Agriculture is a brute ecological intervention that gives nature its identity and civility; and which imparts personality to the land as people cut forests, divert rivers, dam lakes, kill predators; tame, breed, and slaughter animals; and burn, chop, and otherwise eliminate natural growth to replace it with plants that people like. Civilised territory needs poetry, ritual, architecture, outsiders, frontiers, myth, borderlands, landmarks, families, and households -- by means of which power and aesthetics culture the land. War is so prominent in ancient poetry because making a homeland is always violent business. In the long duree of agrarian history, many kinds of hands and minds must combine to make nature into a natural environment. Agrarian territory is thus like cuisine, which it fosters and makes its own. Clearing the land and carving the fields create a place for the nurture and collection of ingredients. Skilled labour selects, cultivates, kills, dresses, chops, and grinds. Using fuel, pots, knives, axes, hoes, mortar and pestle, and many other implements, people cook, combine and spice daily meals and special feasts to sustain work, family, and community. Any cuisine's complexity and refinement develop within elaborate patterns of exchange and specialisation, as materials, ideas, techniques, and tastes come from many sources; but each cuisine emerges in localities of accumulation and experimentation, where people experience their place in the world as they make their own special ingredients into appropriate foods for appropriate occasions.


Radiocarbon dates indicate that people were inventing agriculture in various parts of South Asia by 7500 BCE.1 The Indus Valley civilisation appears abruptly in the third millennium BCE at intersections of huge zones of farming and pastoralism which left behind archaeological remains over a million square miles from Iran to Awadh and Afghanistan to Gujarat. The oldest ploughed field yet to be excavated dates to the early Harappan period, circa 2600 BCE, and at this time, pastoral peoples also moved routinely in the summer from the high mountains in Baluchistan west into Iran and east into the valley of the Indus. Pastoral encampments dominate early archaeological records in the western plains and mountains. Evidence of permanent farming increases during the Harappan period and clusters along the lower Indus and the old Saraswati river. Painted grey ware sites indicate the Saraswati retreated steadly and disappeared during the first millennium BCE. Mohenjodaro was surrounded by small settlements of farmers and herders along networks of trade and migration; and in post-Harappa centuries, agro-pastoral societies expanded their reach and impact. Today, in Saurashtra, earthen mounds rise up on the land in open spaces between wealthy Gujarati farming villages and contain evidence of agro-pastoral settlement and circulation. Prehistoric herders moved their animals among watery places as some dug in to farm the land and produced variously stable farming communities here and there.


The regulation, extension, and elaboration of social power to organise the interaction of farming and herding formed ancient agrarian territories, which come into better view in the last millennium BCE. Ritual was critical, as we can see at Harappa.. Vedic hymns indicate that around 1500 BCE, agro-pastoral people who performed Vedic rituals were moving south from Haryana and east down the Gangetic basin. We can imagine this movement as an extensive pursuit of water and new farm land, but the hymns also record the spread of Vedic rituals among different societies of herders and farmers during an eastward expansion of agro-pastoralism, which eventually moved into the eastern jungles, where it met other social formations. The hymns tell of the fire god Agni burning his way eastward under the patronage of a human lord of the sacrifice, the jajman, who ruled and protected his people. Forty or so generations of farmers must have burned and cut their away into Gangetic forests, carving the rim land and the lowlands of the basin, learning to use iron tools, and inventing a new cuisine full of meat, rice, spices, and vegetables, before documented agrarian history begins in the middle of the first millennium BCE, in the region of Magadha, around the sites of Rajagrha and Pataliputra, in Bihar. It is most probable that a spread of Vedic rituals among ancient peoples occurred alongside migrations by Indo-Aryan language speakers moving down the Ganga basin; but the ritual nexus would certainly have embraced ever more diverse populations as the old rice-growing cultures of the humid tropics in the east made their independent contributions to the rise of agrarian culture in ancient Bihar. As agro-pastoralism and warrior migrations connected the eastern Ganga basin to Iran, Afghanistan, and the Indus-Saraswati cultural complex, the people living in the eastern rain forests and the riverine travellers among littoral sites around the Ganga delta and Bay of Bengal must have contributed to the rise of rice-growing farm societies in Magadha. By the end of the first millennium BCE, Indo-Aryan linguistic evidence mingled with Dravidian and other cultural forms in agrarian sites scattered from the Indus to the Brahmaputra and south of the Vindhyas to Kanya Kumari. A number of distinctive ritual and social complexes emerged, marked by many regionally specific artefacts such as megaliths and burial urns in the southernmost peninsula. The absorption of tribal peoples into the ritual complex that slowly evolved from Vedic rites gave rise to numerous animal deities and blood sacrifices which were missing in the early Vedic texts.


