Marginalized Voices, Indigenous Knowledge Systems and Community Media: A Complexity Framework for Community-Engaged Research in Contentious Times

Abstract

Amid rising authoritarianism, the criminalization of dissent, and the erasure of indigenous expertise from policy discourse, this paper examines community-engaged research as critical public scholarship through the Deccan Development Society’s (DDS) four-decade-plus collaboration with approximately 5000 Dalit and indigenous women farmers in Telangana, India. Against a backdrop of shrinking civic space and technocratic development agendas that delegitimize marginalized knowledge, DDS women’s sanghams have consolidated village-level assemblies into organs of local governance that claim autonomy over land, seed, food, and knowledge—positioning marginalized women as producers of evidence and theory rather than mere data sources. Drawing on complexity theory and participatory action research (PAR), this paper theorizes a community-based inquiry model where marginalized women operate as co-researchers, knowledge producers, and media creators within an emergent system characterized by distributed agency and non-hierarchical information flows. The paper interrogates key methodological, ethical, and political dynamics of critical community-based inquiry for justice and animates the concept of “autonomy as counter-policy” as a contribution to engaged scholarship and epistemic justice.

1. Introduction

Around the world, authoritarian governments are pushing back against communities that hold traditional, local knowledge. Dissent is being criminalized (Gilliard 2025). Indigenous ecological knowledge is being removed from policy. “Indigenous peoples have been marginalized and invisible—both historically and in modern times—including here in the United States. This involves the exclusion of our Indigenous knowledge, as it is often denied or misunderstood by Western culture and the mainstream scientific community (Francis 2021). The space for civil society is shrinking. At the same time, community-engaged researchers face a difficult double task: they need to produce rigorous scholarship and resist the erasure of the very knowledge they are trying to document.

This paper responds to that challenge by examining more than four decades of work by the Deccan Development Society (DDS) in Telangana, India.1 Based in Pastapur, near Zaheerabad, in Telangana, DDS has worked alongside approximately 5000 Dalit and indigenous women farmers, helping them transform small village mutual-aid groups into genuine organs of local governance. These groups—called sanghams—collectively control land, seeds, food systems, and knowledge, in direct opposition to corporate agriculture and government-sponsored technocratic development (DDSIndia.org).

The paper makes three connected arguments. First, it theorizes how the sangham network works as a complex adaptive system—a self-organizing structure where marginalized women function as co-investigators, knowledge producers, and media creators (Muppidi 2001). Second, it documents the ethical and methodological challenges of doing participatory research when the communities’ practices (like saving and exchanging seeds) are under threat of criminalization. Third, it develops the concept of “autonomy as counter-policy” to describe how sangham-generated evidence challenges dominant knowledge systems and reframes policy advocacy around what communities themselves say matters (Wallerstein and Duran 2006).

These three arguments are organized around a single central research question, addressed explicitly in Section 3 and answered cumulatively across Section 2, Section 3 and Section 4: what does community-based inquiry actually look like when it is genuinely accountable to the communities it serves, rather than to the communities’ consent alone? Section 2 develops complexity theory, indigenous and decolonial methodology, and participatory action research as the analytical vocabulary needed to answer that question; Section 3 examines what accountable community-based inquiry looks like in practice, across participatory media, the tensions between academic and community knowledge production, and ethical design under conditions of criminalization; and Section 4 develops “autonomy as counter-policy” as the paper’s theoretical answer to what such accountability makes institutionally possible—not simply an alternative form of knowledge production, but a mechanism through which that knowledge contests and redirects formal policy processes.

1.1. Situating the Research: DDS and the Sangham Model

Founded in 1983 in the Medak district of what is now Telangana, DDS has pursued a model of development grounded in community self-determination. The sangham network—comprising over 5000 women organized into village-level collectives—has become the primary vehicle through which DDS pursues agroecological restoration, food sovereignty, and the revitalization of indigenous seed systems (DDSIndia.org; Lush Spring Prize 2025).

“Their main objectives lie in the consolidation of villages and community governance, an epitome of local needs and priorities. These unsung environmentalists who are Dalits and underprivileged are inspiration for many of us amidst global environmental challenges. The objective which they aim to attain through their model is food security, ecological agriculture and alternate education. The other innovative practices carried out by them are research on non-pesticide management, soil fertility and organic farming. They are enthusiastically engaged in promulgation of green education and green banks.”

“The grassroots agricultural practices evolved by them have significantly contributed towards maintaining ecological biodiversity, turning degraded lands into productive ones and managing the water sources. These farmers of the Deccan region and their novel initiative are all about their way of farming with clear preferences for crops suitable for cultivation in the region. This organization has helped thousands of women to reclaim their farmlands and also generated employment for them”.

(Varun 2020)

Over four decades, sanghams have built grain banks, reclaimed commons, established community biodiversity registers, and produced an extensive archive of participatory media documenting their agricultural expertise.

1.2. Positionality and the Ethics of Engaged Scholarship

The account offered in this paper is, by necessity, a partial one—partial in the sense of being incomplete, and partial in the sense of being situated. The author’s engagement with DDS began long before any formal research agenda. As a graduate student in the Master’s program at the Sarojini Naidu School of Performing Arts, Fine Arts, and Communication (S.N.School) at the University of Hyderabad, India, from 1989–1991—one of the program’s earliest cohorts—the author was first introduced to Pastapur and the Deccan Development Society through the work of P.V. Satheesh, a distinguished television producer and development communication practitioner who had joined the program as a visiting faculty member. The S.N. School, named in honor of the Indian poet and freedom fighter Sarojini Naidu, was itself an experiment in interdisciplinary, integrated arts education, and it was within that intellectually generative environment that DDS first entered the author’s scholarly horizon.

The relationship deepened upon the author’s graduation and later return to the S.N. School as a faculty member. By that point, DDS, with P.V. Satheesh as a co-founder and Director, had become a central pedagogical resource for the development communication curriculum, and the NGO’s experimental use of the jatra as a community communication tool provided a concrete field site for student inquiry. The jatra—traditionally a religious fair characterized by carnival-like communal celebration—had been strategically reimagined by DDS as a vehicle for social and health messaging, delivered across clusters of approximately ten villages at a time. Recognizing this as a significant methodological innovation in community communication, the author was part of a college team of three faculty members that organized four weekend field visits with students to study the jatras as an embedded research project (Krishnatray et al. 1993). Using interviews, focus groups, and ethnographic analysis, the team produced a white paper documenting the sanghams’ structure, their communication practices, the substance of their messaging, and the facilitative role DDS played in coordinating the events. It was during these visits—sitting with sangham members, listening to their accounts of collective decision-making, watching them navigate the tensions between community priorities and external pressures—that the author began to form a ground-level understanding of what the sanghams actually do. The jatra model, it would later become clear, was a precursor to subsequent DDS initiatives including the pantala panduga and other community media experiments that followed (DDSIndia.org).

What follows is the author’s perspective on the work of the sangham women: It was shaped by an initial encounter as a graduate student in 1991, deepened through subsequent conversations with DDS practitioners and sangham members in 1993, and grounded in personal observations gathered across multiple visits to India over the years.

That understanding has been refined, complicated, and occasionally overturned through later conversations and observations during subsequent visits to India. Sangham women spoke with candor about the limitations as well as the achievements of their collectives; DDS staff, especially P.V. Satheesh, offered perspectives, in confidence, that did not always align neatly with the organization’s published accounts; and the communities themselves were never the uniform, harmonious entities that advocacy literature sometimes implies. What the author observed was messier, more contested, and ultimately more compelling than any idealized portrait—and it is that observed reality, filtered through the author’s own disciplinary training and positional limitations, that animates the analysis presented here. It does not claim to represent the sangham women’s own theoretical frameworks, which are richer, more complex, and more internally varied than any single scholarly account can capture. It claims, more modestly, to document what the author witnessed, heard, and came to understand—and to theorize from that position with honesty about its limits.

