After an extended absence, the "social question" has returned to Latin America. In contrast to the early decades of this century, however, a rising and seemingly inexorable tide of popular mobilization is not evident. Strikingly, revolutionary proposals are nowhere to be found, even though democratization, guided by an ascendant neoliberalism, thus far has dashed the widespread expectations for social justice that it raised. Occasional, sometimes large-scale street demonstrations against particular anti-popular reforms or corrupt public officials continue to dot the region's political landscape, to be sure, as popular frustration bubbles up spasmodically. Yet, there are few signs of a sustained, broad-based popular movement against neoliberalism, and much less around an alternative project. Guerrilla movements exist in a few countries, but they are either politically isolated or, like the Zapatistas in Mexico, they have used arms not to seize power but to create new opportunities for peaceful political change. Most disturbing of all, as Carlos Vilas so eloquently details in the introduction, violence is rife in many parts of the region and large numbers of its poorest citizens--whether out of hope or fear--recently have cast their electoral lot with neoliberal reformers.

        Despite the dramatic absence of unified popular political "subjects," social concerns have crept back onto the Latin American political agenda. "Fighting poverty"--if not inequality--has become good politics even for neoliberal reformers. International development institutions have highlighted basic needs; state financial managers need to prevent popular outbursts to avoid scaring off volatile in-flows of foreign capital; and old-fashioned political expediency has its role.

        If one views popular politics through an interpretive lens that looks beyond electoral contests and mass movements (or their absence), moreover, encouraging recent signs of dynamism are ample. Many of the studies in this book point to new activity--the transformation of corporatist structures by innovative unionists in Brazil and Argentina; the reconfiguration of ties between left parties and popular movements in Venezuela and Mexico; experiments in participatory local governance under left parties in Brazil and Uruguay; and new forms of representation in social policy making in Ecuador and Mexico. These examples reflect a trend toward a reconstruction of popular participation--redefining the identities and issues around which the popular sectors organize--and a reconfiguration of popular representation--reordering the structures through which they gain a voice in centers of political power and through which others seek to shape and control them². Activists and poor people continue to "imagine" new causes and identities--like indigenous rights or environmental protection--, organize fitfully around them, and engage with public authorities in novel and proactive ways. Such developments attest to the continued capacity for re-invention of popular politics, even in hostile environments.

        In this concluding essay, we seek to both build on and look beyond the studies in this volume to sketch the outlines of an emerging model of representation. Our analytical point of departure is the notion that the ongoing reorganization and reinsertion of the popular sectors into the political process has not followed the classic model of popular incorporation in the post-Depression era. Nor does it conform to the various versions of the class-based model that popular and left leaders long espoused. Nor do conventional visions of "modern societies" such as the liberal-pluralist framework capture or explain these new developments. Rather, we detect important signs of movement--incipient and partial, but real--towards new and substantially different structures of popular representation that we will characterize as "associative networks".

        In order to explain how some of the developments described in this book point to a nascent type of popular representation, we first discuss the "old" types through a schematic periodization. The first is the period of "popular-sector incorporation," lasting roughly from the 1930s to the 1960s, which was marked by the political "coming of age" of the popular sectors and their decisive incorporation into national polities. The conditions and outcomes varied from country to country (Collier and Collier 1991) but the fact of incorporation was a constant. "Decomposition," the second period beginning roughly in the 1960s, saw a process of either gradual erosion or dramatic dismantling of popular-sector ties with the state. This is followed by the period of our main concern, spanning roughly the last decade, when a "recomposition" of ties began to occur.

        Based on an analysis of these periods and drawing upon the existing literature, we identify four ideal types of structures of representation central to the incorporation period--"clientelist," "populist," "corporatist," and "mass-mobilizational." The new trends differ from all of these in (1) being less tied to a single national structure of power and in (2) placing more emphasis on ties resulting from decisions to associate. Drawing on the empirical evidence from our cases and recent theoretical developments outside the study of Latin American politics, we propose a fifth type of structure of representation--the "associative network". Such a network links state and societal actors--sometimes including popular ones--through interpersonal, media and/or interorganizational ties. Multiple networks process and reshape contending political claims through relatively open-ended and problem-focused interactions. They are distinctive not only in the way they link people with decision-making centers, but also by their multiplicity and relatively rapid reconfiguration over time.

        Popular sectors are not the only, nor the most powerful, of the actors within many associative networks. Indeed, some networks are created without their participation. However, popular participation in them seems to be increasing and, as we explain below, there are reasons why elites draw them in.

        We attribute the rise of these more variegated and less centralized structures of representation to fundamental changes in global and domestic conditions and related shifts in the strategies of elites and popular actors. Especially crucial has been the dispersion of decision making activity away from the centralized state of the earlier era to a more polycentric state, with multiple centers of decision making. Associative networks are constructed around, and further the development of, these new centers.

        Associative networks are transforming popular representation, but only partially as well as unequally across countries, regions, and policy arenas. The evidence is not sufficient for us to make confident statements about the degree to which this ideal-typical formulation has been approximated or realized in practice. Moreover, many poor people evidently are still linked to the polity through clientelistic and state corporatist structures and--as Panfichi suggests in his essay on Peru in this volume--a new form of populism has appeared in some countries. More seriously still, the factors explored in Carlos Vilas' introductory essay have left many individuals and groups enmeshed in situations of grinding poverty and myriad forms of violence with no effective way of engaging with, or protecting themselves from, the state.

        Notwithstanding these sometimes tragic trends in the region, the evidence contained in this book leads us to conclude that there is far more movement in the direction of recomposition of popular representation through the novel forms that we propose than is conventionally perceived. Such perceptions are shared, we believe, even by those observers and participants most identified with the popular sectors, who still tend to think in terms of the recomposition of old forms, not the emergence of new ones. To be sure, substantial fragmentation and social disembeddedness continue to plague the region, but a complex evolution also is occurring that can be overlooked or dismissed as "ad hoc" only if we remain strictly within the confines of existing conceptual categories and analytical tools. Understanding these changes constitutes an important agenda for research.

        We believe the current situation facing the popular sectors is one in which, in varying ways in different countries, associative networks are slowly becoming more relevant. These networks certainly are not inherently more democratic than the earlier forms, although they have the potential to be so. The key point, however, is that unless one understands the new forms of popular representation, it will be impossible to perceive whether progress towards democratization is occurring, or to work for it.
 
 
 
 

REINTERPRETING THE HISTORICAL LEGACY OF POPULAR REPRESENTATION
 
 

        During the populist era in Latin America emerging patterns of popular representation were centered on a complex process of negotiation, accommodation, and confrontation between elites and the popular sectors. With time, subaltern strata were incorporated into national polities, economies, and societies as important actors (Collier and Collier 1991). Popular-sector political incorporation rested on two essential conditions: (1) an expanding economy led by an interventionist state following a basic doctrine of national development and (2) a relatively harmonious--albeit temporary--fit between the political goals and strategies of elite actors, on one hand, and those of popular-sector and left leaders, on the other. Although important cross-national differences in the timing and institutionalization of incorporation existed, there was a broadly similar dynamic across all of the countries considered here except Guatemala.³

        It is worth emphasizing the first of the two points above, the existence of a centralized, interventionist state. Although the state was by no means all-powerful, and significant numbers of land-holders, owners of extractive industries, and foreign companies, among others, retained significant autonomy, that autonomy was not expressed in the autonomy of branches of government, local governments or independent ministries. Decision making in the state was highly centralized in the hands of the president and his or her closest colleagues.

        Within a broader context of inward-oriented, state-promoted industrialization, new elites were led by a variety of motives to cultivate a political relationship with the popular sectors. The latter were believed to be an active or potential threat to political order, requiring measures to preempt and control. The poor were viewed as a potentially powerful social and political base for industrialization, requiring measures to mobilize and co-opt. For their part, popular leaders sought firm relationships with emerging political elites as a propitious, and perhaps the only, way to secure access to the state. Such access was deemed crucial by these leaders not just for obtaining material and organizational benefits, but also for exercising some degree of political power and fostering progressive visions of economic and social change. Underlying the latter goals was a particular set of ideological conceptions and political practices shared by a broad spectrum of popular and left activists. Insofar as state power was central to left and popular conceptions of political and economic life, political opportunism in gaining access to that state became justified and grassroots organizing was viewed largely as an instrument to gain such access. In turn, the vertical and asymmetrical nature of group-state ties tended to be replicated at the level of ties between popular leaders and "the masses."

        Four types of structures of popular representation shaped the emerging new relationship between the popular sectors and the political systems of this era--populism, corporatism, clientelism, and political parties. Populism provided the overarching framework for appeals to a broad abstract entity, el pueblo or la nación, which cut across boundaries of class, ethnicity, gender, and region(Cardoso and Faletto 1977, Coniff 1982, Weffort and Quijano 1973). While the term "populism" has many distinct and often competing usages (e.g., Roberts 1995), we refer to it as a personalistic relationship between leader and masses that tends to be expressed only partially in formal organizations.

        Populism was often accompanied by another, more formal structure of popular representation--corporatism. This concept generally refers to structured relationships tying functionally organized groups in civil society to the state, in which the latter sanctions, oversees, and regulates group formation and activities in the former( Chalmers 1985, Collier and Collier 1979, Malloy 1977, Schmitter 1974). Latin America's state corporatism entailed few, if any, effective restrictions on dominant-class organizations like business, but strong controls on subaltern-class organizations representing workers and peasants (O'Donnell 1977).

        While populism and corporatism were the central structures of popular representation, clientelism and political parties also played an important, albeit subordinate, role. Unlike corporatism, clientelism was based on direct, face-to-face encounters between individual elites and poor people characterized by little or no formal structure (Schmidt et al 1977). While clientelistic practices had deep roots in earlier periods of Latin American history, they became particularly important during the populist era in mobilizing the electoral support of the poorly organized, such as recent migrants and peasants lying outside officially sanctioned associations. Typically, extensive patron-client networks were constructed out of many clientelistic exchanges, in which the ultimate beneficiary of the "upward" flow of support was separated from individuals receiving a "downward" flow of goods and services by a series of regional and local brokers.

        During the populist-corporatist era, political parties claiming to represent popular interests emerged and in many cases flourished, particularly as elections became more open and competitive. Nonetheless, parties generally played a less central role in organizing mass political participation in Latin American polities than they did in the prototypical party systems of advanced industrial democracies. In most countries under consideration here, with the major exception of Chile, parties lacked the ideological and programmatic character and well-defined class basis of their European counterparts. The mass-mobilization model long espoused on the left, whether class or more broadly based, rarely became a reality except in the opposition. Rather, parties tended to be personalistic, pragmatic, and often opportunistic. Links with popular constituents were built on varying mixtures of populist appeals and, as noted above, corporatist and clientelistic exchanges. Given the weakness of party systems and the centralization of power in the executive in many countries, parties tended to be more of an instrument for winning elections than for exercising power and mediating popular demands on an everyday basis (Chalmers 1972).

        At first glance, the highly vertical and hierarchical character of these structures suggests an image of monolithic elite and state control over the popular sectors, much like that generally conveyed by studies of corporatism and incorporation in the region. To be sure, the potential for mass revolutionary uprisings was defused, at least for some time and in most of our focus countries. Yet, it is important to recognize that through the ties elites sought to use to encapsulate the popular sectors, the latter found new and varied points of access to the political system. That is to say, through voting and partisan mobilization, corporatist and clientelistic bargaining, and relationships with populist leaders, poor people acquired an organized political voice for the first time, however limited and imperfect.

