Dignity Through Discourse: Poverty and the Culture of Deliberation in Indian Village Democracies

Employing a view of culture as a communicative phenomenon involving discursive engagement, which is deeply influenced by social and economic inequalities, the authors argue that the struggle to break free of poverty is as much a cultural process as it is political and economic. In this paper, they analyze important examples of discursive spaces - public meetings in Indian village democracies (gram sabhas), where villagers make important decisions about budgetary allocations for village development and the selection of beneficiaries for anti-poverty programs. They examine 290 transcripts of gram sabhas from South India to demonstrate how they create a culture of civic/political engagement among poor people, and how definitions of poverty and beneficiary-selection criteria are understood and interrogated within them. Through this examination, they highlight the process by which gram sabhas facilitate the acquisition of crucial cultural capabilities such as discursive skills and civic agency by poor and disadvantaged groups. They illustrate how the poor and socially marginalized deploy these discursive skills in a resource-scarce and socially stratified environment in making material and non-material demands in their search for dignity. The intersection of poverty, culture, and deliberative democracy is a topic of broad relevance because it sheds light on cultural processes that can be influenced by public action in a manner that helps improve the voice and agency of the poor.


Policy ReseaRch WoRking PaPeR 4924
Employing a view of culture as a communicative phenomenon involving discursive engagement, which is deeply influenced by social and economic inequalities, the authors argue that the struggle to break free of poverty is as much a cultural process as it is political and economic. In this paper, they analyze important examples of discursive spaces-public meetings in Indian village democracies (gram sabhas), where villagers make important decisions about budgetary allocations for village development and the selection of beneficiaries for anti-poverty programs. They examine 290 transcripts of gram sabhas from South India to demonstrate how they create a culture of civic/political engagement among poor people, and how definitions of poverty and beneficiary-This paper-a product of the Poverty Team, Development Research Group-is part of a larger effort in the department to understand local government and citizen-based engagement in developing countries. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at vrao@worldbank.org. selection criteria are understood and interrogated within them. Through this examination, they highlight the process by which gram sabhas facilitate the acquisition of crucial cultural capabilities such as discursive skills and civic agency by poor and disadvantaged groups. They illustrate how the poor and socially marginalized deploy these discursive skills in a resource-scarce and socially stratified environment in making material and non-material demands in their search for dignity. The intersection of poverty, culture, and deliberative democracy is a topic of broad relevance because it sheds light on cultural processes that can be influenced by public action in a manner that helps improve the voice and agency of the poor.

INTRODUCTION
Public deliberation has a long history of being celebrated by political theorists as a hallmark of true democracy, and it is increasingly being adopted as a tool for resource allocation among poor communities in the developing world (Mansuri and Rao 2004). In this paper we examine a major attempt by a country to address poverty through a mechanism of deliberative decision-making which aims to equalize voice and political agency across stratified social groups. We are referring here to the 73 rd amendment to the Indian constitution, ratified in 1992, which vastly increased the role of village councils or gram panchayats (GP) in rural governance. While the ostensible goal of the amendment was to decentralize the functions of government, it did so by explicitly attempting to equalize political power by reserving quotas in political offices for women and underprivileged castes. Scope for the exercise of voice and agency by different social groups was created by mandating that all village councils should hold gram sabhas (GS), public meetings, at regular intervals in every village -acting as the village's parliament.
In these meetings, it was envisioned that citizens would discuss and ratify core decisions made by the GPs on the selection of beneficiaries for anti-poverty programs and on budgetary allocations for the provision of public goods and services. The funds controlled by these village councils and their jurisdictional powers were accordingly increased.
Since then the GS has become, arguably, the largest deliberative institution in human history, at the heart of two million little village democracies which affect the lives 700 million rural Indians. Thus, the practice of democratic politics, its attendant deliberative rituals, election cycles, and political machinations and negotiations, have now become part of the quotidian landscape of India's rural life.
The panchayat system has been widely studied by scholars of Indian democracy interested in the nature and effectiveness of decentralized governance (e.g. Krishna 2002, Besley, Pande and Rao 2005, Jayal, Prakash and Sharma 2006, Bardhan and Mookerjee 2007, Heller, Harilal and Chaudhuri 2007. There is a vast and rich body of research on questions about what this policy initiative has meant for how poor and socially marginalized groups participate in politics and whether this measure has improved the quality of governance. However, the phenomenon of the GS has broader relevance and, in some ways, transcends its geographic location, making it significant worldwide for scholars of poverty and development interested in culture. For these scholars the GS should be of interest as a policy intervention that aims to facilitate a culture of civic, or political, engagement among economically and socially disadvantaged groups, where this engagement is primarily discursive, and is meant to act as a vehicle for guiding and monitoring resource redistribution. From a comparative perspective, it is possible to think of the U.S. welfare system and Indian village democracies as somewhat comparable to the extent that they both operate as mechanisms of allocating benefits to socio-economically disadvantaged groups. There are also parallels between the target populations for federal public assistance in the respective countries. In the US, the poor are disproportionately composed of African Americans, while in India they are largely composed of members of lower castes (also known as dalits, literally meaning oppressed, or scheduled castes 1 ) and tribes, who occupy the bottom of the symbolic and social hierarchy embodied in the caste system. The poor in both these contexts, despite the differences in their histories, share some fundamental similarities. In addition to economic deprivation, both groups have been historically subject to persistent discrimination on the basis of their marginalized identity. Both groups have been targets of negative and essentializing assumptions about behavioral and attitudinal traits. Both groups are the subject of affirmative action policies, which are more extensive in scope in India than in the US. Finally, both groups have been historically blocked from exercising their voice in the public sphere and have lagged behind in their levels of political participation. 2 All of these factors have led to deep "in/equality of agency" (Rao and Walton 2004), i.e., differences in the voice and opportunities for redress that separate these groups from the dominant ones. It is this last deficiency that the Indian government attempts to rectify by reserving quotas for underprivileged castes, tribes, and women in local politics.
However, despite these similarities, there is a major difference between the two redistributive systems in their levels and natures of civic or political engagement required from individuals to gain access to publicly funded benefits. The U.S. welfare system requires minimal public engagement, is highly bureaucratic by nature, and vests complete authority and monitoring capacity in the hands of the state. The individual beneficiary is a powerless subject, left at the discretion of the state-appointed social worker.
Additionally, becoming a beneficiary of the US welfare system comes at the cost of social stigma associated with subscribing to a perceived 'culture of dependency'. In contrast, the Indian GS model requires a high level of political engagement which, at least by statute, requires the active exercise of voice in framing and defending demands in the eyes of the state vis a vis competing individuals and groups. The GS also brings 2 The last two decades have seen a sharp rise in mainstream political activism and engagement by lower castes in India (Varshney 2000). together local individuals and groups who vary in their economic, social, and cultural capital, and, therefore, create a space for public interaction among the dominant and the dominated, the literate and the illiterate, and political veterans and neophytes. The public, including the poor, are simultaneously subject and stakeholder, being vested with the authority to monitor the activities of the local state and demand accountability. The fact that the GS as part of the panchayat system is mapped onto an electoral space -staffed by elected representatives who, for the most part, belong to mainstream political partiesmakes discursive engagement in it free of stigma, a performance of ideal citizenship.
Our analysis of public discourse in the GS, therefore, is an examination of the extent to which a policy of redistribution targeted to the poor, a policy which creates a space for the poor and socially marginalized to exercise voice, can facilitate a culture of civic/political 3 engagement among them. The eventual goal of such an analysis, in spirit, would be to examine if this structural opportunity to exercise voice can alter the "terms of recognition" (Appadurai 2004), i.e., the conditions and constraints under which groups participate in society and negotiate with the social norms that frame their lives.
Accordingly, in our analysis, we ask the following questions: How, if at all, do rural men and women deliberate in these meetings and about what issues? How, for instance, does poverty and its attendant habitus shape deliberation? Do identity and social hierarchy permeate into the political sphere which is meant to guarantee de jure equality? Does providing an arena for citizens to deliberate complicate the state's attempt to have a quantifiable definition of poverty and encourage citizens to challenge it? In answering these questions, we look at how poverty shapes the culture of deliberation, and how public discourse shapes the understanding of poverty by contesting its meaning.
Our goal in this paper is twofold. Theoretically, we wish to contribute to the scholarly literature on culture and poverty (Lamont and Small 2008) by proposing a view of culture that amalgamates sociological perspectives on culture with the human capabilities approach in the interdisciplinary scholarship on development (Sen, 1985).
Empirically, we wish to contribute to the field of scholarship concerning deliberative democracy (Fung 2004;Gutmann and Thompson 2004;Elster 1998;Bohman 1996;Mansbridge 1980), by bringing in rich qualitative data to a field which has been largely restricted to theoretical discussions about the ideal nature and purported virtues of public deliberation. The data are drawn from qualitative analysis of the recorded proceedings of a large representative sample of 290 randomly selected gram sabhas held in four states in South India.
The rest of the paper is divided into six sections. In the second section we outline our understanding of culture and contextualize it within past and present usages of culture in the literatures on poverty and development. In the third section, we outline relevant parts of the deliberative democracy literature and highlight the implicit assumptions about culture embedded within it. The fourth and fifth sections include details about the context and data and method. In the sixth section we present our substantive discussion.
Finally, in the seventh section we conclude by highlighting our important findings and elaborate on their implications.