This positionality carries real obligations: to represent community priorities accurately, to be honest about the power imbalances between academic institutions and grassroots organizing, and to resist the extractive tendencies that have historically characterized research conducted on rather than with marginalized communities (McHugh et al. 2024). Where the author’s interpretation diverges from what sangham members or DDS staff have expressed directly, that divergence is acknowledged rather than papered over. The goal is not a seamless narrative but an honest one.

1.3. Organization of the Paper

Section 2 details the methodological and ethical framework. Section 3 presents findings across three analytical dimensions: participatory media production as counter-ethnography; productive tensions between academic and community knowledge production; and ethical design for communities documenting criminalized practices. Section 4 develops the theoretical discussion, including the concept of “autonomy as counter-policy.” Section 5 draws conclusions with implications for community-engaged partnership, organizing, and policy. Across these sections, the argument moves from theoretical framework (Section 2) to empirical demonstration (Section 3) to theoretical synthesis (Section 4): each section answers the central research question introduced above at a different level of analysis, culminating in “autonomy as counter-policy” as the paper’s account of what genuine community accountability makes institutionally possible.

2. Materials and Methods—Methods and Ethics

As mentioned earlier, this study draws on four decades of DDS’s documented practice, collaborative fieldwork conducted across field visits between 1993 and 2024, participatory documentation of agroecological practice, and systematic analysis of media produced through the Community Media Trust (est. 2001) and Sangham Radio (est. 2008). The methodological framework brings together four traditions: indigenous and decolonial research methodologies, complexity theory, participatory action research, and critical media studies.

2.1. Indigenous and Decolonial Methodological Foundations

This study is situated within a broader body of global-majority scholarship on indigenous knowledges, decolonial methodology, and community-engaged research—scholarship developed largely outside the Indian context but that speaks directly to the epistemic claims sangham women make about their own knowledge. Smith’s (2021) foundational critique in Decolonizing Methodologies: Research and Indigenous Peoples argues that, for indigenous and colonized peoples, “research” is itself a loaded term—historically a technology through which Western academic institutions extracted, classified, and claimed authority over indigenous knowledge while denying indigenous peoples’ own standing as knowers and theorists. Smith’s proposed indigenous research agenda, organized around principles of survival, recovery, and self-determination, insists that communities retain control over what is studied, how it is studied, and who benefits from the resulting knowledge. That insistence directly informs this paper’s positioning of sangham women as producers of evidence and theory rather than as data sources (Section 1.2), and it is echoed methodologically throughout this study’s reliance on DDS’s own community-generated documentation rather than researcher-imposed data collection instruments.

Chilisa’s (2020) Indigenous Research Methodologies extends this critique through a postcolonial indigenous paradigm grounded in relational ontology, epistemology, and axiology—the premise that knowledge is inseparable from relationships among people, land, ancestors, and community, and that a researcher’s accountability runs first to those relationships rather than to disciplinary convention. Chilisa’s framework, developed primarily from Southern African contexts, resonates strongly with the sangham model’s own theorization of knowledge as something produced collectively in assembly rather than extracted individually through interview or survey (Section 2.2, “Collective Deliberative Judgment in Assembly”). At the same time, caste-based oppression in India is not reducible to the settler-colonial and postcolonial dynamics that ground Smith’s and Chilisa’s work; Dalit women’s exclusion operates through an internal, centuries-old hierarchy of ritual purity and social stratification that predates and operates alongside colonial rule. Engaging Smith and Chilisa here is not meant to flatten that distinction, but to situate DDS’s sangham-based knowledge production within, and as a distinct contribution to, a wider global-majority effort to reassert community epistemic authority against knowledge hierarchies—whether colonial, caste-based, or both—that have historically rendered such knowledge illegitimate.

Recent scholarship on community-engaged research in the Global South sharpens the methodological stakes further. Mohammed’s (2025a) analysis of knowledge production in the Global South argues that even self-consciously critical or participatory research often reproduces colonial epistemic hierarchies when Western theoretical vocabularies and institutional metrics remain the implicit standard against which community knowledge is judged; genuine decolonization of knowledge production, in her account, requires centering the conceptual categories communities use to understand their own experience. Mohammed’s (2025b) companion chapter on navigating power dynamics in community-centered research design offers practical guidance largely consistent with the approach adopted here: long-term relationship-building prior to any formal research agenda, iterative negotiation of what is documented and disseminated, and researcher accountability that persists across the life of the relationship rather than terminating at data collection. The decades-long, relationship-first trajectory described in Section 1.2—beginning as a graduate student in 1991 and continuing through repeated returns over more than three decades—reflects this kind of accountability, though the paper is also attentive to Mohammed’s caution that even long engagement does not, by itself, dissolve the power asymmetries between an academically credentialed researcher and non-literate Dalit farmers; those asymmetries are managed, not eliminated, through the community-sovereignty protocols detailed in Section 2.5.

These epistemic commitments carry institutional costs that are themselves documented in the global-majority literature on community-engaged scholarship. Bell and Lewis (2023) identify four structural barriers that discourage community-engaged work even at universities that formally endorse it: epistemic biases that privilege credentialed academic knowledge over community-generated expertise; neoliberal incentive structures that reward metrics-driven publication over sustained public engagement; gendered norms that devalue relational and service-oriented labor; and colonial-racist defaults that marginalize non-Western knowledge systems and the scholars who work with them. Each of these dynamics surfaces later in this paper’s account of the tensions between DDS’s community-facing knowledge outputs and conventional academic metrics of impact (Section 3.2 and Section 4.3); flagging them here, alongside Smith, Chilisa, and Mohammed, situates those later findings within an established body of scholarship on the institutional politics of engaged, community-centered knowledge production across global-majority contexts, rather than treating them as idiosyncratic features of the DDS case alone.

2.2. Complexity Theory as Analytical Framework

Complexity theory helps us understand the sangham network as a self-organizing system where agency is spread across many nodes rather than concentrated at the top. Unlike conventional development models that assume experts pass knowledge down to beneficiaries, complexity frameworks pay attention to feedback loops, non-linearity, and the productive power of apparent disorder (Stacey 2001; Snowden and Boone 2007). Applied to DDS sangham networks, complexity theory lets us analyze how sangham assemblies—locally rooted and individually bounded—produce system-wide effects: the revival of seed diversity across a whole region, the reconstruction of food provisioning systems, and the creation of policy-relevant evidence from the ground up.

Complexity theory is deployed here rather than more conventional social movement or institutional frameworks because it foregrounds precisely the features of the sangham network this paper needs to explain: distributed rather than centralized agency, emergent rather than planned system-wide effects, and resilience that survives the loss or capture of any single node (Section 4.2). Institutional theory tends to explain how organizations reproduce stable structures and would struggle to account for a network that has changed form repeatedly across four decades without central direction; classical social movement theory, with its emphasis on framing, resource mobilization, and political opportunity structures, is well suited to explaining why movements emerge and sustain themselves but is less equipped to explain how a movement’s day-to-day organizational practice can itself constitute the alternative social order it advocates for—the dynamic this paper later develops, following Yates (2015), as prefigurative politics (Section 4.1). Complexity theory’s core commitments—self-organization, emergence, and non-linear feedback between local practice and system-wide effect (Cilliers 1998; Byrne and Callaghan 2014)—provide the analytical bridge between these two observations: it explains both why the sangham network holds together without central control and how its internal organizational logic can itself function as a form of political argument, a connection developed further in the discussion of “autonomy as counter-policy” (Section 4.1). Chesters and Welsh’s (2006) application of complexity theory to social movements is instructive here: they argue that movements characterized by distributed, self-organizing network structures are better understood as complex adaptive systems than as unified strategic actors, precisely because their capacity for large-scale effect emerges from the aggregation of locally autonomous action rather than from centralized command—an argument this paper extends to the specific case of a Dalit and indigenous women’s farming network.

