Restoring Agency to Class: Puzzles from the Subcontinent

Class explains much in the differentiation of life chances and political dynamics in South Asia. Yet in the subcontinent class has lost its centrality as a way of understanding the world and how it changes. Indian intellectuals have been a major force in the eclipsing of class through discursive strategies of constructivist idealism. Formalism in social sciences finds class relations elusive and difficult to measure. Market triumphalism eclipsed concern with rehabilitation of “weaker sectors” and redressing of exploitation as measures of national success. Class analytics, however, continues to serve two critical functions: disaggregating development and explaining challenges to rules of the game. Restoring agency to class requires attention first to relations that structure choice in restricted or expansive ways. Global forces have altered people’s relations to production and to one another, as have changes in the political opportunity structure, with significant effects on tactics and outcomes. Knowing how to aggregate or disaggregate classes is more complicated than ever. Nevertheless, alternative understandings of class structure are more than academic: they reflect the strategies of political actors. The difficulty for class analysis is to illuminate the conditions under which interests of those disabled by particular class systems may be inter-subjectively recognized and acted upon politically at the local and/or international levels. Appropriate and robust sociopolitical theory for this purpose is illusive, but no more so for class than for other bases of difference — caste, community, identity, gender — that likewise seek to explain transformation of locations in social structures to effective collective agency. The Puzzle in Broad Strokes Class analytics was once central to South Asian studies, and South Asian contributions to global discourse around class were prominent. The decline of class is difficult to measure, but beyond dispute. Why has class lost, in the subcontinent Critical Asian Studies 38:4 (2006), 323–356 ISSN 1467-2715 print/1472-6033 online / 04 / 323–34 ©2006 BCAS, Inc. DOI: 10.1080/14672710601072954 and elsewhere, its importance as a way of understanding the world? Can it be that class now explains much less in the differentiation of life chances and political dynamics than previously? Has class analysis been sidelined by competing intellectual fads and political interests? This puzzle is more pronounced when we consider that class differences across the subcontinent are changing form and magnitude, and they are becoming more politically charged with rapid growth and structural change of economies. A growing middle class is celebrated as evidence of increased opportunities. Yet since the Indian government began implementing economic reforms in 1991, per capita income differentials across states have risen, along with inequality within states. These trends have moderated the effects of economic growth on poverty reduction. The postindependence ideal of a secure, protected labor force in India has been supplanted by the more typical phenomenon of capital’s pursuit of cheap, flexible informal workers without state protection; the percentage of workers in the formal sector has dropped by 2 percent since 1990. India’s agrarian sector offers evidence of crisis and armed conflict. In April 2006, India’s prime minister, Manmohan Singh, addressed chief ministers of six states affected by violent agrarian insurgency: “It would not be exaggeration to say that the problem of Naxalism is the single biggest internal security challenge ever faced by our country.” Ranjit Kumar Gupta, former police commissioner of Calcutta, estimated that the Naxalite movement has spread to 20 percent of India’s districts (159 districts in 14 states) since it began in 1967. In neighboring Nepal, Maoist rebels exercised total sovereignty over vast areas with de facto freedom to operate in virtually all rural areas by 2005. Agrarian radicals in armed conflict with police and military claim to be fighting a class war, a war for a classless society. Class analytics itself can and does evoke theological permutations among pure theorists, but the essential perspective is both straightforward and commonsensical. Though much has been made of distinguishing Marxian from Weberian class analysis, for example, both emphasize the primacy of economic assets as differentiating people across classes. Class analysis always takes the material world seriously, and empirically: it is never simply a construction or an imaginary. Class structures relations among people; these relations are critical for understanding not only life chances, but also political behavior. Authors in this collection hold to no orthodoxy. But we all agree on the asymmetric and consequential elements of class location, and thus the essentially structural em324 Critical Asian Studies 38:4 (2006) 1. Deaton and Drèze 2002. 2. NSSO 2001. 3. Members of the Peoples Guerrilla Army, Peoples War Group (PWG), Maoists Communist Center (MCC), Communist Party of India (Maoist), and the Communist Party of India (Marxist-Leninist) Janashakti are called ultra-leftists or Naxals, after the 1967 agrarian uprising in Naxalbari, West Bengal, which was sparked by landlord abuses of tribals. Singh’s comments in GOI 2006. 4. Gupta 2004. 5. Kalyvas 2006, 211. bedding of class analysis. Jon Elster writes: “A class is a group of people who by virtue of what they possess are compelled to engage in the same activities if they want to make the best use of their endowments (i.e., tangible property, intangible skills, and cultural traits).” Class location for Elster predicts and explains “endowment necessitated behavior.” Class determines what people must do, what they have the freedom to do, what they cannot do. It structures the realm of choice. Though “choice” — whether “rational” or otherwise — dominates much of contemporary social science, a structural understanding of class illustrates why there is no choice independent of some matrix of constraints and payoffs external to the individual decision maker. Defining that choice matrix reveals a structure of freedoms, capacities, and compulsions: i.e., the class structure. This structure defines relations of power; exploitation becomes a possibility missing from a choice frame of neoclassical economics. Not everything is a choice. Eric Olin Wright concisely defines exploitation as the antagonistic interdependence of material interests among actors within a set of economic relations. Exploitation of a commodity — labor power — is inextricably linked to exploitation of a person possessing — and forced to sell — that commodity. The purchaser of that commodity must somehow realize more from its use than has been paid for it; this is the systemic imperative facing capital. Class theory then deflects policy from Gandhi to Ambedkar: justice is not a question of reforming the hearts and minds of propertied people, but rather a question of reducing the dependency and destitution that subject those without property to abject subordination, including the rich tapestry of humiliation and degradation implied by “caste.” Obscuring class analysis contributes to thought and policy harmful to those least capable of pursuing their interests via either state or market. Has class analysis declined because of some inherent weakness in explaining either differentiation of life chances under economic change or political responses to inequality? If so, what is class analysis missing? In our effort to recover class, we ask two central questions. First, what is lost in the move away from class? Second, if the move had to do with weaknesses of class analytics to address reality on the ground, what does experience in the subcontinent tell us about needed refinements and limits of class theory? What would it take to recover class? Whatever Happened to Class? It seems especially odd that so obvious a tool for explaining differentiation of life chances, as well as political dynamics that have so often historically developed from moral outrage at inequality, should wither in contemporary South Asia. Some obfuscation is transparently instrumental: it is in the interest of winners in booming economies — and regimes seeking credit for growth — to emHerring and Agarwala / Introduction 325 6. Elster 1985, 331. 7. Elster 1985, 325. 8. Wright 1997, 10. phasize aggregate gains, not class divisions. The campaigns and administrations of George W. Bush in the United States strategically branded any critique of redistribution of income to the already wealthy as “inciting class war.” The 2004 elections in India found the Government claiming success for an aggregate “India Shining”; opposition parties tried to disaggregate the picture into winners and losers in globalization. In the dominant legitimation of global market integration, functionalist accounts held suppression of labor to be necessary for competitiveness. Class divisions were viewed as dysfunctional for national success. Market triumphalism itself was reinforced by the demise of Soviet-style dictatorships; in a rather bizarre non sequitur, Marx was somehow rendered quaint thereby, and, with Marx, the question of class. Claims of rigid autocracies to be “state socialist,” though empty by any Marxian criterion, had married repressive politics and economic disaster to nominally Marxian analysis, tainting the essentially critical nature of the latter with the oppression and inefficiency of the former. Yet, writing on Eastern Europe, David Ost concludes that in the demise of centrally planned economies, class analysis is ironically rejuvenated: “retiring class with communism...just will not do.” With the transition to market society, familiar class cleavages take center stage in Eastern Europe. In India, a neoliberal consensus rejected much of the aspirational Nehruvian project of a “socialistic pattern of society,” and with it the centrality of redressing class dominance. Over time, public discourse turned to empty aggregates without relational content: “the weaker sectors,” “poverty,” and other obfuscations replaced the acute class analysis of early nation-building. In the realm of ideas, global intellectual dispensations have militated against class analysis. At one end, idealist constructivism abandoned the very gritty empirical work that class analytics demands. In constructivist accounts, epistemological relativism challenges — or denigrates as simplistic — empirical science. Science itself, which claims no terrain for itself, becomes “Western science” or “imperialist science.” Science itself has become “an enemy of the people” in critiques of some activist intellectuals. Vivek Chibber calls this aggregate dispensation “post-structuralist/postcolonialist,” but it is hard to pin down; part of the dispensation is precisely the denial of an operational definition. Discursive moves to more micro levels (the body as site of contestation, the subject, the in326 Critical Asian Studies 38:4 (2006) 9. Paul Krugman (2006, A19) noted: “So what’s our bitter partisan divide really about? In two words: class warfare.” For an argument that politics in the United States follows class coalitions that either fatten the rich or attempt mild distribution, see Phillips 1990. 