing how group activity and influence are constrained by the norms of the regime in which these institutions are embedded and by the independent initiative of political leaders for whom the regime's norms are especially important. 34 E. E. Schattschneider, The Semisovereign People (New York: Holt, Rinehart and Winston, 1960); Grant McConnell, Private Power and American Democracy (New York: Alfred A. Knopf, 1967); Theodore Lowi, The End of Liberalism (New York: W. W. Norton and Co., 1969); Peter Bachrach, The Theory of Democratic Elitism: A Critique (Boston: Little, Brown and Co., 1967).

Political Development — 25

of participation through associational channels is not a paradoxical and inevitable product of the pursuit of' 'the phantom of equality'' but a direct and remediable result of specific structural rigidities and biases that narrow the range of policy alternatives by systematically underrepresenting or excluding certain types of citizens and interests. Each American critic focuses on a different structural aspect of group politics to account for its tendency to limit participation and contribute to stable inequality, and each proposes a different set of remedies. For Schattschneider interest groups contribute to a "mobilization of bias" in favor of the status quo, the business community, and established elites. He argues that this bias results from the tendency of group politics to narrow the "scope of conflict'' to covert interactions among a small number of well-organized participants in which the poorer and less powerful generally are unable to effect change by taking advantage of their superior numbers in an open and public confrontation.35 McConnell relies on Madison's Tenth Federalist to argue that the nature of group demands and specific policy outcomes are basically determined by the size and diversity of group constituencies. He maintains that groups with small and homogeneous constituencies tend to be exclusive and promote policies that favor the privileged and harm the general public, whereas groups with large and heterogeneous constituencies tend to be inclusive and to promote policies that are relatively egalitarian and responsive to the public interest.36 Lowi's identification of various "arenas of power" is an attempt to show how the structure of group activity is determined by the scope of the public policy at issue and especially by the divisibility of the relevant benefits. He argues that organized interest group politics focuses on issues with highly divisible outputs, encouraging a pattern of minimal conflict and particularistic arrangements among entrenched interests. He is far more pessimistic than Schattschneider and McConnell in maintaining that this type of "distributive politics" has become so pervasive that American interest groups will be utterly incapable of taking part in more open and conflictive politics in the future. Lowi not only rejects the notion that American group politics is fluid and processual, but sees it as so rigidified that all established groups eventually conform to an "iron law of decadence" in which declining militancy gives way to outright conservatism.37 Bachrach criticizes both pluralist and antipluralist writers who concen35

Schattschneider, pp. 1-47. McConnell, pp. 91-118. 37 Theodore Lowi, The Politics of Disorder (New York: Basic Books, 1971); Theodore Lowi, "American Business, Public Policy, Case Studies, and Political Theory," World Politics 16 (July 1964):676-715. 36

26 — Political Development

trate only on active, visible groups for overlooking the more subtle of "the two faces of power"—the various forms of coercion and repression that result in crucial "nondecisions" by preventing many interests from being represented in the group process however it may be structured. In fact, he goes so far as to question the value of any system of interest group representation, biased or unbiased, as discarding the ideal of equality of power in favor of the concentration of power in the hands of organized elites who ignore the common citizen's fundamental interest in participating in decisions that affect his life instead of merely benefiting from favorable policy outputs.38 The specific remedies proposed by the earlier American critics of group politics incorporate several important assumptions held by Bendix and by pluralist writers. Like Bendix, both Schattschneider and McConnell view interest groups as constituting an alternative set of channels for participation that coexist in a state of ongoing tension with other channels that are more inclusive, less biased, and less supportive of stable inequality. Both propose remedies that would not substantially alter the nature of group politics, but would instead attempt to increase the importance of other channels that rely on larger, more heterogeneous constituencies and provide a wider, more national scope of conflict. For Schattschneider this involves the emergence of "party government," whereas for McConnell it involves the strengthening of "presidential government."39 In both cases we observe a call to check the supposed predominance of functional, associational channels of participation with the countervailing power of reinvigorated plebiscitarian, party-electoral channels. For Schattschneider and McConnell the checks and balances of a fluid pluralism cannot be expected to arise from the spontaneous interaction of interest groups, but they can be expected to arise from the more general interaction of groups with centralized parties and a strong national executive. It is, in other words, as though traditional Madisonian strategy were to be applied against the network of group politics itself in order to compensate for the failure of its own "rigidified process" to conform to Madisonian principles. The remedies proposed by the more recent American critics, on the other hand, aim more directly at superseding group politics or at transforming it altogether. Lowi also considers the possibility of countering group influence through the promotion of an alternative channel of participation. Instead of relying on conventional plebiscitarian channels, he turns to what he regards as the creative anarchy of social movements that are animated by a severe sense of grievance and seek substantial change through political 38 Peter Bachrach and Morton Baratz, "Two Faces of Power," American Political Science Review 56 (December 1962):947-952. 39 E. E. Schattschneider, Party Government (New York: Holt, Rinehart and Winston, 1942), pp. 206-211; McConnell, pp. 336-368.

Political Development — 27

protest. Yet even these kinds of conflict groups, Lowi fears, would eventually yield to ' 'the iron law of decadence'' and, after achieving their initial objectives, become absorbed into a new structure of collusive alliances and compromises. The most effective remedy in Lowi's view is not for "interest group liberalism" to be superseded by the "politics of disorder," but for it to be dissolved and replaced by "juridical democracy" in which government coercion would be employed to enforce clearly stated policies while resisting interest group pressures on behalf of the already privileged.40 Bachrach favors an even more thorough transformation involving, not the reassertion of state authority, but the widespread dispersion of power among quasi-syndicalist units of self-government in the work place and the local community. Bachrach, more than any other theorist of group politics or political development considered here, regards political participation as an end in itself rather than as a means of pursuing other goals. He is by far the most emphatic in maintaining that participation contributes not only to the development of the political system but also to "enhancing the self-esteem and development of the individual." He is also the only theorist who discusses the possibility of restructuring associational life itself in order to achieve a devolution of decision making and thereby more closely approximate the ideal of equality of power.41

A FRAMEWORK FOR THE ANALYSIS OF GROUP EMERGENCE AND DEVELOPMENT

These three theoretical approaches may be useful in forming global hypotheses concerning the various ways in which interest groups can shape political participation and thereby produce different consequences for political development at different historical periods. They do not, however, specify a clear set of variables and interrelationships between variables that might provide a framework for explaining the wide variety in interest group organization and activity that is observable in different political systems at similar stages of development or in individual systems over time. The contributions of these theorists do not substantially surpass Huntington's notion of "institutionalization" or Lowi's "iron law of decadence." Thus, the principle contributions of developmental and group theorists are summarized in their assertions that, at lower levels of development, relatively unstructured interest groups tend to promote excessive participation, which threatens political stability and economic growth, whereas, at higher levels of development, well-organized groups tend to constrict effective participation, contribute to the stagnation of party-elec40 41

Lowi, The End of Liberalism, pp. 287-314. Bachrach, pp. 93-106.

28 — Political Development toral politics, and preserve stable inequality. These shortcomings have been noted and partially corrected by Philippe Schmitter and J. David Greenstone in their recent criticisms of the literature on political development and group theory. Their work has been particularly helpful in this investigator's attempt to frame a description and analysis of the changing nature and consequences of interest group politics in modern Turkey. Schmitter has argued that the emergence and development of interest groups must be understood, not simply in terms of the varying degrees to which their organizations and activities become more highly structured, but in terms of the different ways in which they become structured. More specifically, he elaborates ideal-typical conceptualizations of "corporatism" and "pluralism" as alternative patterns and strategies for structuring interest representation that have very different consequences for the organization of groups, their modes of interaction with the state, and their degree of political influence. Schmitter distinguishes between corporatist and pluralist associational structures in terms of dichotomies in several interrelated characteristics. Within their respective social and economic sectors they are singular or multiple in number and are monopolistic or competitive agents of representation. Their internal organizations are based on compulsory or voluntary membership and are hierarchically or nonhierarchically ordered. They may be officially recognized, licensed, subsidized, or even created by the state; or the state may exercise control over their selection of leadership and articulation of demands and supports.42 Schmitter argues that structural differentiation and economic development are important contributors to, if not necessary conditions for, the emergence of group politics, but that the specific direction of group development according to corporatist or pluralist patterns is more directly determined by two key intervening variables—public policy toward associations and political culture. With regard to the linkages between systems of interest group politics and socioeconomic change, he argues that historical fluctuations between corporatism and pluralism as the predominant mode of representation are understandable in terms of the changing structural imperatives and international context of capitalist development.43 Thus, Schmitter maintains that two considerations are most obviously missing from general discussions of the emergence and development of modern associations. The first is an understanding of the independent contributions of political and cultural variables as opposed to socioeconomic variables. These include the direct impact of the state in shaping 42 Philippe Schmitter, "Still the Century of Corporatism?" in The New Corporatism, ed. Fredrick Pike and Thomas Stritch (Notre Dame: University of Notre Dame Press, 1974), pp. 85-131. 43 Philippe Schmitter, Interest Conflict, pp. 3-20; "Still the Century?" pp. 107-125.

Political Development — 29 Political Culture Structural Differentiation Public Policy FIGURE

1-1. A Framework for the Analysis of Group Development

or even initiating group development and the importance of certain enduring attitudes and values, particularly among the political elite and the associational leadership, in influencing group behavior. The second is an understanding of substructural change and economic development, not in terms of a uniform and universal process, but as a multivaried process that provides different constraints and opportunities to political elites at different historical moments. Schmitter's conceptual framework is especially useful in ordering an analysis of long-term historical changes in the associational politics of modern Turkey, because Turkish public policy and political culture have promoted the emergence of both corporatist and pluralist types of associations. Furthermore, the relative predominance of these associational types has periodically shifted as successive regimes and political elites have adopted different strategies of economic development and political domination. A more explicit statement of the postulated relationships between the art of association and other relevant variables is outlined in the model given in figure 1-1 which will serve as the general framework of this analysis, most notably in Part I of the study. Greenstone argues that group theorists generally have been unable to account for situations in which previously quiescent groups suddenly intervene in politics to demand major social transformations or in which previously active groups become increasingly militant rather than conforming to the "iron law of decadence." He maintains that although group theorists have departed from Bentley's original processual view to emphasize the structure provided to group politics by stable social and economic categories such as race and class, they have, nevertheless, retained Bentley's method of identifying politically relevant groups in terms of subjective interests. According to Greenstone, group theorists who define groups in terms of subjective interests can only account for sudden increases in group activity and militancy through a "stability-disruption-protest" model, which assumes that stable subjective goals are upset by an unfavorable change of circumstances and that group political responses seek

30 — Political Development

to restore a preexisting equilibrium. Greenstone notes, however, that it is often the group goals that change and not merely the circumstances, and that such new goals may include demands for a radical transformation of previous conditions rather than a return to them.44 Using American civil rights and labor organizations as examples, Greenstone tries to show that the group theory approach can be developed by incorporating an "objective interest analysis" into its framework. Greenstone argues that the scope and intensity of conflict generated by group activity vary widely according to the subjective recognition or nonrecognition of objective group interests. Instead of assuming a tendency toward greater group conservatism, he maintains that group activity can suddenly and greatly expand the scope and heighten the intensity of political conflict when previously unrecognized objective interests are perceived and pursued by large stable social categories. Greenstone suggests that there is no reason that group theorists cannot account for such outcomes and also include in a group theory analysis "discussions of class or near-class political activity where the facts justify it." 45 He maintains, however, that they must, first, place greater emphasis on large stable social groups as the main units of analysis and, second, distinguish between changing subjective group goals and enduring objective group interests in substantially improving their social and economic conditions, particularly if such groups have been relatively deprived. Greenstone's introduction of objective interest analysis into group theory as a means of improving its ability to explain class or near-class political activity resembles Nicos Poulantzas's attempt to direct the attention of class analysis toward interest group political activity in order to resolve the conflict between the ' 'historicist'' and ' 'empiricist'' concepts of class.46 44 J. David Greenstone, "Group Theory," in The Handbook of Political Science, ed. Fred Greenstein and Nelson Polsby, 2 vols. (Reading, Mass.: Addison-Wesley, 1975), 2:243318. « Ibid., p. 290. 46 Poulantzas argues that the existence or nonexistence of a social and economic group as a distinct class is determined neither in minimalist terms of its mere presence at the economic level nor in maximalist terms of possessing its own political and ideological organization. Instead, he maintains that a socioeconomic group can be identified as a "distinct and autonomous class" when and only when it has reached "a certain organizational threshold," at which its economic existence is reflected on the political and ideological levels by a "specific presence" that has "pertinent effects" on these noneconomic levels. Poulantzas argues for conceptualizing class organization and the capacity of a class to realize specific objective interests "in the broad sense" in order to distinguish between "a class's practicewith- 'pertinent-effects' " and the more advanced "organization of its power" in an autonomous party. Poulantzas describes the "pertinent effects" of intermediate levels of class organization as "a new element" that "transforms the limits" of existing political and ideological structures and practices and contributes to "important modifications" of them. Specific examples of such pertinent effects include the organization of broad professional

Political Development — 31 Of course, Greenstone is not attempting to substitute class analysis for the group theory approach any more than Poulantzas is attempting to substitute the group theory approach for class analysis. Rather they argue, from different perspectives, that both group theory and class analysis can be improved in defining their basic units of analysis and in explaining certain important kinds of group and class political activity by placing greater emphasis on the degree to which the objective interests of broad social and economic groups are perceived and pursued by large representative associations and especially by "peak associations." This approach is particularly useful in examining interest group politics in the early and intermediate stages of economic development where class formation and organization may be undergoing rapid change, as well as in certain societies (such as the United States and Turkey) where traditionally there have been strong cultural inhibitions to the recognition and pursuit of objective interests among social and economic groups. Greenstone's suggestion that the study of group politics include an objective interest analysis of large socioeconomic categories in order to account for possible sudden increases in group militancy and conflictive political activity provides a valuable perspective for examining some of the most important recent changes in Turkish politics. These changes include the increasing salience of functional cleavages as the basis for group organization, the growing importance of peak associations as the principal agents of interest representation in the major economic sectors, the heightened perception of interest conflict between and within the largest peak associations, and the rising militancy of several groups such as industrialists, segments of the working class, small-scale traders and craftsmen, and technical personnel. Each of these developments will be examined in Part II of this study. The study consists of two parts: "The Bases of Interest Group Politics" and "The Network of Interest Group Politics." The discussion in Part I departs from the orienting hypothesis of Durkheim and de TocqueviUe that association formation is the product of increasing structural differentiation and the universalization of citizenship (chap. 2). Political culture and public policy toward associations are then examined as two key intervening variables that account for the often wide range of variance in the political influence of interest groups and in their strategies of interaction with the state (chaps. 3,4). Social differentiation and formal political equality are considered in terms of their contributions to the emergence and diffusion of a diverse set of interest groups since World War II. Political culture associations, their interactions with governmental and party structures, and especially their contributions to ideological and structural realignments of party systems (Nicos Poulantzas, Political Power and Social Classes [London: New Left Books, 1973], pp. 57-98, 104-114, and Fascism and Dictatorship [London: New Left Books, 1974], pp. 71-88).

32 — Political Development and public policy are considered in terms of their contributions to the ongoing tension between pluralism and corporatism as alternative strategies for structuring interest representation during different historical periods and in different economic sectors. Part II concerns the current network of interest group politics, which is examined from three different perspectives. First, the emergence and diffusion of several types of voluntary associations are analyzed in terms of their changing relationships with various indicators of modernization and with electoral support for the two major political parties (chap. 5). This involves a comparative time-series analysis of aggregate data on eleven different types of association between 1945 and 1970 in order to explore their changing social bases and to identify their roles in the recent realignment of the national party system. Second, recent interest group organization and activity are examined comparatively in four major economic sectors—labor, business, smallscale trade and crafts, and the free professions (chap. 6). The discussion focuses on evaluating the consequences of recent governments' attempts to promote the corporatization of interest representation and to develop stable clientelistic relations with peak associations by fostering quasi-monopolistic representational cartels under the control of cooperative association leaders. Third, an analysis of an attitudinal survey of labor and business leaders describes their general ideological orientations toward interest group politics and attempts to determine whether experience in the leadership of the nation's most important pluralist and corporatist associations tends to promote attitudes toward group politics that can be described as pluralist or corporatist (chap. 7). The responses of labor and business leaders are compared with each other and with those of two control groups having much less direct political experience—university students and young, lowerechelon unionists—in order to specify the major sociological and attitudinal bases of corporatism and pluralism in contemporary Turkish political culture. A summary chapter reviews the major findings of the study and attempts to evaluate the consequences of the growing conflict between pluralist and corporatist strategies for Turkey's future political development (chap. 8). Finally, a concluding chapter compares Turkey's experience with other systems in which the ongoing competition between pluralism and corporatism has played a major role in shaping political conflict (chap. 9). By comparing the findings of the Turkish case study with those of several recent works on interest representation in other societies, we will attempt to show how they can contribute to a reformulation of conventional pluralist and corporatist paradigms of group development.

PARTI

The Bases of Interest Group Politics

CHAPTER 2

Structural Differentiation and Uneven Development: Socioeconomic Bases

of Durkheim and de Tocqueville that associability is a product of the increasing division of labor in society and the expansion of formal political equality provide a useful point of departure for a discussion of the emergence of associational interest representation in modern Turkey. Durkheim's focus on structural differentiation serves as a reminder that the Turkish economy has become wealthier and more productive not simply because of the growth and increased capacity of previously existing economic activities and structures but because of a more fundamental structural transformation toward a more diversified and specialized form of social and economic organization. Similarly, de Tocqueville's expectation that the formal egalitarianism of modern democracy would promote the emergence of diverse forms of collective action is confirmed by the rapid and broad expansion of Turkish associational life soon after the universalization of citizenship. The work of both Durkheim and de Tocqueville assists us in understanding that the combination of broad structural change and liberal democracy in post-World War II Turkey has provided a context within which political participation in general and the art of association in particular have become increasingly oriented toward protecting and advancing the competing interests of newly organized social and economic groups. THE ORIENTING HYPOTHESES

This section discusses the socioeconomic bases of interest group politics by examining such structural transformations as occupational diversification, urbanization, and especially industrialization. The approach here stands in contrast to that of Durkheim and de Tocqueville by regarding structural differentiation within a liberal democratic political framework as a necessary but not sufficient cause of growing interest conflict and its expression in associational interest groups. Especially since 1950, Turkey's social and economic transformation has often tended to reinforce or widen gaps and inequalities between different economic sectors, social classes, and geographic regions, and this pattern of uneven development has, in turn, produced a heightened perception of interest conflict between various groups and regions competing for larger shares of the benefits of economic growth. A major catalyst for forming new political associations and conflict groups

36 — Bases of Politics

has been the growing perception of the inequalities that have been created or exacerbated by the uneven nature of recent structural transformations. Most of the discussion will be devoted to examining the process of industrialization in the postwar period and particularly since 1960. Rapid industrialization consistently has been the major focus of Turkish development policy since the formation of the republic. Its changing problems and imperatives have strongly influenced all other processes of structural change and have provided the major lines of the growing interest conflict that is increasingly straining Turkish democracy. Therefore, it is important to examine carefully the social and political problems involved, first, in the earlier encouragement of industrial diversification and import substitution to develop the home market and, second, in the recent attempt to restructure Turkish industry and to reorient it toward exporting in order to make Turkey competitive with its proposed future partners in the European Common Market.

OCCUPATIONAL DIVERSIFICATION

One useful indicator of structural differentiation and the development of a more complex form of social and economic organization is the increasing diversification of Turkey's occupational structure and the growing relative importance of more specialized occupational categories. Table 2-1 provides an overview of changes in the occupational composition of Turkey's economically active population over nearly half a century. Three dominant trends stand out as evidence of a substantial long-term shift in the structure of the Turkish labor force. First, the most consistent and sizable change has been the relative decline of employment in agriculture from approximately 88 percent to 64 percent of the national total. Turkey remains a predominantly agricultural society, but since 1927 one quarter of its labor force has abandoned agriculture to engage in different livelihoods and since 1965 employment in the agricultural sector also has been declining in absolute terms. Second, Turks continue tofindnonagricultural occupations, but they are not moving into one or two alternative categories; rather, they are filling out a variety of fields to form an increasingly differentiated occupational structure. Every nonagricultural category has at least doubled its share of total employment and the "measure of industrial diversification" devised by Gibbs and Martin to gauge overall occupational differentiation has shown an unbroken pattern of substantial increases.1 Third, the size of the population engaged in economic production has grown much 1 Jack Gibbs and Walter Martin, "Urbanization, Technology, and the Division of Labor: International Patterns," American Sociological Review 27 (October 1962):667-677.

Socioeconomic Bases — 37 TABLE 2-1 Economically Active Population by Major Occupational Groups, 1927-75 (Percent) 1927 1935 1945 1955 1960 1965 1970 1975 Agriculture 87.9 Mining f Manufacturing J 3.7 Construction 1 Trade & commerce 3.1 Transportation & communications I" 5.3 Services & < professions J^ Miscellaneous & unknown — Measure of industrial diversification .220 E.A.P. as percent of total population 67.0

85.3 .2 5.5 .8 2.3 1.3

84.8 .3 4.7 .5 2.4 1.2

77.3 .5 6.0 1.6 2.8 1.6

75.0 .6 6.8 2.2 3.1 1.9

72.0 .6 7.1 2.6 2.9 2.1

67.0 .7 8.4 2.8 6.1 2.6

64.1 .8 10.8 3.3 4.6 3.4

4.6

6.1

4.2

4.3

6.3

11.5

12.0

—

—

6.0

5.1

6.4

.9

1.0

.264

.274

.391

.426

.468

.526

.569

58.8

60.5

50.7

46.9

43.2

40.6

38.7

SOURCE: State Planning Organization,

Kalkinan Turkiye [Developing Turkey] (Ankara, 1971), p. 14; Z. Y. Hershlag, Turkey: The Challenge of Growth (Leiden: E. J. Brill, 1968), p. 330; State Statistics Institute, 1970 General Population Census—Social and Economic Characteristics (Ankara, 1973), p. 570; World Bank, Turkey: Prospects and Problems of an Expanding Economy (Washington: International Bank for Reconstruction and Development, 1975), p. 26; State Statistics Institute, Statistical Yearbook (Ankara, 1977).

less rapidly than the size of the total population. While Turkey's population more than doubled and annual growth rates reached nearly 3 percent, the relative size of the work force above fifteen years of age shrank drastically from two thirds to only two fifths of the total population. Thus, a much larger proportion of a rapidly expanding population is being supported at rising standards of living by a more productive labor force that is distributed over a wider variety of economic activities. It is clear that the pace of occupational differentiation was fairly modest until 1945 and that it increased substantially thereafter. Between 1927 and 1945 only about 15 percent of the productive population was employed in nonagricultural activities. Of these, about one third were found in mining and manufacturing and about two thirds were in the service sector. Ag-

38 — Bases of Politics

nculture's share of the total work force remained quite stable throughout this period despite a slight decrease. World War II appears to have slowed the pace of occupational differentiation and to have diverted its direction somewhat away from the secondary sector toward the tertiary sector. Before 1935, when the state strongly supported the development, first, of private and, then, of public industrial enterprises, most of the decline in agricultural employment was absorbed by manufacturing. But during the war years smaller decreases in the relative importance of both agriculture and manufacturing were absorbed by the temporarily bloated services and professions. Between 1945 and 1960 the prewar pattern of differentiation reappeared, but at a much more rapid rate of change. This was a period in which rapid farm mechanization led to a considerable displacement of the agricultural population, most of which was absorbed by the heavily protected and subsidized private industrial and housing sectors. By I960 one fourth of the economically active population was employed in nonagricultural categories with the most striking increases occurring in manufacturing and construction. In the tertiary sector trade and transportation showed modest growth, while the services and professions quickly dropped from their wartime high point and remained relatively stable. The rate of occupational differentiation continued to increase after 1960, when centrally planned economic development was introduced, especially during the period of Justice Party rule, from 1965 to 1971. Yet the volume of new employment created during Turkey's first two Five-Year Plans was consistently below the rather modest targets, and the pattern of occupational diversification took a very different direction from what had been intended. Though a slight increase in overall unemployment was accepted in advance as a price worth paying for more imported technology and more rapid increases in production, the price was much higher than expected. By 1973 the official estimate of unemployment rose to nearly two million out of a labor force of about sixteen million with almost half of the unemployed in nonagricultural sectors.2 Perhaps even more important than the deteriorating employment situation was the government's failure to continue the desired structural transformation of the Turkish labor force. During both plans there were sizable shortfalls of new employment in industry and construction and large overfulfillments in the service sector. Industrial 2 During the First and Second Plan periods (1963-67 and 1968-72) the economically active population grew at about 4 percent per annum, acquiring about 400,000 new job seekers each year. The plans called for annual increases in total employment of only 3 percent and realized only about 2 percent.

Socioeconomic Bases — 39

employment, which was targeted to grow first at 7 percent and then at 9 percent, increased only by 3 percent and 5 percent.3 The inability of the industrial sector to absorb the displaced agricultural population is in large part a consequence of the sector's growing tendency to substitute capital for increasingly expensive and unionized labor and its basic concern with increasing productivity rather than creating employment. This growing capital intensiveness of Turkish industry is reflected in the recent appearance of certain distortions in the structure of the labor force such as higher levels of total unemployment and the rapid growth of marginal jobs in the tertiary sector. Both types of distortion would undoubtedly be even more evident if recent attempts to accelerate industrialization had not also been accompanied by the massive export of Turkish workers to Europe. The flow of the 815,000 officially registered Turkish workers into the European pool of immigrant labor has been an important safety valve for reducing the pressure of domestic unemployment.4 The remittances from workers abroad have also provided Turkey with a means of compensating for its growing foreign trade deficit, since the rising value of foreign exchange derived from such remittances often has equaled what it earns from all of its exports. Yet the indisputable short-term benefits of this strong dependence on exporting labor may prove to be deceptive if these structural problems persist in the domestic labor force, especially when European economic slowdowns suddenly force Turkey to reabsorb a large part of Europe's exported unemployment.

URBANIZATION

The displacement of much of the agricultural population and its redistribution among a more diverse set of occupations has taken place through massive internal migration and an accelerating process of urbanization. The push of farm mechanization and the consolidation of commercialized landholding in the countryside have combined with the pull of alternative employment and higher living standards in the cities to draw more and more of the rural population into the more complex economies and social structures of urban areas. Before 1950 the modest rate of occupational differentiation received little stimulus from urban migration and was more closely related to internal changes within city economies. Between 1935 3 During the same two planning periods employment increases in the tertiary categories were 4 percent and 9 percent for trade, 5 percent and 7 percent for transportation, and 7 percent and 8 percent for services (Suzanne Paine, Exporting Workers: The Turkish Case [London: Cambridge University Press, 1974], pp. 33-35). 4 Since 1965 about one third of the increment in the labor supply has been absorbed by emigration (ibid.).

40 — Bases of Politics

TABLE 2-2 The Growth of the Urban Population and Its Distribution among Types of Settlements, 1935-75 (Percent)

Urban population as percent of total population Average annual growth rate for five-year intervals

1935 1940 1945 1950 1955 I960

1965 1970 1975

16.3

18.1

18.6

18.7

22.5

26.2

29.8

35.1

42.0

—

3.8

1.8

2.3

6.7

6.1

5.3

6.1

6.4

Distribution of Urban Population

Average Annual Growth Rates

City Size

1950

1960

1970

1975

1950-60

1960-70

1970-75

10,000-20,000 20,000-50,000 50,000-100,000 100,000-500,000 500,000 +

22.6 23.4 10.1 18.8 25.1

14.8 21.2 17.9 17.5 28.8

12.9 17.8 10.9 22.1 36.3

11.0 15.9 12.0

1.8 5.6 12.4 5.4 8.0

4.4 3.5 .6 8.3 7.9

3.2 4.5 4.6

\60.9

{«

SOURCE: World Bank, pp. 465-466; State Planning Organization, Kentlesme Grubu Calismalan [Urbanization Group Studies] (Ankara, 1977); State Statistics Institute, Statistical Yearbook (Ankara, 1977), p. 30. (In this and other tables all columns may not add up to 100% because of rounding.)

and 1950 the urban population grew only slightly faster than the total population and wartime restrictions on internal migration kept the ratio of urban to total population nearly constant throughout the 1940s (see table 2-2). Since 1950, however, urban population consistently has increased more than twice as fast as total population. Between 1950 and 1975 cities with over 10,000 inhabitants more than doubled their share of the national population from 18 percent to 42 percent. Assuming that urban growth will continue at annual rates of 6 to 7 percent, we can estimate that Turkey will make the transition from a predominantly rural to a predominantly urban society by 1985, when 54 percent of the population are expected to be city dwellers.5 During the 1950s this urbanization process resulted in a much more even distribution of the city population among various-sized settlements. 5 World Bank, Turkey: Prospects and Problems of an Expanding Economy (Washington: International Bank for Reconstruction and Development, 1975), p. 30.

Socioeconomic Bases — 41

An extensive program of highway construction contributed to the especially rapid growth of middle-sized settlements and provincial capitals with 50,000 to 100,000 inhabitants. This had the important result of filling out and diversifying Turkey's urban network and providing new linkages between the few large metropolitan centers and the widely scattered provincial towns. During the 1960s, however, the pattern of internal migration and urbanization shifted substantially. Middle-sized cities, which earlier had experienced the fastest rates of growth, remained nearly stable in size and their share of the total urban population declined sharply. The three metropolitan areas continued to expand at about 8 percent per year, but the most striking increases occurred in the seventeen secondary cities ranging between 100,000 and 500,000 residents. Thus, whereas urbanization during the 1950s tended to distribute the population more evenly among a more differentiated network of settlements, after 1960 it tended to concentrate most of the urban population in the twenty largest cities.6 The employment prospects for these new urbanites have been quite limited. Especially since 1960 the pace of urbanization has greatly exceeded the creation of new employment in industry and has been much more closely associated with the rapid expansion of the service sector. For this reason Miibeccel Kiray has characterized recent Turkish urbanization as a process of "fast de-peasantization and slow workerization" through which much agricultural unemployment has been transformed into marginal employment and disguised unemployment in tertiary occupational categories.7 This means that as urbanization has quickened in pace and shifted toward concentrating population in the largest cities, it also has become less possible to describe the process as integrating newcomers into a more differentiated and functionally specific socioeconomic structure. In her important study of Izmir, Kiray argues that recent urbanization has involved not greater structural differentiation and diversification but segmental growth and congestion. The rapid and largely unplanned proliferation of small, inefficient workshops and retailers has made the coordination of production and distribution more difficult. The appearance of a large economically marginal population has put pressure on traditional craftsmen and undermined their standards of training and controlled pro6

In 1959 about 50 percent of Turkey's city dwellers lived in communities of less than 50,000, whereas by 1975 61 percent of them lived in communities of over 100,000. The Ministry of Reconstruction and Settlement expects this concentration to continue and projects that by 1985 75 percent of the urban population (or 40 percent of the total population) will be living in cities of more than 100,000 (ibid.). 7 Miibeccel Kiray, "Squatter Housing: Fast De-Peasantization and Slow Workerization in Underdeveloped Countries,'' paper presented at the 7th World Congress of Sociology, Vama, 14-19 September 1970.

42 — Bases of Politics

duction. The inability of construction to meet the rapid increase in demand for low-cost housing has produced sprawling squatter settlements around every major city. The narrow authority and feeble financial resources of municipal governments has often allowed a deterioration of essential services.8 Despite the obvious poverty of recent urban migrants and slum dwellers compared to the general city population, several surveys indicate that their levels of income, health, and education are substantially higher than those of their village counterparts.9 There is also convincing evidence, however, that the kind of concentrated urbanization that has become predominant actually has widened inequalities in the distribution of income. Hacettepe University's 1968 study of income distribution in Turkey reveals a clear curvilinear relationship between inequality and settlement size. Various measures of inequality are highest for the village population and decline sharply for hamlets with up to 15,000 inhabitants and for small towns with up to 50,000 inhabitants. But income inequality increases slightly for the most rapidly growing secondary cities with more than 50,000 residents and increases sharply for the three metropolitan centers. The study also found that per capita income does not increase uniformly with settlement size. Average income rises gradually from villages through small- and middle-sized towns, but then drops for the secondary cities that have absorbed the bulk of urban migration since I960.10 These aspects of recent Turkish urbanization appear to have mixed implications concerning the propensity of new city dwellers to engage in collective action. Since increases in urbanization no longer involve proportionate increases in socioeconomic differentiation and new urbanites often are not being integrated into more specialized occupations, they might be expected to rely on informal patron-client relations or limited mutual assistance in adapting to city life and benefiting from its greater material advantages. On the other hand, their exposure to new but not necessarily less severe forms of social and economic inequality might be expected to encourge broader varieties of associability, especially since those ine8

Mubeccel Kiray, OrgUtle^emeyen Kent [The unorganizable city] (Ankara: Sosyal Bilimler Dernegi, 1972), pp. 56-108. Economic planners have begun recently to deal with the chaotic pattern of urban production by shifting from predominantly sectoral plans to new areal projects that attempt to consolidate small production units in new "industrial parks." Nevertheless, Turkey's growing urban problems are likely to worsen for some time, particularly since the Third and Fourth Five-Year Plans projected long-term declines in allocations for housing and in the growth rate of local government revenues (World Bank, p. 31). ' For example, per capita agricultural income is about one half of average per capita income and one fourth of industrial per capita income (ibid., p. 175). 10 Tuncer Bulutay, Hasan Ersel, and Serim Timur, TUrkiye'de Gelir Dagilimi [Income distribution in Turkey] (Ankara: Siyasal Bilgiler Fakiiltesi, 1971), pp. 67-117.

Socioeconomic Bases — 43

qualities are increasing in the areas of greatest migration. This suggests that as the nature of Turkish urbanization has changed, it may have provided different kinds of stimuli for association formation. During the 1950s and in middle-sized cities of the 1960s urbanization has probably contributed to associability mainly through its promotion of structural differentiation. But since 1960, particularly in the most concentrated urban areas, urbanization has probably contributed to associability primarily through its promotion of greater inequalities among the city population. INDUSTRIALIZATION

Modern Turkish political elites have pursued a remarkably consistent strategy of economic development that has focused on rapid industrialization and the achievement of greater self-sufficiency within the context of a mixed economy. The emphasis on industrialization has provided a major element of continuity in economic policy despite substantial changes in the relative importance of the state and private sectors and in the particular form of the division of labor between them.11 It is necessary to examine the development of domestic industry since 1950 as the most important structural transformation in the Turkish economy. But this analysis must also indicate, first, how industrialization has contributed to the overall pattern of uneven development by creating or reinforcing imbalances between different economic sectors, regions, and classes and, second, how industrial development has produced some major problems in the structure of Turkish manufacturing that seriously hinder the kind of industrial transformation that has been planned for the next twenty years. When industry's importance relative to other economic sectors is considered in terms of changes in output, investment, and productivity, overall industrial development appears quite impressive (see table 2-3). While the Turkish economy has been transformed, in terms of production, from a predominantly agricultural economy to a predominantly service economy, the fastest rates of growth have been occurring in the industrial sector. 11 Industrialization policy since 1950 can be described in terms of three fairly distinct stages. From 1950 to 1962 import substitution through strong protectionism and large subsidies to private manufacturers was used to provide Turkey with a more diversified industrial infrastructure and an expanded home market. During the second stage, between 1963 and 1969, comprehensive central economic planning was introduced to improve coordination between the state and private sectors and to increase their productivity. Although the emphasis on import substitution continued, several incentives were provided to encourage the consolidation of private sector production and to reorient it toward greater exporting. Finally, since 1970 policy makers have been trying to implement a long-term strategy of restructuring Turkish industry and improving its competitiveness in foreign markets in order to prepare for the planned customs union with the European Common Market by 1995.

44 — Bases of Politics

TABLE 2-3 Distribution and Growth Rates of Gross Domestic Product by Economic Sector, 1950-75 {Percent at 1968 Prices) 1950

1955

1960

1965

1970

1975

49 13 38

40 15 45

41 17 42

33 20 47

29 22 49

26 25 49

Share of G.D.P. contributed by: Agriculture Industry Services

Average annual rates of increase of G.D.P. in: Agriculture Industry Services Total

1950 -55

1956 -60

1961 -65

1966 -70

1971 -75

1950 -62

1963 -75

5.7 6.9 8.1 6.3

5.3 4.7 7.0 6.1

2.3 8.5 6.2 5.1

3.5 9.5 7.5 6.7

4.8 11.7 10.3 9.0

4.8 5.6 7.2 5.8

3.7 10.3 7.8 6.9

SOURCE: World Bank, p. 336; State Statistics Institute, Statistical Yearbook (Ankara, 1977), p. 362.

There have been considerable changes in the pace of agricultural and industrial growth that appear to coincide with the introduction of central economic planning after 1962. Agricultural production, which has shown the lowest overall rates of increase, declined sharply during the 1960s and has only recently regained its earlier levels. The pace of industrial growth in the planning period, however, has nearly doubled and industry has overtaken services as the most rapidly expanding economic sector. Another view of long-term changes in the relative importance of the different sectors emerges from examining global changes in the pattern of investment and the degree of capital intensiveness (see table 2-4). Since 1950 the ratio of domestic savings to total production has doubled and currently approaches 20 percent. As savings have increased, the sectoral distribution of fixed capital investment has changed markedly. Agriculture's share of investment has remained quite stable, at about 10 percent, but industry has been absorbing a much higher share, and services a much lower share, of expanding investment. Each sector is also experiencing important changes in the ratios between capital and labor and between capital and output. Capital intensiveness has increased greatly in agriculture

Socioeconomic Bases — 45 TABLE 2-4 Indicators of Savings, Investment, and Capital Intensiveness, 1950-72

Domestic savings as percent of G.D.P. Domestic savings as percent of G.N.P. Shares of total fixed capital investment in: Agriculture Industry Services

Incremental capital-labor ratios in: Agriculture Industry Services Total Incremental capital-output ratios in: Agriculture Industry Services Total SOURCE:

1950

1955

1960

7965

1970

1972

9.2

13.8

13.0

15.9

19.1

15.5

9.2

13.6

12.4

16.2

20.0

18.4

9 25 66

11 28 61

17 30 53

9 38 53

11 41 48

10 10 80 1955-60

1960-65

1965-70

15.8 31.5 105.7 54.6

63.0 46.6 83.2 66.0

— 122.6 60.4 87.8

.56 1.86 7.73 2.86

1.37 2.48 5.02 3.12

2.52 2.59 3.44 2.97

World Bank, pp. 8, 18.

and especially in industry and has decreased greatly in the services. This appears to reflect important differences in the kind of structural changes occurring in each sector. Agriculture is losing labor and absorbing capital at stable or moderate rates, yet it is not becoming substantially more productive. Industry, which is experiencing the greatest increases in productivity, is absorbing capital at increasing rates and much more rapidly

46 — Bases of Politics

than it is absorbing labor. The services, in contrast, are absorbing capital at decreasing rates and labor at much higher rates, resulting in lower levels of productivity. In other words, the rapid growth and structural change of industry are occurring within the larger context of a relatively stagnant agricultural sector and a bloated and much less productive service sector. The impressive advances of industry have not necessarily been advantageous to other economic sectors or to the economy as a whole. In fact, ever since the end of World War II, a recurrent criticism of Turkish economic policy from various foreign advisers and consultants has been that its tendency to identify economic development with rapid industrialization has diverted attention from the development of other areas such as agricultural exports and mining and, especially, tourism and private financial institutions.12 These problems have become even more serious in view of the recent shift of industrialization strategy toward greater capital intensiveness and importation of the most advanced technology. The new long-term strategy of gradually restructuring Turkish manufacturing and opening it to foreign competition requires that greater emphasis be given to coordinating the development of all economic sectors and gearing them to large-scale industry's growing needs for new investment. Thus, as Turkish developmental policy has become increasingly dominated by the imperatives of preparing for prospective Common Market membership, Turkey is confronted not only with serious weaknesses in the structure of industry itself, which arose during the earlier stages of import substitution, but also with growing imbalances between the major economic sectors. The major achievement of import substitution since 1950 has been the increasing diversification of Turkish manufacturing, especially in the private sector, involving an important shift away from consumer goods industries toward intermediate and investment goods industries. Table 2-5 illustrates this change in the general structure of industry in terms of the relative distribution of employment and production in firms with ten or more employees in various branches of manufacturing. Between 1955 and 1970 the share of consumer sectors in total manufacturing dropped sharply 12 World Bank, pp. 32-46; Max Thornburg, G. Spray, G. Soule, Turkey: An Economic Appraisal (New York: Twentieth Century Fund, 1949). The relative neglect of these areas during an extended period of import-substituting industrialization is one of the major reasons for Turkey's growing trade deficit and frequent shortage of foreign exchange. The failure to exploit these potential sources of foreign exchange and to mobilize domestic financial resources adequately has forced Turkey to rely heavily on foreign aid and more recently on the remittances of emigrant workers in order to finance the growing volume of imports required by the developing manufacturing sector.

Socioeconomic Bases — 47 TABLE 2-5 The Diversification of Manufacturing, 1955-70 (Employment and Gross Value Added by Industrial Sector) (Percent) Employment 1970 1963

1955

J

1955

Value Added 1970 1963

Food Beverages Tobacco Textiles Clothing Wood Furniture Paper Printing Leather Rubber Chemicals Petrochem. Nonmet. mins. Basic metals Metal prods. Nonelec. mach. Elec. mach. Trans, equip. Miscellaneous

13.8 1.4 18.9 32.7 .8 1.7 .2 1.5 1.3 .4 2.2 5.0 .3 5.1 3.6 3.6 1.0 .1 5.8 .3

20.0 1.4 5.2 32.2 .9 1.9 .3 2.3 1.6 .6 1.9 4.4 .6 6.3 4.1 6.6 1.6 1.8 5.5 .7

15.0 1.8 6.8 25.6 1.6 1.7 .5 2.4 1.9 .5 1.7 6.0 .4 7.2 6.1 6.6 4.0 1.9 6.5 1.6

22.0 4.8 12.3 30.7 1.0 1.1 .2 1.6 1.5 .5 2.1 6.6 .2 4.4 3.9 2.6 .9 .2 2.8 .5

18.8 3.3 6.5 22.0 .5 1.5 .2 3.3 2.0 .5 1.0 7.1 5.2 5.6 6.1 6.1 2.6 2.4 3.5 .6

12.6 3.7 9.7 13.8 .5 .9 .2 2.6 1.8 .2 1.7 6.6 15.1 5.2 10.6 4.3 4.0 1.4 3.1 1.5

Consumer goods Intermediate goods Investment goods

67.6 21.3 10.5

59.7 24.0 15.5

50.8 28.6 19.0

70.8 22.1 6.5

51.1 32.5 14.6

40.4 45.2 12.8

SOURCE: State Statistics Institute, Statistical Yearbook, 1964-65 (Ankara, 1965), p. 456; State Statistics Institute, Statistical Yearbook, 1977 (Ankara, 1977), pp. 207-209. from about two thirds to one half.13 The greatest shifts were from such traditional branches as food and tobacco processing and textiles to more specialized sectors such as petrochemicals, basic metals and metal prod13

Of course, even this substantial degree of diversification still leaves Turkey with a relatively primitive industrial structure by European standards, since the share of consumer industries among E.E.C. members ranges from about one fourth for West Germany to only one third for Belgium and Holland (Giilten Kazgan, Ortak Pazar ve Tiirkiye [The Common Market and Turkey] [Istanbul: Gercek Yayinevi, 1970], p. 145).