Complexity theory’s application to social systems requires careful differentiation from its antecedents in classical systems theory. Where systems theory—as developed through the cybernetic tradition (Wiener 1948) and later structural-functionalism (Parsons 1951)—tended to privilege equilibrium, boundary maintenance, and top-down regulatory feedback, complexity theory attends instead to emergence, non-linearity, and the generative potential of disorder. In social contexts, this distinction matters enormously: a systems-theoretic reading of the sangham network might reduce it to a functional subsystem performing integration tasks within a larger development apparatus, whereas a complexity reading reveals it as a self-organizing formation that produces genuinely novel social arrangements—new forms of governance, new knowledge infrastructures, new policy-relevant evidence—that cannot be predicted from or reduced to the properties of its individual components. Emirbayer and Mische’s (1998) relational account of human agency is particularly useful here. Their argument that agency is neither a fixed property of individuals nor a simple effect of structural position, but rather a temporally embedded process oriented simultaneously toward past habits, present judgments, and future projections, maps onto the sangham model with considerable precision: sangham women draw on locally inherited agroecological knowledge, exercise collective deliberative judgment in assembly, and orient their organizing toward visions of food sovereignty and epistemic recognition that do not yet fully exist. Agency, in this account, is not something the sanghams have; it is something they continuously produce (Muppidi 2001).

Collective Deliberative Judgment in Assembly

The sanghams function as what DDS explicitly describes as “vibrant organs of primary local self-governance.” This is not just aspirational language—it describes a specific decision-making structure. Each village sangham is a regular assembly where members deliberate collectively about farming, resource use, credit, health, and food distribution. Decisions emerge through bottom-up participatory processes rather than top-down directives, which means women who are largely non-literate and from Dalit (formerly “untouchable”) communities are the actual authors of collective choices (Muppidi 2001).

A revealing example of what this looks like in practice is the sangham’s response to the Indian Public Distribution System (PDS). The World Bank successfully pressured the Indian government to scale back the PDS in 1992, restricting access and reducing its scope—a decision made at the highest levels of global financial governance, affecting millions of poor rural households. Rather than simply endure this, the sanghams collectively deliberated and decided to build their own alternative (Personal conversation with P.V. Satheesh). In response, the sangham women organized their own Alternative Public Distribution System (APDS), based on their locally administered Community Grain Funds, involving over three thousand women in fifty villages who produce more than a million kilograms of extra sorghum annually and distribute it among themselves. This is a direct institutional counter to a top-down policy decision—arrived at through assembly, not hierarchy (Mookerjea 2010).

Inherited Agroecological Knowledge

The “inherited” knowledge at stake here is very specific: it concerns knowledge about the cultural and geological context, which crops to grow, in what combinations, under what rainfall conditions, what health issues consuming them help prevent/address, how to store and share seed varieties, and what nutritional value different millets carry (Krishnatray et al. 1993). This knowledge is embodied and oral, held primarily by women, and has been systematically devalued by Green Revolution-era agricultural policy that pushed wheat and rice monocultures. DDS’s seed catalogue documentation effort captures sangham women’s knowledge about when to sow each variety based on weather conditions, what foods can be prepared from each crop, and the multiple social and ecological values of each cultivar—knowledge that is described as “fast vanishing away from villages and agrarian landscapes” (Mookerjea 2010).

The sanghams have institutionalized this knowledge through Community Gene Banks (also called Gene Funds). In sixty villages, more than eighty species of cultivars have been retrieved from extinction and conserved as collective public property, through an Agro-biodiversity Register designed by the women themselves for non-literate farmers—which was subsequently adopted into the Indian National Biodiversity Strategy and Action Plan. That last detail is important: knowledge produced by non-literate Dalit women farming in a semi-arid region was formally recognized as authoritative at the level of national policy (Mookerjea 2010).

Women farmers set up seed banks where they collect and store many varieties of millets and other locally grown food crops, loan seeds to members, and establish repayment mechanisms in the form of seeds rather than money. This seed-lending system encodes ecological knowledge (which varieties suit local conditions) within a social institution (mutual obligation among members) (Dutta and Thaker 2020).

Visions of Food Sovereignty and Epistemic Recognition “Not Yet Fully Existing”

The phrase “not yet fully existing” is meant to capture something real about the sangham project: it is not simply restoring a past practice but working toward a form of agricultural and social life that does not exist anywhere fully, combining traditional knowledge with new organizational forms in a world where the political and economic conditions for real food sovereignty have not yet been achieved.

The sanghams articulate this as a “cycle of autonomies”—not a single goal but an interlocking set of aspirations: food sovereignty includes community gene banks, control over natural resources, control over markets, and a “green education system” that includes community preschools, a school for working children, intensive workshops for adult women, and village night schools—all understood as parts of a single integrated vision (International Newsletter for Sustainable Local Development Newsletter #94 2012).

The epistemic recognition dimension means the sanghams are not just farming differently; they are asserting that their knowledge counts as knowledge. Sangham Radio—India’s first community radio and first all-women Dalit-run radio station—focuses on food sovereignty, seed sovereignty, and retrieving local culture and traditions in food, farming, health, and ecology, explicitly positioning indigenous agricultural knowledge as something worth broadcasting and preserving, not just surviving quietly. The Community Media Trust has developed what they call the “Sangham shot” as opposed to the “Patel shot”—where the Patel shot refers to the top-down, power-imbalanced approach to media production, and the Sangham shot refers to media creation and consumption guided by equality (personal conversations with Sangham members, and P.V. Satheesh). The terminology is telling: even the camera angle is theorized as a political question.

Read together, these examples illustrate the complexity-theoretic dynamics introduced above rather than simply cataloguing sangham activities. Collective deliberative judgment in assembly exemplifies self-organization: the Alternative Public Distribution System emerged from village-level deliberation with no central directive, yet aggregated into a regional system moving more than a million kilograms of sorghum annually. Community gene banks and the Agro-biodiversity Register exemplify emergence and non-linear feedback: locally maintained seed knowledge, accumulated independently across dispersed villages, became the input for a national policy template that no single village could have produced alone. And the “cycle of autonomies”—the refusal to treat food sovereignty, resource control, and epistemic recognition as separable goals—exemplifies the kind of system-wide, mutually reinforcing effect that complexity theory, rather than a linear causal model, is suited to describe. It is this same self-organizing, distributed architecture, examined at the level of organizational form rather than outcome, that Section 4.1 develops as prefigurative politics and, subsequently, as autonomy as counter-policy.

Finally, it is worth noting that this vision remains contested and incomplete. Recent reports indicate that resource constraints and generational changes have reduced the number of active sanghams, and some programs like the alternative PDS no longer function—with the organization currently deliberating about how to combine old and new elements, including exploring livelihoods that blend traditional agriculture with opportunities in manufacturing and services. This ongoing uncertainty is itself part of what makes the claim that agency is “continuously produced” rather than possessed a meaningful one: the sangham project is not a settled achievement but an ongoing practice of collective self-determination under real constraints.