        Though formally circumscribed, channels of popular representation under the classic "populist-corporatist" model nonetheless enabled critical segments of the popular sectors to gain some leverage over elites, particularly in the more extensively industrialized countries of the region. Underlying the growth of popular-sector political leverage was a style of popular demand-making and bargaining that we call "insider pressure politics." This style entailed the following key elements: a rhetorical emphasis on abiding loyalty to particular elites; the casting of demands in the populist language utilized and legitimized by leaders; the use of mobilization as a flexible tool alternatively to extract concessions from and express support for leaders; and, in extreme cases, threats to withdraw support or transfer it to other elites.

        What are we to make of the costs and benefits for the popular sectors of this populist-corporatist model of popular representation? In recent political discourse there is a tendency either to condemn classic popular representation for its manifest limitations or to wax nostalgic about its perceived virtues vis-à-vis the current state of affairs. We see truth in both views. In economic terms, benefits were selective, to be sure, generally favoring the organized working class and emerging middle strata over the unorganized poor, as well as urban-industrial over rural-agricultural interests. Yet, notwithstanding this selectivity and the deepening inequalities, it is undeniable that in many countries fairly broad segments of the popular sectors gained some measure of upward social mobility through transfer payments as well as improvements in literacy, nutrition, sanitation, and health care.

        In political terms, the loss of organizational autonomy and grassroots participation clearly had negative long-term consequences for the popular sectors, despite the immediate material advantages that were secured through links with the state. Nonetheless, the popular sectors acquired an important organizational presence, a crucial degree of political recognition, as well as an intangible but extremely important subjective sense--an identity--of possessing political and social rights within emerging national polities. Over time, these organizational and symbolic resources enabled popular actors increasingly to challenge and contest not only the boundaries of their incorporation, but also to some extent the elite domination of the underlying political and socio-economic systems.
 
 

The Decomposition of Historical Structures
 
 

        From roughly the 1960s through the 1980s, various pressures contributed to what we call the decomposition of traditional, populist-era structures of representation. While we recognize that some new forms emerged and attempts were made to reform and reinvigorate old patterns during this period, the four structures of popular representation mentioned in the previous section deteriorated as channels of popular political access at a much more rapid rate than they could be reformulated or replaced by new ones. Decomposition was a gradual, though by no means steady or even, process that stemmed from pressures that were both exogenous and endogenous to structures of representation.

        From the vantage point of the mid-1990s, it is tempting to assume that historic patterns of popular representation simply were overwhelmed by the major structural shifts that have transformed political, economic, and social systems around the globe--amounting to a sort of Latin American functional equivalent to the crisis of the Keynesian welfare state in advanced industrial countries (e.g., Offe 1984). To be sure, events and processes that are, strictly speaking, exogenous to traditional patterns of popular representation--such as military coups and governments followed by redemocratization, the end of the Cold War, economic globalization, structural adjustment and market reform, and the growing informal economy--did much to erode structures of popular representation. Yet, they already were experiencing considerable flux and strain before these global and national trends emerged. The source of these ongoing strains, evident as far back as the late fifties in some countries, was twofold: (1) growing intra-elite rivalry and division over the political role of the popular sectors as well as (2) rising popular-sector and left challenges to traditional institutions of representation. The context for these strains was growing centrifugal pressures generate by the region's model of development. (O'Donnell 1973, Cardoso and Faletto 1979). We contend that it was the juxtaposition of these existing strains with subsequent "structural" pressures--which served to exacerbate them--that together produced the decomposition of popular representation.

        As suggested above, increasing pressures from below gradually began to undermine nascent or established ties of popular incorporation. Popular demands for growing wages, benefits, and rights proved difficult to accommodate. Incorporated groups became increasingly adept at playing--and pushing the limits of--insider pressure politics; groups that were excluded struggled to gain comparable access; and, all the while, economic conditions turned more and more volatile. During this period, popular leaders and their constituencies began to confront more openly the restrictive dimensions of incorporation. In part, this departure in popular-sector behavior reflected strategic and ideological trends within the partisan left--namely, the decline of "collaborationist" thinking and the rise of nationalist and radical ideologies and even guerrilla strategies, particularly following the Cuban Revolution. Together, all of these developments combined to heighten popular-sector militancy and the general confrontational tone of politics.

        Popular actors were not alone in challenging the established terms of their incorporation into national polities. The growing political salience of the popular sectors, along with the emerging new socio-economic realities of state-led, dependent industrialization, also generated new dynamics among elite actors. Political and ideological divisions over the course of national development began to emerge, increasingly pitting reformers, radical nationalists, and conservatives against each other (Cohen 1994). In Brazil and the Southern Cone, where populism was most strongly rooted, an outright elite backlash against populism occurred in the form of a series of military coups between 1964 and 1976. The ensuing new authoritarian regimes closed channels of popular representation, either by dismantling popular and left organizations and the arenas in which they operated or, alternatively, by turning them into narrow instruments of coercive state control (Schamis 1991, O'Donnell 1973). Even in those countries where existing regimes weathered similar pressures (Venezuela and Mexico) or where late populist breakthroughs occurred under the auspices of reformist military regimes (Ecuador and Peru), elite counter-pressures--through coup-plotting, capital flight, and the like--narrowed the range of maneuver for distributive policies and economic nationalism. Moreover, broadly inclusionary systems, like those in Mexico and Venezuela, became more rigid, unable to accommodate smoothly the rapid expansion of social forces.

        Hence, by the mid-1970s, spaces for popular representation in our focus countries could be characterized as "forcibly closed" (under conservative military regimes), "restricted" (under inclusionary regimes turned oligarchic), or "nascent" and "fragile" (under reformist military regimes). In most countries, popular representation was manifestly weak--and in some cases moribund--and had been so for a number of years when exogenous shocks, economic restructuring and regime change began to exacerbate this process of decomposition.

        Together, economic restructuring and democratization have dominated the region's political praxis and scholarly reflection on the region since roughly the early 1980s. Yet, it is infrequently understood that the impact of these twin processes on popular representation has been decidedly contradictory in nature. In an immediate sense, the ongoing process of decomposition reached its culmination, as major elite-dominated economic and political changes left the popular sectors more distant and disengaged from political decision-making than they ever had been since the onset of incorporation. We argue below, however, that from a long-term perspective the seeds for a new structure of popular representation were sown even in the midst a terminal phase of decomposition.

        A complex array of domestic and global forces has rewoven the economic and social fabric of the region in the past quarter century. Among them are the oil "shocks" of the 1970s; the post-1982 debt crisis; bouts of megainflation and stabilization; market and state reforms; increased integration into the world economy; the globalization of production and capital flows; and changes in the geography, social structure, and models of capitalist production. During the 1980s, these trends generally had a devastating impact on the material conditions and traditional political organizations of the popular sector. In the cities, formal-sector jobs declined in number, security, and remuneration, while the informal sector mushroomed. In the countryside, wage labor in capitalist, export-oriented agriculture generally expanded, although pay levels and security remained precarious and subsistence farming declined. Unions and peasant organizations found not only their memberships and bargaining strength diminished (if not decimated), but also their traditional access to decision-making arenas concerning social services and benefits cut off or several eroded. More generally, claims on state resources by popular organizations generally fell on the deaf ears of increasingly austerity-minded governing elites.

        As economic restructuring further undermined already weak ties of popular representation, processes of democratization were having much the same overall impact. Belying general expectations, the democratic transitions or openings that occurred in most of our focus countries neither stimulated the emergence of new patterns of popular representation advocated by many--such as social pacts and "democratic class compromise"(e.g. O'Donnell and Schmitter 1986, Roxborough 1992a)--nor reinvigorated old populist-era patterns and aspirations, centered on various forms of mass mobilization feared by many. Rather, popular actors found themselves unable to forge strong links of whatever type with democratizing movements and the governments of either new democracies or entrenched political systems. Among other factors, the post-transition legacy of elite-dominated regime change, the dire nature of economic realities, and the growing predominance of neoliberal thinking in domestic and international quarters all played a role in reducing the space for popular representation.

        Yet, while efforts to rebuild popular representation in the democratic transitions were decidedly unimpressive, popular participation did show considerable signs of dynamism in some countries. Building on the lessons of earlier periods and changing left models within and outside Latin America, popular activists began to place unprecedented importance on grassroots organizing, autonomy from the state and political parties, and traditionally displaced or ignored issues and collective identities. (Eckstein 1989, Alvarez & Escobar 1992). These newfound values and concerns nurtured not only struggles for greater participation within traditional peasant and labor organizations, but also the mushrooming of a vast new array of movements and associations--neighborhood-based, church-linked, gender-oriented, human rights-oriented, indigenous, environmental, and so on. Many associations began to take the novel form of non-governmental organizations, or NGOs, analytically distinct from membership-based popular organizations, social movements, and political parties.

        However much civil society was growing independently of the state, the mid-eighties was generally a period of very weak structures of representation linking those groups with the state. The debt crisis, stabilization programs, and market reforms exacerbated this situation, severely undermining living standards and capacity for sustained and organized popular mobilization. In some cases, the decay of popular representation was so profound, and the conditions so hostile that institutions became extremely fragile and on the local level at least, order broke down.

        The view that a general political "demobilization" of popular actors seems to suffuse contemporary discourse in and about the region. Some observers believe that civil society has degenerated into "anomie" and "disarticulation" (Tironi 1986, Zermeño 1989) Traditional bases of sociability and solidarity like religion, family and community structures, and work are held to have undergone rapid decline, contributing to the destruction of capacities for collective organizing and action. A weak civil society, divided against itself, further undermines historically weak political institutions and provides fodder for the personalistic appeals of often unscrupulous politicians. However formally democratic they remain, their democratic content is drained by chief executives who run roughshod over rules and institutions, ruling by a mixture of decree, charisma, and old-fashioned patronage. Whether these political forms are labeled "exclusionary," "delegative," or "neopopulist" (Acuña 1994, O'Donnell 1994, Roberts 1995) the common image that emerges is one in which nominally democratic rule rests on the absence--and even the active destruction--of links within civil society, and between it and the state.

        We characterize such a type of political system as "socially disembedded," where remaining popular organizations are few and often politically isolated and ineffective--when not outright repressed, albeit by ostensibly legal means. The exercise of citizenship is thus reduced for the popular sectors to voting, receiving messages conveyed through the elite-controlled media, and waiting passively for the benefits of state policies to trickle down--perhaps--from on high. This is a vision of the logical completion of decomposition.

Several of our chapters--Yashar on Guatemala, Burt and Panfichi on Peru, Pinheiro on Brazil, and the introductory chapter by Vilas--highlight important manifestations of disembeddedness in contemporary Latin American polities. While we agree that these are important in some countries and present in all, other chapters highlight countervailing tendencies. Moreover, we believe that there are reasonable grounds for expecting that, as medium- to long-term issues of growth, sustainability, and equity begin to overshadow largely crisis-driven macroeconomic reforms on the region's policy agenda--and in the absence of authoritarian reversals--greater opportunities will emerge for a reembedding of politics linking state and societal actors⁴.

        We find that there is variation in the degree of decay or reconstruction between different issue areas, producing noteworthy dissimilarities between countries, or even within a single one. For instance, macroeconomic policymaking throughout the region--on issues like trade liberalization, monetary and general fiscal matters, privatization, and deregulation of foreign investment--continues to be dominated by elite actors and technocratic styles, excluding popular sectors from all but diffuse influence through the ballot box. Yet, the essays by Piester and Fox on Mexico and Segarra on Ecuador underline a broader, though uneven, trend toward greater involvement of NGOs and popular actors in the implementation, and sometimes the shaping, of social policy. While their involvement is unlikely to produce dramatic redistributive shifts that favor the popular sectors, it does provide opportunities for voicing claims and extending debates on social policy issues. Here and elsewhere contradictions are apparent. For instance, if we juxtapose growing popular access to decision-making in Brazil over sectoral industrial policy (Martin), environmental issues (Hochstetler), and small-town local governance (Nylen), with Pinheiro's sobering account of the absence of basic civil liberties and genuine rule of law for the poorest Brazilian citizens, we gain a vivid image of the kind of contradictory realities that may co-exist within a single country. Finally, if we contrast the perverse combination of social disarticulation, personalistic rule, and decay of political institutions documented by Burt and Panfichi in Peru--a case that has been very influential for the aforementioned analyses of social disembeddedness--with trends highlighted in other countries, that Andean country emerges seems more an exceptionally severe case of institutional and social decay than a "paradigm" for contemporary Latin American politics.