When imagining the oldest periods of the agrarian past -- for which empirical evidence is steadily increasing, forming a vast puzzle with an unknown number of missing pieces -- we must contend with the old view that ancient states evolved with the progress of Aryan conquest, during Aryan elite differentiation, with the incorporation of native peoples into an Aryan political and social order, described in Sanskrit texts. Most scholars have discarded this narrative but it still appears in many textbooks. There were actually no Aryan people as such (defined either as a race or as a linguistic or ethnic group). Rather, what we have is a number of texts that reflect the linguistic elements that scholars classify as "Indo-Aryan"; and these texts, spread over many centuries and locations, convey a number of ritual, prescriptive, descriptive, and narrative messages, whose authorship, audience, influence, and cultural coherence remain debatable. Archaeological evidence constitutes an increasing proportion of our evidence on ancient sites, and indicates that a number of cultures were developing separately in various parts of South Asia. Various trajectories of historical change can be proposed using available evidence; and various indigenous peoples with their own histories in the eastern Gangetic basin, adjacent hills, nearby coastal areas, and across the Bay of Bengal must have played a significant part in the rise of ancient Magadha. Ancient India's many histories intersected, diverged, and travelled independently, so that instead of a linear trend connecting the RgVeda and the Mauryas, we can generate many open-ended hypotheses to account for shifts among various forms of socio-political order. Increasingly complex forms of social organisation -- state institutions and imperial dynasties -- did evolve in the last half of the first millennium BCE, but the shift that Romila Thapar has called a movement from lineage to state societies did not constitute a general or comprehensive evolutionary shift toward state formation and away from older forms of lineage society. Very old and very new forms of social, political, and economic organisation coexisted and interacted, as they would continue to do in agrarian history. Ancient South Asia was a universe of small societies in which some of the more powerful groups left us records to indicate some prominent features. Famous ancient states arose in the eastern flood plain at the intersection of trade routes, and territorial markers in ancient texts indicate sites and peoples around them. One continuity between the Indus Valley and the Mauryan empire seems to be the importance of trade and migration among diverse local sites in South Asia, Central Asia, and the Indian Ocean basin.


In this context, agriculture expanded in named territories called janapadas and mahajanapadas, whose names appear in texts well into the medieval period, after 600 CE. In ancient Magadha and the Maurya heartland, agriculture seems to have been more intensive -- combining more labour and supporting more non-farming elites per unit of land -- over a larger territory than anywhere else in the subcontinent. Two lines of development converged here. Technological change in metallurgy, irrigation, plant breeding, and farming techniques facilitated more intensive farming; and in this, the old rice cultures around the Bay of Bengal and minerals from Chota Nagpur and Jharkhand would have been significant. Alliances among warriors, traders, ritualists, and farmers formed the state institutions that connected settlements to one another, connected farms to sources of iron ore, and disciplined labour (for farming, fighting, building, hauling, mining, smelting, forest clearing, and other work) to produce an expanding agrarian territory around the central urban sites of dynastic authority in the eastern Ganga basin. Ancient material and social technologies of agrarian power spread together. State authority and intensive agricultural production moved together and depended upon one another. Dynastic capitals rose along routes of trade and migration across landscapes filled with many types of farming settlements, which we can eventually see historically in all the agrarian landscapes of South Asia. Kautilya's Arthasastra reflects a compilation of elements that pertain to the Maurya core zone over a period of about seven hundred years, down to the time of the Guptas, and indicates that here state institutions did exert direct power in agriculture. But not so much elsewhere. Most agrarian territories that felt the fleeting impact of Maurya power were inhabited predominantly by pastoralists, shifting cultivators, and small settled farming communities. Agro-pastoral societies along the model of the post-Harappa sites in Gujarat must have remained a typical form throughout dry landscapes of the north, west, and central peninsula well into the first millennium CE. In most janapadas, the Mauryan empire seems to have consisted of strategic urban sites on routes of trade and conquest, connected loosely to vast hinterlands.