48 — Bases of Politics

TABLE 2-6 The Structure of Manufacturing in State and Private Sectors, 1955-71 (Firms, Employment, and Value Added by Industrial Sector) (Percent) Firms

Employment

Value Added

1955

1963

7977

1955

7965

7977

1955

7965

7977

State Sector Consumer goods Intermediate goods Investment goods

54.3 28.6 17.1

55.8 30.4 13.8

55.3 25.7 18.8

60.8 18.5 20.7

55.8 24.6 19.6

51.9 31.6 16.3

78.0 13.2 8.1

50.2 37.6 11.7

49.0 44.9 5.6

Private Sector Consumer goods Intermediate goods Investment goods

58.5 35.3 6.2

59.6 26.6 13.8

46.0 30.8 17.8

72.8 23.5 3.7

60.5 26.7 12.8

47.7 28.5 20.6

63.8 30.5 4.7

52.2 28.4 18.0

39.6 36.0 21.9

SOURCE: State Statistics Institute, Statistical Yearbook, 1964-65 (Ankara, 1965), p. 456; State Statistics Institute, Statistical Yearbook, 1977 (Ankara, 1977), pp. 207-209.

ucts, machinery, and transportation equipment.14 This structural change in manufacturing is also reflected in long-term changes in the volume and composition of Turkish imports. Between 1950 and 1970 the dollar value of imported goods nearly tripled and the share of consumer goods in total imports declined from 20 percent to only 5 percent.15 Thus, rapid import substitution and industrial diversification have resulted in near self-sufficiency in consumer products and far greater reliance on imported raw materials and investment goods. Another important aspect of Turkey's industrial transformation has been the rapid development and changing structure of large-scale private manufacturing. Table 2-6 presents changes in the relative importance of the three general categories of manufacturing in the state and private sectors. The general tendency is for the originally more differentiated structure of state manufacturing to remain fairly stable while the originally more primitive structure of private manufacturing changes substantially and eventually converges with that of the state sector by 1971. The major change in private industry is from consumer to investment goods; hence, the 14 Relative changes in production tended to be slightly faster and larger than changes in employment, indicating again that industrial diversification has also involved greater capital intensity and productivity. This is particularly evident in the intermediate categories of petrochemicals and basic metals, where very large increases in value added have been accompanied by only modest increases in employment. "World Bank, p. 221.

Socioeconomic Bases — 49

relative importance of the three industrial categories becomes very similar in the public and private sectors.16 The structural transformation of private manufacturing in Turkey is even more evident in terms of the increasing concentration of ownership, the growth in the size of private firms, and their greater capital intensity and productivity. Tables 2-7 and 2-8 present a more detailed comparison of changes in the structures of private and public manufacturing.17 These data clearly demonstrate the consistently much larger scale of public sector industry, its high concentration of labor and capital, and the very large increases of productivity in several consumer and intermediate categories. They also contain some important evidence of a trend toward greater organization and rationalization of the relatively fragmented private sector. First, it appears that private ownership has become considerably more concentrated. A clear process of consolidation and merger was evident during the 1960s in several important private branches such as food and tobacco processing, textiles, wood, chemicals, and construction materials. Second, the size of the average private firm has grown rapidly, so that by 1971 four industrial categories averaged over one hundred workers per firm and six other categories averaged over seventy workers per firm. As a general rule, the degree of growth in the scale of production tended to increase with the degree of industrial specialization and reliance on assembly line methods. The average size of a firm tripled in intermediate categories and quadrupled in investment categories. Finally, rapid productivity increases in all types of categories indicate an overall trend toward greater capital intensiveness in private industry. In fact, productivity in the most rapidly expanding areas of private manufacturing—machine building and transportation equipment—has surpassed that of the generally more organized state enterprises. Taken together these data provide a general overview of the impressive 16

Regarding the distribution of value added, on the other hand, we observe some evidence of a recent divergence between the state and private sectors. In both sectors consumer industries' share of total production has dropped sharply, but the major shift in production has been toward investment goods in the private sector and toward intermediate goods in the state sector. This suggests that the growth and diversification of private manufacturing have been accompanied by a gradual alteration of the original division of labor between public and private industry in Turkey's mixed economy. As the relative importance of investment goods increases, the state is assuming greater responsibility for supplying industrial raw materials to large private producers. 17 Due to a lack of data on the degree of capital concentration or market domination for individual enterprises, the rather crude indicator of the concentration of ownership employed here is simply the number of manufacturing establishments in each category of industry. Size is measured in terms of the average number of workers per firm and productivity is indicated by the amount of gross value added per worker.

Food Beverages Tobacco Textiles Clothing Wood Furniture Paper Printing Leather

1271 33 188 856 20 233 33 12 108 53

1955

1955 16 15 160 59 10 10 12 17 23 19

1971 1093 55 31 812 115 144 109 61 190 92

822 24 40 689 32 76 40 24 115 63 43 29 70 105 19 32 23 25 35 32

1963

Workers per Firm

1963

Firms

41 94 203 118 40 36 26 49 43 40

1971 7 11 2 7 9 7 9 15 10 10

1955

1971 42 66 24 42 31 43 28 85 72 45

14 25 13 14 14 15 17 20 29 16

T.L) 1963

(1,000

Value Added per worker

TABLE 2-7 Concentration and Productivity in Manufacturing, 1955-71—Private Sector (Number of Firms, Workers per Firm, and Value Added per Worker by Industrial Sector)

253

719 376

1607

820

2106 1411

4571

2775

4106

2368 1434

113 250 8 282 162 320 227 125 148

99 153 1 119 29 209 80 47 40

72 475 2 388 58 88 102 8 55

11 35 45 17 26 18 32 29 27 17 15 18 24

7 11 10 6 12 10 6 20 6 7 6 9 9

67 85 100 98 67 73 79 103 89 72 74 66 82

60 30 20 120 31 57 49 119 55 65 70 68 62

72 21 10 28 10 31 13 25 11 34 42 22 19

42 64 54

58 86 21 54 59 45 58 54 63 50

SOURCE: State Statistics Institute, Statistical Yearbook, 1964-65 (Ankara, 1965), p. 456; State Statistics Institute, Statistical Yearbook, 1977 (Ankara, 1977), pp. 207-209.

Consumer Intermediate Investment

Rubber Chemicals Petrochem. Nonmet. mins. Basic metals Metal prods. Nonelec. mach. Elec. mach. Trans, equip. Total

Food Beverages Tobacco Textiles Clothing Wood Furniture Paper Printing

1963

63 18 27 22 2 18 — 2 6

1955

24 13 18 23 2 10 — 2 4

Firms

103 18 29 30 3 23 — 6 9

1971

200 — 3450

150

1700

125

1380 1150

1120

850 210 —

460 217 515

1963

495 207 792

1955

Workers per Firm

189

2200

735 182 —

1050

374 256 890

1971

21 30 11 8 9 2 — 7 4

1955

(T.L.

25 51 19 14 11 18 — 29 17

1963

1,000)

Value Added per worker

TABLE 2-8 Concentration and Productivity in Manufacturing, 1955-71—State Sector (Number of Firms, Workers per Firm, and Value Added per Worker by Industrial Sector)

70 246 213 37 28 22 — 35 27

1971

80 42 25

Consumer Intermediate Investment

132 72 33

— — 13 6 18 9 14 3 1 15 237 183 85 62

— — 11 7 22 7 12 19 2 29 330 710 405 780

— — 200 167 200 890 530 330 — 1170 605 600 490 840

415 300 316 1185 650 400 100 1000 590 558 735 520

663 642 441 3320 550 205 550 712 600 13 7 4

6 10 8 3 9 — 3 10 22 35 15

191 21 32 20 40 11 11 25

101 151 36

70 1163 48 115 36 108 35 23 106

SOURCE: State Statistics Institute, Statistical Yearbook, 1964-65 (Anki ι, 1965), p. 456; State Statistics Institute, Statistical Yearbook, 1977 (Ankara, 1977), pp. 207-209.

— — 9 3 5 9 11 3 — 11 156

Leather Rubber Chemicals Petrochem. Nonmet. mins. Basic metals Metal prods. Nonelec. mach. Elec. mach. Trans, equip. Total

54 — Bases of Politics

growth and structural transformation of Turkish industry, and especially of the rapid development and increasing importance of private manufacturing. Nevertheless, the pattern of rapid industrialization in Turkey has failed to alleviate some longstanding structural weaknesses and has, in fact, promoted additional problems that will seriously hinder implementation of the new and much more ambitious long-term strategy of economic development. For example, the greater diversification and rationalization of manufacturing have not broken the traditional compartmentalization of the public and private industrial sectors. Turkish industry still is not really ' 'mixed'' either in terms of serious intersectoral competition within specific categories of manufacturing or in terms of collaborative enterprises that pool public and private capital. In this sense, both competition and collaboration between the sectors are still quite rare.18 This compartmentalization of industry is reinforced by the highly specialized nature of Turkey's major financial institutions, which tend to force credit flows through rigid channels both within the industrial sector and in the economy as a whole. Long-term financing is managed primarily by a number of "special law" banks, which specialize in lending to particular sectors and conduct relatively few transactions with one another.19 Just as the compartmentalization of the public and private sectors has inhibited domestic competition, strong protectionism has insulated Turkish industry as a whole from the pressures of foreign competition. Consequently, industry has preserved its orientation toward costly, inefficient, and relatively small-scale production for the home market. The system of import controls has brought chronic problems in allocation and high costs for both producers and consumers. Frequent shortages of raw materials have caused production delays and a common underutilization of capacity. Recurrent bottlenecks have forced producers to assume higher inventory 18 Food processing and wood are the only categories that consistently have been divided fairly evenly between public and private enterprise, and clothing and transportation equipment are the only categories in which there has been any substantial trend toward increasing competition. AU other categories remain firmly dominated by either public or private enterprise and the dominant sector commonly accounts for 80 percent or more of total production. 19 For example, the State Investment Bank lends exclusively to public enterprises and draws most of its resources from workers' compulsory insurance funds. Two other banks, the Turkish Industrial Development Bank and the Industrial Investment Credit Bank, provide medium and long-term credit to private industry. But given the limited nature of Turkey's private capital market, it is estimated that bank credit accounts for less than 40 percent of private industrial investment. Most private firms are still very closely held, so that the sale of corporate securities plays a negligible role in their financing. Over 60 percent of private industrial investment is financed directly from the entrepreneur's own profits (World Bank, p. 19).

Socioeconomic Bases — 55

costs and drawn them into continuous conflicts with traders, hoarding middlemen, and black marketeers. The high costs and inefficient nature of Turkish manufacturing also are attributable to the existence of a large number of scattered firms operating at uneconomic levels of production. Small-scale enterprises, employing fewer than ten workers, still account for 30 percent of total employment in manufacturing and about 25 percent of production. Even larger-scale Turkish industry remains highly fragmented despite considerable consolidation in the private sector since 1960. The continuing predominance of the small firm in Turkey's industrial structure is particularly striking when contrasted with the scale of production units in various E.E.C. countries.20 Fragmented and poorly coordinated production continues to characterize even some of the most important and rapidly expanding branches of manufacturing. For example, the fast growing automotive industry is still basically an assembly operation in which numerous and scattered subcontracting firms contribute different parts to be added to imported engines and transmissions. While the locally produced components of transportation vehicles have risen from about 20 percent to about 60 percent, the units of production often remain small and dispersed and the process of production is often poorly integrated. Problems of scale, productivity, and marketing are particularly serious in Turkey's important textile industry.21 Even in the state sector, where industrial enterprises enjoy far greater advantages of scale, high production costs and low profitability have been persistent problems. Planners have continued to hope that the State Economic Enterprises would generate substantial surpluses to finance future public investment, but their profits have not been commensurate with the vclume of their operations or the capital resources used. Throughout the planning period, savings and investments have fallen below their targets in the state industrial sector while exceeding their targets in the private 20 Less than 7 percent of large Turkish firms employ 100 workers or more. Firms of similar scale constitute 40 percent of all industrial establishments in Italy, SO percent in France, Belgium, and Holland, and 60 percent in West Germany (Kazgan, pp. 161-162). 21 Textiles are still the leading manufacturing category in terms of employment and production. They have received the greatest portion of the government's new export incentives and have also benefited from the tariff liberalization provisions of the protocol with the E.E.C. They now account for the largest portion of Turkey's nonfood industrial exports and are acquiring much greater access to an expanding market in Europe. Turkey's high quality cotton and low labor costs give her textiles an advantage in Western markets that is even greater than that of her more distant Asian rivals. Yet, the realization of these considerable advantages is still limited by poor marketing efforts and by the low productivity (between SO and 60 percent of British and German levels) of small and underutilized textile plants (World Bank, p. 258).

56 — Bases of Politics

sector. As a result, public industries have increasingly required budgetary support to finance operating losses.22 The principal reason for the low profitability of public establishments has been the policy of maintaining rigid prices in the face of rapid inflation in the rest of the economy. This pricing policy has aimed at easing the short-term impact of inflation on consumers and private industrial producers, but it has actually contributed to further inflation by forcing heavy deficit financing through credits from the Central Bank. Losses of public enterprises, together with policies maintaining high agricultural price supports and subsidizing agricultural inputs, have become key contributors to a pattern of deficit spending and structural inflation.23 Given the nature and structure of industry that has developed with the encouragement of import substitution policies, it is not surprising that Turkish manufacturers have tended to direct production toward the highly protected and more profitable home market and to neglect the development of industrial exports. Since 1950, protection and subsidies have been largely responsible for the rapid development of private industry, but they have also contributed to serious price distortions. The large differences between domestic and export prices have generally discouraged industrial exports and new investment in export industries. The already poor international price competitiveness of Turkish industrial goods has deteriorated even 22

Actually the performance of state manufacturing probably is not as poor as is often assumed. There is a very wide range in the profitability of the various public enterprises, with most of the total deficit resulting from the large losses of railways, maritime transport, and coal mining, while several manufacturing concerns show modest earnings of about 10 percent (ibid., pp. 90-95). Even this level of efficiency is considerable, since most public enterprises have been used as instruments for achieving such social and political goals as developing backward regions, providing employment, and ensuring the availability of many consumer goods and raw materials at stable prices. The pursuit of these extraeconomic goals has created several common and persistent problems among all varieties of public enterprise— political interference with management and a serious loss of technical personnel to the private sector, relatively high labor costs and padded payrolls, and high transportation costs to and from some remote locations. 23 The financial difficulties of state industry have reached a critical point for two major reasons. First, planners are placing increased emphasis on public industrial investment, including very large allocations to highly capital intensive sectors such as steel, where investment costs are particularly high, and petrochemicals, which are expected to be carriers and accelerators of technological development. Second, this more ambitious investment strategy is being pursued while future deficits are forecast for the workers' compulsory insurance fund, which has been the traditional capital resource of the State Investment Bank. Thus, it is very unlikely that public sector investment can approach its planned levels unless there is a major transformation of the operational policies and financing practices of state enterprises. Several proposals have been formally adopted for reforming the public enterprises, but as yet little has been done aside from consolidating a portion of their debts and limiting their ability to borrow (ibid.).

Socioeconomic Bases — 57 further during periods of rapid domestic inflation due to the common policy of maintaining fixed rates of exchange and permitting the overvaluation of the lira.24 In general, then, as Turkish industry has grown it has come to display more clearly the greater burden of its own structural weaknesses. The new long-term strategy of economic development outlined in the Third FiveYear Plan focuses on trying to expand and reshape Turkish industry in order to overcome these weaknesses in the shortest possible time. Having accepted the self-imposed deadline of 1995 for Common Market membership, the framers of the new strategy in effect call for compressing within a period of about twenty years the kind of major industrial transformation that required over fifty to one hundred years in Western Europe. The new emphasis on rapid export-oriented industrialization is likely to intensify interest conflict between industry and other economic sectors. It is also likely to increase the already large disparities in the distribution of the benefits of economic growth across different geographic regions and social classes.

INEQUALITIES IN REGIONAL DEVELOPMENT AND INCOME DISTRIBUTION

Several nationwide surveys, including those conducted by the State Planning Organization and Hacettepe University, indicate the persistence of large and longstanding disparities in the levels of social and economic development of different regions. Although there are insufficient data to describe confidently the long-term trends in these disparities, there is considerable evidence that the rapid urbanization and industrialization of the last three decades have widened differences in development and wealth both between regions and within them. Studies by the State Planning Organization have used over fifty social and economic indicators to construct a composite index of development for ranking individual provinces. In general, the distribution of development and modernization follows a clear regional pattern, according to which the Marmara and Aegean provinces are the most developed, the eastern region is the most backward, 24 Several governments have attempted to reduce the strong dependence of Turkish industry on the domestic market by enacting drastic currency devaluations in 1958, 1970, 1978, and 1980, as well as by offering an increasingly generous set of incentives and subsidies to industrial exporters since the mid-1960s. The recent devaluations have produced upsurges in exporting, especially of textiles, leather products, beverages, chemicals, and petroleum products. But as yet Turkey still has no major export industry. Between 1960 and 1970 industry's share of total exports actually declined from 25 percent to only 17 percent, and by 1972 exports still accounted for less than 3 percent of the total value of manufacturing production (ibid., p. 221).

58 — Bases of Politics

and provinces along the Black Sea and Mediterranean coasts and in the central Anatolian plateau occupy an intermediate position.25 The Hacettepe University study considered differences in development between urban and rural settlements as well as between the rural populations of different regions. While the study found large inequalities along both of these dimensions, it observed that the most pronounced disparities were between the urban and rural populations within regions rather than between different rural populations. The Hacettepe group explained these findings by attributing the relatively narrower rural regional differences to "survivals of the cultural and historical past" and attributing the greater urbanrural differences to "contemporary and on-going processes of industrialization and economic development." It concluded by predicting "a progressive contraction in rural regional differences" as traditional cultural factors become less salient and "a progressive aggravation of rural-urban differentials" (and presumably also of regional differentials) in the intermediate phase of economic development.26 During the one-party period, especially when etatist economic policy was predominant, industrialization was conceived as a key means for transmitting modernization throughout the Anatolian hinterland and integrating it into an expanded national economy. The dispersion of statesponsored industrial enterprises promoted regional growth centers in about thirty-five cities, so that Istanbul and Izmir, which enjoyed a virtual monopoly on industrial activity during the early 1920s, accounted for only one third of Turkey's total manufacturing production by 1950. During the 1950s the dispersion of public industrial investment continued, but was offset by the much stronger countertendency of growing private industry to reconcentrate resources in the already most developed areas. Between 1950 and 1960 private entrepreneurs received generous subsidies and credits from the new Industrial Development Bank, which was established to channel American economic assistance to Turkey's rapidly expanding private business sector. Of the 401 projects it helped to finance, 63 percent were in the Marmara region (205 in Istanbul alone) and another 12 percent were in the Aegean (39 in Izmir).27 Since 1960, Turkish planners have professed great concern in slowing or reversing this tendency toward regional polarization. But neither the pattern of state investment nor the new incentives offered to private industrialists for locating in relatively back25 Devlet Planlama Teskilati, Tiirkiye'de liter habariyle Sosyo-Ekonomik Gelismelik Endeksi [Socio-economic development index in Turkey by provinces] (Ankara: Devlet Planlama Teskilati, 1970). 26 K. S. Srikantan, "Regional and Rural-Urban Socio-Demographic Differences in Turkey," Middle East Journal 27 (Summer 1973):275-300. 27 World Bank, pp. 163-164.

Socioeconomic Bases — 59

ward areas have reduced the general flow of labor and capital from less developed regions to more developed ones. Munro, Robinson, and Shorter have examined Turkey's massive internal migration and pointed to its long-term contribution to widening disparities in regional population structures and dependency ratios. Munro's analysis of the 1965 census reveals twenty-five major routes of migration, through which more than 800,000 persons changed residence from one province to another. About 60 percent of the migrants settled in Istanbul and about 25 percent settled in Ankara. The particularly strong attraction of Istanbul is indicated by the report in the 1965 census that only 963,000 of its 2,155,000 inhabitants were born in that province.28 Robinson notes that although the majority of emigrants are agricultural day laborers, the emigrant population also generally includes the most capable laborers between fifteen and forty years of age, leaving the poorer region's work force with an even more burdensome dependent population. The large urban centers—principally Ankara and Istanbul—are magnets for the more dynamic elements of the population elsewhere, namely, the ambitious, the wealthy, the intellectual, the skilled, the politician, the creative, the frustrated, even the criminal. The growth element is thus being siphoned off the rest of the country, and the disparity between these centers and the balance of the country widens.29 Shorter argues that such migration is particularly harmful in already depressed areas, since their dependency ratios are also rising as a result of higher birth rates. Consequently, he emphasizes the increased importance of family planning to prevent a further deterioration of savings and investment potentials in such regions.30 Suzanne Paine has suggested that the recent flow of Turkish migrant labor to Europe may also be contributing to a further deterioration in the regional distribution of income. This is so, she contends, because a majority of those who work abroad come from the richer provinces originally and also because a significant proportion of those from poorer areas have moved to richer ones on their return, taking their repatriated earnings with them.31 While migration has reduced the ability of less developed regions to generate savings, the proliferation of local branches of national banks has 28 Gustave Schachter, Regional Dualism in Turkish Development (Boston: Northeastern University Department of Economics, 1970), pp. 10-11. 29 Richard Robinson, High Level Manpower in Economic Development: The Turkish Case (Cambridge: Harvard University Press, 1967), pp. 28-29. 30 F. C. Shorter, "The Economics of Turkish Population Policy Since the Founding of the Republic," Hacettepe Bulletin of Social Sciences and Humanities 1 (June 1969):63. 31 Paine, pp. 102-106.

60 — Bases of Politics TABLE 2-9 The Regional Distribution of Credits from the Turkish Banking System, 1971

Marmara Aegean Mediterranean North Central South Central Black Sea East Central North East South East

Total Credits per capita (T.L.)

Credits as Percent of Deposits

3250 1282 1165 3996 478 1016 252 267 349

91 98 126 178 64 123 43 61 141

SOURCE: World Bank, pp. 177-178.

encouraged the flight of capital. Savings accumulated in the less developed provinces are transferred to areas of greater profitability and lower risk, depriving the poorer regions of a large part of their own scarce and independently created capital resources. Table 2-9 presents figures from the Turkish Banking Association regarding the regional distribution of bank credits and the ratio of credits to deposits.32 The Marmara and North Central regions enjoy an overwhelming predominance in terms of per capita credit and together receive about two thirds of all credit extended. When credits are compared with deposits, it is clear that the flow of funds through the banking system has been away from the poorer regions of the east and toward the more developed provinces of the Mediterranean, the Black Sea, and north central Anatolia.33 Schachter has argued that even much public investment tends to reinforce the "regional dualism" in Turkish economic development despite the increased emphasis planners are placing on geographically balanced growth. 32 In both sets of data the share of the North Central region is undoubtedly exaggerated somewhat, since credits for nationwide public investments such as transportation and energy are oftenregisteredfor Ankara, where their management is located. 33 The major exceptions to this pattern are the South East and South Central regions. The South East has benefited recently from state-sponsored investment, though only to a very modest degree as is indicated by that area's level of per capita credit, whichremainsone of the lowest in the nation. The South Central area, on the other hand, is the only moderately developed region that has experienced a net capital loss. Businessmen in these provinces have taken the lead in organizing their colleagues in the more disadvantaged eastern regions to demand new legal limitations on private bank transactions between provinces.

Socioeconomic Bases — 61 He characterizes Turkish regional development policy as "vacillating and confused" and shows that during the 1960s public investment was concentrated in Istanbul, Ankara, and Zonguldak, while the poorest areas benefited very little. He notes that the new Keban Dam on the Euphrates River—one of Turkey's largest public projects in decades and the largest ever undertaken in the east—was constructed primarily to provide electrical power to Istanbul and Ankara rather than to the eastern provinces and was not planned for local irrigation or land reclamation. In Schachter's view, "It seems that even an investment which occurs in the East is only for the benefit of the West." 34 Schachter is also critical of the numerous regional studies for planning "growth poles" and establishing concentrations of industrial activity in selected locations, but he finds this policy generally preferable to the still ineffective alternative of subsidizing the few private entrepreneurs who are willing to relocate. He argues that the kind of investment projects that are planned for the east are those that are least likely to create substantial new employment, since they are primarily in capital-intensive heavy industry. More important, he claims that policy makers have become discouraged about regional development and that the actual implementation of these "growth pole" plans indicates a shift toward further encouraging industry in the three largest cities and in the booming East Marmara cities of Bursa, Izmit, and Adapazan in order to realize the highest short-term returns.35 The high concentration of economic activity can be appreciated by observing the relative contributions of Istanbul and its four East Marmara neighbors—Kocaeli, Sakarya, Bilecik, and Bursa—in several economic fields (see table 2-10). In 1965 these five provinces included about 12 percent of the total population and only 3 percent of the national territory. Yet they also accounted for about one quarter of the gross domestic product, about one third of total value added in trade, banking and insurance, and other services, and one half of all large-scale manufacturing. When the data on industrial activity are broken down by sector, the extreme concentration of private manufacturing in Istanbul is particularly clear. Whereas Istanbul contains only about 15 percent of state industrial activity, it accounts for one half of all private sector firms and workers and two thirds of private industrial output. There is also some indication that the already enormous relative economic importance of the East Marmara region has increased even further since 1960. Writing even before the construction of the costly and highly controversial suspension bridge over the Bosporus, Schachter concluded that Turkish planners, in effect, had adopted an unannounced strategy of preserving and further dividing a dual economy. 34 35

Schachter, p. 23. Ibid.

62 — Bases of Politics

TABLE 2-10 The Relative Economic Importance of Istanbul and the East Marmara Subregion (Percent of Nationwide Totals) East Marmara

Istanbul

Population Labor force Value added per capita (Turkey = 100) Value added per worker (Turkey = 100) Value added by economic sector Agriculture Mining Manufacturing Construction Electricity, water Transportation Trade Banking, insurance Other services Shares of large manufacturing (1967) Firms Workers Wages & salaries Value added

1960

1965

1960

1965

6.8 5.8

7.3 6.4

12.2 11.0

12.5 11.1

165

208

143

173

208

258

159

190

1.6 2.5 25.8 12.7 27.5 14.9 12.3 22.9 18.0

1.1 .7 25.3 23.3 24.0 14.9 24.7 31.7 22.2

7.1 4.0 34.0 20.2 28.7 22.7 18.9 27.6 23.6

6.0 1.1 34.6 30.7 27.7 22.0 31.7 36.0 27.8

Private 46 50 64 65

State 13 15 18 15

Total 43 36 41 39

Total 51 45 50 50

SOURCE: World Bank, p. 457; Sabahaddin Zaim, Istanbul Sanayi Bolgeleri [The Istanbul industrial regions] (Istanbul: Istanbul University Economics Faculty, 1971), p. 95.

"By concentrating efforts in areas where development has already occurred, the poorer areas are further drained of their potential in terms of existing resources—labor and capital—decreeing their perpetual backwardness. This corresponds to an overt policy of having two countries: one prosperous with a modem industrial sector; the other a primitive agricultural society."36 Besides increasing disparities in regional levels of development, rapid 36

Ibid., p. 32.

Socioeconomic Bases — 63 TABLE 2-11 Distribution of Personal Income and Landownership, 1963-73 Percent Personal Income Percent Population Poorest 20 Next poorest 20 Middle 20 Next richest 20 Richest 20 Size of Landholdings (Hectares) Landless 1-20 21-50 51-100 101-200 201-500 500 +

1963 (SPO)

1968 (Boratav)

1968 (Hacettepe)

1973 (SPO)

4.5 8.5 11.5 18.5 57.0

4.2 6.4 10.7 17.7 61.0

3.0 7.0 10.0 20.0 60.0

3.5 8.0 12.5 19.5 56.5

Percent Owners

Percent Land

1963

1968

1973

1963

1968

1973

9.0 36.8 25.5 16.4 8.6 3.0 .5

17.5 36.1 28.1 11.5 5.0 1.3 .7

19.4 35.5 22.5 13.4 5.6 2.1 .7

0.0 11.3 17.7 22.2 22.2 15.9 10.7

0.0 9.4 24.7 20.4 16.8 9.4 19.3

0.0 8.4 17.9 22.6 19.5 16.2 15.4

SOURCE: State Planning Organization, Gelir Dagihmi Araqtirmasi [Study of income distribution] (Ankara, 1963); Korkut Boratav, YUz Soruda Gelir Dagilimi [Income distribution] (Istanbul: Gercek Yayinevi, 1969), p. 194; Tuncer Bulutay, Hasan Ersel, and Serim Timur, Tiirkiye'de Gelir Dagihmi [Income distribution in Turkey] (Ankara: Ankara University Political Science Faculty, 1971), p. 16; World Bank, pp. 432-433; State Planning Organization, Gelir Da|iltmi Arastirmasi [Study of income distribution} (Ankara, 1973), p. 24; State Planning Organization, Kirsal Refah Politikasi [Rural prosperity policy] (Ankara, 1973).

urbanization and industrialization in Turkey also have been associated with a highly unequal distribution of personal income (see table 2-11). Several nationwide studies of income distribution in Turkey since 1963 have concurred in estimating that the richest 20 percent of the population received about 60 percent of total income, while the poorest 20 percent received only about 4 percent.37 Taken together, the four income studies indicate 37 For the nation as a whole, each study derived a coefficient of inequality (the Gini coefficient) of about .56. This is roughly comparable to the highly unequal distributions of such Latin American and Middle Eastern countries as Brazil, Ecuador, Iran, and Lebanon. It is a degree of inequality that is somewhat higher than the levels of Western European

64 — Bases of Politics a slight deterioration in the relative position of the poorest group, but it is not clear whether the greatest relative improvement has been in the position of the richest or the second richest group. A study of Turkish economic development by the World Bank reaches similar conclusions about the probable relationship between greater prosperity and greater inequality. It may be that inequality has, in fact, increased over the last few years as a result of rapid industrialization, urbanization and inflation, though no definite statements can be made regarding trends. But it is also likely that the proportion of the population below an absolute poverty line has, in fact, decreased as a result of economic growth, emigration and the government's social policies.38 Income is far more unevenly distributed in the agricultural sector than in the nation as a whole, and is evidently becoming even more skewed in view of recent trends in the distribution of landownership. In 1973 the largest landholders constituted only 3 percent of all farm owners, but they controlled about 30 percent of all cultivated land and received 33 percent of total agricultural income. There is also some evidence of an increasing polarization of large and small holders between 1963 and 1973. During this period there was a substantial rise in the number of landless peasants and agricultural wage laborers, a clear trend toward the fragmentation of small holdings under twenty hectares, and a general concentration of medium and large holdings over fifty hectares. Given the generally close correlation between landownership and agricultural income, it is very likely that inequalities of wealth are growing among the rural population. A general view of the degree of income inequality between and within various occupational categories is provided by the Hacettepe survey, which classified eleven occupations into four subgroups. Average annual income was found to be about T.L. 19,300 for large industry, trade, professions, and esnaf; T.L. 15,000 for higher and lower civil servants; T.L. 9,400 for technicians and skilled and unskilled workers; and T.L. 8,100 for farm operators and laborers. Within these categories inequality was lowest among public employees and wage earners; higher civil servants earned about twice as much as lower civil servants, and technicians and skilled workers earned about twice as much as unskilled workers. Inequality was far greater among agriculturalists, with operators receiving over four times the income of nonowners, and it was greatest among the commercial and service nations and of most other Latin American and Middle Eastern nations. It is considerably higher than the levels of the United States, Eastern Europe, and the Scandinavian countries. 38 World Bank, p. 175.

Socioeconomic Bases — 65 occupations, with large businessmen and professionals earning five to seven times as much as the esnaf.39 It is, of course, difficult to determine with any accuracy long-term changes in the relative status of these occupational groups or the varying degrees to which they have benefited from economic development. Even so, the wide disparities in growth rates of different economic sectors since 1950 and the clearly differential impact of inflation and taxation make it possible to form some tentative conclusions about the kind of social restratification that has occurred in the context of uneven development in Turkey. Agricultural income probably has experienced the greatest relative deterioration despite persistent government efforts to maintain or improve its absolute level through low taxation and high price supports. This is due to the substantially lower growth rates of real per capita output in agriculture relative to other economic sectors and the sharp decline in the terms of trade for agricultural exports during periods of rapid inflation. Public employees have suffered from a recurrent pattern of periodic declines in real income followed by tardy and only partial adjustments of their fixed salaries to match the rate of inflation. This has resulted in a growing exodus of technical and professional personnel to the private sector and in demands for the right to unionize and engage in collective bargaining. The recent "reform" of the civil service structure has ignored those demands in favor of substantial pay raises, whose principal effect was to accelerate inflation further. Wage earners shared civil servants' vulnerability to high taxation and inflation during the 1950s, but since I960 they have benefited from new minimum wage regulations and the much greater bargaining power of labor unions. Since 1963 the real wages of organized workers covered by social insurance have risen at about the same pace as productivity in manufacturing and slightly faster than inflation. Comparable data are not available for unorganized and uninsured workers, but they are very unlikely to have fared as well as their organized counterparts. The greatest relative gains in income probably have been among commercial and professional groups, since they have been the principal beneficiaries of the redistributive impact of inflation while also being most able to escape the redistributive effects of taxation. In the first place, Turkey's income tax structure is only moderately progressive, taxing the poorest brackets at a rate of about 13 percent and the highest brackets at a rate of only 20 to 33 percent.40 In the second place, tax laws have always provided numerous loopholes and opportunities for evasion to entrepre39 40

Bulutay et al., pp. 119-125. World Bank, p. 177.

66 — Bases of Politics

neurs, self-employed professionals, and esnaf. It is perhaps most difficult to generalize about long-term changes in the relative position of the esnaf, since the nature and composition of that group has changed greatly with the recent advances and changes in Turkish capitalism. At this point we can note that the esnaf, perhaps even more than agriculturalists, appear to have become polarized between those capable of consolidating their enterprises on a larger scale and those who abandon their trades to become lower-level civil servants or wage earners. In the most general terms, then, it is possible to summarize the impact of uneven economic development on social stratification in the following manner. The greatest relative improvements in income have occurred among commercial-professional groups and the greatest relative deterioration has occurred among agricultural groups. These are also the sectors within which internal polarization and income inequality have increased the most. In the meantime, public employees and wage earners have been running fairly close races with inflation, in which poorly organized civil servants have tended to lag behind while better organized workers have been able to hold their own and make occasional advances.

THE " N E W STRATEGY" of LONG-TERM DEVELOPMENT AND GROWING INTEREST CONFLICT

As uneven development has reinforced imbalances and inequalities between economic sectors, social classes, and geographic regions, the perception of interest conflict has sharpened between groups on the rise and groups in decline. This increasing perception of interest conflict has provided a major stimulus for the rapid proliferation of associational organization and counterorganization throughout every economic sector and geographic area. Both increasing interest conflict and its greater expression in associational interest groups have encouraged a recent fragmentation and realignment of the national party system. This overview of Turkey's social and economic transformation will be concluded by summarizing the new longterm strategy of economic development and indicating briefly how it is likely to increase the intensity of interest conflict during the next two decades and, therefore, to impose still greater strains on the political system. The Third Five-Year Plan, announced during the celebration of the republic's fiftieth anniversary and highly publicized as the beginning of Turkey's "New Strategy" of economic development, was both more ambitious and more realistic than its predecessors.41 It was more ambitious 41

The First Plan placed nearly equal emphasis on accelerating economic growth and on

Socioeconomic Bases — 67

because it was not merely a short-term program but the first phase of a new long-term development strategy aiming at a rapid economic transformation that will bring Turkey by 1995 to the level of development reached by Italy in 1970, thereby allowing full Turkish membership in the European Economic Community. The Third Plan was more realistic than the other plans in two important respects. First, it recognized that achievement of its long-range goals required not merely a stronger emphasis on industrial growth but also a major restructuring of Turkish industry, large imports of advanced technology, and the further coordination of other economic sectors in accordance with industry's growing needs for investment. Second, the new plan also recognized that the achievement of its economic objectives would require a de-emphasis and possible postponement of several social objectives. Its relatively blunt demand for long-term "sacrifices" to achieve higher rates of savings, investment, and productivity indicated a much greater willingness to confront the need for trade-offs between the conflicting priorities of accelerated economic growth and greater social equity.42 A general overview of the kind of structural transformation envisaged in the new plan can be seen in a summary of its major macroeconomic targets. The plan conceived of the period between 1973 and 1995 as a critical phase during which Turkey would undergo the transition to an urban-industrial society. Rapid industrialization was expected to raise annual per capita income about five times, from $320 to about $1,500. By 1995 industry's share of the gross domestic product was expected to rise from 25 percent to 37 percent, while agriculture's share was expected to drop from 26 percent to only 12 percent.43 These projected changes in the structure of production were matched by equally important changes in the achieving greater "social justice." It was particularly concerned with promoting a more equal distribution of increasing national prosperity in order to rectify the inequities between different regions and social groups that had worsened during the period of unplanned economic growth in the late 1950s. The Second Plan tended to view the distribution of the benefits of economic growth as unproblematic, assuming that accelerated development would lead to substantial increases in the absolute living standards of all groups and regions and, therefore, reduce the salience if not the magnitude of relative differences in wealth. Particular emphasis was placed on providing incentives to the private industrial sector as the pace setter in creating greater national prosperity from which all other groups would ultimately benefit, directly or indirectly. 42 Yalcin Kiiciik, "Ucimcu Bes Yilhk Plam" [The Third Five-Year Plan], Yanki, 25 September 1972; Yalcm Kiicuk, Planlama, Kalhnma ve Tiirkiye [Planning, development, and Turkey] (Istanbul: Gercek Yayinevi, 1971), pp. 237-284. 43 Manufacturing output was projected to increase at over 12 percent per annum, with the intermediate and investment sectors growing at 14 and 17 percent, respectively, or more than twice the rate of the consumer sector. By the 1980s the share of the nonconsumer sectors in total manufacturing output was expected to grow from 49 percent to about 70 percent.

68 — Bases of Politics rate of savings and the structure of investment. The rate of marginal savings was planned to more than double from 18 percent to over 40 percent by the 1980s, reaching about one fourth of the gross national product. By 1995 manufacturing's share of total investments was expected to rise from 37 percent to 50 percent, about four fifths of which would be allocated to the intermediate and investment sectors. The shift from import substitution to the encouragement of more export-oriented industry was also expected to produce a major change in the composition of foreign trade. The share of manufactured goods in total exports was expected to rise from 17 percent to 42 percent, while investment goods were expected to rise from 55 percent to 70 percent of total imports. Some brief indication of the social costs that were projected as a necessary and acceptable price of this industrial transformation can be seen in the growing neglect of employment and social services. The rapid introduction of advanced technology for the production of capital goods and industrial raw materials was expected to boost unemployment from its already high rate of 11 percent to more than 15 percent, with no decline being projected until after 1987. A further sign of the shifting priorities of Turkish economic planning was the gradually declining proportion of total investment targeted for basic social services. The share of investment to be allocated for housing, education, and health dropped from 29 percent in the First Plan to 26 percent in the Second Plan and then to only 22 percent in the Third Plan.44 A World Bank study of the Third Five-Year Plan and its long-term strategy generally characterized its goals as "ambitious but feasible," yet it also strongly criticized several provisions as likely to contribute further to Turkey's longstanding pattern of uneven development. Considering the development strategy and targets that Turkey has opted for . . . and the institutional framework within which the strategy and targets are pursued . . . strong tensions and disequilibrating forces are likely to appear and persist in several parts of the economy in the medium term. . . . While rapid economic growth is expected to continue in the medium term, various social problems may be accentuated.45 In concluding this discussion of the social and economic bases of interest group politics, it is useful to note those features of current developmental policy that, in light of the World Bank report, are most likely to heighten interest conflict in the next two decades. First, the greater emphasis on rapid and heavy industrialization will « World Bank, p. 32. « Ibid., p. 47.

Socioeconomic Bases — 69 preempt resources that could be used to increase production in agriculture and in more labor-intensive sectors that aim at creating more employment and improving the social environment. This is likely to lead to a further deterioration of the employment situation, an increase in the absolute gap between urban and rural living standards, larger areas of urban squatter housing, and a deterioration in the level of social services. Second, a continuation of rising incomes and rapid urbanization is likely to encourage an actual pattern of investment and production that is very different from what is planned. During the first two plans private demand was stronger than expected for consumer goods, housing, and services, so that investment and production exceeded their modest targets in these sectors while there were shortfalls in industrial performance. Similar deviations are certain to reappear unless future governments are able to prevent the overheating of the economy by implementing unpopular policies to curb demand and restrict private consumption. Third, there is a danger that intensive industrial investment may promote, not the development of competitive export industries, but merely a more sophisticated and expensive form of import substitution that will continue to require a high degree of protection. Investment in protected industries that have no chance of becoming competitive by 1995 would greatly increase the already serious conflict between different kinds of industrialists concerning the conditions of E.E.C. membership and could lead to a basic revision or even rejection of the proposal. Finally, one of the most serious problems of the new planning strategy is its inadequate provisions for mobilizing resources needed to finance the growing volume of investment. Financing difficulties are likely because the new plan greatly underestimated the growth of commodity imports and because Turkey's reserves of foreign exchange were depleted by the recent ban on the emigration of workers to Germany and the sharply rising cost of imported fuel. Preventing investment shortfalls in private sector industry is dependent on creating a private capital market and reforming the banking system to encourage more capital allocation to industry and less to shortterm commercial and real estate ventures. Both measures have been prevented for several years by the combined resistance of banking and trading interests. It is unlikely that investment shortfalls can be avoided in the public sector, where marginal savings have been targeted at an extremely high rate of 60 percent in contrast to the more realistic goal of 20 percent for the private sector. Previous plans also suffered from overestimations of the savings potential of the public sector and sought to compensate for the deficits of the State Economic Enterprises by relying on inflationary Central Bank credits and on "captive capital" accumulated in the workers' social insurance and pension fund. Several means are being pursued to narrow the gap between industrial

70 — Bases of Politics

investment targets and available capital resources. These include higher agricultural taxation, a new value-added tax on the commercial sector, and greater efforts to attract foreign investors, especially for technologically advanced projects. Perhaps the most important innovation is the growing interest in forced domestic savings that are being accumulated by several new compulsory social insurance programs.46 Although these funds are to be administered by various professional associations rather than by the government, they are clearly intended to support the establishment of new industrial enterprises that will be approved by the State Planning Organ­ ization. In other words, the recent proliferation of these insurance programs represents a major effort to establish an entirely new economic sector based on massive funds of captive capital that can be used to finance Turkey's ambitious long-term industrial investments. It is apparently an attempt to compensate for probable shortfalls in public savings by encouraging the leaders of the largest professional associations to participate directly in industrial development as investors and entrepreneurs. The prototype for this new associational entrepreneurship appears to be the Turkish armed forces' impressive experiment in military capitalism through the Army Mutual Aid Society (Ordu Υατάιηύαφηα Kurumu). OYAK was established soon after the 1960 coup and aimed originally at creating a pension program that would rescue career officers from the economic insecurity they had suffered during the rapid inflation of the 1950s. In fact, the military managers of OYAK have consistently held retirement benefits well below the rate of inflation in order to plow back the profits of their growing investments and quickly expand the fund's assets.47 Following this strategy, Turkish military commanders have transformed OYAK into the country's largest and most diversified industrial conglomerate. Its cur­ rent assets are estimated at over $300 million, with investments concen46 In recent years Justice Party and martial law governments have promoted the notion of comprehensive social insurance funds to be supported by compulsory contributions from all of the major occupational categories. They have revised and expanded the insurance program for workers, established new compulsory funds for civil servants, esnaf, and professionals, and proposed a similar program for agricultural producers. 47 "The Army Conglomerate," Time, 11 September 1972, p. 55; "Ordu Is Aleminde" [The army in the business world], Yanki, 11 June 1973, pp. 14-15; "Kalkinmanm Finansmani, HaIk Tasarruflan ve OYAK ornegi" [Financing development, mass savings, and OYAK's example], Yeni Ortam, 19 March 1973. All regular commissioned and noncom­ missioned officers in every branch of Turkey's armed forces—totaling about 80,000 men— are required to contribute 10 percent of their salaries to the fund. Upon retirement they receive what they paid in plus only 5 percent interest. Until 1970 OYAK's management insisted that interest payments be kept low in order to provide greater benefits for future generations of officers. Since then they have responded to growing dissatisfaction among lower-ranking members and raised payments to 7 percent and recently to 10 percent, but still nowhere near the increasing rate of inflation.