        In sum, then, violence, social disarticulation, and personalistic modes of leadership clearly are very present in today's Latin America, and particularly rife in some countries. These characteristics of Latin American societies and polities continue to cry out for scholarly reflection, normative critique, and practical solutions. Yet, they by no means exhaust the full range of recent and emerging trends in popular participation and representation as documented in this book. In order to perceive and understand these trends we need to look beyond the old structures of representation, however, and recognize the factors that are producing a new structure.
 

Forces Shaping the Recomposition of State-Popular Sector Links
 

       Recomposition comes about for a variety of reasons, in part the same ones for which popular sectors mobilized in earlier periods. Popular actors participate in politics in large part in order to overcome the exceptionally high rates of inequality, unemployment, and physical and economic insecurity. Political and economic elites search for ways of setting limits on popular politics because of the fear that "the dangerous classes" will disrupt stability in the face of sacrifices demanded by economic restructuring, particularly by making skittish international and domestic investors flee⁵. Yet, these usual reasons are not producing, on any large scale, the classic responses in the form of the rise or strengthening of corporatist institutions and populist parties. Although a neo-populist style does characterize some current (Fujimori and Menem) and past (Salinas and Collor) presidential leaders, these experiences of leadership seem much more media-based, election-oriented, and episodic than the institution-building breakthroughs associated with classic populism⁶.

        Nor have traditional motives for popular mobilization yielded substantial steps towards a mass movement to counteract efforts to control popular sectors, an outcome long sought--or feared--by many. Some conjunctures do suggest the possibility of a broad mobilization, but the movements they have generated have been either transitory or narrow in scope. For example, the rioting in Caracas in 1989, the mobilization of support for Cárdenas in the 1988 Mexican election, and the recent civic pro-impeachment campaigns in Brazil and Venezuela were dramatic and had a substantial social base, but they did not produce enduring mass movements, party-based or otherwise. The very broad protests by the indigenous movement in Ecuador described in Selverston's essay were an important national event, but chiefly for consolidating the political role of an indigenous organization and carving out a sphere for indigenous rights. Even in cases where the appeal of a popular movement captures national and international attention, as in the 1994 Zapatista uprising in Chiapas, Mexico, the sustained popular mobilization appears to be local, even though its democratizing effects may turn out to be national⁷.

        Despite the absence of enduring mass movements and of any large-scale reformulation of classic structures of popular representation, nonetheless recomposition is occurring and it is taking the specific form that we call associative networks. Changing conditions and shifting goals of elites, popular sectors, and other relevant actors are leading to the forging of links between groups of people and the state along new lines. Although there is no way at this time to predict with certainty--nor even measure with accuracy--the degree to which associative networks will become dominant, there is good reason to believe that they will be of growing prominence because the forces producing them are powerful. These forces can be summarized as follows:
 

• the dispersion of political decision making,

• the impact of new resources for communicating and acquiring knowledge,

• the influence of ideas advocating greater societal and popular involvement in decision making in order to increase competitiveness and the cost-effectiveness of public administration,

• the consequences of the internationalization of many policy spheres and political arenas, and

• political learning by elites and popular actors induced by the perceived failure of established strategies and institutions to respond to changing political and economic realities.

        The "centralized state" model that historically characterized Latin American polities under both populist-democratic and authoritarian regimes allowed many observers to discuss politics with reasonable accuracy as if it took place in one singular, unified "space." Set geographically in the national capital, this space was centered in the informal but widely recognized sphere dominated by the interactions (coalition-building, deal-making, intrigues, maneuvers, etc.) of the president with a particular set of powerful cabinet members, personal advisers, party officials, legislative leaders, and military officers who were directly tied to the president, together, under some circumstances, with foreign investors, bankers, ambassadors and church leaders⁹. One should not overemphasize this centralism, as there were obviously important spheres of business activity and rural enclaves where de facto public power lay in private hands. Yet, in the populist-corporatist era, the single national arena of decision-making was the place where political actors, especially from the popular sectors, directed their demands.

        If we think of the state as the set of authoritative decision-making centers of a polity backed by a legitimate monopoly of the use of coercion, along with the rules and "rule-enforcing" agents who regulate those arenas, there was, by and large, only one such center in the pyramidal structure of political authority of the centralized state. All the various structures of representation, particularly popular representation, were firmly anchored at the top of the pyramid in either direct or indirect fashion, flowing downward from it¹⁰.

        The papers in this volume, along with a growing body of other evidence, provide indications that--by default and by design--there has been in recent years a significant shift of effective decision-making activity away from the executive-centered political coterie lying at the apex of the pyramid. A variety of more or less autonomous administrative, legislative, subnational, and external sites have emerged and become the foci of political action.

        The most straightforward examples of dispersion of decision-making activity concern administrative and political decentralization to subnational units. Decentralization, in the sense of shifting authority over budgets and policy decisions to state/provincial and municipal levels, is on the reform agenda of many Latin American countries and is already a significant, if incomplete, reality in some. (López Murphy 1995, Morris and Lowder 1992; Reilly 1995). Movement in this direction flows from many wellsprings: increased political contestation at the subnational level; the reformist ideas of "good government" advocates, such as the World Bank, and neoliberals; the opportunism of national-level officials who want to rid themselves of seemingly intractable problems; and practical concerns about stimulating local economic development and reflecting the greater territorial distribution of economic activities and population that now characterizes the region. Constitutional reform has played a part in decentralization in a few countries, as in the 1988 Brazilian document that mandated (the now quite controversial) revenue-sharing provisions and the new Colombian constitution's creation of elected municipal governments.

        The subnational dispersion of decision making creates a new political reality for those interested in influencing public policy. While it is conceivable that a unified national party representing the popular sectors could act in each of these centers, the essays on innovative municipal administrations in small towns in northeastern Brazil by Nylen and in the Uruguayan national capital by Winn and Ferro-Clericó suggest the difficulties attendant to such efforts. These researchers' inquiries were initially guided by a concern for what the respective national parties could accomplish when given a chance to govern and how local "lessons" might reshape national politics. Yet two intriguing implications emerge from the two chapters: (1) greater local success comes not from implementing national party decisions and doctrines but rather from developing a creative political praxis centered on local needs and (2) local progressive governments do not necessarily translate into either success at the polls or clear guidance for party platforms at the national level, nor should they be assessed strictly in such instrumental terms.

        The relative success at the local level of these progressive experiments was predicated on the capacity of local administrations and party activists to take advantage of and further foster the autonomy of local arenas of decision-making from national control, whether exercised by central governments or by national party structures. Of course, opposition governments at the state or municipal levels are a more widespread and growing phenomenon in the region--one with, it bears emphasizing, relatively few historical precedents under the centralized state. This trend currently encompasses centrist to center-left governments in major cities and/or states in at least two additional focus countries (Argentina, and--as discussed by López-Maya in this volume--Venezuela) and center-right opposition governments at the state and municipal level scattered throughout much of Mexico (Rodríguez and Ward 1994).

        The dispersion of state decision making, however, entails not only a shift to subnational territorial units, but also the development of relatively distinct policy domains within the state, principally (but not only) at the level of the central government¹¹. In these domains, some group of officials, often reaching out to some societal actor or actors, makes policy with a significant degree of autonomy. A trend towards the establishment of independent central banks is one major example of how this may happen in a formal way (Maxfield forthcoming). Social policy (as discussed in the essays by Segarra on Ecuador and by Piester and Fox on Mexico) has also emerged as a field in which specialized segments of the state apparatus--various agencies, ministries, legislative committees, and the like--play a central, and moderately autonomous role, in shaping policy. One could easily add other policy domains, like environmental regulation (Hochstetler in this volume) or industrial promotion (Evans 1995), in which state policymakers sometimes call on the expertise, manpower, political support, or other forms of cooperation that non-state actors may offer.

        The broad phenomenon that is currently grouped under the broad label of "privatization" can sometimes be equated with this sort of dispersion. Privatization can mean many things, including simply selling off government-owned firms, which then become part of the market. Yet, even in these cases, states frequently maintain leverage and influence over newly privatized firms through a variety of policy instruments, ranging from minority ownership and close ties to the new management to regulations in areas like anti-trust, trade, and labor and environmental matters. More generally, and as this example suggests, much of what is called privatization entails, in fact, less the abdication of public authority than the dispersion of public decision making, because both major firms and "privatized" social security entities make authoritative decisions that have public imprimatur and a force similar to that of public law. Government officials often continue to be part of the decision-making process, formally or informally.

        Whether the dispersion of decision making takes place through creating autonomous teams in government ministries, establishing independent agencies, or (partial) privatization, the net impact is to stimulate political actors, including those from the popular sectors, to target those varied, and often changing, centers of decision making. Further, the relatively narrow range of policy issues being considered in any particular center suggests the importance of political strategies focused at least in part on policy debate, rather than simply making broad demands that can only be realized through electing a whole new government or, at the extreme, seizing power.

        A final direction of dispersion is outwards. Chalmers (1992) refers to the US Congress, international organizations and even the decision-making bodies of major foreign private organizations as "secondary arenas" of domestic politics, when they become the locus of conflict not only among the citizens or members of those foreign institutions but also for domestic actors from the country affected by those decisions. It is clear that in some countries, most dramatically El Salvador, the decision-making processes that shape the actions of the UN peace-keeping presence are an important arena for many political actors within the country. A more complicated case concerns US decision making bodies.

        In the discussion of the decision on the North American Free Trade Agreement (NAFTA) in Cook's essay, it is clear, for example, that the US Congress was an arena for making decisions significant for Mexican citizens. This is nothing new, of course, since what goes on in "Washington" has long been of major importance. The difference lies in the way in which politics is played out in that arena. In the classic view (and in international law), such centers as the US Congress do not become arenas of contestation for citizens of foreign countries, and their influence is limited to exchanges of notes between the heads of state and their representatives. Although this norm was ignored by many businessmen, revolutionaries, and governmental lobbyists, adherence to this formality was emphasized by many Latin American governments--Mexico in particular--as a device to prevent retaliatory intervention by US actors in Mexican domestic politics.

        In the new situation, illustrated by the NAFTA debate, not only did the Mexican government mount a very broad campaign to influence the US Congress, but also Canadian and US NGOs and labor unions allied with Mexican groups and parties opposed to a purely market-oriented treaty. The net result was to make the US Congress a secondary arena of Mexican decision making. Furthermore, this case illustrates the interaction of the formation of networks and the dispersion of decision-making centers, in that these actors helped to create new secondary arenas, i.e., the trinational institutional arrangements for monitoring and dispute settlement on labor and environmental issues set up through the side agreements to NAFTA.

        It is important not to conflate the dispersion of decision-making activity with other related but analytically distinct aspects of the configuration of state power. For example, the distribution of decision-making activity is different from the question of where final legal authority lies. Whether a subunit has the legal authority to make the final decision, or whether it remains in the hands of the president, may or may not be important for the strategies of popular actors. The question of where, in the normal course of events, public policy is discussed, bargained over, and shaped, is arguably more important for strategy than this legal issue. When political actors strive to influence decisions or gain concessions, it is to these centers that they direct their efforts, and it is around these centers that networks are built.