That production, however, cannot be fully theorized without attending to the epistemic dimensions of complexity in contexts of colonial and postcolonial power. Santos’s (2014) concept of the “epistemologies of the South” provides a critical complement to complexity theory by insisting that the suppression of non-Western knowledge systems is not merely a cultural loss but a structural feature of global power—what he terms “epistemicide,” the deliberate destruction of alternative ways of knowing in the service of colonial and neoliberal hegemony. The DDS sangham network operates precisely in the space that epistemicide has tried to foreclose: it preserves, systematizes, and politically asserts indigenous agroecological knowledge at the moment when national and international corporate seed regimes and technocratic development agendas are most aggressively seeking to render that knowledge obsolete or illegal. Mignolo’s (2011) concept of “decolonial epistemology” sharpens this further, arguing that genuine epistemic alternatives require not merely the inclusion of marginalized knowledge within existing frameworks but the delinking from the colonial matrix of power that structures what counts as knowledge in the first place. Read together, Santos and Mignolo suggest that the complexity of the sangham network is not simply organizational—it is epistemic. The network’s distributed, non-hierarchical architecture is not merely an efficient design choice; it is a structural refusal of the knowledge hierarchies that colonial and postcolonial development institutions have sought to impose at various levels from the local to the global. Complexity theory, in this application, is not a neutral analytical lens but a theoretically committed one: it is being used to make visible what linear, hierarchical, and extractive frameworks are designed to obscure.

2.3. Participatory Action Research

PAR methodology treats community members not as research subjects but as co-investigators who define research questions, collect and interpret evidence, and control how findings are shared (Fals Borda 1987; Cornwall and Jewkes 1995). Within DDS, PAR has taken the form of sangham-led agricultural experiments, farmer-run seed trials, community nutrition surveys, and participatory video documentation. The research reported here builds on this accumulated body of co-created inquiry, supplemented by the author’s own fieldwork and archival analysis of DDS materials spanning several decades (Krishnatray et al. 1993; Muppidi 2001).

2.4. Community Media as Methodological Infrastructure

The Community Media Trust and Sangham Radio—India’s first community radio station and its first Dalit-run, all-women radio station—are both research outputs and methodological tools. Sangham women filmmakers have produced community videos, engaging local communities in debates around food and seed sovereignty, autonomy over natural resources, and the role of markets and media, and offering critical evaluations of state policies—meaning the sanghams do not just implement decisions; they collectively reason about and contest the policy frameworks that govern their lives. Their programs, scripts, production processes, and audience engagement records provide a rich body of material for analysis. These media forms produce, circulate, and preserve community knowledge (On-Site observations and personal conversations with Sangham members and P.V. Satheesh).

The Community Media Trust’s video archive represents one of the most extensive participatory media collections produced by and for a marginalized farming community anywhere in the Global South. Produced primarily in Telugu and local Telangana dialects—with select materials subtitled or dubbed for wider distribution—the archive encompasses documentation of seed diversity trials, millet cultivation techniques, traditional water harvesting systems, community nutrition practices, women’s governance deliberations, and sangham assembly proceedings. Distribution has operated through multiple channels simultaneously: screenings at neighboring village sanghams as popular education; submission to state and national agricultural policy bodies as advocacy evidence; circulation through development sector networks and NGO partnerships; and, more recently, digital archiving through institutional collaborations that are available via DVD’s and streamed on the NGO’s website at DDSindia.org. This multi-channel distribution strategy is not incidental—it reflects a deliberate architecture of knowledge circulation designed to serve community mobilization, peer learning, and policy intervention as parallel rather than sequential goals. The archive is not a repository waiting to be activated; it is itself an ongoing act of epistemic assertion.

Sangham Radio extends and deepens this architecture into the domain of live, interactive broadcast. Sangham members meet to listen to moderated radio broadcasts and also individually, at other times. The station broadcasts to the sangham communities in Telugu and local dialects across programming categories that include agroecological knowledge sharing, seed sovereignty advocacy, women’s health and nutrition, community governance updates, cultural programming, and call-in formats that enable real-time listener feedback and peer exchange. That feedback mechanism is methodologically significant: it transforms broadcast from a one-to-many transmission into a many-to-many dialogic infrastructure, creating continuous loops between producers and audiences who are members of the same communities being documented. Sangham Radio (and DDS’s broader community media infrastructure) is documented in academic and advocacy literature. Michel Pimbert’s work on food sovereignty and community media, published through IIED (International Institute for Environment and Development), references DDS’s community radio programming in the context of participatory research methodology (Pimbert 2008). The FAO has also referenced DDS’s community media work in publications on indigenous knowledge documentation. This dynamic resonates directly with Rodríguez’s (2001) theorization of citizens’ media—communication forms that are distinguished not by their technical properties but by their capacity to recodify social identities, reconfigure power relations, and enact new forms of citizenship from within communities historically excluded from media production. Dagron’s (2001) foundational account of community radio further contextualizes Sangham Radio within a global tradition of participatory broadcasting oriented toward social transformation rather than audience delivery—a tradition in which the process of production is inseparable from the outcome, because the act of communities telling their own stories is itself a form of development. What distinguishes Sangham Radio within even this tradition is the specificity of its ownership: it is not community radio administered on behalf of a community by outside practitioners, but radio produced, broadcast, and governed by the women whose lives, cultural traditions, community-specific knowledge, political struggles and hyperlocal solutions constitute its entire content.

Community media, in this framework, is not merely a data source but a structural mechanism of the complex system described in Section 2.2: the Community Media Trust archive and Sangham Radio broadcasts function as some of the sangham network’s most visible feedback loops, the channels through which locally produced knowledge circulates across nodes, aggregates into system-wide effects, and returns to inform local practice (Section 4.2). Where Section 2.2 theorizes complexity at the level of governance and knowledge production, community media is the infrastructure that makes those dynamics observable and, as Section 3.1 details, is itself a primary site of the counter-ethnographic knowledge production this paper investigates.

2.5. Ethical Framework and Research Design

Research done with communities whose practices are politically contested demands careful ethical design that goes well beyond standard institutional compliance frameworks. My field visits and informal conversations, over several decades, were in accordance with visits developed collaboratively with P.V. Satheesh, other scholars, and sangham representatives.

This ethical design draws explicitly on the indigenous and decolonial methodological commitments outlined in Section 2.1. Following Chilisa (2020), the guiding standard here is relational accountability rather than procedural compliance: obligations run to the sangham women and to DDS as an institution, not only to an Institutional Review Board, and those obligations are renegotiated continuously rather than fixed at the point of informed consent (see also Section 4.3’s discussion of the limits of IRB-centered ethics). Following Smith (2021), community control extends across the research process rather than being confined to its outset—communities determine not only whether to participate but what is documented, how it circulates, and when documentation stops, a standard reflected in DDS’s own protective principles detailed in Section 3.3.

Mohammed’s (2025b) framework for navigating power dynamics in community-centered research design is instructive for naming what this ethical approach does not resolve. Sustained relationship over decades reduces, but does not eliminate, the asymmetries between an academically credentialed outside researcher and non-literate Dalit sangham women whose livelihoods and legal exposure are directly at stake in what gets documented. In practice, this means treating disagreement and divergence—between the author’s interpretation and what sangham members or DDS staff express directly—as information rather than as noise to be smoothed over (Section 1.2), and it means accepting community veto power over dissemination even when that limits the completeness of the academic account (Section 3.3).

Finally, this ethical framework is offered with an awareness, informed by Bell and Lewis (2023), that community-engaged research of this kind is structurally disadvantaged within the academic institutions that formally endorse it. The extended, non-linear timeline required to build the relational trust described above—and the corresponding reliance on community-defined rather than disciplinary metrics of value—runs directly against the epistemic, neoliberal, gendered, and colonial-racist defaults Bell and Lewis identify in conventional academic reward structures. That tension is not incidental to the ethics of this study; it is, following Mohammed and Chilisa, part of what a genuinely community-accountable ethical framework in a global-majority research context requires acknowledging openly rather than resolving away.