10. Ost 2005, 19. 11. See Corbridge and Harriss 2000, 143–72, and passim; Herring and Mohan 2001. 12. On early analyses of how India’s urban workers’ movements reflected the strength of the nation’s democracy, see Fisher 1961; Kennedy 1958. On early official agrarian class diagnosis for policy in India, see Herring and Edwards l983, 17–49; 153–79; Herring 1983, 17–49; 153–79. 13. Chibber 2006. On the assault on science, see Nanda 2003. On the postmodern move in history in particular, see Richard Eaton’s “(Re)imag(in)ing Otherness: dividual) or more macro (globalization, modernity) both exclude precise specification of class actors and the problematic of collective action: under what conditions do individuals with similar interests unite to promote common goals? At the other end of the intellectual spectrum, positivist and formal-theoretic turns in the social sciences have obscured class. First, the most extreme forms of positivism privilege the measurable over the real, even if available measurements poorly tap the concept one requires and data are acknowledged to be deficient. Second, the implantation of U.S.-style positivism has emphasized measurable stratification over less easily measured relational variables. Social relations, central to class analytics, give way to easier-to-measure proxies, such as income and education, which emphasize comparison among status groups, not interaction. Finally, the aggregationist language of developmentalism has privileged wholes over parts. “Strengthening civil society” replaces improvement in income distribution or agrarian reform. GDP and change therein become operative measures of success, whatever the distributive consequences. Place and context became less important to explanatory models in neoliberal prescriptions of the “Washington consensus,” where one medicine famously treats all diseases, regardless of cause. The move away from class has been supported by apparent failures of classbased organizations and political parties. Even if true, failures would be no indictment of class analysis, any more than any particular recession indicts macroeconomics. Instead, class analytics points to the obvious comparative question: why do class-based parties and organizations succeed electorally in Bengal to a remarkable extent, but fail in Gujarat — or Pakistan? Why did social democracy succeed in Europe but there is famously “no socialism in the United States?” Theorists of social movements have found, not surprisingly, that strategy, success, and failure depend a great deal on the political opportunity structure faced by movements. What niches are available for mobilization? How open is the system to new actors? What allies are available, and what resources do they have? How likely and effective is repression? Historically, expressing left politics in Pakistan could get one killed; the same behavior in Kerala or Bengal could lead to a comfortable career. Political party systems then determine what interests can be mobilized, with what effect. With the decline of the “Congress system” in India, the left in Sri Lanka, and the absence of a competitive party system in Pakistan, the rise of nonprogrammatic parties characterized the subcontiHerring and Agarwala / Introduction 327 A Postmortem for the Postmodern in India,” in Eaton 2000; see also Hutnyk 2003. 14. Drèze and Sen 1996, Chap. 1; Harriss 2001. 15. Stiglitz 2002. 16. Bardhan 2001; Bardhan and Rudra 2003; Guha 2001. See also, Chibber 2003 on left failure politically. 17. Sombart 1976. 18. For example, see Tilly and Tarrow 2006, 57–67. 19. Even under the “Islamic socialist” regime of Zulfikar Ali Bhutto. See Candland 2007. nent. Nonprogrammatic parties have great latitude in adopting symbols to mobilize support, deploying social identities rather than class-relevant programs. Within India, disintegration of the Congress system made for a politics of opportunistic alliances that rendered all political parties less programmatic and simultaneously less relevant to redistributive politics. Under these systemic conditions, the possibilities for obfuscation of interests via alternative identities ramify: nation, region, caste, community. Indeed, the mass political energy of the past two decades in the subcontinent has seemed to be more accurately represented by “identity politics” (struggles of women, “tribals” (adivasi), and dalits [oppressed]). Gruesome communal conflict suggested the extreme salience of an identity politics that overwhelmed all interest-based accounts of political action. Though sophisticated analysts recognized the interpenetration of class and identity, class was largely submerged in favor of identity. This was an odd outcome; class locations have always been inhabited by people with multiple identities, and must necessarily remain so. Moreover, as identity politics was sweeping academic treatments in the 1980s, a strong leftist coalition led by the Communist Party of India-Marxist (CPI-M) was consolidating what has turned out to be the longest running democratic government in India — and one of the longest in the world: in West Bengal, a state of 82 million people, now twenty-eight years and counting. No one denies the significance of identity politics, but privileging identity over class raises three analytical problems. First, attributing political behavior to identity too often takes constructions of political entrepreneurs at face value, thus obscuring material interests behind a claim of ascriptive solidarity. Second, identities exhibit the same explanatory ambiguity in relation to politics and collective action as do classes. Which of the competing identities do individuals choose as a basis for collective action? For this very reason, constructivist idealism is attractive as academic practice and not very useful as social science. Finally, the either-or formulation of identity vs. class deflects attention from the additive dimensions of inequality. Class produces divisions in ascriptive identities that reduce potential for collective action (rich women and poor women, for example), just as ascriptive identities may divide — or in rare cases — solidify potential solidarities of class (for example, anti-brahmin movements in South India historically united disparate castes horizontally against privilege). No robust theory answers these questions, in part because of the necessarily overlapping and additive effects of hierarchy in social systems. Class mobilization is also increasingly in competition with civil society organizations. Nongovernmental organizations (NGOs) claim to represent single-issue topics (such as the environment or poverty) or constituencies defined by ascriptive characteristics (women, tribals, children). Hindu, Christian, and Mus328 Critical Asian Studies 38:4 (2006) 20. John 2005; Basu and Kohli 1998. 21. Gail Omvedt’s Reinventing revolution (1993) is tellingly subtitled “New Social Movements and the Socialist Tradition in India.” On material interests in ethnic politics, see Herring and Esman 2001; Chandra 2004. lim fundamentalist organizations the world over provide support for the NGO sphere, as do development organizations. Class-based organizations receive little funding and media attention relative to the exploding array of “grassroots organizations.” Women’s micro-finance organizations based in the slums of Bombay, Bogota, or New York City use fashionable policy prescriptions of individual behavior and self-sufficiency to complement neoliberal cuts in public welfare support. This template is sufficiently attractive that Citibank, Goldman Sachs, HSBC, Standard Chartered, ABN Amro, Deutsche Bank, and Morgan Stanley have all begun to join the micro-finance effort. Building “social capital” is more attractive to capital — which can influence these flows — than organizing workers for enforcement of health and safety standards or minimum wage statutes. As John Harriss shows, multiple sources of material support for NGOs are augmented by a tendency of middle-class Indians with concerns for activism and social change to think of all politics as a “dirty river,” reinforcing the ranks of “grassroots” organizations with talent and energy. 22 Though the NGO phenomenon has blunted and complicated class organization, class analysis may aid in understanding new developments. Harriss shows that different classes participate in activist organizations in different ways. Moreover, he finds that part of “being middle class” in India now often includes social activism. Being middle class facilitates activism both through the cultural capital of that class and the freedom from the dull compulsion of economic necessity that hems in other classes. 23 Ronald Herring finds the same to be true of mobilization against new technology in agriculture; Vandana Shiva can afford to label Bt cotton seeds “suicidal,” then “homicidal,” and finally in 2006 “genocidal” because she is free from dependence on cotton production for her livelihood. Cotton farmers, who lack this freedom, but face compelling economic pressure, experiment with Bt seeds, and in the aggregate adopt them rapidly. “Operation Cremate Monsanto” failed in part because activists misunderstood class interests in biotechnology, its property configuration and relations: particularly the capacity of farmers to appropriate the technology under the radar screen of both Monsanto and Delhi. An approach that analyzed this movement without reference to class, property, and the cognitive screen of science would lose explanatory power; as important, the failure of movement leaders to take a class perspective hindered their ability to represent the class they claimed. Class analytics has also been sidelined by the rise of alternative intellectual models. Gary Becker won the Nobel Prize for his Economic Approach to Human Behavior in 1992; the recognition was indicative of the pervasive spread of rational-choice theory into social science disciplines to explain a wide range of behavior and relations, including marriage, education, immigration, families, crime, and elections. The promise of methodological individualism to provide robust and parsimonious explanations for long-debated and perennially mudHerring and Agarwala / Introduction 329 22. Harriss 2006a. 23. Ibid. 24. Herring 2006. dled phenomena held great