CULTURE IN POVERTY AND DEVELOPMENT
Scholarship in the fields of culture and poverty, which has a long tradition in the US, and on culture and development, which is more internationally focused, have contentious histories. We touch upon these fields briefly here only to highlight a commonality in the way in which culture has been mis/used in some works in both fields.
In the former, culture has been loosely understood as patterns of behavior, or norms and values, regarding work, marriage, family, and childbearing. Often actual behavior and underlying attitude or ethics have been confusingly conflated. Culture and race have also been used interchangeably. The combined effect of this way of thinking, best encapsulated in the "culture of poverty" theories of the 1960s, has been to suggest that some groups, particularly the poor, have a culture which is self-defeating and acts as a vicious cycle to trap members of these groups in persistent poverty, even in the face of opportunities to escape it. 4 On very similar lines, some development scholars studying the question why some countries develop versus others lag behind have suggested that some communities have a culture, beliefs and practices, which is detrimental to progress (e.g.: Harrison and Huntington 2001). They have argued that "culture matters" because societies immersed in traditional cultures are ill-suited to market-driven development and are critically hindered in their pursuit of economic growth and progress. This thesis bears close kinship to the notion of an adverse "culture of poverty." Departing from these approaches, we propose a view of culture as a relational phenomenon -among individuals within groups, among groups, and within ideas and perspectives. Culture is concerned with identity, aspiration, symbolic exchange, coordination, communication, and structures and practices that serve relational ends, such as ethnicity, ritual, norms, and beliefs (for a fuller discussion see Rao and Walton 2004).
It is a set of contested, malleable, and quotidian attributes, constantly being re-created and re-imagined by the changing dynamics of human connections and interactions. Thus, communication is at the heart of culture, and patterns of communication both shape and are shaped by economic, social and political inequalities. The struggle to break free of poverty and inequality is, consequently, as much a cultural process as it is political and economic. It is centrally related to voice, modes of discourse, and the degree of access to the public sphere. Thinking along these lines, we believe, has important implications for policy since it suggests that in addition to equalizing opportunity, public action to address poverty would need to find ways to equalize voice and agency.
Among recent conceptualizations, the notion of culture as repertoires of practices, beliefs, and attitudes that individuals mobilize at the time of action is useful for our purposes. Particularly instructive is the idea of culture as a "tool kit" of habits, skills, and styles from which people construct "strategies of action" (Swidler 1986). This understanding of culture closely resonates with the idea of culture as a capability -the constraints, technologies, and framing devices that condition how decisions are made and coordinated across different actors. We add to this tool kit of cultural capabilities by arguing that styles of political discourse, i.e. ways of articulating demands in the process of political participation, is another key component of culture. Discursive styles are shaped by poverty and its associated deprivations, like lack of literacy, lack of access to the public sphere, scarcity of resources, etc. But, at the same time, particular forms of state intervention, like the GS, require the poor to strategically deploy discursive styles to seek resources that are expected to ameliorate poverty. The goal of the paper, therefore, is to understand the discursive styles used by the poor in their civic/political participation, or what we call the "political culture of poverty" 5 .