3. Results—Findings

The analysis produces three interconnected clusters of findings, all organized around the central question: what does community-based inquiry really look like when it is genuinely accountable to the communities it serves?

This question is not a departure from the three arguments introduced in Section 1, but their empirical grounding. Section 3.1 examines the complex-systems claim from Section 2.2 at the level of practice, showing how participatory media inverts the conventional relationship between researcher and researched; Section 3.2 examines the methodological claim from Section 2.3 and Section 2.4 by documenting what accountability costs in practice—a divergence from conventional academic knowledge production that this paper treats as evidence, not as a limitation to be minimized; and Section 3.3 examines the ethical claim from Section 2.5 under the specific conditions of political and legal risk that make community accountability urgent rather than aspirational. Read together, these three findings answer the research question empirically, ahead of the more explicitly theoretical answer—“autonomy as counter-policy”—developed in Section 4.

3.1. Participatory Video and Radio as Counter-Ethnography

The most structurally significant finding is that DDS’s media infrastructure flips the conventional ethnographic gaze. Instead of outside researchers documenting sangham women’s practices and translating them for external audiences, the women themselves have been trained as videographers, scriptwriters, and radio producers. They document their own communities, narrate their own expertise, and circulate that knowledge through channels they own and control. They have even adapted the grammar of television production through their own lens and cultural relevance. The terminology for television shots in terms of Sangham shot and Patel shot is a good example of that.

The Community Media Trust’s participatory video methodology has documented and developed archives on seed diversity, millet cultivation, traditional water harvesting, community governance, and women’s health and nutrition. These materials serve simultaneously as community archives, popular education resources for neighboring villages, and advocacy evidence submitted to state and national agricultural policy bodies.

Sangham Radio, broadcasting in Telugu and local dialects, reaches sangham members in the DDS Communities and beyond through multiple channels of distribution. Its significance as a methodological site lies in what it makes possible:

• Knowledge production in community-owned formats, not solely in academic outputs legible only to credentialed audiences;
• Circulation of evidence through horizontal peer networks rather than top-down dissemination;
• Continuous feedback loops between producers and audiences who are members of the same communities being documented;
• Living archival infrastructure that preserves indigenous agricultural knowledge at risk of erasure.

3.2. Productive Tensions Between Academic and Community Knowledge Production

DDS operates with an explicit hierarchy of priorities: materials generated for community mobilization, popular education, and policy advocacy come first—academic publication comes second. This creates real and sometimes uncomfortable tensions for engaged scholars, whose institutional incentive structures reward journal articles and conference presentations over movement-facing outputs. Three aspects of this tension deserve attention.

• Aspect One: Primary Knowledge Outputs Are Community-Facing, Not Academic

First, the primary knowledge outputs of the collaboration are not academic texts but training films on millet cultivation, radio dramas about seed bank governance, illustrated guides to participatory biodiversity assessment, and sangham assembly proceedings. Academic papers are secondary translations of these materials, not the other way around. DDS’s community media trust, operating since the early 1990s, produced documentaries—made entirely by sangham women—which documented community resistance to displacement, genetically modified seed promotional policies, and the viability of dryland food systems. Several of these films circulated through food sovereignty networks internationally and were screened in policy forums, functioning simultaneously as popular education, community mobilization, and international advocacy. It was not produced for or reviewed by academic audiences. Similarly, DDS produced illustrated seed catalogues in Telugu, designed for non-literate or partially literate sangham members, documenting the agronomic properties, culinary uses, and ecological roles of dozens of traditional millet and sorghum varieties maintained in community grain banks. These catalogues were functional tools for seed exchange governance, not academic texts.

Radio dramas addressed seed bank governance, women’s land rights, and pesticide hazards in narrative formats accessible to communities without formal education. The scripts were developed collaboratively with sangham members, drawing on their own experiences, and broadcast schedules were organized around agricultural seasons and sangham meeting cycles.

Academic Translations (Secondary):

Academic engagement with DDS’s work has largely taken the form of researchers drawing on DDS’s community-generated documentation to produce journal articles and book chapters. Michel Pimbert and colleagues produced academic analyses of the Prajateerpu process published in development studies journals, explicitly acknowledging that the primary knowledge outputs—the jury proceedings, the community testimony, the verdict documentation—were produced outside academic frameworks and that the academic papers were interpretive translations of those materials. Similarly, work published in journals such as Agroecology and Sustainable Food Systems and edited volumes on participatory research has cited DDS’s community biodiversity documentation as primary evidence, with the academic text functioning as a secondary frame around community-produced content.

• Aspect Two: Timeline Governed by Community Organizing Calendars, Not Academic Cycles

The timeline of knowledge production is governed by community organizing calendars—planting seasons, cultural festivals, sangham assembly schedules, policy submission windows, and advocacy campaign timelines—not by academic research cycles or publication deadlines. Engaged scholars must develop the patience and institutional flexibility to follow these rhythms.

The most concrete illustration is the structure of DDS’s annual seed sovereignty festival, which brings together sangham women from across villages to exchange seeds, share cultivation knowledge, and collectively assess which varieties require urgent conservation attention. The knowledge produced at this event—about seed performance, variability under climate stress, and community preferences—feeds directly into grain bank governance decisions and biodiversity register updates. This event occurs on an agricultural calendar logic: after harvest, before the next planting cycle, when women can assess the season’s results and make collective decisions about the following year (personal conversations with Sangham members and P.V. Satheesh).

An engaged scholar embedded in this process cannot submit a grant proposal in September, begin fieldwork in January, complete data collection by April, and submit findings in August—the standard academic project cycle. The festivals, sangham events, and celebrations happen when they happen; sangham assemblies are called when community governance requires them; policy submission windows are determined by government consultation timelines that do not consult academic schedules. Researchers working with DDS have described the practical consequence: years-long engagement before sufficient relational trust exists to participate meaningfully in knowledge production processes, and publication timelines that lag community outputs by two to three years, creating difficulties with academic performance review systems that measure annual productivity, if that was the guiding criterion.

The policy submission dimension is equally concrete. When the Indian government opened consultation periods on the Biological Diversity Act’s implementing regulations, DDS’s advocacy response drew on community biodiversity documentation assembled over years of sangham practice. The submission deadline was fixed; the knowledge infrastructure that made a credible submission possible had been built on community timelines entirely independent of that deadline. Academic partners who had not maintained continuous engagement with DDS’s documentation processes could not have contributed meaningfully on short notice. Even this contribution is based on several decades of formal and informal observations and interactions.

• Aspect Three: Success Measured by Capacity Building and Empowerment Indicators

DDS measures success in terms of capacity building and empowerment-very differently from academia. Hectares of millet under cultivation restored, number of traditional seed varieties maintained in community grain banks, women elected to local governance bodies, and documented policy changes won are the primary indicators of impact. DDS’s own reporting documents the restoration of sorghum and millet cultivation across several thousand hectares in Zaheerabad region villages over the course of its community food sovereignty programs. These crops had been displaced by Green Revolution wheat and rice varieties and, subsequently, by Bt cotton.2 The restoration figure is not incidental—it represents a reversal of a decades-long policy-driven agricultural transition, achieved through sangham-organized seed distribution, training, and collective market arrangements that made millet cultivation economically viable again for smallholders. No academic publication produces an equivalent outcome.

Traditional Seed Varieties Maintained in Community Grain Banks:

DDS’s community grain banks collectively maintain over 80 traditional varieties of sorghum, millet, and other dryland crops—varieties that national and international gene banks had not systematically collected because they lacked commercial significance by Green Revolution criteria. The number of varieties actively maintained—as opposed to frozen in ex-situ collections—is itself a measure of living knowledge infrastructure, because maintaining a seed variety requires cultivating it, selecting it, and re-integrating it into community agricultural practice each season. This is an impact metric with no academic equivalent.