attraction. As Jon Elster eloquently demonstrates, rational choice theory and class-based analysis are not mutually exclusive. There can after all be no choice outside some structure of constraints and payoffs. Though rational-choice theory and class analytics are often pitted against one another, synergies abound. The healthy contribution of rational-choice theory to South Asian studies is its rejection of Orientalist assumptions about difference; the imprinted cultural other becomes instead an agent making decisions within (often binding, and certainly class-differentiated) constraints. Methodological individualism in turn becomes robust only through specification of structures that bind, guide, and differentially reward choice. Methodological positivism, unless tempered by a realist philosophy of science, has a difficult time with relations as opposed to discrete values of a variable. There are few good proxies for relational constructs. In practice, the enumerated takes precedence over the theoretically important or empirically significant: class is but one example. Moreover, a purely positivist dispensation — precisely because it takes the world as presented by existing indicators and has no method to probe deeper — often fails to recognize that data are themselves products of social interaction, their relation to reality varying with conditions of their production. For want of precise indicators of class, proxies are deployed as both independent and dependent variables: income, education, and skill level. Stratification studies operationalize hierarchy by means of easily measurable, independent categories with no attention to interaction among status groups. Stratification theory treats a relation of comparison, not interaction. Postmodernist theory from the humanities undermined class analysis through rejection of both causal theory based on demonstrable mechanisms — the core of class analytics — and empirical referents as a measure of the truth value of statements of fact — the core of positivism. This literature resisted attempts to distinguish and define specific social categories according to transparent and consistent criteria. Positivism and class analytics can share a platform, though it is understood that there will be arguments about appropriate proxy measures, about valid indicators, about the priority of naked data over multidimensional confirmation of empirical statements. This is a nonantagonistic contradiction. The same meeting grounds of ontology and epistemology 330 Critical Asian Studies 38:4 (2006) 25. Becker 1976. 26. Elster 1985. 27. Wade (1988) offers an example of how careful attention to material forces and rational decisions by farmers explain some interesting puzzles in water and land control in South India, to his satisfaction, controlling for cultural variables. 28. Herring 2003b. 29. See Elster 1985, 335–41 for a detailed discussion on this point. There is a parallel to gender studies: tables showing numbers of males and females in various occupations do not tell us much about gendered relationships on the job or in society generally. 30. An excellent treatment of the “qualitative-quantitative” divide, and synthetic methods, is Kanbur 2003. are not readily apparent between constructivist idealism and class analytics, with arguably deleterious effects on progressive scholarship. What Explanatory Work Do We Ask Class To Do? Too much has been expected of class theory. The hubris of mono-causal grand theory is not limited to class; social science feeds on and from long cycles of master narratives. Nevertheless, bold claims to a universalist framework making strong predictions rendered class theory uniquely vulnerable. Overreaching political prediction from class theory especially undermined the scientific aspirations of the enterprise: why are there no revolutions among workers? Why is voting only imperfectly correlated with class? Dueling orthodoxies and partisan theoretical product differentiation have exacerbated these critiques. . The authors in this collection agree fundamentally with Eric Olin Wright’s observation: “Class analysis is based on a conviction that class is a pervasive social cause, and it is worth exploring its ramifications for many social phenomena. This also involves understanding the limits of what class can explain.” If we expect the framework to explain a broad range of phenomena, four discrete components are essential: class structure (class-in-itself for Marx), class consciousness (understanding by individual actors of their class interests), class formation (collectively organized actors of similar structural position, constituting a class-for-itself in Marx), and class struggle (collective practices of actors for the realization of class interests against interests of other classes). Specifying how class structure interacts with societal development is now a more complicated task than in the time of Karl Marx, or even Max Weber. Few doubt that ownership and control of property explain changing forms and degrees of inequality and common political responses over historical time: there would be no social democracy in Western Europe otherwise. It is one thing — and a valuable, too often slighted, thing — to take class into account in explaining different life chances during economic change. It is much more demanding to expect class — or any notion of social structure — to predict or explain politics that drive policy. Class structure defines positions for individuals, based on their relationship to economic assets; these class positions in turn differentiate objective material interests: landlords and tenants, workers and owners. Under certain conditions, these interests may be recognized, mobilized, and acted upon — thus ultimately explaining collective action in which people attempt to improve their life chances through politics and policy: land reform, minimum wages, welfare transfers, income redistribution. Though it is true that ownership of the means of production cleaves a fundamental division in society, it seems equally clear that two giant classes — bourHerring and Agarwala / Introduction 331 31. Chibber 2006. 32. Wright 1997, 1. 33. On useful distinctions among these concepts, though somewhat different from the text, see Wright 1997, chaps. 1, 2, 10. geois and proletarian — are inadequate for understanding politics of contemporary class systems, especially in semi-agrarian nations. In India, for example, a “new middle class” evokes intensive interest. However much or little physical capital members of this class may control, they bring important forms of cultural capital to the marketplace. Yet most work for bosses, many have servants, some face extreme insecurity and financial pressure; others not only aspire to, but live, a globally cosmopolitan life. How finely does one divide such a stratum? Or should we think of it, as Fernandes and Heller argue, a genuine class in the making, a class tentatively, unevenly for itself? 34 This multi-jointed path from class structure to class consciousness to class formation to politically efficacious class politics immediately suggests limits and contingencies of class theory. Interests are not always transparent, to individuals or observers. The paternalistic attribution of “false consciousness” is currently out of favor, but it is hard to accept the romanticization that subordinates always see through the ideologies of superordinates — i.e., are not mystified. All interests are necessarily filtered through cognitive screens. What is the workers’ interest in monetary policy and exchange rates, WTO rulings, and Trade-Related Intellectual Property Rights (TRIPs)? Does the answer vary by sector? Over what time period? Cognitive screens process “interests” such that they can be recognized, evaluated, given meaning, ordered, and thus rendered actionable by individuals. These screens necessitate modesty in reading claims of interests off structures, as much classical class theory did. Some workers believe party-affiliated unions to be an appropriate vehicle for guarding their interests, others in the same structural location think not — for many reasons. Informal laborers, who are increasingly doing the same work as formally employed laborers, are often willing to accept insecure jobs for wages that are far below the hard-fought minimum wages in return for relatively small welfare benefits from the state. A deep irony of “farmers’ movements” for higher agricultural prices is that very poor farmers and landless workers often join in, though the net effect of higher prices is a heavier burden on food-deficit households — i.e., virtually all of the rural poor. Thomas Frank puzzles over U.S. working-class support for economic policies that are not in their interest: What’s the Matter With Kansas? His answer is that their attention has been effectively shifted to “moral” issues that create validation and solidarity, but distract from the reality of income redistribution to the very rich at the expense of the average worker. Through the 332 Critical Asian Studies 38:4 (2006) 34. Fernandes and Heller 2006. 35. Wright 1997. 36. Scott (1985) comes very close to this position. Much of the resurrection and celebration of “local knowledge” likewise assumed inherently superior Volk information and wisdom. 37. See Tietelbaum 2006 for the worker’s dilemma in choosing between well-institutionalized (perhaps co-opted) unions and very militant (perhaps counterproductive) unions. 38. Agarwala 2006. 39. Frank 2004. lens of filth on television, burning flags, abortion rights, and gay marriage, political entrepreneurs of the right have been able to change the subject in U.S. politics. Political strategy and framing thus confound any direct deduction of class interests; good rational choice theorists understand the problem, as did Marx. With technical and economic integration characteristic of globalization, scientific knowledge becomes asymmetrically distributed, both globally and within social movements, between leaders and those represented. Different cognitive screens cause divergent perception of interests in such critical science-embedded problematics as climate change or genetic engineering. Biotechnology offers a particularly contentious case, as many NGOs present the interests of farmers as threatened even as farmers themselves find the technology to be so much in their interest that an underground market in transgenic seeds develops. Cognitive screens are not given by class position, but are conditioned in important ways: activists are free to use junk science for dramaturgical effect, and indeed agitational politics creates selection pressures for junk science. More generally, political ecology produces interest complexities unfamiliar to longestablished class routines, from landscapes to genomes. Nature presents special problems for analysis of interests: interests of many primary producers are embedded in local nature, yet are contingent on dynamics of larger biophysical systems nowhere fully understood. Moreover, state property in nature dominates private property in nature; Nancy Peluso calls the results “secret wars and silent insurgency” inimical to conservation. As in the case of informal sector workers, demands are often targeted on the state rather than capital, yet property is still the crux of conflict. Variable class structures also affect political ecology outcomes. Ramchandra Guha notes that the relative absence of class inequality was a necessary condition for both conservation and collective protest in Uttarakhand forest mobilization. Some objective interests are then both difficult to understand and subject to alternative cognitive screens, or framings. Not all who are objectively members of a class may recognize that position; not all who recognize their location in a Herring and Agarwala / Introduction 333 40. Jon Elster (1985) produced the most explicit attempt to meld these traditions. 