Socioeconomic Bases — 71

trated in the largest and most modern industrial sectors.48 While extending its control over a wide variety of large industrial enterprises, OYAK's management also has compiled one of the nation's most enviable investment records. The rate of return on their total investments was reported as 23 percent for 1972 and as averaging 17 percent per annum for the preceding twelve-year period. It is still too early to determine whether or not OYAK's brand of associational capitalism can be replicated in other occupational sectors. But perhaps the most interesting aspect of the OYAK experiment, and especially of the desire to extend it to all other major social groups, is that it represents the most obvious of several recent attempts by the state to enlist the support of the most important professional associations in implementing its policies of economic development. This involves an important expansion of the responsibilities of the largest associations, which are increasingly expected to operate as collective bargaining agents that promote collaborative class relations, as instruments of involuntary capital accumulation, and potentially as direct participants in investment and in the process of production. But this attempt to enlist the aid of professional associations in carrying out economic policy is creating serious problems for both government leaders and the leaders of Turkey's major interest groups. The decision to opt for a more severe type of uneven development is likely to heighten the tension between the elite's ambitious goals for rapid economic growth and mass demands for greater political participation and economic redistribution. In other words, the government's "New Strategy" of economic development appeared so certain to exacerbate interest conflict and to encourage its greater expression through associational interest groups that the government appeared also to require an additional "New Strategy" for managing social and political conflict if its goals were to be pursued without recourse to an overtly authoritarian regime. Such a new political strategy, in fact, began to emerge several years before the announcement of the far better publicized new economic strategy and was observable in the attempts of several governments to transform Turkey into an "organized democracy" by corporatizing associational life, especially in the major economic sectors. This amounted to a general policy of trying to promote semiofficial, compulsory, and hierarchically organized representational cartels while imposing greater controls and restrictions 48 Yeni Ortam, 19 March 1973. OYAK now owns controlling shares of three major automotive plants and distributorships, including one that assembles International Harvester trucks and tractors; two large cement plants; an insurance company; and a food-canning firm. It also owns 42 percent of Renault's new automobile factory; 40 percent of a fertilizer and pesticide distributorship; 20 percent of a large new petrochemical corporation; 8 percent of the state-run oil refineries; and 7 percent of Goodyear's Turkish subsidiary.

72 — Bases of Politics over the vast and competing pluralist network of private voluntary associations. The Justice Party, technocrats, and military commanders apparently hoped that by creating or strengthening such representational cartels they could induce key associational leaders to cooperate in implementing economic policy and in managing interest conflict by voluntarily limiting the political and economic demands of their memberships. Thus, the policy of corporatization aimed not merely at structuring and coordinating mass political and economic demands but at moderating and reducing them at a time when weak and unstable governments were unable and unwilling to satisfy them. In general, then, the new corporatist political strategy can be described as an attempt to resolve in a relatively noncoercive fashion certain tensions in Turkish political development that have arisen due to the rapid emergence and diffusion of associational interest groups and their tendency to express growing political conflict within a more differentiated social and economic structure. The principal aims of this strategy can be summarized as, first, the promotion of rapid economic growth at the expense of political participation and economic redistribution; second, the promotion of associational and functional channels of interest representation at the expense of party-electoral and plebiscitarian channels; and, third, the promotion of corporatist interest groups at the expense of pluralist interest groups in structuring and regulating associational life.

CHAPTER 3

Political Culture: Attitudinal Bases

have recognized the legitimacy of interest group organization and political activity only recently and tentatively. The leaders of the major political parties still have only halfheartedly attempted to accommodate the newly differentiated social and economic groups that have arisen as a result of their own developmental policies. Republican, Democratic, and Justice Party ruling elites have each regarded the growing tendency to express interest conflict through associational activity as a threat to their ability to retain control over the political consequences of rapid social and economic change. Further, while in power, each party has responded with attempts to maintain political control through some form of manipulation or generalized repression of associational life. While the pattern of uneven economic development in the context of liberal democracy explains the overall emergence of the art of association in Turkey, it does not explain the still limited legitimacy of interest group politics, the recurrent repression of associability, and the frequent attempts to channel it into structures that are susceptible to the manipulation of the state. The relationships between associations and the state have been more directly influenced by political culture and by the pattern of public policy toward associations. Political culture and public policy may be considered as key intervening variables that account for the often wide range of variance in the political influence of associations in a given political system or in different systems that have reached similar levels of social differentiation. More specifically, variations in group influence can be seen as dependent upon the degree to which political culture and public policy contribute to the development of pluralism or corporatism as alternative strategies for structuring associational interest representation in different systems or in different sectors of a single system. The concern here is with earlier but still relevant elite attitudes toward the relationship between state and society and toward legitimate interest representation.1 The thrust of the argument is that there is much in Ottoman TURKISH POLITICAL AUTHORITIES

1 By focusing on elite attitudes and practices concerning social organization and interest representation I am dealing only with certain specific aspects of political culture and am employing that concept in a much narrower fashion than is common in current social science literature. Such a strategy involves some clear advantages. First, unlike many studies of

74 — Bases of Politics

and Turkish political tradition to support corporatism and very little to support pluralism. This will be illustrated through an examination of both premodern social organization and certain recurrent themes in the attitudes of successive ruling elites toward political representation. The discussion will draw on and attempt to synthesize the work of several leading students of Ottoman and Turkish history on the following five topics: the role of municipal and professional institutions in traditional Islamic society, the nature of Ottoman statecraft, the development of center-periphery tensions in the Ottoman Empire, the corporatist nationalism of Ziya Gokalp, and the Kemalist concept of populism.

THE "CORPORATIVENESS" OF TRADITIONAL ISLAMIC SOCIETY AND THE OTTOMAN GUILDS

Orientalists have long debated the question of whether traditional Islamic society supported any counterpart to medieval European notions of civil society or autonomous corporate privilege. The major portion of this debate has centered on the concepts of the Islamic city and the Islamic guilds from the medieval period of Middle Eastern history and on the degree to which such concepts are relevant to later Ottoman social organization. The "culture and personality," it makes no pretense to being a general analysis of an unchanging Turkish "national character" or a "modal personality," and in no way implies that such things even exist. It avoids a historical or psychological determinist explanation of more recent policy and behavior and does not divert attention from the importance of varying structural and political factors. Another advantage is that such specficity prevents political culture from becoming a poorly delimited, catchall concept that can be used merely as a device for introducing any evidence that could not otherwise be included in exclusively political or structural models. It avoids the common tendency to define the concept so broadly that it is impossible to determine what one would not investigate in studying political culture. A third advantage is that it encourages sensitivity to both the ideal-expressive and the realbehavioral aspects of culture, and to possible tensions and contradictions between them. Such a distinction is particularly useful in dealing with historical descriptions of premodern Middle Eastern society, since many of them are highly idealized perceptions produced both by native writers and by European Orientalists. It is often possible for the analyst to juxtapose such idealized perceptions with other historical data in order to identify both legitimate and nonlegitimate aspects of political culture. There are, however, two obvious disadvantages to this conceptualization of political culture. First, it may tend to emphasize traditional elements at the expense of transitional elements, thereby exaggerating the degree of historical continuity in political culture. Second, by focusing on the attitudes and practices of political elites rather than on comparing them with those of nonelites, it may also tend to exaggerate the coherence of political culture. A more balanced view of political culture should include a fuller description of the degree to which traditional views have been replaced by or coexist in tension with more recently acquired views at both elite and nonelite levels. Chapter 7 will examine the political subculture of contemporary association leaders and two other nonelite groups.

Attitudinal Bases — 75

discussion of both concepts reflects a serious division between early and more recent scholarship as Orientalists have tended, first, to advance and, then, to refute claims that corporative organizations not only existed in traditional Islamic society but may have antedated and influenced similar institutions in Europe. Several early French students of medieval Islamic society, such as Sauvaget, Marcais, and Massignon, argued that urban Muslims possessed a distinctive propensity for social organization and a continuous tradition of civic spirit that allowed a common form of city life to assert itself as a major force in Islamic history despite frequent and abrupt shifts in political and military domination.2 The work of Xavier de Planhol incorporates the arguments of these writers on the centrality of urban life in Islamic religious practice and the broad structural similarities in its organization over an extensive geographic area. De Planhol asserts that "the mark of Islam has been impressed on the life and appearance of its cities more indelibly than anywhere else" and that the creation of cities (tamsir) has been "an affirmation of the Islamic personality itself." Like his compatriots, de Planhol portrays the city as being essential for the realization of the social and religious goals of Islam and indispensable to the dignity of life that it demands. The communal nature of prayer performed in a permanent mosque and the basic rhythm of religious practice that is often rigorous and prudish are supposedly most suited to urban middle-class life. "By virtue of its social constraints as well as its spiritual demands Islam is a city religion. The city is a military fortress, but, more importantly, it is a pillar of the faith and a framework within which to live the good life." 3 But the most distinctive aspect of de Planhol's argument is his denial that either the importance of the urban ideal in Islam or the regular structure of municipal organization produced among urban Muslims a serious measure of solidarity and cooperation or any form of institutionalized privilege. 2 A. H. Hourani, "The Islamic City in the Light of Recent Research," in The Islamic City: A Colloquium, ed. A. H. Hourani and S.N. Stern (Philadelphia: University of Pennsylvania Press, 1970), pp. 9-24. 3 Xavier de Planhol, 7"Ae World of Islam (Ithaca: Cornell University Press, 1959), pp. 12, 7, 9-14. He also describes this urban framework as possessing a common and distinct general morphology. Urban organization is regularly centered on the chief mosque, around which are arranged such common components as the main bazaar, khans, hammams, the citadel and seat of government, residential and semirural districts, and cemeteries. De Planhol emphasizes that the outward order of the Islamic city obeys a number of well-defined general rules that give the impression of a high degree of structure in social organization along lines of status, occupation, religion, and ethnicity. The concentric arrangement and hierarchical division of the urban quarters generally reflect the relative social position of their inhabitants; commercial districts are characterized by the hierarchical ordering and spatial partitioning of the various trades; and residential areas clearly delineate and segregate various ethnic and religious communities.

76 — Bases of Politics On the contrary, he asserts that the Islamic city was not an organic community with an independent sense of municipal consciousness, but a fractured mosaic of disconnected, self-contained, and mutually antagonistic quarters. "The [Islamic] city is a collection of cities living under the haunting fear of general massacre." The apparent order and structure of urban life are obscured by an "inextricable entanglement of detail," so that, in fact, "irregularity and anarchy seem to be the most striking qualities of Islamic cities." De Planhol contends that the absence of municipal consciousness is seen in the lack of supervision of urban growth or planning. Since there are no intermediate institutions to foster political interest in the urban community as such, nothing intervenes to temper the capricious absolutism of the prince or the usurpation of the public domain by private interest. "In the face of this arbitrary power from which nothing protects him, the Moslem conceals his private life and that of his family, so far as he can, behind the forbidding walls of his house, in the maze of alleys and back streets which half seal off his residential district.' ' 4 De Planhol concludes that, in fact, the Islamic city represents a basic regression of social organization when compared to the urban institutions that it inherited from antiquity. The fundamental institutions of urban life have retained only the form given them by antiquity. Islam has been able to substitute for them nothing of its own. It simply c o p i e s . . . . The effect of Islam is essentially negative. It substitutes for a solid and unified collectivity a shifting and inorganic assemblage of districts; it walls off and divides up the face of the city. By a truly remarkable paradox this religion that inculcated the ideal of city life leads directly to a negation of urban order.5 Ira Lapidus goes even further than de Planhol in denying that Islamic cities were unique or self-contained social units defined by characteristically Muslim qualities. Instead, Lapidus emphasizes the ties between the medieval cities and their hinterlands, which formed broad regional units within which social conflict more commonly appeared along religious-communal lines rather than urban-rural lines. Although Lapidus recognizes several dimensions of community organization and integration in medieval Islamic society, he does not include municipalities among them. He argues that cities were simply the geographical locus of groups whose membership and activities were either smaller or larger than cities themselves. In the period which followed the collapse of the 'Abbasid empire and preceded the consolidation of the Ottoman and Safavid empires, there 4 5

Ibid., p. 8. Ibid., pp. 22-23.

Attitudinal Bases — 77

are four kinds of groups or communities: neighborhood bodies; fraternities, including youth gangs and Sufi brotherhoods; religious communities such as the Muslim schools of law and sectarian groups; and, finally, states or empires. None of the bodies was a city community, except insofar as cities were naturally the headquarters of all groups. Some of these groups such as families, quarters and fraternities were usually small segments of city populations while religious communities and states were much larger than individual cities. As opposed to the ancient polis or medieval European commune, there were no geographically defined communities in the Muslim world.6 Within the city Lapidus points to socially homogeneous, villagelike quarters as the keystone of social organization and argues that the sharply divided city populations enjoyed few institutions that cut across quarter boundaries to bind them together. He specifically denies that the medieval guilds or any occupational groupings performed such an integrative function. These, he claims, were extremely weak and were usually created to meet the fiscal and regulatory needs of the state rather than reflecting the autonomous interests of their members. He credits the fraternities and mystical brotherhoods with greater relative importance, especially in linking urban and rural populations. But he believes that the most powerful force in overcoming parochial loyalties and shaping more highly ordered communities was the ulema and their schools of religious law. These alone recruited adherents from all quarters and classes of the population.7 Thus, the concept of the Islamic city has undergone a major transformation involving a fundamental reassessment of the role of urban life in premodern Middle Eastern social history. De Planhol retains the concept, but interprets it paradoxically to argue that the original Muslim quest for community solidarity was actually defeated by subsequent municipal disorganization, which produced a net decline in the cohesion of the region's urban life. Lapidus suggests abandoning the concept altogether because it perpetuates the misconception that Muslim cities were somehow analogous to the autonomous European communes and municipalities and because it 6

Ira Lapidus, "Muslim Cities and Islamic Societies," in Middle Eastern Cities, ed. Ira Lapidus (Berkeley: University of California Press, 1970), p. 33. 7 He argues that throughout the Middle East from the eleventh century onward the ulema were able to assume many of the functions of the older administrative and landowning classes and to become a political and social as well as a religious elite. Leading ulema families merged with landowning, bureaucratic, and merchant families and used their control over offices and incomes from land and waqfs to provide the principal component of both urban and broader regional leadership for several centuries before the reemergence of major bureaucratic empires (Ira Lapidus, "The Early Evolution of Muslim Urban Institutions," Comparative Studies in Society and History 15 [1973]:21-50).

78 — Bases of Politics

detracts attention from more powerful integrative links forged by the ulema between different quarters and classes and between the cities and their hinterlands. The conflict between Orientalists over the existence of autonomous corporate privilege in medieval Muslim society has focused even more clearly on the issue of the guilds. The most far-reaching assertions of the importance of the Islamic guilds have been made by Louis Massignon and Bernard Lewis. Massignon claimed that traditional Islamic society was essentially corporative in nature and that its main units of economic and social organization were the professional guilds. He believed that from the ninth century onward the Islamic guilds generally enjoyed great autonomy as armed civil defense forces that combined religious heterodoxy with political protest to check arbitrary rule. He also contended that their example spread to the West, where they influenced the formation of European guilds and Freemasonry.8 Lewis emphasizes the transmission of these traditions to Anatolia through the associations of the Ahis. The Ahis, he claims, were the first to merge guild structures withfictuwwa—chivalric fraternities—and Sufism, producing a social movement that had "solidarity and hospitality" as its code, the artisan class as its social basis, and "the slaying of tyrants and their satellites" as its task.9 * According to Massignon, the Islamic guild was created as a weapon of Qarmatian propagandists in their struggle to weld the laboring classes into a force capable of overthrowing the 'Abbasid caliphate. Their initial expansion supposedly resulted from their participation in the political, religious, and military movements of the Ismailis, and under Fatimid rule they are said to have acquired formal recognition by the state, possessed considerable privileges, and contributed to an important commercial revival. Although Massignon recognizes that during the post-Mongol period the guilds were generally persecuted and strictly controlled by Sunni governments, he claims that their heretical and insurrectional traditions spread to Anatolia in the thirteenth and fourteenth centuries and throughout the Muslim world thereafter (Louis Massignon, "Sinf," in Encyclopaedia of Islam, 1st ed., pp. 436-437). 9 Bernard Lewis, "The Islamic Guilds," Economic History Review 3 (November 1937):2537. Building upon Massignon's description, Lewis specifies four distinctive qualities of the Islamic guilds that, he contends, persisted throughout the premodem Muslim world, and he clearly implies that these qualities made guilds even more important features of Islamic society than they were in medieval Europe. Lewis argues, first, that Islamic guilds were a spontaneous development from below. Their privileges and political importance did not derive from official recognition and the performance of public services but from their ability to mobilize the hostile, mistrustful, and antiauthoritarian sentiments of the laboring masses. Second, the Islamic guilds retained their popular and egalitarian character. Since they did not experience the inner divisions of production changes that split European masters and journeymen, their ranks remained of the same class. Third, they performed a major social integrative function by virtue of their interconfessional and mystical orientation. Fourth, they retained a stronger inner spiritual life that made them more than mere professional organizations, since ideology and a strict moral code were taught in the same degree as the craft itself.

Attitudinal Bases — 79

Recent writings of Claude Cahen and S.N. Stern present a fundamental challenge to all of these assumptions about the Islamic guilds. Cahen argues that Massignon and Lewis have exaggerated tenuous and transitory relations between the Ismailis, the fraternities, and the guilds that were, at best, confined to certain specific regions, resulting in an incorrect description of the evolution and diffusion of guilds that misrepresents their limited role in Islamic society.10 Cahen also dismisses claims concerning the insurrectional and paramilitary activities of the guilds, saying that urban unrest was largely the creation of formless mobs within which the fraternities and not the guilds occasionally supplied minimal organization. He concludes by asserting that the medieval Islamic guilds generally were not spontaneous and private associations that represented professional interests but "organisms emanating from the administrative apparatus of the state." Stern expresses a much stronger and categorical rejection of the original conceptualizations of the Islamic city and the Islamic guilds, including a particularly severe critique of Massignon's theses, which he describes as "fantastic" and "bogus." For Stern "one of the most essential characteristics of the Islamic city is the looseness of its structure, the absence of corporate municipal institutions." To the extent that he recognizes any parallel between the Islamic and Western guilds, he sees them as existing in the late Roman and Byzantine structures that were created and controlled by the state rather than in the medieval European pattern. Stern describes "a groping towards the evolution of guilds" in late medieval Islamic society, but insists that compared to their European counterparts they were less monopolistic, less institutionally organized, and deprived of the right of self-regulation or any autonomous privilege.11 The weakness of these municipal and professional institutions, Stern argues, is merely a particular illustration of the more general absence of corporative institutions in Islamic society. "One of the striking differences between the society of medieval western Christendom and Islamic society was this: that whereas in the former all sorts of corporate institutions 10

Claude Cahen, "Ya-t-il eu des corporations professionnelles dans Ie monde musulman classique?" in The Islamic City, pp. 51-64. Cahen claims that the chivalric fraternities always drew members from a wide variety of occupations and remained structurally distinct from the guilds. Cahen recognizes a late and brief convergence of fraternal and guild activities around the thirteenth century that was limited to the Irano-Turkish plateau, but he denies any original or subsequent interconnection between them even in these areas. Similarly, he acknowledges a brief relationship between Ismaili movements and the guilds only in Iran and Iraq, but notes that even this was primarily a rural phenomenon that failed to take root in the major cities. 11 S. N. Stem, "The Constitution of the Islamic City," in The Islamic City, pp. 25-50.

80 — Bases of Politics

proliferated, in the latter they were entirely absent."12 According to Stern, ' 'that. . . no corporate civic institutions ever developed within the framework of Islamic civilization, which started with an incomparably more advanced urban life than Western Europe, is paradoxical, but it is undeniably true." 13 A similar rift between early and more recent scholarship exists regarding the guilds of the Ottoman Empire. Gibb and Bowen accept Massignon's and Lewis's description of the corporatist nature of medieval Islamic society and project it, with minor qualifications, onto their outline of Ottoman social organization until the nineteenth century. They claim that within the empire's urban areas the guilds provided the main basis for collective action and cohesion among otherwise fragmented and hostile quarters. "[The city's] characteristic feature was the corporation, whose social importance can hardly be overestimated. If religion was the cement of the Islamic structure, the corporations were the bricks of which it was built. . . . The hara (quarter) system was superimposed on the corporative system, but did not conflict with it." 14 A very different image of the Ottoman guild system emerges from the recent studies of Gabriel Baer. He contrasts the associations of Turkish craftsmen and traders with those of their counterparts in other areas of the empire and concludes that the Ottomans imposed far more control and hierarchical organization over Turkish guilds than over the esna/elsewhere. He argues that from its inception the Turkish guild system was molded from above rather than drawing its impetus from the spontaneous organization of its membership.15 He describes the Turkish guilds as being 12 He notes, for example, that while juridical personalities exist in Islamic law, they are narrowly limited to the estate of a deceased person, the pious endowment (waqf), and the public treasury, but do not include municipalities, professional associations, universities, or religious orders (ibid., p. 47). 13 Ibid., pp. 32-33. 14 H. A.R. Gibb and Harold Bowen, Islamic Society and the West, 2 vols. (London: Oxford University Press, 1950), 2:277. Gibb and Bowen concede that the powers of the Ottoman guilds were narrower than those of the European corporations, but argue that their comprehensive coverage of nearly all occupations and their persistent autonomy explain the surprising stability and adaptability to changing political circumstances shown by industry in all parts of the empire. They recognize that the consolidation of Ottoman rule was accompanied by attempts to impose stricter government control over guilds, especially in the city of Istanbul, in order to curb their insurrectional tendencies. But they claim that even in the capital such attempts were often frustrated by the close association that artisans developed with the troublesome janissaries and the Bektaski mystics and that the Ottoman guilds generally preserved the organization and privileges of the Ahis and their medieval predecessors throughout the eighteenth century (ibid., pp. 288-289). 15 Baer insists that the popular and autonomous Ahi fraternities of the early empire were not professional organizations and that Turkish guilds did not appear until the first half of

Attitudinal Bases — 81

closely connected with the state apparatus, functioning primarily as an administrative link between the urban population and the ruling institution, and providing an important means of supervision and control over that population.16 The guild privileges that Gibb and Bowen present as exemplifying Ottoman corporatism are treated by Baer as isolated vestiges within a larger network of bureaucratic control and central domination. He makes it clear that the representation of esnaf interests was a very minor function of the Turkish guilds and that they were allowed only a narrow scope of operation even in fulfilling their primary responsibilities of transmitting government directives and reporting legal violations of their members.17 A few of the largest and most loyal guilds, such as the shoemakers and tanners, were granted the right to punish their own members for their offenses. But these were exceptional cases in which the privileges appear to have been a reward for their support of the sultan against the janissaries. In addition to their primary function as administrative links between the state and craftsmen, the guilds were also responsible for aiding and protecting their members.18 Nevertheless, Baer concludes that, with the exception of mutual assistance, Turkish guilds engaged in relatively few activities that were not related to the regulation of commerical life and control of the urban population, especially when they are compared to guilds in other parts of the empire. Contrasting the Turkish guilds with the sixteenth century, when the state began to replace the disruptive fraternal associations with a comprehensive set of centrally controlled professional structures. 16 Gabriel Baer, "The Administrative, Economic, and Social Functions of Turkish Guilds," International Journal of Middle East Studies 1 (January 1970):28-50. 17 Ibid., pp. 33-42. The chief officers of the guilds (the kethiidas or kahyas) were generally agents of the government rather than spokesmen for the guilds, and were commonly recruited from retired civil servants. Unlike their Egyptian and Syrian counterparts, the Turkish guilds were deprived of fiscal functions and had no role in tax collection. The Turkish guilds were charged with supervisory responsibility for implementing detailed government regulations concerning production methods, quality control, weights and measures, fixed prices (nahr), and wages. But in each case the authority of the guilds was generally limited to denouncing violators to the government; direct enforcement and punishment were carried out through the official market inspectors (the muhtesibs) or the Islamic court judges (kadis). 18 They were responsible not only for supplying the government with specific goods and labor power but also for distributing important raw materials needed by craftsmen and artisans. The arbitration of disputes among members and the regulation of their personal behavior were commonly conducted by guild leaders in order to prevent the intervention of the kadis. The particularly widespread mutual assistance programs and provident funds of the Turkish guilds may have been a survival from the original fraternal organizations. Although Baer doubts the common assumption that the Turkish guilds maintained close connections with the Sufi tarikats, he does recognize their role in promoting solidarity among their members through common dress and prayers, annual processions and festivities, and occasional pilgrimages.

82 — Bases of Politics

those of Macedonia, North Africa, Damascus, and Egypt, Baer concludes that in the Ottoman provinces guilds were generally more autonomous, that they commonly exercised wide judicial powers, that they often served as genuine representatives of esnaf interests vis-a-vis the local administrations, and that they preserved much stronger social and religious traditions.19 The recent reassessments of the nature of municipal and professional institutions in the pre-Ottoman Middle East suggest that it is highly doubtful that the Ottomans inherited from medieval Islamic society even limited analogues to the European corporations and estates. Furthermore, Baer's comparative examination of Middle Eastern guild systems during the Ottoman period clearly indicates that the Ottomans did not independently develop such institutions. The emergence of new professional associations after the fifteenth century reflected the increasing penetration of the administrative apparatus of the state rather than the development of civil society. Baer's specific description of the relationship between the Turkish guilds and the state raises the more general question of Ottoman views of social organization and political representation.

OTTOMAN STATECRAFT AND PATRIMONIALISM

Several students of Ottoman history have emphasized the importance of Ottoman political theory in understanding the empire's early social and economic organization and political institutions.20 inalcik, Mardin, Rosenthal, Thomas, and Itzkowitz interpret Ottoman statecraft as combining pre-Islamic Turkish custom, the Persian tradition of siyasetnames, and medieval Islamic political philosophy in a technique of imperial rule that departed significantly from classical Sunni theories of political legitimacy in its broad acceptance of social stratification, bureaucratization, and man19 Ibid., pp. 46-50; Gabriel Baer, Egyptian Guilds in Modern Times (Jerusalem: Israel Oriental Society, 1964). 20 Such discussions often argue that the principles of Ottoman statecraft had little practical relevance after the classical period because of the increasing gap between the ideals stated by Ottoman political writers and the reality of deterioration in the empire's political institutions from the mid-sixteenth century onward. An equally common view is that these principles gradually hardened into an idealized vision of an imaginary Golden Age and that their continuing importance had the adverse effect of diverting a complacent elite from necessary attempts at fundamental reform. Our own discussion of Ottoman statecraft portrays it neither as an accurate description of the actual practices of Ottoman rule nor as a rigid and irrelevant set of precepts that became an obstacle to change. Instead, we seek to specify certain key concepts and themes of Ottoman statecraft that constituted a kind of legitimizing formula that provided a flexible but enduring guide for successive ruling elites in their attempts to order social and political life under varying historical conditions.

Attitudinal Bases — 83

made law (kanuri). Each of these historians describes the legitimation of Ottoman patrimonialism as based upon an organic and hierarchical view of social organization in which order, harmony, and equilibrium (nizam, ahenk ve denk) are to be guaranteed by the sultan's prudent and personal application of distributive justice {adalet) directly among his individual subjects rather than through intermediary representatives with autonomous rights. A general overview of Ottoman notions of the relationship between state and society can be provided by examining briefly three related themes that characterized the most important works of premodern Turkish political literature—the four social pillars of the state (erkan-i erbaa), the "circle of justice," and the use of biological metaphor in political analysis.21 Many early Islamic political writings reflected a tension between the original egalitarian image of the umma (the Muslim community), in which believers were differentiated primarily according to their degree of personal piety, and the government's need for well-defined and stable social bases to organize production and support the state. Under the classical caliphates and especially during the medieval period the latter tendency was often expressed in attempts to specify various "natural" or "universal" social categories that were reminiscent of earlier Sassanian views of stratification and of ancient Greek classifications. The model of social organization most commonly employed by Ottoman writers was based on the "four pillars," or fundamental classes outlined by Nasir al-din Tusi and adopted by the fifteenth-century Ottoman historian Tursun Bey. The four major Ottoman social orders were the "men of the sword" and the "men of the pen" (who together constituted the askeriyya, or the ruling class) and the "men of negotiation" and the "men of husbandry" (who together constituted the reaya, or the flock). The relative status of each class was clearly reflected in their different tax obligations, economic functions, and rights of domicile and dress. A fundamental principle of Ottoman social policy was the strict separation of these social orders, especially the exclusion of the reaya from the privileges and exemptions of the askeriyya, and the unchallenged authority of the sultan alone to determine an individual's status.22 The Ottoman social orders were considered to be divinely sanctioned, but they were not regarded as "estates" either in the sense of possessing 21 Halil Inalcik, "The Nature of Traditional Society," in Political Modernization in Japan and Turkey, ed. Robert Ward and Dankwart Rustow (Princeton: Princeton University Press, 1964); §erif Mardin, The Genesis of Young Ottoman Thought (Princeton: Princeton University Press, 1962); E.I.J. Rosenthal, Political Thought in Medieval Islam (London: Cambridge University Press, 1962); Norman Itzkowitz, Ottoman Empire and Islamic Tradition (New York: Alfred A. Knopf, 1972). 22 inalcik, pp. 42-^5; Itzkowitz, pp. 37-62.

84 — Bases of Politics

political rights independent of the crown or in the sense of forming a contractual network of groups with hereditary rights and duties toward one another. Instead, they were regarded as functionally interdependent and hierarchically ordered categories that were universally subordinated to the will of the sultan. They were seen as making necessary though unequal contributions toward the common purpose of serving the royal dynasty and the religion it protected and, therefore, as being entitled to unequal priv­ ileges and rewards. The maintenance of social peace and harmony was seen as dependent on the state's ability to keep each individual in his appropriate social and economic position and to ensure his continued con­ tribution to the service of the sultan.23 The principal aim of Ottoman statecraft was to increase the state's territories and revenues as much as possible without endangering the pros­ perity of the productive classes or upsetting the balance between the tra­ ditional social orders. These goals are clearly summarized in the "circle of justice," a simple but all-encompassing set of rules of prudence that are repeated in Turkish political writings from the Kutadgu Bilig of the eleventh century, to Kimhzade's sixteenth-century Ahlak-i AM, to the proclamations of the nineteenth-century Tanzimat reformers. These max­ ims, designed to illustrate the interdependent relationships among the var­ ious classes and their appropriate functions in the well-run state, "em­ bodied the ethical, political and social values of the Ottoman class." 1. 2. 3. 4. 5. 6. 7. 8.

There can be no royal authority and kingdom (miilk) without the military (askerler); There can be no military without wealth and treasure {mal ve servet); Wealth and treasure are produced by the reaya; The sultan maintains the reaya by making justice (adalet) reign; Justice requires harmony (ahenk) in the world; The world is a garden, its walls are the state (devlet); The state's support is the religious law (qeriat); There is no support for the religious law without royal authority 24 and kingdom.

23 Until the end of the Suleymanic era this was largely achieved through the timar-tahrir system of land tenure and the ghulam system of meritorious recruitment and education of the military and bureaucratic elite. The timar-tahrir system aimed at maintaining direct state control over the peasant and the soil by placing responsibility for continuous cultivation and revenue collection in the hands of local cavalrymen (sipahis) who were agents of the central administration. The ghulam system similarly aimed at stratifying the members of the askeriyya and channeling them into specialized and hierarchically graded career patterns as the personal servants or slaves of the sultan. 24 Itzkowitz, p. 88; Haul inalcik, "Kutadgu Bilig," in Resit Rahmi Arat Ιςίη (Ankara: Turk Kulturiinu Arastirma Enstitiisii, 1966), pp. 259-271.

Attitudinal Bases — 85 In the Ottoman texts these statements were generally arranged around the circumference of a circle, thus demonstrating how the last statement led directly back to the first, and forming a schematic overview of the relationship between state and society. The harmonizing and coordinating role of the state was the principal means by which Ottoman patrimonialism sought to discharge the sultan's responsibility as the "father of his subjects." Even late Ottoman writers sought to legitimize this notion of the state's function by reference to the immutable principles of the feriat. But these writers were, in fact, instrumental in reasserting the independent legislative prerogative of the sultan, since they were also responsible for formulating state policies that, especially in periods of crisis, were promulgated as kanun or nizamname issuing directly from the sovereign will of the ruler. Thus, the circle of justice blended traditional Turkish-Mongol reliance on rule through imperial decree with Indo-Iranian concepts of social hierarchy. The writings of the most famous Ottoman chroniclers, Katib (Jlelebi and Naima, also drew on medieval Islamic political philosophy's cyclical view of history and incorporated it into their descriptions of the empire's military setbacks and political unrest during the seventeenth century. Like Ibn Khaldfin, whom they generally imitated, these writers portrayed the state, like the human body, as passing through the stages of birth, growth, and decay. They also identified the grand vizier with the physician who attempts to restore good health through the skillful application of his art or practical science. Unlike Ibn Khaldun, however, the Ottoman chroniclers made these organic metaphors a central part of their analysis and welded them to their concern with the application of distributive justice among the unequal traditional social orders. Institutional breakdown and social dislocation were identified with physiological dysfunction caused by the disequilibrium of society's four classes, which were depicted as corresponding to the four humors of the body. Both Katib ^elebi and Naima employed vivid medical and organic analogies in their descriptions of the indiscipline of the janissaries and the breakdown of the classical system of land tenure. In summarizing the writings of these two chroniclers, Itzkowitz aptly describes their style of political analysis as "the pharmacology of Islam." The four classes are compared to the four humors of the body: the Men of the Sword to phlegm, the Men of the Pen to blood, the Men of Negotiation to yellow bile, and the Men of Husbandry to black bile. Of principal concern in the problem of maintaining equilibrium were phlegm and black bile. Old age in the body is characterized by an excess of phlegm and in the state by an overexpanded military establishment. It

86 — Bases of Politics

is extremely difficult to keep the phlegm (military) under control. The black bile is disturbed through undernourishment. In the early periods the sultans had protected the reaya from injustice. In later periods the reaya were the victims of oppression and mismanagement. This, in turn, affects one of the central organs of the body, the stomach, which in the state is the treasury. If bile cannot enter the stomach, or wealth cannot enter the treasury, undernourishment results, and the body, or state, suffers.25 Mardin and inalcik note that this formal model of Ottoman patrimonialism was more an ideal than a reality. They contend that one key to the expansion and endurance of the empire was the willingness of the elite to stop short of insurmountable organizational tasks and accept the decline of its authority in favor of local elites as distance from the capital increased. Mardin, in particular, distinguishes between legitimate and nonlegitimate features of Ottoman social organization; between the static view of patrimonial society based on the equipoise of hierarchically ordered, interdependent elements and the actual fluidity of Ottoman social structure; and between the superficial appearance of an oriental despotism based on a slave-staffed executive and the more complex reality of continuous centerperiphery tensions throughout the empire.26 Mardin argues that even in the heartlands of the empire serious exceptions to the model of central control existed during the empire's formative and classical periods. Several remnants of earlier Byzantine "noble" families and rival Muslim princes who joined the Ottomans rather than fight them later merged into a new type of landed gentry into which flowed all types of former privileged residues. The sipahi cavalrymen did not enjoy the formal rights of land ownership, bequest, or alienation, since most land was considered the personal property of the sultan. But their control over timars was, in fact, generally hereditary and they often became allies of the "noble" families in attempting to perpetuate their tenure. As early as the sixteenth century the need for a permanent standing army and declining revenues from the dwindling overland trade led the state to resort to tax-farming (iltizam) and office selling to meet its growing need for cash. With the corruption of the land regime and of the recruitment process for the classical institutions of the state, the sultan was forced to tolerate and bargain with, though not necessarily legitimize, a number of privileged 25

Itzkowitz, pp. 102-103. §erif Mardin, "Historical Determinants of Social Stratification: Social Class and Class Consciousness in Turkey,'' Siyasal Bilgiler Fakiiltesi Dergisi 22 (December 1967): 111-142; Serif Mardin, "Power, Civil Society, and Culture in the Ottoman Empire," Comparative Studies in Society and History 11 (June 1969):258-281; Serif Mardin, "Center-Periphery Relations: A Key to Turkish Politics?" Daedalus 102 (Winter 1972):169-190. 26

Attitudinal Bases — 87

and relatively autonomous groups behind the fagade of an inflexible commitment to centralized control and a denial of institutionalized privilege.27 Mardin contends that a key factor that permitted the Ottoman state to preserve its often tenuous dominance and prevented the emergence of anything resembling civil society was not the absence of private privilege but the general precariousness of such privilege and its dependence on the prestige and authority of the state. Effective power and the ability to accumulate wealth were generally dependent on control of positions that were in one way or another positions of state service rather than on the direct and independent control of the means of production. Since autonomous privilege was seldom institutionalized and never legitimized, it might be tolerated during periods when the state was relatively weak only to be retracted later if it was flouted openly and aggravated the elite's lingering fears of political disintegration. Accordingly, he believes that the overriding dimension of political conflict in the empire was a continuous tension between center and periphery. The Ottoman state relied on precarious privilege without trusting anyone in a position of power for very long, while its officials and agents, especially in the provinces, persistently sought to extend their tenure and transform their de facto powers into permanent and legitimate rights.28 Mardin contrasts this Ottoman pattern of unidimensional political confrontation between the state and all other groups possessing even transitory, de facto autonomy with the more diverse set of conflicts that characterized postfeudal European society. In Europe the breakdown of feudalism in favor of the rise of capitalism and the emergence of centralizing state bureaucracies involved multiple and occasionally crosscutting cleavages between church and state, between centralizing nation-builders and local aristocracies, and between owners and nonowners. In the Ottoman Empire, however, conflict continued to be defined by a primary axis dividing the 27 The ulema, whom the early sultans had organized within a highly structured system of religious education and meritorious promotion, became a self-recruiting caste that enjoyed a virtual monopoly over education and the administration of justice. The janissaries, who had originally constituted the core of the ghulam system, increasingly came to be recruited from freeborn Muslims who violated the traditional prohibitions against marriage, sought additional income through moonlighting and enrolling in esnqf guilds, and often incited urban mobs to depose both viziers and sultans. Amirs and beys in charge of provincial administration frequently sought to extend their tenure by allying with local elites (esraf) and janissary garrisons and denying revenues to the state. Tax-farmers (m<azims), who paid handsomely for their positions, increased the exploitation of the peasantry to extract greater surplus and gradually transformed much state land into de facto private holdings. The ayan and derebeys established vast local dynasties during the eighteenth century and then turned their troops against the weakened central government, briefly forcing the sultan to recognize and guarantee their usurpation of political and economic control in the provinces. 28 Mardin, "Center-Periphery Relations," p. 175.

88 — Bases of Politics

state from all social strata rather than by multiple axes dividing social strata from one another. The discontinuity in the Ottoman view of power as belonging exclusively to the sultan and his executive machinery led to the creation of a view of strata in the Ottoman Empire as political, and to a conception of the game of politics as a ' 'zero sum game,'' an individual being by definition either top or under-dog. The saliency of these strata replaced the European saliency of strata connected with the production and distribution of goods and services.29 Mardin argues that one of the clearest examples of this unidimensional pattern of conflict was the failure of peasant revolts to polarize the peasantry against the local gentry.30 Mardin characterizes the major Ottoman "revolutions"—whether they involved the janissaries and esnaf'm the capital or the ayan and peasantry in the countryside—as "flashes in the pan" because they were brief and sudden eruptions of accumulated grievances whose aim was to gain temporary concessions and relief from the state rather than basic clashes between broad social and economic groups enjoying substantial political rights. Even when immediately successful, rebellious groups were unable to consolidate their gains and in the long run were subdued by the bureaucracy's renewed efforts at centralization.31 The traditional tolerance of Ottoman statecraft provided a necessary flexibility in managing the often severe tension between center and periphery over a vast territory whose population included a heterogeneous collection of religious, ethnic, and linguistic groups. Yet it also prevented the development of civil society, since even the most conspicuous deviations from the ideal were merely tolerated and never formally legitimized. The emergence of European bureaucratic states was characterized by a series of confrontations and compromises with various social forces—the nobility, the Church, municipalities, burghers, and later the industrial work force—whose gradual integration into new nation-states involved some recognition of their autonomous rights and status. But such compromises on the part of the Ottoman state were merely tactical and temporary, so M

Mardin, "Historical Determinants," p. 130. Peasant protest was commonly incited and led by disgruntled local landholders whose grants were being revoked (such as sipahis who were dispossessed by tax-farmers) or by ambitious local officials and ayan attempting to capitalize on economic chaos in order to create an autonomous force in a remote province. The peasantry understood and resented the Ottoman reliance on precarious tenure, which often increased their exploitation at the hands of insecure officials and landholders. Yet they often shared with their more immediate oppressors a common antagonism toward the center and were willing to join them in resistance if offered protection against the taxing, conscripting, and confiscating agents of the state. 31 Mardin, "Center-Periphery Relations," p. 180. 30

Attitudinal Bases — 89

that center-periphery tensions persisted as the major dimension of political cleavage. THE EXACERBATION OF CENTER-PERIPHERY TENSIONS

A major turning point in the history of center-periphery relations occurred in the early nineteenth century, when Sultan Selim Ill's initiatives toward defensive modernization provoked fundamental challenges to the authority of the state from several groups seeking to defend their traditional status. The sultan believed that the Ottomans' long series of military setbacks in Europe could be reversed only by forming a modern standing army independent of the increasingly corrupt and inept janissary corps. But his attempts to raise such a force, and to provide it with the necessary financial support and training, met with open rebellion from the janissaries, the ulema, and the ayan. After the janissaries and ulema had incited the populace of Istanbul to join in deposing the "infidel sultan" and executing his supporters, the combined troops of the ayan occupied the capital and seized control of the central government. The ayan enthroned their own candidate, Mahmud II, and forced him to sign a formal agreement guaranteeing their positions in the provinces. Inalcik notes that such pledges on the part of the sultan were unheard of in Ottoman history and reflected the clear diminution of the state's power, especially vis-a-vis the provincial magnates. The Covenant of Union (Sened-i Ittifak) between the sultan and the ayan "was drawn up in the form of a regular contract according to the shari'ah, the parties being the state and its officials, ulema, generals of the military corps in the capital, and representatives of the provincial 'dynasties.' " Its main provisions pledged that as long as the "dynasties" did not infringe upon the central authority, the government had to respect their acquired status as permanent and irrevocable. [All parties] promised always to respect the sovereign authority of the sultan and the orders of the Grand Vezir, who represented in his person the sultan's absolute power, and take united action against any rebellion. . . . Just as the signatories promised collectively to safeguard the person and authority of the sultan and the order and security of the state, so the security of the provincial "dynasties" was to be protected by joint action of the signatories in the event that any law-abiding "dynasty" was assailed by the "state" or by vezirs in the provinces.32 The document specifically stipulated that the covenant was to be ' 'perpetual and unalterable" and that it would be signed by every grand vizier and 32

inalcik, "The Nature of Traditional Society," p. 52.