        Efforts to establish checks and balances among the branches of government are also important, but probably more for preserving than for shaping popular representation. Achieving a separation of powers entails structuring decision-making processes in such a way that the legislature, executive, and judiciary can check each other. Even with built-in checks and balances, this is a single process, with the three powers jointly shaping public policy through a combination of conflict, competition, and cooperation. The dispersion of decision making concerns, instead, the fact that the actual work of forging policy out of competing demands and claims takes place in multiple centers. Establishing effective checks and balances is a major concern for popular representation because of the haunting danger of the collapse of democracy itself¹². If Fujimori and Menem can ignore or manipulate congress, then any structural changes in popular representation might conceivably be wiped out in a new wave of legalized repression. We are inclined to believe, however, that short of that outcome, the dispersion of decision-making authority is rooted at least in part in the requirements of government (and governance) in a complex society, and is likely to continue, even if the separation of powers is badly flawed.

        It should be clear that the trend towards multi-centric decision making is not the same as democratization in the sense of shifting power to "the people." Democratization entails a change in the rules of political competition and participation. The dispersion concerns the units of decision making, which may or may not be democratic. The transfer of policy making to a relatively autonomous ministry or agency, in itself, is only a shift from one set of elites to another. A new sub-center of policy making may even be less democratic than the larger system, for example, if power is devolved to state governments where a narrow group of elites monopolizes control; the "subnational authoritarian regimes" identified by Fox are a clear example. It may even be suggested that, in some cases, the "farming out" of policy making may be a way to remove it from the sites likely to be influenced by popular groups. One logic of creating independent central banks, for example, is to insulate them from politically motivated decisions which might be "too favorable" to (short-term) popular needs.

        On the other hand, it is also not true that dispersion of decision making is always non-democratic. Decentralization may bring decision making physically closer and more accessible to concerned actors, and the disaggregation of policy areas may have the desirable effect of making policy debate more substantive and meaningful from the point of view of the particular popular groups involved, say, indigenous groups, specific sets of workers, or participants in the informal economy. There are thus important potentials for democratization.

        Besides dispersion of decision making, another broad factor affecting the shape of popular participation concerns the lowering costs for political actors of establishing contact with others sharing similar interests, coordinating with them, exploring possible solutions and seeking out resources for action. It is clear that the availability of less expensive mass media--print, radio, television--has had an important impact, but there now are also more specialized communication--inexpensive publication of pamphlets, books, and journals; access to discussion groups in electronic media; and less expensive telephones, fax machines, and electronic mail.

        The increasing number of mass media sources and, especially, the escalating availability of more specialized media have a massive impact on political organization, though its precise nature is hard to pin down. At the very least, these developments have made possible the establishment of links beyond community and organizational boundaries to diverse people at some distance around relatively special interests and concerns. At the same time, new and expanded means of communication vastly increase access to knowledge--and misinformation--as it becomes easier and easier to obtain competing sources of information about many different kinds of issues. In addition to diffusing through more efficient and accessible technologies, information also flows more readily in today's Latin America through the activities of universities, research centers, NGOs, and many other organizations. Obviously, access to these new technologies and ideas is not evenly distributed and the possibility for new informational inequalities is real. Overall, the impact of changes in communication on the shape of political organization, while diffuse, is profound.

        A third force that shapes popular representation is the emerging sets of beliefs held by elites about how they should relate to those popular actors with whom they have most to do. Traditional elitist attitudes suggesting distance, the appropriateness of clientelistic relations, or authoritarian postures are no doubt still common. Yet, there are growing currents--and not only in Latin America--that promote a different approach. An increasingly wide variety of actors perceive the need for complex social coordination to make and implement effective decisions in the particular settings of most concern to them, whether they be firms, government agencies, or other kinds of institutions. Whether the goal is sustainable development, producing competitive goods, combating extreme poverty, making efficient use of scarce state resources, or some other, such social coordination--often referred to as "governance" ¹³--requires more effective cooperation between, for example, public and private actors or workers and managers.

        Insofar as some set of popular actors is often deemed essential to any particular instance of coordination--and can hinder it if they are excluded or coerced into participation-- the search for collaborative solutions contributes however unintentionally to the emergence of new forms of popular representation. Such new forms, therefore, grow out of the need to solve problems of social coordination and governance, especially to provide scarce public goods and manage common-pool resources (Ostrom 1990, 1994). Mass participation is seldom the resulting form, but rather more diffuse participation focused around relatively specific purposes.

        A fourth factor shaping new forms of popular representation is new patterns of the internationalization of Latin American politics. Foreigners of one kind or another have long had a direct interest in mass politics. The economic presence of multinational capital has, of course, long been emphasized by observers. In an earlier period, international solidarity movements supported mobilization by various left and center-left parties sympathetic with political movements like Communism, Social Democracy or Christian Democracy. For its part, the United States Government encouraged conservative counter mobilization through a combination of development assistance, support for anti-Communist political forces, and covert action. Recently, Washington has promoted "governance" and building "free markets". However fiscal crises and ideological turns in the US itself have diminished its once hegemonic impact in the region. And the collapse of the expansionist socialist bloc has radically changed global political competition.

        Partisan and government influences by foreign actors on popular representation have come to be overshadowed by those of international organizations, foreign NGOs, and the diverse needs of international corporations. These actors are pushing in a variety of political directions, but a prominent thread is the emphasis on project or sector-specific involvements and some version of the governance doctrines noted above. The World Bank and the Inter-American Development Bank often promote some kind of participation by program beneficiaries and other local actors in order to rationalize project administration. They have been joined by other international organizations such as the Organization of American States and various United Nations agencies as well as semi-governmental donors, such as the European foundations, the various branches of the National Endowment for Democracy in the US, the international representatives of labor organizations, and a host of private donors. The overall impact, we believe, has been away from promoting mass movements and towards project-oriented mobilization focused on delimited groups of people, more particular objectives, and some form of interaction among government officials, foreign and domestic experts, and service recipients. Although there are enough reports to know that such interactions are not always successful and, more to the point, to the satisfaction of supposed beneficiaries, the broad impact seems to be fostering specialized groupings within network structures.

        Finally, as many countries in Latin America have recently experienced obvious failures of political institutions, a complex learning process seems to have gone on that may also contribute to nascent forms of popular representation. In Brazil and the Southern Cone, many new political organizations emerged out of the opposition to military governments. In countries in transition like Mexico or in breakdown (and transition?) such as Venezuela, groupings associated with the popular sectors have had to build outside established, but weakening party and governmental organizations (See Fox, Bruhn and López-Maya in this volume).

        While some may argue that many new organizations were the product of special circumstances and will fade, we do not believe that is always true. NGOs emerged during the fight against authoritarian regimes, neighborhood organizations arose in response to crises such as earthquakes (especially in Mexico), and election monitoring has been important in establishing "founding elections". Yet, the presumption is that once the crisis is over, the groups will disappear, or be absorbed into political parties and government agencies. The evidence is clearly not definitive, but there is some ground for another conclusion, namely, that the activists from these organizations have struggled--sometimes successfully--to find new roles for themselves and/or their organizations after the immediate crisis has passed, without returning to what are often perceived to be failed or much weakened parties, electoral processes and local government. The legacy of earlier activism is thus sometimes both a (sometimes) sense of accomplishment when objectives have been achieved at least partially, and the capacity to acquire the knowledge, skills and resources necessary to act politically around new issues, in new venues, or through new organizations, while remaining independent.

        With regard to the learning by movements on the left, the apparent failure of mass-mobilization models seems to have led people to explore alternatives other than a return to party mobilization. There are many reasons: their re-valorization of formal democracy following the experience of repression under authoritarianism; the critique of left instrumentalism and vanguardism; the growing recognition of multiple sites of domination (gender, ethnicity, race, class, etc.); and the decline or demise of once-attractive international models (social democracy and state socialism respectively). In addition, more immediate needs as well as global shifts in progressive thinking have fostered individual identification with groups much less broad than el pueblo or the working class. As a result, initiatives from below have a more delimited focus in which collective action centers around localities, the workplace, or previously neglected problems, such as environmental degradation, gender exploitation, and racial and ethnic discrimination.

        These and other factors shaping the reconfiguration of popular representation could and should be the subject of much more extensive analysis than space permits. Yet, we are convinced that the combination of trends in governance, information technology, ideas about collaboration in societal domains, internationalization and other basic forces in contemporary society are weighty enough to go a long way toward changing political action in many Latin American countries. And if those underlying trends are broadly convergent, then the chances of seeing a strengthened new political model emerge are quite significant. We believe that that is the case, and that the associative network is the centerpiece of that model.
 
 

A NEW ANALYTICAL MODEL

        We will present associative networks as an ideal type in Weberian sense. That is, we are not presenting this as achieving some normative ideal, but rather as an elaboration of what the characteristics of a polity would be if the model were widespread. We recognize that in any extant polity in Latin America elements of this new structure of popular representation co-exist with vestigial or modified elements of older structures like corporatism, clientelism, and populism, along with classic forms of territorial representation based on elections and political parties.

        The other sense of the term "ideal", that is, as a norm or a goal, is not irrelevant to the particular model of an associative network that we develop, as will be apparent. Every form of popular representation embodies some ideal condition that may or may not be achieved. Corporatists aspired to an ordered harmony, and even clientelism suggests a kind of caring paternalism. The norms associated with associative networks will evidently be a combination of flexible adaptation induced by sheer pragmatism, together with the possibility of more "discursive" or "deliberative" democracy (Dryzek 1990, Mansbridge 1983, Habermas 1987). But our goal is far from that of promoting the ideal, or even its analysis. Rather, we see the evolution of associative networks as a fact--by no means fully realized, but partially so--and our intention is to lay out those characteristics. We take no position on the desirability of this form over others, although we will call attention to its advantages and disadvantages from the standpoint of the popular actors it links to the state, and try to elucidate, in the concluding sections, the conditions under which these advantages may be realized more fully.
 
 

Associative Networks
 
 

        What sorts of structures of representation connect the popular sectors with the state?¹⁴ Increasingly, we argue, they are a particularly flexible kind that we call "associative networks." Let us first clarify the general category of which we see associative networks as a particular type.

        By "structures of representation," we mean first the sets of individuals or organizations that make claims and second, the mediating entities that debate, reshape, and transmit claims and pressures to authoritative decision-making centers. For example, a series of NGOs and grassroots groups may be linked to social policy-making centers through a government consultative council, as in the case of Ecuador discussed by Segarra. Or rural workers may be organized into rural unions in Northeast Brazil, as discussed by Pereira. The latter's links to decision centers take place in part through the relatively formal ties of those unions with the government in Brazil's corporative labor structure, though they also participate in a network of organizations and official agencies that are involved in issues of agrarian policy; rural unions' contribution to the latter is mediated through public statements, as well as strikes and demonstrations.

        We do not mean only an "interest group" and its channels for making demands when referring to structures of representation. We are concerned also with linking procedures and organizations that resolve conflicts between competing interests and in which debate or discussion takes places, leading to clarification of preferences, examination of the conditions affecting any policy decision that may emerge, and interpretation of the meanings of the actions to be taken. Thus we are interested not only in demand-making and bargaining, but also in "cognitive politics." Cognitive processes relevant to politics include perception, social learning, and communication, which go beyond the strategic bargaining based on fixed interests to encompass the consideration and sometimes the resolution of competing claims. A structure of representation is thus a set of actors that are linked to decision-making centers through a set of procedures and organizations in which bargaining and cognitive exchanges occur, influencing the policies of that center.

