Spectres of Agrarian Territory

David Ludden

December 11, 2001

The national imagination has had a long, productive career as our guide to historical research, but other modes of thought now need more nurturing. National maps mechanize research by putting all our data in their pre-assigned place. Spaces that elude the national sensibility disappear when scholars heap all data from all times and places into national containers. All the histories of all the peoples in the world currently appear in the cage of some national past or another, but some need their own space. It is a pressing challenge to imagine at least some history in non-national terms, and particularly for scholars who want to write about old geographies that became spectres in a world of nations.

These old geographies are spectral in several senses. Archaic and out of place in the present, they seem imaginary and only make sense inside routines of national mapping. Some are quaint and benign but others are scary spooks that conjure up places outside the national order of things. Eerie ghosts emerge when old geographies refuse to die yet resist substantiation. Some old and barely visible regions of human activity remain vital for people inside them. Spaces that offend national sensibilities stimulate intense cartographic anxiety, as for example among the Indian officials who censor and regulate the circulation of maps depicting "border areas" and "sensitive regions." Paxtun territory, Bengali Assam, and Tamil Indo-Lanka are but three of the many old historical spaces whose living legacies haunt nations in South Asia.

Many old geographies did however die in the past and their death haunts the nation with the prospect of mortality. Relics of old territories have been caged in national states to serve as spectacles of national heritage, but their archaic forms embody the future of national landmarks as relics of a by-gone day. National efforts to unearth old geographies must therefore prevent the new discoveries from undermining the nation's power to inhabit all the ages of history. Thus the national past comes to life in a suspended vitality. The spectres that roamed earlier times insert annoying question marks willy-nilly into national history's confident narrative. In the past as in the present, archaic spaces acquired new meaning and substance over time as ruling ideologies and authoritative textual practices endeavored to bury old geographies under the weight of records we use to write history. In each epoch, human living spaces lurk in shadows cast by the same light that illuminates the past.

The spectral qualities of archaic geographies should not deter us from exploring their history. Reading historical data into spaces that are not depicted on modern maps is most imperative for scholars who wish to understand disorderly worlds of human activity that almost disappeared inside the mechanical certainties of national modernity. Appropriate data abounds. Pre-modern space did not disappear when modern minds began to standardize and homogenize geography; in fact, early nineteenth-century texts shed new light on old territories that still survived then to haunt a new regime that was digging their grave.

I explore the history of pre-modern spaces in three ways: by tracing their modern legacy and contemporary vitality; by peeling back the temporal layers of historical documentation that cover old geographies in each epoch with the ideas that cage them in new environments; and by locating the origins of specific forms of geographical order and tracing their transformations over time. In this essay. I do a little of each by focusing on the geographical grounding of social identities inside agrarian territories in India's southern peninsula.

The Substance of Territory

The historical geography of agrarian India began to assume its contemporary forms in the 1950s, when the Census of India mapped ecological and demographic regions, in 1952,[1] and Daniel Thorner mapped agrarian regions of economic development, in 1959.[2] In the 1970s, ecological regions gained prominence, particularly in southern India, where Burton Stein traced medieval geographies of social activity connecting regions, integrating regions, and defining networks around temples and kings.[3] In 1983, Ranajit Guha called "territoriality" an elementary aspect of peasant insurgency in colonial India, and his account of geographical elements in tribal revolts added new impetus to the cultural reconstruction of historical geography.[4]

In the study of historical space, three methods are most prominent, which we can call distributional, cultural, and social. Like Daniel Thorner and the Census, most geographers locate variables on maps and describe spatial distributions.[5] Like Ranajit Guha, cultural historians depict representations of space by inhabitants and by people who seek to control territory with coercive force, words, surveys, maps, and other technologies.[6] Like Burton Stein, social and economic historians describe spatial circuits and networks of activity and relationships; that is, geographies of mobility, settlement, transaction, alliance, and institutions like states, temples, and markets.[7] Agrarian geography looks different in each perspective and they are most usefully combined. For instance, the contrasting geographies of tribal regions in western India offered by Ajay Skaria and by Sumit Guha provide more insight than either would in isolation.[8]

My goal is to spatialize social and economic history in geographies of human activity that are also described by distributions of variables and by cultural representations. I make a basic distinction between "space" and "territory" in agrarian history. I define agrarian space as being an area in which all the various elements combined in agriculture circulate geographically. Such spaces can be described even when they cannot be mapped or otherwise controlled by humans. Agrarian territory is that part or a feature of agrarian space under some human control: it needs be marked and bounded, physically and culturally; and its boundaries mark spatial domains for organizing social power in everyday agricultural production. In the world of representation, words can capture any space that humans imagine but maps in the mind or on the page are for controlling space inside territory. Traces of maps-in-culture are strewn across a vast variety of historical documentation from ancient times to the present. Using the various traces to depict space-in-time poses many methodological challenges that I have not yet overcome.

We can begin this exploratory exercise by proposing that agrarian territory is produced, reproduced, and transformed over time by social powers constructed culturally in social institutions that leave records behind for charting agrarian territory-in-time. Social power is distributed unequally in amount, in quality, in society, in space, and also in time. For example, the tinai in Sangam poetry represent spaces that seem to have been territorialized primarily through the operations of the poetry itself. Medieval nadu represent substantially more territorial force and social inequality but in very small territories compared to those in later centuries. Social powers that define territory appear in various forms and combinations in records depicting negotiation, alliance, exchange, conflict, and accommodation as well as command, control, and subordination. The spatial institution of social power defines the geographical contours for agrarian territory by using texts, landmarks, and social identities attached to the land.