90 — Bases of Politics $eyhulislam upon his accession to office. The ay an had not only forced the sultan into the unprecedented act of transforming revocable privileges into permanent rights, but had also arrogated the right to take collective action against anyone, including members of the sultan's government, who violated the agreement. Mahmud's decision to renounce the covenant at the earliest possible moment precipitated an even more severe exacerbation of center-periphery tensions, for he combined his renunciation, first, with a universal assault on nearly all of the autonomous centers of privilege that had been permitted to operate in the twilight of legitimacy and, second, with renewed efforts to reconcentrate power in the administrative apparatus of the state.33 For subsequent generations of bureaucratic reformers the commitment to subordinate all partitive interests to a more centralized and expanded state bureaucracy soon became identified with modernization itself. This remained the principal aim of Ottoman reformism until the empire's destruction. Chambers notes the enduring impact of Mahmud's style of reformism, including a rather pejorative reference to "interest groups" that is faithful to common Turkish usage even today. Mahmud sought to restore the authority of the sultan over his empire by breaking down the power of entrenched interest groups and replacing their institutions with new ones of his own. He pursued that end by centralizing and modernizing the machinery of government and at the same time extending its operations. Bureaucratic centralization and the expansion of the bureaucrat's sphere of activity thereafter constantly attended all efforts to modernize the Turkish polity.34 Mardin argues further that the rise of the secularist bureaucrats and their long control over reformist policy was accompanied by the dominance of a new bureaucratic mentality that departed significantly from the goals and 33 Mahmud encouraged a janissary counterattack on the ayan that succeeded in dispossessing many of them and restoring the state's authority in Rumelia and Anatolia. Before long he was also able to destroy the janissaries as well, removing the other major obstacle to his desire to reorganize the institutions of the state and reassert central control over the empire. The ulema were gradually divested of their control over education, the administration of justice, and the lucrative pious foundations. Military force was used to subdue all of the rebellious pashas and notables except for Muhammad AIi in Egypt. The timars, which had been gradually transformed from personal grants of the sultan intofieflikehereditary preserves, were abolished and tax-farming was greatly reduced. And, perhaps most importantly in the long run, the supremacy of the administrative bureaucracy was firmly established over the rival military and ecclesiastical bureaucracies, preparing the way for the secularist Tanzimat reformers to reorganize the central government, the entire provincial administration, and the system of land tenure and tax collection. 34 Richard Chambers, "The Civil Bureaucracy," in Political Modernization in Japan and Turkey, p. 313.

Attitudinal Bases — 91

practices of traditional Ottoman statecraft. He describes the nineteenthcentury bureaucracy as renouncing its traditional slave status and constituting an administrative aristocracy trained on the model of the French Grandes Ecoles and guided solely by "interests of state.'' He characterizes their renewed concern with shoring up the center and expanding the penetration of the state apparatus as "less patrimonial and sultanic" than "rationalizing" since it aimed, not at restoring balance among immutable social orders, but at establishing direct relationships between the state and the citizenry without intermediary agents or representatives.35 Mardin concludes that the modernizers' rigid notion of the national interest and their rejection of intermediaries made the Ottoman elite intolerant of the social mosaic over which it ruled. Their preoccupation with saving the empire from destruction by its internal and external enemies led to a constant, and often exaggerated, fear of anarchy, rebellion, and treason as a leitmotif of modernizing Turkish statecraft. This served to strengthen the traditional cultural and political alienation of the ruling elite from society as military and bureaucratic leaders viewed themselves as the sole disinterested defenders of the public interest in an era of pervasive corruption, privilege seeking, and foreign intervention. As this hardened bureaucratic mentality emerged, the flexible and accommodative nature of traditional Ottoman rule was replaced by an indiscriminate hostility toward all representation of partitive interests—whether based on religion, culture, language, class, or region—as a threat to national unity and the security of the state. While the bureaucratic reformers proceeded to reinterpret the traditional ideals of Ottoman statecraft as demanding the subjection or elimination of all centers of private power, a small group of minor bureaucrats and journalists known as the Young Ottomans were trying to reinterpret the actual practice of Ottoman statecraft as supporting the establishment of representative government. The Young Ottomans claimed that by concentrating greater powers in their own hands the administrative elite was not only usurping much of the sultan's traditional authority but was also reversing a supposedly long and clear Ottoman tendency toward limited and constitutional government. Namik Kemal, for example, asserted that a form of "social contract" or unwritten constitution had, in fact, emerged from the numerous checks on royal and bureaucratic tyranny provided by such "populist" groups as the ayan, the janissaries, and the esnaf. Certain of these limitations already had been formalized in the Sened-i Ittifak— sometimes referred to as an Ottoman Magna Charta—and, according to the Young Ottomans, should have been institutionalized in a new constiMardin, "Center-Periphery Relations," p. 182.

92 — Bases of Politics tution and parliament. The Young Ottomans urged, in other words, that previously tolerated, group-specific privileges be expanded into guaranteed, universal rights, that relatively autonomous centers of private power not be subordinated to the state but serve as popular representatives in a consultative democracy. This, they believed, merely amounted to a return to the original intent of the early Islamic community and a natural evolution from recent Ottoman practice, both of which were ignored by the secularist and tyrannical bureaucratic elite.36 From the Tanzimat to the Young Turk era the modernizing militarybureaucratic elite did engage in limited experiments with political representation as part of their unsuccessful attempts to foster Ottoman nationalism.37 But their overriding concern continued to be strengthening the crumbling power of the state apparatus and they were never seriously willing to replace the traditional checks on their authority with new ones. If the Young Ottomans exaggerated the degree to which privilege and autonomy were tolerated traditionally in order to justify the recognition of universal rights, the modernizing elite exaggerated the scope and danger of private power in order to justify more centralized bureaucratic rule. The Young Ottomans' argument that constitutional government could be supported by the checks and balances of various social and economic groups may be seen as an early attempt to derive a vague and incipient pluralist view of representation from their reinterpretations of Ottoman political tradition. But when the leadership of the constitutional movement passed from Ottoman nationalists to Turkish nationalists, corporatism began to emerge as a much clearer ideological tendency.

THE CORPORATIST NATIONALISM OF ZIYA GOKALP

Ziya Gokalp is best known as the major ideologue of modern Turkish nationalism, but he is also notable as Turkey's first professional social scientist and its first clear advocate of corporatism. The bulk of Gokalp's writings were a definition and celebration of Turkish culture, which, he argued, formed the basis of a unique national identity that would allow the Turks to adapt the most advanced economic and technological aspects of European civilization without losing their supposed moral superiority. Gokalp's famous distinction between culture and civilization {hars ve medeniyei) has been discussed at length by Heyd and Berkes, but relatively little attention has been given to his attempts to incorporate the concepts * Mardin, The Genesis of Young Ottoman Thought, pp. 308-313. Roderic Davison, "The Advent of the Principle of Representation in the Government of the Ottoman Empire," in Beginnings of Modernization in the Middle East, ed. William Polk and Richard Chambers (Chicago: University of Chicago Press, 1968), pp. 93-108. 37

Attitudinal Bases — 93 of Durkheim's sociology into an evolutionary theory of the social and economic development of the Turkish nation.38 This effort became a consistent secondary theme throughout Gokalp's work. If his primary purpose was to define the cultural bases of Turkish nationalism in order to show that modernization was compatible with retaining a distinctive Turkish identity, his secondary aim was to show that the Turks also could overcome the problems of delayed, dependent economic development without promoting antagonistic social classes. Gokalp's unusual blend of Durkheimian sociology and Turkish nationalism provides an important ideological link between traditional Ottoman statecraft and modern Turkish corporatism. Unlike the Ottoman chroniclers, Gokalp's philosophy of history was evolutionary and positivist rather than cyclical. Yet he retained much of the Ottoman view of organic and hierarchical social organization, while refashioning it and providing it with what he thought were more modern philosophical and scientific bases. In the short term his writings anticipated several social and economic policies of Kemalism such as "populism," etatism, and corporatist experiments with professional associations. In the long run they also helped sustain an enduring interest in corporatism among later generations of Turkish intellectuals, especially through the faculties of sociology and law at Istanbul University. After the I960 revolution these Istanbul intellectuals emerged briefly as the foremost advocates of corporatism and functional representation in the Constituent Assembly that revised Turkey's constitution. The scattered and often vague nature of Gokalp's political writings makes it difficult to describe his ideology satisfactorily. Even a brief examination of his view of professional associations and their relation to the state makes it difficult to accept Berkes's description of him as a "syndicalist."39 Gokalp, himself, used the term "solidarism" (tesanutqiiliik) to describe the program of political reforms that he viewed as the practical consequence of his sociological studies, and he was clearly trying to present himself as the founder of a Turkish variant of the French solidarist movement.40 Steven Lukes has described the French solidarists as "seeking a 'middle path' between laissez-faire liberalism and revolutionary socialism" by ameliorating glaring social abuses while maintaining untouched the bases of existing capitalist society in private property and freedom of enterprise.41 38 Uriel Heyd, The Foundations of Turkish Nationalism (London: Luzac, 1950); Ziya Gokalp, Turkish Nationalism and Western Civilization, ed. and trans. Niyazi Berkes (London: G. Allen and Unwin, 1959), pp. 13-31. 39 Gokalp, p. 15. 40 Ziya Gokalp, The Principles of Turkism, trans. Robert Devereux (Leiden: E. J. Brill, 1968), chap. 9. 41 Steven Lukes, Emile Durkheim, His Life and Work (London: Allen Lane, 1973), p. 249.

94 — Bases of Politics This characterization seems applicable to Gokalp as well, since he also sought to create the appearance of partially embracing socialism in theory in order to neutralize it in practice. For Durkheim, who remained outside the mainstream of solidarism, this involved occasional references to his own ideas as a form of socialism. Gokalp, on the other hand, was both more timid in his approach to socialism, which he tended to equate with communism, and more aggressive in his attacks on capitalism, which he closely identified with European imperialism. But what most distinguishes Gokalp from the earlier advocates of such European "middle paths" as solidarism, classical pluralism, and guild socialism was his emphasis on the subordination of occupational associations to the state and on the state's leadership of a command economy. In this sense his frequent use (or misuse) of the major concepts of Durkheim's sociology provided the basic outline for the ideological justification and the institutional infrastructure of Kemalist experiments with state corporatism as an authoritarian response to the problems of economic underdevelopment. Gokalp made it clear that his attraction to Durkheim's work stemmed not merely from an interest in the scientific study of society but also from a desire to refute the sociology of Marx and to demonstrate that modernization in Turkey would not and should not involve the emergence of class conflict.42 Gokalp regarded Durkheim almost entirely as a theorist with a collectivist image of society who focused on subjective group self-perceptions as the basic data of social analysis. His most frequent references to Durkheim are to provide support for his own contentions, first, that society exists as a force independent of and external to individuals that constrains their internal drives and, second, that social life is made up of consciousness of various collective representations. Hence his characterization of Durkheim as the founder of "social idealism." Gokalp attempted to use the concept of collective representations as the 42

"There are two sociological systems, at once both similar and dissimilar, that may be used in the interpretation and explanation of social phenomena, namely, historical materialism and social idealism; the first was founded by Karl Marx, the second by Emile Durkheim. Superficially the two schools of thought seem to be closely related, for both accept the principle that social phenomena are the results of natural causes and are subject to natural laws. . . . Beyond this point, however, the two systems begin to diverge. . . . For Marx, all social phenomena other than economic phenomena are in the nature of epiphenomena. . . . Marx committed another error by extending this determinism from theory to practice. According to him, the 'people' consist only of the working class, which therefore has the obligation to abolish all other classes. However, since the word 'people' means everyone, it means the totality of all classes which accept the fact that they are equal before the law. . . . AU those who accept that principle are part of the 'people,' irrespective of the professional group to which they belong" (Gokalp, Principles, p. 49).

Attitudinal Bases — 95

unifying theme of his own scientific history of the Turks, which he described in terms of an evolutionary development of their collective sentiments' and their corresponding forms of political organization. The Turks, he claimed, had retained a coherent cultural tradition while identifying with various international civilizations—first, with Eastern civilization; then, with Islamic civilization; and, most recently, with Western civilization. In the course of their historical shifts from one civilization to another, their political organization had evolved as well—from tribalism to leadership of an international empire and, finally, to the nation-state. Gokalp used this evolutionary view as a backdrop for elaborating several hierarchical modes of social organization. He believed that as the Turks developed the common sentiment of nationalism—the most advanced state of collective consciousness—it was essential that they also develop a clear perception of the ' 'natural hierarchies'' that existed between different types of societies and that ordered relations between men and institutions within societies. He frequently described the sociologist as one who corrected the disarray in men's minds by revealing to them both their true national identities and their proper places in relation to other men and institutions. His own sociological studies elaborated three kinds of "scientific classification," all of which were presented as hierarchical typologies—a hierarchical classification of types of societies, of types of nations, and of subgroups within nations. First, he made it clear that the nation is the highest form of society, but that its formation must be preceded by more primitive stages of social organization in clans and multinational empires.43 Gokalp believed that in the latest phase of the evolution of Turkish society from empire to nation the increased division of labor played a critical role. But he also believed that its consequences among the Turks were very different from its consequences in the West. He accepted Durkheim's judgment that in Europe the greater division of labor often led to at least an immediate erosion of common sentiments and the appearance of anomie and class struggles. But he argued that in the Ottoman Empire the division of labor eroded only one type of common sentiment (religion) while strengthening another (culture and language).44 Thus, Gokalp presented the division of labor as the 43 "AU real societies are nations. Ethnic groups cannot suddenly become nations, for they must first pass through the childhood stage of social life as clans and then through a long apprenticeship stage as a community [i.e., a part of an international empire]. Finally, after having become tired of the oppression of the imperium, they leave the community to live independent lives [as nation-states]. . . . Just as a human being must live through periods of childhood and youth, an ethnic group must pass through the stages of clan and community . . . before becoming a [real] society and nation" (ibid., pp. 57-58). 44 "Since a division of labor produces occupational groups which, in turn, produce occupational [rather than class] consciousness, the two collective consciences that had dominated

96 — Bases of Politics

midwife of Turkish nationalism. By promoting consciousness of occupation (meslek) instead of consciousness of class (simf) the increasing division of labor did not disintegrate common sentiments, but served as a catalyst for their transformation into a more advanced and stronger variety.45 In a second hierarchical classification Gokalp ranked nations—the highest form of society—according to their types of social organization. He listed "feudal" nations with a village basis, "communal" nations with an urban basis, and "compound" nations with a mixed village and urban basis. Gokalp concluded this taxonomy by asserting that "the highest form of civilization . . . [is] . . . the corporate nation," whose basic units are ' 'corporate bodies which have a national character.'' Turkey's recent adoption of advanced Western civilization instead of decadent Islamic civilization would require that ultimately it also adopt corporatism as the most advanced type of national organization. "The most advanced nations of Europe are developing in this direction. In light of this classification, it will be seen that the Turkish nation [currently] belongs to the communal type and that in the future it will develop into a corporate nation."46 Finally, Gokalp noted that as Turks progress farther up the evolutionary hierarchy of societies toward the corporate nation, they must also acquire a clear understanding of the relative importance of men and institutions both between nations and within them. First, he described what may be called the "external hierarchy" concerning relations to men of other nations. Gokalp referred to Durkheim's discussion of the conflict between patriotism and cosmopolitanism, or "patriotic morality" and "civic mothe Muslim and Christian communities began to weaken. . . . Whereas every individual previously had viewed his religious community as a social organism and himself as an inseparable part of it, he now began to view as a social organism only his own linguistic society and to consider himself an inseparable part of it alone" (ibid., p. 54). 45 This belief in the benign consequences of the division of labor led Gokalp to a treatment of two other important Durkheimian concepts—mechanical versus organic solidarity and the collective versus individual consciences—that departed significantly from Durkheim's own understanding. For Gokalp the growing division of labor did not involve a fundamental change in the basis of social solidarity from a consciousness of shared similarities to a consciousness of interdependent differences. Rather, it facilitated a change from consciousness of religious similarities to a consciousness of cultural and linguistic similarities. In this sense, then, he did not really envisage a development from mechanical to organic solidarity but from one type of mechanical solidarity to another. Similarly, Gokalp did not share Durkheim's expectation that the collective conscience iyicdari) would recede in relative importance and allow greater freedom to individual consciences (suur). When Gokalp spoke of the division of labor and the transformed collective conscience as resulting in the liberation of the individual, it was not in terms of Durkheim's belief in the reduced need for uniformity and externally imposed consensus but in terms of throwing off the "false consciousness" of Ottomanism in favor of an authentic and latent Turkish identity (Gokalp, Turkish Nationalism, pp. 124, 153, 161-163). 46 Ibid., p. 125.

Attitudinal Bases — 97 rality." This is perhaps the most striking example in which he employs Durkheim's own vocabulary to contradict him. The principle must be adopted that the value of the individual human decreases as he moves from the center to the periphery within the realm of civic morality and increases as he moves towards the center from the periphery. In other words . . . we must view in descending order of importance our fellow countrymen, our co-religionists, our civilization partners, and all mankind, and we must understand them according to the order of their importance (emphasis added).47 Gokalp also emphasized the need for an accurate perception of the "internal hierarchy'' that is formed by various social groups and institutions within nations. The failure of many European nations to understand such internal hierarchies had often led to "an unnatural conflict" between social groups as the division of labor increased. Like the Ottoman chroniclers who compared the role of the statesman to a physician responsible for maintaining equilibrium among the immutable social humors, Gokalp compared the role of the sociologist to the peacemaker responsible for containing and preventing class conflict. There may be contradictions of a pathological nature between the values of the sub-groups within the whole culture group. . . . Class consciousness in Europe gives rise to the feeling of enmity against the fatherland and military service. It is the duty of the sociologist to convert this unnatural state of warfare to a peaceful condition by discovering the natural hierarchy between the social groups (emphasis added).48 Gokalp's description of this natural hierarchy included a pivotal role for professional associations as intermediate structures between the family and the state. There are three principal kinds of social groups: family, political and occupational. Of these, political groups are the most important because a political group is an independent or semi-independent entity which exists for its own sake. Family and occupational groups are merely segments or parts of such an entity. In other words, a political group is a separate social organism of which family groups are the cells and occupational groups are the organs. . . . Thus, when the situation demands, the family should subordinate itself to the corporation, and the corporation should subordinate itself to the state.49 47 48 49

Gokalp, Principles, pp. 64-65. Gokalp, Turkish Nationalism, p. 100. Gokalp, Principles, pp. 57, 105.

98 — Bases of Politics Unfortunately, Gokalp never wrote extensively on the role of professional associations in the corporate nation of the future. He produced nothing comparable to Durkheim's famous "Remarks on Professional Groups," but he did outline the most important aspects of the corporations' structure and functions. He described these in the most general fashion as "the establishment by each occupational group of an internal professional morality after the group has developed a national, country-wide organization." 50 He believed that the traditional guild system was a totally inadequate prototype for such organizations, since it consisted of locally oriented and disconnected units that hindered the development of a national market economy. Accordingly, he strongly approved of the Young Turks' decision to abolish the vestiges of the old guilds. But he did not agree that they should have been replaced with the new private, voluntary associations that Young Turk governments haltingly promoted among the esnaf and business sectors. What he envisaged instead was a far more comprehensive and centralized network of associations that would be established first among the esnaf to integrate regional economies and that would then be extended to coordinate all other occupational categories in a state-directed national economy. Gokalp's outline for the proposed structure of the corporations resembled Durkheim's in suggesting that they be singular and all-inclusive organizations. Unlike Durkheim, however, he favored an extreme degree of organizational centralization. His unusual description of the corporations' structure suggested a series of federations that would be capped by a single, multiprofessional superconfederation. [Lodges] must be organized in every city with a secretary general rather than a shaykh at the head. In each city there should also be organized a central committee composed of delegates of all the different lodges in the city; this committee will be called an is borsasi (business exchange) and will have the task of supervising the common affairs of the lodges in the city and regulating the city's economic life. . . . After lodges have been organized in each city, they will organize themselves into a federation and will establish a headquarters in the national capital . . . [and] organize a confederation and elect members of its General Assembly. Members of the various intellectual pursuits will also create their respective professional federations and join the confederation. Once this has been done, all these professional groups will have been united in the form of a regular army (emphasis added).51 50 51

Ibid., p. 65. Ibid., pp. 106-107.

Attitudinal Bases — 99 As for the functions of these new professional associations, the activities listed by Gokalp did not depart substantially from those cited by Baer regarding the traditional Turkish guilds. Gokalp described at some length the new corporations' expected contributions to professional morality, mutual assistance, and economic development. But he was totally silent on their role in representing the political and economic interests of their members. His writings included no hint of Durkheim's proposals that functional representation replace geographic representation in the legislature or that the corporations serve as intermediaries in a two-stage electoral system. The most important function of the associations was to support "solidarism," which was the term he used to describe his vision of the corporatist state in Turkey. He argued that professional associations would advance solidarism in two major ways. First, they would develop professional ethics and morality in such a way as to promote among their members the consciousness of interdependent occupations rather than the consciousness of antagonistic classes. Second, they would support and help implement a state-directed policy of planned economic development that would focus on heavy industrialization and protectionism to rescue Turkey from permanent economic dependence on the West. Gokalp believed that this kind of corporatist nationalism would amount to a unique Turkish model of social and economic development that would differ from both capitalism and socialism by not discriminating against the interest of any particular class. Private ownership would be preserved and increased, but its wealth would be invested by the state for the benefit of society as a whole. Only the expanded economic role of the state, supported by new corporatist professional associations, could guarantee social solidarity by harmonizing conflicting interests and balancing individual freedom with social equity. As the Turks love freedom and independence, they cannot be communists. But as they love equality, they cannot be individualists. The system most suited to Turkish culture is solidarism. . . . Together with private ownership there must also be social ownership. . . . The new Turkey, which has to introduce the latest and most developed techniques of Europe, cannot afford to wait for the spontaneous rise of the spirit of enterprise among individuals in order to industrialize. . . . Turks are temperamentally etatists. They expect the state to take the initiative in everything new and progressive. . . . The state . . . must become an economic state. . . . The modern state . . . is like a business concern . . . that will at the same time perform a moral service because the rise of a new class of speculators will be prevented. . . . If we accept the

100 — Bases of Politics principle of state capitalism, we will be able to prevent the rise of insatiable and predatory capitalists in our country (emphasis added).52 Gokalp's writings on occupational associations and their role in an etatist strategy of economic development were a better general guide to subsequent Kemalist practice than were his writings on most other topics. Kemalist experiments with state corporatism as an authoritarian response to the problems of economic underdevelopment were more modest and less sustained than Gokalp's proposals, but they were generally faithful to his original formulations. In one sense, however, Gokalp was quite unwittingly prophetic. He believed that the modern state would become' 'like a business concern" and that future professional associations would be organized "in the form of a regular army." But he could not have foreseen that when Turkey's second major initiative toward corporatization arose after 1960, it would involve an important new entrepreneurial role for semiofficial occupational associations whose prototype would be supplied by the Turkish army itself. THE KEMALIST CONCEPT OF POPULISM

Kemalist ideology is commonly dismissed either as a purely pragmatic response to Turkey's specific economic needs during the collapse of world capitalism or as an eclectic and superficial imitation of German sozialpolitik, Italian corporatism, and Soviet economic planning. But Turkish insistence that Kemalism represented something distinctively native can be taken more seriously if we appreciate the degree to which it incorporated several elements of Ottoman statecraft and Gokalp's corporatist nationalism. This is particularly evident in the changing interpretations of the concept of populism (halkylik), which became the main Kemalist principle underlying social policy and associational life during the one-party era. The term "populism" was first used by Ataturk during the War of Independence in place of the word cumhuriyet (republic) in order to express support for popular sovereignty without offending his more conservative supporters. In this period populism was simply equated with nationalism and antiimperialism or used as a shorthand term embracing all of the goals of the liberation movement. But even before the declaration of the republic, nationalist leaders began to define populism more specifically as a description of the sociological bases of the new nation-state. Ataturk's early descriptions of the national community and its sources of solidarity were very similar to those of Ziya Gokalp. At first he spoke merely of solidarity based on common cultural identity, as in his characterization of the new 32

Gokalp, Turkish Nationalism, p. 312.

Attitudinal Bases — 101

Turkey as "a society united in terms of race, religion and culture, filled with feelings of mutual respect and sacrifice, and having a common fate and interest.'' By 1923, however, Atatiirk, like Gokalp in his later writings, became increasingly concerned with promoting solidarity based on the functional interdependence of different social and economic groups. "In my opinion, our nation does not possess various social classes that will pursue interests that are very different from one another and that will, accordingly, come into a state of struggle with each other. The existing classes are necessary and indispensable to one another." With this statement of the concept, "populism" became increasingly distinct from popular sovereignty and cultural solidarity and was identified instead with the effort to encourage economic development and preserve social peace by maintaining equilibrium among the producing classes.53 For a brief time it appeared that this new interpretation of populism would involve the creation of permanent institutional mechanisms for harmonizing interest conflict and promoting collaborative class relations. Materials recently collected by Gunduz Okciin indicate that this was commonly assumed to be the reason for organizing the Izmir Economic Congress of 1923. The congress was an unprecedented "economic summit meeting," arranged by the government and attended by over eleven hundred delegates, who were freely elected from four major economic sectors—agriculture, trade, artisans and craftsmen, and labor. The delegates of each sector met, first, separately and, then, in a General Assembly to hammer out specific proposals for economic reform that were to be submitted to the government and parliament as part of a nationwide debate on developmental policy. The common program that was finally adopted by the congress was referred to as the Economic Pact (Misak-ilktisadi) and was publicized as a collective commitment to strive for economic independence in order to supplement and complete the earlier commitment to formal political independence symbolized by the famous National Pact {Misak-i Milli) of 1919.54 Although several of the Izmir Economic Congress's recommendations were eventually incorporated into the government's brief attempts to encourage private enterprise, the experiment in formal, multisectoral consultation was never repeated. Only two years after the Izmir congress, labor unions and nearly all other voluntary associations were closed down and the remaining economic sectors were organized around singular and compulsory corporatist structures whose purpose was to increase govern53 Ismail Arar, Atatiirk'iin Halkgihk Programi ve Halkqihk Ilkesinin Tarihfesi [Atatiirk's populist program and the history of the principle of populism] (Ankara: Baha Matbaasi, 1963). 54 Gunduz Okciin, Tiirkiye lktisat Kongresi, 1923-Izmir [The Turkish Economic Congress, Izmir—1923] (Ankara: Siyasal Bilgiler Fakultesi, 1971).

102 — Bases of Politics

ment regulation and control rather than promote associational consultation. Turkish governments did not resume such economic summit conferences on even an irregular basis until the adoption of comprehensive central planning after the 1960 revolution. The long hiatus between these modest attempts of the state to foster collaborative class relations coincided with a revised understanding of populism whose widespread acceptance among Turkish political leaders has been challenged seriously only in the last decade. The revision of populism was a subtle but very important one whose enduring and limiting effects on interest group politics are still observable in the attitudes and behavior of current party and association leaders. The reinterpretation can be described as a change from the recognition that certain fundamental social classes existed but that their conflicting interests were reconcilable to the assertion that Turkey currently was and always should remain a classless society in which either a single party or the state itself could simultaneously represent and regulate all economic interests. People's Party publications often declared, "The Turkish nation is the most classless nation in the world. . . . We are a nation without classes, without differences."55 One of the clearest expressions of this view by Atariirk was the following statement to a group of People's Party candidates in 1931, shortly before the principle of populism was incorporated into the constitution. The statement began with a repetition of Gokalp's call for occupational solidarity and ended by linking social peace to the state's application of distributive justice among all social groups. One of our basic principles is to consider the people of the Turkish Republic not as composed of separate classes but as a community divided among various occupations with regard to the division of labor for individual and social life. Farmers, esnaf, laborers and workers, free professionals, industrialists, merchants and civil servants are the main working groups comprising Turkish society. The labor of each of these is indispensable to the life and well-being of the others and society in general. The goal at which our party aims with this principle is to obtain social order and solidarity instead of class conflict and to establish harmony among interests so that they will not injure one another. Interests will be balanced according to their degree of capability, knowledge and contribution (emphasis added).56 Ataturk did not indicate any explicit hierarchical ordering of occupations, but this famous description of populism departed significantly from his 55

Walter Weiker, Political Tutelage and Democracy in Turkey (Leiden: E. J. Brill, 1973), p. 218. 56 Arar, p. 31.

Attitudinal Bases — 103

own earlier endorsements of social egalitarianism and recalled instead Ottoman statecraft's traditional concern with increasing economic prosperity by maintaining a "weighted balance" between unequal functional categories. This understanding of populism was accompanied by a shift toward etatist economic policy and provided much of the ideological justification for Kemalist experiments with state corporatism during the 1930s. Between 1932 and 1934 official sponsorship was given briefly to a form of corporatist nationalism promoted in the journal Kadro by such intellectuals as Yakup Kadri Karaosmanoglu and Sevket Siireyya Aydemir. The Kadro group elaborated on Gokalp's earlier social and economic writings and portrayed the Kemalist revolution as leading an international liberation movement struggling against capitalism and imperialism. They argued that "for an under-developed country such as Turkey the main question was not the reconciliation of economic interests, as seen in economically developed countries, but the creation of an advanced technology, and of a society without social contradictions." They claimed that in Turkey neither capital accumulation nor class struggle had developed significantly, and they believed that Turkey would provide a unique historical example of the state accomplishing the former while avoiding the latter. In other words, the Kadro group portrayed the issue of class harmony in Turkey as being so unproblematic that even the Izmir Economic Congress's ad hoc organization and aggregation of divergent interests were seen as unnecessary. Social classes were described either as nonexistent or as in such a rudimentary stage of formation that they could be both represented and regulated by a unified party-state apparatus, led by an elite or "cadre" that would form a kind of "universal class." 57 The Kadrists' support of corporatist strategies of development and representation did not prevent their eventual suppression by rivals, who detected the threat of communism even in their basically anti-Marxist discussions of class relations. But their basic strategy of etatist corporatism was continued under the leadership of Recep Peker, who had headed the opposition against them. Under Peker's chairmanship the People's Party and the state were eventually merged into a single authoritarian apparatus that was to serve both as the supreme arbiter in social relations and as the chief organizer of economic enterprises. The People's Party soon invoked the concept of populism to describe itself as "the synthesis of the people" and as the sole authoritative interpreter of the national interest. Populists were described as "those who do not accept privileges for any family, 57

73.

Kemal Kaipat, Turkey's Politics (Princeton: Princeton University Press, 1959), pp. 70-

104 — Bases of Politics

class, group or individual."58 The representation of such specific interests through alternative, nonparty, or nonbureaucratic channels was regarded as unnecessary and as encouraging the acquisition of illegitimate privilege. Accordingly, Turkish associational life was virtually nonexistent between 1925 and 1946 with the exception of a few modest attempts to sponsor corporatist professional organizations. During the transition to multiparty politics the People's Party again revised its view of populism. Still avoiding the word "class," the new definition nevertheless recognized that "various social groups make up the whole of the nation" and described the party's task as "harmonizing" the interests of those groups.59 At the same time the Republicans amended the Law of Associations by eliminating the ban against associations based on class interest, thereby allowing the reestablishment of labor unions and the emergence of employers' associations. Although populism was now seen as justifying class-based voluntary associations, it continued to reject the notion of class-based political parties. Long after the rapid expansion of associability during the multiparty period, associations were formally prohibited from engaging in "political activity" and party leaders remained very reluctant to appeal overtly for the support of distinct social and economic groups. Republican, Democratic, and Justice Party leaders have commonly tried to portray their organizations as "national parties" or "mass parties" that represent the interests of all social groups and seek to reconcile them with each other and with a higher national interest rather than as "class parties" that represent the interests of specific groups against those of others. Although this understanding of populism is a legacy of the People's Party's early pretension to provide singular and universal representation, it eventually became even more characteristic of the ideological positions of the two leading bourgeois parties. Fuad Kopriilu's early description of the Democratic Party's views on political representation often have been repeated by Justice Party leaders, especially in criticizing the People's Party's recent shift to the "left of center." We recognize the social classes. We completely accept that they have special interests and that if not regulated these interests may contradict each other. We find it useful and necessary for various classes in a modern society to organize themselves for work and activity in professional associations, trade unions, and cooperatives . . . but we cannot accept the fact that the social classes have irreconcilable interests and 58

Arar, p. 15. "Karpat, p. 311.

Attitudinal Bases — 105 have to struggle with each other. Such a conception in our view is outdated and baseless. . . . This is the sole reason why the Democratic Party is not a class party producing conflicts of interests among the social classes, but on the contrary is a "national party" assembling around itself all those citizens believing in the above principles.60 Of course, the persistence of this notion of populism has not prevented the emergence of an increasingly clear correspondence between the interests of different social groups and the programs and policies of different parties. But such avowed hostility to bargaining with or representing partitive interests has tended to obscure the process, to make it fitful and protracted, and to lessen the ability of most interest group leaders to form stable alliances with party leaders. Karpat's description of the ambivalence of Turkish politicians toward seeking support from specific social and economic groups during the earliest years of the multiparty period is still relevant even during the recent realignment of the national party system. "Whenever a political party made a move toward defining more clearly its own social basis, the next move was an immediate reaction in the opposite direction that left it, socially, more confused than ever." 61 In conclusion, then, it is clear that changes in elite views of legitimate interest representation generally have not matched the far more rapid changes in socioeconomic structure and association formation. Throughout the multiparty era much of the political elite has continued to share a lingering fear that unless partitive interests are repressed, closely regulated, or prudently harmonized, divisions along such lines as class, religion, and region will threaten both the unity of the nation and the authority of the state. Such current apprehensions are very similar to those of the reformist bureaucrats during the period of defensive modernization and of the early nationalists during the first years of the republic. Their persistence may not seem warranted in view of the very different objective circumstances under which contemporary politicians operate. The population of modern Turkey is relatively homogeneous culturally and linguistically. Republican regimes have been highly successful in promoting widespread identification with the new secular nation-state. The effective control and penetration of the state's administrative apparatus are unchallenged within the nation's territory, and a generally peaceful international context combined with impressive military strength has provided long-term national security unknown to earlier political elites. Nevertheless, the enduring hostility of much of the political elite to the 60 61

Ibid., p. 312. Ibid., p. 310.

106 — Bases of Politics independent organization and expression of partitive interests has preserved the salience of center-periphery tensions as a key dimension of political conflict in modern Turkey. Of course, such tensions are no longer the result of attempts to shore up the crumbling power of the state apparatus against the imminent dangers of social disintegration, political separatism, and foreign aggression. Instead, center-periphery tensions are increasingly manifested in a series of confrontations and compromises between the modernizing state pursuing ambitious goals of economic development and the representative associations of newly emergent social and economic groups that favor conflicting distributions of the benefits and sacrifices of modernization. Not only has the strength of the political center increased considerably, but the nature of demands from the periphery has also changed from attempts to usurp state authority and threats of secession to demands for greater participation in decision making and for larger shares of the nation's growing economic surplus. In other words, the prototype for current center-periphery relations is not the Covenant of Union imposed upon Sultan Mahmud by the ayan but the Economic Pact presented to the new nationalist government by the delegates of the Izmir Economic Congress. One aspect of current center-periphery tensions, however, is strikingly similar to the premodern pattern—the continuing tendency of the state to rely on grants of precarious and revocable privilege to those who cooperate in fulfilling its economic goals and to resist the attempts of those enjoying such privileges to extend them and transform them into permanent and secure rights. In this sense corporatist strategies of development attempt to provide new institutional structures for offering precarious privilege. Modern political elites are seeking to organize and manipulate representative associations as relatively noncoercive mechanisms for preserving the dominance of a strengthened political center over an increasingly differentiated periphery. Previous techniques such as the repression of voluntary associations and their colonization by ruling political parties certainly have not been discarded. But they are gradually being supplemented by attempts to enlist the support and cooperation of still weak and insecure interest group leaders in implementing economic policies without substantially increasing their political autonomy or widening their roles in policy formation. In general, this has involved state sponsorship of compulsory, hierarchically organized, semiofficial representational cartels as a means of organizing and regulating various economic sectors. As we shall see in the following discussion of public policy regulating associations, in many ways such structures are analogous to the premodern Turkish guilds. But their expanded responsibilities and their gradual extension to most other major

Attitudinal Bases — 107 occupational categories have provided contemporary elites with far greater resources to realize the traditional aims of Ottoman statecraft—coordinating and balancing various sectoral interests while differentially rewarding them according to their unequal contributions to preserving social order and increasing general prosperity.

CHAPTER 4

Public Policy Toward Associations: Formal-Legal Bases

toward associations has been influenced both by political culture and by the changing problems of economic development. The general impact of political culture has been to create an unfriendly official attitude toward associational interest representation and political activity. The freedom of association has had a fitful and uneven history in modern Turkey and the legitimacy of interest group politics, as distinguished from party politics, has been recognized only recently and tentatively. Ruling elites have promoted a formally bifurcated set of associational structures containing both corporatist and pluralist elements. But group legitimacy has always been more easily extended when interest representation has been based on the precarious privileges of semiofficial corporatist associations rather than channeled through the more autonomous private associations. The impact of economic development on public policy toward associations has not been a simple matter of allowing increasing social differentiation to become more clearly reflected in the expansion and greater variety of associational life. At times the state has encouraged associational organization so that it could more clearly represent conflicting interests within a changing social structure. But at other times the state has sought to contain or prevent interest conflict and to restructure society to match elite visions of the future. The result has been periodic shifts between pluralist and corporatist policies as successive governments have attempted to shape associational life according to their changing strategies of economic development and political domination. In this section we will highlight the eclectic and fluctuating nature of the state's associational policies. We will see not only that divergent policies have been adopted in different periods but that different policies have existed toward various groups in the same periods. The discussion will focus on three principal topics: the history of the freedom of association in Turkey; the formal-legal bifurcation of associational life into pluralist and corporatist sectors; and the manner in which periodic fluctuations between pluralist and corporatist strategies have been linked to the changing imperatives of delayed capitalist development. Our purpose is to provide PUBLIC POLICY

Formal-Legal Bases — 109

a descriptive summary of historical changes in public policy toward all associations, but general points will be illustrated by more extensive references to the experiences of the labor movement and esnaf associations.

A HISTORY OF THE FREEDOM OF ASSOCIATION IN TURKEY

Public policy toward voluntary associations has been characterized by abrupt fluctuations in which the increasingly modest promotion of the freedom of association has been followed by increasingly severe limitation. The history of associational freedom in Turkey is basically a combination of tentative constitutional guarantees with recurrent statutory abridgments. This can be seen through a comparative analysis of successive constitutional provisions concerning the freedom of association and of various Laws of Associations adopted between 1909 and 1972. These have provided a common legal framework for the regulation of all voluntary groups and, therefore, represent the most comprehensive overview of associational policy. The first Ottoman constitution, in 1876, contained no guarantee of associational freedom because of Abdiilhamit's objection. Seeking to evade the sultan's network of spies and informants, opposition groups organized abroad, not as self-declared political parties, but as secret societies and masonic lodges. Thus, in its earliest form, associational activity was commonly identified with foreign-based, conspiratorial groups that thinly disguised themselves as voluntary associations while agitating for the overthrow of the government and the reinstatement of the constitution.1 This experience reinforced the original Ottoman suspicion concerning the potentially subversive nature of secondary associations, and the legacy of that suspicion is apparent in the recurrent tentativeness with which subsequent constitutions have included the right of association among fundamental freedoms. Even when the most successful of the early conspiratorial "societies," the Committee of Union and Progress, reestablished representative government, it was particularly careful to provide for broad limitations on associational freedom. The 1908 constitution was the first to state forthrightly that "Ottomans possess the right of association." But this "guarantee" was immediately qualified by the vague and sweeping prohibition against "the formation of associations that violate the territorial integrity 1 Ahmet Yiicekok, Tiirkiye'de Orgiitlenmi$ Dinin Sosyo-Ekonomik Tabam [The socioeconomic basis of organized religion jn Turkey] (Ankara: Siyasal Bilgiler Fakiiltesi, 1971), pp. 104-105; Bernard Lewis, The Emergence of Modern Turkey (London: Oxford University Press, 1961), pp. 156-166.

110 — Bases of Politics

of the Ottoman State, that operate to change the [constitutional] regime and government, and that are contrary to civility and order." 2 The two constitutions promulgated during the Republican era were considerably more emphatic in endorsing the principle of associational freedom, but they also included explicit statements recognizing the need to abridge that freedom. While the restrictive language used in the republican constitutions has been generally less urgent and harsh than that used by the Young Turks, it hardly has been less ambiguous. The constitution of 1924 proclaimed freedom of association to be "a natural right of the Turks." It discarded the 1908 constitution's list of prohibited associations, but included no criteria to guide the application of its own proviso that "the limits of associational freedom shall be set forth by statute." The liberal 1961 constitution provided by far the most explicit protection of associational freedom, stating that "everyone possesses the right to establish an association without obtaining prior permission." This time the endorsement of the right of association included a modest attempt to limit the conditions of its statutory abridgment: "This right can be limited by law only in order to protect public order and morality." But, again, expression of the need to restrict the right seems at least as strong as expression of the right itself.3 If it can be argued that this series of constitutional changes reflected a 2

Yucekok, p. 104. Ibid., p. 105. Perhaps the most important innovation of the 1961 constitution regarding associational life was its explicit differentiation between voluntary associations, labor unions, and political parties and the requirement of separate sets of regulatory legislation for each type. The framers of the constitution sought to bolster the legitimacy of each type of organization, but they considered political parties rather than interest groups to be the most effective and reliable channels of representation. In their concern to prevent the reemergence of the bitter bipolar factionalism of the 1950s, the framers were eager to widen the scope of social and economic interests represented in parliament. While there was general agreement that factional violence might be contained by encouraging cross-pressures and checks and balances among a more diverse set of interests, there were two clearly opposing choices for structuring the representation of those interests. One group sought to break the legislative monopoly of the parties and career politicians by injecting greater professional and technical expertise into parliament. They supported the adoption of functional representation so that deputies from political parties in the National Assembly would be supplemented by deputies chosen directly from occupational associations in the new Senate. But the Constituent Assembly ultimately rejected functional representation in favor of attempts to strengthen and diversify the party system. Proposals for proportional representation were approved on the grounds that it would promote "more principled" and more cohesive party organizations and that these would be able to manage interest conflict more effectively than either the old, rigid two-party system or the still weak and untested network of professional associations. C H . Dodd, Politics and Government in Turkey (Manchester: Manchester University Press, 1969), pp. 106-115; Walter Weiker, 7"Ae Turkish Revolution 1960-61 (Washington: The Brookings Institution, 1963), pp. 64-81. 3

Formal-Legal Bases — 111

long-term tendency, however gradual and tenuous, toward the greater legitimization of associational life, that tendency was certainly reversed by the constitutional amendments that were demanded by the military junta in 1971. The principal amendment concerning associational freedom retained the 1961 constitution's guarantee of association formation, but revived the restrictive and vague language of the 1908 constitution and, for the first time, gave the government the fundamental right to suspend associational activity. Every individual is entitled to form associations without prior authorization. The forms and procedure to be applied in the exercise of the right are regulated by law. The law may place restrictions for purposes of safeguarding the integrity of the State with its territory and people, national security, public order and public morality. No individual can be coerced into becoming a member of an association or into retaining his membership. Associations can be closed down by court judgment in cases prescribed by law. In instances where delay is deemed prejudicial for the [abovementioned] purposes . . . such associations may be prevented from operating by order of the competent [executive] authority clearly designated by law until such time as a court judgment is made.4 The most novel aspect of the 1971 constitutional revisions was not the further jeopardy in which they placed the already precarious freedom of association but the clear intent to make that precariousness permanent by resurrecting discarded statutory restrictions on associational activity and raising them to the level of fundamental law. If the intent of the 1961 constitution was to discourage associational legislation from becoming arbitrarily restrictive, the intent of the 1971 amendments was to prevent such legislation from becoming too permissive. Voluntary organizations have been regulated according to four major pieces of legislation—the Laws of Associations adopted in 1909, 1938, 1946, and 1972. These special codes demonstrate the uneven history of associational freedom more directly than the omnibus constitutional guarantees they supposedly supplement and clarify. The first and third statutes are commonly recognized as relatively liberal, since they place only min4 Constitutional Amendments (Ankara: State Information Organization, 1972), p. 6. Another amendment applied similar language to labor unions and opened the way for broad interference in unions' internal affairs: "The regulations, management and operation of trade unions and federations shall not conflict with democratic principles.'' Still another amendment revoked the right of civil servants to join either trade unions or political parties and required that violations be punished by permanent dismissal from public service (ibid., pp. 8-9, 1213).