        Such structures of representation become institutionalized in different ways. One common, but not exclusive, basis anchors the structure of representation around a single decision center. The most obviously institutionalized structure of representation is that built around legislatures with geographic constituencies and representatives interacting through election campaigns and political debate. This is the structure of representation most likely enshrined in democratic constitutions. Legislatures are the classic representative institution, so there is a presumption that they are the unique structure of representation shaping the entire polity. Yet it has long been recognized that, in all systems, a great deal goes on outside of this formal institution, particularly in highly centralized states such as those that have prevailed until recently in most of Latin America. Much of this activity is "extra-constitutional" and sometimes ignored in looking for the institutional basis of politics. Yet, if "institutionalization" means more than enshrining a pattern in the constitution, and more than those patterns which are legally established, then there are many more institutionalized structures of representation¹⁵.

        One special but influential structure of representation that was not defined by it relationship to a single decision making center was that made up of class-based and programmatically-oriented political parties, which, particularly in Europe, became an ideal of "mass parties". They constitute a structure of representation in themselves, in that they define a membership or constituency and establish procedures for transmitting claims and demands, organizing debate, building consensus and control, and finally, bringing influence to bear on decisions in many decision-making arenas¹⁶. The "glue" holding the structure together, in addition to the organizational commitments of the members, is the common concern with the issues and program of the party.

        Beyond these structures, there are many other possibilities--not only organized around centers other than the legislature, but also around major issues which are not the exclusive domain of a political party. The notion of "issue-" or "policy networks" (Atkinson and Coleman 1992, Heclo 1977, Katzenstein 1985, Keck and Sikkink forthcoming , Marin 1991) starts from a different point, emphasizing the personal relationships and shared commitment to ideas within the networks rather than the notion of representation. In order to fit more closely our conception of associative networks, such formulations would have to be adjusted to pay attention to their relationship with political claims--in this case those of the popular sectors. Yet there is a common thread between our approach and these in terms of a concern with the way in which political space is structured. From our point of view, there are many different actual or potential structures of representation in any given polity, and this is an area where significant evolution seems to be taking place not just in Latin America, but around the world.

        We have already discussed a series of structures of representation that have been common in the region: clientelism, populism, corporatism, and mass mobilization. Clientelism is a structure of representation where the actors are individuals, and the links are exchanges with brokers and patrons. The patron resolves competing claims in this structure. Populism is one in which the actors are primarily individuals (in large clusters) and the links are found in the interplay of the leader's charisma, promises, etc. and the moods and demands of the masses. In the classic forms of populism (e.g., Perón in Argentina and Vargas in Brazil), this kind of link was strengthened by organizations (the leader's political party) and often reinforced by clientelistic "machines." In either case, the "rules" of bargaining and debate are heavily centered on the populist leader. Finally, corporatism is a structure of representation in which the units are (usually) officially sanctioned labor, business, or professional organizations and the links are the organizational and legal ties between these and with the state bureaucracy, as well as, in some cases, the governing party¹⁷.

        In the increasingly polycentric states in the region, popular representation--and political representation more generally--is increasingly taking place through a different structure of representation, the associative network. We use the term "associative" to call attention to the way in which these structures arise out of and rest on purposive, non-hierarchical "acts of association." As a particular way of coming together to form a framework of social interaction, associating or "associationalism" is both distinct from and historically parallel in importance to the two major principles of modern social and political organization--"hierarchies" (like centralized states and corporations), with their trademark relations of authority and dependency, and ideal-typical "markets," with their faceless competition among atomized agents (Cohen and Rogers 1992, Hirst 1994)¹⁸. We use the term "network", in the way that social scientists--particularly sociologists¹⁹--use it, to characterize social ties without making assumptions about the nature of the norms guiding participants' interaction. Within any given network, there may or may not be competition, conflict, ascriptive features, domination and dependency, or cooperation. The concept of networks also possesses the attractive feature of connoting a purposeful interconnectedness. The associative networks we are concerned with are structures of representation, and therefore their central purposes is that of shaping public decisions and policy. Putting the terms together, a working definition of associative networks emerges: "non-hierarchical structures formed through decisions by multiple actors who come together to shape public policy.

        Organized around issues and one of the new decision making centers, such networks are numerous and changing. The distinctive characteristics of this new model are the following:
 

• First, the form of any single associative network is likely to be characterized by a diversity of organizations, individuals, and other participants.

• Second, any particular network is likely to be reconfigured over time as issues, decision making rules, participants and opportunities change.

• Third, the associative network entails strong emphasis on what we have called cognitive politics, involving debate and discussion of preferences, understandings, and claims, in addition to--and potentially transforming--more conventional bargaining over demands and interests²⁰.

• Finally, while associative networks can and often do involve actors with sharply unequal resources, there are typically more chances to escape or shift the ground to avoid a direct test of strength with an unequal competitor. These chances derive, we believe from the lesser importance of rigid, hierarchical authority relations (compared with party or corporatist forms), the shifting and multiple patterns of identity (compared with clientelism and populism) and the more open-ended character of cognitive politics. This results not so much in more equality as less rigid inequality among the participants.

        We will discuss and illustrate these four characteristics by reference to the two basic elements of a structure of representation--the set of actors who participate in it, and the links between them and the decision making arenas, where debate and bargaining take place.
 
 

Popular Actors In Associative Networks
 
 

        The increasing vitality and diversity of popular sector-actors is apparent, but there is also a kind of fragmentation and specialization, as well. New identities have become the basis for associations, alongside the "traditional" ones based on class, locality or ties to particular political leaders. As shown in the articles by Cook, Martin, Murillo, and Pereira, well-established popular organizations such as labor unions are struggling with, but also adapting to, the new demands of more open economies. For the most part, the project of representing a broad social class in a long, historical, and potentially revolutionary process has been replaced with concerns for the workers' well-being and role in a rapidly changing economy²¹.

        In addition, there are many more organizations based on ethnic identities, identification with ecological causes in almost all countries, gender-based groupings among all classes, neighborhood organizations in many cities, and a variety of grassroots organizations in rural areas. While some of these might be considered "new social movements" (Alvarez and Escobar 1992), there are few signs that "post-modern" causes are taking over, and the older organizations based on "traditional" identities are still very much with us. Diversification and a search for new forms of collective action seem more the norm than a replacement of one type of identities with another.

        Government organizations and officials, as well as concerned party activists, legislators, and businessmen are also among the potential non-popular participants in associative networks. Frequently, issue-specific networks will involve those government officials responsible for that issue, interacting with the organized groups who make relevant claims and stimulate discussion among them. Thus associative networks should not be thought of as sitting in civil society, separate from the government, but rather as connecting segments of civil society with the state. The growth of associative networks is not the "growth of civil society," but the growth of its connections with the state.

        Another set of potentially participating organizations that is neither popular nor official is usually called "non-governmental" or "non-profit". To this pair of negatives, one might add "non-party", to emphasize that not only are these organizations not a part of the state apparatus, nor the business community, but they do not function like a party (for example, fielding candidates for elective office)²². Choosing the designation most common in Latin America, non-governmental organizations (NGOs), we give that term the meaning employed by Piester and Segarra in their chapters in this volume, roughly suggested by the notion of "intermediary organizations" as used by Thomas Carroll (1992) or "grassroots service organizations." These are organizations of professionals or semi-professionals who, when they are involved with the popular sectors, help to organize projects among them, develop policy proposals relevant to them, generate information about them, help to mobilize them, and provide many other services. They clearly are directly relevant to these sectors, although they are often neither of these sectors themselves nor directly engaged in representing their interests in the way, say, a union organization represents its membership. The issue of the (lack of) representativity of NGOs is, in fact, often a source of tension. Yet, they certainly belong in any description of structures of popular representation because of their role in shaping claims and making it possible to bring pressure to bear on decision-making centers.

        This brings us to an aspect of the model that underpins not only the diversity but also the ongoing reconfiguration of the representative structures. which is another central characteristic of associative networks. A political party of the idealized type described above presumably does not need much reconfiguration, since the same groups within the party develop positions on a very wide range of issues, provide skills needed to mobilize support, establish bases for resolving disputes, generate a media presence and use their local affiliates to understand how people are reacting to events. In a more polycentric state where alliances vary greatly from issue to issue, parties are less able to do this than they ever were. The electoral-representative links usually provided by parties and party systems are even more under strain. More and more individuals and organizations become involved in order to meet the demands of the political process, and to develop and use skills independent of particular interests or particular policies.

        NGOs, both domestic and internationally based, have begun to play some of these diverse roles, even those that enter the field intending only to implement specific projects. They become involved in mediating between the state and grassroots organizations and begin variously to mobilize, bargain and debate, and press for action. Moreover, popular groups need and sometimes obtain from NGOs, organizational resources, including spaces to meet, communications equipment, information to make strategies and formulate proposals, training, organization, openings to make contact in the government, and funds to pay salaries and buy equipment.

        To some extent NGOs have become part of a professional "political service sector," which includes media specialists, academic policy experts, newspaper pundits, pollsters, grass-roots organizers, training facilities, think-tanks, and funders (Chalmers 1992). Not only are they better trained for these things, but they often provide organizational resources to a wide range of groups. Their connections with groups may be enduring, or on a contract basis. One of the changes in politics more generally in Latin America, as elsewhere, is the emergence of such professionals. Election campaigns have become the domain of pollsters, public relations, image consultants and all kinds of "handlers." Policy making is another area of professionalization. Policy think-tanks, given a major push when they provided a space for intellectuals during periods of repression (Puryear, 1994), have emerged as ongoing centers for policy discussion and formulation, sometimes linked to the government, sometimes not (Goodwin and Nacht 1995).

        Many of these professionals are not, of course, oriented towards the poor--who are less likely to be able to pay. Yet, whether it is because of social commitment, a chance to use their skills, career advancement or long term political ambition,(or all of these) there are significant numbers of such professionals working with such populations as the urban and rural poor or the indigenous. Many of the groups we identify as NGOs appear to think of themselves in this way.

        A noteworthy aspect of this political service sector is the presence of a substantial number of people and organizations from abroad, perhaps because they are likely to have specialized skills and also because their non-citizen status is likely to push them away from a more "direct" role in the political process. Pollsters and economic policy specialists may be hired international experts. International sources may remain behind the scenes as funders of projects managed by domestic actors. They may pay for local opinion or policy studies or for meetings of activists. They may arrange for publication of local experts. For good or ill--and neither can be taken for granted--the international presence has helped to change the political economy of organizing and acting for the popular sectors.

        While NGOs may lower the costs of resources for political organization, technical factors do as well. Buying time in the mass media is very expensive in Latin America as elsewhere, but the immediacy of television, together with strategies for getting free or cheap political coverage may--in those countries with uncontrolled media--make some kinds of appeals much less expensive than they might have been without the media. In general, of course, information and communication technology, crucial resources for any kind of organization, have become notoriously cheaper, even when they remain beyond the reach of many. The ease of communicating, including the enormous reduction in person to person communication afforded by computer networks, has played a role in shaping organizations, as is clear from the story about the building of a transnational coalition among citizens of Canada, Mexico and the US recounted by Cook.

        The net result of the diversification of popular-sector (and other) organizations, as well as the addition of new organizations that provide the resources to mobilize, articulate claims and interact with like-minded organizations, is not, as some might have expected, to create a dense solidaristic web of popular organizations in a broad movement. Rather what emerges is a highly differentiated and rapidly shifting set of popular groups struggling, and sometimes succeeding, in forming a cluster of organizations equipped to recognize, analyze, debate and bring pressure on issues. As a consequence, popular representation increasingly depends on the participation of societal and state actors in a shifting set of associative networks.
 