Women Elected to Local Governance Bodies:

DDS tracks the number of sangham women who have stood for and won election to gram panchayat bodies—village-level governance institutions. The organizational capacity, public speaking experience, and collective confidence developed through sangham participation directly enabled women who had been structurally excluded from public deliberation to enter formal governance (Muppidi 2001). Several women who began as sangham participants in DDS’s food sovereignty programs subsequently served as elected panchayat members and used those positions to advance community land and water governance priorities. The causal pathway from DDS organizing to formal political representation is traceable and documented in DDS’s own program records, though it appears in no academic impact metric.

Documented Policy Changes Won:

As noted in earlier examples, DDS’s evidentiary interventions contributed to the Andhra Pradesh government’s Bt cotton moratorium recommendation and to the adoption of the Community Biodiversity Register format as the national template. These are discrete, attributable policy outcomes. By contrast, an academic paper documenting the harms of Bt cotton or the value of community biodiversity knowledge—however rigorous—produces no equivalent traceable policy consequence. The difference is not one of knowledge quality but of institutional positioning: DDS’s knowledge infrastructure was built to intervene in policy processes, not to contribute to disciplinary literatures.

3.3. Ethical Design When Documenting Criminalized Practices

India’s 2001 Protection of Plant Varieties and Farmers’ Rights Act provides some legal protection for seed sovereignty—but implementation has been uneven, and subsequent proposed seed legislation has at various points threatened the informal seed-saving and community exchange practices that DDS sanghams have maintained for generations (Gopalakrishnan 2018). Corporate agricultural interests have in some instances treated traditional seed exchange as illegal competition with registered varieties (Hunga 2026).

This legal and political context means that questions of documentation, data custody, dissemination strategy, and community exposure are primary ethical concerns—not afterthoughts. The DDS collaboration has developed several protective principles:

• Community sovereignty over which documented practices are made publicly accessible versus maintained in internal archives;
• Collective sangham decision-making about participation in public advocacy versus management of legal risk;
• Strategic use of community media channels rather than academic databases for materials that could be weaponized against the community;
• Ongoing legal literacy programs so sangham members can assess risks and make informed decisions independently.

These are not merely precautionary measures. They are expressions of the research design’s core commitment: that the communities bearing the risks of documentation must retain control over how that documentation is used.

4. Discussion—Reflections

The findings illuminate several theoretical dynamics with implications for engaged scholars working in contexts of political contention, knowledge suppression, and contested resource governance.

4.1. Autonomy as Counter-Policy

The central theoretical contribution of this paper is the concept of “autonomy as counter-policy.” This describes the mechanism by which DDS sanghams use the infrastructure of local self-governance—farmer-led research juries, women-run grain banks, community biodiversity registers, and movement-generated evidence—to contest state-corporate knowledge hierarchies and reframe policy advocacy on community terms.

“Autonomy as counter-policy” is not mere resistance or refusal. It is generative and institutional. As Appadurai (2006) notes, “it is worth regarding research as a right, albeit of a special kind. This argument requires us to recognize that research is a specialized name for a generalized capacity, the capacity to make disciplined inquiries into those things we need to know, but do not know yet. All human beings are, in this sense, researchers, since all human beings make decisions that require them to make systematic forays beyond their current knowledge horizons” (p. 167).

Sanghams produce evidence, develop theory, train new researchers, and intervene directly in policy processes—not by adopting the epistemic frameworks of the institutions they contest, but by asserting the validity of their own knowledge systems as a precondition for meaningful policy engagement. The farmer-led jury model, termed Prajateerpu, in which sangham women evaluate agricultural research findings against their own multigenerational observational knowledge, is the clearest expression of this dynamic: it inserts community epistemology into the apparatus of knowledge validation itself. As documented by Pimbert and Wakeford (2002), the following is a detailed example of how the jury model operates, from an event organized in 2001 at the DDS krishi Vigyan Kendra (Farmer Liaison Center). It was organized jointly by DDS, the International Institute for Environment and Development (IIED), the Institute of Development Studies (Sussex), the University of Hyderabad, and the All-India National Biodiversity Strategy and Action Plan. The Prajateerpu report, published jointly by IIED and IDS and cited in the Economic and Political Weekly, is the fullest documented account of the jury process. It remains the primary reference for this model.

The Prajateerpu Process

The Prajateerpu process, convened in Andhra Pradesh, India, constitutes a landmark case in participatory democratic deliberation and the redistribution of epistemic authority within technocratic governance. Designed to evaluate the World Bank- and DFID-funded “Vision 2020” development strategy—which proposed large-scale agricultural mechanization, contract farming, and GM crop adoption—the process explicitly challenged the unidirectional flow of knowledge from credentialed institutions to affected communities.

Central to Prajateerpu’s deliberative design was its representational politics. Rather than pursuing statistical cross-section sampling, organizers privileged the most structurally marginalized constituencies: small and marginal farmers, predominantly Dalit and Adivasi, with women constituting the majority. This compositional choice reflects a commitment to communicative justice, recognizing that formal participatory equality, absent structural intervention, reproduces existing social hierarchies.

The process further departed from conventional consultation models by positioning jurors not as passive recipients of expert testimony but as active epistemic agents. Participants cross-examined government officials, corporate representatives, and researchers using experiential, multigenerational knowledge of local agroecological systems as legitimate counterevidence. Critically, the evaluative framework jurors applied—encompassing soil health, seed sovereignty, women’s labor, and community cohesion—operated outside the technocratic criteria endorsed by proposing institutions, thereby contesting the very terms on which Vision 2020 sought to be judged.

The jury ultimately rejected land consolidation, displacement, and biotechnological intensification in favor of localized self-reliance and community resource governance. Beyond the verdict itself, however, Prajateerpu generated enduring institutional impact: the Community Biodiversity Register, developed by sangham women to document indigenous seed varieties and ecological practices, was subsequently adopted as the national template for biodiversity documentation in India. This institutionalization of knowledge, produced outside formal credentialing structures, exemplifies what scholars of deliberative democracy might term the structural recognition of subaltern epistemic contributions—a reversal of conventional authority that extended well beyond the deliberative event itself (DDSIndia.org; Pimbert and Wakeford 2002).

This concept extends and specifies existing theoretical frameworks. As mentioned earlier, Santos’s (2014) “epistemologies of the South” provides the broad framework; “autonomy as counter-policy” names a specific institutional mechanism through which epistemic justice is not merely claimed but put into practice.