112 — Bases of Politics

imal restrictions on association formation and leave associational activity generally free from executive control. But the closely related statutes of 1938 and 1972 place numerous and detailed limitations on association formation, internal organization, and activities while allowing extensive official interference and surveillance. This body of legislation reflects a continued suspicion toward associational life, especially through its common sensitivity toward organization along lines of class, religion or ethnicity, affiliation with political parties and international organizations, and any activities not specifically defined in association charters. The first Law of Associations, adopted in 1909, is the most straightforward and least restrictive of the four statutes. Its major prohibition simply states that no association may be secret or rely on separatist appeals to the empire's various nationalities. This led to the closing of several Greek and Bulgarian societies in Rumelia, but associations founded for cultural, literary, or philanthropic purposes were still allowed to express demands for greater autonomy for the empire's non-Turkish and nonMuslim populations. In the earliest years of the republic the formal freedom to establish associations remained unhindered. A revision of the Law of Associations in 1923 gave the executive control over associational activities, but this interference was limited to cases in which associations were in clear violation of the law and the responsibility to suspend an association's operations was reserved for the judiciary.5 Although the formal freedoms of the 1909 law were not officially abrogated until shortly before Ataturk's death, they were effectively emasculated by the "independence tribunals" established under the Law for the Maintenance of Public Order between 1925 and 1929. Opposition parties, religious orders, and trade unions were the principal targets of these extraordinary chambers. After 1924 the major form of association promoted by the regime was the People's Houses, which disseminated nationalist and secularist propaganda through lectures, libraries, concerts, sports events, and social aid groups. These were intended to supplant and absorb preexisting associational life and were formally integrated into the structure of the People's Party in 1931. By 1935 most other private associations had closed in anticipation of the drafting of a new Law of Associations and most of those that remained were innocuous local sporting clubs. Some groups, such as the traditionally suspect Masons, chose to demonstrate their patriotism and solidarity with the regime by dissolving their organizations voluntarily.6 The Law of Associations of 1938 introduced detailed restrictions on the 5 6

Yiicekok, pp. 105-106. "Masonlar" [Masons], Yanki, 4 September 1972.

Formal-Legal Bases — 113

freedom of association. Association formation was, in effect, impossible without prior permission from the government. Association founders were required to present copies of their constitutions and bylaws to the provincial governor and to obtain an official certificate of registration. The list of illegal associations was lengthened considerably to include: groups whose goal is to destroy the territorial integrity of the state and political and national unity; groups based on religion, sectarianism or mystical orders; groups based on family, community, class or racial loyalties; groups that are secret or that conceal their goal; political groups that pursue the goal of regionalism or carry a regionalist title . . . associations of civil servants.7 Associations were required to have only a single goal, clearly stated in their constitutions, and were prohibited from "engaging in political activities." Direct permission from the Ministry of Interior was necessary for the establishment of federations, affiliation with international organizations, and participation in international meetings. The financial resources of an association were limited primarily to membership dues and these could not exceed T.L. 120. Associations in violation of these provisions were threatened with revocation of their official registration, suspension of their activities, and confiscation of their property. Stiff fines and prison sentences were also provided for convicted association leaders. Thus, while the 1938 code contained relatively few provisions for direct government surveillance of associational life, it granted the Ministry of Interior very broad discretion in controlling association formation and virtually summary powers to close down voluntary groups.8 The 1946 Law of Associations preserved the bulk of the 1938 statute, but included several liberalizing amendments that considerably encouraged associational development throughout the postwar era. Its most important provision freed association formation from government control. Associations could henceforth be established without prior permission simply by presenting a copy of the group's constitution to the Ministry of Interior and publishing public notice in a newspaper within fifteen days of the organization's formation. A few of the most restrictive definitions of illegal associations in the 1938 law were repealed. The section prohibiting groups that destroyed "political and national unity" was discarded. The ban on "class-based" associations was lifted and labor unions were again per7

"Cemiyetler Kanunu" [Law of Associations], Resmi Gazette 14, no. 3512 (July 1938). «Ibid.

114 — Bases of Politics

mitted to organize. Civil servants were finally allowed to form professional associations, though they were prohibited from joining labor unions. Associations were still prevented from engaging in activities outside their originally declared goals, especially from "political activities," but the punishment for such violations was reduced from forced closure and possible imprisonment to light monetary fines. Another key provision of the 1946 code restored to the courts the sole authority to dissolve associations.9 The relative ease with which associations could be established led to an immediate upsurge in associability in the major cities immediately after 1946. When the Democratic Party assumed power in 1950 voluntary associations multiplied at unprecedented rates, spreading throughout nearly every geographic region. The rapid emergence of a diverse collection of voluntary groups during the period of multiparty politics is often cited as evidence of the success of public policy in stimulating the development of a "pluralist democracy." In fact, the promotion of such a large number of rival, fragmented, and financially feeble groups was part of a general policy of discouraging well-organized and effective collective action, especially in the major occupational sectors. Throughout the postwar era the new freedom of association formation has been coupled with the continuation of broad restrictions on associational activity, so that voluntary groups have remained dependent on some form of political sponsorship and highly vulnerable to the retaliation of hostile governments. Although this policy has been more liberal than the 1938 Law of Associations, it seems most appropriate to characterize it as a policy of "debilitating pluralism" designed to promote a multitude of weak and manipulable organizations. Whereas the 1938 code sought to prevent the organization of factions through direct repression, policy in the multiparty period has applied a perverse version of the Madisonian strategy of trying to contain factionalism by encouraging the proliferation of fragmented associations whose questionable legitimacy makes them all insecure. Even the most liberal Turkish associational regulations tend to be so complex, limiting, and unclear that it is virtually inevitable that assocation leaders will violate them sooner or later and thus be subject to official reprisal. On the other hand, official cooperation or sufferance is often necessary if an association is to perform even its narrowly defined legal activities. 9

"Cemiyetler Kanunu," Resmi Gazette 10, no. 6329 (June 1946); Tiirkiye Biiyiik Millet Meclisi Dergisi, 8 May 1952, no. 5927; T.B.M.M. Dergisi, 6 July 1964, no. 490. This guarantee of due process, however, was soon qualified by an amendment in 1952 allowing courts to suspend associational operation and to sequester property before rendering a final judgment. A 1964 amendment revived the right of direct executive intervention by allowing the police to enter association headquarters without a court order in cases where "delay would be prejudicial to national security or public order."

Formal-Legal Bases — 115 Every ruling party has used public funds to sponsor favored groups. The result has been a common dependence on personal influence (torpit) to get things done along with possible punishment for seeking that influence or for doing anything at all. Associations are deterred from "political activity," yet remain constantly vulnerable to party colonization. Every government is able to exploit the limited legitimacy of associational activity by closing its eyes to the infractions of groups it supports or finances while retaining the right to attack or close down groups it finds objectionable. During the 1950s Democratic Party governments had little difficulty in maintaining this fagade of associational freedom while exploiting the vagueness of the law to reward its supporters and harass its opponents. But after 1960 ruling parties had far less maneuverability in their attempts to control and manipulate voluntary groups. Arbitrary and self-serving applications of the law became more difficult in view of the 1961 constitution's obvious intent to encourage the representation of diverse social and economic interests through private associations and to impose further checks on majority party power. This became apparent as an increasingly independent judiciary frequently decided in favor of association leaders appealing against the questionable intervention of Justice Party governments. Furthermore, the gradually more open attempts of the major occupational groups to influence policy and public opinion through lobbying, electoral endorsements, and public demonstrations received such clear encouragement from all of the competing political parties that the formal ban on associations engaging in political activities became increasingly irrelevant and unenforceable. After 1965 Justice Party governments became alarmed by the growing tendency of private associations to engage in interest group politics and to collaborate with opposition parties. Government spokesmen argued that this represented not the development of a more truly pluralistic regime but a threat of anarchy and a challenge to the authority of the state. Yet the Demirel governments generally avoided the severe repression of associational life that had been common under earlier regimes, mainly because they did not want to provide the military with the same excuse it had used earlier in overthrowing the Democrats. Instead, the ministers of justice and interior drafted a new Law of Associations that closely resembled the 1938 code. They led a publicity campaign against what they claimed was an uncontrolled proliferation of subversive activity that was exploiting the freedom of association in order to destroy majority rule and deny freedom to others. The ministers dramatized the increasing ideological polarization and violence among student organizations in order to justify their demands for reestablishing sweeping government controls over all types of voluntary

116 — Bases of Politics

groups.10 Because the Justice Party realized that such restrictions on associational freedom were likely to be nullified by the courts even if accepted in parliament, they emphasized the issue as one of the principal reasons that voters should grant them the two-thirds legislative majority necessary for constitutional revision. Of course, when pressure finally did arise for redrafting the liberal 1961 constitution, it came not from an outraged electorate but from a military junta, and the statutory limitations imposed upon the freedom of association were far more severe than anything the Justice Party could have expected to achieve unilaterally. The 1972 Law of Associations is the most detailed and comprehensive piece of legislation ever drafted for the regulation of Turkish voluntary groups. It incorporates for the first time in a single statute several provisions that previously were scattered in the Civil and Penal codes, the Political Parties Law, and the Law on Public Assemblies and Demonstrations. While preserving the general right to establish associations without prior permission, the new law revived and considerably extended the original 1938 list of illegal associations and prohibited activities. This new list represents a summary statement of the major attacks recently leveled against Turkish private associations and includes several points presented in the indictments against political prisoners who were tried in the special military tribunals after March 1971. Section 1 of the law states that It is prohibited to establish associations that have more than one goal or that pursue any of the following goals: a) to destroy the indivisible integrity of the Turkish State and nation; b) to destroy the Republic whose characteristics are described in the Constitution by relying on differences of language, race, class, religion and sect; c) to destroy human rights and freedoms; d) to engage in activities that are contrary to law, public order and general morality; e) to provide a dominant or privileged position for their members over others by relying on loyalties of region, race, class, religion or sect; f) to establish the domination of one social class over others or to destroy a social class, or to overthrow any established fundamental economic or social systems or to destroy the State's political and legal systems; g) to make the State's fundamental social, economic and legal systems rely upon religious teachings, even partially, or to exploit or abuse religion, religious sentiments or things considered holy in whatever manner for the purpose of achieving political or personal interest or influence; 10

Cumhuriyet, 10 October 1968.

Formal-Legal Bases — 117

h) to eliminate or change the reformist laws secured in the Constitution of the Turkish Republic or to revive matters prohibited by these laws; i) to defame or ridicule the personality, activities or memory of Ataturk; j) to support or oppose a specific political party, or to establish cooperation among political parties, or to support or prevent the victory of a political party or any of its candidates or of independent candidates in elections for Parliament, local or precinct offices, or to establish cooperation among these candidates; k) to defame or ridicule the personality of the Turkish State; 1) to commit crimes or to encourage or suggest their commission.11 The major innovation of the new law was to grant unusually broad government authority for the continuous surveillance and inspection of associations' internal organization and operations. The Ministry of Interior was directed to establish a special Bureau of Associations under the chief of police in every province to supervise the registration of all existing and new associations. The statute specified in great detail the kinds of internal committees that each association must establish, including a Committee on Discipline; the responsibilities of each committee; the methods for convening them; the places in which they may meet; procedures for conducting the meetings, determining their agenda, and electing officers; procedures for establishing branches, federations, and confederations; and official inspection of financial records no less than once every three years. Local officials were authorized to appoint "government commissioners" to act as observers in association meetings. It was the commissioner's responsibility to monitor the meetings, to determine the identity of all persons in attendance, and to ensure that there are no violations of the prescribed procedures. He was authorized to tape-record, photograph, and film all proceedings, to postpone the meeting in case of public disorder, and, if necessary, to request police assistance in enforcing his decisions.12 11

"DerneklerKanunu," Resmi Gazette 2, no. 14379 (December 1972), section 1 of law. Other sections concerning illegal associational activities prohibit meetings and publications that praise governments, parties, or individuals, living or dead, who advocate subversive ideologies. They also prohibit opposition to the legal rights of armed forces personnel, military or civil defense training and the use of special uniforms and armbands, the possession of arms and explosives, and receiving financial assistance from political parties, labor unions, or employers' associations. Several other activities were made dependent on prior government approval. Contact with the representatives of foreign associations, in Turkey or abroad, requires approval of the Ministry of Interior. Use of the words "Turkish," "Turkey," "national," "republic," and "Ataturk" in the name of an association requires the consent of the Council of Ministers. Finally, copies of all announcements to be broadcast on radio or television must be filed in advance with the office of the local public prosecutor (ibid., sections 6, 9 of law). 12 Ibid., sections 7, 8 of law.

118 — Bases of Politics

Even before martial law was suspended and arrangements were made to hold the 1973 parliamentary elections, the People's Party moved to have the Constitutional Court overturn the new Law of Associations. Eventually the court did declare that certain provisions violated the rights of assembly and travel, that they imposed a form of indirect censorship, and that the entire statute would have to be redrafted. But the revised law did not seriously alter the expanded list of illegal associational activities or the executive's broad powers of surveillance and intervention. Thus, throughout the 1970s, coalition governments led by the Justice Party were able to revive and expand the old Democratic Party practice of selective repression by strictly enforcing the law against troublesome student and labor groups while tolerating or encouraging its violation by supporters of its right-wing coalition partners. The fitful history of the freedom of association in Turkey has involved a particular sensitivity toward preventing and containing class conflict that deserves special attention. This is seen most clearly in the shifting pattern of public policy regulating workers' organizations.13 Motivated by the constant fear that economic development would produce heightened class antagonisms, Turkish labor policy has repeatedly sought to delay or control the development of collective action among workers and to isolate the labor movement from all other associations. To accomplish these ends Turkish labor policy has alternately employed techniques of repression, preemptive co-optation, fragmentation, and, most recently, corporatism. From the late nineteenth century until the 1960s labor policy consistently hindered collective action among wage earners by maintaining that disputes between employers and workers involved individual citizens each of whom had equal protection under the law.14 State paternalism, which was to become another continuing aspect of Turkish labor policy, characterized the first specific piece of labor legislation—an 1865 regulation that imposed minimal safety standards to protect miners in the foreign-owned Egreli coal fields. The expansion of state intervention in labor relations during the republican era involved an implicit recognition that the individual work 13

The continuing disorganization of the Turkish labor movement and the extraordinary reluctance of its most important leaders to describe social and political conflict in class terms are, in part, the result of Turkey's still moderate level of structural differentiation and the political culture. But the common emphasis on these two variables underestimates the more direct and immediate impact of public policy in shaping the development of the labor movement. 14 Modern Turkish labor unions have not evolved out of the network of esnaf guilds (loncas) of the Ottoman Empire. Although esnaf guilds survived until the Young Turk era, their traditional form of mutual responsibility between owners and workers was abolished in 1860 by the Mecelle, which introduced the notion of the individual work contract as the basic principle regulating labor relations.

Formal-Legal Bases — 119

contract was an insufficient means for protecting workers' interests, but only in the last twenty years have workers' organizations been granted any significant role in defending those interests.15 The most important evidence of union activity before 1946 was a brief series of strikes immediately after the 1908 revolution that were directed at foreign concessions controlling transportation and public utilities in Istanbul and the Aegean. These strikes were soon suppressed by the Young Turks and were quickly followed by the first clearly antilabor measure— the Work Stoppages Act of 1909. Although the law prohibited strikes and unions only in public services and utilities, it effectively destroyed the empire's nascent labor movement, since the bulk of the nonagricultural work force was still concentrated in the service sector. Most unions were replaced by mutual aid societies or officially approved workers' representatives operating within a state-controlled system of conciliation. During the War of Independence large strikes again occurred among transport workers supported by Turkey's earliest socialist parties. But by 1925 the Law for the Maintenance of Public Order prohibited all union organization, leaving only a small number of mutual aid societies.16 Soon after this initial repression of the labor movement, preparation began for drafting a comprehensive labor code to anticipate the dangers of class conflict that might arise from rapid industrialization.17 At the time, the regulation of industrial relations was not viewed as an urgent matter and the bill was prepared with great leisure before the emergence of a sizable industrial proletariat. Serious work on the statute was not begun until after 1932, when Turkey joined the International Labor Organization and sought the advice of several German and American experts. The most important of these I.L.O. advisers, such as Gerhard Kessler and Oscar Weigert, were German social democratic 6migres who opposed Hitler's regime, but who strongly endorsed state coordination of industrial relations to prevent strikes and maintain social harmony. They had a major influence not only in shaping the Labor Code of 1936 but also in training those who 15

Sumner Rosen, "Labor in Turkey's Economic Development" (Ph.D. dissertation, Massachusetts Institute of Technology, 1959), chap. 11; Oya Sencer, Tiirkiye'de I§ci Sinifi [The working class in Turkey] (Istanbul: Habora Kitabevi, n.d.). 16 Sencer, pp. 172-295; Orhan Tuna, "Trade Unions in Turkey," international Labor Review 40 (November 1964):413-431. 17 As early as 1924 the government submitted a draft proposal to the National Assembly in which it stated: "Turkey will be endowed with a national industry in the near future. To give this budding industry more vigor, and to obtain the maximum output from the working masses, it is important to place the various branches of industry under the control of law, and to prevent conflicts between capital and labor" (Sumner Rosen, "Turkey," in Labor in Developing Countries, ed. Walter Galenson [Berkeley: University of California Press, 1962], p. 268).

120 — Bases of Politics

would later head the new Ministry of Labor and on students in Istanbul University who pioneered the study of industrial relations in Turkey. The guiding principles of Turkish labor policy during the 1930s were most clearly summarized in an article prepared by Weigert to praise the new labor code before an international audience. He explained that Turkish industrial development was intentionally kept within certain bounds to prevent the rapid growth of an industrial work force. This was possible because new technology permitted Turkish industry to attain higher output with fewer workers than had been the case at similar stages of development in Europe. Weigert argued that the disorganized nature of the work force was a natural concomitant of Turkey's primitive industrial conditions and that this necessitated maintaining the principle of the individual work contract and expanding compensatory state protection.18 Weigert insisted that even though the absence of unions forced the worker to "face the employer as an individual . . . [the worker's] legal protection in Turkey exceeds labor legislation in countries that allow collective agreements."19 Besides the benevolence of the state, the principal means for guaranteeing workers' rights and securing industrial peace was not to be collective bargaining but the case-by-case enforcement of individual work agreements through compulsory conciliation and arbitration. Turkish etatism could not have found a more forceful exponent than the foreigner who applied its principles to the republic's industrial relations. Besides the original two parties, employers and [unorganized] workers, the State now appears as a third and immediately powerful factor in economic life, compelling private interests to justify themselves by the standards of common justice and the public good. At the same time many conditions and relations which originated in the mists of tradition or in the interplay of unconscious forces will be forced into the clear light of rational regulation.20 Two of the most important sections of the 1936 code prescribed minimal standards regarding working hours, health and safety, and female and child labor to be enforced by inspectors from a Directorate of Labor in the Ministry of Economic Affairs. Two other sections concerned industrial 18 Of course, Weigert ignored the clearly inhibiting impact of previous policies toward labor organization and failed to indicate that one purpose of the new legislation was to discourage unionization through somewhat less coercive means. In this regard he was certainly less candid than Prime Minister Inonii, who at almost the same time stated his hope that "the new Labor Law will sweep away the clouds which make possible the birth and life of class consciousness" (ibid., p. 270). 19 Oscar Weigert, "The New Turkish Labor Code," International Labor Review 35 (June 1937):761-764. 20 IbId., p. 772.

Formal-Legal Bases — 121

disputes, which were divided into individual and collective categories. There was no provision for collective action in disputes concerning work contracts. Such disagreements were to be settled in special labor courts. The only matter in which collective action was permitted was wage claims. If wage disputes involved at least one fifth of the workers in a given enterprise, they could be submitted by elected workers' representatives to provincial arbitration boards whose members were selected by the government. The code prohibited strikes and lockouts. Although it was silent on the issue of union organization, this was specifically outlawed by the Law of Associations adopted two years later.21 In practice, the comprehensive system of paternalism envisioned by the drafters of the Labor Code was restricted to employees in the largest state enterprises.22 Even this narrow application of Turkey's supposedly enlightened labor law was suspended during World War II. The National Defense Act of 1940 sought to provide a stable supply of labor to industrial and mining enterprises in order to maintain production rates and avoid work stoppages. It declared that "for this purpose paid work obligations may be imposed on citizens" and prohibited employees from leaving their work sites without an acceptable reason. A 1944 amendment to the same act likened the work obligation to military service and authorized the use of security forces for the forceable return of those who fled their places of work. Finally, the imposition of government wage controls suspended the sole issue on which collective action had been permitted.23 Immediately after the war the government attempted to redeem a few of the welfare pledges made in the 1936 Labor Code. A poorly financed Ministry of Labor was established together with the beginnings of the Social Insurance Organization and the State Employment Service. The most important change in labor policy—the decision to permit the revival of union organization—did not involve any major legislative initiative comparable to the careful preparation of the 1936 code. The rebirth of the labor movement was made possible simply by striking from the Law of Associations the ban against groups based on class. The number of or21 Between 1936 and 1946, when unions werefinallypermitted, even this limited machinery for collective dispute settlement was hardly used. Only ten cases reached the central arbitration board. In 1950 the labor courts were transformed into tripartite tribunals, but the same system was not extended to arbitration boards until 1958. After 1950 unions were granted limited authority to initiate disputes and the volume of cases increased rapidly. Nevertheless, labor unions still had no role in selecting the workers' representatives who participated in the deliberations (Rosen, "Turkey," pp. 269-271). 22 Coverage originally excluded the majority of workers because it applied only to firms that employed ten or more wage earners. Enforcement in the private sector was lax because of the minimal staffing of the Directorate of Labor. 23 Ibid., pp. 271-272.

122 — Bases of Politics

ganizations that arose in the six months immediately after adoption of the 1946 Law of Associations clearly surprised the government. The People's Party had expected that the new Labor Ministry and its own Workers' Bureau would allow it to direct the development of organized labor at a gradual pace. Alarmed by the rapid success of opposition leaders, especially of new left-wing parties, in establishing alliances with labor, the government began as early as December 1946 to close down several unions and arrest both labor and party leaders. In order to reassert its control over union formation and activity, the government hastily adopted a separate Trade Unions Act. The 1947 Law on Workers' and Employers' Unions applied the fragmenting and isolating strategy of "debilitating pluralism" to the labor movement more directly than the Law of Associations did to voluntary organizations in general. Rival unionism was encouraged by permitting an unlimited number of organizations in the same work branch and allowing multiple membership without providing any authoritative mechanism for settling disputes over representation. The principle of voluntary membership was strictly applied to prevent the formation of union shops. Federation or affiliation with any type of multiunion organization required a formal endorsement from two thirds of the membership. Several provisions also attempted to isolate the leadership of the labor movement from undesired political contamination. Nonworkers, white-collar employees, and civil servants were prohibited from joining labor unions. Unions could not "engage in politics, political propaganda or publication" and could not "become the instrument of a political organization." Affiliation with international organizations was permitted with approval from the Council of Ministers, but this was not granted for affiliation with the International Confederation of Free Trade Unions until I960.24 Labor organizations were clearly expected to subordinate their own economic and institutional interests to national goals. The statute declared that "unions are national organizations and cannot engage in activities that are unpatriotic or contrary to national interests." Strikes were, of course, illegal and participating unions could be closed down by the courts for up to one year. The role of unions in settling industrial disputes was limited to providing advisory opinions to courts and arbitration panels and petitioning against employer attempts to reduce wages. Unions were given the right to negotiate collective agreements for their members, but the prohibition against strikes provided employers with no incentive to engage in face to face bargaining. Thus, the clear intent of the law was to promote 24 Toker Dereli, The Development of Turkish Trade Unionism (Istanbul: Istanbul University, Faculty of Economics, 1968), pp. 61-79.

Formal-Legal Bases -— 123 a multitude of competing, small local unions, deprived of any substantial economic functions, prohibited from forming open political alliances, and dependent upon the good will of the government in promoting the welfare of their members.25 Throughout the 1950s the Democrats contained and controlled the labor movement, first, by expanding state paternalism and, then, by arresting union leaders who moved too close to the opposition. Democratic governments continually refused to honor the party's original pledge to legalize strikes. Instead, they sought worker support by ignoring the union leadership and directly distributing material benefits to the rank and file. The Ministry of Labor promoted expensive programs to build low-cost housing, establish minimum wage levels, and grant pay bonuses and tax exemptions. It also provided a wide variety of fringe benefits including health care, clothing, meals, transportation, recreational facilities, and child allotments. Rosen estimated that such fringe benefits amounted to about 25 percent of the salaries of workers in the State Economic Enterprises.26 In the meetings of the Social Insurance Organization, the major instrument for managing welfare programs, the government often exceeded the demands of union representatives for more generous benefits while consistently rejecting their requests for a greater role in decision making and program administration. The Democrats' approach toward unionists soon changed from exclusion and avoidance to direct harassment and manipulation. The national labor confederation, Turk-I§, was never able to establish itself as the recognized spokesman for the labor movement. Its leaders were told by several ministers of labor that the government had no need for'' intermediate channels'' in addressing workers. As the confederation's frustrated leaders increasingly relied on the support of the People's Party, the government began to interpret virtually all union criticism of its policies as illegal political activity. The confederation came into direct confrontation with the Democrats, dividing the already fragmented union leadership into partisan factions. The government itself encouraged several unions to help organize the mass demonstrations that led to the anti-Greek riots of 1955, and then took advantage of the subsequent period of martial law to arrest its most outspoken labor critics. Several regional federations (birliks) were closed, new union formation was tied to prior permission from the Labor Ministry, and strict financial controls were imposed on existing unions. Finally, in the confederation's 1957 presidential election Democratic loyalists replaced Ismail inan, leader of the militant People's Party faction, with the more pliable Nuri Beser. The change in leadership allowed Turk-Is to regain its 25 26

Ibid., pp. 80-107. Rosen, "Turkey," p. 276.

124 — Bases of Politics

subsidy from the government and to preserve its symbolic position as the chief peak association of the labor movement. Yet, the confederation remained as powerless as ever.27 Labor policy shifted again after adoption of the liberal 1961 constitution with its emphasis on both social justice and planned economic development. After more than a decade of government and employer resistance, the rights to strike and engage in collective bargaining were finally endorsed in order to redress the imbalance of economic power that had resulted from Democratic discrimination in favor of private business. The constitution's framers also felt that their renewed concern for accelerated industrialization required a more centralized system of industrial relations that would enlist the support of reliable labor leaders. Tripartite commissions, including representatives from government, employers' associations, and labor unions, drafted two key pieces of legislation—a new Trade Unions Act and a Law on Strikes, Lockouts, and Collective Agreements. These statutes were adopted in 1963 and currently regulate Turkey's system of industrial relations and union organization. This legislation was characterized by a considerable ambivalence toward the notion of pluralism as it had come to be understood in the labor movement since 1946. On one hand, it was commonly recognized that reliance on the voluntary and spontaneous formation of competing organizations had seriously handicapped the labor movement. Unions were still too weak to protect worker interests effectively and assist in the urgently felt need to further coordinate economic life. On the other hand, there was great reluctance to allow any group of unionists to concentrate too much power in their hands. A continued plurality of union representatives was seen as necessary to provide the rank and file with a choice of leaders and to protect ambitious industrialization policies from the potentially disruptive impact of a "labor aristocracy." The government and employers were eager both to prevent radical socialists from gaining a foothold in the labor movement and to lay the basis for centralized, industrywide collective bargaining. Thus, moderate union representatives in the tripartite commissions were able to gain several concessions that greatly strengthened workers' organizations. For the first time, unions were granted responsibility for selecting workers' representatives and shop stewards. Union shops were still prohibited, but greater financial security was provided by adopting a checkoff system for dues collection. The common practice of free riding was discouraged by requiring nonunion members to pay "solidarity contributions" amounting to two thirds of regular dues in order to enjoy the benefits of collective 27

Rosen, "Labor in Turkey's Economic Development," chap. 25.

Formal-Legal Bases — 125 agreements negotiated by the union. White-collar workers were finally permitted to form unions and the minimum age for membership was reduced to sixteen. Special protection was provided for the jobs of union officers in order to discourage employers from union busting. Lockouts were permitted, but strikebreaking and scabbing were not. In order to promote industrywide unionization, geographic federations were eliminated and conditions for affiliation were reduced. Unions were still prohibited from receiving financial aid from political parties, but they no longer needed government approval for affiliation with international organizations. Unions were authorized to allot 5 percent of their income to worker education programs and to make financial investments provided that they did not distribute profits.28 All of these provisions for strengthening union organization and encouraging industrywide bargaining agents closely paralleled Tiirk-is's aim of establishing itself as the coordinating representative of a more centralized labor movement. After a decade of isolation and harassment the confederation leadership hoped that, at last, it might achieve with statutory assistance what it could not accomplish through its own limited resources— acceptance by the government and the unions of its claim to act as the authoritative spokesman for Turkish labor. Nevertheless, several key aspects of the new Trade Unions Act were bound to frustrate Tiirk-is's goal of "fewer unions and more members" because they preserved rival unionism and maintained the government's ability to fragment labor organization. First, the law still permitted an unlimited number of unions in each work branch and allowed multiple memberships. Second, the delimitation of work branches that structure union organization remained entirely in the hands of the Ministry of Labor. The government insisted on retaining a highly differentiated list of about thirty-five industrial categories, which inhibited effective organization in several important sectors such as transportation, public services, and utilities. Turk-Is endorsed a German model of centralized unionization with as few as fifteen national federations, but it was not even successful in persuading the Labor Ministry to accept a modest alternative scheme based on twenty-eight categories. Third, although the statute was designed to promote exclusive bargaining agents, it provided no clear procedure for determining how a union qualified as such. The encouragement of multiple organizations in highly fragmented work branches made it likely that there would be a large volume of authorization disputes, but no mechanism was established for their settlement other than lengthy court fights.29 28

Dereli, pp. 110-135. »Ibid.,pp. 138-379.

126 — Bases of Politics

Within this problematic legal framework, Turk-is/s efforts to establish a single national union in each industrial sector quickly alienated unionists who refused to accept the confederation's discipline, and led to several jurisdictional disputes with militant independent and left-wing organizations. A major confrontation arose as early as 1966, when Turk-Is, condemned an unauthorized strike in the Pa^abahce glassworks. When TurkIs, began proceedings to expel the glassworkers' union, several federations came to the strikers' aid and formed the nucleus of DISK, a rival socialist confederation that soon became allied with the Turkish Workers' Party. DISK openly advocated political struggle as indispensable for advancing the interests of the working class. It denounced Turk-is/s "above party, bread and butter" unionism as a mask for collaboration with bourgeois politicians in attempts to suppress class struggle. DISK gradually established itself as the principal organizational alternative to Turk-is, and was particularly successful in radicalizing more skilled workers in several manufacturing sectors around the Istanbul-Izmit area and in Adana.30 As the rivalry between Turk-Is, and DISK intensified the older confederation's relations with the Justice Party government improved substantially. Although Turk-Is/s attempt to portray itself as labor's sole authoritative representative was becoming less credible, the government became more willing to grant official recognition to the confederation's claims. Turk-Is, leaders were commonly accepted as labor's exclusive representatives in mixed consultative boards and regulatory commissions in the State Planning Organization, the Minimum Wage Commission, the Social Insurance Organization, the Supreme Arbitration Board, the State Economic Enterprises, the National Productivity Council, the Turkish Standards Institute, and in ad hoc consultative committees preparing labor legislation. A clear sign that the Justice Party government desired to move away from the ambivalent pluralism of the 1963 legislation and opt for more unitary labor organization under Turk-Is, control was its sponsorship in 1970 of two ill-conceived and obviously discriminatory amendments to the Trade Unions Act and to the Law on Strikes, Lockouts, and Collective Agreements. The two most important amendments aimed at granting Turkis, an effective representational monopoly at the national level and undermining its rivals even in those areas where they already had organized the greatest number of workers. The new Trade Unions Act provided that "for a trade union to be recognized as a national union [or a federation] it must 30 Alpaslan Isikli, Sendikacilik ve Siyaset [Unionism and politics] (Ankara: Siyasal Bilgiler Fakiiltesi, 1972); AniKJicen, Turkiye'de Sendikacilik [Unionism in Turkey] (Ankara: Ozgur Insan Yayinlan, 1973); Bruce Millen, ' 'Factions of the Turkish Labor Movement Differ Over Political Role," Monthly Labor Review (June 1969):31-35.

Formal-Legal Bases — 127

represent at least one-third of the insured workers in its trade. . . . Confederations must comprise at least one-third of the trade unions and federations . . . and have a membership of at least one-third of the unionized workers of Turkey." 31 This meant, of course, that Turk-Is would be the only labor organization authorized to use the title "confederation" and that its informally acquired privileges of exclusive national and international representation would become official and permanent. The proposed amendment to the Law on Strikes, Lockouts, and Collective Agreements went even farther in attempting to dismember DISK'S socialist unionism by statute. It provided that "[only] a federation or union which is active throughout Turkey . . . and which represents at least thirtyfive per cent of the insured workers in a given sector of economic activity shall have power to conclude a collective agreement." This meant that a federation recognized as an industrywide bargaining agent would negotiate for all workers in a given industry, even though a particular plant might have no affiliates of the bargaining union and a union not affiliated with the bargaining agent might already have organized the majority of workers in the plant. Massive workers' riots in Istanbul and Izmit and the judgment by I.C.F.T.U. investigators that the I.L.O. convention on freedom of association had been violated prompted the government to temporarily withdraw the measure on collective bargaining. The new Trade Unions Act was overturned by the Constitutional Court two years later, but Turkic soon received even stronger right-wing support for efforts to eliminate its rivals with discriminatory legislation.32 Thus, since 1960 policy toward labor unions has become increasingly distinct from policy toward voluntary associations in general. Previously, special treatment for labor involved the use of repression, paternalism, and fragmentation in a particular attempt to disorganize and isolate the labor movement while most other voluntary associations were permitted more room to maneuver and greater relative autonomy. Since 1960, however, rigid restrictions on voluntary associations have been accompanied by the first serious efforts to strengthen and centralize labor organization. In each 31 Some of the less controversial changes aimed at bolstering the more established union leadership against both newer rivals and the rank and file. The terms in office of union officials were lengthened, and officials were given more autonomy from general assembly meetings. Officers in executive positions were required to have worked for at least three years in the same work branch. All members who wished to resign from one union and join another were required to file notarized affidavits. These measures did nothing to resolve the growing number of jurisdictional disputes between collective-bargaining agents, but they were generally well received as long overdue steps to encourage strong union leadership ("Changes in Trade Union Legislation Stir Controversy," Labor Developments Abroad [March 1971]:30-33). 32 Ibid., p. 32; MUliyet, 17-22 June 1970.

128 — Bases of Politics

case policy makers appear to be abandoning the framework of debilitating pluralism, but they are attempting to replace it with different alternatives. For voluntary associations in general this has involved establishing more direct government surveillance and creating greater opportunities for the use of selective repression. For labor unions, on the other hand, it has involved establishing clientelistic relations with favored unionists and coopting them through grants of selective privilege that traditionally have been reserved for semiofficial, public associations. In other words, while debilitating pluralism is being transformed into a revival of state control over most voluntary associations, it is being abandoned in favor of corporatizing initiatives toward the labor movement.

THE FORMAL-LEGAL BIFURCATION OF ASSOCIATIONAL LIFE

The Turkish associational universe has long been divided between two major legal types of organization—private associations (dernekler) and public associations (kamu kuruluqlari). These have coexisted and expanded as parallel associational networks to produce a system of interest representation that is formally bifurcated into what may be described as pluralist and corporatist channels. Understanding corporatism and pluralism as alternative models for structuring systems of interest representation, we can use Schmitter's definition of these concepts to compare public and private associations in Turkey. Schmitter states that Corporatism can be defined as a system of interest representation in which the constituent units are organized into a limited number of singular, compulsory, non-competitive, hierarchically ordered and functionally differentiated categories, recognized or licensed (if not created) by the state and granted a deliberate representational monopoly within their respective categories in exchange for observing certain controls on their selection of leaders and articulation of demands and supports.33 Alternatively, Pluralism can be defined as a system of interest representation in which the constituent units are organized into an unspecified number of multiple, voluntary, competitive, non-hierarchically ordered and self-determined (as to type or scope of interest) categories which are not specifically licensed, recognized, subsidized, created or otherwise controlled in leadership selection or interest articulation by the state and which do 33

Schmitter, "Still the Century?" pp. 93-94.

Formal-Legal Bases — 129 not exercise a monopoly of representational activity within their respective categories.34 In periods when the freedom of association has not been openly repressed, private groups have represented the closest approximation in Turkish associational life to the pluralist paradigm. Yet we have seen that even in the most liberal times since 1946, private associations have departed from the pluralist model in several important respects. They have never been free to determine the type or scope of interests they represent. They have continually been required to limit their representational scope to a single, clearly defined interest. Although not officially licensed by the state, they have had to comply with often difficult procedures for registration and certification. While not directly created by the state, often they have been established with the aid of ruling party organizations and commonly are subsidized from the public treasury. Leadership selection and interest articulation seldom have been directly controlled by the state, but the government always has retained broad powers of intervention either through direct executive action or through a more or less manipulable judiciary. Public associations, on the other hand, adhere more closely to the corporatist paradigm. Each public association is created by special legislation. In each case the statute describes in detail the association's relationship with parent ministries, the scope and characteristics of its membership, its generally hierarchical organizational structure, its range of permitted activities (the most important of which are "to serve the public interest" and "to secure professional discipline"), and provisions for periodic government inspection of its records and finances. Association officers are not formally recognized as civil servants, but their salaries are paid from public funds and they are punishable for abuse of office with the same severe penalties that apply to government officials who violate the public trust. In exchange for accepting this lesser degree of formal autonomy, public associations enjoy, by virtue of their semiofficial status, a considerable number of privileges that are denied to the dernekler and are highly attractive to the associational leadership. The most important of these can be summarized as follows: (1) recognition as the most authoritative and legitimate spokesman for the interests of a given sector, conferring a formal monopoly of representation versus rival associations; (2) compulsory membership and dues from all who work in the economic or professional area represented by the association, thereby providing it with a guaranteed minimal income; 34

Ibid., p. 96.

130 — Bases of Politics

(3) deputation to perform certain public services that would otherwise be the responsibility of the government itself, in return for which the association is entitled to collect revenues from its membership and the general public; (4) authorization to distribute among its membership credit and investment subsidies as well as coveted import licenses and quotas allocated by the Ministries of Trade and Industry and the State Planning Office; (5) ex officio representation in public regulatory and planning committees, providing guaranteed access and consultation; (6) increasing regulatory power over its membership by virtue of the delegated authority to make "binding professional decisions" regarding professional customs and ethics, licensing, dispute settlement between members or between members and the public, and even appropriate forms of dress in the case of the legal profession; (7) the right to administer newly established compulsory social'insurance funds and to invest the accumulated capital in such projects as consumer cooperatives, specially licensed import firms, credit cooperatives, and new industrial enterprises approved by the State Planning Organization. In the combination of control and privilege that characterizes Turkish public associations there is a clear example of what Schmitter calls "the hallmark of corporatism: the exchange of a legal monopoly on representation and a guaranteed access to decision-makers in return for compliance with certain limitations on behavior."35 There are, however, some important features of this exchange that have made the Turkish brand of corporatism a rather "impure" and unstable type. While the increasing scope and value of privileges accorded to public associations have made this variety of representation quite attractive to cooperative association leaders, the traditional precariousness of those privileges has not changed substantially. In fact, the recent extension of more generous privileges has been accompanied by greater government determination to control leadership selection and interest articulation. Consequently, the exchange relationships that have emerged between governments and public associations often seem to be, not mutually beneficial clientelisms, but thinly disguised attempts of the state to cultivate and reward cooperative association leaders without allowing them to acquire independent authority. On the whole, the network of public professional associations operates, not as a centralized system for harmonizing and balancing similarly organized functional units, but as a series of channels for selectively distributing unequal privileges and rewards within and among differentially organized, competing social and economic sectors. Schmitter, Interest Conflict, p. 111.

Formal-Legal Bases — 131

Governments have been able to manipulate public associations through a number of techniques. They have controlled the timing of association elections through ministry-ordered postponements. They have infiltrated association conventions with party liaisons who circulate electoral lists including favored and acceptable candidates. They have precipitated financial crises in associations by temporarily suspending their key revenueearning functions and reassigning them to government agencies. Since public associations are constituted as branches of one or more parent ministries, much of their internal organizational structure is determined by executive regulations or ministry-prepared statutes. This has been particularly important in allowing the government to manage conflicts between different segments within each occupational sector. At times ministryinitiated reorganization has aimed at smothering internal interest conflict by further centralizing associational structures despite the increasing functional differentiation of the membership (as with businessmen and the esnaf). At other times it has aimed at frustrating the expression of common associational interests by imposing extreme organizational fragmentation on a membership whose common goals supersede their narrowly defined functional differences (as with architects and engineers). In either case the attempt has been to bring the policies of public associations into line with those of the ruling party by structuring the organizations so that favored interests are overrepresented and other interests are isolated. Finally, in exceptional cases when none of these techniques is sufficient, a ruling party can use its control over parliament to strip an association of its semiofficial status and reduce it to a common voluntary organization. Another destabilizing aspect of Turkish corporatism is the imperfect nature of the representational monopolies enjoyed by most public associations. The Turkish associational universe has always been a mixed system of representation combining corporatist, public structures with pluralist, private ones. It is possible to distinguish certain historical periods and certain economic and social sectors in which one type of representative structure has had a clear predominance relative to the other.36 On the whole, however, public policy has been quite eclectic and changeable. Over time, it has promoted both types of associational structures and the total system has remained heterogeneous. Thus, it is not uncommon for an individual M The vast majority of Turkish associations have always been of the private variety. These are mainly small, short-lived community groups with little political or economic influence. In the major economic sectors, such as business, the professions, and agriculture, however, representation has been structured primarily through public associations. Labor unions, as we have seen, have been treated as a special category similar to private associations, and esnaf organizations have gone through several reincarnations as both public and private entities.