 

Linkages in Associative Networks
 
 

        What kinds of interactions bring these diverse actors together to debate, reshape, and transmit claims and pressures to decision-making centers? How does interaction among the organizations and individuals take place, and how is the output--a strong policy recommendation, a signal of a deep division, or readiness to accept the status quo--produced? In the formal constitutional system, representatives debate and vote on laws in legislatures. In a strict corporatist model, these processes would be found in the formal meetings among designated representatives. In the idealized party version, it lies in the deliberative bodies of the party organization.

        Recalling their distinctive characteristics, the mediating processes in associative networks (a) relate diverse types of organizations and people, (b) frequently change in response to new situations, (c) generally privilege cognitive politics and (d) rarely exhibit a sharp hierarchical form. Where does this happen?

        In the cases covered in this book, some networks are centered in formal institutions. If popular actors simply were coopted into a hierarchical administrative agency or government party, this situation clearly would not fit within the notion of an associative network. However, governments in the region have been experimenting with "demand-based" and interactive administrative forms, particularly for public works and social programs, which take some steps towards qualifying as associative networks (Graham 1994). In the face of potential political unrest as a result of economic restructuring, governments have sought to reach out to the poor, as with the National Solidarity program during the Salinas administration in Mexico. The innovation that moves it in the direction of the associative network model, is that local groups are established and share in some degree in setting priorities for spending on social and public works projects through local committees. While these are no doubt often manipulated by the government--and the evidence is by no means all in²³--it is hard to believe that at least some of the local groups created will not have achieved some distinctive organizational base and identity, and find ways of negotiating with--or confronting--the government or other entities, whether the program survives in some form or completely disappears. Indeed, Fox highlights the paradox that manipulatively conceived programs may end up engendering or politicizing popular claimants.

        Another model centered in official bodies were the experiments in local government launched by leaders aligned with left opposition parties in Brazil and Uruguay, described by Nylen and Winn/Ferro-Clérico. There and elsewhere, local governments created political/administrative structures to involve popular organizations in both decision making and policy implementation. Another partial example concerns the Ecuadorian experimentation with creating consultative bodies including both public and private groups (NGOs) in the Social Emergency Investment Fund program, discussed by Segarra²⁴. Finally, there is the example Martin describes concerning the metal worker unions' entry into the sectoral chamber in the automobile sector in Brazil. This chamber should not be seen as reproducing the corporatism of the past, but rather to have constituted one step toward creating centers for discussion of the challenges that competition present for all its participants. Although this took place within a formal, government sponsored institution, the agenda and negotiating dynamics were much more open-ended that in, say, a conventional collective bargaining situation. Finally Martin suggests that the network ties among actors in the auto sector have persisted and in some cases, even grown in the wake of the state's formal retreat from the auto chamber, and example which illuminates nicely the adaptation and reconfiguration of networks in response to changing political and economic environments.

        More typical of the associative networks then the organizational links which are generated within these formal institutions, however, is a pattern of interaction which is less formalized, and a process of producing outcomes which is less guided by formal institutions. Even those which are principally centered in an organization seem very likely to develop links beyond the organization as well.

        As the term network implies, these interactions can take the form of personal links among the individual members from various organizations. Even when the network is centered in a government agency, the links seem to spread out beyond the boundaries of the organization to other group representatives, experts, and so forth. The personal linkages are often a crucial element, particularly in the process of arriving at some formal position.

        On the other hand, one of the striking characteristics of the networks we see emerging in many countries is their creation of a public space that is not sustained as much by personal contacts as by the existence of some kind of common media--journals, television, specialized publications, radio talk shows and increasingly, computers. Communication between computers represents a mixture of the two--with direct communication between people through email, and a generalized medium with newsgroups and the rapidly growing World Wide Web.

        The importance of personal communications, mass media and electronic communication is easy enough to observe (although harder to measure). A more interesting question arises, though, when one asks how this sort of network produces an output, pressure on decision makers, policy proposals or other messages with political clout. There is ample research about public opinion and policy in the US, although less in Latin America. These impacts are probably more successful in one kind of arena--involving general public opinion about highly visible public policy, mediated through elections--where research in Latin America is even more scarce.

        The associative networks we see are more difficult to analyze, at least when they are not centered on a formal organizational structure. In the latter case, votes are taken or key meetings force some sort of consensus, but where the patterns of communication are personal or occur in the increasingly common public spaces created in the media, the process is more diffuse. Opinions are expressed, claims are made, ideas are put forward and challenged, assumptions are made and questioned. In some, threats are made, symbolic violence is enacted (perhaps even terrorism), dramatic public gestures are staged. And through it all, tendencies appear and become reinforced, interpreted and rejected or adopted. It begins to be said that "most of the activists think", "the experts believe", or simply that "everyone says..." Through multiple actors and their links the message--the de facto political output-- gets assembled and delivered, and, perhaps, decision makers act on it.

        We mean to emphasize the cognitive aspects of this process--the discussion of proposals and ideas--but not to exaggerate their importance. When, for example, the individuals and organizations in the network have well-established priorities and understandings, and their preferences are reasonably fixed, this interplay of ideas and interpretations may be seen as simply the confrontation of rationalizations, and the real process as a bargaining game. Or for those who see a sharp underlying conflict of interest between, say, the poor and the wealthy or some other division, the interplay of ideas may be seen as a process of overcoming false consciousness. In either case, something politically meaningful is occurring, even if it not an intellectual debate. Yet, especially in the type of associative networks where the target is specialized decision-making centers and diverse groups participate, including professionals and other political service providers, and the questions at stake concern how to deal with serious social problems lacking simple answers, a model of cognitive politics seems appropriate, even when very real and understood interests are also present.

        How should we characterize this open-ended process, which is at least a part of the linking processes of many associative networks? It is more like a conversation, the trading floor of a major stock market, or like street theater, than the ordered processes in a legislative body or a staff meeting. It is a kind of interactive field, in which the participants are reasonably clear but the rules according to which the outcome is determined are much less so.

        Such interactive fields are not exclusive to the associative networks that we have been describing. There is a long-running show that involves public opinion at the broadest level. The selective use of violence by guerrilla activists, the political use of dramatic public events (such as soccer championships), and many other techniques required playing out a strategic game in a public space aimed at influencing policy outcomes. Moreover, populist leaders have long played on the lack of a transparent process inherent in this sort of interactive field, making the broadest possible appeals to limit their accountability. Through the force of their personality or their monopoly over public discourse, they appropriate the right to determine what "the people" want. While modern neo-populists also clearly utilize a mass communications-driven style (Cavarozzi 1994), it seems less dangerous because of the transitory quality of its effect but more dangerous because of its volatility²⁵.

        We think the diffuse nature of the linking process in interactive fields is a very important feature of some associative networks and, therefore, of the emerging structure of popular representation. The fluidity allows both a maximum flexibility and room for innovation and clear thinking and on the other hand, seems easily manipulated, making accountability difficult. There are other difficulties, too. It may be easy for experts to dominate the process, and they may be remote from the popular sectors. Certainly, popular criticism about the insensitivity of experts to indigenous values, to the poor's non-developmental economic goals, to the costs of dislocation, all suggest that a system enshrining the values and opinions of experts--a possible but not necessary outcome of associative networks--does not always benefit the popular sectors.

        It is worth noting that the phenomenon to which we are pointing through the concept of associative networks is not the same as classical or interest-group pluralism. Aside from the plurality of societal actors involved, associative networks have little in common with classic understandings of political representation under pluralist approaches. Unlike the pluralist model of competition among interest groups and demand-making on the state through pressure politics, the type of interaction found within an associative network involves state and societal actors engaging in cognitive processes--of making, countering, and trying to resolve competing claims--that can in practice lead to the redefinition of interests as well as bargains among established ones. Moreover, the narrow connotation usually attached to the term "interest groups" fails to capture the variety of analytical categories which the popular and other societal actors participating in associative networks encompass--social movements and movement organizations based upon newly politicized collective identities, non-governmental organizations, private firms, as well as the classic membership-based interest groups organized around the socio-economic categories. Lumping these categories together not only blurs the distinctiveness of each but also negates any possibility of problematizing the relationship among them, such as the often tense ties between NGOs and popular actors. Finally, the pluralist vision is of a political world made up of individual citizens and interest groups, separate from the politicians and policymakers who aggregate their demands only through parties and elections and the apparently individualistic notion of lobbying. Classic pluralism minimizes the intermediate structures linking state and society aside from territorial representation and party systems²⁶, to which we call attention with the notion of associative networks. Such structures fuse--or at least blur greatly the boundaries between--activities of "interest articulation" and "interest aggregation," of "demand-making" and "policy making".

CONCLUSIONS: REQUIREMENTS FOR MORE

EFFECTIVE POPULAR EPRESENTATION

        A "better deal" consists of progress on three dimensions: establishing stable and effective government, overcoming the still drastic inequality in the region, and third, building spaces where people are able to participate effectively and have their claims taken seriously. The meanings of the first two of these are relatively straightforward²⁷, but a few words about the third are in order.

        Progress in creating meaningful participation is obviously a democratic objective, but the nature of associative networks has prompted us to look for more than the opportunity to vote and express opinions freely. The notion of the associative network embodies the possibility of a politics that goes well beyond--without supplanting--these traditional forms. Participation in such networks also focuses attention beyond the strength of positions in strategic bargaining to maximize interests--always important but not the only dimension. In associative networks, the challenging task for any group or individual is to define or reshape one's identity and goals and, through interaction with others, come to collective decisions. Democracy involves meaningful participation in that cognitive dimension as well as the strategic bargaining dimension of representation.

        The distinctiveness of this political style is captured by the fact that the characteristic political "tools" of each set of actors in other political situations change their significance in the dynamics of associative situations. Administrative fiat by government actors has little control over others; monolithic technocratic pronouncements must confront those of other experts; log-rolling and compromise by politicians tends to be very visible to the participants; demand-making backed by mobilization or the threat thereof on the part of popular actors in civil society is a blunt tool. In associative networks all these conventional political tools are more evidently in tension with the ideal of a shared search for solutions. This is not to say that such tactics no longer are used, and sometimes effectively. However, the emphasis on cognitive politics means that adopting such tactics can become counter-productive or suboptimal for both actors and the network-defined collectivity later, and/or in reconfigured associative networks.

        The political encounters in associative networks hold the promise of generating new understandings of the nature of the problems confronted and, by extension, new solutions. This may be true whether the problem is how to balance sustainability and economic efficiency in the use of natural resources; how to insure that social benefits are tailored to the specific needs of both particular groups of recipients and the labor market niches which they could fill; or how to promote greater cooperation among workers and managers in the interest of fostering competitiveness in higher value-added market niches. Hence, the role played by such cognitive resources as the ability to marshal information and arguments and use persuasive speech is crucial, which shifts the basis of inequality at least partially away from their structural position. If (and, of course, it is a big "if") popular sectors in associative networks are able to develop the cognitive skills, and if elites can develop the capacity to hear what they are saying, such networks represent arenas in which outcomes will not be determined uniquely by the fundamental political resources associated with class, property, social status, or access to the means of coercion.

        What, then, are the requirements that need to be met so that this new structure will promote stability, redistributive justice and democratic participation? While this is a broad topic, there are a few points that flow from the nature of the structures of representation we are describing, which we offer as the beginnings of a research agenda.

        We see three sets of requirements: (a) organizations and procedures that coordinate the multiple decision centers and constantly changing networks, (b) a framework of rights that makes participation in associative networks possible, and (c) popular-sector strategies that make the most of the opportunities and avoid the pitfalls of the new form.
 