CULTURE IN DELIBERATIVE DEMOCRACY
The relationship between culture and democracy is not obvious or made explicit in much of the theoretical and empirical scholarship on democracy. In theory, democracy represents a set of procedures governing decision-making that ought to operate effectively regardless of culture and context. We argue here that, in reality, the notion of democracy, particularly of the deliberative kind, is based on implicit assumptions about culture and that a concern with culture is intrinsic to achieving a nuanced understanding of the nature of public deliberation.
A democratic system, "radical democrats" envision, will go far beyond keeping a check on the power of the few and attempting to respond to the demands of various constituencies. Participation in a democratic system is expected to lead to a process of positive self-transformation by catalyzing a set of desirable changes in the individual.
These are the following: enhance the individual's facility for practical reasoning; make people more tolerant of difference; and more sensitive about the need for reciprocity; enhance people's ability to think and act with autonomy on the basis of their own preferences; and to engage in moral discourse and make moral judgments (Warren 1995).
Some scholars also expect the deliberative process of democracy to produce a consensus 5 We hope to make it sufficiently clear that our relational, communicative perspective is orthogonal to the general understanding of the term "culture of poverty" originally coined by Oscar Lewis but for which he has been incorrectly demonized (Later authors were responsible for using the term to generate unfortunate stereotypes that suggested that the poor had a culture that caused their poverty --see Arizpe (2004) for a history). on preferences regarding final ends and means, that is, a "unanimous preference" determined through the power of reasoning (Elster 1986: 112). It has indeed been shown (Dryzek and List 2003;List et al 2006) that deliberation makes individual preferences more "single-peaked" and amenable to aggregation. In theory, then, participation in a deliberative democracy is expected to produce more cognitively competent and wellinformed people with an enhanced capacity for consensual action.
The fairness of this system is guaranteed by the existence of an "ideal speech situation", a social structure based on discursive equality. This structure is composed of a "public sphere" where individuals are a priori equal and free of the distorting effect of inequalities or coercion; every person with the competence to speak and act is allowed to participate; and every person is allowed to question and introduce assertions, and to express his opinions, desires, and needs without fear of repercussion (Habermas 1981). In cultural terms, the kind of community envisioned is "the ideal of a moral community, one whose norms and practices are fully acceptable to those subject to them, a society based not on imposition, but on the agreement of free and equal persons" . In such a community, social behavior is never "agonistic", i.e., behavior that is observed among opposing entities with unequal means, e.g. fighting, threats, attack, appeasement, submission, and retreat, are not present.
The theoretical scholarship cited above implicitly suggests that, as far as the proper functioning of democracy is concerned, there is a right (conducive) and a wrong (not conducive) kind of "culture." The "right culture" -complete and voluntary acceptance of social norms by all in a system free of hierarchy-is expected to provide a fertile basis on which democracy will sprout and flourish. The "wrong culture," a hierarchical system marked by disagreement and contestation around social norms, on the contrary will serve as a fractious basis on which democracy, even if introduced by external intervention, will perish. Moreover, much of this scholarship does not contend with the existence of inequality. In fact, some scholars like Lyotard (1984) andLuhmann (1982) have criticized Habermas's emphasis on reaching consensus via deliberation, interpreting such consensus as a new form of hegemony that may potentially mask the interests and narratives of marginalized groups. However, despite such theoretical criticisms, the ideal of deliberative democracy tends to maintain its intellectual dominance in thinking about how democracy should indeed operate in the real world.
This dominance is reflected in the few existing empirical studies, like the extremely insightful ethnographies of Manbridge (1980: rural Vermont) and Baiocchi (2005: Porte Allegre, Brazil), which emphasize deliberative, consensual agreement.
However, a useful cultural approach to democracy, we argue, has been suggested by scholars of communication (Barnett 2003) who have pointed out that, in cultural terms, democracy itself is an 'artful practice' in that it involves "the formal and informal cultivation of competencies of judging, reasoning, appreciating, performing and responding" (199). Scholars have duly emphasized the importance of deconstructing "representation" -asking "Who speaks?", given that speaking "depends on an individual's position within regulated systems of discourse" (Barnett 2003: 16) -and focusing on the "technologies and techniques of persuasion" used in speaking (Morris 1998: 230). This idea of democracy as an artful cultural practice reinforces our argument in favor of considering discursive styles as part of a cultural tool-kit and political participation through discursive engagement as a capability, both being fundamentally influenced by poverty. Therefore, our main empirical goal is to examine the claims of deliberative democracy against actual evidence, a task so far undertaken by few (Delli Carpini 2004), and analyze public deliberation through the lens of culture and poverty. In keeping with our goal, we raise and answer the following questions: how exactly does the empirical reality of India's social and cultural life deviate from the ideal view? How is representation in the supposedly deliberative institution of the GS patterned by inequality and identity? If they are not deliberative in the ideal sense, do these discursive negotiations further democracy or the cause of the poor in any manner at all? We examine who speaks, what they say, and how they speak, i.e., the modalities through which narratives are presented.