132 — Bases of Politics

sector to contain both public and private associations as parallel, though obviously not equally effective, channels of representation. As a result of this representational dualism, there continues to be considerable room to maneuver for groups that become dissatisfied with the kinds of clientelistic relationships established between the government and the leaders of public associations. If those relationships are perceived as yielding insufficient benefits for the association membership or as exclusively serving the interests of a favored, dominant faction, disgruntled leaders can resort to the alternative channel of voluntary associations to promote their specific interests in a more independent manner. In such cases private associations can be transformed from weak and insignificant extensions of their more privileged and influential corporatist counterparts into more specialized, competing channels of representation that may form political alliances with opposition parties. Private groups that succeed in establishing such alliances cannot, of course, compensate for the absence of official privilege and may, in fact, prompt a hostile response from the government. But even if such attempts prove self-defeating in the short run, they can have the effect of seriously undermining the representational monopolies of dominant, public associations and jeopardizing their position as the most authoritative spokesmen for an entire occupational or economic sector. The origins of modern corporatist representation through public associations go back to the attempts of nineteenth-century Ottoman reformers to combine administrative centralization with the rehabilitation of declining Ottoman industry.37 In 1876 a new Council on Trade and Agriculture made several recommendations for economic reform, including the establishment of "consultative associations" of industrial and commercial leaders. The first of these, the Istanbul Chamber of Industry, was founded in 1879. The chamber's structure was a modified version of the traditional guilds. It 37 This originally amounted to a brief and unsuccessful attempt to reverse the free trade policies of the early Tanzimat, which had abolished the traditional system of commercial licensing and monopolies for the esnaf and granted generous concessions to European merchants and their agents. As early as 1864 the Ottoman state established a new Commission for Industrial Reform composed of leading craftsmen and merchants whose primary objective was to organize tradesmen in new associations and producers' cooperatives. The commission's charter asserted, "There is no basis for and no advantage in the individual and separate endeavor of the tradesmen as of old, and their success is dependent upon their common and unified efforts in arranging for the organization of their interests and skills in collective enterprises having as much capital as is necessary" (Nizamettin AIi Sav, Sanayi Iktisadi ve Tiirk Sanayi [Industrial economy and Turkish industry] [Izmit: Seliiloz Basimevi, 1950], p. 129). By 1874 only seven such associations had been founded and the government abolished the commission, attributing its failure to tradesmen's "lack of understanding of the advantages of collective commercial practices as opposed to individual endeavor" (ibid., pp. 132-134).

Formal-Legal Bases — 133

was clearly intended to serve as an administrative link that would extend td larger merchants a less rigid type of state regulation than had prevailed among artisans and tradesmen. Several features of its organization and operation define it as the direct forerunner of modern corporatist occupational associations.38 Even in this first businessmen's association, interest representation soon was complicated by the imposition of a unitary and hierarchical organizational structure on a heterogeneous and divided constituency. The Chamber of Industry was originally founded to organize Ottoman manufacturing and support it in competing with foreign capital. Responding to Abdtilhamit's request for advice on economic policy, the chamber quickly prepared several recommendations for developing new markets and increasing exports, including opposition to the proposed linking of European and Anatolian railway lines. But as the sultan became increasingly dependent on European credit, economic policy abandoned its encouragement of native manufactures. Control of the chamber soon shifted into the hands of importing interests who openly supported the Administration of the Public Debt. This was accomplished by expanding the association's scope of representation to incorporate interests that conflicted with those of its original constituency and by direct government intervention to tip the balance in favor of those who supported the return to "free trade.'' The chamber was soon renamed the Chamber of Commerce and Industry, and eventually of Agriculture as well. Chamber-sponsored commercial exhibits that earlier displayed Ottoman products replaced them with foreign-produced goods. Finally, in 1896 a new minister of trade complained that the association's organization and method of leadership selection were among "the causes preventing various Chambers of Commerce from rendering all the useful services that the nation has a right to expect from these institutions." The organization that had been "uniquely created to defend the interests of Ottoman trade" became increasingly dominated by the agents of European 38

It was established as a semiofficial association under the Ministry of Trade, which controlled its election of officers and organization of branches. Membership was, in effect, compulsory and an attempt was made to register all Ottoman merchants, including those who had previously acquired privileged status as beraths. The chamber performed numerous functions of regulation and surveillance, acting as an auxiliary enforcer of the decisions of the new Commercial Courts. As a deputized agent of the Ministry of Trade it provided the government with information on market conditions and prices, established norms for professional conduct, and transmitted ministry regulations through its own publications. As the officially recognized representative of private business it also advised the government of merchant grievances and prepared recommendations for economic reform (Margaret Hoell, "The Chamber of Commerce of Istanbul," paper presented at the 9th annual meeting of the Middle East Studies Association, 20-22 November 1975, Louisville, Kentucky).

134 — Bases of Politics

manufacturing and finance, who contributed to doubling the empire's trade deficit in the decade before World War I.39 While modern public associations preserve many features of the traditional guilds and the early Chambers of Commerce and Industry, they are a more direct legacy of Kemalist attempts to establish a network of corporatist occupational groups as supporting structures oietatism. Originating in the business and agricultural sectors, corporatist representation has expanded in scope to include several of the free professions as well. Under Ataturk public associations were created to organize and regulate industrialists and merchants (1925), the esnaf'(1925), exporters (1936), lawyers (1938), and journalists (1938-46). The Democrats later included such categories as doctors (1953), architects and engineers (1954), veterinarians (1954), and pharmacists (1956). Of course, workers and civil servants have been consistently excluded from the network of corporatist privilege, and when they have been allowed to organize at all, it has been only through voluntary associations. Historically the application of corporatist representation has been most consistent among large-scale merchants and industrialists and most irregular among the esnaf. The generally eclectic and shifting nature of public policy toward associations and the recurrent tendency to mix and remix both corporatist and pluralist structures in changing proportions are most clearly evident in the long series of arrangements that have been devised to structure the interests of large and small businessmen. Under the Young Turks these interests were organized separately in voluntary associations.40 In 1925 both sets of associations were entirely reorganized. The Chambers of Commerce and Industry regained their status as public associations and have retained it ever since. Although the esnaf associations technically remained private organizations, they were reshaped as compulsory and hierarchical structures that resembled public associations in all respects except that they were denied special status and privileges. They were, in effect, incorporated into the chambers as a special section. Many of the regulatory powers over esnaf that had formerly been given to municipalities were transferred to the chambers and the Ministry of Trade.41 39 Margaret Hoell, "The Ticaret Odasr. Organization, Functions, and Activities of the Chamber of Commerce of Istanbul, 1886-99" (Ph.D. dissertation, Ohio State University, 1973), pp. 202-209. 40 The Chambers of Commerce and Industry established in 1910 lacked formal semiofficial status, but they were delegated a number of regulatory functions and were defined as "on the one hand, the representative before the government of occupational members and, on the other hand, a consultative organ of the government in professional matters." In contrast, the esnaf associations created in 1909 were restricted to mutual assistance and internal dispute settlement, and were closely controlled by municipal governments (Sav, pp. 136-137). 41 Nabi Dincer, Esnaf ve Sanatkdrlarm Sosyal ve Ekonomik Sorunlan Arastirmasi [Study

Formal-Legal Bases — 135

The ambiguous legal status of esnaf associations continued even after 1943 when the public esnaf groups were created within the centralized structure of the Chambers of Trade, Industry, and Commercial Exchanges. This change clearly subordinated the interests of the esnaf to those of owners of larger enterprises. Members of the new esnaf chambers were also required to join and pay dues to the Chambers of Trade and Industry even though they had no effective voice in those associations. Since the older network of "private" (but still compulsory) esnaf associations had not been abolished, many esnaf were actually committed to the simultaneous financial support of three different associations. Even within their own semiofficial chambers esnaf leaders had little effective influence; the association's general secretary was an appointed ministry representative, authorized to block the enforcement of chamber decisions.42 An unusual renewal of concern with esnaf organizations arose shortly after adoption of the 1947 Trade Unions Act and the revival of the labor movement. Policy makers began to consider the possibility of rescuing the esnaf from their precarious status as the stepchild of the private sector and promoting them as a broad and stable middle class of small owners that might act as a buffer against the antagonisms of organized labor and big business.43 A new attempt was made to define the heterogeneous and widely dispersed esnaf as a distinct social category. In fact, the new legal definition of the esnaf did little to clarify their ambiguous status, since it employed multiple and conflicting criteria, including the size of enterprise, the branch of industry in which the enterprise operated, and the degree to which production methods could be considered "assembly line operations." The major intent of policy makers was clearly to promote the notion of a broad residual category, intermediate between and outnumbering antagonistic workers and employers, rather than to provide a more effective representational structure for a highly diversified constituency. Thus, the most commonly used definition of the esnaf was also the vaguest, embracing all tradesmen and merchants who "combined a small amount of capital with a large amount of labor." The main postwar innovation in structuring esnaf representation was the creation, in 1949, of a set of esnaf associations and regional federations of the social and economic problems of esnaf and craftsmen] (Ankara: Devlet Planlama Teskilati, 1971). 42 Ibid.; Sav, pp. 145-148. 43 Speaking in support of yet another proposal to restructure esnaf associations, the minister of labor declared: "we are creating a new type of society. With the Trade Unions Law we have created an organization with the characteristics of a large social class. . . . Along with the proletariat there comes into existence a balancing power. This is why it is a good undertaking to preserve the esnaf as a whole and derive social benefits from them" (Karpat, p. 316).

136 — Bases of Politics independent of the Chambers of Commerce and Industry. These organizations were granted more autonomy from the Ministry of Trade than either their predecessors or the chambers, and regulatory power over the esnaf was largely restored to local municipalities. But the increased autonomy of esnaf associations was accompanied by the loss of their status as public organizations and their reconversion into voluntary associations. The legislation that created these groups broadly defined their goals as improving production, achieving professional cooperation and solidarity, and advancing the interests of small businessmen. Yet it failed to provide them with even their traditionally limited official functions or financial base. Instead, the new statute provided for unlimited and competing association formation, encouraged affiliation on local rather than national and functional lines, and granted a long overlevied membership the welcome opportunity to remain unorganized and pay no dues at all. In contrast, during the very next year (1950) the set of merchant and industrialist associations was capped with a new central Union of Chambers, which further consolidated corporatism in the private business sector. At the same time that esnaf associations were deprived of semiofficial status and transformed into fragmented pluralist groups, big business acquired the first nationwide peak association to be formally recognized as the sole legitimate spokesman for an entire economic sector. Even the creation of two distinct organizational structures possessing such obviously unequal privileges and resources failed to disentangle the representation of esnaf interests from those of larger enterprises. The esnaf were no longer compelled to enroll in businessmen's chambers, but they faced the new problem of trying to exclude intruding big businessmen. It became a fairly common practice for large- and medium-sized enterprises that were already enrolled in the Chambers of Commerce and Industry to take advantage of the vague legal distinction between the esnaf and businessmen to enroll in esnaf voluntary associations so that they could qualify for credits, subsidies, and special import quotas that had been earmarked for small traders and artisans. The continuing weakness and fragmentation of esnaf organization became a serious concern of economic planners in the early 1960s, when attention began to focus on industrialization policy. During the same period, labor leaders had benefited from government and employer desires to promote industrywide collective bargaining in their own efforts to secure new legislation strengthening union organization. Similarly, esnaf leaders also sought statutory assistance to resolve their chronic organizational problems and acquire greater economic concessions for the faltering enterprises of their membership. Although the esnaf were more successful than unionists in overcoming the fragmentation of debilitating pluralism,

Formal-Legal Bases — 137

they were required to sacrifice the formal autonomy of their associations in the bargain. The Esnaf and Small Artisans Act of 1964—the fifth major overhaul of esnaf associations in half a century—effectively removed esnaf organizations from the pluralist arena and once again placed them within the network of corporatist occupational associations. The statute that created the current esnaf associations continued to refer to them as dernekler, but its unusually detailed provisions specified nearly every aspect of associational structure and operation, producing perhaps the most hierarchically organized and tightly regulated of all Turkish public associations.44 While the new Esnaf Confederation roughly resembles the peak associations of other public organizations, it has far greater direct administrative and financial control over its affiliates and was clearly established to transmit directives and resources from the key economic ministries. The cost of organizational centralization has been not only greater government control but also a reduction in the opportunities for widely divergent interests to be represented through differentiated channels. The esnaf law has promoted heterogeneous local organization in order to discourage challenges to confederation control from specialized functional federations. Affiliation is narrowly confined by law to the boundaries of municipal governments except in unusual cases where a specific occupational leadership has organized an exceptionally high 60 percent of its category. Thus, the most diverse and idiosyncratic collection of occupational groups has been circumscribed within the most hierarchical and undifferentiated set of representative associations. This most recent organizational reshuffle has done little to reduce the longstanding alienation of the esnaf from their own associations or to manage more effectively the sharp conflicts between different strata and functional groups. Compulsory membership has provided esnaf associations with a stable financial base, but generally employees and independent operators join the owner-dominated organizations simply because it is the only way to obtain required licenses and work certificates. The meetings of these associations are commonly marked either by very low attendance or by violent confrontations between different occupational groups. Yet the only innovation of the new esnaf law for managing internal conflict is to require that all such gatherings include a government-appointed commissioner-observer, who is authorized to order police intervention and suspend the meetings. In fact, given the failure of several governments to fulfill their common pledges to reverse the economic deterioration of small businesses, the 44 Nejat Izmirlioglu and Muzaffer Uzunyayla, Izahh Esnaf ve Kiiqiik Sanatkarlar Kanunu [The Esnaf and Craftsmen Law] (Ankara, 1964).

138 — Bases of Politics recent reorganization of esnaf associations has resulted in the alienation of much of the es/ia/leadership as well. In contrast to the labor movement, where favored confederation leaders are being granted privileges generally reserved to public associations without being required to relinquish the formal autonomy of their private organizations, esnaf leaders have been forced into one of the most tightly regulated public associations without acquiring the expected quid pro quo of a more favorable economic policy. Esnaf leaders themselves have come to regard these new associations as mechanisms of economic and political control through which the state and big business are imposing a policy of industrial and commercial concentration on small enterprises, either by converting them into specialized suppliers for larger units or by driving them out of the marketplace. Representative structures that the esnaf had expected to strengthen occupational solidarity and to channel badly needed material assistance are increasingly perceived as agents for their manipulation and potential destruction. STRUCTURAL CONDUCIVENESS AND FLUCTUATIONS BETWEEN PLURALIST AND CORPORATIST POLICIES

The numerous fluctuations in public policy toward labor and esnaf associations are not unique or isolated examples of eclecticism in the regulation of Turkish associational life. Historically, policy in these sectors has been somewhat more changeable than in other sectors, because successive regimes have displayed a common sensitivity to the political and economic importance of workers and small businessmen while employing different techniques to organize or disorganize the representation of their interests. But frequent policy shifts in these two sectors are simply unusually clear examples of broader periodic shifts in which ruling elites have alternated between pluralist and corporatist strategies for structuring associational life to suit their changing strategies for political control and economic development. Schmitter has pointed to the linkages between systems of interest representation and the changing constraints and opportunities that political actors find in the economic system. He suggests that the corporatization of interest representation is related to certain structural imperatives of capitalist development. He also argues that historically the emergence of different varieties of corporatism are related to "differences in the specific nature of these imperatives or needs at different stages in the institutional development and international context of capitalism, especially as they affect the pattern of conflicting class interests." Accordingly, Schmitter elaborates two separate theories to explain the emergence of modern corporatism. He distinguishes between two corporatist subtypes—"societal

Formal-Legal Bases — 139

corporatism" and "state corporatism"—which when viewed statically and institutionally exhibit a basic structural similarity, but when viewed in motion appear as products of very different political, social, and economic processes and serve as vehicles for different power relations. In a nutshell, the origins of societal corporatism lie in the slow, almost imperceptible decay of advanced pluralism; the origins of state corporatism lie in the rapid, highly visible demise of nascent pluralism. . . . Societal corporatism appears to be the concomitant, if not ineluctable component, of the post-liberal, advanced capitalist, organized democratic welfare state; state corporatism seems to be a defining element of, if not structural necessity for, the anti-liberal, delayed capitalist, authoritarian, neo-mercantilist state. . . . The core of my speculation about structural conduciveness rests on the problems generated by delayed, dependent capitalist development and non-hegemonic class relations in the case of state corporatism, and advanced, monopoly or concentrated capitalist development and collaborative class relations in the case of societal corporatism.45 In effect, what Schmitter contributes here is the outline of a developmental model for analyzing periodic fluctuations between pluralist and corporatist associational policies with reference to underlying economic change. A more explicit statement of the model would involve four distinct, though not necessarily sequential, phases: incipient pluralism, state corporatism, advanced pluralism, and societal corporatism. Although our Turkish case departs from the last two phases of the model in some important respects, the long-term data display a strong enough parallel to make the scheme useful in organizing a historical overview of the seemingly bewildering associational policies of modern Turkey. In general, it is possible to summarize the history of policy changes toward Turkish associations by identifying four distinct policy periods in which different political regimes and strategies of economic development have been linked to different systems of pluralist or corporatist interest representation.

Incipient Pluralism (1908-24) This period includes the Young Turk regime, the War of Independence with its era of dual government, and the earliest years of the republic. Politically, it was characterized by intense organizational and ideological competition between a wide variety of political parties and groups proposing conflicting solutions to such fundamental issues as the definition 45

Schmitter, "Still the Century?" pp. 107-108.

140 — Bases of Politics of the political community and the nature of the regime. These early crises of identity and legitimacy were eventually resolved by the triumph of the Turkish nationalist movement and the establishment of a constitutional, secular republic. Insofar as one can speak of a coherent economic policy during this period, it can be described as a modest liberalism that sought to promote domestic industry by sporadically encouraging the development of a native middle class. But, like the simultaneous experiments with parliamentary democracy, liberal economic policy was frustrated and undermined by nearly continuous mobilization for war and by the imperialist threat to national independence. Policy toward associations was basically pluralistic. Both the new Ottoman corporatist structures regulating big business and the vestiges of the traditional guild system were abolished and replaced by voluntary associations. A nascent labor movement began to organize with the support of small left-wing parties and mounted two major series of strikes against foreign concessionaires immediately after the 1908 revolution and during the War of Independence. Yet, despite an unusually strong endorsement of the freedom of association in general, the persistent insecurity of political elites promoted the recurrent repression of associations among workers and ethnic minorities. Limited State Corporatism (1925-46) This was the period of tutelary, one-party rule under the Republican People's Party. Politically, it was marked by the abolition of all autonomous opposition, the consolidation of the R.P.P. as the "synthesis of the nation" and the "guardian of the republic," and the eventual merger of state and party into a centralized, authoritarian apparatus. Economic development policy gradually abandoned reliance on the private sector in favor of rapid, state-directed, import-substituting industrialization. This was supported and financed through a combination of quasi-mercantilist state control over foreign trade, heavy protectionism, and primitive accumulation in the agricultural sector. On the whole, it was a period of drastic reduction in political participation in favor of consolidating the authority of the new nation-state and extending its governing capabilities. Public policy toward associations was a limited type of state corporatism that aimed, primarily, at political demobilization and, secondarily, at filling out organizational space with semiofficial occupational associations as supporting structures for etatist economic policies. There was an abrupt and universal crackdown on existing voluntary associations, which were replaced gradually by party-controlled umbrella groups. Trade unions were outlawed, while the state attempted the preemptive co-optation of a still

Formal-Legal Bases — 141

limited industrial work force through paternalism. Corporatist occupational associations were reestablished to regulate industrialists and merchants and expanded gradually to include the esnaf, exporters, lawyers, and journalists.

Debilitating Pluralism (1946-60) The last four years of People's Party rule brought a major attempt to move away from authoritarianism by encouraging a transition to free elections and multiparty competition. But this was followed by a decade of Democratic Party governments in which electoral support of the newly participant rural masses was gradually translated into the tyranny of the majority over the former ruling party and its urban supporters. Political participation became universal, but was largely restricted to direct exchanges between parties and voters in the electoral arena, since secondary associations were either excluded from political activity or controlled by the ruling party. A new policy of economic liberalism, sustained by massive foreign aid, encouraged the expansion of the private sector through generous subsidies to agriculture and private industry. The highly unequal impact of inflationary policies resulted in a general redistribution of resources from urban wage earners and public employees to a coalition of urban businessmen and the provincial middle class of small, independent producers and traders. Public policy promoted broad freedom of association, but restricted associational activities in such vague terms that group legitimacy was in constant jeopardy. The proliferation of thousands of competing, feeble, and politically manipulable voluntary associations facilitated government domination of associations allied with the ruling coalition and frequent repression of groups allied with the opposition. At the same time that Turkish policy was promoting this disorganizing and fragmenting variety of pluralism, the U.S. foreign aid mission in Turkey was beginning to develop a policy that may be termed A.LD. pluralism, in which American economic assistance would be used in creating business, professional, and labor organizations as the building blocks of pluralist democracy.46 In fact, the Democratic governments channeled much of this assistance into creating a series of new corporatist occupational structures, thereby extending the precarious privileges of public associations to most of the free professions. Thus, even during the period when public policy most closely approximated the pluralist paradigm, group legitimacy was more easily ex* Kenneth Redden, Legal Education in Turkey (Istanbul: Fakiilteler Matbaasi, 1957).

142 — Bases of Politics tended when interest representation was structured through less autonomous public associations. Emerging Societal Corporatism (1960-80) The period from I960 to 1980 was characterized politically by a series of weak coalition or majority party governments, punctuated by military coups and coup attempts. Proportional representation in parliament encouraged the emergence of several smaller parties filling out the ideological spectrum to the left and right of the two major parties, and there was a growing tendency for conflicting social and economic interests to be represented in different parties as well as through a more diversified network of secondary associations. Economic policy placed increasing emphasis on planned development. The principal goals of economic policy were greater coordination of the public and private sectors, more rapid and concentrated industrialization to produce manufactured goods that can be competitive in foreign markets, and ultimate integration into the European Economic Community by the 1990s. After 1960 the general thrust of developmental policy also underwent a gradual, but clear, shift from its early concern with increasing "social justice" through immediate improvements in living standards to its current emphasis on immediate "sacrifice" in order to achieve higher rates of savings and investment. On the whole, a series of weak governments in pursuit of increasingly ambitious economic goals were confronted with growing demands from a politically more organized and conscious citizenry, resulting in simultaneous crises of participation and distribution. As developmental policy came to endorse deferring a broader distribution of the benefits of economic growth until the end of the century, public policy toward associations became a key element in the attempt to reconcile rapid economic development with democracy and to avoid a return to overt authoritarian rule. Policy toward associations sought to encourage cooperative interest group leaders in the major economic sectors to share the responsibility for implementing developmental programs and to voluntarily moderate demands for political participation and economic redistribution. The principal means through which governments attempted to achieve these goals was a general corporatization of associational interest representation, which was promoted, especially by the Justice Party, in three major forms. First, the state attempted to foster collaborative class relationships and to coordinate nationwide collective bargaining through vertically structured, quasi-monopolistic peak associations in the key economic sectors. Second, there was a significant expansion of the already substantial privileges and economic activities of the public associations, including new provisions

Formal-Legal Bases — 143

for their direct participation in industrial production. Third, important new restrictions were placed on the political activities of voluntary associations in order to inhibit uncooperative and dissatisfied interest group leaders from exploiting noncorporatist channels of representation and forming alliances with opposition parties. Although we shall deal with the problem more fully in the second part of this study, we should note here that it is difficult to describe the emergence of societal corporatism in Turkey as the gradual decay of "advanced pluralism'' in the context of "advanced, concentrated capitalism.'' Turkish capitalism is not yet advanced and concentrated, but advancing and concentrating. Because of its internal structural weaknesses and unfavorable international context, Turkish capitalism continues to be plagued by the problems of delayed, dependent development. Furthermore, Turkish pluralism, such as it is, has never been characterized by a highly institutionalized system of collective bargaining among a well-organized network of collaborative groups or between those groups and the state. Instead, we are dealing with a collection of growing, but still weak, insecure, and manipulable associations, eager for official recognition and privilege in order to improve their political and economic positions immediately and in the long run. It seems reasonable to conclude that the continued weakness of Turkish capitalism and the peculiarly restricted and fragmented Turkish experience with associational pluralism have resulted in a much less clear-cut distinction between the earlier form of state corporatism and the emerging form of societal corporatism in Turkey than has been observed in Europe and Latin America.47 The corporatization of Turkish associational life is not simply an example of the emergence of societal corporatism but rather a highly unusual and unstable mixture of both state and societal corporatist 47

Drawing on European and Latin American historical experiences, Schmitter focuses on two important variables that define clear differences between state and societal varieties of corporatism—the degree to which representative associations are subordinated to the state and the pattern of class relations that public policy seeks to promote. He suggests that state corporatism has been characterized by associations that are created by, dependent upon, and penetrated by the state, whereas societal corporatism has included associations that emerge spontaneously, retain considerable autonomy, and themselves penetrate the state. The abrupt demise of incipient pluralism and its forceful replacement by state corporatism have involved the enforcement of "social peace" by repressing and excluding the autonomous articulation of subordinate class demands in a situation where the bourgeoisie is too weak, internally divided, externally dependent, and short of resources to respond effectively to these demands within the framework of the liberal democratic state. On the other hand, the gradual decay of advanced pluralism and its subtle displacement by societal corporatism have involved the need to strengthen and stabilize bourgeois-dominant regimes by co-opting and incorporating subordinate classes and groups more closely within the political process (Schmitter, "Still the Century?" pp. 100-105).

144 — Bases of Politics types. It is clear that the initiative for the corporatization of interest representation has come primarily from the state's attempt to co-opt reliable clients who are willing to share responsibility for policy implementation, and not from a confident and aggressive set of interest group leaders seeking to translate preexisting organizational strength into a greater role in policy formation. Schmitter describes the emergence of societal corporatism in Europe and North America as an "osmotic process whereby the modern state and modern interest associations seek each other out.'' In elaborating the basic motivation for this mutual interest in developing stable clientelistic relationships, he notes that the more the modern state comes to serve as the indispensable and authoritative guarantor of capitalism by expanding its regulative and integrative tasks, the more itfindsthat it needs the professional expertise, specialized information, prior aggregation of opinion, contractual capability and deferred participatory legitimacy which only singular, hierarchically ordered, consensually led representative monopolies can provide. To obtain these, the state will agree to devolve upon or share with these associations much of its newly acquired decisional authority.48 After 1960 Turkish policy makers were unable to rely on such an osmotic, evolutionary process. Previous policies of debilitating pluralism had prevented the emergence of strong representative associations possessing the political and organizational resources that the state now required to achieve its expanded economic goals. Since I960 the economic development programs of the ruling elites have been too ambitious, the level of mass political and economic demands has been too high, and the institutional structure of associational life has been too weak and disorganized to permit reliance on a spontaneous and evolutionary process of corporatization. Instead, successive Turkish governments have provided the major push to create singular, hierarchically structured representational cartels in sectors where they had not existed previously (as among the esnaf and in the labor movement) and to bolster them in sectors where they had been created during the earlier phase of state corporatism but remained weak (as in business and several of the free professions). Consequently, several aspects of the process of corporatization and the most important associations that it has spawned appear to be a continuation of the earlier phase of limited state corporatism. Official recognition of an association as the authoritative spokesman for a given sector is generally granted from above rather than imposed from below. Representational monopolies are legislatively created 48

IMd., p. 111.

Formal-Legal Bases — 145 rather than independently conquered, and association leaders often seek the elimination of their rivals through state harassment rather than through superior organizational competition. Hierarchical organizational structures are the result of statutory decree and dependence on the administrative apparatus of the state rather than of an internal process of bureaucratic consolidation. Controls on leadership selection and interest articulation are imposed by ministry or party leaders rather than produced by bargaining and mutual consensus. And, finally, participation in decision making is generally formal and consultative rather than effective and collaborative. Turkish attempts to encourage the development of societal corporatism display similar evidence of the continuing relevance of earlier state corporatist practices in the pattern of class relations they have sought to promote. The process of corporatization can be described as an attempt to consolidate the tenuous political hegemony of a weak and internally divided bourgeoisie that is simultaneously confronted with the continuing problems of delayed, dependent economic development and new popular demands for social justice. The major proponents of corporatization apparently regard it as a potential means for incorporating the associations of subordinate classes, especially labor unions and esnaf groups, within the framework of liberal democracy in order to strengthen the unstable bourgeois-dominant regime. Although Turkish corporatists consider this to be an attractive alternative to previous policies of repression and exclusion, they do not intend that corporatization be accompanied by the establishment of a universal welfare state as it was in Europe. On the contrary, their purpose is to postpone redistributive initiatives for another generation in favor of a stronger effort toward industrialization. Turkish corporatism is attempting to restructure interest representation in order to promote collaborative class relations and policies of economic rationalization. Yet the planning process it is intended to serve has been geared, not to thefine-tuningand stabilization of advanced industrial capitalism, but to a more protracted and severe phase of uneven development. In conclusion, then, it appears that despite the reservoir of attitudes and values in Ottoman and Turkish political culture that might be tapped to support it, the attempt to promote a Turkish variety of societal corporatism is severely hindered by the weakness of capitalist development and by the absence of a preexisting tradition of strong associational pluralism. The long-term consequences of this state-initiated and state-directed experiment with societal corporatism are likely to be very different from the professed intent of its proponents. It is likely that, over time, the clearly discriminatory impact of such policies will increase rather than moderate the per-

146 — Bases of Politics ception of interest conflict both within and between all of the major economic and occupational sectors. In this type of situation corporatist policies could quickly become selfdefeating and produce the unintended consequence of increased conflict that is expressed through competitive, pluralist associations. This means that although the proponents of corporatization recommend it as a means of preserving the formal framework of liberal democracy and avoiding a return to prolonged authoritarian rule, in the long run it is doubtful that their policies can be successful unless they are willing to employ the kind of coercion that they claim to disavow. If, on the other hand, they are unwilling or unable to employ such coercion, the eventual dissolution of this unusual and unstable mix of state and societal corporatism may provide Turkey with its first opportunity to experience a more genuine approximation of associational pluralism—or to experiment with alternatives to both pluralism and corporatism.

PARTII

The Network of Interest Group Politics

CHAPTER 5

The Emergence and Diffusion of Interest Groups: The Art of Association as a Dimension of Modernization

THIS SECTION will present the results of a quantitative analysis of the emergence and diffusion of various types of private, voluntary associations from 1946, when the Law of Associations was amended to encourage association formation, through 1971, when military intervention accelerated attempts to restructure associational life. The main purpose of this chapter is to provide an overview of the kinds of voluntary associations that have arisen in the context of Turkey's "debilitating pluralism" and to explore their underlying socioeconomic and political bases. As such it will present a backdrop for our discussion in chapter 6 of the recent emphasis on corporatist associations as alternative channels of interest representation. We will present a time-series analysis of broad changes in associational life over a twenty-five-year period and investigate the relative impact of public policy, socioeconomic development, and political culture on the emergence and diffusion of eleven types of voluntary associations. The discussion will focus on four major topics. First, we will compare the growth trends of each variety of association to determine their relative rates of emergence and their periods of greatest proliferation. Then, we will examine the changing social and economic bases of association formation by observing long-term relationships between various types of association and selected indexes of modernization. Third, we will describe the regional distribution of various associations in order to determine the degree to which ecological and cultural variables have encouraged or impeded their diffusion. Finally, we will investigate the changing relationship between various types of association formation and support for the two major political parties to uncover indirect evidence of the interpenetration of party and associational organizations and the role of certain associations in the gradual realignment of the national party system. The data used in this study were acquired from a number of official Turkish sources. Figures on association formation are drawn from Ahmet Yucekok's Turkiye'de Dernek Geliqimleri (The development of [private]

150 — Network of Politics

associations in Turkey).1 Yucekok's booklet is the first public disclosure of official Ministry of Interior records of all formally registered voluntary associations. It contains the number of associations (divided into fourteen separate categories) found in each of the sixty-seven provinces for every year from 1946 through 1968. In order to examine the changing socioeconomic and political correlates of association formation over the twentyfive-year period, national and provincial data were obtained on population size and on four selected indexes of modernization (urbanization, industrialization, mass media exposure, and literacy) for each five-year interval between 1945 and 1970. The data base also includes election returns for all seven national parliamentary elections between 1950 and 1973.2 It is helpful to make some preliminary and cautionary remarks about the general problems of inference involved in aggregate data analysis as well as specific questions raised by the nature of the data in hand. One problem is the danger of committing the "ecological fallacy."3 On the most fundamental level, it is clear that a correlational analysis whose basic unit consists of provinces cannot by itself provide adequate evidence of the phenomenon that is at the center of our interest—the propensity for an individual possessing certain specified characteristics to participate in associational life. Since ecological correlations cannot be used as substitutes for individual correlations, our conclusions must be limited to statements about collectivities in provincial units rather than inferentially extended to individuals.4 Nevertheless, in view of the small number of other studies of associational development in Turkey and the restricted opportunities for survey research on political issues, such indirect evidence can still be 1 Ahmet Yiicekok, Tiirkiye'de Dernek Gelisimleri [The development of associations in Turkey] (Ankara: Siyasal Bilgiler Fakultesi, 1972); published as an appendix to the author's study of religious associations. 2 Although excellent Turkish electoral data are widely available, acquisition of detailed statistics on comparable indexes of modernization over such a broad time span would have been impossible without the careful assistance of several officials in the State Planning Organization, the State Statistics Institute, the Ministry of Reconstruction and Settlement, and the General Directorate of Post, Telegraph, and Telephone. 3 Erwin Scheuch, "Cross-National Comparisons Using Aggregate Data: Some Substantive and Methodological Problems," in Comparing Nations: The Use of Quantitative Data in Cross-National Research, ed. Richard Merritt and Stein Rokkan (New Haven: Yale University Press, 1966), pp. 131-168. 4 For example, the observation that across provincial units the formation of religious associations is negatively related to urbanization and positively related to mass media exposure does not, in the absence of supporting supplementary evidence, prove that within provinces such associations are being established or joined by residents of small towns and villages or by owners of radio receivers. Similarly, the finding that formation of esnaf associations is strongly related to the Justice Party vote does not independently establish the support of such associations or their members for the Justice Party.

Art of Association — 151

valuable in an exploratory effort to generate and specify hypotheses that can be taken up with more appropriate methods in future research. To claim more for such a study at this time would be to exaggerate its utility in understanding the broad questions of social and political change that concern us. But to attempt less would amount to forgoing an opportunity to learn what we can from the available data and to provide some clues for other researchers with more specific interests and greater opportunities to use survey techniques in studies of particular regions and communities. Regarding the quality of the data, there are several problems of reliability, comparability, and validity. The greatest difficulty concerns our measure of the dependent variable—associability. Since data on association membership are generally lacking or unreliable on both national and provincial levels, we are forced to conceptualize associability here in narrow terms of the propensity to establish associations rather than the propensity to join and participate in them. Our measure of associability is a simple ratio of the number of a given type of association to the relevant population. In the strictest sense, then, this is a study of association formation, not association membership. The distinction between these two dimensions of associability is itself an important one that would merit careful examination if comprehensive data were available on membership.5 At this point, however, I can only alert the reader to the implicit tendency for this portion of the study to overestimate the degree of spontaneous associational participation as a consequence of measuring associability in terms of the mere number of organizations.6 3

Private, voluntary associations have commonly had to deal with persistent problems of low membership rates among their organizable constituencies, financial difficulties due to members accepting membership in name only and refusing to pay even minimal dues, and intense rivalry among leaders who establish new organizations to compete for the same limited and often apathetic membership. These problems have been chronic and they have repeatedly influenced the policies and strategies of association leaders in dealing with governments and political parties. 6 This tendency toward overestimation is further compounded by the record-keeping practices of the Ministry of Interior. Although law requires all new voluntary associations to register with the ministry at the time of their formation, subsequent monitoring and surveillance by the ministry are unusual, and no official agency has ever conducted a nationwide census of associations. Although we can be reasonably certain that the ministry has scrupulously recorded associational "birth certificates," we have no comparable information about' 'death certificates.'' The problem is less important for occupational associations, which tend to have a wide range of activities and large potential constituencies among social classes, than for the far more numerous, but relatively short-lived, community associations, which are commonly oriented toward specific, short-term goals of construction or civic improvement in a particular neighborhood or village. Yet, the fact remains that we know very little about the survival rates of any of these organizations. Consequently, we are inevitably counting many corpses without being able to distinguish them from live specimens.

152 — Network of Politics

Our measure of the dependent variable can lead to both overestimation and underestimation of associability in another way. In effect, we are constantly in danger of "overrewarding" provinces and sectors where a given associational leadership is highly fragmented and producing organizational proliferation and "underrewarding" provinces and sectors where the tendency is to fill in existing associations with more members rather than to redistribute a stable membership among squabbling rivals. It is not always accurate to infer that high or low rates of association formation indicate the presence or absence of strong and active associational life.7 A final problem with our measure of associability is the set of fourteen categories used to classify associations. These categories seem to have been arbitrarily determined by Yucekok during his inventory of ministry files. In general the classifications are quite straightforward. Six of them may be described as occupational or class-based associations—labor unions, employers' unions, and professional, esnaf, agricultural, and civil servants' organizations. Five more are community and civic improvement associations^—religious, cultural, beautification, sport, and social aid organizations. Two categories, foreign and "other," are too ambiguous and heterogeneous for meaningful interpretation, and they have been excluded from most of the analysis. A final type, associations for the struggle against communism, is too recent in vintage and too insignificant in numbers for our purposes. It, too, has been excluded, especially since the author expresses doubt about the reliability of the ministry's enumeration. The remaining eleven categories are still quite heterogeneous and in general we suffer from the inability to disaggregate them or to reformulate them on a new basis.8 7

For example, the observation that when population size and level of economic development are controlled, Zonguldak has a higher propensity for labor union formation than Istanbul does not necessarily indicate that the labor movement is stronger or better organized in Zonguldak. In view of the recent history of labor violence, wildcat strikes, and union corruption among the Zonguldak coal miners, the opposite inference is far more persuasive. Istanbul's apparently more modest performance in union formation is likely attributable to the relative success of its labor leadership in containing disputes over organizing and determining recognized collective-bargaining agents. On virtually any other quantitative or qualitative measure of associability, Zonguldak would rank well below Istanbul in labor organization. Delwin Roy, "The Zonguldak Strike, 1965: A Study of Industrial Conflict in Turkey" (Ph.D. dissertation, Purdue University, 1968). 8 Because the author was primarily concerned with religious organizations, he divides only that class into subgroups that can be examined separately. As we shall see, disaggregation of this category does reveal a clear tendency for mosque-building associations to decline over time and be replaced by other varieties of religious organization such as groups for studying the Quran and for educating and financially aiding preachers. Similar tendencies undoubtedly

Art of Association — 153

For the indexes of socioeconomic development validity is a less serious problem than reliability and comparability. Even though the quality of Turkish national census data is certainly well above the international average, reliable data collected in each province according to consistent categories comparable over a twenty-five-year period are extremely rare.9 As a result, we have occasionally been forced to sacrifice some degree of accuracy in order to gain wider scope and have had to make some corrections of and interpolations from official data to improve comparability. In most official Turkish sources urbanization is measured by the proportion of total population living in legally constituted "municipalities." Since this is a legal-administrative definition rather than a demographic one, we have chosen instead as a stricter but more meaningful measure of urbanization to use the proportion of the total population residing in settlements of more than ten thousand inhabitants.10 Industrialization is measured by the percentage of the economically active population employed in manufacturing and mining." As our index of mass media exposure we exist within other categories, but are not detectable given the present arrangement of the data. The ability to disaggregate categories would have been particularly useful in examining occupational associations. For example, we could have distinguished between esnaf involved mainly in manufacturing and those involved mainly in commerce; between agriculturalists who market different types of crops that are differentially affected by government price supports and supportive purchasing arrangements; between workers in manufacturing and workers in other economic sectors that have been far more resistant to union organization such as construction, communications, and services; between "professionals" who practice in what are commonly recognized as the free professions and those who are more accurately classified with businessmen and employers; between civil servants employed by the central ministries (many of whom could also be classified as professionals) and those in the more numerous but far less important local and municipal unions (many of whom could more aptly be considered workers). 9 Before 1955 the categories according to which demographic and socioeconomic data were collected and reported often shifted from census to census and departed from internationally accepted classifications. Furthermore, since 1960 the proliferation of newly specialized ministries and agencies has produced not only more numerous and competent means of data collection but also more inconsistency and conflict between competing sources. Traditionally accepted official sources cannot always be considered the most reliable, but their competitors seldom have comprehensive figures at the provincial level extending beyond the last ten to fifteen years. 10 This was possible by disaggregating census figures for the earlier periods and by relying on a recent series of regional settlement studies by the Ministry of Reconstruction and Settlement. 11 Unfortunately comparable data are not available for the important year 1950. In that year the State Statistics Institute inexplicably expanded its definition of the economically active population from persons fifteen years of age and older to persons five years and older, but then reported figures only for areas with five thousand or more inhabitants. The result

154 — Network of Politics

have chosen the number of radios per thousand population.12 Literacy is measured by the proportion of the total population six years of age and older that is able to both read and write by census standards.13 Electoral data consist of the proportion of votes received by each political party from the total number of valid votes cast in each of the seven nationwide parliamentary elections between 1950 and 1973.14 of this double transformation was to make 1950 data on the national work force incomparable with similar data for all other years, and they have been excluded from our analysis. Occupational data for 1945 present some difficulty, since they were reported originally for the total population rather than the economically active population. We have been able to improve comparability by calculating the occupational distribution of the work force over fifteen years of age, but our figures for that year can only be intelligent estimates. 12 These data are the most suspect in terms of reliability, since they include only formally registered radio receivers. They omit many of the increasingly popular transistor sets as well as radios recently brought into the country by workers returning from Europe. These data also present the greatest difficulties in terms of reconciling the conflicting claims of official sources. The State Planning Organization has conducted mass communications surveys that estimate the total number of radios to be about 15 percent greater than the figure reported by the State Statistics Institute and the P.T.T. (see Frederick Frey, 7"Ae Mass Media and Rural Development in Turkey [Ankara: State Planning Organization, 1966]; Devlet Planlama Teskilati, Ikinci Bes Yillik Kalkmma Plam Radyo ve Televizyon Ozel lhtisas Komisyonu Raporu [Ankara: D.P.T., 1966]). Even though the State Planning Organization estimates are undoubtedly more reliable, we have preferred to use the official P.T.T. figures because of their comprehensive coverage of all provinces for the entire twenty-five-year period. 13 Here again the data for 1945 present a problem of comparability, since in that year reading ability alone was accepted as sufficient evidence of functional literacy and the population base consisted of those seven years of age and older. 14 These data are generally of higher quality than those for any other variable considered here. But in this study, as in all other examinations of contemporary Turkish politics, these data cannot be exploited fully. The relatively retarded state of the study of the social bases of Turkish political parties does not stem from problems with electoral data but rather from two other problems that still prevent social scientists from using socioeconomic characteristics as effective predictor variables in electoral research. The first is the nonexistence or inaccessibility of socioeconomic data below the provincial or county (i7ce) level. The second problem is the still restricted range of opportunity for native and foreign social scientists to conduct public opinion surveys involving political issues. As a result, previous studies of the changing social bases of political party support have been limited to the upper and middle levels of party leadership, focusing on the social backgrounds of members of parliament, cabinet ministers, and candidates for positions on party electoral lists. But little is yet known about changes in the parties' broader constituencies or their relationships with other representative associations. For example, see Frederick Frey, 7"Ae Turkish Political Elite (Cambridge: M.I.T. Press, 1965); Frank Tachau, "Social Backgrounds of Turkish Parliamentarians," in Commoners, Climbers, and Noteables: Social Ranking in the Middle East, ed. C.A.O. van Nieuwenhuijze (Leiden: E. J. Brill, 1975); Ergun Ozbudun, Social Change and Political Participation in Turkey (Princeton: Princeton University Press, 1976); Dennis Williams, "Local Level Politics: A Study of Political Parties in a Turkish Province" (Ph.D. dissertation, Johns Hopkins S.A.I.S., 1976); Muzaffer Sencer, Tiirkiye'de Siyasal Partilerin

Art of Association — 155 GROWTH TRENDS IN ASSOCIATION FORMATION

The increasing importance of associational life in Turkish society and politics is only partially reflected in the rapid numerical proliferation of voluntary groups. We will be able to observe the remarkable development of the art of association in Turkey from several points of view, including the rich diversity, broad geographic distribution, and expanding activities of secondary associations. Initially, however, it is useful to examine the extraordinary growth of the sheer number of voluntary associations, even while remembering that the encouragement of such a huge volume of organizations has often been part of an intentional official policy of hindering some types of collective action by encouraging intense associational rivalry and fragmentation. Under the influence of the highly restrictive 1938 Law of Associations, adopted in the final year of Kemalist rule, the total number of associations grew slowly from a handful of 205 organizations in 1938 to a modest 820 groups in 1946. Given the explicit prohibition against associations based on social class, associational life in this period consisted mainly of local sporting clubs that were concentrated in the metropolitan centers of Istanbul, Ankara, and Izmir. During the last few years of the one party era, from the liberalizing amendments to the Law of Associations in 1946 until the 1950 elections, the total number of groups grew rapidly to over 2,000. The end of the formal ban on class organizations was followed by a notable increase in occupational associations during this period. Although Turkish associational life accordingly began to acquire a somewhat more diversified character, it was still largely confined to the three largest cities. The greatest explosion of association formation clearly occurred during the decade of Democratic Party rule, when the number of associations multiplied approximately eight times to exceed 17,000 by 1960. This was a period of rapid growth and extensive geographic diffusion for a wide variety of associations. It signals the emergence of associability as a nationwide phenomenon, extending broadly though still unevenly over several aspects of social and economic life in nearly every region. During the decade between the Turkish military's two successful coups of 1960 and 1971 this already large organizational base again multiplied two and onehalf times. By 1970 the total number of associations reached an estimated 42,000. Though the vast majority of these were still small and politically insignificant local community organizations, several of the largest occuSosyal Temelleri [The social bases of political parties in Turkey] (Istanbul: Gecis Yayinlan, 1971).