To appreciate what is theoretically distinctive about “autonomy as counter-policy,” it is necessary to distinguish it carefully from adjacent concepts with which it might otherwise be conflated. The closest apparent parallel is Scott’s (1985) concept of everyday resistance—the quiet, individualized, often invisible acts of non-compliance through which subordinate groups navigate domination without directly confronting it: foot-dragging, feigned ignorance, pilfering, gossip. Everyday resistance is structurally defensive and deliberately inconspicuous; it operates in the gaps of power rather than against its architecture. The DDS sangham model is something categorically different (Muppidi 2001). Sanghams do not operate in the gaps—they build institutions. They do not avoid visibility—they seek it, through community radio, participatory video, farmer-led research juries, and direct submission of community-generated evidence to state policy bodies. Where everyday resistance leaves dominant knowledge hierarchies intact while carving out survivable space beneath them, autonomy as counter-policy, as evidenced in the Prajateerpu process (detailed above), directly contests those hierarchies by generating alternative epistemic infrastructure and asserting its policy relevance on community terms. A second distinction is required from prefigurative politics, as theorized by Yates (2015) and others in the tradition of social movement studies. Prefigurative politics describes the practice of embodying, in the present organization of a movement, the social relations and values that the movement seeks to bring about in the world—living the alternative as a form of political practice. This is precisely the self-organizing, distributed dynamic introduced through complexity theory in Section 2.2 (see also Chesters and Welsh 2006): a political practice that does not require central coordination to produce coherent, system-wide social form. The sanghams are undeniably prefigurative in this sense: their internal governance structures model the horizontal, women-led, collectively deliberative arrangements they advocate for externally. But prefigurative politics, as a theoretical category, tends to remain at the level of internal movement culture and is often explicitly skeptical of engagement with state institutions and formal policy processes. Autonomy as counter-policy goes further: it is not content to model alternatives internally but actively deploys community-generated evidence to intervene in, reframe, and where possible redirect the policy processes it contests. The distinction is between enacting an alternative and using that alternative as leverage as evidenced in the below examples:

Bt Cotton and the Andhra Pradesh Pesticide Debate

DDS sangham women went above merely practicing non-GM, non-pesticide farming as an internal community alternative (even though they did start with that in the initial years)—they systematically documented the comparative performance data and presented it as counterevidence in state-level agricultural policy deliberations. When the Andhra Pradesh government was evaluating whether to expand Bt cotton promotion programs, DDS mobilized multi-year records of input costs, yield variability, soil degradation, and debt accumulation from sangham villages alongside adjacent villages that had adopted Bt cotton. The women filmmakers of the Community Media Trust even made documentaries after interviewing affected farmers in the region, beyond their own and distributed them on DVD’s and other platforms (personal observations and interactions with Sangham members and P.V. Satheesh). The community-generated evidence did not simply demonstrate that another way of farming was possible; it reframed the policy question itself—shifting the burden of proof onto the government to justify continued Bt cotton promotion against documented evidence of smallholder harm. This contributed to Andhra Pradesh becoming the first Indian state to formally recommend a moratorium on Bt cotton, a policy outcome directly traceable to community-produced evidentiary intervention rather than academic or NGO research alone (DDSIndia.org).

The Community Biodiversity Register as Policy Leverage

Sangham women developed the Community Biodiversity Register initially as an internal practice—documenting seed varieties, ecological knowledge, and cultivation histories for community use. However, DDS subsequently deployed this register as an active intervention in national biodiversity policy deliberations, specifically in debates around the implementation of India’s Biological Diversity Act and the rights of communities over genetic resources. The register was not presented merely as evidence that communities could manage biodiversity effectively; it was advanced as the institutional model that national policy should adopt, effectively forcing a renegotiation of where authoritative knowledge about biodiversity resided. The Indian state’s eventual adoption of the community register format as the national template represents a measurable policy redirection—the alternative became the standard.

Prajateerpu as Direct Policy Process Intervention

The citizens’ jury itself exemplifies the distinction most sharply. DDS and its partners did not organize Prajateerpu solely to demonstrate that smallholder, self-reliance-based agriculture was viable—that could have been accomplished through internal documentation alone. The deliberate choice to structure the process as a formal policy evaluation, invite government and World Bank representatives as witnesses, produce a verdict in the language of policy recommendation, and disseminate findings through international development networks was a strategic decision to use the alternative model as direct leverage against Vision 2020. The process generated sufficient international pressure—including critical responses from DFID-affiliated researchers and coverage in development policy forums—that the Andhra Pradesh government faced reputational costs for proceeding with the Vision 2020 framework unchanged. The alternative was weaponized as a policy intervention (Pimbert and Wakeford 2002).

Community Radio and the National Broadcasting Policy Debate

DDS’s community radio stations, operated entirely by sangham women, functioned simultaneously as local knowledge infrastructure and as a challenge to India’s restrictive community radio licensing framework, which at the time prohibited NGO and community organizations from holding broadcast licenses. DDS actively used the demonstrated success of its unlicensed community broadcasts—the agricultural knowledge shared, the women trained, the local governance strengthened—as evidence in national advocacy for policy reform (DDSIndia.com). This contributed to the eventual revision of India’s community radio policy in 2006, which expanded licensing eligibility to civil society organizations. The community practice was not separable from the policy campaign; the practice was the argument.

The concept also requires separation from conventional policy advocacy, with which it shares the goal of influencing formal governance, and from Mignolo’s (2011) epistemic disobedience, with which it shares the commitment to decolonial knowledge production. Conventional policy advocacy operates by translating community interests into the epistemic language of the institutions being lobbied—commissioning technical reports, producing policy briefs, engaging consultative processes on terms set by the state or by international development bodies. It accepts, implicitly or explicitly, the validity of dominant knowledge frameworks as the medium through which claims must be articulated to be heard. Autonomy as counter-policy refuses this translation imperative.

The DDS farmer-led jury model is the clearest illustration: rather than presenting sangham women’s agroecological knowledge as data to be evaluated by credentialed experts, it constitutes the sanghams themselves as the evaluative authority—inserting community epistemology directly into the apparatus of knowledge validation, rather than petitioning for entry into someone else’s apparatus. This is where the concept converges with, but ultimately extends, Mignolo’s (2011) notion of epistemic disobedience—the refusal to think within the terms set by unequal social structures, colonial and neoliberal knowledge regimes. Epistemic disobedience, as Mignolo theorizes it, is primarily a philosophical and scholarly posture: a refusal enacted in theory and in the critical intellectual work of delinking from Eurocentric epistemology. Autonomy as counter-policy names something more institutionally concrete: it describes a mechanism through which epistemic disobedience becomes organizational, durable, and policy-effective. The alternative public distribution system of the DDS is a good example of this (DDSindia.org). The sanghams do not merely refuse dominant knowledge frameworks in principle—they build the grain banks, biodiversity registers, and community media archives that make that refusal materially consequential. What is structurally distinctive about the DDS model, then, is precisely this combination: the internal logic of prefigurative self-governance, the external ambition of policy intervention, the epistemic refusal of Mignolo’s decolonial tradition, and the institutional durability that none of these frameworks, taken alone, fully theorizes.

4.2. Complexity, Distributed Agency, and Systemic Resilience

The sangham network’s resilience across forty-plus years—surviving drought, displacement, local and social pressures, corporate encroachment, and political hostility—is best understood through a complexity framework. Unlike hierarchical development organizations that are vulnerable to collapse or capture at their institutional center, the DDS network’s distributed architecture means that knowledge, resources, and organizing capacity exist across multiple nodes simultaneously. Feedback loops between local practice and broader advocacy are continuous, adaptive, and mutually reinforcing. Here are some examples of how the Sangham women’s groups responded to some of these instances.

• Distributed Architecture/Resilience Against Collapse or Capture

When the Andhra Pradesh government shifted toward pro-GMO policy under corporate lobbying pressure, DDS’s response capacity was not dependent on a single advocacy office or leadership figure. Sangham groups across dozens of villages independently maintained seed banks, continued biodiversity documentation, and sustained local food sovereignty practices regardless of the state-level political climate.

2.

Resources Distributed Across Nodes

Deccan Development Society’s community grain banks and seed lending systems operate at the village level with local governance, meaning financial or logistical disruption to the central DDS organization would not dissolve the material resource base sustaining individual communities.

3.

Knowledge Existing Across Multiple Nodes

The Community Biodiversity Registers were not stored in a central institutional repository but maintained by sangham women across dispersed communities. When the Indian government sought a national template for biodiversity documentation, it had to engage with a distributed knowledge infrastructure—no single point of capture or deletion existed.

4.