Independent but connected agrarian histories were underway in many areas during Mauryan times; this is indicated by the many new centres of power that enter the historical record in the first centuries of the Common Era. In Maurya times, Sri Lanka and Cambodia would have formed an outer rim of inter-connected, rice-growing territories. Puskalavati, Taxila, and Gandhara arose along the upper Indus and Kabul Rivers, along overland trade routes in the north-west. Satavahana inscriptions show new state authorities rising in the peninsula, around Pratisthana in Berar, Girinagara in Saurashtra, Amaravati in the Krishna-Godavari delta, and Vanavasi in the southern Deccan. In the far south, Sangam literature reflects another emerging agrarian culture. Buddhist and Jain texts depict many pre-Gupta urban sites along old trade routes from north to south, which like Kanchipuram thrived in agricultural settings; and by the second century CE, we can see Buddhist sites in most riverine and coastal areas in which medieval dynasties would later thrive. In the fourth century CE, when the Guptas sought to extend their own empire into their historical environment, they faced stiff opposition, and they never did conquer the Vakatakas, who succeeded the Satavahanas in the north Deccan.


The Gupta empire produced a new kind of articulation between state institutions and social power in agriculture. The Mauryas had thrown impressive land bridges across many janapadas, among islands of farming in the sea of pastoralism, and they concentrated their power in urban sites along extensive routes of trade and transport. Gupta imperium launched a conquest of the janapadas by farming. Mauryas travelled a more extensive empire. Guptas were more down to earth. A Gupta core zone of intensive agriculture expanded westward to include not only Maghada and adjacent Vaisali and Videha but also janapadas around Prayaga (near Allahabad) and Ayodhya (Kasi and Kosala). Gupta agrarian power also expanded north and south toward the hills on both sides of the Ganga basin as it embraced a larger number and diversity of farming peoples. Gupta rituals sanctified agrarian kingship. State-sponsored religious institutions (temples), elites (Brahmans), and sacred texts sanctified land as they incorporated local community leaders. Historians have shown how the Gupta system empowered rising elites outside the imperial court in a ritualised state which had expansive capacities for political inclusion. Gupta ritual techniques for alliance building were adopted widely and adapted to many local conditions from the third century onward. Saiva, Vaishnava, and Buddhist cult institutions combined with state authority to create powerful but flexible agrarian alliances among farmers, warriors, merchants, ritualists, kings, and literati.


Sanctification constituted real social power with tangible benefits for its participants, and its social production and meaning changed over the centuries as sanctity attached to more and more land in culturally distinct, interconnected territories. With scattered evidence from literary sources, we can dimly see how Vedic rituals helped to organise agrarian power in a world of agro-pastoralism. Vedic ritual pacified conflict between nomadic pastoralists and sedentary farmers and formed stable structures of alliance by sanctifying the performance of ritual and the social composition of community. Ancient myth depicts battles between herders and farmers as being supernatural struggles between devas and asuras; and ritualistic gambling -- performed with the injunction, "Play the cow for rice!" -- may represent "a sacrificial contest [that] could also be put to work to regulate and sanction conquest, tribute levying, overlordship and generally, state formation."2 Expanding the scope of agrarian territoriality involved the elaboration and adaptation of ritual negotiations in countless new competitive settings in each agrarian landscape. Rituals spread and changed form as they proved effective in creating stable alliances in a world in which communities of settled farmers were certainly a minority.