The national state is one kind of agrarian territory. Even at the national level, agrarian territoriality -- like nationality -- is a project, rather than a fixed fact of life, because all the moving elements in agriculture resist control; and within territory, control is always relative. Natural phenomena like monsoons, topography, evaporation, photosynthesis, and soil type resist control absolutely. Prices, knowledge, beliefs, and migration can be as hard to control as wind, but controlling them is an ancient ambition. Humans do control things like crops, wages, marriage, occupation, taxes, property, and agrarian settlement patterns. Controllable items mark the territoriality of agrarian space. Agrarian territories are filled with ideas that people use to control the order of things in space-and-time. Many kinds of agrarian territories have emerged historically in South Asia and national states have encompassed all others to make them appear as constituents of the national past.

Agrarian territory thus includes elements which have been separated from one another intellectually in dominant disciplinary methods for studying Indian "society" and "economy," respectively. This alienation emerged in the late nineteenth century and still marks a boundary between methods attached to political economy and socio-cultural studies. India's economy came to be studied though its markets, classes, and development indicators; while Indian society came to be studied through its cultural traditions. This division has split agrarian history into separate intellectual spheres inhabited by the likes of Amartya Sen and Ranajit Guha. Though Sen's "entitlement approach"[9] theorizes the market as a realm of social power that Guha's "subaltern approach"[10] would treat as a domain of subordination and resistance -- a useful combination of ideas -- their ideas have marched off separately into separate spheres in the disciplinary embrace of economics and anthropology.

Agrarian territories emerge in geographies of entitlement and subalternity that organize agricultural production in physical and cultural sites. We can take from Sen the idea that markets produce a kind of power called "exchange entitlement" that people use to buy, sell, consume, and produce. Exchange entitlements rest on another kind of entitlement, which Sen calls "direct" and arises outside the market in social institutions that give people power over things they take to the market to sell. In Sen's work, direct entitlements are legal property rights, including tenant rights, but the idea of direct entitlement can be extended to include powers over things that people gain through various social institutions, such as inheritance, kinship, and caste. In Sen's analysis of famines, herders have direct entitlements over animals, as farmers do over land; but craft producers also have direct entitlements, and family producers in general have direct entitlements to the family labour they use to make commodities. "Directness" is a feature of entitlement arising from its origin outside the market; it comes from social relations of power among people who gain capacities to control things through institutions other than the market.

Direct entitlements animate all markets. Property rights are most visible, but agrarian societies also produce an array of other direct entitlements that sustain agriculture as well as markets. In capitalist economies, direct entitlements from an array of non-market institutions sustain capital accumulation.[11] Direct entitlements typically involve complex transactions and cultural symbolism. In southern India, for example, entitlements to temple honours and ritual symbols of prestige sustain extensive networks of commodity exchange and capital formation as well as everyday productive activities on the farm.[12]

Spatial combinations of exchange and direct entitlements erase intellectual boundaries between political economy and culture. Thus they allow us to explore the history of capitalism in South Asia against the grain of classical theories that equate capitalism with Western culture. Agrarian South Asia included markets from ancient times, but markets are absent from theories of Indian tradition, where markets seem auxiliary, marginal, and external to Asian culture. This theoretical alienation of commodity production and capital accumulation in markets from ideas about Indian tradition has made alien force seem necessary to explain the rise of capitalism everywhere in Asia, particularly in the countryside, where the grip of tradition is still unbroken.

Historians have made the antiquity and pervasiveness of markets in agrarian Asia more visible, and thus we need to rethink relations between markets and societies. It is no longer apt to imagine that originally and essentially non-market societies evolved into or were forced into modern capitalism. Markets are as ancient and pervasive as caste, temple worship, and kingship. In addition, critical work on modern construction of tradition and invention of "village India" now makes the "traditional Indian village" an artifact of our intellectual past. No account of Indian society is now complete without indigenous markets.[13] Spatial histories of entitlement permit us to follow social power outside market as it moves into the market, where exchange entitlements attain real cultural force in social life. Spatial histories of entitlement also accentuate the role non-market power in capitalist development everywhere; they may eventually replace national maps that control the world history of capitalism with subtler geographies.

Amartya Sen defines entitlements in the framework of modern state law. We need to extend this idea into a wider realm of legitimate social power, and to expand legitimacy beyond legality in modern states. Legitimate "direct" (non-market) entitlements include those sanctioned by custom and by cultural institutions like the Hindu temple, where gifts to god entitle donors to honors that can be "cashed" to form exchange entitlements. Marriage customs provide direct entitlements to gifts and inherited property that enter into market exchange. Patriarchal authority provides entitlements to control female and subordinate male labour. Caste rules entitle high caste families to direct the labour of lower caste people. Such "direct" and culturally embedded entitlements move legitimately into markets where they sustain exchange entitlements, even when they are not explicitly sanctioned by law.

Agrarian territory is thus a physical space where people obtain entitlements in market and non-market institutions that have a definite geographical reach. In such territories, subalternity enters entitlement. Ranajit Guha and his Subaltern Studies colleagues insist that all entitlements include subordination and resistance. Entitlements subordinate and exclude. Property rights exclude non-owners and subordinate labour. Inheritance excludes and subordinates people who are not entitled to inherit. Coercive institutions like the army and police maintain entitlements. Markets operate in discourses of power and resistance that secure entitlements with law and custom. The legitimacy of entitlement represents power that is not inherently stable but is held in place ideologically and by the use and threat of force, as it is constantly met by resistance.