156 — Network of Politics

pational associations had become active and influential participants in Turkey's expanding network of interest group politics. The enormous increase in voluntary organizations is no less striking when we consider the doubling of the national population between 1940 and 1970. Allowing for the effects of Turkey's population explosion, we see that the number of associations per million people grew from 40 in 1946 to over 100 by 1950, to 620 in 1960, and then to 1,200 by 1970. The unusual proportions of these increases in association formation can also be seen by placing them in the larger context of other aspects of social change in the last thirty years. Figure 5-1 juxtaposes associational growth to increases in several indexes of modernization since 1945, using that year as a base from which to calculate figures for subsequent five-year intervals. While all indicators have risen at an overall rate at least two or three times that of total population, none approaches the sixty-fold increase in community associations or the fifty-fold rise in class associations.15 Especially after 1955 the growth of voluntary organizations appears increasingly more advanced than, and removed from, underlying social and economic change. These findings, however, do not necessarily support the repeated claims of Justice Party leaders, martial law commanders, and right-wing politicians that Turkish society is experiencing a mounting and uncontrolled proliferation of voluntary groups. This has been a blanket charge leveled against all associations as a justification for imposing new restrictions on association formation and activity regardless of the differential impact of groups on the political system.16 A closer examination of associational growth trends suggests that, in fact, such claims seriously exaggerate the degree of Turkey's "associational inflation" and underestimate the degree to which changes in the political system have repeatedly influenced association formation as various political authorities have tried to encourage or discourage it. Figure 5-2 presents two summary views of associational growth trends between 1946 and 1968—one in terms of the annual number of total associations and the other in terms of the annual percent of increase 15

The upsurge in associability is somewhat exaggerated here by our selection of 1945 as a base year, because associability had been generally repressed until then. It could be argued justifiably that association formation is a dimension of modernization that had long been intentionally retarded and whose apparent leap ahead of other dimensions of social change might actually be revealed to be a mere matter of catching up with other processes of change if a more suitable base year were chosen. But even if we narrow our time frame to the twenty years after 1950, we still find an impressive twenty-three-fold increase in community associations and a seventeen-fold rise in class associations compared to only ten times as many radios, four times as many periodicals and potential readers, and three times as many city dwellers and industrial workers. 16 Cumhuriyet, 10 October 1968.

Art of Association — 157 community associations/

6,000

4,000

2,000

1,000

500

literate population,

urbon

300

population

100

industrial workforce

1950

I960

1970

5-1. Association^ Growth and Socioeconomic Change, 1945-70 (Selected Indexes: 1945 = 100)

FIGURE

in the number of total associations. The points on the graphs represent the actual observations for each year and the solid lines are the lines that are best fitted to the data points according to the principle of least squares. The overall pattern for the number of associations is best described by an extended S curve whose slope steepens during the early 1950s, but then declines in the 1960s. The pattern for annual rates of increase in the number of associations follows a downward curve of gradually diminishing slope. Average annual rates of increase fall from between 25 and 30 percent in the early 1950s to about 15 percent at the end of the decade, and then to around 10 percent during the 1960s. Both trends indicate a clear deceler-

158 — Network of Politics

4Oh

Annuol number of associations (thousands)

1950

1955

FIGURE

I960

1965

Annual percent of increase in number of associations

"

1950

1955 I960

1965

5-2. Growth Trends in Association Formation, 1946-68

ation in the national rate of association formation that contradicts the widespread impression of sustained and extreme organizational proliferation. It appears instead that associational formation in general has passed its peak and is gradually approaching a point of saturation in which there is a tendency to fill in and consolidate existing organizations rather than to establish new ones. Looking beyond general trends, we see some clear deviations from the overall patterns, revealing specific periods when association formation has been unusually accelerated or retarded. For example, the points indicating the annual number of groups describe, not a simple S curve, but a curve whose steepest slope appears in the late 1950s, after which it seems to be depressed sharply following 1960 and then to spring back again in the late 1960s. Periodic fluctuations are even more evident in the graph of annual rates of increase. In that figure we find a periodic phasing of deviant points in what appears to be a series of downward and upward oscillations whose rhythm closely corresponds to the timing of national elections and major changes in government. The greatest downward deviations occur just before and after the change of regime in 1950 and immediately after the I960 coup. The greatest upward deviations occur during the period of consolidation of Democratic Party rule in the early- to mid-1950s (most spectacularly on the eve of the important 1957 elections) and immediately after the takeover of the Justice Party in 1965. The picture that emerges is hardly one of uncontrolled associational development. On the contrary, the evi-

Art of Association — 159 TABLE 5-1 The Diversification of Associational Life, 1946-70 (Number of Associations per Million Population)

Class Associations Labor Employer Professional Agricultural Esnaf Civil servant Community Associations Religious Cultural Beautification Social aid Sport Foreign Other Total

1946

1950

1955

1960

1965

1970

.7 — 1.5 1.1 5.5 2.2

4.3 — 2.5 2.3 12.1 7.9

11.4 .1 4.9 6.0 60.8 14.5

14.7 .8 6.8 9.7 99.0 24.0

20.7 3.4 8.3 18.1 114.3 46.2

31.7 6.2 7.3 20.4 103.9 76.2

.6 5.6 1.8 3.8 18.2 1.9 .7

7.3 13.6 4.9 10.0 33.3 3.6 1.7

45.4 45.8 13.7 17.8 57.2 4.8 3.4

184.0 90.3 30.7 22.8 121.5 7.9 8.3

230.0 138.3 85.2 35.2 135.0 7.8 17.0

340.0 199.5 152.2 47.3 161.8 5.8 20.0

43.6

103.6

285.8

620.5

859.5

1172.8

dence suggests a continuing susceptibility of voluntary groups to frequent interference by political authorities, who shift back and forth between sponsorship and repression. Our discussion thus far has treated associability as a single variable. In fact, we are dealing with a complex summary variable whose separate components should also be examined individually. Therefore, we have adopted a dual strategy of conceptualization in which associability will be treated both in summary and disaggregated forms. Such a strategy allows us to make useful generalizations about associability as a whole and to explore comparatively some important differences among the specific varieties of association. Table 5-1 presents a more differentiated picture of the growth of the art of association in which the total national figures are disaggregated according to association type at five-year intervals. It reports the number of each type of association per million population, providing a global view of changes in the variety of associational life in which population increase is controlled. The particular years at which the data are reported also provide a rough impression of the differential impact of major political events and policy changes on various types of association,

160 — Network of Politics since they conveniently demarcate different regimes—the final years of the one-party era (1945-50), the more liberal and the more repressive years of Democratic Party rule (1950-55 and 1955-60), the period of coalition governments (1960-65), and the years of Justice Party rule (1965-70). Interestingly, the most striking general observation provides evidence of considerable continuity rather than change in the composition of Turkish associational life. The highest rates of association formation have consistently appeared among community associations. The overwhelming majority of Turkish voluntary groups has always consisted of small, locally oriented and politically marginal types. Associations most directly related to underlying structural and economic change and to the representation of diverse class interests have never constituted more than one third of the total. The recent decline in the relative share of class associations to a mere one fifth of all groups indicates that, at least in terms of sheer numbers, organization based on and expressing conflicting economic interests is falling even farther behind organization based on and expressing local community solidarity. Among community associations there have been some clear changes in relative importance over time. The relative shares of religious, cultural, and beautification groups have all increased, while those of sport, social aid, and foreign groups have steadily decreased. The most striking reorientation is the replacement of sport associations by religious, cultural, and neighborhood beautification groups as the dominant types of community associability. The period of Democratic Party rule witnessed the beginning of a basic reorientation of community associability that became even more pronounced during the 1960s; the art of association was increasingly directed toward the expression of religious and cultural identity among the dominant Turkish-speaking, Muslim population and of local pride among fellow townsmen. Patterns are commonly much more fitful and broken among class associations than those of community groups. The continuous decline in the relative importance of professional groups is the only example of a clear and unbroken long-term pattern among class associations. The common appearance of uneven and discontinuous growth patterns among labor, civil servant, industrialist, and esnaf associations suggests that these groups have been particularly susceptible to the alternating incentives and inhibitions of the political system. Whereas the changing structure of community associability has proceeded in a fairly clear and constant direction, trends in occupational associability are often reversed from one change in government to the next. The long-term growth patterns of each type of community and class association are summarized in figures 5-3 and 5-4. A comparative ex-

.Mosque Building

_ Sport

.Cultural

46 50 55 60 65 46 50 55 60 65 6,000 - Beautification

Religious

46 50 55 60 65 Socio! Aid

4,000

2,000

• n • *TJ

46 50 55 60 65 46 50 55 60 65 46 50 55 60 65 FIGURE 5-3. Annual Number of Community Associations, 1946-68 (Semilogarithmic Scale) Esnof

Labor

Civil Servont

100 k*

46 50 55 60 65 46 50 55 60 65 46 50 55 60 65 Agricultural

Employer

Professional

2,000 L....1..

46 50 55 60 65

55

60 65 46 50 55 60 65

FIGURE 5-4. Annual Number of Class Associations, 1946-68 (Semilogarithmic Scale)

162 — Network of Politics

amination of the best-fitting lines allows us to identify more clearly specific phases of associational growth and to isolate particular years that have been important turning points between different phases.17 In general we can observe three types of lines, indicating three different associational growth patterns. First, the three associational categories that have historically rivaled one another as the most numerous—mosque-building, sport, and ew/ia/groups—describe an S curve that resembles the pattern for overall associability. Each of these types appears to have passed through three distinguishable phases—an early period of moderate growth, a rapid acceleration during the 1950s, and a relative decline during the 1960s. The second type of growth pattern is the most common in that it includes the greatest variety of associations as well as the greatest number. This is a curvilinear trend of low to moderate growth in the first phase followed by much more rapid increases in the second phase. In many cases the two growth phases are sharply demarcated by a clear turning poinfc—often a single year—that signals an abrupt takeoff of organizational activity. The third and most unusual growth pattern is the almost perfect linear tendency of professional associations. The exceptionally uniform and poor performance of professional groups seems to reflect the generally retarded state of voluntary associability in those sectors where semiofficial corporatist structures are the principal means of interest representation. The other two occupational sectors that until recently had similarly regular and unimpressive levels of voluntary associability—agriculture and business—are also areas in which corporatism has been historically predominant. In examining the relatively muted development of voluntary associations in these three sectors we are also observing, indirectly, one aspect of the generally stifling impact of corporatism in Turkish associational life. The general regularity with which abrupt takeoffs appear in such a wide variety of associational growth curves suggests the utility of separately analyzing each distinguishable growth phase. Thus each growth curve, except for the clearly linear professional pattern, was broken into two or three distinct "phase lines" for the purpose of segmental regression (table 5-2). The approximate years at which growth curves passed from one phase to another are shown in the left-hand columns and the slopes of the lines in each phase are shown in the right-hand columns. The slopes can be interpreted as the average number of each type of association established annually during the specified time periods. This provides crude but convenient and easily interpretable measures of the relative intensity and timing of successive waves of associational organization. 17

The category of religious associations has been disaggregated here into two subtypes— mosque-building organizations and other religious groups devoted to the study of the Quran or to the education andfinancialassistance of preachers.

Art of Association — 163 TABLE 5-2 Phasing and Slopes of Associational Growth Curves Timing of Phases 1-2 Cultural Mosque building Sport Beautiflcation Religious Social aid

1952 1956 1956 1962 1963 1965

Esnaf

1952 1961 1962 1964 1965 —

Labor Agricultural Employer Civil servant Professional

2-3 1961 1962

1959

Slopes ofPha.ses I

2

45 104 110 63 30 43

335 111 360 535 430 155

50 31 19 4 55 12

362 69 52 25 315

3 462 234

127

If these dates are regarded as turning points in the modern history of Turkish association formation, that history can be summarized roughly in terms of the emergence of two successive waves of intense organizational activity. Each wave includes a different collection of association types whose takeoffs occur together or in close sequence. The first includes cultural, mosque-building, sport, and esnaf groups; the second includes all other varieties except professional groups, which never enter a phase of accelerated growth. The second wave begins to rise about the same time that the first begins to break, and the revolution of 1960 appears to be a pivotal event precipitating the change. In fact, the long-term impact of the 1960 revolution on the overall structure of associational life should not be exaggerated. One of the principal intentions of those who drafted the new constitution in 1961 was to encourage the representation and organization of social and economic groups whose interests had been neglected under Democratic Party rule, especially wage earners and public employees.18 The second wave of associational organization inaugurated by the reformers does include notable takeoffs among several occupational groups, but they are quite moderate when compared to the simultaneous and far more dramatic rise of new community groups. When we consider the far greater contribution of beautiflcation, "Dodd.pp. 106-115.

164 — Network of Politics

religious, and social aid organizations, the second wave of association formation appears to be not so much a radical departure from the declining first wave as a continuation of it in a modified form. If during the 1950s community leaders solicited private (and, often undisclosed, public) contributions primarily for the construction of mosques, stadiums, schools, libraries, and dormitories, they shifted their efforts during the 1960s toward imam-hatip schools, Quran courses, bridges and irrigation projects, statues and monuments, tourism, and institutional philanthropy.19 Thus, although it is clear that the I960 revolution roughly demarcates two successive waves of association formation, there is no persuasive evidence that it produced more than a modest improvement in the relatively retarded state of class-based associability. Figures 5-5 and 5-6 provide a clearer picture of short-term changes in associational growth trends in terms of the annual percent of increase in the number of each type of voluntary organization. In general, community associations and class associations appear to respond to different kinds of change in the environment. Short-term upward and downward deviations from the overall trends of community associations are commonly greatest before or after election years (especially 1950, 1957, and 1965) and at the time of the 1960 coup. Similar short-term deviations appear among class associations immediately before and after the election of 1950, when newly founded occupational groups were particularly sensitive to direct government incentives and inhibitions and a major change in regime was taking place. After this important early election, however, the growth rates of class associations deviate much less during election years than those of community associations. As a rule both short-term and long-term changes in the annual growth rates of occupational groups seem to be either responses to innovations in public policy direcdy concerning the groups in question or general adaptations to changing economic conditions. The impact of new enabling legislation is evident in the steady proliferation of labor unions after the adoption of the 1961 constitution and in the brief rise in civil servant groups immediately after the passage of the 1965 Public Employees' Union Law. The most striking response to this more favorable climate for union organization was the rapid expansion of private industrialist associations during the early 1960s. The peak of associational activity among employers coincides exactly with the drafting and adoption of two key pieces of legislation in 1963—the revised Trade Unions Act and the new Law on Strikes, Lockouts, and Collective Agreements. Thus, despite their relatively late entrance into associational life, private employers quickly managed to catch up with labor organization 19

Cumhuriyet, 15 August 1968, 9 December 1970.

Sport

• Mosque Building

60

Cultural

40

20-

50 50

55

60

50

65

Beoutification

55

60

50

65

J

Aid

Socio

Religious

30

—

·•

L

55 60 65

•

Φ

ι

J

50

I

J 55

L

55 60 65

I 60

L 65

j

••

I

101-

50

•

I

I

55

60

I

65

FIGURE 5-5. Annual Percent of Increase in Number of Community Asso­ ciations, 1947-68 Esnaf

60

.

Civil Servant

Labor

40 20 0 -5ci 5

I

I

50 55

I

I—

J

20

J

L

50

L

50 55 60 65

55 60 65

Professional

•

I * ·\

-

_L

I

55 60 65

- Employer

Agricultural

40

50

60 65

/ J f· I

·\ \

* I

I

55 60 65

I

I

I

1·

50 55 60 65

FIGURE 5-6. Annual Percent of Increase in Number of Class Associations, 1947-68

·

166 — Network of Politics

when changes in the legal structure of industrial relations provided them with the necessary incentive. Although recent changes in public policy have promoted increased organization among workers, civil servants, and industrialists, the continuation of corporatism has produced the opposite result among agricultural and professional groups, where steadily diminishing annual rates of growth have fallen near to zero. The impact of long-term economic change is reflected most clearly in the precipitous decline of esnaf associability that began soon after the end of the boom of the early 1950s and deepened as private enterprise became more concentrated in larger firms. Considered together these trends in class associability appear to reflect fundamental economic changes that distinguish the periods before and after 1960, i.e., the increasing competitive disadvantage of small-scale producers and the growing polarization of private capital and organized labor. THE SOCIAL AND ECONOMIC BASES OF ASSOCIABILITY

The extreme diversity of Turkey's ecology and the rapid but very uneven diffusion of modernization present a laboratory for comparing intranationally changes in social and economic structure with changes in the art of association. Table 5-3 outlines the changing parameters of four selected indexes of modernization and associability between 1945 and 1970. The extreme concentration of most aspects of modernity in the three largest metropolitan areas makes it useful to present two sets of statistics—one including Istanbul, Ankara, and Izmir, and one excluding them.20 The uniform increases in the mean values of all of these variables demonstrate an impressive degree of social change over these two and onehalf decades. Nevertheless, large differences in absolute values and rates of increase confirm our earlier impression that certain aspects of modernity are diffusing far more rapidly than others. Industrialization and urbanization clearly lag behind all other processes of social change. Especially during the 1960s rural migration to the cities was exceeding the capacity of the industrial sector to absorb it with new employment. Overall there is no doubt that the differentiation of social and economic structure is experiencing relatively modest rates of change in contrast to sharp increases in all other indexes. Literacy expands consistently at a much higher rate and radio ownership spreads with remarkable spurts in the early 1950s and early 1960s. Again we see the most spectacular change occurring in the 20 Such a strategy gives us a clearer understanding of the special contribution of these areas to the general processes of modernization and associability. It also allows us to see to what degree our conclusions about interrelationships between these variables are determined merely by highly skewed distributions that result from the inclusion of these special cases.

TABLE 5-3 Parameters of Indexes of Modernization and Associability, 1945-70 j

All Provinces

Excluding Istanbul, Ankara, and Izmir

Mean

Standard Deviation

0-80.0 0-84.0 0-84.6 0-87.0 0-89.0 0-90.5

11.2 11.2 13.6 15.8 18.2 20.4

7.9 7.8 8.4 9.6 10.4 11.3

0-30.0 0-31.0 0-36.0 0-41.6 0-44.5 0-47.4

3.4

0-17.5

4.2

3.2

0-17.4

4.9 5.2 5.7 4.7

0-30.0 2.1-29.7 1.8-36.7 .8-28.3

5.7 7.2 6.9 5.8

3.8 4.2 4.0 3.7

0-18.9 2.1-22.4 1.8-24.2 .8-19.1

Mass Media Exposure 1945 5.4 1950 9.1 1955 31.1 34.0 1960 60.5 1965 75.6 1970

7.9 13.1 28.1 28.3 33.2 34.6

.7-58.0 1.9-99.2 7.2-205.0 7.1-207.0 23.0-215.5 21.5-180.2

4.0 6.9 26.7 29.4 55.8 72.3

3.2 5.1 15.3 15.5 24.3 31.0

.7-18.6 1.9-27.0 7.2-78.0 7.1-82.6 23.0-121.8 21.5-133.4

Literacy 1945 1950 1955 1960 1965 1970

23.9 26.3 34.3 34.8 43.9 51.4

13.0 13.9 14.1 12.9 13.4 12.7

5.9-68.0 9.0-70.6 11.5-73.4 11.4-72.9 17.6-77.2 22.8-79.4

22.5 25.0 33.0 33.6 42.7 50.3

11.3 12.5 12.9 11.6 12.5 12.0

5.9-48.1 9.0-51.4 11.5-59.1 11.4-57.9 17.6-68.8 22.8-72.0

Associability 1945 1950 1955 1960 1965 1970

24 66 233 550 798 1216

32 62 130 251 338 501

0-192 0-389 0-730 161-1143 188-1579 255-2373

19 56 222 537 789 1209

17 38 120 249 343 507

0-83 0-150 0-730 161-1143 188-1579 255-2373

Urbanization 1945 1950 1955 1960 1965 1970

Mean

Standard Deviation

13.1 13.2 15.7 18.0 20.5 22.9

12.3 12.8 13.2 14.3 15.0 15.7

4.4 6.3 7.8 7.6 6.4

Industrializati·on 1945 1950 1955 1960 1965 1970

Range

Range

168 — Network of Politics

index of associability, which begins at a modest level and then soon outdistances all other indicators of change by a wide margin. The special position of the three largest metropolitan regions can be seen in the fact that their inclusion in the case base produces substantial increases in both the range and standard deviation of most variables. This provides evidence of the continued concentration of modernity and also suggests that as modernization proceeds, regional disparities and inequalities are widening even further. Similar results appear even when Istanbul, Ankara, and Izmir are excluded. Only literacy displays even the slightest tendency toward more even distribution across the remaining provinces. Thus, the overall pattern of modernization in Turkey is basically exacerbating longstanding inequalities between regions rather than lessening them. The data on associability, however, suggest strongly contrasting conclusions regarding the pattern of diffusion of voluntary associations. After 1950 the inclusion of the three major cities has virtually no impact on any of the descriptive statistics for associability. Although associability began, as did all other indexes of modernization, in a concentrated regional base, the flurry of association formation triggered by the rise of the Democratic Party quickly expanded beyond the major metropolitan centers.21 While the current regional distribution of associations appears to be no less skewed than those of the other indexes, provincial disparities in associability seem to cut across rather than reinforce the uneven distributions of other aspects of modernity. Thus it seems reasonable to infer that although associability originated as one of several closely interrelated and geographically concentrated elements of modernization, it has become an increasingly independent dimension of social change. The correlation matrices in table 5-4 allow us to explore these hypotheses more directly by examining the changing interrelationships between our five selected indexes over the twenty-five-year period. The most striking feature of this table is the presence of so many high rates of covariance. Even considering the tendency of data aggregated in large units to produce artificially high rates of covariance, this is strong evidence that the discrete dimensions of social change are all highly interrelated and constitute a general process or "syndrome" of modernization. Nevertheless, a closer examination of changes in these bivariant relationships reveals a gradual tendency for the process of modernization to become more differentiated 21 The proportion of total associations in Turkey's three largest cities dropped from 47 percent in 1946 to about 19 percent in 1970. Istanbul's share fell most sharply from 27 percent to only 7 percent and Izmir's share declined continuously from 9 percent to 4 percent. Ankara's share also dropped from 12 percent to only 6 percent in 1960, but then rose slightly to 8 percent in 1970 as many federations and nationwide organizations began establishing their national headquarters in the capital city.

TABLE 5-4 The Changing Social Bases of Associability (Correlation Matrix) (Indexes of Modernization and Associability, 1945-70) Excluding Istanbul, Ankara, and Izmir

All Provinces I.

2.

3.

4.

1.

2.

3.

4.

1. Urbanization 1945 1950 1955 19j60 1965 1970 2. Industrialization 1945 1950 1955 1960 1965 1970

.63 — .76 .76 .81 .72

3. Mass Media Exposure 1945 1950 1955 1960 1965 1970

.84 .86 .86 .85 .71 .50

.56 — .80 .78 .81 .65

4. Literacy 1945 1950 1955 1960 1965 1970

.62 .58 .58 .61 .54 .48

.65 — .63 .66 .64 .62

.71 .67 .73 .75 .81 .91

5. Associability 1945 1950 1955 1960 1965 1970

.80 .81 .58 .40 .22 .12

.57 — .70 .61 .41 .40

.86 .85 .69 .54 .56 .65

.57 — .56 .60 .64 .62

.67 .71 .69 .75 .74 .68

.57 .59 .64 .64 .42 .33

.60 — .65 .66 .64 .52

.40 .38 .44 .44 .38 .35

.57 — .52 .55 .55 .55

.72 .73 .77 .78 .80 .90

.48 .50 .45 .33 .21 .11

.50 — .65 .59 .50 .54

.51 .60 .71 .63 .70 .75

.51 .63 .62 .74 .77 .73

170 — Network of Politics

as specific variables are distinguished from a tightly interwoven cluster into increasingly independent dimensions of social change. In general, we observe a gradual disaggregation of these four indexes into what appear to be two more or less distinct dimensions of modernization as seen in the strengthening correlations between urbanization and industrialization on the one hand and between mass media exposure and literacy on the other.22 Over the twenty-five-year period there have been major shifts in the social bases of association formation. The most precipitous declines occur in the relationships between associability and the measures of urbanization and industrialization. The rapid and sustained fall in the influence of urbanization since 1950 illustrates the great extent of geographic diffusion of associability after the transition to multiparty politics. Especially after 1955 there is a fundamental reorientation in the bases of association formation from urbanization and industrialization toward literacy and mass media exposure. This tendency is even more apparent when we exclude the three largest cities from our calculations. Here we see the early emergence of associability originally having nearly identical rates of moderate covariance with all indexes of modernization. But then as the modernization indexes emerge as increasingly distinct dimensions of social change, association formation becomes less related to urbanization and industrialization and more related to literacy and media exposure. Figure 5-7 graphically illustrates these long-term shifts in the correlates of associability using both the total case base and the sixty-four province sample. Although this pattern of bivariant relationships strongly suggests major changes in the social bases of association formation, zero-order correlation alone is not adequate for evaluating the hypothesis and cannot accurately reveal the independent contributions of the predictor variables in determining levels of associability. Therefore, multiple regression was used both as an additional check on our preliminary conclusions and as a technique for examining the relative impact of each independent variable on associability while holding all other independent variables constant (table 5-5). Although multiple regression does improve our ability to predict early rates of association formation, the regression equation steadily loses its 22 A more complex multi variant analysis of Turkish socioeconomic development by Albaum and Davies has also revealed the existence of two principal independent dimensions of modernization, which the authors term "urban development" and "rural development." Both sets of findings point to the emergence of two interrelated but distinct aspects of socioeconomic change—the slow but accelerating expansion of industrial capitalism in the major urban centers and the more widespread diffusion of commercialized agriculture and small-scale production in the provincial towns and villages. Melvin Albaum and Christopher Davies, "The Spatial Structure of Socio-Economic Attributes of Turkish Provinces," International Journal of Middle East Studies 4 (July 1973):288-310.

Art of Association — 171

(All Provinces)

Excluding Istanbul, Ankara and Izmir)

FIGURE 5-7. Correlates of Associability, 1945-70 (Total Associations)

predictive power after 1950. Again we find strong evidence that associability originally emerged as part of a highly interrelated cluster of modernization variables, but that as it expands and diffuses it is becoming an increasingly independent dimension of social change. Regarding the relative contribution of each of the modernization indexes to associability, multiple regression emphasizes even more than did bivariant analysis the dramatic decline of urbanization and rise of literacy as predictor variables. In fact, when all other variables are controlled, increases in urbanization have a consistent and strengthening negative relationship to increases in the art of association after 1950. Literacy steadily improves in predictive power during the 1950s and is by far the strongest contributor to associability from I960 onward.23 23 Although thesefindingssupport the conclusions about the overall shift in the correlates of association formation, the evidence is still confounded by the problem of multicollinearity— the high degree of interrelatedness between all of our supposedly independent predictor variables. Hubert Blalock, "Correlated Independent Variables: The Problem of Multicollinearity," Social Forces 42 (December 1963):233-237; Edward Tufte, Data Analysis for Politics and Policy (Englewood Cliffs, N.J.: Prentice-Hall, 1974), pp. 148-155. In order to minimize the impact of this serious departure from the causal assumptions of multiple regies-

172 — Network of Politics TABLE 5-5 The Changing Social Bases of Associability (Beta Coefficients from Multiple Regressions) (Indexes of Modernization and Associability, 1945-70) 1945

1950

1955

7960

1965

1970

Urbanization Industrialization Mass media exposure Literacy

.2269 .0379 .5849 .0922

.2831 — .4634 .2064

-.1166 .3846 .2183 .3578

- .2752 .4078 - .0746 .7107

-.4014 .1764 .0548 .8041

-.4152 .2161 .1958 .5774

Multiple R Multiple R2

.8779 .7707

.8857 .7844

.7835 .6139

.7995 .6392

.7915 .6265

.7483 .5600

Urbanization Literacy

.6248 .2842

.6012 .3565

.2713 .5325

-.0922 .8161

-.2595 .8896

- .2850 .8269

Multiple R Multiple R2

.8323 .6927

.8609 .7411

.7258 .5268

.7632 .5825

.7791 .6070

.7285 .5308

Industrialization Literacy

.2274 .5253

— —

.4406 .4110

.1930 .6316

-.1145 .8215

- .0361 .7079

Multiple R Multiple R2

.6956 .4838

— —

.7706 .5939

.7733 .5980

.7532 .5674

.6858 .4704

It is interesting to note that Schmitter's study of Brazilian corporatist syndicates between 1940 and 1960 produced exactly the opposite results. Using similar techniques of bivariant correlation and multiple regression, Schmitter found that total associability was increasingly related to indexes of modernization. He also observed clear long-term shifts in the social bases of associability, but again they are the opposite of the changes we see in Turkey—the impact of industrialization and urbanization improved substantially and they become by far the best predictors of Brazilian synsion, we have repeated the calculations using only two modernization indexes instead of all four. Since literacy is the variable that is consistently least related to either urbanization or industrialization, it was paired with these two variables in the duplicate regressions shown at the bottom of table 5-5. Although we still expect one variable to be credited with an exaggerated predominance over the other, it is crucial to our hypothesis to determine if there is a clear shift in the relative importance of the variables over time. Again, such a shift appears most dramatically in the displacement of urbanization by literacy as the key predictor of associability after 1950. In both sets of regression equations we see a long-term decline in the coefficients of multiple determination, indicating associability's increasing independence from all dimensions of modernization even as the relative importance of individual indexes changes.

Art of Association — 173

dical membership, while literacy and mass media exposure changed from good predictors to poor or even slightly negative ones. Schmitter interprets this reversal in the relative importance of modernization variables in explaining associability as "evidence of the contrived nature of early syndicalization—its concentration in Rio, its lack of connection with prior differentiation of social and economic structures, and its essentially political motivation. Subsequently, what was originally a creative, co-optive, preemptive maneuver on the part of political authorities began to lose some of this artificiality."24 Here again Turkey provides a contrasting pattern. As we have seen, the thrust of post-1946 Turkish associational policy did not follow the Brazilian strategy of co-optation through "artificial corporatism'' but rather a strategy of fragmentation through "debilitating pluralism." Early association formation was not preemptive but restrained, not officially contrived but relatively spontaneous, not artificial but closely tied to socioeconomic development behind which it had previously lagged. The increasing independence of Turkish associations from their original social bases suggests that, unlike their Brazilian counterparts, they have become more politicized rather than less so. The observation that the diffusion of associability has involved a corresponding shift in its social and economic bases is, of course, relevant only at the most general level, since we have not yet disaggregated the global index of association formation into its various components. Given the highly disproportionate contribution of a few very numerous associational categories to the measure of total associability, it is possible that these findings disguise the existence of divergent patterns of associational diffusion. David Collier and Richard Messick distinguish between two interrelated yet different patterns of diffusion—hierarchical diffusion and spatial or geographic diffusion. In the case of hierarchical diffusion innovation and change originate in the most advanced or largest centers and are then adopted by successively less advanced or smaller units. More developed units are presumed to be more disposed to innovation due to their greater resources and higher levels of information about innovation once it begins to spread. Less developed and smaller units are presumed to imitate the example of earlier innovators. An alternative type of diffusion is spatial diffusion along lines of geographic proximity or along major lines of communication. According to this pattern change may originate part way down in the developmental hierarchy rather than at the top and then diffuse rapidly among units at widely differing levels of development to produce fairly clear patterns of regionalism. Hierarchical and spatial 24

Schmitter, Interest Conflict, pp. 173-175.

174 — Network of Politics

patterns of diffusion may, of course, often overlap, since socioeconomic development itself generally diffuses over a regular, geographic pattern. Thus, diffusion that is ordered in terms of a hierarchy of development may be spatially ordered as well.25 Indeed, this seems to be the case regarding the general diffusion of the art of association in Turkey. The originally high and then gradually declining relation between total associability and modernization indicates extensive hierarchical diffusion from more developed to less developed provinces. But the particularly precipitous drop in associability's relation with urbanization after 1950 also suggests a very broad geographic diffusion that rapidly covered regions at different ends of the developmental hierarchy. Though we cannot expect to determine definitively the relative importance of hierarchical as opposed to spatial diffusion, a comparison of the changing social bases of each type of associability can illustrate whether different types of association formation have tended to follow different patterns of diffusion. We have, therefore, examined the changing correlates of each type of voluntary association over the entire twentyfive-year period. The changing impact of urbanization and of literacy is illustrated graphically for community associations in figure 5-8 and for class associations in figure 5-9. Correlations between associability and urbanization are shown by the solid lines and correlations between associability and literacy are shown by the broken lines. The pattern of extensive hierarchical diffusion that characterizes total associability is most clearly observable for cultural and social aid groups and to a lesser degree for esnaf and sport associations. Each of these association types originates in the more developed and urbanized areas and then expands into the less developed provinces, so that literacy eventually surpasses urbanization as the principal correlate of associability. Cultural and social aid groups provide the best examples of extensive hierarchical diffusion, since they were originally highly concentrated in the largest metropolitan centers and eventually became two of the least urban-oriented association types. Esnaf and sport associations appear to have diffused, not from the top of the developmental hierarchy, but from more intermediate levels and have not become quite so unrelated to urbanization as cultural and social aid groups. Sport groups have descended the hierarchy in a gradual, steplike fashion, whereas esnaf groups display one of the most abrupt and rapid expansions beyond their areas of origin. In contrast, the remaining association types display comparatively greater stability in their social correlates, suggesting either much less extensive 25

David Collier and Richard Messick, "Prerequisites versus Diffusion: Testing Alternative Explanations of Social Security Adoption,'' American Political Science Review 69 (December 1975):1299-1315.

Art of Association — 175

•6r-

Religious

Beautif ication

-

.4

•p*

.2

-

•Oh

I

45

I

I

50 55 60 65 70 45 50 55 60

I

1

65 70

Cultural

45 50 55 60 65 70 45 50 55 60 65 70 Sociol Aid

45

50 55 60 65 70

FIGURE 5-8. Correlates of Associability, 1945-70 (Community Associations)

hierarchical diffusion or spatial duffusion. Limited hierarchical diffusion is most evident among labor, employer, and professional associations, which have been unique in maintaining consistently stronger relationships with urbanization and industrialization than with literacy and media exposure. Within these three key occupational sectors the diffusion of associability has remained closely tied to the moderate increases in the division of social labor and the gradual emergence of a more differentiated, class-based stratification system in the major urban-industrial centers. The diffusion of professional groups appears to have proceeded farther down the developmental hierarchy than the diffusion of labor unions and employers' associations, which are still highly concentrated in the largest and most advanced provinces.26 Employers' associations follow a highly un26

This would be expected in view of the recent tendency for nationwide employment to expand more rapidly in the tertiary sector than in the secondary sector, but it may also be a

176 — Network of Politics Employer

Labor .8 .6 .4 45

_L J_ 50 55 6 0 65 70 Professional

55 -

60 65

70

Esnaf •

.,'- —

\---

ι

j_

ι

I

I

I

45 50 55 60 65 70 45 50 55 60 65 70 Agricultural

45

50 55

60 65 70 45

Civil Servant

50

55

60 65

70

FIGURE 5-9. Correlates of Associability, 1945-70 (Class Associations) usual pattern of diffusion up the developmental hierarchy until 1965 and diffusion down the hierarchy afterward. This is largely the result of the relative indifference of Istanbul industrialists toward collective action until the late 1950s, when their delayed but rapid organization produced an extreme concentration of employers' associations in that province. Spatial diffusion is most evident among religious, beautification, agri­ cultural, and civil servant associations, which have had consistently weak relationships with all indexes of modernization, especially with urbani­ zation. These types of association appear to have originated, not at the top of the developmental hierarchy, but quite far down it and have become even less related to urbanization as they diffuse. This spatial diffusion is most pronounced for religious groups after 1950, for civil servant groups reflection of corporatist dominance among professionals, since their voluntary groups closely parallel their more important and geographically extensive semiofficial organizations.

Art of Association — 177

between 1950 and 1960, and for beautification and agricultural associations after I960.27 In general, slow and limited hierarchical diffusion is the characteristic long-term pattern for those occupational groups most directly linked to the emergence of the modern capitalist mode of production. Although none of these association types has expanded much beyond the largest urbanindustrial centers, professional groups have diffused the most and labor unions the least. Spatial diffusion is the characteristic long-term pattern for associations of the provincial middle class of small-scale producers and lower-level official functionaries as well as religious and beautification groups. It is among these predominantly provincial groups that we would expect to find the greatest evidence of regionalism in shaping associational development. Finally, esnaf associations and the remaining types of community groups (cultural, sport, and social aid) have followed a mixed hierarchical and spatial pattern in which diffusion originates from the highest or intermediate levels of development, but then rapidly and often abruptly includes provinces at all other levels. These are the associations that have experienced the most radical reorientation in their social bases and that come closest to bridging the gap between metropolitan and provincial life.

REGIONAL PATTERNS OF ASSOCIABILITY

The increasing independence of associability from each of the selected indexes of modernization indicates that association formation has diffused geographically in a different manner from other aspects of social change. Though the various provinces are becoming more disparate in terms of the propensity of their populations to form voluntary associations, the regional distribution of that propensity cuts across the distribution of socioeconomic development. The graph in figure 5-10 provides an overview of the relationship between total associability and socioeconomic development in 1965.28 27 Civil servant groups have departed from this pattern somewhat during the 1960s in favor of the unusual pattern of upward hierarchical diffusion. This is undoubtedly the result of the relatively late organization of public employees in the central ministries in the capital city compared to the earlier activity of their highly dispersed counterparts in local municipalities. 28 The measure of overall development used here is a weighted, multivariant index formulated by the State Planning Organization. While this index has the advantage of conveniently summarizing the rank of each province on a wide variety of measures of modernization, its weighting clearly reflects the State Planning Organization's tendency to conceive of modernization primarily in terms of industrialization. Consequently, the relative independence of associability and modernization is more exaggerated in this illustration than it would be if we had been able to use an index that gave equal weight to nonindustrial variables.

178 — Network of Politics

50 100 150 200 300 325 State Planning Organization Index of Development, 1965 5-10. Associability and Development, 1965 Overassociaters: 1 Ankara, 2 Kocaeli, 3 Bursa, 4 Konya, 5 Burdur, 6 Denizli, 7 Rize, 8 Nevsehir, 9 Cankin, 10 Gumushane Underassociaters: 1 Istanbul, 2 Izmir, 3 Icel, 4 Elazig, 5 Hatay, 6 Gaziantep, 7 Ordu, 8 Siirt, 9 Diyarbakir, 10 Mardin

FIGURE

The overall relationship between associability and development follows a clear curvilinear pattern. The propensity to form voluntary associations increases sharply in the middle ranges of development, but then tends to stabilize among the most developed provinces. Again the evidence indicates that association formation tends to reach a threshold or saturation point beyond which organizational efforts shift from the establishment of new groups toward filling out existing groups with more members. The curve fits considerably better in the lowest and highest ranges of development than in the middle ranges, where deviations are the most striking. In order to provide a preliminary view of these deviations we have isolated ten pairs of provinces whose performance in the general art of association is not commensurate with their similar levels of development. We have termed the provinces listed in the upper columns "overassociaters," since they display a greater propensity for the art of association than they "deserve" for their levels of development, whereas their "underassociating" counterparts in the lower columns display the opposite tendency at equivalent developmental levels. In order to learn whether these deviant cases are part of broader regional tendencies toward overassociability and underassociability, we created a subfile structure in which provincial groupings correspond to Turkey's nine agricultural regions. To improve our ability to identify long-term changes in the pattern of regional deviance we have used the standardized residuals from our regressions of modernization variables on associability. These residuals provide us with convenient and comparable measures of the

Art of Association — 179 TABLE 5-6 Overassociability and Underassociability by Region (Total Associations) (Regional Means of Standardized Residuals from Regression of Development on Associability) 1945

1950

7955

1960

1965

7970

.212 .337

.475 .510

.102 .112

.019 .195

-.301 .309

-.044 .067

Mediterranean North Central South Central Black Sea

-.775 .157 -.756 .432

-.522 .110 -.691 .329

-.121 .183 .047 .028

.035 .032 1.005 .227

-.358 .166 1.136 .183

-.357 .170 .662 .333

East Central North East South East

-.309 .005 .153

-.660 -.275 .417

-.478 -.405 -.002

-.615 -.550 -.129

-.581 -.364 -.074

-.760 -.340 -.193

Marmara Aegean

degree to which association formation in each province exceeds or falls below the level that would be predicted from its degree of development. They are, in effect, measures of overassociability and underassociability controlling for the impact of population size and socioeconomic development. Table 5-6 presents the regional means for these residuals at each five-year interval. The pattern of residuals indicates clear long-term changes in the relative propensity for association formation in the Marmara, Aegean, and South Central regions. Marmara and the Aegean, the two most developed regions, begin as moderate overassociaters and strengthen that position through 1950; after 1950 their levels of total associability become roughly commensurate with their levels of development. For these regions modernization variables have become increasingly accurate predictors of associability. The most dramatic contrasting change is the rapid jump by the South Central provinces from being strong underassociaters until 1950 to being the unchallenged leaders in associability during the 1960s. Throughout the entire twenty-five-year period the South Central area demonstrates the most persistent regional disparity between association formation and development. Associability lags far behind development until 1950, quickly catches up by 1955, and then surges ahead during the next fifteen years. Thus the strongest direction of the geographic diffusion of the art of association has been from the most developed regions near the European border and along the Aegean coastline toward the central Anatolian hin-

180 — Network of Politics

terland, which occupies an intermediate position on the scale of socioeconomic development.29 It is important to examine separately the regional distribution of each associational category; therefore, a series of six regressions was run for every type of voluntary association to match the series for total associations. Table 5-7 shows the regional means of the standardized residuals for class associations, and table 5-8 presents the same data for community associations. Together they provide us with a broad view of each region's distinctive pattern of associational life and allow us to observe long-term changes in regional propensities for various types of associability. First of all, it is obvious that the meteoric rise of the South Central provinces to become the nation's most conspicuous overassociaters is entirely the result of advances in only three types of community association— religious, cultural, and beautification groups. In contrast to its clear predominance in these specific areas, the performance of the South Central region has been consistently mediocre and occasionally quite dismal in all other varieties of association formation. Among the occupational sectors, labor, civil servants, and professionals have been particularly underorganized and their residuals in the South Central region are commonly some of the most negative in the nation. Only by 1970 did agricultural producers show even the slightest tendency toward overassociation—certainly an unimpressive showing for those who cultivate the heart of Turkey's wheat belt. Thus, the well-known religious and social conservatism of Konya and its closest neighbors seems to be reflected in two different kinds of deviance regarding association formation. Though no other region has experienced a comparable upsurge of religious associability, no other region has been as persistently hostile and resistant to class-based collective action, especially among wage-earning and salaried groups. The Black Sea region resembles the South Central area in its pattern of community associations, but shows considerable variation in the performance of its class associations. After 1950 the Black Sea region's propensity to form religious associations is generally second only to that of the South Central provinces, while its propensity to establish beautification groups 29 The remaining six regions show greater continuity in maintaining their relative status as overassociaters or underassociaters. The North Central and South East provinces display the greatest overall congruence between development and association formation. The Black Sea appears to be a region of consistent moderate overassociability. By 1970 the Black Sea was second only to the South Central region in its propensity for collective action. Three other areas—the moderately developed Mediterranean and the still underdeveloped East Central and North East—have been the most persistent underassociaters. In each of these regions increases in associability have lagged far behind increases in other dimensions of social change, but no area has remained so resistant to collective action as the East Central provinces.