Pastoralism occupied much of the subcontinent until later medieval times. When Indus valley urbanism disappeared, its culture dispersed into mostly pastoral surroundings. Rgvedic society was pastoral. The Mahabharata depicts a society full of pastoralism. Krishna was born among Yadava cattle herders and many gods have similar origins among hunters and pastoral people. Sangam poetry describes five ecological regions and only one is sedentary. The ancient Tamil mountain, forest, wasteland, and seashore regions were the home of tribes, hunters, gatherers, nomads, and travellers. Hero stones across the peninsula record the pre-eminence of cattle raiding as a political activity in the first millennium. Old Tamil society was probably conquered from the peninsular steppe by nomads called Kalabhras in the fourth century. South Asia is in fact part of a vast larger historical space in which pastoralism is very prominent -- stretching from Mongolia across Central Asia, Syria and Egypt to the Maghreb and Sahel. In this wider world of arid climates, pastoralism has historically surrounded and permeated agrarian landscapes in which farms cluster around water sources along trade routes. South Asia is a borderland between this world of pastoralism and humid Southeast Asia, where dense forest and intensive agriculture exclude nomads herding large flocks. In the wet lands, rice typifies agriculture, natural forest is dense, large domestic animals are scarce (when farm land is surrounded by tigers), and nomads are rare. But in the west, north-west, and the peninsular interior, dry agrarian space is more like south-west and central Asia, where millets and wheat dominate field crops, thirst and drought preoccupy society, lowland forests are predominantly scrub, herds abound, and nomads pervade agrarian history. Ancient states in South Asia arose at the intersection of these two different worlds of agrarian ecology.


In the first millennium, the creation of landscapes of settled agriculture moved ahead more rapidly, as agrarian institutions promoted ritual negotiations to solve conflicts among farmers, pastoralists, warriors, merchants, forest dwellers, and many others. Agrarian territories expanded when conflicts could be resolved routinely under stable institutions of social power and authority. War could destroy the social routines that stabilised territory and thus allow the jungle and wild animals to invade farm land that nurtured piety and nobility. Farming communities became increasingly populous and complex as agrarian territories evolved which could embrace larger and larger populations of people on the move. Among institutions for negotiating competing interests, war was always important, and preventing war from destroying structures of agrarian stability was a secret of success. In the sixteenth century, the A'in-i Akbari and other sources support speculation that nearly twenty percent of the population depended on fighting for their livelihood, which of course meant travelling much of the time. Armies would pillage some farms to provision warriors who returned with their loot to their own farms when battles were done. The dry season was always a time for fighting; drought sent villages out to fight for food; and armies subsisted in turn on pillaging drought-stricken villages, causing communities to flee to fortress towns or to go out in search of new land. Migrations by whole communities were common and many agricultural sites have thus been settled and resettled, historically, over and over again. Herders heading to the hills in the summer and back down to the lowlands with the monsoon, seasonal worker migrations, people fleeing war and drought, army suppliers and camp followers, artisans moving from town to town, farmers moving into new settlements looking for new land, traders, nomads, shifting cultivators, hunters, pilgrims, and transporters would have added up to perhaps half the total population at most any point in most regions during centuries before 1800. What we call "sedentary agriculture," therefore, was not really sedentary. Reigning social powers settled, inhabited, identified with, and controlled territories of agricultural investment and political order, but farmers worked within institutions that embraced many conflicting social forces, many of which were constantly on the move.