Subaltern resistance is also a positive force defining entitlements; though this feature of subalternity is not well developed in Subaltern Studies. In market exchange, unequal buyers and sellers resist each others' demands in conflicts of bargaining. Direct entitlements emerge historically at the same time as subordinates resist superiors' demands in conflicts of negotiation. For instance, the tenant rights in Bengal and Bangladesh that stabilized peasant entitlements during the famines studied by Sen were historically contested entitlements produced by legal, quasi-legal, and illegal force, as well as by resistance. Such subordinate entitlements arise in a culture where social conflict over land is customary. Entitlements emerge in social interactions where conflict is not a zero-sum game. Customary and conflictual social interaction alter entitlements over time. Workers slack off on the job, tenants refuse to pay rent, and debtors default or abscond; such resistance produces subaltern entitlements within a cultural framework of antagonism. In historical perspective, the legitimacy of entitlements is always contingent: it includes changing legal prescriptions and powers as well as customary forms of subversion, subterfuge, and deception.

Agrarian subalterns typically struggle against domination and for entitlements at the same time. Agrarian territories thus emerge as geographical spaces defined by social power and resistance that together produce and transform entitlements. Property rights are paradigmatic subaltern entitlements in agriculture. People with higher ranks and more authority grant, confirm, legalize, and secure property rights for people below them who resist constraints and fight for rights at the same time as they use entitlements obtained by subordination and loyalty. The history of landed property includes a vast array of subaltern struggles for entitlements among people who both accept and resist proprietary claims. Historical change in proprietary institutions involves authoritative state action and also subaltern struggles against domination and for legitimate entitlements. Thus, Dipesh Chakrabarty describes subaltern consciousness inside a “composite culture of resistance to and acceptance of domination and hierarchy.”[14] One illustrative case of entitlement-changing subaltern action is the Agrarian League of Pabna in nineteenth century Bengal.[15] Gunnel Cederof has recently documented a more ambiguous case of "lost" -- rather than "broken" -- bonds of pannaiyal servitude among Madhari workers in twentieth century Kongu country (in Tamil Nadu), where drought, shifting employer options, technological change, and missionary influence combined with effective subaltern struggles to alter the composition of entitlements for landless workers.[16]

Territories in Time

In India's south-eastern peninsula, in what is today Tamil Nadu, institutions of agrarian entitlement have territorial histories from the middle of the first millennium. The three kinds of territory that I consider here each have their own time in historical space.

Well-documented medieval territories emerge in stone and copper inscriptions in the ninth century on routes of river drainage near the Indian Ocean coast. The first inscription to describe (rather than merely indicate the existence of ) territorial institutions comes from Manur (near Tirunelveli) around 800 CE, and the most elaborate early account comes from Uttaramerur (near Kanchipuram) a century later.[17] Irrigated paddy production typified these territories and set them apart from their surroundings. Their definitive documentation is epigraphy produced in transactions among Brahmans, farmers, and kings that assert entitlements to agricultural resources. The term nadu is prominent in their geographical nomenclature. It appears that irrigated rice cultivation and inscriptional transactions occurred only inside the medieval nadu. The nadu seems to be a territory of ethnic identity, agricultural expansion, royal authority, ritual, and institutional development. Inside the nadu, segregated settlements of Brahmans and Vellalas, called sabha and ur, respectively, controlled most recorded entitlements; and in some but not all nadu territories, medieval inscriptions describe nadu assemblies and leaders called nattar.[18]

Medieval spectres survived to modern times, but medieval territories faded away. Inscriptional records of the nadu decline rapidly after 1300. The last inscription recording an action by the best-documented sabha, in Uttaramerur, near Kanchipuram, came in 1434.[19] In later centuries, nadu legacies entered larger agrarian territories, as agricultural expansion accelerated, new warrior alliances conquered the coast, and old entitlements entered new warrior regimes. The title of nattar remained, and in 1801, the Nattawars of Uttarmerur were three Brahmans who may have had ancestors in its sabha.[20] But medieval nadu, ur, and sabha were engulfed in the expansion of Nayaka power under the umbrella of Vijayanagar Rayas and by the formation of later medieval ethnic territories. The term nadu acquired new meanings, for example, in the organization of professional Kaikkolar weavers.[21] Medieval inscriptions and oral traditions were used to negotiate entitlements with new state authorities in later regimes, but the early medieval nadu was a time-bound territory of agrarian entitlement.

New territorial forms emerged uniquely in later medieval centuries. Their creation is not so well documented but some inscriptions and substantial lore indicate that they were built after 1300 on frontiers of agricultural expansion in upland and interfluvial areas by lineages and clans embracing warriors and farmers. Such territories developed distinctive ethnic identities. Some were built by peasant warriors who moved out from old nadu on the coast, like the Kammas and Kapus who colonized the Krishna-Godavari uplands and Kongu Vellalars who settled the Kaveri River uplands below Mysore. Others were built by people who had ancient tribal backgrounds as hunters and pastoralists who now became settled farmers and warrior kings, like Boyas north of the Kaveri in Karnataka and Kallars and Maravars south of the Kaveri in Tamil Nadu. Still other new territories were built by warriors who conquered the nadu under the Nayakas of Vijayanagar and then settled their families, retainers, and peasant allies in compact frontier territories in land outside the old nadu.

New territories of this later medieval ethnic kind also appeared in the Deccan and in north Indian frontiers like those colonized by Jats. Their definitive products in the peninsula were typically not rice but rather dry crops like millets, oilseeds, and cotton. They did build irrigation tanks, but in dry places like those east of Madurai occupied by the Maravar Setupati rajas in Ramanathapuram, tanks provided a precarious basis for paddy cultivation, which was typically mixed with a heavy dose of dry cropping. As in Maravar and Kallar country and in Kongu, Brahmans did not typically own land in these new frontier domains; Brahmans were not nearly so important for systems of entitlement as in the nadu. Temples, mosques, and gurudwaras were important for agrarian entitlements in later medieval India, but military alliances and leadership were much more critical.[22]

Ethnic territories of warrior-farmer frontier colonization had long lives. They obtained their spectral appearance when they entered British India, when their leaders became Zamindars (as in Ramanathapuram, or Ramnad), Native State rulers (as in Pudukkottai), and Ryotwari landowners (in most of Madras Presidency). The visibility of ethnic territories declined as entitlements inside them came under the writ of modern property law, but early Company documents provide a good indication of their vitality before 1850. Some have died since then, but others continue to live with a vengeance, as we will see.