Continuous Feedback Loops Between Local Practice and Broader Advocacy

When sangham women documented the failure of Bt cotton on dryland soils through their own observational records, that granular, place-specific data fed directly into DDS’s national and international advocacy against GM crop expansion—making their lobbying evidence empirically grounded in ways that institutional research often was not. Conversely, when advocacy work surfaced new threats (land acquisition legislation, pesticide approvals), information flowed back to village-level sanghams, enabling preemptive local organizing.

5.

Mutually Reinforcing Loops

DDS’s community radio stations, run by and for sangham women, served simultaneously as local agricultural knowledge-sharing platforms and as documentation infrastructure whose recorded content became usable in policy advocacy contexts—collapsing the distinction between grassroots practice and institutional intervention.

This has direct implications for research design: engaged scholarship with complex adaptive systems must itself be designed with flexibility, distributed authority, and iterative feedback built into its structure. Linear research designs are fundamentally inadequate for documenting emergent, non-linear social processes.

4.3. Implications for Engaged Scholarship in Contentious Times

The DDS case challenges several assumptions that continue to structure mainstream academic research: that peer-reviewed publication is the primary measure of scholarly impact; that separating research from advocacy is a marker of methodological legitimacy; that institutional IRB frameworks are the main site of ethical deliberation; and that community-produced knowledge requires academic translation to become policy-relevant.

Standard IRB protocols govern researcher conduct toward research subjects—informed consent, confidentiality, harm minimization—but they do not address the structural question of who owns the knowledge produced, who controls its dissemination, or whether the research agenda serves community interests. DDS’s sangham-governed research process embedded ethical deliberation at every stage: communities decided what would be documented, retained ownership of biodiversity data, controlled how findings were represented externally, and could withdraw participation from advocacy uses of their knowledge. No IRB framework requires or even contemplates this level of community epistemic sovereignty. The contrast is illustrated concretely by cases in which academic researchers extracted indigenous seed knowledge through formally IRB-compliant processes, published findings under institutional authorship, and left communities with no rights over the resulting intellectual property—technically ethical by institutional standards, substantively extractive by any community-centered measure.

This finding is not unique to the Indian context. It resonates with, and extends, arguments developed about indigenous epistemologies and community-engaged scholarship in other global-majority regions: Smith’s (2021) account of Kaupapa Māori research in Aotearoa/New Zealand, Chilisa’s (2020) postcolonial indigenous paradigm developed largely from Southern African contexts, and Mohammed’s (2025a, 2025b) analysis of community-centered research design in West Africa all converge on the same structural point argued here: that Western institutional ethics frameworks, however well-intentioned, were not built to recognize community epistemic sovereignty, and that communities across otherwise very different colonial, postcolonial, and caste-based histories have independently developed their own mechanisms—Kaupapa Māori protocols, relational accountability, sangham-governed documentation—for asserting that sovereignty where formal ethics review does not.

In each dimension, DDS’s practice demonstrates that these assumptions serve institutional interests more reliably than they serve justice. Engaged scholarship that genuinely centers community priorities will look structurally different from conventional academic research—and should be evaluated by different criteria accordingly. The Community Biodiversity Registers produced by sangham women were never published in agronomy or development studies journals, yet they became the foundational template for India’s national biodiversity documentation policy. By conventional academic metrics, this work would be invisible—no citations, no impact factor, no institutional authorship. Yet its policy reach exceeded that of most peer-reviewed scholarship on agrobiodiversity published during the same period. Similarly, DDS’s community films documenting seed diversity and women’s ecological knowledge circulated internationally through food sovereignty networks, directly shaping advocacy frameworks at the FAO level, without ever entering academic publication infrastructure.

Research-Advocacy Separation as Methodological Legitimacy

Mainstream agricultural economics research on GM crop performance in dryland India was largely conducted by institutions—ICRISAT, World Bank-affiliated researchers—that formally separated their empirical work from policy recommendation, presenting this separation as a guarantee of objectivity. Yet that research consistently underweighted variables central to smallholder survival—soil degradation timelines, women’s labor burdens, seed re-use economics—because those variables did not fit existing modeling frameworks. DDS’s openly advocacy-oriented documentation of Bt cotton failure, by contrast, captured precisely these dimensions because the research questions were set by communities whose livelihoods depended on accurate answers. The advocacy commitment did not distort the findings; it determined which questions were worth asking.

For communities like the DDS sanghams—whose media programming, land tenure, seed sovereignty, and food systems are continuously subject to contestation by corporate and state actors who deploy technical knowledge as an instrument of dispossession—the right to document, produce, and research is not an abstract democratic ideal but an urgent practical necessity.

The standard assumption is that community knowledge must be validated, systematized, and reframed by credentialed researchers before it can enter policy deliberation. The Prajateerpu citizens’ jury directly refutes this: marginalized farmers’ multigenerational observational knowledge of soil behavior, seed performance, and agroecological interdependence was presented as primary evidence in a formal policy evaluation process, without academic intermediation, and that evidence was sufficient to support a verdict that influenced national debate on GM crop policy. Equally, DDS’s community radio broadcasts of local cultural and agricultural knowledge—entirely self-produced, without academic involvement—were recognized by the UN Food and Agriculture Organization as models of indigenous knowledge preservation, not because academics certified their validity, but because their practical and documentary rigor was self-evident to other practitioners and policymakers engaging with them directly.

What the sangham network has built over more than four decades is precisely the institutional infrastructure through which that right is exercised: the capacity to pose questions about their own agroecological systems, to gather and interpret evidence through their own methodological frameworks, to produce findings that carry authority within their communities and credibility in policy arenas, and to control how that knowledge circulates and to what ends it is put. Appadurai (2006) frames the right to research as a horizon toward which progressive education and development practice should move; the DDS case demonstrates that, under the right conditions of sustained collective organizing, it is a horizon that communities can reach on their own terms, with or without academic accompaniment.

5. Conclusions—Considerations for Community-Engaged Partnership, Organizing and Policy

This paper has theorized the DDS collaboration as a model of community-based inquiry in which Dalit and indigenous women function as co-researchers, knowledge producers, and media creators within a complex adaptive system. Drawing on personal formal and informal interactions, observations, participatory fieldwork, community media production, and movement-generated evidence, this analysis has developed three principal arguments.

• First, participatory video and community radio—when designed, produced, and controlled by the communities they document—can constitute primary sites of knowledge production that challenge extractive research paradigms by eliminating the separation between the knower and the known that extractive research depends upon.
• Second, the productive tensions between academic and community knowledge production—in terms of priority, timeline, and metric—are not deficits to be managed but analytical resources that reveal the institutional conditions under which academic research reproduces the hierarchies it ostensibly critiques, and that point toward the structural changes required for genuine engagement.
• Third, “autonomy as counter-policy” names a mechanism through which marginalized communities use the infrastructure of self-governance to contest epistemic hierarchies and assert their standing as legitimate producers of policy-relevant knowledge. This mechanism is transferable: it describes a strategic logic available to other communities engaged in struggles over knowledge, resources, and recognition.

For practitioners of community-engaged research in contentious times, the DDS case study suggests a set of concrete commitments. Community control over research outputs is not merely an ethical nicety but a structural requirement of genuine partnership. However, while “Civil society can force markets and states to be democratically accountable,… it can also collude in the reproduction of oppression and inequality, absorbing suffering and diffusing resistance” (Burawoy 2004). Hence, ethical protocols must be developed and always with, not on behalf of, communities. Impact must be measured by community-defined criteria. And scholars must be prepared for the possibility that the most significant contribution they can make will be an archive, a radio broadcast, or a seed bank—maybe not a peer-reviewed article.

In a period when expertise is increasingly weaponized against the communities it claims to serve, these commitments constitute not only a methodological position but a political one. Engaged scholarship that takes epistemic justice seriously cannot be neutral.