CONTEXT: POVERTY AND THE TECHNOLOGY OF GOVERNANCE
The 73 rd amendment to the Indian constitution that transformed village democracy built upon a legacy of legislation that stemmed from the 1882 Resolution on Local Self-Government initiated by the then Viceroy Lord Ripon. Ripon's main intention behind this legislation was to facilitate "political education" and "training in the work of representative institutions" in a manner that built on indigenous systems of village government (Tinker 1967). Ripon's tenure as viceroy did not last long enough for him to ensure that his reforms were sustainable, and subsequent colonial administrators deemphasized democratic village government. But during India's early 20 th century struggle for independence, gram swaraj (village self-rule) became a key tenet of Mahatma Gandhi's political philosophy and was consequently seared into Indian nationalist ideology. Based on the notion that it would concretize Gandhi's vision of village self-rule, the 73 rd amendment received widespread support across all Indian political parties and regions. The amendment made deliberative processes via the GS a cornerstone of village government, thereby creating a state-sponsored public sphere.
Thus, in rural India, this sphere was not organically derived but was rather mandated from above by national legislation. Therefore, public participation in discursive negotiation toward problem resolution is a governmental technology deliberately instituted and managed by the state ─ a fact that blurs the boundary between civil society and the state (Gupta 1995).
Overall, the cultural context within which the deliberative institution of the GS is embedded reminds us how far removed it is from any idealized notion of the public sphere. Rather than being a monolithic moral community, India encompasses a pluralism of values, ethnicities, faiths, and caste-groups represented by communities varying in size, symbolic social prestige and economic and political power. These communities possess unequal economic, cultural, and social capital and compete for political power. In GS meetings across India, such negotiations are targeted to achieve a means of survival for individuals and families and, at best, a path to upward social and economic mobility.
In this cultural context marked by inequalities and interdependencies, negotiators use their established caste and community identities as resources ─ a kind of capital ─ to stake claims to their due share.
One notable exception is the state of Kerala, one of the four states in our sample.
In Kerala, which enjoys near universal literacy, a well established cadre of Communist Party workers conducted a "People's Campaign" to train citizens in deliberative planning processes, which resulted in remarkably effective local government (Heller, Harilal and Chaudhuri 2007). This campaign built on years of progressive left-leaning rule and effective land reform, sharp reductions in social inequality, high levels of civic participation, and extremely effective human development investments. In Kerala, moreover, where it is mandated that 40 percent of the state's budget be allocated via GPs, these local bodies have substantial resources. Yet in Kerala, as elsewhere, the GS's primary function as a part of the nationwide governance system is to select beneficiaries and allocate public goods.
In terms of their location in the structure of governance, GPs and GSs are the lowest level in a hierarchy that reaches upwards from the village to the county (Block), the district (Zila), the state, and the central government in Delhi. This entire system is staffed by elected representatives and works within the framework of the Indian constitution, which adopts affirmative action policies in order to address social and economic inequalities. The state, therefore, plays an active redistributive role. Wealth is redistributed to the poorest citizens via a technocratic process using lists of "scheduled castes" and "below poverty line" families assessed via surveys. Quotas for elected positions within the government, including seats for GP presidencies and ward representatives, are reserved for SCs in proportion to their population in the village, and a third of the seats are also reserved for women. Quotas are provided for SC enrollment in educational institutions and for government jobs. In addition, a slew of anti-poverty programs implemented through GPs are meant to allocate resources (such as concrete houses, toilets, and small plots of land from common property resources) exclusively for SCs.
Additionally, several other benefits are allocated to people defined as "below the poverty line," known now in every Indian language as "BPL." 6 Depending on the state, families classified as BPL get access to houses, toilets, subsidized food, jobs, cheap credit, and scholarships. BPL criteria, which vary from state to state, include landlessness, unemployment, quality of housing, etc. 7 These "objective" criteria are typically assessed on the basis of a questionnaire designed by the state government and implemented by the GP. The creation of these lists is the government's attempt at establishing a process of commensuration, transforming different qualities into a common metric of poverty (Espeland and Stevens 1998). It is a policy response to the complex task of measuring deprivation in order to redistribute resources and emphasizes economic criteria rather than caste identity. Consequently, it disqualifies people of disadvantaged identity groups if they are relatively better off. However, in order to counter-balance the GPs power over this process, most states require that the list of poor families be ratified by the public at the GS meeting. Thus, the GS is placed in an adjudicative role, and public deliberations on this issue reveal an interesting tension between a desire to 6 A few examples: women over eighteen years of age in BPL households are given Rs. 500 to cover the delivery costs of up to two childbirths; 450 grams of food are given to each house having a child under oneto-three years of age; subsidized housing; subsidized electricity hook-up. 7 There are several criteria specified and used by the Government to identify households falling below the poverty line. Some of these criteria, like annual household income below Rs. 11,500, are applicable nationwide, while others are state-specific. For example, in Kerala the criteria are as follows: (i) families that do not have shelter and have less than ten cents of land, (ii) those who do not have houses, (iii) income below Rs. 300, (iv) those without access to sanitation facilities, (v) the unemployed and those having jobs for less than ten months in a year, (vi) female-headed household, (vii) households with mentally or physically handicapped members, (viii) SC and ST households, and (ix) illiterate. Families having any two characteristics from vi, vii, and viii qualify as BPL. participate in and perfect the process, and, occasionally, interrogating the validity of some of the criteria used or the people selected.
Overall, citizen-state relationships in rural India exist more in the matrix of a gift economy than in the realm of rights and responsibilities. Poor accountability mechanisms, lack of resources, and the identity-based nature of electoral politics result in a culture of supplication and benefaction (Gupta 1995;Mehta forthcoming). Political parties engage in the politics of patronage and maintain well-oiled networks to exchange public and private goods for electoral support (Bardhan 1988). The vast majority of rural residents do not pay their taxes, and a GP's financial resources are mostly derived from grants from higher levels of government. These grants are almost all "tied," in that they are required to be allocated to particular groups of people (SCs, BPL, etc.) or to be used for particular purposes (such as building private toilets or houses). Untied funds are practically non-existent. A GS's ability to make allocation decisions is therefore limited.
It can decide where to locate a road, a water pipeline, or some other public good. It can ratify the selection of beneficiaries for targeted benefits. And, it can serve as a clearing house for information. Most discussions in the GS, therefore, arise in the form of a demand or supplication. Villagers ask the GP to provide a public good in a particular location or to recognize someone as poor enough to deserve private benefits.
Since this discursive space, this "public sphere", is embedded within an electoral space, with politicians attempting to curry favor with different groups, political incentives preclude attempts to suppress voice. By silencing a person, a local politician may risk alienating an entire vote-bloc. Consequently, the GS creates a relatively level discursive field. It briefly releases people from durable inequality traps and allows them the freedom to speak. Such freedom has the tendency to spill over outside the GS into everyday life. Therefore, GSs are troublesome for local elites who at times try to ensure that they are not held. In 2001 twenty-five per cent of GPs did not hold even one GS over a year-long period (Besley, Pande and Rao, 2005).

DATA AND METHODS
Data for this study are drawn from tape-recordings of 290 GS meetings in four South Indian states: Andhra Pradesh (AP), Karnataka (KN), Kerala (KL), and Tamil Nadu (TN). The tape recordings were conducted by one or two field investigators who were instructed to dress in a simple manner and locate themselves in an unobtrusive spot at these meetings, after having taken the permission of the GP president. The recordings were then transcribed into the corresponding local language, and then translated into English to facilitate coding. Each transcript also includes information on attendance, and the caste, gender, official designation, and social position of speakers (elected representative, school principal, villager, etc.). Table 1 provides summary information from the transcript data. The average GS lasts about 84 minutes and is held about an hour after the scheduled time (which is not untypical of public functions in India). About 83 people attend on average, a tiny fraction of the village population which ranges from 2000 to 10,000 depending on the state. Besley, Pande and Rao (2005) report results from a regression analysis of household survey data from the same sample and show that, after controlling for household characteristics and village fixed-effects, illiterate individuals, SCs, the landless and, and the less wealthy are more likely to attend the gram sabha, while women are less likely to attend them. This is primarily because of the GS role in selecting BPL beneficiaries which the poor, disadvantaged families are more likely to benefit from. However, Besley Pande and Rao (2005) also show that this extreme form of selection is less acute in villages with higher literacy levels, where GSs have more representative participation. Table 1 shows that a third of the attendees, on average, are women and 37 per cent are SCs. However, women and SCs do not speak much at the meeting. The "indicator" variable has a value of 1 when any person in a category speaks in a GS, while the "intensity" variable is the time that any person in that category speaks as proportion of the total length of the gram sabha 8 . With this metric we see that 68% of GS had at least one woman speak, but women spoke on average for 9% of the GS's length.
Similarly 60% of GS had at least one SC person speak, but they spoke, on average, for 11% of the time. 9 The typical GS meeting begins with a presentation by the president or the secretary of the GP. This is followed by a public discussion open to all participants during which, typically, villagers mention their demands and grievances, and the secretary or a member of the GP responds to them. These discussions generally center on routine problems, which are summarized in Table 1. The discussions are dominated by issues related to drinking water and village roads, followed by education, electricity, housing, and health. Concerns about employment and agriculture feature less prominently. Discussions also address such complex problems as the legitimacy of 8 Strictly speaking it is the proportion of the number of lines in the transcript spoken by the category divided by the total number of lines in the GS transcript. 9 The SC data are imperfect because we were able to identify SC speakers in only about a third of the sample, which may result in some significant downward bias. Also, the data do not allow us to identify other discriminated castes, like the "backward castes" or BCs. paying taxes when obligated funds fail to arrive and the fairness of caste-based affirmative action as a principle of resource allocation.
In this paper we highlight a limited selection of public discourse culled from a vast array that includes discussions on diverse themes. We focus mainly on discussions about distributive justice, including caste-based affirmative action and criteria for BPL selection. In our view, this is one area where the voice of the poor is most crucial.