Gupta-era institutions developed new capacities to control territory by sanctifying the land and by establishing rules of dharma that disciplined labour for the co-ordinated performance of all the activities of agriculture. In Sri Lanka, Anuradhapura was the centre of a Buddhist empire of irrigated agriculture that expanded across the dry north of the island in the first millennium, at the same as the Guptas began seriously to sedentarise Bharat. Maurya conquest had first defined the territory of Bharat as a triangle with its apex in the eastern Ganga, in the sacred precincts of Maghada, Kasi, and Kosala, and with its base in the fertile parts of Rajasthan. The northern leg of the triangle ran west-north-west up across submontane Punjab and the Khyber pass; and its southern leg ran west-south-west down the Narmada into Gujarat. The western frontiers of ancient Bharat thus ran north-south; and at the base of the triangle lay Gandhara in the north and Nasika in the south. The Gupta's version of Bharat was concentrated in the agrarian lowlands. Samudragupta's fourth century Allahabad inscription divides Gupta conquests into four categories, which correspond roughly with the literary geography found in the Puranas. In the territory called Aryavarta, the inscription says, rulers were subdued and territories brought under Gupta administration -- in the Ganga plain, Naga domains (Bundelkhand and Malwa), Kota territory (around Delhi and Bulandshahr), Pundravardhana and Vanga (in Bengal) -- territories of direct Gupta power. These became Puranic desa. Here, Gupta cities -- Prayag (Allahabad), Benares, and Pataliputra (Patna) -- provided ideological reference points for the sacred geography of Bharat. The sanctity of Bharat would bolster agrarian power in many medieval territories. But Puranic desa did not explicitly include the highlands around the Ganga basin, nor the Indus valley, Punjab, and western Rajasthan. Puranas describe the desa of Bharat as Purva-desa, Madhya-desa, and Aparanta desa, which embraced the Ganga lowlands, north Bengal, the Brahmaputra valley, Avanti (Malwa), Gujarat, Konkan, and the Deccan around Nasik. Old janapadas which lay outside the land of the desa would have been frontiers and peripheries of the Gupta regime. The western plains, Punjab, high mountains, central mountains, and coast and interior peninsula outside Nasika-Konkana are not called desa in Puranas, but rather asreya, patha, and pristha.3


Gupta imperium fell apart in the late fifth century as new dynasties detached Saurashtra, Malwa, Bundelkhand, and Baghelkhand; as Vakatakas expanded from the northern Deccan into Dakshina Kosala, Baghelkhand, and Malwa; and as Hunas conquered the lowlands along routes running south and east from the north-western highlands. Puranic authors called this Kali Yuga, but the idea that a classical age collapsed with the fall of the Guptas pertains at best to Gupta core regions and their ruling elites. Many historians describe the second half of the first millennium as an age of political fragmentation and regionalization, but this imagery fits only janapadas in Bharat and Puranic desa in Aryavarta. Gupta centres may have been the wealthiest in the subcontinent but most people lived outside Gupta territory. In fact, agrarian history outside Bharat comes into much better focus after the Guptas, as social powers disperse and develop which had been nurtured in the Gupta realm. Many new regimes now took up the project of protecting dharma and formed a cultural basis for medieval dynasties. As regimes of royalty and ritual multiplied after the fall of the Guptas, they produced new historical documentation. Inscriptions on stone and copper provide raw material for medieval historiography and their interpretation continues to be filled with unresolved debates. Two debates are most important here. One concerns "the Indian state" in medieval centuries. Should it should be understood as bureaucratic, feudal, segmentary, patrimonial, or something else? The other concerns the mode of production and specifically whether European models of feudalism or Marx's model of the Asiatic mode of production apply in South Asia. Both debates hinge on the effort to reconstruct typical or characteristic institutional forms in medieval South Asia. But instead of looking for "the medieval state," we can examine the range of institutions that organised social power during the expansion and intensification of agriculture. Instead of describing "the mode of production," we can try to outline the working of social power in agriculture, keeping in view the great diversity of agrarian conditions.