A uniquely early-modern form of agrarian territory emerged after 1500 and became more visible after 1750 in texts produced by the English East India Company. Territories around urban centres of capital accumulation combined direct and exchange entitlements in compact territories strategically placed in expansive networks of military and market activity. Cotton cloth was the product that best typified urban territories that brought together social power in agriculture, manufacturing, war, taxation, and trade. This early-modern textile urbanism thrived where military and financial power combined to secure entitlements to land and labour amidst social institutions of capital accumulation at the intersection of farming and textile production. Two urban territories surrounded Kanchipuram and Tirunelveli, respectively, in the ambit of ports at Madras and Tuticoryn, in areas which had contained wealthy and well-documented nadu; they also incorporated parts of late medieval ethnic territories, but they comprised distinctively urbane ethnic territory of an entirely new sort.

However robust they might have been in 1800, early-modern urban territories were spectres from the outset, because the same documents that record them also break them up into administrative localities called "villages." This classification guided modern territorialism and erased the urban complexity of the early-modern economy. The concentration of locational privilege and resource control in big cities during the nineteenth century erased the old urbanity of the country around Kanchipuram and Tirunelveli. The ruralization of agrarian space then laid intellectual basis for India's national agrarian history by establishing the comprehensive dominion of village tradition everywhere and for all time.

Though all the spectres of these three kinds of archaic territory inhabit present-day Tamil Nadu, they elude the history of Tamil Nadu, which is a national artefact. In 1956, India produced linguistic states by partitioning territory according to the distribution of "majority languages." No linguistic state defined agrarian territory before 1956, but all became permanent territories of culture and history thereafter.[23] Agrarian entitlements in India's four southern states had been defined previously in other domains, but after 1956, historical research on Tamil Nadu covered the state with a single history.[24] Spectres became harder to see. To see them better, we need to remember that histories of Tamil-speaking peoples have spilled in all directions across borders of Tamil Nadu and into Sri Lanka, Malaysia, and South Africa. The Tamil country had no borders before the nineteenth century. India's linguistic partitioning of national territory ignored multi-lingualism and multi-cultural mixing, using the same geographical methods employed at Partition in 1947. History in what became Tamil Nadu includes countless non-Tamils, multi-lingual peoples, and cultural forms that evade Tamilness. Telugus, Kannadigas, Malayalis, Marathas, and Europeans were crucial in pre-modern history. Maratha kings ruled Tanjavur for two centuries before it entered Madras Presidency. Nayakas were mostly Telugus, including the national hero Kattabomma Nayakar, who was later Tamilized as Vira Pandiyan. Madras was also a Telugu city. All the countless non-Tamil elements in the history of Tamil country became spectral when one official linguistic Tamil identity covered the territory that was enclosed for the first time in 1956 as a component of Indian national territory.

Territories in Space

We can peel back the layers of time to reveal many non-national geographies, some still quite alive. In the early nineteenth century, literature and folklore indicate that medieval terms were still in use. Three names found in medieval texts describe parts of the coastal plains: tondaimandalam spread north of the Kaveri, with its core in the old Pallava heartland, around Kanchipuram; cholamandalam surrounded the Kaveri delta, with its old Chola centre at Tanjavur; and pandyamandalam spanned Vaigai and Tampraparni river basins, with its centres in Madurai and Tirunelveli. The kongunadu region covered Kaveri uplands on the borders of Tipu Sultan's Mysore, from which it was ceded to the East India Company.[25] The distinctive ethnic character of kongunadu appears in unique documentation,[26] and its unique economic history is dramatized today by Kongu Vellalar industry around Tiruppur.[27] Though ethnic domains like Kongu country became less visible as such in Company times, District-level administration recognized the territories of the Nayakas and Maravars inside the Pandya country, where they retained their own very lively textual identity.[28]

Some spectres visible in 1800 died thereafter. Accounts from Tirunelveli show that Tamil Vellalars then uniquely used mandalam terms to designate their own ethnic identities, which were attached respectively to the Pandya country, Chola country, and Tondaimandalam, as well as to Karaikkadu (in Marava country) and Tulu country in the interior mountains. Such naming practices reflected ethnic divisions among immigrants and "natives" in the Tirunelveli region of the southern Pandya country. Names of this kind faded in later times, but the spectre of the mandalam did not die. Narendra Subramanian has recently shown each of the old regions has made its own distinct contribution to Tamil ethnic identity through its particular pattern of electoral politics.[29] (More on this below.)

New spectres of old geographies came to life in the 1890s, when U.V.Swaminathaiyar recovered texts dating from early centuries of the Common Era, which came to be called Sangam literature. These texts were then turned into "classical Tamil," though they long pre-date the separation of Tamil and Malayalam. Turning Sangam space into "ancient Tamil Nadu" prevented ancient spectres from disrupting the nation, but other readings of the ancient poems can also be used to give their spectres life.

Sangam authors seem to have lived in tribal societies scattered across the peninsula, from coast to coast, mostly south of the Kaveri. Their poems assert that social identities live in particular places: mountain people, plains people, forest people, seacoast people, and wilderness people have distinctive characters that represent their home environment. Social identity is part of the land. Each Sangam environment (tinai) appears in the poetry as a kind of cultural island. The poets and their subjects move among islands and also outward into the wider world. The wider world of travel and danger is an integral feature of poetic imagery but remains outside the tinai as an open space into which people embark, leaving loved ones to wait anxiously at home. Though accounts of travel to the outer world most often depict inland journeys across the wilderness (palai), where fearsome hunters and robbers live, the seacoast (neythal) is the place where worried lovers pine. The poets of ancient tribes depict a habitus in the peninsula where the coast and the sea merge with the land.