41. Roy, Herring, and Geisler 2007; Herring l985. With the divergences in local knowledge of agriculture in north India documented by Akhil Gupta (1998), one sees the complexity of farmers’ dilemmas in sorting means to ends in agriculture. 42. Herring 2006. 43. Gadgil and Guha (1995) propose an alternative class system for ecology, based on relation to nature and consumption. For a more complicated look at the interests of their “eco-system people,” see Baviskar 1995. On state property, see Herring 2002. 44. For a powerful case for the importance of narratives of environmental change and policy in differentiating interests of individuals, see Dryzek 1997. As evidentiary rules differ across paradigms, closure is cognitively difficult. 45. Peluso l992, 235–36. 46. Agarwala 2006. 47. Guha 1983, 33; see also Baviskar 2005. class structure will find that particular dimension of inequality most salient, or most amenable to change; not all who seek to alter the terms of their class position will find sufficient colleagues to make collective action feasible; and not all class-based collective action will be effective: much will be suppressed, bought off, tactically flawed, or ignored by political actors with alternative support bases. Empirical evidence of unrecognized or mis-recognized interests leads logically to cultural, psychological, and situational variables that intercede between class location and behavior. James Scott’s great accomplishment in Weapons of the Weak was to work through empirically and theoretically the limitations on class collective action in a particular village in Malaysia at a particular point in time. Class anger and envy remained offstage, deflected to alternative channels, few of which enhanced interests of the poor materially. Yet Scott’s celebration of Malaysian farmers’ ability to penetrate hegemonic ideologies does not resonate with patterns of self-defeating behavior in many other settings — such as Kansas. Indian Exceptionalism: Isn’t Class Irrelevant Because of Caste? In Indian studies, informed by an Orientalist focus on the exotic nature of otherness, religio-cultural overlays on class stratification have understandably loomed large. Moreover, economic class has always seemed too simplistic and materialist a concept for the richly stratified layers and dimensions of privilege and deprivation in Indic society. Rather than class, every student learned, “caste” dominated. On reflection, we find this a puzzling construction. First, economic compulsion and social subordination in the caste system tended to correlate: low or “untouchable” (avarna) status almost always meant property-less existence and degrading labor. All class systems valorize behaviors and characteristics of super-ordinates, denigrate the language, dress, manners, living conditions, and mores of subordinates. All class systems, to greater or lesser extent, structure interaction among classes: whom one lives among, eats with, goes to school with, marries. In India, distinctions are explicitly codified in caste-based norms — though the reification and timelessness of caste was certainly influenced by colonial rule. The word “caste” itself is from the Portuguese casta suggesting “race, lineage, breed” — among animals as well as people; it connoted “pure or unmixed (stock or breed).” Pure or unmixed derives from the word’s origins in Latin: castus meaning pure or unpolluted, from which English also derives chaste. Darwin used the term for classes among social insects, such as ants. Before 1800, the spelling in English was “cast,” illustrating more clearly its deriva334 Critical Asian Studies 38:4 (2006) 48. Scott 1985. 49. India’s first Backward Classes Commission, appointed in 1953, offered empirical evidence of the well-understood social reality: inferior position in the caste hierarchy was the main determinant of social and economic backwardness. 50. Dirks 2001, esp. 43-60, on “the ethnographic state.” tion from the verb: i.e., “a throw or stroke of fortune; hence, fortune, chance, opportunity; lot, fate.” From the beginning, English-language usage and colonial practice reinforced Hindu ideology’s focus on fate, purity, heredity: one is cast into a class position, from which movement is sanctioned. Members of the avarna or “untouchable” castes in South India were defined by their lack of land and their denigrating work as serf castes or slave castes even into the twentieth century; socially, they were forbidden to cover some parts of their body with clothing, to use certain forms of address and grammar, to travel public roads, to enter temples. Class disabilities and distinctions were mapped onto social clusters glossed ideologically as caste: i.e., a class location for which there is a reason. Indology makes much of the rich particularity of caste’s extra-economic distinctions, yet an argument for uniqueness is hard to sustain. Shakespeare was by law forbidden to wear the fine clothes reserved for gentry when he left his own stage, where he had portrayed characters of higher classes than his own — to which he aspired by pursuing his father’s failed dream of acquiring a coat of arms. His caste was hereditary, but accident of birth could be corrected with cash — suggesting how little purity of birth line had to do with it in the first place. Shakespeare was told that his family was so cast that it fell below the gentleman class, but he knew there were ways to turn resources into mobility, despite accident of birth. Where race differs from hereditary caste is the difficulty of passing, or erasing, evidence of birth. Increasingly, South Asians flock to cities where losing heredity is much easier, easier to shed in fact than the disabilities of class. One can disguise village origins, but cannot pass as capitalist without capital. Indian exceptionalism in social science has often used caste as exhibit one; but the dichotomy is false. The challenge for Indian class analytics is to demonstrate how extra-economic aspects of class location — e.g., caste — aid in understanding political responses. Dimensions of class subordination are elaborately codified — how many steps must the “untouchable” stay from “clean” castes? What body parts may be covered? These dimensions of subordination affect the freedom and dignity of particular class locations, as well as for efficiency of labor markets and educational capital. Deprivation of freedom and dignity in turn may well fuel moral outrage more than mere economic deprivation. Moreover, the sociological reality of caste groupings may affect potential for collective action. Common deprivations and common interests generated by stratification are conjoined with a social basis of organization: caste (jati) is ultimately a local phenomenon, with propinquity of members, marriage connections, and authority. Castes thus may exhibit more potential for political solidarity than do Herring and Agarwala / Introduction 335 51. The text uses “caste,” as is commonly done, to mean jati. All etymological derivations based on the Oxford English Dictionary. 52. For example, Saradamoni 1980. For a treatment of political implications in Palakkad District, see Herring 2001b; more broadly, Herring 1988. 53. Greenblatt 2004, chap. 2. classes of the potatoes-in-a-sack variety. On the other hand, subtle caste distinctions and competition among subordinates in similar class positions also generate obstacles to horizontal solidarity, much as race or ethnicity divides subordinate classes in many societies. Finally, the ideological underpinnings of caste almost perfectly explain away exploitation: one’s station in life is justified by one’s previous conformity with dharmic law, which enjoins adherence to dharmic law, which means accepting subordination as both inevitable and justifiable. Assuming for a moment that caste is more than the “euphemization of class,” the common assumption that caste is the bedrock of political behavior is seldom tested and increasingly problematic. Anirudh Krishna concludes from his surveys in North Indian villages: “Caste continues to be a primary source of social identity in these villages, people live in caste-specific neighborhoods, and the clothes that they wear reveal their caste identity. Yet insofar as political organization is concerned, caste no longer has primary importance.” Krishna finds that primordial loyalties are very much still in place, but that patronage and material advantage figure prominently in political connections, as Kanchan Chandra’s work on castes in Uttar Pradesh likewise illustrates. Class analysis sensitive to caste social embeddings has proved an indispensable conceptual tool in explaining differences in redistributive policy in India, as indicated by such measures as poverty reduction or agrarian reform, and resultant differentials in life chances at the bottom across Indian States. Barrington Moore Jr., in his classic Social Origins of Dictatorship and Democracy, posed a challenge to Indologists: given the extensive misery at the bottom of Indic society, why have there been so few effective radical challenges to the social order — particularly in comparison to China? The answer is complicated, but must in part ride on politically critical variations in caste-class relations of dominance. Moral outrage at the injustice of hierarchy in some cases drove ameliorative politics; in the long development of left movements and parties in Kerala — class analytics provide an explanation for exceptional enhancements in quality of life and protection of the weakest classes through public policy. More commonly, dominance prevented or co-opted lower-order mobilization. Variations in political opportunity structure and party systems have mattered fundamentally. Atul Kohli’s early work contrasted Karnataka with West Bengal to illustrate how a committed and disciplined left-of-center party could effect poverty alleviation in ways not possible in states lacking this political resource. John Harriss complements Kohli: as important as electoral politics is a decisive break of “caste-class dominance.” Successful redistributive parties nurture and 336 Critical Asian Studies 38:4 (2006) 54. Krishna 2003, 1071; Chandra 2004. 