Marmara Aegean Mediterranean North Central South Central Black Sea East Central NorthEast SouthEast

Marmara Aegean Mediterranean North Central South Central Black Sea East Central NorthEast SouthEast

'50

'55

'60

'65

'70

'50

'55

'60

'65

'70

-.09-.17 .03-.06-.21 .11 .23 .11 .03 .33 .18 .29 - . 6 0 .00 .38 .57 .44 .36 - . 0 3 .01 .54 .38 .30 .15 -.21-.44-.18-.09 .00-.26 .52 .16-.60-.47-.34-.18 -.27-.40 .19 .00-.07-.34 .43 .11 - . 3 2 - . 2 8 - . 0 9 - . 3 8 -.19 -.12 -.42 -.46 -.21 -.29

'45

Esnaf

.34 .12 .07 .00 .00 .02 - . 14 - .35 - .25 - .29 - .40 - .49 - . 3 0 .11 .51 .57 .17 - . 2 3 -.12 .07-.12 .05 . 1 4 - . 2 2 .03-.37-.10-.44-.33-.27 .24 .00 - .05 .03 .27 .45 - . 1 1 .18 . 0 2 - . 0 2 .08 .45 .15-.07-.09-.01 .03-.08 - . 0 2 .14 .05 .04-.11-.17

'45

Labor

-.06 .07 .01 -.03 -.08 .21 -.28 .14 -.02

-.08 .09 .56-.16 -.25 .38 -.15 -.10 -.18 -.07 -.04 .07 -.08 -.20 .00 .10 .05 .00

-.12 .47 -.06 -.28 -.16 .41 -.25 .00 -.10

'70

'50 .06 -.26 .79 .15 -.48 .16 -.47 -.42 .16

•45 .05 .05 .07 -.06 -.17 .34 -.25 -.18 -.03

.32 .26 .54 -.25 .14 -.21 -.38 -.32 .05

'55 .32 .42 .41 -.29 -.17 -.27 -.12 -.28 .00

'60

Agricultural

'65

-60

'55

Employer

-.32 1.21 .02 -.40 -.17 -.27 .06 -.19 -.07

-.07 1.05 .04 -.70 .30 -.14 .00 -.18 -.08

'70

.07 -.10 -.49 .15 -.12 .06 -.15 .25 -.05

-.02 -.08 -.30 .17 -.25 .04 -.03 .25 .03

'65

'50

'45

'50

.15 .08 -.70 .32 -.39 -.19 -.16 .49 .09

'65 .26 -.41 -.40 -.05 -.33 -.23 -.19 .40 .14

'70

'60

'65

'70 -.30 -.41 -.37 -.41 .13 - . 0 4 - . 1 5 .00 -.40 -.26 -.34 -.28 .23 .27 .46 .13 .20-.08-.49-.59 .31 .38 .09 .53 - . 1 8 - . 0 9 .26 .25 -.22 -.08 -.14 -.27 - . 0 5 .00 .09 - . 1 2

'55

Civil Servant

.33 -.30 -.42 .14 -.16 .16 -.30 .11 .23

'60

-.20 -.38 .01 .53 -.38 -.16 .27 - . 0 7 -.38-.24 .21 .20 .20 .08 -.09 -.45 -.08 .04

'45

.11 -.11 -.33 .08 -.16 .17 -.23 .26 .08

'55

Professional

TABLE 5-7 Overassociability and Underassociability by Region (Class Associations) (Regional Means of Standardized Residuals from Regression of Development on Associability)

Marmara Aegean Mediterranean North Central South Central Black Sea East Central NorthEast SouthEast

Marmara Aegean Mediterranean North Central SouthCentral Black Sea East Central NorthEast SouthEast

·50

·55

'60

'65

'70

'45

'50

'55

'50

'55

'60

'65

'70

.09 .28 .50 .18 .33 .55 .67 .50 .56 .28 .50 .66 .00 - . 0 2 - . 1 0 .36 .02 - . 0 9 - . 17 - .30 - .03 - .24 - .32 - .52 - .62 - .46 - .62 - .34 - .27 - .49 -.05 .10-.09-.17-.32-.29 -.29-.40-.35-.35-.17-.40 -.42-.40-.47-.12-.29-.43 .40 .31 .38 .30 .35 .28

'45

Sport SO

.48 .07 -.17 -.08 -.56 -.39 .04 .39 -.35 -.40 .22 - . 1 3 -.09-.09 .17 - . 1 2 .02 .26

'45 .02 .28 -.35 .14 -.13 -.07 -.20 -.16 .15

'55 - .27 .50 .09 .03 -.23 -.06 -.26 -.25 .13

'60

Social Aid

-.17 .14 -.12 .12 .49 -.05 -.25 -.26 .09

'60

Cultural

- . 4 0 .21 .12 .70 - . 0 9 - . 1 2 .01 .04 - . 2 4 - . 4 2 .46 .02 - . 1 3 - . 1 8 - . 2 4 - . 1 8 - . 0 2 .08 .08 - . 2 2 - . 3 2 - . 2 5 - . 4 5 - . 3 6 - . 6 2 - . 6 1 - . 1 1 .47-.25-.27-.05 .11 .34 .23 .19 .30 -.32-.35 .35 1.31 1.31 1.14 .03 - . 1 9 .07 .15 .25 .36 .41 .44 .31 .18 .08 - . 0 8 - .20 - .44 - .38 - .42 - .42 - .64 .13 - . 1 4 - . 3 5 -.16 . 0 0 - . 1 9 - . 3 7 - . 1 4 .14 .09 - . 0 3 .05 - . 1 1 .08 . 0 0 - . 1 4 - . 2 0 - . 3 5 -.05 .28 .10

•45

Religious

'70

-.04 -.29 -.42 .35 .91 -.08 -.38 -.14 .04

'70

- .47 - .59 .70 .76 .23 .32 .01 .20 -.28 -.29 -.15 -.21 -.31 -.71 -.17 -.30 .11 .23

'65

-.24 .11 -.13 .21 .56 -.14 -.30 -.21 .15

'65 .28 .18 -.21 -.35 -.30 .23 -.02 .75 -.32

'45 .15 .06 -.24 -.31 .07 .71 -.43 .10 -.10

'55

'65

.14 - . 2 0 .10 .20 -.40 -.37 -.22 -.07 .35 .92 .71 .43 -.44 -.39 .06-.18 -.22 -.13

'60

Beautification

.13 .39 -.19 -.52 -.37 .59 -.26 .49 -.22

'50

TABLE 5-8 Overassociability and Underassociability by Region (Community Associations) (Regional Means of Standardized Residuals from Regression of Development on Associability)

.00 -.01 -.31 .09 .34 .77 -.68 -.22 -.15

'70

Art of Association — 183

is commonly the highest in the nation. Whereas increasing development in the South Central region has had little impact on that area's relatively sluggish class associations, in the Black Sea area there have been unusually strong fluctuations in occupational associability, which seem to reflect fundamental structural changes in the local economy. As agricultural production in these provinces became increasingly dominated by cash crops of tobacco, hazel nuts, and tea there has been a growing tendency for landholdings to break into small marginal units, which are highly vulnerable to price fluctuations and often indebted to purchasing agents and local merchants. The expansion of state and private industry in Zonguldak and Samsun has placed similar pressure on the esnaf, resulting in a broad outmigration of small, independent producers seeking alternative employment in the major cities or as part of the huge European migrant work force. The impact of these economic changes is observable in the gradual displacement of declining esnaf and agricultural groups by labor, employer, and civil servant associations.30 In no other region has the changing pattern of class associability so clearly coupled the deterioration of small-scale producers with the advancement of large-scale industry and government employment. Our consistently underassociating regions also show interesting tendencies to promote certain types of association and neglect others. The East Central provinces are a particularly striking contrast to the South Central region. While the South Central area has generally resisted class-based associations, the East Central region has remained even more persistently immune to all types of community associability. No other region has such strong and uniformaly negative residuals among community associations. Conversely, the occupational sectors that have shown the clearest recent evidence of overassociability in the East Central region—labor and civil servants—are the same sectors that have remained the most retarded in the South Central provinces.31 30 Associability among small, independent producers has fallen most in Zonguldak, Samsun, Trabzon, and Rize. 31 There are two unusual features of the ecology of the East Central region that appear to explain this combination of a peculiar aversion to community associations and relatively strong organization among workers and public employees. First, the region includes several provinces that have high concentrations of religious and ethnic minorities. Sivas, Tokat, and Amasya contain the highest proportion of Alevi Shiites. If Konya is the recognized center of Sunni proselytizing, these three provinces are the preserve of Alevi heterodoxy. Similarly, Elazig, Malatya, Tunceli, and Adiyaman contain the westernmost distribution of Turkey's large Kurdish-speaking population. Low levels of community associability in these provinces do not necessarily indicate less communal solidarity among these minorities than exists for the dominant Turkish-speaking, Sunni majority. But the expression of that solidarity is far more likely to be informal rather than organized in groups that are required to register officially

184 — Network of Politics

The distribution of associability in the Mediterranean region provides an even sharper contrast to the South Central pattern. Here, too, we find an unusually strong and enduring aversion to religious, cultural, and beautification groups.32 Regarding class-based organizations, the Mediterranean shows an even stronger and longer tendency toward overassociability than the East Central region. Early organization of the industrial work force in Adana, Mersin, and iskenderun made the Mediterranean's labor movement one of the most active in the country since the 1950s.33 Employers show no comparable regional tendency of overassociability, but since 1960 the residuals for Adana have been the highest of any province in the country. In that province both labor and private management have been unusually active in association formation. Interestingly, the Mediterranean's concentration of industrial investment has not seriously weakened its highly diversified small-scale industries. In contrast to their counterparts in the Black Sea area, the craftsmen of Gaziantep, Hatay, and Adana have developed one of the nation's largest networks of local esnaf associations. Again in contrast to the Black Sea area, increasing production for the market and greater mechanization have reinforced the traditionally high concentration of landholding. Although the Mediterranean appears as the area of most intense agricultural association throughout the 1950s, this is almost entirely a reflection of the activity of the large cotton cultivators in the Cukurova region and the citrus growers of Antalya. Since 1960, large cultivators from Adana have been particularly important in the leadership of new national organizations whose with the Ministry of Interior. The evidence suggests that, in practice, the continuing legal prohibition of associations based on religious, linguistic, and cultural differences strongly discriminates against minority populations. Thus, whatever its stated intent, the Law of Associations has helped to preserve and perhaps reinforce religious and cultural differences. The second unusual aspect of the East Central ecology concerns its economic structure. In spite of the region's relatively low level of rural development, selected urban centers such as Sivas, Malatya, and Elazig have received large influxes of state industrial investment. Actually, this peculiar pattern of economic development is indirectly related to the region's religious and ethnic composition, since much of the state investment was a reward for longstanding support from local minorities for the secularist policies of the People's Party. In any event, the predominance of the public sector in local industry together with the benevolent paternalism of the State Economic Enterprises seems to account for the recent rise in labor and civil servant organization, while private employers remain relatively underrepresented. 32 Again, as in the East Central area, it appears that a sizable local minority—in this case the Arabic-speaking population close to the Syrian border—has a major impact on the nature of community associability. 33 In the early 1960s the largest of these so-called southern unions strongly criticized the moderate policies of the national confederation and tried unsuccessfully to establish a rival organization at the national level. The area has become a stronghold of independent and militant left-wing unionism, which has been closely courted by the People's Party.

Art of Association — 185

primary goal is to resist proposals for land reform. The clearest gap in the Mediterranean's otherwise impressive pattern of class associability appears among professionals and civil servants. While occupational groups in the manufacturing and agricultural sectors have shown high propensities for engaging in collective action, those in the tertiary sector have been consistently underorganized. Thus, the overall pattern of class associability in the Mediterranean shows strong organization among the traditionally important esna/and agricultural sectors, very weak modern middle sectors, and what appears to be an increasing polarization between big labor and big capital. In the Aegean we observe another distinctive pattern of associability. The Aegean consistently has been the national leader in sports associations and, since 1955, in social aid groups as well. Collective action in local community improvement projects is at least as intense in the Aegean as in the most overassociating regions, but it is relatively unrelated to religious and ethnic solidarities. What makes the area most distinctive is the unique organizational strength of its provincial middle class. The Aegean is the only region in which craftsmen have been consistently well organized throughout the entire period. The sharp drop in esnaf associability in Izmir since 1950 has been more than offset by gains in the Aegean hinterland, especially in Balikesir, Aydin, Isparta, and Burdur. The region's highly diversified agriculture supported a dense network of agricultural groups in Izmir and its neighboring provinces during the 1950s. After 1960 these groups rapidly spread through the rest of the region and the national focus of agricultural associability dramatically shifted from the Mediterranean to the Aegean. By 1970 only the still isolated province of Mugla and relatively infertile Isparta remained immune from this explosion of Aegean agricultural organization.34 In contrast to this unusual degree of organizational activity among its 34

New organization was most concentrated in provinces where small holdings are more common than large holdings, where cultivation centers on labor intensive tobacco and fruits rather than on capital intensive cotton, and where the cultivation of opium has been threatened recently by stricter government regulation. This impressive second wave of Aegean agricultural organization may, therefore, be principally the response of small growers to the new precariousness of their traditional crops in the marketplace caused by the sharp drop in international demand for Turkish tobacco and the steady restriction of the opium poppy. This suspicion is strengthened by growing criticism from small tobacco cultivators of the semiofficial Chambers of Agriculture and Commodities Exchanges, which traditionally have been dominated by the combined interests of large growers and Izmir exporters. These latter groups have been demanding lower production levels and stricter quality control to liquidate the government's huge tobacco stockpile. They also oppose longstanding government pricing and purchasing policies that support what they regard as inefficient and uncompetitive producers.

186 — Network of Politics

provincial middle class of small producers in manufacturing and agriculture, the Aegean displays a marked imbalance in the relative organization of labor and big business. The Aegean provinces have traditionally rivaled and recently surpassed the South Central region in resisting the development of a local labor movement. The relative retardation of labor organization has been a consistent and general regional phenomenon, but it is most pronounced in Izmir. Ever since 1950 Izmir has had a greater gap between its level of development and its level of union formation than any other province in the nation.35 Industrial employers, on the other hand, are well organized in Izmir. Since 1960 they have joined managers from Adana and Ankara in filling out a national network of employers' associations that were formerly concentrated in Istanbul. In fact, the young leader of Izmir industry has recently emerged as a strong rival to the traditionally dominant Istanbul employers for the role of big industry's most prominent national spokesman.36 This analysis of residuals has given us a fairly detailed view of changing patterns of association formation both within and between regions. It has allowed us to observe important variations in the distribution of several types of associability across regions and to identify shifts in the relative importance of different associations within each region over time. It is useful here to attempt a descriptive summary of these complex and changing trends in association formation for the nation as a whole.37 In order to 35 In virtually every other province where 10 percent or more of the economically active population is employed in the industrial sector, the gap is in favor of unionization. 36 In the remaining four regions certain types of association occasionally have high or low residuals, but there is no evidence of overall regional patterns of associability as there is in the five most deviant regions discussed above. We can briefly summarize these secondary deviations in regional associability as follows. The border provinces of both the most developed Marmara and the least developed South East prefer sport groups to religious and cultural associations. The high concentration of professionals in Istanbul and civil servants in Ankara accounts for occasionally high residuals for those groups in the Marmara and North Central regions. The persistence of traditionally strong pockets of esnaf organization in such provinces as Cankin, Usak, and Kiitahya boosts residuals for the North Central area, while esnaf groups in the North East and South East have gradually deteriorated. Agriculture in the less prosperous North Central section of the wheat belt is even more underorganized than in its wealthier South Central extension. And finally, the peculiar overrepresentation of professionals in the remote North East suggests the development of a bloated local service sector as other occupational groups remain relatively underorganized. 37 Thus far provinces have been grouped arbitrarily according to the conventional ecological classification of the nine agricultural regions rather than according to their more relevant propensities to foster similar types of associational life. We have not yet determined how many separate dimensions of association formation we are dealing with. We have conceptualized associability either in overaggregated terms of total association formation or in highly fragmented terms of eleven independent dimensions. If we can improve our conceptualization of the varieties of associability, we can also improve our understanding of regionalism.

Art of Association — 187

sharpen our understanding of the interrelation between various types of voluntary association and to check our intuitive classifications we have used the data-reducing capabilities of factor analysis. Factor analyses were performed for each five-year interval from 1946 to 1965 using the eleven types of association as variables and the total sample of sixty-seven provinces as the case base (table 5-9).38 By 1965, we find that nine of our eleven types of association are clearly grouped into three independent dimensions of associability. The first dimension is composed of community improvement associations reflecting religious, ethnic, and localist ties among the dominant Turkish-speaking, Sunni population (religious, cultural, and beautification groups). Second are the occupational groups representing those classes most related to the emergence of large-scale, modern capitalism (labor unions, employers' associations, and professional groups). The third dimension includes occupational groups representing independent, small-scale producers and one type of community association, which appears to be a common channel for their economic surplus (agricultural, esnaf, and sport associations).39 In order to examine these dimensions of associability while controlling for socioeconomic development we have used the standardized residuals derived from regressions of the State Planning Office's index of general development on each of the three sets of factor scores. In figure 5-11 we provide an overview of the geographic distribution of high and low propensities to engage in these three subtypes of the art of association by mapping all residuals with absolute values of .40 or greater.40 The distributions of residuals for all three types of associability reveal several examples of provincial clustering that permit a greater specification of regional patterns in association formation. For religious, cultural, and beautification groups overassociability is most highly concentrated in south central Anatolia along what may be called Turkey's Quran belt extending from Denizli through Konya and as far as Kayseri. Two other smaller pockets of overassociability also appear along the western and eastern shores of the Black Sea, where Cankin and Rize rival the extreme deviance 38 Table 5-9 shows all associational variables that loaded at the .400 level or higher on each of the major factors extracted after a varimax rotation. The proportion of total variance accounted for by each of the Ave sets of factors is a stable 65 to 70 percent. Within each factor set the relative importance of the factors uniformly declines with the first generally explaining about 40 percent of the total variance, the second about 20 percent, and the third about 10 percent. 39 The remaining two types of association—social aid and civil servant groups—have no strong relation to any of these dimensions. Social aid groups load moderately on the first and second factors; civil servant groups load moderately on the first and third. 40 A complete listing of the factor scores and standardized residuals for each province and region appears in Appendix A.

TABLE 5-9 Changing Dimensions of Associability, 1946-65 (Factor Loadings for Five Varimax Factor Analyses) Factor 3

Factor 2

Factor 1

Cultural Social aid Professional Labor Esnaf Sport

1946 .910 Agricultural .895 Esnaf .850 Beautification .802 Sport .570 Labor .484

Professional Cultural Social aid Labor Esnaf Sport

.894 .878 .814 .807 .555 .473

Professional Social aid Labor Cultural Employer

1955 .921 Religious .849 Beautification .768 Civil servant .767 Sport .463

Employer Professional Labor Social aid Cultural Beautification

1960 .907 Religious .906 Beautification .783 Civil servant .760 Cultural .524 .442

Beautification Religious Cultural Social aid Civil servant

.889 Employer .863 Labor .709 Professional .420 Social aid .396

1950 Religious Civil servant Beautification Agricultural Sport Esnaf

Civil servant Religious

.801 .779

.893 .843 .608 .544

Esnaf Agricultural Sport Employer

.760 .746 .553 .482

.852 .701 .692 .603

Agricultural Esnaf Sport

.888 .783 .699

.844 .816 .770 .442

Agricultural Sport Esnaf Civil servant

.833 .789 .639 .420

.708 .635 .522 .457 .421

.814 .758 .653 .604 .596 .418

1965

Art of Association — 189 Factor 1 (Religious, Cultural < Beautification)

Factor 2 (Labor, Employer ξ Professional ]

Factor 3 (Agricultural, Esnaf < Sport)

Residuals of 1.00 s.d. or more Residuals between .40 and 1.00 s.d.

=

Residuals of-I.OOs.d. or more -

Residuals between -.40 and-1.00 s.d.

FIGURE 5-11. Regional Distribution of Three Dimensions of Associability, 1965 of Konya. Resistance to this variety of associability is most pronounced in the highly developed coastal provinces of the northwest between Istanbul and Izmir and particularly among the southern and eastern border prov­ inces, which form an unbroken line extending eastward from Icel to Hakkari and northward from Hakkari to Kars. A somewhat weaker cluster of re­ ligious and cultural underassociaters appears among the East Central prov­ inces along a line running southeasterly from Coram and Amasya through Tokat and Sivas to Elazig and Diyarbakir. Labor, employer, and professional groups follow more closely the pat-

190 — Network of Politics tern of socioeconomic development, so ihat fewer provinces display extreme deviance for this dimension of associability than for the other two. Furthermore, the geographic distribution of deviance for these groups is somewhat less clearly regional in nature. Overassociability is most concentrated in only four uncontiguous provinces—Zonguldak, Ankara, Adana, and Diyarbakir—and their neighbors. The deviance of these four areas primarily reflects intense conflict and rivalry between local labor leaders, while the overassociability of their neighbors is more commonly the result of disproportionately large service sectors. Overassociability also appears, though on a more modest level, in Istanbul, Samsun, and Kirsehir, where there are large industrial work forces, but where the labor movement is less fragmented. The distribution of underassociability, however, does reveal strong evidence of regionalism in the Aegean hinterland and the west central area. Provinces lying within the triangle formed by Izmir, Eskisehir, and Burdur display a strong general resistance to occupational organization among groups engaged in modern capitalist production, and especially to labor unions. The distribution of residuals for agricultural, esnaf, and sport associations reveals the clearest and simplest patterns of regionalism. Overassociability is very highly concentrated along the southern coast of the Sea of Marmara and in the Aegean area. Underassociability stretches across the entire length of the Black Sea area from Istanbul in the west to Rize in the east and straight through central Anatolia from Zonguldak in the north to icel in the south. The only breaks in these otherwise continuous lines of provincial middle-class weakness appear in Kastamonu and Nigde, where traditionally strong esnaf organizations have persisted. A secondary cluster of underassociating provinces lies within the eastern triangle formed by Malatya, Siirt, and Erzurum. Clear differences also appear in the social and economic bases of the three types of associability. The second dimension—labor, employer, and professional groups—has especially high rates of covariance with urbanization (r = .80) and industrialization (r = .76). Mass media exposure trails these with a coefficient of .66 and literacy is a distant fourth at .44. In contrast, the factor scores for the other two types of associability have no significant relationship to either urbanization or industrialization. Instead, the first and third dimensions of associability have moderate correlations with literacy (r's = .46 and .49) and slightly weaker relationships with mass media exposure (r's = .28 and .39). Since the three dimensions of associability were extracted after an orthogonal rotation, their factor scores are, of course, statistically independent. But when socioeconomic development is controlled and the re-

Art of Association — 191

suiting residuals are examined, there is a significant inverse relation between the residuals from factor two and the residuals from both other factors (r's = - .23 and - .26). Thus the propensity to establish occupational associations in the expanding modern capitalist sector is inversely related to the propensity to form associations based on religious, ethnic, and localist ties as well as the propensity to form other occupational groups representing the provincial middle class of small-scale producers. Dimensions of associability that are related to different aspects of modernization are also inversely related to one another. These findings correspond with and supplement Alan Dubetsky's observations concerning the inverse relation between primordial ties and class associability in small- and medium-sized Istanbul factories. Dubetsky describes the manner in which factory owners commonly rely on ascriptive criteria (especially regional origin, kinship, and religious sectarianism) in recruiting industrial workers. He emphasizes that such recruitment practices help employers maintain work discipline and control over the production process. He argues that the major consequences of these practices are to structure factory organization in personalized terms of worker loyalty to patriarchal employers, divide the industrial work force along nonfunctional lines, and inhibit collective action through unionization. While Dubetsky regards ascriptive recruitment as at least temporarily effective in isolating less skilled workers from the contamination of the labor movement, he finds that it is impractical in more technologically advanced work sites, where employers often cannot afford to train inexperienced workers and are unable to rely on primordial ties as mechanisms of control and isolation.41 Our own observations suggest that the findings of such industrial anthropology may be extended to nationwide trends in association formation and illustrate the manner in which associability based on primordial ties can inhibit class-based associability in general and unionization in particular. The inverse relationship between modern class associations and provincial middle-class groups probably reflects the dissolution of provincial middle-class strength in the face of expanding large-scale capitalist production. This certainly seems to be the case along the Black Sea and in parts of the Marmara and North Central regions, where esnaf and agricultural associability have been deteriorating as industrialization proceeds. On the other hand, it may also be indicative of the ability of a strong provincial middle class to retard the emergence of modern class organi41

Alan Dubetsky, "Kinship, Primordial Ties, and Factory Organization in Turkey: An Anthropological View," International Journal of Middle East Studies 7 (July 1976):433451.

192 — Network of Politics

zations through their influence over regional industrialization and their local political hegemony. This is particularly likely in the Aegean, where agricultural and esnaf interests have been relatively successful in utilizing associational ties with the ruling parties to obtain high price supports and investment credits and at least temporarily avoid a serious deterioration in their position. Dennis Williams's recent study of Aegean party organization clearly documents the continuing influence of these groups in the Justice Party in Izmir and the Aegean hinterland.42 This persistent political strength of the Aegean provincial middle class may account for that region's unusual resistance to modern class associability, especially labor organization, in spite of its increasing industrialization. In the Mediterranean the continuing strength of large agriculturalists has more commonly been exercised in directly controlling industrialization rather than trying to contain it. The close interconnection between local landed, industrial, and financial interests may be responsible for the Mediterranean's unusual combination of radical labor unions with weak professional and civil servant groups. Finally, we should note the general absence of any relationship between organizations of the provincial middle class and groups based on primordial solidarities, especially religious groups. In his study of religious associations Yiicekok argues that a key factor in their extraordinary proliferation has been the increasing economic insecurity of the esnaf, which finds an outlet in religious reaction.43 Our own analysis suggests that this explanation may be appropriate for the Black Sea region and certain of the North Central provinces, but there is no convincing evidence that it is a nationwide phenomenon.44 In conclusion, then, it appears that the emergence of class-based associations among workers, employers, and professionals is generally inhibited by the organization of primordial solidarities based on religion, ethnicity, and localism as well as by the continued organizational strength and regional political hegemony of small-scale producers making up the provincial middle class. We have seen previously that these three dimensions of associability are related to different aspects of modernization and have experienced different patterns of diffusion. We shall now attempt to 42

Williams, pt. 2. Yiicekok, Orgiitlenmis, pp. 175-192. 44 In the Black Sea area the expansion of industry and commercialized agriculture has been accompanied by a clear decline in esnaf and agricultural associability, and this has, indeed, been followed by a sharp increase in religious and beautification groups. On the other hand, in the South Central region, where religious associability is strongest, all types of class-based associations, including esnaf ana agricultural groups, have always been weak. Conversely, in the Mediterranean and Aegean, where esnaf and agricultural organizations have been consistently strong, religious groups have been quite weak and the predominant form of community associability has been sport and social aid groups. 43

Art of Association — 193 establish whether they also have differing relationships with support for the two major political parties. ASSOCIABILITY AND MAJOR PARTY VOTE

Ever since the liberalization of the Law of Associations in 1946 both major parties have attempted to cultivate informal alliances with all types of voluntary association. One of the major aspects of Turkey's "debilitating pluralism" has been the encouragement of a multitude of competing, fragmented, and financially feeble groups whose range of officially permissible activities is narrowly circumscribed while political parties are left free to penetrate and colonize them. The vague legal ban on associational political activity is commonly violated, most often by the party in power itself, and enforced only sporadically and selectively when the ruling party feels that a particular associational leadership has become too close to the political opposition. We have already observed that the growth trends and diffusion patterns of several types of association have been seriously influenced by the impact of electoral campaigns, changes in regime, and shifts in associational legislation. Here we shall explore more directly the connection between association formation and party competition by examining the changing relationships between electoral support for the two major parties and various types of associability. Such an analysis can provide indirect evidence of the relative success of these parties in mobilizing the support of newly emergent groups and also suggest the role of voluntary organizations in the recent realignment of the national party system. It is helpful to begin by comparing the relationships between party vote and our indexes of modernization. Table 5-10 presents the correlates of electoral support for the two major parties in the seven national parliamentary elections between 1950 and 1973. These data clearly confirm the principal findings of several other Turkish electoral studies—the consistent and nearly uniform superiority of the Democratic Party and its Justice Party successor over the People's Party in terms of the relative relations of their electoral support to all aspects of modernization. In spite of the People's Party's longstanding self-image as the vanguard of Turkish modernism and the main bulwark against "traditionalists" and "reactionaries," it is the Democratic and Justice Party coalitions that have been most successful in politically mobilizing those areas that have experienced the greatest degree of social and economic change in the last thirty years. Nevertheless, there have been some substantial changes in the socioeconomic bases of party support since I960. While Democratic Party vote had moderate relationships with all modernization indexes, Justice Party

194 — Network of Politics TABLE 5-10 Correlates of Major Party Vote, 1950-73 DPIJP

RPP

.23 .16 .15 .33 .29 .20 .15

-.09

Mass Media 1950 1954 1957 1961 1965 1969 1973

.21 .21 .23 .32 .52 .64 .58

-.09 -.02 -.05 -.03 -.07

Associability 1950 1954 1957 1961 1965 1969 1973

.30 .29 .20 .54 .69 .71 .55

.00 .04

Urbanization 1950 1954 1957 1961 1965 1969 1973

DPIJP

RPP

1950 1954 1957 1961 1965 1969 1973

.21 .23 .56 .44 .45 .31

.01 - .05 .06 .06 .30 .23

Literacy 1950 1954 1957 1961 1965 1969 1973

.53 .17 .23 .52 .66 .62 .47

.07 .04 -.09 .00 - .02 .34 .37

Industrialization .02 .07 .11 .03 .29 .28

.34 .36

-.15 -.05 -.17

.24 .14

support has had particularly strong relations with literacy, media exposure, and associability. The extensive social and political mobilization of the provincial and rural population initiated by the Democrats thus appears to have become even more central to the support of its successor party in the 1960s and 1970s. On the other hand, since 1969 the People's Party has considerably improved its relationship to all aspects of modernization, including associability, and now leads the Justice Party in its relationship with urbanization. This suggests that the parties are becoming increasingly competitive in the most developed regions of the country. Of course, this greater competitiveness since 1969 coincides with the fragmentation of the Justice Party coalition and the gradual consolidation within the People's

Art of Association — 195

Party of new leadership committed to left-of-center policies.45 Finally, it is interesting to note that although total associability has gradually become more independent of all other aspects of modernization, since 1960 it has become substantially more related to support for both major parties. As socioeconomic variables decline in their ability to predict associability, political variables are improving. The strengthening relationship between associability and major party vote suggests that associations may be playing an important role in political realignment as both parties gradually revise their electoral strategies from broadly based, aggregative, mass party appeals to more overt attempts to represent specific and conflicting socioeconomic interests. In order to determine which specific types of association have been most related to which parties we have correlated major party vote with each associational category for all seven national elections. Table 5-11 shows two sets of results—one in which socioeconomic development was controlled and one in which it was not. When development is not controlled, the Democratic and Justice parties solidly and uniformly dominate the People's Party in the relative strength of their relationships with all types of associability in virtually every election. The only exceptions appear in the 1969 and 1973 elections, when the People's Party's improved performance in the largest urban areas is reflected in their relatively stronger association with labor, employer, and professional groups.46 When development is controlled, we observe interesting fluctuations in the relationship between total associability and Democratic and Justice Party vote, which seem to reflect the building and breaking of coalitions between these two parties and various associations. Associability has no independent relation to support for the Democratic Party at the time of its initial victory in the 1950 elections. The relationship first emerges during the Democrats' strongest electoral showing in 1954, when religious, beautification, sport, agricultural, and civil servant groups are all significantly correlated with the expanded Democratic majority. The expansion of the Democratic Party provincial organizations paralleled and stimulated the 45 Frank Tachau and Ergun Ozbudun, "Social Change and Electoral Behavior in Turkey: Toward a 'Critical Realignment'?" International Journal of Middle East Studies 6 (October 1975):46(M80. 46 These results reflect the persistently stronger association of the Democratic and Justice parties with the process of modernization in general, but they do not provide a clear view of the independent relationship between associability and political party support. Since both associability and major party vote are often closely related to various aspects of modernization, it is impossible to evaluate their degree of interdependence without controlling for the level of development. Only in this way can we attain some degree of confidence that correlations between association formation and party vote are not simply the spurious product of an antecedent variable.

.30 .29 .20 .54 .69 .71 .51

.21 .23 .16 .34 .49 .49 .32

.11 .12 .00 .35

Total 1950 1954 1957 1961 1965 1969 1973

Religious 1950 1954 1957 1961 1965 1969 1973

Cultural 1950 1954 1957 1961

DPIJP

-.08 .05 -.15 -.02

.03 -.02 -.16 -.15 -.23 .03 -.10

.00 .04 -.15 -.05 -.17 .24 .14

RPP

No Control

-.14 .01 -.14 .07

-.04 .19 .09 .15 .22 .26 .13

-.04 .22 .03 .22 .28 .31 .24

DPIJP

-.03 .03 -.19 -.05

.00 -.04 -.07 -.12 -.24 -.02 -.23

.04 .02 -.14 -.07 -.22 .04 -.09

RPP

Controlling for Development

TABLE 5-11 Associability and Major Party Vote, 1950-73

— .09 .10 .19 .19 .15 .11 .08 .08 .10 .14

—

Professional 1950 1954 1957 1961

.17 .08 .11 .31 .27 .08 .00

1954 1957 1961 1965 1969 1973

Employer

Labor 1950 1954 1957 1961 1965 1969 1973

DPIJP

No Control

-.11 -.08 -.01 -.15

.04 .09 -.02 -.09 .00 -.01 -.08 -.04 .00 .00

— —

-.02 -.03 .01 .08 .03 -.17 -.22

DPIJP

-.01 .03 .11 .04 .30 .22

-.08 -.06 .09 .09 .00 .35 .45

RPP

-.03 -.11 -.11 -.13

-.02 -.08 -.01 .03 .15 .05

—

-.02 -.10 .05 .03 -.02 .18 .31

RPP

Controlling for Development

.28 .29 .17 .41 .53 .54 .37

.32 .34 .30 .48 .52 .43 .45

.17 .14 .03 .34 .36 .51 .41

Sport 1950 1954 1957 1961 1965 1969 1973

Social Aid 1950 1954 1957 1961 1965 1969 1973

.46 .52 .38

Beautification 1950 1954 1957 1961 1965 1969 1973

1965 1969 1973

-.05 .04 -.10 .06 -.06 .09 .03

.06 .03 -.06 .08 .03 .35 .39

.01 -.03 -.12 -.10 -.14 .13 .05

-.06 .16 .22

-.16 .00 -.17 .03 .08 .26 .23

.03 .29 .20 .10 .08 .09 .17

.17 .24 .09 .20 .21 .28 .16

.02 .13 .07

-.06 .02 -.15 .02 -.16 -.10 -.19

.10 .00 -.05 .11 .07 .22 .20

.03 -.05 -.14 -.13 -.14 .03 -.07

-.05 -.05 -.01

Civil Servant 1950 1954 1957 1961 1965 1969 1973

Agricultural 1950 1954 1957 1961 1965 1969 1973

Esnaf 1950 1954 1957 1961 1965 1969 1973

1965 1969 1973

.26 .26 .07 .34 .41 .34 .36

.19 .25 .23 .40 .44 .48 .42

.23 .09 .02 .39 .59 .52 .49

.11 .08 .07

.06 .17 .05 .03 -.08 .25 .24

.05 .10 -.08 .06 .02 .17 .08

.00 .09 -.07 .03 -.16 .27 .15

-.11 .32 .26

.05 .22 -.02 .14 .09 .04 .14

-.05 .19 .13 .13 .19 .19 .21

.07 .02 -.10 .17 .25 .21 .27

-.11 -.11 -.07

.05 .16 .14 .07 -.07 .10 .07

.02 .08 -.03 .07 .05 .04 -.08

.05 .08 -.05 .01 -.18 .09 -.07

-.17 .09 .15

198 — Network of Politics rapid diffusion of these associations throughout the countryside and in provincial towns. Already in the mid-1950s an extensive network of local community and occupational groups appeared as part of the Democratic coalition among the peasantry and provincial middle class. But the instability of that coalition and the relative autonomy of voluntary associations from the dominant party are evident at least three years before Menderes' downfall. The narrow Democratic victory of 1957 again finds total associability unrelated to Democratic vote. Evidently the ruling party's extraordinary investment in promoting a wide variety of community and occupational organizations immediately before the election did not produce the anticipated quick return in the form of support at the polls. By 1957 Democratic support is associated only with a core of sport and agricultural groups, and on a weaker level than before. Relationships that arose in 1954 with several other types of association disappear and esnaf groups show a slight negative relation to Democratic vote. Thus, the government's further encouragement of the already rapid growth and diffusion of voluntary associations did not prevent voters in areas that benefited from that aid from withdrawing support when the economic boom of the early 1950s ebbed and the government's inflationary policies came under increasing domestic and foreign attack. After the overthrow of the Democratic Party in 1960 a similar prodess is observable with the Justice Party. During its first electoral contest in 1961 the Justice Party appears to have reestablished the relationship with total associability that the Democrats enjoyed in 1954. That relationship is strengthened during the next two elections, reaching its peak in 1969, but then it drops substantially in 1973 to about the same level as when the Justice Party first emerged and began its attempt to reconstruct the old Democratic coalition. The relationship between Justice Party support and associability does not disappear as it did for the Democrats in 1957, even though the electoral losses of the Justice Party have been much heavier than the earlier Democratic setback. Indeed, the weakest rates of covariance for the Justice Party are about the same as the strongest rates for the Democrats, so that the overall trend is toward a stronger relation between vote for the dominant party and total associability. Nevertheless, in both cases the dominant party's failure to achieve a majority coincides with its weakening relation to association formation. The specific types of voluntary organization associated with Justice Party support are slightly different from those related to Democratic vote. The clearest examples of continuity are religious, beautification, and agricultural associations. But the most distinctive aspects of Justice Party support are its solid and continuing relation to esnaf associability and its emerging

Art of Association — 199

negative relation to labor unions. Regarding the esnaf, the Justice Party appears to have succeeded where the Democrats failed. Although Democratic governments provided substantial encouragement to the expanding network of esnaf associations in the early- and mid-1950s, it was the Justice Party that took the lead in establishing the first "esnaf bureau" within its party structure to act as a liaison between party and esnaf organizations. Regarding labor unions, on the other hand, the Justice Party appears to be failing where the People's Party is beginning to succeed. Until the 1969 elections union organization had no clear relation to support for any of the major parties, but since 1969 unionization has become inversely related to Justice Party vote and more strongly related to support for the People's Party than has any other form of associability.47 The graph in figure 5-12 gives a detailed view of the relationship between Justice Party vote and associability when that relationship was at its strongest, just after the 1969 elections and just before the 1971 coup. In order to illustrate the differential relationship between Justice Party vote and various types of associability we have identified those provinces that have unusually high propensities for fostering each of the three distinct dimensions of associability extracted from our factor analysis. Provinces identified here by special symbols are the same provinces that were identified earlier as overassociaters in the maps infigure5-11. Although the overall relationship between Justice Party support and total associability is quite impressive (r = .71), Justice Party vote is positively related to only two of the three dimensions of associability and is negatively related to the other dimension. The associational bases of Justice Party support are limited to factor one, the very numerous local religious, cultural, and beautification groups (r = .24), and factor three, the agricultural, esnaf, and sport groups of the provincial middle class (r = .22). As we see from the listing below the scatterplot, the coincidence of strong Justice Party support with high propensities for these two types of association formation is clustered primarily in the provinces of central Anatolia, the Black Sea, and the Aegean. On the other hand, factor two—labor unions, 47

This unusually clear polarization of major party support in relation to union organization suggests a widening gap between the voting behavior of the rank and file and the continuing insistence of the major labor confederation's leadership that they remain "above party." Recalling again that our index of unionization is based upon the number of organizations rather than the number of members, it is very likely that these results reflect the growing alliance between the People's Party and the smaller socialist confederation (DISK) in areas where union rivalry and proliferation are most intense. Other examples of associability having opposing relations to voting for the two major parties occur among professional groups, where the People's Party has recently gained a slight advantage, and religious and social aid organizations, where the Justice Party enjoys a more considerable edge.

200 — Network of Politics

FIGURE 5-12. Associability and Justice Party Vote, 1969 Factor 1 (o): 1 Kocaeli, 2 Rize, 3 Nevsehir, 4 Cankin, 5 Konya, 6 Isparta, 7 Artvin, 8 Kayseri, 9 Trabzon, 10 Afyon, 11 Kiitahya, 12 Antalya, 13 Erzincan, 14 Giimu^hane Factor 3 ( • ) : 15 Balikesir, 16 Kirklareli, 17 Aydm, 18 Bursa, 19 U$ak, 20 Manisa, 21 Izmir Factors 1 & 3 (A): 22 Denizli, 23 Bilecik, 24 Edime, 25 Kastamonu, 26 Nigde, 27 Burdur Factors 2 & 1 ( ® ): 28 Zonguldak Factors 2