DISCUSSION
This section is broadly divided into two segments. In the first segment we show how the contextual realities of poverty and social inequality shape the culture of deliberation in the GS, i.e., the themes introduced and the discursive styles observed. In the second segment we explore how deliberation acts as a vehicle for questioning governmental definitions of poverty and creating a shared, intersubjective 10 understanding of what it means to be poor. In both segments, our focus is on how poverty influences the concerns and discursive styles of the poor and the implications of this participatory exercise for cultivating their voice.

How Poverty Shapes Deliberation and the Discursive Styles of the Poor
Let us begin with an excerpt from a GS meeting in Dharmapuri (TN), which reveals the patterns of participation and the identity of the usual interlocutors in a typical 10 This term is used primarily in phenomenological sociology to refer to the mutual constitution of social relationships and reality. It suggests that people can reach an agreement about their understanding of what they have experienced in their life-world and create a shared world based on their subjective understandings. There is no objective reality that exists outside of this mutually shared subjective understanding of phenomenon. As the term has been used here, it means a mutually shared and constructed meaning of poverty.
GS. This meeting of the GS, serving a village with a voting-age population of 563, was attended by only seven people: three elected members of the GP (the stand-in president, clerk, and an office holder) and four villagers. All of the seven attendees were men.
Among the villagers, Mariappan and Muniraj are both SCs, lowest in the local socioeconomic hierarchy. Of the two, the latter embodies the agonistic voice and highlights the pervasive, historically rooted inequality that is perpetuated even within the current democratic structure. A third villager, Jayaraman, belongs to the "other backward castes" (OBC) category. Finally, there is the "President-husband", who is a member of the "most backward castes" (MBC). In constituencies reserved for female candidates, the female president is often only a statutory head, replaced in her seat of authority by her husband. 11 The excerpt reveals the issues that are of importance and the modalities through which they are presented in the narratives of ordinary citizens. 12 Jayaraman [male, villager, OBC]: There are 45 families in our village. None of us have any land. We work for meager daily wages. Whatever little we get we spend on our children's education. But it's impossible to educate our children up to high school because we don't have the money…. So we request the Government to do something.... Our whole area is dirty. Even the water is muddy, and that's what we drink…. How many times we have requested for a road near the cremation ground and for the supply for clean water?! We can only request and apply. The rest is up to you.

President-husband [MBC]:
If there are 20-25 houses in an area, a ward member should be appointed to represent the area. That ward member should listen to your problems and must do something to help you…