55. One comparative study that presents the quantitative case carefully but goes beyond numbers to social relations and political history is the excellent work by Jean Drèze and Amartya Sen (1996). 56. Frankel and Rao 1989; Kohli 1987; Harriss 2003. 57. Ramachandran 1996; Heller 1999, 51–117; Herring 1988; Herring Forthcoming a. 58. Kohli 1987. reinforce this decisive break, spurring and building on horizontal mobilization in civil society. Such mobilization then energizes reforms: uncovering malfeasance, pressing for timely action, intimidating colluding bureaucrats. The conclusion is that successful redistribution requires a break in both caste and class dominance. Kerala’s early success in poverty reduction unambiguously developed from successful political mobilization of subordinate classes. Congress hegemony was replaced by a distributive anti-elite populism in Tamil Nadu, and by aggressive left coalitions in Kerala (first) and West Bengal (more enduringly). As importantly, in Kerala the Congress and its permutations responded to electoral competition with left forces, particularly on land policy, making the party less conservative than most Congress state units in India. An alternative pattern was produced in States with middling-caste/class-dominated regimes — such as Maharashtra, Karnataka, and Gujarat — where Congress was effectively challenged but did not collapse. In these states, the politics of accommodation vis-à-vis lower-class interests has worked effectively, especially in Maharashtra and Karnataka, and less so in Gujarat. There has been no decisive rejection of caste-class dominance of the old social order through politics and public policy, and redistributive policies have not been robust. In states where caste-class dominance has survived unchallenged, poverty alleviation has been correspondingly weak. What do we learn from these divergent patterns of subnational societal development? First, successful class mobilization in India deviated from European orthodoxy in favor of organic class theory: grown up from the ground, based on political experience. The move was from urban proletarians of received theory to coalitions of the despised, insecure, and less-privileged. Any Marxian notion of bimodal classes was replaced in practice by extensive differentiation of targets of mobilization, building on locally understood forms of exploitation. And Herring and Agarwala / Introduction 337 59. John Harriss (2003) finds that a decisive defeat of the dominant class/caste social coalition has occurred in Kerala, Tamil Nadu, and West Bengal. In these states, the umbrella accommodationist Congress Party lost its dominance early. Congress hegemony was replaced by anti-Brahmin populism in Tamil Nadu, and by aggressive Left coalitions in Kerala (first) and West Bengal (more enduringly). Despite the differing organizing principles used, namely, caste in Tamil Nadu and class in Kerala and West Bengal, all three state parties initiated successful redistributive policies. See also Frankel and Rao 1989. 60. Ramachandran 1996, 180–200; Herring 1988. Mass energy in social movements was organized as the Congress Socialist Party [Communist Party after 1942], which eclipsed conservative Congress elements by sheer weight of numbers and activity. 61. This understanding is not restricted to the subcontinent. If one were to ask, for example, why there is no working-class party in the United States, and thus no social democracy of the European variety (Sombart l976), one would have to consider ethnic tensions generated from waves of immigration, which created intense competition for jobs and deflected anger to class-fellows of different ethnicity. These dynamics help explain the comparative absence of class consciousness or class formation that make the United States anomalous among industrialized societies. finally, multiple dimensions of humiliation and oppression embedded in caste dimensions of class mattered politically. Piling on multiple dimensions of inequality — and not simply terms of the commodity exchange — fueled challenges to the class-caste system. Caste theory provides a useful reminder for materialists: good class analysis is inherently multidimensional. Living bearers of a commodity — labor power — are subjected to the impersonal forces of market capitalism that treats them as the commodity — which may or may not be in sufficient demand to earn a decent living, or provide dignity and security to individuals. Exploitation of a commodity — labor power — is inextricably linked to exploitation of a person possessing that commodity. Exploitation between classes then describes and enables a system of power, not just one of privilege, rooted in unequal access to economic assets. “Moral outrage” is more explainable in the multidimensional world of oppression and dominance than in the discrete step world of stratification. Theories that reduce inequality to mere stratification miss this relational component; by doing so, they de-emphasize dynamics of domination, power, and exploitation. Lumpers and Splitters of Classes How many classes are there? Where are the boundaries? As often, there are lumpers and splitters. Karl Marx boldly lumped layers of stratification together into two classes: bourgeois and proletarian. Dividing society into two megagroups (the exploiter and the exploited) creates “classes” of great heterogeneity but periodic subjective reality — nosotros los pobres; hum garib log. If giant classes are to act politically, coalitions across subgroups must be hammered out. Modern stratification theory tends to produce dimensions that can be represented by a continuous variable — e.g., income, education — creating an infinite number of classes, or none at all. Some literature divides the population into a very large number of fine-graded categories based on occupation, in the name of “class.” The analytical problem is real. David Grusky and Jesper Sorensen raised the question starkly with their 1998 article, “Can Class Analysis Be Salvaged?” The “big class assumption” — that there are a small number of big classes, generated by the forces of industrialism — is one held primarily by academics and rarely by class members themselves. Instead, Grusky and Sorensen argue that structure 338 Critical Asian Studies 38:4 (2006) 62. In the New York Times series entitled “Class Matters,” mobility is connected to cultural capital in the United States. See Correspondents 2005. 63. Karl Polanyi is perhaps the most noted theorist to insist on the unnatural nature of the process of commodification of human beings as labor. Responses to insecurities and indignities of commoditized labor in market society — child labor, for example, or discarding of the elderly when their labor has no market value — drive his vision of social policy that hems in the market. See Polanyi 1957. 64. Wright 1997a. 65. Grusky and Sorensen 1998. at the site of production is comprised of much smaller classes defined by functional positions in the division of labor. Grusky and Sorensen argue that these institutionalized occupation groups have greater explanatory power than the “big classes” in terms of collective action, as well as group identification, interests, and culture. This perspective mirrors common perceptions of caste: an exceptionally finely graded hierarchy of occupation and ritual ranking. But if identities and occupational interests become infinitely divisible, why not speak merely of occupations, dropping altogether the term “class”? At the other end of the spectrum, Pranab Bardhan argued that India was controlled by three “dominant proprietary classes,” in an “uneasy alliance” that maintained power by sharing out spoils of patronage and subsidies through log-rolling across issue areas. Two of Bardhan’s classes were familiar to students of both class analysis and of India: landlords and industrial capitalists. But the third — the “professional class” — was more problematic. In one sense Bardhan’s professional class owned the state, and the incalculable rents appropriable from command of office. Bardhan attributed the power of this class to “the scarcity value of education” — a value reproduced over time by the state’s failure to democratize schooling. Yet it would seem today that the “scarcity value of education” is too narrow for all the advantages of that messy but real category of a “new middle class” in India. Pierre Bourdieu popularized the use of a concept of cultural capital that is broader than education, though certainly congruent with command of knowledge and intellectual skills: slippery assets such as attitudes, status, expectations. We would add connections. Likewise, it is clear that command over people and resources now has a broader ambit than ownership of positions in the license-permit-quota raj (rule) of India’s pre-liberalization era. One new set of class positions is generated by the intersection of the international politics of globalization and the cultural capital of sectors of the metropolitan middle class: the NGOs, consultants, and global activists who are disproportionately important in driving media and public policy in India. Lumping and splitting stratification systems in India offers almost infinite possibilities. For Bardhan, the proof of the theory revolves around power in shaping development policy and state discretion: who gets what and how? This is the core developmental question. As the developmental state assumed central importance in driving economic change and mitigating consequences, Bardhan argued that it was necessary to separate those who essentially owned the state from those without access to the state. Is the middle-class activism in civil society emphasized by Harriss, and the retreat of middle classes from politics, predicated on the declining importance of the state in a neoliberal political economy? Herring and Agarwala / Introduction 339 66. Portes 2000; Wright 1997b. 67. Bardhan 1984. 