 

Institutions for Coordination
 
 

        In the face of increasingly polycentric decision making and multiple, diverse, and sometime diffuse and constantly reconfiguring associative networks, the meta-institutional framework that organizes them is crucial. We might refer to these top-level institutions as constitutional, although many of the norms, rules, and monitoring devices necessary are indicated only very partially in written constitutions. Although there are many elements in this coordinating structure, we only touch on a few that require an approach somewhat out of the ordinary.

        In our analytical model, the key institutional elements in liberal democracy--elections, legislatures and party systems--should be considered as much or more meta-institutional elements, rather than simply structures of representation. Although the rhetoric of liberal democracy is to speak of the way in which all three effectively represent their voters, constituents, or party members, in all three cases the rules and procedures governing them have a significant impact on how sensitive decision makers are to the diverse pressures from associative networks, and how well they are coordinated in producing policy.

        Recent efforts at legislative reform in the region suggest a fairly widespread concern with establishing a coordinating function for law-making bodies. The issue is not only legislative independence from the executive, but also whether legislative institutions can "rise above" the immediate pressures of representation (which may involve clientelism, corruption, "special interests," and localism). The isolation of decision makers to ponder only collective goods is also not the relevant concern here. Rather, the issue is how to tie institutionally the legislatures into the constantly changing pattern of associative networks, whether by having members of the legislature become part of the networks, or by granting members of the networks access to the legislative process. The goal is that the input from each network be given its due in the formulation of a coherent public policy.

        Elections are another institution that is often thought of in terms of representation alone (e.g., whether the votes are correctly counted, or the clarity and binding quality (or lack thereof) of the mandate that emerges.) These are very important issues, particularly in countries that have experienced manipulated elections for many years. Yet it is less common to address elections in the light of the problem of coordination. A classic view that is relevant here suggests that the ideal elections are those between teams of leaders in which the electorate is called upon to judge with respect to their overall competence, not so much their specific programs (Schumpeter 1942). Yet, in a polycentric state more systematic attention should probably be paid to the capacities of leaders to respond to, and perhaps generate, associative networks28.

        Parties are even more often seen as a structure of representation. Each party, and most especially the party system, however, is also a major component of the meta-institutional framework that coordinates the various decision-making centers and the multiple associative networks. Mainwaring and Scully (1995)²⁹ have shown there are a number of countries where parties are well institutionalized, in the sense of having stable rules of competition and established organizational and voter bases. Beyond this minimum, however, we suggest that analyses of political parties will have to investigate how political parties are responding to the rise--and the challenge--of associative networks.

        The old model in which the party took on all the aspects of a structure of representation seems pretty much dead in the rapidly changing and reconfiguring world of more polycentric states. The old pluralist vision that appropriated and made a virtue of the notion of the "catch-all party" converges with the limited notion of democratic elections in which the main task was simply to elect competent teams. There seems little doubt that political parties need to be accessible and responsive to associative networks. Yet, as coalition builders responsible for organizing debate and bargaining among the various actors in the top-level decision-making centers, they need some way of building in not just flexibility, but meaningful interaction with the constantly changing associative networks. In one sense, parties are losing (or failing to maintain) their dominance within processes of representation, yet, if they can become indispensable coordinators among decision-making centers (a tall order, to be sure), they could become more central in the end.
 
 

The Rights and Resources Necessary for Participation
 
 

        If a new, more flexible, more complex and, perhaps, more powerful set of structures of representation is emerging, then the conditions for popular actors becoming involved will shape strongly whether the popular sectors will be represented effectively or not. Since these are not formal institutions based on individuals, such as elections, equalization of rights to participate must go beyond any single institutional or legal change. Guarantees of rights are very important, however, and take many forms.

        A first requirement, suggested by the chapters on the stubbornness of violence, is the guarantee of personal security. This may seem an odd place to start, but the formation of associative networks clearly requires some level of personal security, not only from state repression, but also from a breakdown of order, or the kind of situation outlined by Paulo Sergio Pinhiero, in which government agents and paramilitary groups behave like vigilantes.

        A second set of institutional requirements for rendering an associative network just is that the formation and activities of associations, NGOs, movements, etc., be as free as possible. People should not be subject to private or public measures designed to prevent group formation, such as penalizing workers for joining a union. In addition, a legal status for organizations that recognizes their existence and gives them standing in the courts, legislative hearings and other arenas is important. For instance, one of the relevant struggles in Latin America recently has been to find the right legal status for NGOs. Yet, the emphasis we have given to the importance of resources involves not only the legal right to exist and be heard, but also the facilitation of their access to resources, without--as Jonathan Fox points this out in discussing clientelism--conditioning that access on political loyalty. These resources include the right to information from a variety of sources, particularly about the operations of government.

        A third set of requirements concern the establishment of arenas of discussion, debate, and policy formation that are accessible to popular groups. An associative network model raises not only the issue of the relative power of popular groups, but, since so much of the politics in this model takes place in the shaping of policy arenas, their ability to shape agendas. The establishment of media which is independent not only of the state but of particular economic interests and accessible and attentive to popular voices is clearly a central task. In many countries, universities have also been a place where agendas can be shaped. The growth of independent universities, research institutes and other centers of discussion imbued with a healthy respect for popular concerns, is an important requisite.

        A fourth requirement relates to the waves of financial scandals that currently seem to plague the region. A challenge to the effectiveness of government clearly lies in the corruption and rent-seeking of public officials. And one can probably credit at least part of the new trend towards exposing such corruption as the grudging elite rejection of such practices as detrimental to their role--and reputation--in the world. Yet, there is an important question about distributive justice involved here, too. The old clientelistic and populist patterns sometimes actually used corruption as a tool for redistribution, but in the associative network model, such a possibility is absent. Corruption becomes merely self enrichment, not the building of a machine based on a favored clientele. In this connection, the challenge for redistributive justice in the current wave of anti-corruption efforts is that they be done in such a way as to establish meritocratic standards and public accountability for officials.

        A fifth set of requirement centers on the need to protect group rights. If individuals continue to be subject to tacit of official discrimination based on race, ethnicity, gender, sexual orientation, or other aspects of their background and beliefs, associative networks will face obstacles in both their social reach and their ability to secure greater stability, justice and participation for the popular sectors. Specific guarantees that recognize the disadvantages faced by specific social groupings--including special provisions for the political and cultural autonomy and economic welfare of the region's many indigenous peoples--are another meta-institutional requirement for a better deal.

        Some of these rights can be achieved by legislation, but this is a more complex process and its full realization will depend also on changing norms and beliefs among both citizens and officials. Any structures of representation need to be embedded in society and, as a relatively new one, associative networks require learning a new set of standards and behaviors.
 

Popular-Sector Strategy: Towards a New Politics of Inequality
 

        The final set of requirements concerns the principles guiding the actions of popular actors and their political allies. The problems they face include the need to prevent the multiplicity of associative networks from becoming fragmentation. Many associative networks create a serious potential for the splintering of political issues--and hence of organizations--into arenas that become disconnected, depoliticized, and narrowly circumscribed.

        Further, popular-sector participants enter into associative networks with unequal resources. Indigenous activists confronting government authorities and local landowners over land-tenure and development issues are less likely to have access to education budgets, land-use studies, or powerful political ties that are often important. Neighborhood organizations concerned with urban services and housing are likely to be much more limited in their ability to have access to architects and engineers, the money or social prestige to place ads or influence the local television, or the means of coercion available to city officials, building contractors, and local businessmen with whom they have to deal. Unions trying to organize in restructuring industries may have to develop considerable capacity to understand market trends, global commodity chains, and trade regulations in order to interact meaningfully and effectively with government economists and management. Moreover, as Carlos Vilas writes in the introduction, there are still significant areas where actors are willing to use violence--Guatemala for much of this century, the interior of Brazil, the shantytowns of Lima. If violence becomes the norm, the formation of networks is impossible. But if violence is used selectively, as it often is, as a means of gaining tactical advantage, then unless popular sectors groups are armed, they will be at a disadvantage compared to the landowners, state officials, or others who can command or hire the necessary means of coercion.

        In sum, although associative networks present real possibilities for improving the conditions of the popular sectors, it is clear that they will not automatically do so. Simply changing the type of structure of representation obviously does not mean that the popular sectors will be taken more seriously or that their claims necessarily will be satisfied. Even if the shift to associative networks were complete and universal, it would not necessarily mean a perfect democracy. Associative networks, like clientelistic relations, corporatist institutions, class-based parties, and populist movements, can be manipulated by elites to entrench or reinvent inequalities or mystify people about their real interests. Their fluidity and reliance on people who frame agendas and offer information present distinctive opportunities for such manipulation. Yet, we argue, when their inherent possibilities are realized by resourceful actors, associative networks do offer new ways of clarifying and reformulating strategic goals and finding new resources and allies to pursue claims; to at least some extent, popular actors are already using these emerging networks in such a way.

        If our view of the dynamics of new structures of representation centered on associative networks has any validity, a new politics of inequality becomes possible. For example, the elaboration of associative networks changes the context of political action, in part because it encourages an elite view of popular sector as partners in finding solutions, as opposed to being either enemies or objects of all-or-nothing co-optation. This does not make them equal partners, but it opens up new opportunities. Since specific popular actors can influence decisions of particular concern to them, taken together, they have the potential to shape public decision making more directly than ever before. In these arenas, "experts" may become privileged, but in associative networks popular actors may seek the help of technically sophisticated allies, and, further, bring their "local knowledge" to bear.

        The diversification of arenas and the weakening of broad ideological claims make it possible to what would have been contradictory before, "having it both ways" (like the auto unionists in Brazil, who negotiated with Collor government officials on sectoral issues while continuing to be political opponents of that government and harsh critics of its overall economic policies.) Despite continued diversity and discussion, there are trends within the left which are consistent with the rise of associative networks: the embrace of proactive styles; the erosion of hierarchical organizational forms like vanguardism and democratic centralism; the movement away from instrumental party attitudes toward popular organization (including the formation of movement-based parties such as the PT and the Causa R); and the valorization of formal democratic institutions. And while the broad political mobilization associated with Social Democracy has not emerged, some popular sector activists have been influenced by some of its core values, like "concertation," "codetermination," and "social solidarity." Taking other streams of ideas, "radical democracy," with its emphasis on deepening democracy and extending it into diverse spheres of social life, and various notions of "citizenship," generally link procedural issues of democratic rights with substantive issues of equity and equality.

        Some efforts to reformulate popular strategy along these general lines have centered on "individualistic" mechanisms of direct democracy like plebiscites and referenda. We believe that even in an ideal situation they can and should play only a supporting, rather than central, role in reformulating representation along progressive lines. They encourage a "one-time-only" electorally-centered mobilization and a "yes or no" formulation of often complex issues which is in sharp contradiction with the many-sided quality of associative networks and their comparatively deeper discussions. As Filgueira and Papadopulos suggest in their essay in this volume, in Uruguay plebiscites proved useful only in the limited--although no means trivial--sense in which they enabled popular and left actors to mobilize public opinion to block technocratically conceived neoliberal reforms; it was simply not in the nature of these mechanisms to facilitate the proactive formulation of programmatic alternatives or social mobilization around such proposals.

        The broadest implication of the growth of a polycentric state and multiple, changing associative networks is that securing redistribution and meaningful participation will not be the work of some single central popular leadership, but will be accomplished, if at all, by many individuals and organizations operating in many different arenas. There will be no "popular-sector strategy" to deal with the impacts of neo-liberal policies, rather there will be social policy coalitions, progressive unions, indigenous and racial movements, and others, going their own way, and each calling on resources, allies, and assistance from a wide, shifting and sometimes international range of professionals. Political action based on solidarity and horizontal coordination among popular groups within and across these shifting networks and arenas--to take common action without a unified and centralized organization--remains a challenge.