Muniraj [SC]:
That way [if they have a ward member] we will have the guts to enter this room [where the GS meeting is taking place]. If the required ward members are not with us, to whom can we voice our woes? Who will represent us? ... If the ward member belongs to another community, he won't even listen to our problems. Earlier there was a time when a backward caste person was not even allowed to sit in the same area with others! The officers and leaders who come here [to the GS meeting] already have a pre-set plan about what to do and say. You come, sit on the chair, say something, decide among yourselves, and go away. What's there for us to do?! You've enjoyed power for all these years. Why don't you let us have a turn? ... We don't want any problem at the communal level. For us, whether Subban comes or Kuppan comes [common names], it is the same. We vote, but what happens later? Whereas other people get water even before they ask for it, we have to ask endlessly, and even so, our demand is not fulfilled…. We don't want to fight with anyone. But at least there should be someone to listen to our problems. We've been without water supply for the past one month. Even the president knows it. He has promised to send water. But the ward member is not allowing us to take water. The water is sent to all his relatives. We cannot do anything to stop it… President-husband: … In any competition it's a rule that one should win and the other should lose. There's no community-based discrimination or problem. If all of you in booth no. 1 join and vote for me, I become the President. On the other hand, if everyone in the other booths vote for another person, then he'll become the president. And then what'll matter is what he can do for those booths that voted for him. Today, among youngsters, the level of public awareness is very high. Anyone can become a leader.… Even though there are problems between you two groups, I try to mediate. I don't encourage communal riots.… Muniraj: Everyone should be treated equally. No one should be treated as inferior to others. We should also be given a chance to sit on the dais [where the leaders sit]. Why should we be denied that right? Just because I talk like this, it doesn't mean that I fight with you or disrespect you. I am simply voicing my feeling.
The voices heard and unheard in this excerpt reveal the three fundamental cleavages that fracture the Indian public: the social-symbolic divide of caste hierarchies, which is distinctively Indian, and the more generally prevalent economic and gender divide. In order to understand who exercises voice, how they are positioned, and the significance of their speech, it is first necessary to understand these social divides.
Caste-based divisions have deep historical roots and still manifest in such practices as physical distancing and symbolic deference. It is noteworthy that these traditional cultural scripts, which until India's independence legitimized inequality, are now being openly challenged in GS meetings, as witnessed in Muniraj's angry complaints. Such challenges are not completely new given the history of caste reform movements in South India. The Lingayat movement dates back to the twelfth century, and many more such challenges to the status quo emerged during the colonial and postcolonial periods. What makes challenges voiced in GS meetings, like the above instance, different is that, unlike most of their predecessors, they derive not from the educated elite or spiritual leaders, poets, and philosophers on the margins of society but rather from ordinary villagers embedded in everyday, local structures of inequality. These castedisadvantaged ordinary individuals now have a stake in political participation, even in time consuming ones like GS meetings. In effect, identity and resource politics have merged into a common struggle in a scenario where caste-based affirmative action is the principle for distributing resources. And reason-giving as a process of claims-making, claims about appropriate ends and means, is frequently replaced by a competition for predetermined resources based on the severity of need, and caste and communal identity.
The economic divide in rural India is illustrated by disparities in land ownership which frequently coincides with position in the caste hierarchy. Upper and middle castes (which include OBCs in South India) may own a significant portion of cultivable land in a village, while the lower-caste SCs work as daily agricultural wage laborers on that land.
To a large extent, economic disparities determine attendance patterns in these meetings.
Typically, the upper castes do not see any value in attending the GS meetings, which they regard as the government's vehicle to benefit the lower castes by selecting beneficiaries for subsidized goods. When they are present, upper castes often try to dominate the meetings by demanding that precedence be given to their needs. The gender divide is best embodied in the GS in the figure of the President-husband and the absence of women.
Even when women are present (often in large numbers, especially in the South Indian states, due to their membership in self-help, or microcredit, groups), they generally do not participate as actively or enjoy the same rights as their male counterparts. This fact is evident in instances when they are silenced and their contribution discounted by male authority figures.
Despite the fractured terrain of identity, GSs in India function as Durkheimian "sacred spheres" marking the conjunction of civil society and the state. The ritualized interaction in this sphere gives rise to a community of citizens and a brief moment of "collective effervescence." This moment allows individuals with disadvantaged identities (lower caste and poor), like Muniraj, to momentarily discard the stigma of their ascriptive identities and low economic status and to slip into their sacred identity as citizens with equal rights in the eyes of the state. In fact, the more downtrodden an individual is in real terms, the more sacred he or she is in the eyes of the democratic state, who's avowed mission is to guarantee equality and preferential treatment for the disadvantaged. As illustrated in the example, the ritualized interactions in GSs have the potential of challenging the nature of entrenched social relations. They serve to make the covert "weapons of the weak" overt, expose such "hidden transcripts" (Scott 1990) as the physical segregation of lower castes, and provide a means to challenge them.
The process of making the kinds of claims and complaints cited above may seem quite ordinary on the surface, but it acquires deeper significance as a vehicle through which poor, low caste individuals imbibe a sense of possessing equal recognition as citizens. This nascent consciousness occasionally bubbles to the surface, as in Muniraj's vehement request for equal treatment. Such expressions of the "politics of dignity", which have a deep resonance in Indian political life (Varshney, 2000), underlie many material and non-material demands made in the GS. They are demands to reverse adverse "terms of recognition"; to be recognized as social and political equals, with a concurrent desire to improve material well-being.
The work of maintaining, defending, and challenging hierarchical social and symbolic boundaries (Lamont and Fournier 1992) is prominent in the discursive landscape of the GS. In contrast to the open confrontation noted above are subversive strategies which test the boundaries of distinction, without calling for immediate change.
An example is this interaction in a GS in Medak, AP, where a lower caste man uses the offer of water from its community well to tempt upper caste members to relax the rules of purity and pollution governing their consumption.
Villager (male, SC): In our SC-ST colony, there's a bore-well that's plush with water. This water is being misused. If a tank is constructed, this water can be useful for the entire village.
The offer is made with the knowledge that members of the upper caste are most likely unwilling to drink lower caste water, scoring an effective political point about discrimination. Challenges and boundary-testing such as the ones described above are civic capabilities that are being newly cultivated by poor, low-caste individuals as they experience participation in this new arena.
What kind of reaction do these assertions meet, particularly from the upper-castes at whom these are directed? In other words, if these assertions are an exercise of voice, then do they encounter attempts at being silenced? In GS deliberations, the upper castes (the "general" or "forward" castes, referred to below as GC) are most notable for their absence. When they do participate, they may try to dominate the discourse and establish their privileged claim on resources, a privilege primarily derived from their traditionally recognized superiority. For example, in one GS meeting (Chittoor, AP), a general caste person states his demands and uses brute vocal force to assert their primacy over other claims.
Villager: (male, GC): We need cemented roads in Brahamana Veedhi [Brahmin street]. We're not bothered about the expenses incurred by the panchayat. Our problem must be addressed. They work as coolies [day laborers] along with other people [non-SCs] who also have no other option than to work as coolies. However, whereas all the harijans get their dues and facilities, the others who do the same job do not get the same reward as his fellow worker. The government does not give any sort of concessions to these poor coolies, whereas the harijans get all sorts of concessions from the government.
All of these challenges are, in reality, pleas to the government, masked in adversarial language, for a larger share of private and public goods, underscoring the government's benefactor-to-beneficiary relationship with the public. But in the democratic system, the public as beneficiary is endowed with a moral claim. The following example suggests that villagers' understand this as they make a deferential request as if asking for alms, then express indignation at being passed over in favor of others' demands.
Velusamy [villager, male, OBC]: I have been residing in this village through several generations and I've been asking for a house to live in. They say: 'today, tomorrow,' but so far, nothing's been done…. I am sitting here at the mercy of my fate… GP Clerk: Till now houses have been allotted only for the SCs … It hasn't come for OBCs.
Velusamy: They say that it has come only for the SCs, only for them! Is it that only they are humans? And are we people not human beings? How can you say such a thing! What kind of a panchayat is this? We can't go directly and meet the officer. We can only make kind requests to our President, whom we believe in. Make some arrangements for me … Poor, low caste women also engage in making demands, but their discursive style is distinctly different, bordering more on demonstrating their helplessness and beseeching the state for personal assistance. They vary between unabashedly making private demands, in the most desperate cases, to presenting their demands as public concerns. In these cases, the GS becomes a site for the continual masking and unmasking of private needs disguised as public needs and becomes a perilous terrain, where "hidden transcripts" are continually dragged into the discourse. The following example from Coimbatore (TN) illustrates these discursive strategies and shows how figures of authority question intentions, debunk claims, and pass moral judgments on individual needs.
Mailaatha [villager, female, SC]: You must lay a road for our street. It's very difficult for the children to walk. You must help us.

President [male]:
Your street is not a busy street! Your husband is ill and disabled. So you have to hold and assist him [to walk]. We will lay roads if more funds are available. … Savithri [villager, female, SC]: I don't have a husband, and I'm suffering a lot with a child. I've got to pay for six months house rent. I want a house. Despite the brisk refusal with which both of these demands are met it is important to note that individuals such as these who bear a triple burden of disadvantage, being poor, lowcaste, and women, are learning the art of making demands in a public arena -which was inaccessible to them even a few decades ago. In addition, they are also seen cultivating the artful practice of using a public goods framework to present private needs.
In terms of the concerns that are expressed in the GS, distributional equity is prominent. However, this concern is not restricted to the imperatives of survival and bread-and-butter goods, like free meals and uniforms for children and subsidized sanitation facilities. It also extends to issues of socio-economic mobility, such as obtaining greater opportunities for education and employment. Poor, socially marginalized groups use the GS to broach such subjects, which fall far beyond the practical scope of the GS. For example, in the following excerpt from Coimbatore (TN), the villager and the GP president debate the un/affordability of an English medium education and whether that hinders a person's chance to get into the highly coveted civil services, in the context of caste-based affirmative action.

Villager (male):
A girl who studies in Mallanad 13 Panchayat corporation school [Tamil medium] cannot match a boy who studies in an English medium convent school in Coimbatore. Parents need to spend a lot for English medium education for their children.