68. Fernandes and Heller 2006; Harriss 2006a. 69. Bourdieu and Jean-Claude Passeron first used the term in Cultural Reproduction and Social Reproduction in 1973. On the concept and applications to India, see Fernandes and Heller 2006. An illustration of the power of this insight comes through the prize-winning work of Aseema Sinha. Sinha asks the profound question: if economic interventionist policies from Delhi were so bad for growth, as mainstream economists, international financial institutions, and many citizens of India have concluded, why is there so much variance among subnational states within India? The national average growth rate was low over the period l947–1985, but some states grew rapidly and experienced significant structural transformation; others stagnated or experienced structural retrogression. One conclusion is that the developmental state literature is right, but at the wrong level. In a large federal nation, developmental state dynamics happen below the level of Delhi (and scholars working at that level). Gujarat does what much of the developmental-state literature suggests is important for promoting growth, and it works. One common reading of the original developmental-state theory based on Japan (Chalmers Johnson’s original work) was that the special genius of the Japanese state was to ascertain what capital needed to grow and then to do it. This is the structural power of capital: every state must promote “business confidence,” whether or not particular capitalists are politically active or not. India’s “license-permit-quota raj” had done quite the opposite: discouraged capital in general while treating well-connected capitalists with boons through embedded particularism. This system was condemned for stifling growth, but without consideration of relations between particularities of the capital-state relationship, where the state resides in Gandhinagar, not Delhi. Capitalists must matter; it would be puzzling indeed if so powerful a class were helpless before mere bureaucrats or did not know what it needed, or how to get it. That relations between provincial-level states and a partially formed national capitalist class differ in so diverse a country as India is not surprising; that much mainstream literature misses this critical determinant of economic growth is. The Agrarian Question: Classes and Politics Rural India illustrates themes developed above: the nature of the class system was theoretically disputed and politically contested; lumpers and splitters found evidence for divergent interpretations; and multiple dimensions of degradation and privilege defined rural power relations. Moreover, these differences mattered to politics. Theorists, political entrepreneurs and activists have variously assumed identity, class, regional, and sectoral dimensions to be the key to rural mobilization. Debate around rural class relations in independent India grew from intimations of agrarian unrest: might the “green revolution turn red”? This possibility 340 Critical Asian Studies 38:4 (2006) 70. Sinha 2005. 71. See Woo 1999, Chap. 1; Herring and Mohan 2001. 72. Herring 1999. 73. Chibber puts capital in a more active political role vis-à-vis the state in 2003, especially Chaps. 5, 6, and 8. 74. This section owes much to John Harriss, who bears no responsibility for its flaws. seemed pressing in the wake of the Maoist-influenced Naxalite movement, and of widespread “agrarian tension” — as officially described in a report of the Ministry of Home Affairs in 1969. Might agrarian tensions be aggravated by the development of agrarian capitalism, spurred by the Government’s support of technological change and “building on the best” in rural areas? The great puzzle of the twentieth century for Marxists was that agrarian upheavals brought down governments and overturned societies, not working-class revolutions. India became the center of a global debate around models of peasant economies and what becomes of them under capitalism. One line of theorizing, derived from Lenin’s work The Development of Capitalism in Russia projected differentiation of peasant producers into distinct classes. In theory, the development of capitalism would eventually create a rural society of largely dichotomous and antagonistic classes: agricultural capitalists and an agrarian proletariat. Proletarianization would plausibly lead to labor-capital conflict, perhaps revolution, in rural areas. In direct opposition to the class polarization model was that of a “peasant economy,” derived largely from the work of the Russian economist A.V. Chayanov. In this model, household production has its own very specific economic characteristics, with economic drivers quite different from maximization of profit at the margin. Rather than being exploited by superior classes, peasant families engaged in “self-exploitation” — using family labor intensively at submarket rates of return. Self-exploitation in theory allowed the peasant household to survive in circumstances that would be irrational for a capitalist farm. Chayanovian logic suggested that rural household production could continue indefinitely despite the development of agrarian capitalism. Small farmers could reproduce themselves by exploiting their family labor, propped up unevenly by government aids in extremis (loan melas, debt cancellations, input subsidies, public works, etc). Ironically, stagnation and relatively stable agrarian structure was also the implication of a largely Marxian perspective on “semi-feudal” agrarian structure. Semi-feudal landlordism could survive indefinitely, enabled by abject dependence of labor on landlords, and hence great opportunities for exploitation. Landlords would not necessarily follow the incentives of market capitalism for investment and technical change in agriculture, because they could reap more surplus through domination of the peasantry by usury, unpaid labor (begar), extortionate rents, and starvation wages. Another alternative to development of thoroughly capitalist relations in agriculture was posed by Lloyd and Susanne Rudolph, who projected the emergence of powerful “bullock capitalists” rather than full-scale development of capitalist relations in agriculture. Herring and Agarwala / Introduction 341 75. Wolf 1969; Paige 1975. 76. For an excellent treatment of the interaction of ideology and structure in reference to this line of reasoning, see Harriss 1982. 77. Contesting theorists launched what became known as the “mode of production debate,” mainly but not entirely through the Economic and Political Weekly. See Harriss 1980; Thorner l982; Mukhia 1999; and Herring 1985. In retrospect, rural capitalism has generated no radically polarized class structure. There has generally been what Byres referred to as “partial proletarianisation”; small and marginal peasant producers have continued to reproduce themselves. Chayanovian self-exploitation is now joined by migration (both rural-rural and rural-urban, both short-term and long-term), associated remittances, and diversification of employment in rural areas outside agriculture to sustain small-scale farming. Marginal farmers have been partially sustained by subsidized institutional credit, as in most countries, though market criteria are now increasingly important in agricultural credit via neoliberal logic. Constantly shifting development “schemes” contribute as well — for example, micro-finance programs, rural public works — and by state welfare provision. Broader processes of economic change, much accelerated since liberalization in 1991, have altered the relationship between land ownership, power, and poverty. Diversification of rural livelihoods and the increased importance of nonagricultural employment among rural people, both locally and in distant places, has altered the relation between agrarian capital and rural labor. Though the “patron-client relationship” has been much romanticized in academic work, face-to-face contact over time between families is waning with differentiation of the rural economy and market-rational behavior. Where the incidence of rural poverty has declined, it has been because of purposive public policy or the tightening of labor markets, both of which may reduce abject dependence of labor on capital. Land ownership may still mean considerable wealth, and remains an asset for local status and power. Economic rationality drives much investment out of agriculture, toward education and activities with higher rates of return. In other cases, as the profitability of agriculture declined, for status distinction, landowners have moved out of agriculture. As ascriptive status converts less easily to political power, one finds emergence of a new generation of local leaders from among educated but often unemployed younger men. For all these reasons, the class power of “rich farmers” of higher ranked castes appears to have declined significantly. Collective action of farmers reached a high point in the l980s, built on multi-class, sometimes caste-based, mobilization around costs of production and prices of outputs. Farmers’ movements have been less significant since the early 1990s. Ironically, some of the largest collective actions in recent years have come in protest against Delhi’s banning of genetically engineered (Bt) cotton in 2001. Here the issue was, for a significant farmers’ movement, autonomy from the corrupt and costly interference of the 342 Critical Asian Studies 38:4 (2006) 78. Rudolph and Rudolph 1987. 79. Byres 1981; Harriss 1994; Harriss-White and Janakarajan 2004. 80. Bhaduri 1986; Harriss-White and Janakarajan 2004. 81. Gaiha 1996; Herring and Edwards l983; Harriss-White and Janakarajan 2004, 195–295. 82. Mayer 1996; Harriss 2006b on the former, and Gidwani 2000 on the latter. 83. Krishna 2002; 2003. 84. See Brass 1995; Lindberg 1995; Omvedt 1993; and Teitelbaum Forthcoming. state in agriculture, exemplified by the denial of access to knowledge farmers wanted — biotechnology — Delhi sought to “bottle up in the cities.” Ties of dependency of the rural poor in class relations have loosened, but their interests are only weakly articulated politically — except in regional pockets and in the leftist states of Kerala and West Bengal. There is abundant evidence for what Frankel and Rao described as “the decline of dominance” — referring to “the exercise of authority in society by groups who achieved socio-economic superiority and claimed legitimacy for their commands in terms of superior ritual status.” Democratic deepening has enabled much greater assertiveness among subordinate rural people. There are large regional differences; uprisings against Brahminical power preceded by decades mobilizations of particular “backward-caste” groups in Bihar and Uttar Pradesh especially, and of dalits. Globalization and accompanying neoliberal policies have increased pressures on the agricultural sector, but means of class formation remain illusive. One formulation has been a politically imagined super and inclusive class — i.e., agriculture as a sector, Bharat as a place. The Bharat-vs.