Sangam texts invite us to read the later history of agrarian territories that arose inside their space with ancient spectres in mind. They have three specific messages. The composition of the landscape spans the peninsula, from coast to coast. The localization of ethnicity in small islands of collective identity defines social life. The openness of the land and sea all around forms a space for travel, return, departure, and arrival that is part of everyday life.

Reading historical sources with ancient spectres in view deflects us from the habit of heaping all our data into national geography. In southern Pandya country, we immediately notice that early medieval inscriptions make one mandalam that is not in Tamil Nadu (and not in Madras Presidency) quite prominent: cheramandalam, the realm of Chera kings in Kerala. South of the Vaigai, Chera dynasties seem more influential than northern Tamil dynasties in early medieval times. In later centuries, land west of the mountains continued be an integral part of agrarian history in the east. Routes through mountain passes intimately linked east and west coasts overland. In the eighteenth century, banking and textile trades ran along east-west routes more than along north-south routes. These east-west routes led to and from Sri Lanka. In the eighteenth century, the coast south of Ramesvaram was commanded by Dutch based in Sri Lanka. The east-west blending of political landscapes continued after the English conquered the far south. Shencottai and Kanya Kumari remained in Travancore until they came into Tamil Nadu. In 1981, many Dalits in the village of Meenakshipuram, near Shencottai, converted to Islam, causing a national uproar in India,[30] and the networks of religious fraternity that sustained them ran along old pathways into Kerala -- and also overseas to the Middle East.

The open space of land-and-sea spaces in the southern peninsula contrasts sharply with the landlocked Sanskrit and Persian geographies that anchor ideas about Indian tradition, which make the sea foreign territory. India's the coastal environs are much better documented than the interior in "foreign sources" that are actually native to the open world of the sea and the coast.[31] Early medieval Geniza manuscripts in Cairo depict a steady flow of migrant traders from west Asia into the “India” they found around Cochin.[32] Muslim, Christian, and Jewish traders formed significant local resident populations in early medieval Kerala.[33] In Sri Lanka as well as in Chera and Pandya countries, pre-modern farmers lived in spaces that included maritime trades, migration, and cultural influence extending naturally across the Indian Ocean, east and west. Ports along the southern coasts were still their own best trading partners in the early 1800s, when more trade went by sea than by land between the Native State of Travancore and Tinnevelly District in Madras Presidency.

Moving in these open spaces, Tamils settled naturally in Sri Lanka for all the centuries before this one. Tamil Nayakas formed a dynasty in eighteenth century Kandy. Fishing families and palm tree cultivators moved and settled everywhere along southern coasts. In 1807, James Cordiner reported from Sri Lanka that "none of the Cingalese are divers, and scarcely any of them engage in the other active parts of the fishery ... which occurs immediately off the coast just south of governor's house at Aripo at 8.47 north latitude..... [and for which] boats with their crews and divers come from Manaar, Jaffna, Ramisseram, Nagore, Tutakoreen, Travancore, Kilkerry, and other parts of the coast of Coromandel ... completely equipped, and furnished with everything necessary to conduct the business of the fishing."[34] Arabs and Europeans settled ashore with encouragements from pre-modern rulers. Religions came with them. Coastal societies domesticated many overseas elements. The 1931 Census counted populations of Muslims, Jews, Armenians, Chinese, Eurasians, and Christians who represented “foreign elements” in India but were natives of the Indian Ocean coastal lands from Kerala to Sri Lanka.[35] Before 1850, combined local populations of Christians and Muslims seem to have been more prominent here than anywhere else in South Asia. Muslims became prominent merchants and many fisherfolk and toddy tappers who had always lived on the fringes and outskirts of agrarian territory converted to Catholic and Protestant Christianity.[36]

Territories of Identity

The land-and-sea of India's southern peninsula constitutes a distinctive agrarian space where seas, mountains, and lowlands mingle, and where, in pre-modern times, Kerala and Sri Lanka were more intimately connected to southern Pandya county than the northern Tamil country. Spatial history along the southern shore eludes national history.

The localization of ethnic identity in Pandya country became increasingly elaborate over the centuries. This (and other) features of its spatial history also characterized northern regions, but spectres of the ancient tribes tell us to resist the habit of reading early medieval data into the maps of modern states. The nadu territories in early Pandya inscriptions were small, scattered, and clustered around rivers running west-to-east from mountains to the sea. Mountain forests never lay more than a few day's walk from the beach. Surrounded by untamed land filled with untamed people, who often appear in epigraphy as enemies and threats to farmers, the nadu embraced tiny islands of farming. All around, scrub and mountain forests held hunters, pastoralists, and nomads who must have outnumbered farmers. In this context, inscriptions seem to record the settling down of lineages and clans to paddy cultivation. Dynastic alliances recorded in the inscriptions protected the nadu and extended its power over people and land in its vicinity. Rituals of patronage for Brahmans and deities held alliances together and also produced the cultural basis for controlling a workforce drawn from conquered tribes whose descendants became the underclass of the nadu.

The nadu in Pandya country seems originally to have been a novel ethnic territory in which jati and varna emerged from and overlaid lineage, clan, and tribe. In the nadu, Vellalars, Brahmans, and Pallars invented new social identities in archipelagos of farming where leaders secured and expanded community power under the banner of Pandya kings in Madurai. Pandya wars and rituals of patronage became headline news in medieval texts. The local ethnicity of clans in the nadu faded from view as history became a tale of kings. The micro-politics of nadu territory faded from the records of clan leaders who became patriarchs in locally dominant Vellalar jati families, whose status became second only to Brahmans in the varna scheme.