Peasantry


Most information for medieval history comes from inscriptions that record donations of land, animals, and other assets to Brahmans and to temples to support Vedic knowledge, dharma, and rituals for Puranic deities. Donations typically come from named, titled individuals, acting under dynastic authority; and they typically name donors, recipients, protectors, and asset holders who are often members of farming communities. Donative inscriptions often depict the transfer of land entitlements to Brahmans in the name of -- or at the behest of -- a king. They represent a transactional nexus that involves dynastic royalty (warrior-kings and their families, officials, and retainers), Brahmans (individually and in groups, Vedic scholars, ritualists, and temple administrators) and agricultural communities (farmers, herders, artisans, and merchants). Brahmans are pivotal figures and the most obvious beneficiaries, and in other ways, also, the agrarian power of Brahmans is very apparent in the second half of the first millennium. As farm territory multiplied and expanded, Brahmans produced more agricultural literature. One elusive persona, Kasyapa -- perhaps a mythical authority rather than a single author -- wrote that " ... for pleasing the gods and protecting the people, the king should take keen interest in agriculture," and further he said, "Agriculture should be practised by priests, Brahmanas and ministers particularly." He tells the king to mine "iron, copper, gold, [and] silver," to have agricultural implements made by "expert iron smiths, cutters, and goldsmiths in villages and cities," and to "distribute these among the village people."4 The role of the good king in linking together various agricultural activities is clear in these injunctions, and kings in Sri Lanka, Nepal, and many places in medieval South Asia seem to have followed this advice. Dozens of dynasties emerged from the sixth century, complete with growing centres of production and rising aggregate farm yields where Brahmans recorded, created, and propagated agricultural knowledge.


Krisiparasara, Kamba Ramayana, Krisisukti, Vriksa Ayurveda, and Paryayamuktavali are among the texts that describe irrigated tracts in the south, east, and north. The distribution of inscriptions also leads to the conclusion that in the early medieval period, the organised social effort to build agrarian territories was concentrated spatially in irrigated tracts in the lowlands, near river beds throughout the northern basins, the coastal plains, and the Deccan, Maharashtra, Gujarat, Malwa, and Rajasthan. Inscriptions record investments in fixed assets -- irrigation tanks, dams, wells, channels, paddy fields, temples, towns, markets, and cities -- and transactions in networks of exchange, marriage, ritual, and dynastic authority, which connected settlements to one another. Inscriptions describe a world of kings, Brahmans, and temple deities that constituted medieval agrarian territory physically, socially, morally, and mythologically. Inscriptional prasastis (preambles) narrate dynastic genealogies (vamsavali) and map royalty into social territory, and devotional poetry and temples likewise brought the gods into the farming landscape. Medieval Tamil poems like the Tevaram depict a sacred geography of Shiva temples that sanctified the land much more extensively and intensively than did the Sangam poetic accounts of Murugan cult sites in Tirumurugattruppadai or post-Sangam accounts of Buddhist centres in the Manimekalai (in the Gupta age). Territorial power and symbolism are more definitely documented in early medieval literature and inscriptions; and intensive, sedentary farming -- particularly using irrigation -- required more control over land and labour, as farms advanced forcefully into space inhabited by pastoralists, nomads, forest people, hunters, wild animals, and malevolent spirits. Building agrarian territory was difficult and contested. It was not peaceful. Farms carved up nature, enclosed open land, and commandeered the physical world to constitute civilisation on frontiers of farming. Taming the landscape meant displacing forms of land use and forms social life other than that represented by kings and gods, who spread the rule of dharma. Agrarian territory involved disciplining workers, co-ordinating their activities, and reorganising the allocation of resources. Medieval inscriptions recorded events in this process -- as a technology-of-record -- in compact agrarian territories.


Many types of agrarian societies came into being. A general contrast emerged between the wetter eastern landscapes and the coastal plains, on the one hand, and the drier west and interior peninsula, on the other, which was based on broad differences between wet and dry cultivation. In the humid wetlands, wild animals, disease, dense jungle, forest people, and floods posed the worst obstacles to the expansion of permanent field cultivation. In semi-arid regions, by contrast, the worst battles were waged against pastoral people and warrior nomads, whose income was readily enhanced by raids on farming villages; whose grazing lands were being converted into farmland, their herds, captured and domesticated. In the drier landscapes, settlements were more scattered and pastoral nomad warriors, more prominent. Walled towns were more common, and long distance trade was more visible in dynastic core settlements where military activity was a permanent adjunct to farming. In the wetter landscapes, farmers needed more labour to carve out fields from jungle; the higher nutritional output of paddy fields also sustained denser populations and a higher proportion of non-cultivating elites. Dry regions grew millets, and in the north, wheat; their population was thinner and elites depended on trade and wide systems of exchange and expropriation. Wet, dry, agro-pastoral, forest, fishing, and other kinds of settlements were generally mixed together in agrarian territories, which would have at their centre a central place of power and authority. Inscriptions often reflect a cultural hierarchy that distinguished the more cultivated central settlements from surrounding hamlets that were part of the territory but less cultured and privileged.