-India formulation attempted cultural and economic synthesis of rural India as a class for itself, focused on rolling back “urban bias” in development policy. The framing of farmer movements — and the Bharat against India construction in general — failed not so much for reasons of primordial loyalties, as Ashutosh Varshney suggested, but for reasons of intra-sectoral differences of interest, conflicting economic ideologies, and factionalism. After the 1980s, the ability of this imagined aggregate rural class to influence prices and subsidies was diminished by the failure of political organization. Moreover, globalization produced a new focus of mobilization, to which farmer organizations responded in divergent ways, partly for difference in interests, partly for differences in ideology. For example, the Shetkari Sanghatana — one of the largest farmer organizations in India — took a pro-liberalization, pro-technology, anti-state program even to small farmers and landless laborers; other farmer organizations (e.g., Nanjundaswamy’s KRRS) prominently joined the anti-globalization and anti-technology forces. Whatever their program, most farmer organizations have been dominated by commercial and larger farmers. The aggregative class strategy — a rural class for itself, coalitional agrarianism for sectoral advantage — largely failed, as did efforts to organize “classes within the peasantry” — again with a partial exception of India’s “red belt.” Kerala’s Left organized agricultural workers as a class, not as a subset of an organization of farmers who pay their wages. West Bengal’s Left preferred the national kisan Herring and Agarwala / Introduction 343 85. Zoya Hasan (1009) has suggested, with regard to Uttar Pradesh, that this form of mobilization of class interests has been eclipsed by the politics of Hindutva. 86. Omvedt 2005; Herring 2005. 87. For the analytical perspective of a farmer organization leader, see Joshi 2001; on the politics of the organization itself, see Omvedt 2005. 88. Frankel and Rao 1989, 2; see also Mendelsohn 1993. 89. Varshney l998; see also, Lindberg 1995; and essays in Brass 1995. sabha construction that assumes interests in agriculture to be complementary, not antagonistic. Bengali communists papered over the class contradiction between labor-hiring farmers and labor-selling workers; Kerala communists accepted the reality of conflict and built organizations accordingly. The Bengal model of rural class cooperation worked better politically than Kerala’s confrontationist model. In official understandings of rural India, the cycle has returned to the 1969 report of the Home Ministry on “agrarian tension.” Naxalism again threatens law and order — and thereby investment and growth. The grievances of rural “Maoists” build upon additive dimensions of subordination developed above: landless workers of “tribal” or outcaste standing suffer from social oppression, political exclusion, and economic exploitation. Reciprocally, the state faces a rising tide of mobilization against its integration with the global economy generated by middle-class activists with great skills and connections. Anti-globalization campaigns portray severe and generalized agrarian crisis, as indicated by what is held to be a rising tide of suicides. As a sector, agriculture continues to account for a smaller percentage of the workforce each year, as in all industrializing societies. But the class structure of agriculture is not static. In the first systematic data of Independent India, in l951, 82.7 percent of the population was rural; 71.9 percent of rural people were cultivators, 28.1 percent were agricultural laborers. The ratio is that of a peasant society: dominated by farmers, with a significant proportion of rural proletariat. By 2001, only 72.2 percent of the population was “rural,” but farmers were a bare majority: 54.4 percent of the rural population. Agricultural workers — now much more diversified in employment — constituted 45.6 percent of the rural population. This is the truly awkward class: largely unattached to anyone’s land, selling labor power as a commodity in an unpredictable market, often uprooted by pushes and pulls of market forces, and largely without representation. Recovering Class: Contributions and Puzzles What do we lose from the marginalization of class analysis? At the aggregate level, representation of whole societies by summary compressions such as GDP per capita loses variance. Class analytics depends on disaggregation, of moving beneath aggregate presentations of economic well-being to the level where people live, where life chances are still, perhaps more than before, unequally 344 Critical Asian Studies 38:4 (2006) 90. Patrick Heller (1999, 237–48) rightly emphasizes the resultant pressure on the Kerala communists to make their own class compromises, and he thinks the class compromise may produce economic development. 91. Beyond the scope of this essay, it must be said that such generalizations are weak on evidence, and the connection to transgenic technology imaginary. See Herring 2006. For a broad overview of the crisis and farmer suicides, see Vaidyanathan 2006. Some sectors are doing better than others, some years are better than others: 2002–03 was a hard year for agriculture; 2003–04 was a banner year, with growth in excess of 9 percent; 2004–05 was a year of weaker performance. 92. GOI 2004. distributed. It is not surprising that a rising and prosperous middle class, along with rapid technical change and explosion of consumption opportunities should overwhelm seemingly quaint Nehruvian concerns with “weaker sectors” and inequality. Yet class itself certainly does not go away with wealth. Most bluntly, how long one can expect to live, and how well, even in the richest countries, depends on the lottery of class. The loss of class analysis to fads of developmentalist or constructivist aggregations obscures deprivations of those who have little power to defend themselves in either markets or politics. “Development” shorn of class is a lazy and ideological distortion. The more ambitious explanatory project is to explain how class structure influences political behavior and thus social change. As with identity — and other broad characterizations of groups — class enables explanation of collective action that may change the rules of the game at a societal level. Class analysis has given us appreciation of historical junctures and path dependency in explaining contemporary variation; of the agency of subordinate classes in the face of binding constraints on action; of the impact of new production relations on new strategies of workers. Our historical understanding of the subcontinent would be impoverished without studies of class forces in challenging colonial rule, establishing independent states, devising development strategies. Understanding new state-society relations under conditions of globalization will depend on better class analytics, encompassing strategic choice as a fundamental element, with attention to variable political opportunity structures created by party systems and states and new modes of integration with the global economy. Neoliberal reforms deregulate the workplace as a matter of policy, and simultaneously rein in welfare functions of government. Recent literature concludes that with globalizing neoliberalism informal workers are replacing formal-sector workers; contract labor that operates outside the protection of labor laws increases in train. The decreasing proportion of formally employed workers the world over is held to signify a decline in class-based organization, undermining labor-union power and membership at the global level. As capitalists become less recognizable in a world of subcontracting and informalization of work, inHerring and Agarwala / Introduction 345 93. Marmot summarizes findings from multiple studies. He notes: “When traveling along the distance of nearly twelve miles on the Washington, D.C., Metro from downtown to Montgomery County, Maryland, life expectancy of the local population segment rises about a year and a half for each mile traveled. Poor black men at one end of the journey have a life expectancy of 57 years, and rich white men at the other end have a life expectancy of 76.7 years.” Marmot finds more generally that “In rich countries, such as the United States,...persistently, those at the bottom of the socioeconomic scale have worse health than those above them in the hierarchy” (Marmot 2006, 1304). 94. See, for example, Chakrabarty 2000; Chandravarkar 1994; Chibber 2003. 95. That neoliberal magic can work for the poor if they can only be em-propertied by legal changes is developed in Soto 2000. On pressures from multilateral agencies in the form of aid conditionality, see Herring and Esman 2001; Stiglitz 2002. formal sector workers seek to translate their collective position as a fragmented class-in-itself to a politically effective class-for-itself. The informal economy illustrates the necessity of original class analysis and possibilities for rethinking theory. It is not that informal-sector workers fail to organize for class interests, but rather that new structures of production alter their strategies. First, because capital takes the form of constantly changing employers, who may even be unknown at the point of production, worker organizations take their demands to the state, rather than to capital. Second, demands for expansion of citizenship rights focus on welfare benefits (such as health and education), rather than workers’ rights (such as minimum wage and job security). Third, because neither employers nor workplaces remain constant, informal workers organize around the neighborhood, rather than on the shop floor. These strategic changes have an impact on class identification: a unique class identity that simultaneously asserts workers’ informality and their position within the working class. Informal workers employ a rhetoric of “citizenship” and mobilize votes to institutionalize rights. Without understanding the changing structure of class under globalization, and its relational nature, and without a focus on the strategic aims of labor within a particular political opportunity structure, our understanding of outcomes in this growing sector would be much impoverished. Workers must interpret their interests to make a strategic choice under altered conditions about the most promising collective action; aggressive and disruptive union tactics may be attractive to some workers, but depress investments when compared to unions that buy in to institutional labor compacts. Objectively there is an argument for abandoning class confrontation; that many workers see the world differently illustrates the critical nature of cognitive screens in mediating between interest and action. That these differences exist reduces the ability of a class for itself to form. Formal-sector workers have been much studied — though misread, as Teitelbaum demonstrates — but most workers are relegated to the informal sector. Changes in structures of production — informalization, outsourcing — have facilitated, often necessitated, new strategies as discussed above. The “middle class” — or classes — have especially multiplex interests; their allegiances have been historically opportunistic. Evidence from the subcontinent suggests a significant leavening of celebration of the middle class as the bedrock of democracy and economic dynamism. Governments present an expanded middle class to foreign capital as reason for investment. In turn, the new middle class deploys a range of strategies to protect their privileges in the face of political empowerment and the growing assertiveness of the poor. Dem346 Critical Asian Studies 38:4 (2006) 96. Tilly 1995; Western 1995. 