Because inscriptions across Tamil country (and beyond) deploy one textual method, and because modern history strives to cover national territory with a single narrative, historians typically use inscriptions from all over to compose unified account of medieval history. In such accounts, nadu territories seem to cover the land in a coherent agrarian system across mandalam domains defined by medieval dynasties. Commonalities among inscriptions can seen however to reflect the spread of elite modes of documentation rather than a uniform agrarian system. The nadu territories differed inside and among mandalam areas. Medieval wars resemble wars among tribes with intense local ethnic attachments to land. These attachments appear in names and titles that denote individual identities for people in the nadu. Honorific titles reflect local entitlements among people whose names put them in very particular places. Spectres of this medieval ethnicity survived in naming customs designate a man's father, native place, and jati title. The original goal of this practice was to identity a person's place inside local institutions of entitlement. In the inscriptions, hundreds of thousands of local names and titles indicate the intense localism medieval territories filled with spectres of ancient tribes.[37]

After 1300, ethnic territory became more complex as warriors from the upland interior of the peninsula conquered the coast from north to south under the banner of Vijayanagar. Waves of warrior and peasant migration marching south along the coast subdued tribes and took over the old medieval frontiers. Meanwhile, medieval warrior tribes settled down to farming in land they had long occupied outside the nadu. By the fourteenth century, inland routes of mobility stretching across the Silk Road of Central Asia extended their reach south to the tip of the peninsula, where they met networks of mobility running east-west, from coast to coast, and outward into the Indian Ocean. Rulers from the interior peninsula lorded over the coast. The Rayas of Vijayanagar made the novel claim to be “Lords of the Eastern and Western Oceans.”[38] Vijayanagara and the Nayakas remained inland powers, however, and the Vijayanagar title seems to have been substantiated by symbolic powers to the Rayas with precious commodities from overseas, including China.[39]

South of the Vaigai, the land outside the old nadu was occupied by new agrarian territories that emerged after 1300 under the Nayakas of Madurai. The old Pandya nadu elites and centres thrived, however, under initially foreign rulers from Andhra who assimilated quickly. Telugu Nayakas displaced Pandya kings and usurped their role as protector of local Pandya elites in the land of nadu. On former medieval frontiers, outside the nadu, two sets of ethnic alliances occupied most of the land: one was composed of Telugu warriors, peasants, and retainers who came south as Nayaka allies; the other, of Maravars. The term maravar has a spectral history all its own. In Sangam texts and medieval epigraphy, it denotes fearsome people. The maravar were hunters, robbers, and warriors. They lived outside the nadu. They were not incorporated into the family alliances among Vellalar jati in the nadu. The medieval exclusion of outsiders from nadu marriages and from residence in the nadu had maintained the tribal purity of Vellalar and Brahman lineage entitlements. The fearsomeness of the maravar threatened rulers of nadu and their Pandya kings. After 1300, Maravas fought Nayakas to establish kingdoms of their own, as Marava tribes settled down to farming, conquered other tribes, and built ethnic territories from Ramesvaram up the Vaigai to the mountains and south along the base of the mountains past the Tambraparni.

Telugu Nayaka and Marava warriors and peasants carved out separate ethnic territories. Their major leaders eventually became palaiyakkarar (Poligars) of the Madurai Nayakas, who could not defeat them and made them subordinate allies. Maravars and Nayakas held their own territories as Madurai Nayakas maintained personal authority in the old Pandya territories along routes of drainage irrigation and intensive paddy cultivation. The new ethnic territories that covered most of the land between the Vaigai and Tambraparni divided agrarian space in an east-west pattern, with Maravars mostly in the west and Nayaka warrior-peasants entirely in the east. These separate ethnic territories remained in place in 1801, when rebel Poligars finally fell to the British, who were based in the old Pandya territories and met no resistance from the old Pandya Vellala and Brahman elites. (See below.)

After 1800, Maravar and Nayaka ethnic territories took different routes to the present. Telugu Nayaka farmers remained on the land, but the ethnicity of Nayaka territory died. Telugu ethnicity became irrelevant for entitlement under the British and became a positive disadvantage in modern Tamil territory. By contrast, Maravar (Tevar) territory thrived. Many of its rulers became Zamindars, the largest being the Raja of Ramanathapuram, king of all the Tevars. Among the Maravars and Kallars, "little kings" used transactions of fealty and gifting to build authority inside their own state segments; and jati sub-castes and lineages controlled specific sub-regions within royal domains.[40] After 1920, leading Maravarss became paragons of Tamil ethnic identity and leaders in the politics of Tamil territorialism. In 1938, the radical nationalist Ramanathapuram Raja, Muttu Ramalinga Tevar, founded the South Indian Branch of Subash Chandra Bose's Forward Bloc; brought Maravars into the Indian National Army; and, in opposition to Congress, led Tevars into the Dravidian Movement.[41] Since 1980, a Tevar voting bloc has supported the All India Anna Dravida Munnetra Kazhagam. By 1990s, the Kallar Raja of Pudukkottai and Tevar Raja of Ramanathapuram were official exemplars of Tamil kingship.[42] By that time, Maravar warriors and patriots had also inspired the career of Veluppillai Pirapakaran and became heroes for Liberation Tigers of Tamil Eeelam. Spectres roam the Marava past, however. The Tamil term maravar continues to connote hunters and robbers, and the British made Maravars and Kallars "criminal castes" in 1911, a label they lost only in 1947.