The medieval states that produced inscriptions had a basic commitment to the expansion of permanent field cultivation as the foundation of their power, and dharma was the moral code that stabilised their territory. The weakness of agrarian territorialism and thus of the rule of dharma is apparent throughout the first millennium, when many wars recorded in the inscriptions no doubt reflect a breakdown of territorial institutions during violent conflicts among sedentary farmers, pastoralists, shifting cultivators, hunters, warriors, and forest dwellers. Pastoral and tribal polities often opposed the rule of dharma successfully. But pastoral and tribal peoples also became powerful in lowland territories of settled cultivation and their role was particularly pronounced in the western plains, central mountains, Punjab, western Gangetic basin, and the interior peninsula. Rajput rulers came to recognise Bhil chiefs as allies, for instance, and an 1890 account depicts the central role of Bhil chiefs in Rajput coronation ceremonies.5 As permanent field cultivation conquered agrarian landscapes, farm by farm; pastoralism, nomadism, and forest cultivators were increasingly pushed to the margins, and many herders, hunters, nomads, and tribal people also entered agrarian society, becoming labourers, farmers, craft producers, animal breeders and keepers, transporters, dairy producers, soldiers, traders, warriors, sorcerers, and kings. This transformation of the land involved very long transitions and subtle changes in social identity, which further differentiated agrarian societies. It also involved a lot of violence, which can be seen refracted in mythical stories about the conquest of demons by gods. Many vamcavalis depict battles against tribal peoples who are viewed as enemies of civilised society. The Ramayana was reproduced in many forms, attesting to the wide relevance of its central theme, the struggle and triumph of civilisation in a land of demons and mlecchas. The Mughals would take special precautions to protect farmers against hill tribes as they pushed farming into the higher valleys. As agricultural territories expanded and multiplied, they came to include more diverse populations, not only many different kinds of farmers (including families who worked their own plots and families who used others to cultivate their fields) but also non-farming groups whose work and assets were essential for farming (artisans, cattle herders, hunters, transporters, traders, collectors of forest produce, well-diggers, priests, engineers, architects, healers, astrologers, and mercenaries). Many people who came to work in agricultural sites came from lands that were being newly incorporated into agrarian territory. Without the skills, assets, and labour of erstwhile outsiders, agricultural expansion could not proceed, and their incorporation was a major social project. Open spaces around all farming settlements also provided plenty of opportunity for groups to set out on their own to establish independent communities.


Medieval agrarian space came to consist of (1) hundreds of small agrarian territories with permanent field cultivation, diverse, changing populations, and dynastic core sites, (2) thousands of scattered settlements of farming families in the hills and plains, on the outskirts or margins of dynastic territory, and (3) vast interstitial areas in which farms were absent or temporary, covered with dry scrub-forest or dense tropical jungle and filled with tribal societies and polities. Almost all of our documentation pertains to the dynastic territories of agrarian expansion. This land was endowed with the best supplies of everything needed for agriculture. It was prize territory and required the most intense internal controls and protection. Medieval kings concentrated on controlling this land, to protect their people and their prosperity, which involved coercion as well as cultural powers to inculcate deep beliefs in principles and values that sustained agrarian order. Around core dynastic sites swirled all the activity of territorial expansion; and as populations in core sites increased in number, some of their number would strike out to expand agrarian power. They formed scattered settlements that became new dynastic centres, conquered other farming communities, and fought for land and labour with pastoral and forest peoples. Non-farming populations in the hills and plains often settled down to farming in the lowlands, forming their own distinctive communities. A separate agrarian history unfolded in the high mountains, of which we have little record.