97. Agarwala 2006. 98. Ibid. 99. Teitelbaum 2006. 100. Rueschemeyer, Stephens, and Stephens 1992. 101. Herring and Mohan 2001; Fernandes and Heller 2006. ocratic deepening strengthens ties between the poor and political parties. The empirical work of John Harriss and colleagues suggests that middles classes increasingly cede the public sphere to the poor and their “dirty-river” politics; the affluent defect to private provisioning as the state declines in authority and capacity. Participation in NGOs is important for the middle class, as it is for many of the poor, but participation itself takes on forms differentiated by class. For example, the movement to “cremate Monsanto” reinforces Harriss’s findings on the “activist” nature of middle-class identity and class-differentiated interests in collective action. The failure of that movement in part reflects the radical freedom of those with cultural capital and connections based on middle-class brokerage positions to ignore the compelling facts of material production: what farmers face in their fields. Activists are free to adopt discourses tuned to global coalitions that offer authenticity rents but are divorced culturally and materially from those they claim to represent. The brute facts of biological processes confront direct producers: farmers are constrained by their role in production to skepticism about claims of “suicide seeds” — or “genocidal” seeds — and forced into a grounded empiricism that cannot afford junk science. As systems of production and distribution ramify globally and technically, these cognitive screens are of increasing importance to understanding interests. Framing has the power attributed to it by social movement theorists, but within limits, many of which are mediated by class structure. In recovering class, we find a useful rethinking of the deductivist and macrohistorical logic of dominant versions of European class analytics. Uncovering mechanisms takes priority, and the uncovering must be an empirical process. How do things actually work? Mechanisms focus attention on agency of historical actors, recovering in the process human agency, both from its obscurantist obliteration in ideational constructivism and its a-contextual incarnation in certain brands of rational-choice theory. We find that complexities of class structures, and their interpretations from specific class positions, necessitate a less determinative intellectual architecture than The Communist Manifesto. Historically, class analysis has tended to focus on explicit moments of the articulation of class interest (elections, insurrections, repression, etc.). These moments are of course important: indeed, in Pakistan, institutions for expression of class interests are so feeble and discontinuous that working-class power has been expressed primarily in convulsions that brought down political systems, but could not maintain a continuous presence in struggles for class interests. To assume that these convulsions were the only relevant aspect of class in Pakistan would be naïve. At the micro level, where all of us live, are the day-to-day practices through which classes define and reproduce themselves. Old class analysis was not so interested in these struggles in civil society. Marx himself was Herring and Agarwala / Introduction 347 102. Harriss 2006a. 103. Ibid. 104. Herring 2005. 105. Candland 2007. convinced that the point of production was decisive as a determinant of class formation and collective action. Yet class struggles also take place in communities and local institutions. Where aggregate developmentalism posits civil society as an organic entity, class analytics provides more finely grained understandings of the divisions, tensions, and conflicts in civil society. Examples would include efforts to claim and horde education/science/culture, to secure legal sanction (reservations, definitions of formal vs. informal sector, labor regulation) or to secure institutional support (conflicts over space, or government policies). These struggles have important material effects; emphasis on the material forms of property alone deflects attention from the importance of cultural capital and the role of the developmental state in distributing life chances. In pursuing these ends, individuals are strategic, and objects of larger strategies; struggles often take highly euphemized forms, as in struggles over caste, identity, and culture, that are then often interpreted as Indian exceptionalism. For reasons that this essay has explored, the decline or rise of political organizations on the left has little to do with the validity of class analytics; indeed, a full understanding of how class works politically and socially aids in understanding why subordinate classes remain subordinate. The complexity and overdetermination of electoral and organizational outcomes creates puzzles for analysis, not refutation of an approach. There are many slips from perception of interests to mobilizational strategies to coalitional tactics that result in political success 348 Critical Asian Studies 38:4 (2006) 106. Fernandes and Heller 2006. Ishwar Chand makes nylon ropes outside his shanty in Bombay, India. On the left is a promotional poster of a Bollywood movie. “Since the Indian government began implementing economic reforms in 1991, per capita income differentials across states have risen, along with inequality within states.” (AP Photo/Aijaz Rahi. 17 January 2006) or failure. Moreover, assessment of political success or failure is highly time-dependent. After decades of decline, the resurgence of the Left in Latin America through new coalitions along horizontal lines, largely defined by winners and losers in neoliberal policies, underscores this point. Left parties in India had been largely written off before the national elections of 2004, when the communists had their best-ever national showing. Simultaneously, Delhi rediscovered a sporadic agrarian class war that had become more widespread and deadly. The spectrum in the subcontinent moves from Pakistan to regional social democracies in India. In Pakistan, working class power has been episodic in expression, but unable to institutionalize itself. In Kerala and West Bengal, class formation is advanced and politically competent. But there is also Bihar, as the conventional wisdom goes. Amid the general assertion of nations as operative political identities, the regional isolation of Left parties in West Bengal and Kerala reminds us of the need to disaggregate to the level where political action can be effective. This is why one does comparative research: under what conditions does Bihar become more like Kerala, or vice versa? Unfortunately, neither class theorists nor anyone else has robust political theory. Were economics judged by outcomes, at least half the profession would be disbarred. Political outcomes are messier still. This is true because of the complexity of cognitive mediation in decisions of individual actors concerning class identification, class interests, and likely political outcomes of class formation and class struggle. It may not be rational for the individual to fight a landlord over crop shares or an employer over wages if the outcome is to lose access to the means of production. That tenants accept illegal and extortionate production relations is a common finding; the finding itself is explained by both power asymmetries of the class structure and the choice set facing the landless. Knowing the class structure in agriculture cautions against mistaking acquiescence for legitimacy, or being surprised by agrarian insurgencies. Likewise, the larger political opportunity structure affects the rationality of expressing class interests collectively: there may be no point in voting for a party that cannot win, nor going to the streets if the army shoots protesters. The working classes of Pakistan, suppressed and divided though they have often been by ideology and primordial loyalties, were critical in convulsions that ousted very powerful military dictators, though they ironically failed to benefit from democracy. For such low-probability events, no one has good predictive theory. Herring and Agarwala / Introduction 349 107. For an explanation of cleavages without class politics before the resurgence of the left, see Roberts 2002. 108. States within India matter in differentiating the life-chances of citizens. Consider Drèze and Sen 1996; Harriss 2003; Heller 1999, 7–10, and Sinha 2005. Class interests may well be served by patronage systems in which carrots are plucked from a common patch but rewards sustain both local elites’ political power and some floor guarantees for the poorest of the poor. See Herring and Edwards 1983. 109. Herring 1981. 110. Candland 2007. These indeterminacies are not overcome by a claim that identity politics has replaced class politics: we are equally unable to provide compelling theory on the conditions under which some identities will be chosen over others, or why answers vary over time. Nor do we know how some identities produce successful politics when others do not. We do know that such choices are embedded in structures of constraints and opportunities. However one’s identity gets constructed, economic dependency matters fundamentally. Probing class structures and the choice sets of rational — or at least reasonable — actors recovers individuals from the cultural dopes of Orientalism or stomachs attached to calculating machines of economism. The really difficult tasks concern specification of the conditions under which interests of those disabled by the class system can be inter-subjectively recognized and acted upon politically. These conditions cover a wide range: from basic political rights to cognitive mediation of interests to structural power of capital. The transition first noted by Marx — from a class-in-itself to a class-for-itself — is not particularly predictable, but, as with many unpredictable outcomes, the consequences for the subcontinent are profound.