President (male, GC):
Your concern about how an ordinary student will compete with convent-educated students is valid. That's why they've kept caste as a selection criterion. On the basis of caste, if SCs and STs have lower marks and are older, they'll still be preferred over forward caste candidates in the IAS (Indian Administrative Services) selection process. So, with regard to education, nowadays preferential treatment is given on the basis of caste. We should change this system in favor of income-based affirmative action. This is my opinion.
The interesting fact about discussions of this nature is that such matters lie far beyond the purview of the GS. Therefore, in pragmatic terms, it is futile to discuss them in this venue. The villagers know that an extremely complex and contested parliamentary process determine affirmative action policies. Yet the GS provides ordinary citizens a place to think about and voice their concerns about broader policy issues and abstract principles that closely touch their lives. Through this discursive engagement poor villagers participate in democracy and, hence, perform their citizenship.
As illustrated through the various examples cited above, poverty, shot through with material and symbolic inequality, inevitably undermines the idealized neutrality of democratic discourse in the GS and shapes the culture of deliberation. Most notably, an improvised vernacular style of verbal negotiation is emerging, a style which is created as citizens compete for resources, challenge hierarchical social boundaries, and critique 13 Name changed to protect anonymity. principles of affirmative action and the status of distributional equity. These styles include assertive demands and complaints, beseeching for benefits, even demonstrating helplessness, and more directly confrontational or acrimonious styles. However much they may deviate from the idealized rational argumentation envisioned under the deliberative democracy framework, these discursive styles represent the transformation of the GS into a "level discursive playing field", which encourages a culture of competitive participation where the politics of dignity are played out, boundaries of caste and class transgressed, and the political power of the poor displayed. Voice, agency, and dignity are publicly re-imagined and renegotiated, and, as a consequence, the capacity of the poor and marginalized to be full and equal citizens is demonstrated and at least momentarily realized.

How Deliberation Shapes the Meaning of Poverty
The inadequacy of identifying the poor simply on the basis of ascriptive categories (such as caste) has led India's central Government to adopt a quantifiable, poverty-based measure to achieve distributive equality. Yet the definition of poverty is hardly obvious or unproblematic, as can be seen in the discursive exchanges taking place in GSs in the four South Indian states we studied. In these exchanges, elected GP representatives and the public make a joint effort to understand the definition of poverty and the category of the "BPL beneficiary" as laid out by the Government.
Government representatives use the GS forum to keep the public abreast of the state's efforts to fix poverty by pegging it down to certain objective criteria and to translate it to a common metric of numerical scores by using human and mechanized technologies, like surveys, computerization, and color-coded cards. 14 This complicated process determines who gets counted as poor, how different degrees of deprivation are ranked, and who gets excluded from receiving government benefits. Public response to this process ranges from contesting the selection of certain beneficiaries, to demanding a finer calibration of the process, and even critiquing it for fundamental flaws. People often propose accounts of poverty for strategic purposes, revealing their covert desire to be   Rs. 12,000. 15 This is the guideline that we've been given to decide the poverty status. We also have the details of households having telephone connections or mobile phones. Those who have these cannot be considered as BPL.
GP representatives also use discussions in the GS to help the public negotiate the labyrinth of government scores attached to each disadvantaging characteristic, such as having a physical disability or a socio-economic handicap (as, for instance, having multiple unmarried daughters). For example, a household with a physically disabled person is allotted 10 points; a household with two unmarried daughters is awarded 15 points; and an SC/ST household is awarded 10 points. The goal of this exercise is to help individuals convert their subjective experience of deprivation into a poverty score that can be aligned in a rank order of privation, with benefits going to applicants with the highest score.
In GS discussions, participants also weigh the merits and demerits of different methodologies for determining the status of the poor. In the following example from Palakkad (KL), we observe a president attempting to explain the shift from determination by local knowledge and personal preference to impersonal, objective criteria expressed as numbers.
GP President (male): Now, marks are allotted to each applicant. Previously, when Vasu and Chaclo Chetan were presidents, we used to give benefits according to our wish. We knew who the poor people were, and we used to give them the benefits. But now the government has made some rules and regulations based on which marks are allotted to applicants. It's not like [school] teachers giving extra marks to children they like. Here there are rules. And only based on that, marks are allotted for each benefit.… If you have any doubts with the marks allotted to you and others, then we can certainly check it out....
In politically mature contexts, GP presidents and ward members sometimes use the discursive space of the GS to critique these federally defined poverty parameters and point out the flaws in the beneficiary selection criteria. In the following excerpt from a GS in Dakshin Kannada (KN), a specific aspect of the poverty parameters comes under criticism.
We can ask that the Government may have gifted a phone to a poor, aged man. Looking at this phone you can't declare that he's well off. 16 They should try to give the benefits to the right person.
Similarly, in Kasargod (KL) a discussant points out a fundamental flaw in the housing allotment policy. This policy, which allocates house-building grants to those who do not possess a shelter, ironically overlooks landless people.
Ward Member (male): There are many defects in selecting beneficiaries. Usually, we give houses and toilets to persons who have land. If we select a person who doesn't have land, then we can't give them a house. Even if people say that they'll own land after partition It serves, perhaps, as a stepping stone to inculcating skills of rational and critical argumentation in the manner imagined by the rational democrats who believe in the merits of deliberation.
Overall, in the Indian GS, frequently, competition prevails over consensual deliberation. As frameworks of decision-making regarding public goods allocation and determination of ends and means choices, competition and deliberation differ dramatically at least in the following four broad ways: (i) The competition in this case is governed by rules of commensuration and selection set by the Government, rules which award priority to certain characteristics over others. In some cases, caste identity trumps economic indicators (on the assumption that there is a positive association between the two, which is generally a correct assumption in India but ignores the question of the upper caste poor). In other cases, like BPL lists, economic criteria are given precedence over caste. The Government occupies the role of referee, setting and enforcing the rules.
While these rules are not open to negotiation, from time to time they are questioned by groups and individuals who are left behind. In contrast, the deliberative process generously assumes all people to have equal capacity to articulate their arguments, grants everyone equal rights to do so, and therefore privileges none. (ii) In the competition for resources as carried out in GS, citizens generally address their demands and pleas to the GP president, who represents the Government. Although not physically present, the Government's paternalistic authority is recognized as the invisible power deciding winners and losers and bestowing goods and services accordingly. Members of the public usually do not address each other as they would in a true deliberative structure, where participants are expected to establish a dialogue and to jointly evaluate each another's arguments. (iii) Seeking to achieve advantage in the competition for resources, participants in GS meetings use a plethora of articulation strategies. Argument based on reasons is only one of many such strategies observed in GS meetings and is used less frequently than some others. More popular are blatant personal demands, pleas, and deferential requests for private and public goods. In deliberative democracy, however, reason is the only acceptable form for discursive negotiation. And the question of what constitutes reason is unproblematic. (iv) In a competition, unlike deliberation, it is not necessary to arrive at a consensus about the ends and means to be pursued or the fairness of the final distribution. In fact, the logic of competition is contrary to any attempt to recognize the merits of the opponent's demands, which could only weaken one's own claim. In addition, the government's financial constraints and programs determine the parameters of the discussion. No decision can be implemented without the approval of the higher authority, regardless of the strength of the consensus.
In such a setting, we have a case not of deliberative democracy but of discursive competition that requires individuals and groups to declare their demands in the hopes of being heard. Those aspiring to be heard, recognized, and responded to employ various discursive means: pleading helplessness, drowning out competing voices, arguing raucously, threatening protest, and discussing a well-considered list of reasons. This is the version of democracy that actually prevails in the grassroots of India, where different caste groups live and fight cheek-by-jowl and suffer from varying levels of economic deprivation. In this scenario, competitive speech acts represent a vernacularized style of participation in democratic decision-making, which departs from the idealized deliberative style. The scene most resembles a courtroom, where the goal is to win by influencing the opinions of the judge and jury, who in turn determine who will win and who will lose. But even though it departs greatly from the ideal-typical model of deliberative democracy, this exercise toward discursive engagement in the redistributive mechanism is still very valuable as a way of cultivating a capacity for civic and political engagement among the poor and socially marginalized. This engagement also has the potential of helping them to build a wide repertoire of discursive styles over time, which may lead to their greater voice and agency and to better governance as a whole.