I discovered Maravar territory in 1975, in a large village in southwest Tirunelveli, where my host put his hand on my shoulder, turned me to face north, and said, “Look, from hear to Madurai is Marava country!” This country appears on no official map. Drawing on experience and historical research, my guide described a land where Tevars have now exercised personal power locally for centuries. Relics of Tevar territory include old fort towns of palayakkarar; its lore recounts battles against the British in the Poligar Wars.[43] In the nineteenth century, Tevar Zamindars, most famously Chokkampatti Raja, became notorious for resisting the British territorial order. Territory preoccupied Zamindari Maravars who fought pitched battles for land that officials took to be worthless. By 1900, Marava country also included a particular pattern of caste conflict -- a local version of communalism -- pitting Tevars against upstart competitors for entitlements, honours, and privileges. In the 1890s, caste riots broke out in market towns as Tevars fought Nadars who were fighting in the courts and on the streets to put spectres of their own toddy-tapping Shanar past to rest by gaining new rights of temple entry.[44]

Soon after I discovered the old Marava country, conflicts broke out in its current environs between Tevars and upstart Dalits. Dalit conversions to Islam at Meenakshipuram occurred in this context. "Statue wars" erupted in the 1990s when Tevars broke statues of Dr.Ambedkar and Dalits broke statues of Muthu Ramalinga Tevar. Struggles typically concern Dalit entitlements protected under the law, which Tevars seek to control and pass on to the next generation in Marava country. Youth are at centre-stage in caste conflicts, which often erupt in school, on the school bus, or on the road to and from school. As Pamela Price says, "honour is a preoccupation in Marava country."[45] Some Tevars are still living and fighting inside their spectral territory.

Urban Territory

Another kind of spectre appeared after 1763, when the Arcot Nawab addressed his solvency crisis by selling the English Company a rich Jagir in tondaimandalam (Chinglepet District). Company officers gathered detailed local data for revenue administration in the Jagir and special investigations followed its Permanent Settlement (which turned out to be temporary) in the 1790s. [46] At the same time, intense local inquiries also ensued in areas taken from Tipu Sultan in Rayalaseema (the Ceded Districts of Bellary, Anantapur, Kurnool, and Cuddappah) [47] and Kongu (the Baramahal Districts of Coimbatore and Salem).[48] For a few decades after 1801, Company District Records and surveys produced similar data elsewhere in Madras Presidency.[49] One of the largest detailed accounts came from a survey and census of Tinnevelly District in 1823.[50] Data of such minute detail as appear in early Company texts disappeared after the Raj established itself on a firm bureaucratic footing.

Early Company records describe territories of a uniquely early-modern, pre-industrial kind, which had evolved for several centuries to embrace economically specialized localities in compact areas of pre-industrial capitalist development. Cities in these territories later became landmarks of modern Indian urbanism, but spectral urbanity of their surroundings faded into rusticity. Early-modern urbanism -- always a spectre -- faded away in the light of modernity, where cultural elites conceived all the countryside as a world where village people worked their farms inside the "social framework" of Indian tradition.[51]

Tradition worked hard inside early-modern urbanism. Old techniques of entitlement remained effective. Individuals held direct personal powers to control economic resources inside ethnic groups, called jati or caste, many of which were defined outside the varna scheme by region, language, locality, rank, religion, sect, race, and occupation. These named social groups mostly lived in segregated locales, where their families settled in residential concentrations to form communities, and where their elites enforced entitlements in places they called home. Social rank, ritual status, and symbolic honours conferred in temples and courts provided direct entitlements that people took into the market and then reinvested in temples and courts in circuits of cultural capital formation. Physical force and resistance secured and reshaped entitlements. Temples and rulers acquired and spent large amounts. The novelty of early-modern urbanism came from the range and complexity of entitlements among diverse, locally concentrated, and economically specialized ethnic groups, many of whom had migrated in recent memory into compact areas of dense population, where intricate divisions of labour animated new institutions of commercial capital formation.

In 1800, culturally indigenous commercial capitalism permeated agrarian societies in the peninsula. In the dry interior of Rayalaseema -- covering almost 20,000 square miles -- there were no major urban centres, but market towns typically served of about fifteen of its 4,000 villages. All major villages had permanent bazaars, supplemented by 187 weekly markets. Rich in commercial minerals, Rayalaseema was also a textile country. In 1810, land-owning farming families were well outnumbered by the total of craft workers (19%), merchants (12%), and soldiers (10%). Manufacturing and trades together employed 34%; the textile sector alone employed 18% in weaving, dying, cotton cleaning, loom making, and other jobs; and most landed families (23%) grew cotton. Thomas Munro described its farmers thus:

Almost every ryot has an account with a bazaar-man and a balance against himself. This account often runs through 2 or 3 generations and is rarely paid off entirely. It usually originates in small advances by the bazaar-man who probably gives seventy or eighty rupees and takes a bond for a hundred with interest at 2 1/2 p.c. monthly (or 30 p.c. annually). The ryot in return makes payment in grain, cotton and other articles, which are usually valued against him and he receives occasionally from the bazaar-man small sums for the discharge of his Kists ..... [In the case of default] the creditor has only to produce [before the court] a bond; an order for distraint usually follows and the ryot is at once stripped of his cattle, grain and implements of husbandry and will most likely never again rise above the rank of common labourer.