        Successful popular-sector strategy will not depend primarily on dramatic legislation, mass campaigns, or plebiscites. It will also, and more importantly, depend on the development of the skills of many potential popular participants enabling them effectively to exploit the opportunities of associative networks. The form by no means guarantees success, but much can happen if activists can build effective and accountable organizations, obtain accurate information and analysis, and develop ways of utilizing contacts with the media, government officials and experts without becoming subordinate to them. While this task seems daunting, it places a premium on the resourcefulness and experience of popular actors in imagining and crafting new strategies and new forms of social organization.

        The vitality of popular organizations that has been recounted in this volume suggests that even in inhospitable environments progress in forging new spaces for making political claims can occur. Only time will tell whether this will be enough to secure justice.

1. We wish to acknowledge the very helpful comments of our fellow editors as well as Robert Kaufman, Daniel Mato, Al Montero, Eric Hershberg,  and the members of the Workshop on Identities, Institutions, and Representation of the Center for the Social Sciences and of the Latin America Ph.D. group, both at Columbia University. Return to Text.
2.  Although the term "representation" points to the bottom-up qualities of making claims, these structures are always simultaneously  channels for top-down processes of securing compliance, often through control.  We mean always to refer to both dimensions. For popular actors (or any other actor for that matter,) representation involves having access to decision-making processes and thus being "heard" and "taken into account," but not  necessarily that their views or positions are reflected in the final policy outcome.  In this sense, representation does not necessarily yield "representative" outcomes, for it is inherently a process of mediation among competing interests, demands, and claims. Return to Text.
3.  For instance, incorporation occurred considerably earlier in Mexico, Brazil, and the Southern Cone, on one hand, than in Ecuador, Venezuela, and Peru, on the other.  Moreover,  Guatemala constituted an clear exception to the general rule among the countries studied here, for reformist efforts to integrate the popular sector in that Central American country were quashed decisively following the 1954 military coup. Return to Text.
4.  In this connection, Haggard and Kaufman (1995, 309-334) make an important distinction between the "politics of initiation" of market reforms in developing regions like Latin America, typically involving a substantial degree of concentration of discretionary authority in the executive, and the "politics of consolidation" of market reforms and democracy, wherein two key tasks come to the fore:  first, how to "depersonalize" executive authority and make it "accountable"(335) and, second and more generally, how to embed liberal political and economic norms in formal and durable institutions (363-364).  Of course, they also note that whether or not such tasks will be fulfilled effectively remains an open question. Return to Text.
5.  On the increasing political and economic importance of the "management of international expectations," see, for instance, Roxborough (1992c). Return to Text.
6.  The label “neo-populist” has become increasingly commonplace in the literature (e.g. Castro Rea 1992, Dresser 1991, Roberts 1995).  Yet, given the ignominious political fates of the impeached Collor and the subsequently discredited Salinas, as well as the notable contrasts between their heavy-handed governing styles and the more consensus-oriented approaches of their respective successors (Franco and Cardoso in Brazil, Zedillo in Mexico), and, perhaps, the recent electoral difficulties of Menem and Fujimori at the state and/or local level, at the time of this writing (January 1996) there is considerably less reason to see neo-populism--or its analytical "cousin," delegative democracy--as an ineluctable region-wide trend than there seemed to be in the late 1980s and early 1990s. Return to Text.
7.   See the epilogue of the chapter by Fox, which emphasizes the national impact of the rebellion, even though direct mobilization was limited. Return to Text.
8.   There are many terms to describe the process and the end result. Ostrom (1990, et.al. 1993, 1994) uses the term "polycentrism" to describe a system with sub-centers, and her analysis is relevant,  but we have opted not to use the term because it suggests too complete a separation of the centers, and if used to describe a state, too evenly balanced centers. "Decentralization" is useful, but often refers only to dispersion to subnational units.  "Deconcentration" is less specific, and has some currency in contemporary debates, and we use it occasionally. Dispersion to external sites is more controversial, and phrases such as "loss of sovereignty" come to mind.  See Chalmers (1992) for a brief discussion of sovereignty in an internationalized environment. We have settled on "dispersion of decision making activity" as being the most descriptive. Return to Text.
9.  There always have been exceptions, of course. There are some non-trivial aspects of federalism in Brazil and Mexico, new state agencies often have been created to allow some modest arenas of autonomous  decision making, and military institutions have, on occasion, operated as a state within the state. Return to Text.
10.   We do not mean to imply anything about the relative "strength" of the state or about political stability when referring to the centralized state.  In fact, we believe that as a general analytical category centralized states cut across distinctions like "strong" versus "weak" states,  stable versus unstable regimes, and even authoritarian and democratic regime types. Return to Text.
11.  This aspect of dispersion is similar to Evan's (1995) notion of "embedded state autonomy," which embodies a mixture of what he calls "connectedness" to society and "insulation" from particularistic pressures.  Yet, we do not share Evan's strong analytical preoccupation with the issue of autonomy, although we recognize the important normative and practical issues raised by tight collaboration between private actors and public authorities. Return to Text.
12.  Separation of powers has become a major topic, because of the serious concern that executives in some countries have been ruling by decree rather than through effective coordination with, and coalition-forming in, the national legislature. Argentine President Carlos Menem, now in his second term, has been rather high-handed in his dealings with congress, as was Fernando Collor of Brazil during most of his two-and-half year presidency (cut short by impeachment in 1992); "high-handed" would be a charitable characterization of the governing style of Alberto Fujimori of Peru, who closed congress in 1992--to considerable popular claim--and then re-opened it under more restrictive rules making it more malleable, which were approved in a procedurally suspect national plebiscite in 1993.  These experiences have given rise to such terms as "delegative democracy" (O'Donnell 1994). Scholars have speculated that they may well become a permanent fixture on the region's political landscape.   The same mixed record can be seen with regard to progress in strengthening the judiciary and legal system, which historically have been executive-dominated.  Even in traditionally legalistic Chile, military commanders have defied openly the edicts of civilian courts and authorities. There are, however, some signs that traditionally weak legislative arenas are becoming more vibrant in Brazil and Mexico, which are reinforced by the orderliness of the unprecedented presidential impeachment proceedings in Brazil, Venezuela, and Colombia over corruption issues. Yet there seem to be contrary trends in countries like Uruguay and Chile, where traditionally strong national legislatures may have lost some ground in the post-transition era. On the Chilean case, see Garretón (1994).  On the Uruguayan, one interpretation of Filgueira and Papadopulos' essay in this volume is that national referenda proved necessary to break political deadlocks over issues that parties could not resolve in Congress. Return to Text.
13.   This is a large field. See among others, Atkinson and Goldman 1992, Campbell 1991, Cohen and Rogers 1992, Hay and Jessop 1995, Hirst 1994, Hollingsworth et. al. 1994, Malloy 1991, Nagel 1994, Ostrom 1990, Schmitter and Streeck 1985, Thompson 1991. Return to Text.
14.  As indicated above, we use the term  "representative" although the linkage is always two way control is also a possibility, and even strongly implied. Further, "representation" may involve debate and discussion, and not just transmission of demands. A structure of representation specifically includes institutionalized spaces where, for example, preferences are discussed and shaped. Return to Text.
15.   This particular conception of institutionalization resonates more with current formulations in sociology (e.g., Powell and DiMaggio 1991) that with the traditional political science emphasis on the consolidation of formal state and/or regime institutions. We are indebted to Kelly Moore for helping us clarify this point. Return to Text.
16.  Such a vision is an idealized form of political party. Many parties are clearly only parts of structures of representation, or, perhaps, parts of the meta-structures which regulate other ones. (See below.) Return to Text.
17.  Corporatism may be a form of state--with rules establishing chambers of corporations as the authoritative decision making centers, and appeared so in many idealization of the form. As many have pointed out, however, it is  almost always more important as a structure of representation relating some specific set of interests to the government, often in order to control them, rather than as a 'form of regime' (Chalmers 1985). Return to Text.
18.   In works following in the Tocquevillian tradition (e.g., Putnam 1994) as well as in those of former students of neocorporatism like Streeck and Schmitter (1985), the term "association" refers more narrowly to the act, or organizational by-product of, individuals who share a similar status vis-à-vis the larger collectivity joining together to engage in collective action.  That similar status might be, for example, a profession, a place of work, a place in the division of labor, or a common relationship with a particular public agency (e.g., welfare recipients).  In this context, we extend the noun "association" and the adjective "associative" to characterize a particular class of interactions among individuals and groupings who have markedly different status and at least one of whom embodies the formal public authority of the state. Return to Text.
19.   For a useful critical survey of network approaches, see Emirbayer and Goodwin 1994. Among the now classic works in network analysis are White et. Al 1976, Burt 1982, and Granovetter 1973. Return to Text.
20.   We note that this particular notion of the properties of an associative network seems to dovetail with a new current in network analysis labeled "structuralist constructionism" (Emirbayer and Goodwin 1994) that emphasizes the interplay of structural and symbolic-discursive elements. Return to Text.
21.   Pereira in this volume notes that resurgent rural unionism in Northeast Brazil has had more of a class character recently than before military rule, and that it relies on a government sanctioned, and therefore at least quasi-corporatist institution. He treats this as an exception to the rule, but it may also be evidence of actors that create a wide identity and utilize links to the state to deal with practical problems, a development  which may have little to do with at least the old ideal of corporatism. Return to Text.
22.   Another term, 'private voluntary organizations' seems particularly inappropriate here, for although there is some experience with charitable and strictly volunteer work, the NGOs we are concerned with are more likely to be staffed by paid personnel - however badly and erratically paid they might be. Return to Text.
23.   The general conclusion of the volume edited by Cornelius, Craig and Fox (1994) is the political impact of this presidentially-directed program was influenced by local and regional dynamics and varied greatly across the many program areas. Return to Text.
24.   Segarra notes that although the intention of the founders may have been to establish an associative network involving many kinds of groups in the Social Emergency Fund, a hierarchical and clientelistic structure eventually emerged, lacking the kind of flexibility that many had hoped for it. This example highlights the need for an analysis of the dynamics of associative networks. Return to Text.
25.   Marcelo Cavarozzi also emphasizes the importance of the media to neopopulists, in their desire to build support or acquiescence in “Politics a Key for the Long Term in South America,” in Smith, Acuña and Gamarra 1994a, pp. 127-156. Return to Text.
26.  Indeed, it is precisely the absence of such "linking structures" that led Heclo (1977) to propose the notion of "issue networks" connecting actors across legislatures, executive agencies, and interest organizations.  His work was in large part a reaction to the notion of "iron triangles" linking such actors together for particularistic, rent-seeking purposes within American politics, which itself was also a response to the absence of intermediate institutions in pluralist formulations.  Finally, of course, the whole corporatist paradigm (e.g., Schmitter  and Berger    ) particularly in the study of advanced industrial countries, was centered on an explicit critique of the absence of any acknowledgment and exploration of concrete structures of "interest intermediation" in classical pluralist approaches. Return to Text.
27.   Once, discussion would have had to account for two possible scenarios--steps taken to bring down the system in anticipation of a revolutionary reconstruction and a second involving challenges from below to bring about evolutionary changes. We will not consider the former. Return to Text.
28.   This point runs parallel to Heclo's (1977) discussion of a new breed of politicians-cum-managers he calls "policy politicians". He argues that they frequently emerge from, and negotiate between, issue networks. Return to Text.
29.  They argue that Venezuela, Costa Rica, Uruguay, Chile, Colombia and to a 'lesser degree' Argentina have well institutionalized party systems. Chapter 1. Return to Text.