CONCLUSION
This paper highlights a relatively neglected aspect of the relationship between culture and poverty; of culture as a relational, discursive process, which is both affected by economic, social and political inequalities and, in turn, influences them. If pro-poor policy is largely driven by the "equality of opportunity" principle, we argue that this needs to be supplemented by "equality of agency" (Rao and Walton, 2004) which facilitates processes that give voice and agency to the poor. Indeed, cultural processes can be shaped by public action to ameliorate poverty and inequality. The focus of our paper is an important constitutional amendment in India that attempted precisely such a transformation by instituting deliberative forums, called gram sabhas, in all 2 million Indian villages. They were empowered to make important decisions on the selection of public goods and beneficiaries for anti-poverty programs. Within the context of durable inequalities of caste, gender, and religion, these public forums, which can be thought of as a state-engineered public sphere, provide a deliberative space where the boundary between state and civil society is blurred. Challenging several western notions of the "public sphere," the gram sabha is hardly a place where participants engage in rational negotiation to reach a consensus with single-peaked preferences. Public deliberation in the gram sabha is rather a competition between groups and individuals who want a piece of the public pie and employ a wide repertoire of discursive techniques to make their voices heard. Participants are not interested in reaching a consensus but rather seek to stake their particular claims to the "gifts" of the state.
Each petitioner argues that he or she most deserves the benefit, whether private necessities or public goods. Each strives to make the decision-makers hear and grant their plea. The result is a competition based on identity rather than reason. In this way poverty and social inequality have shaped a vernacular style of competitive discursive negotiation within gram sabhas. The competition for state sponsored benefits, moreover, takes place in the context of changing governmental imperatives, which shift with amendments to policies designed to correct social and economic inequalities. Federal and state "schedules," for instance, are used to target benefits to discriminated castes, and quantitative surveys to identify citizens living below the "poverty line." Yet as Scott (1999) argues, the rush to classify human populations results in a Cartesian logic that forces geometric patterns on categories that are inherently fuzzy. The gram sabha allows this fuzziness to be expressed and observed, if not by the "high state" far away in the state capital, then, at least, in the more proximate gram panchayat. The gram sabha also allows those affected by these policies to express their dismay when state categories fail to take into account the realities they are familiar with. Since one of the functions of the gram sabha is to ratify BPL classifications and to voice complaints about the denial of benefits, it provides a forum where public discourse shapes the meaning of poverty, discrimination, and affirmative action.
As a vehicle for expressing rural India's understanding of poverty and the state's attempts to address it, the gram sabha does not always accord with the proximal interests of the state. It rather creates a new "political culture" located within the intersections of the state, the village, and the local matrix of embedded social relationships. By providing a space where opinions can be voiced with relative freedom, (a temporarily level discursive playing field), these local forums also help teach people to engage and debate, and to question decisions and definitions of the Government and the gram panchayat. In this sense, the gram sabha has become an arena where poor, low-caste villagers, male and female, participate and seek dignity. This process resembles the nature of the higherlevel democratic institutions in India, which, some have argued, have been more effective as vehicles for acquiring social dignity rather than economic mobility (Varshney 2000).
What can India's experience with the gram sabha teach us about policies to improve the deliberative capacity of the poor? First, it teaches us that rituals such as the gram sabha work when they are predictable and regular. Regularity ensures that interactions between people and groups have to accommodate this new space where all citizens, regardless of class, caste, or gender, can voice their opinions publicly in way that holds the local state accountable. If deliberative forums were temporary, ad-hoc, events then they would be much more easily ignored, manipulated, and rendered ineffective.
Second, it shows that participatory forums, in order to provide the right incentive for participation, must have clout. For instance, gram sabhas are empowered to discuss village budgets and select beneficiaries for public programs. This makes them worth attending, and the more clout gram sabhas acquire the higher the incentive for citizens to attend them, which reinforces their credibility. Third, it indicates that from the perspective of inculcating voice, any policy effort must pay attention to the relation between the deliberative and the electoral space. Gram sabhas are a deliberative space embedded within an electoral space forcing local politicians to allow all groups to speak lest they lose votes.
However, deliberative rituals are potentially contentious. By allowing marginal groups the space to voice their concerns, they permit previous hidden transcripts to become public, forcing public discussion on issues that people would rather avoid. They can also shift political power by creating political coalitions between like-minded groups that social norms might have previously prevented from collaborating. All this can disturb the social equilibrium, and increase the potential for conflict. In some instances they may also have the ironic effect of making village's less governable by reducing the effectiveness of benevolent local dictators.
Culture therefore is not, as we are often told, a primordially fixed, historically endowed, explanatory variable that is highly resistant to change. It is a relational, communicative process that can be influenced by public policy in a manner that can result in both psychic and material advantages for the poor. Human relations are inherently stratified and these stratifications are reinforced by acts of exclusion and discrimination that create adverse "terms of recognition" for the poor, depriving them of social, political and cultural capital. If we recognize such non-economic capitals to be equally important as human and physical capital in creating durable inequalities, then it is imperative to acknowledge that equalizing voice and civic/political participation require just as much attention from public policy as improving access to education and employment. But, processes that equalize voice and agency are necessarily cultural; they are relational and communicative, creating the capacity for the disadvantaged to crossover from being passive recipients of public largesse towards becoming active participants in determining their own destinies 17 . Finding effective mechanisms to give the poor voice and political agency is, therefore, an important way by which a cultural lens can help inform public policy. 17 Gibson and Woolcock (2008) call this "the capacity to engage".