Commercial vitality did not guarantee prosperity. After centuries of growth under Vijayanagar and Mysore regimes, Rayalaseema's population was certainly in decline when Company officers arrived. All accounts agree that eighteenth century war destroyed its economy.[52] Out-migrations of weavers and merchants continued for decades.[53] As non-agricultural sectors declined and merchants departed, all entitlements revolved more tightly around landed patriarch financiers.[54]

In Rayalaseema and elsewhere, agrarian commercialism expanded in territories where ethnicity structured entitlement. In Rayalaseema's relatively open farming frontiers, hunters and shepherds had much more space to work they did in more intensively farmed places in the coastal lowlands. Like other regions of warrior-peasant colonization, Ralayaseema shows a spatial concentration specific landowning jati groups. In 1810, Kapus were the largest group (48%), concentrated in the east; Okkaligas were second (19%), spread more evenly but living mostly in the west. Reddys (10%) lived only in the west. As among Kallars, Maravars, and Vellalas in the Tamil country, named sub-groups among Rayalaseema farming jatis had more pronounced concentrations: all Kapus in the west called themselves Velanatis, and two-thirds of those in the east called themselves Gonas; and only Kapus in the east omitted sub-caste epithets. Ethnic identities also organized entitlements among all the many non-farming specialist groups who moved among early-modern locales and settled most numerously in expanding urban territories. The weavers in Rayalaseema were organized in twenty named ethic groups.[55] Prasannan Parthasarathi has shown that weaver jati mobility and caste order gave professional weavers bargaining power and capacities for collective action that boosted their productivity, wages, and social status.[56]

Unlike Rayalaseema, urban territories attracted many ethnically organized specialists to live in compact areas which attained unprecedented population densities. These were privileged territories, protected by dominant military force, where commercial capital accumulated like nowhere else. Such urbanism depended on local agricultural productivity anchored by irrigation. Tanks, wells, and canals watered farming that was more intensive and diversified than anywhere else, financed by commercial capital, protected by armies, and controlled by old village elites. These urban territories were not ravaged during eighteenth century wars, like Rayalaseema and all the Kaveri basin below Mysore, including kongunadu and cholamandalam. They were not recurring scenes of rampant war. They were atypical in their time and they did not survive the nineteenth century. Company documents obscured their existence because they were compiled not to record economic organization but rather to measure assets, confer entitlements, and collect state revenues. Nevertheless, these archaic spectres had solid empirical substance.

One very large urban territory developed in tondaimandalam, in what became the Company's Jagir, and another, much smaller one developed in southern pandyamandalam, along the Tambraparni, around an urban cluster composed of Tirunelveli, Palamkottai, Pettai, and Melapayalam. Both were defined as territories of state authority under the Arcot Nawab and then Company (though Tirunelveli remained under Madurai Nayakas to the 1750s).. They were both based in old irrigated areas of continuous agrarian development stretching back to early medieval times; and both were privileged, protected, and governed subsequently by rulers who depended upon and invested in the wealth of the old nadu elites, Vellalas and Brahmans. Both had old connections with long-distance and overseas trade, and boasted major temples with huge endowments, many pious patrons, and widely scattered agricultural "estates." These were not urban centres confined by cities; but rather, urbane territorial agglomerations of big and small towns, villages, and hamlets in close proximity, where specialist communities lived separately in ethnic enclaves that structured direct entitlements. Because I have elsewhere described Tambraparni urbanism,[57] I will concentrate here on the Jagir.

Farming, manufacturing, war, finance, and commodity trades came together in a coherent geographical pattern in the Jagir, but the pattern different than around Tirunelveli. The Tambraparni River is a ribbon leading east and west, and early-modern agro-manufacturing, trade, and banking communities along the river looked in both directions, from coast to coast. Though attached to a major port at Tuttukudi (Tuticoryn), controlled by the Dutch -- based in Ceylon, for all but a few years of the eighteenth century -- Tambraparni trades also moved overland in and out of Travancore and Mysore. Sri Lanka was in its immediate economic environment, and many small ports, most notably Kayalpatnam, led to coastal networks on the west coast at least as far north as Mangalore. Mountain forest lays close at hand; iron and other minerals ay at the base of the hills; and cotton-growers worked land close by; so that all the resources most critical for agro-manufacturing were close-by for communities that could count on stable, growing food supplies under Tambraparni irrigation. Local military alliances and protection from Madurai prevented any serious military disruption in rich Tambraparni villages for a thousand years before 1800. The Tambraparni lay outside theatres of Carnatic wars and Poligar Wars enriched the valley as Nayaka, Nawabi and Company troops garrisoned at Palayamkottai provided defence and local demand for commodity traders.

The Jagir was bigger and depended on tank irrigation that did suffer distress during wars that visited tondaimandalam regularly because of its strategic geographical position. A basic similarit is however symbolized in the specialization of central Tambraparni settlements. Tirunelveli was a temple town, endowed with rich Brahman agraharam communities, and a capital ruled by Pandya Vellala and Tondaimandalam Vellala elites under the Nayakas. Pettai was a protected market town filled with rich merchants. Palayamkottai was a fort filled with soldiers. Melayapalam was a textile centre whose weavers and traders were mostly Muslims. A similar combination enriched the Jagir, where Kanchipuram was a major temple town and a textile centre, and where Madras was the major military centre and the largest of many ports. Tondaimandalam Vellalas (Mudaliyars) and Brahmans were central figures in wide commercial networks of textile trades, primary product markets, and taxation. The Jagir's urbanism spread out over a large area within twenty miles of the coast, where Nayaka rulers and Portuguese, Dutch, French, and English maritime trades had combined with inland networks to stimulate manufacturing for long-distance trade inside centres of agricultural expansion since 1500.[58] It had twenty-seven big temples receiving revenues from villages up and down the coast; only the largest were in Kanchipuram, where Varadarajaswamy, Yekambaresvarasamy, and Kamakshiyamma received the equivalent of twenty-four rich villages's average annual tax revenues in the 1790s.[59] Most big villages supported weaving communities in their suburbs.

Land revenue was big business in urban territory. Direct and exchange entitlements sustained on