In this diversity of agrarian social forms, a "peasantry" is hard to define. Unlike Europe, South Asia contained tropical conditions suitable for intensive paddy cultivation, expanses of arid and semi-arid plains, high-quality soils that could produce nutritious millets with relatively small labour inputs, vast tropical mountains and jungles, and large areas dominated by pastoralism -- all of which sustained very different types of agricultural expansion and intensification, leading to various configurations of agrarian society. In South Asia, there was no analogue to the Roman Empire or Catholic Church under which a feudal nobility could establish itself and define the peasantry as a category of subordinate subject. Unlike China, agrarian states in South Asia evolved significantly within, among, and out of pastoral cultures and they integrated pastoral and forest people into forms of agrarian society that were not embraced by the classificatory system of a single imperial (and ethnic, Han) heritage. Modern images of the peasant that come from western and eastern Eurasia -- which describe a rude rustic living under the jurisdiction of urban elites who embody high culture and civilisation -- do not fit medieval South Asia.


The term "peasant" can be useful to refer in a general sense to family farmers, but the theory of the peasant family farm -- developed by A.V.Chayanov, in the 1920s, to counter V.I.Lenin's theory of capitalist development in Russia -- is not implied by its usage here. Rather, the role of kinship in organising agricultural work and agrarian assets is being highlighted, and as we see in this chapter, elaborations of kinship organised much of medieval agrarian space in lineages, clans, jati, sect, and varna groupings, which included both farmers and kings. The peasant as a family farmer has no fixed class status. Class divisions between peasants and lords took many forms and medieval farmers were encumbered by many types and degrees of subordination, ranging from mere tax or rent obligations for land entitlements to intimate personal servitude. Institutions of control and subordination are the subject of the remainder of this chapter. The most intense subordination of farm families appears to have arisen where very low sudra and untouchable caste (jati) groups worked under Brahman and kshatriya domination in the rice-growing Gupta core territories and in early medieval lowlands along the coastal plains. But this is not a general pattern. Farm families enter the ranks of local ruling elites in many regions of militant peasant colonisation. In agro-pastoral and tribal settings, family farming was a communal enterprise which included military control over mobile resources and shifting farm territories.


The term "peasant" makes the most sense when agrarian social strata are clearly defined by states and when status depends upon strictly ranked entitlements to land. This situation became more common in the second millennium. It began earlier in territories of warrior colonisation, for example, by Gurjara-Pratiharas, when conquest formalised the ranks of lord and peasant. After 1500, social ranks in some parts of South Asia came increasingly to resemble Europe, and after 1820, European categories came into vogue under British rule. In the twentieth century, many political activists call themselves "peasants," modelling their usage on revolutionary Russia and China. As we see in Chapter Four, this usage appears primarily in tenant struggles against zamindari landlordism, where ranked entitlements to land are at issue. Here the term is ideological and normative, rather than being accurately descriptive. As a translation of kisan, "peasant" has been deployed where "landholder," "farmer," "village petite bourgeoisie," or even "tribal" could also apply; and it is usually more accurate to refer to so-called peasant groups by the ethnic or jati terms that they use to refer to themselves. Raiyat, which might also translate as "peasant," attaches to people with various types of entitlements and class positions, as we will see. No term translates strictly as "peasant," with precisely the same cultural connotations, in any South Asian language. In this context, we can aptly consider the rise of the utility of the category of "peasant" in South Asia as a product and component of modernity and use this term to discuss the power position of small farmers and tenants in opposition to landlords and states.


The term "gentry" is not widely used in South Asia but it does have utility. Multi-caste agrarian farming elites were formed by the interaction of state elites and local patriarchies in many places, by expanding family alliances, upward mobility, and the imposition of state-enforced, ranked entitlements to land. The term "gentry" has had no place in official terminology in South Asia, as it has had in China, but an important sector of the village farming population in South Asia more resembles Chinese gentry than European peasants. I consider the gentry to consist of relatively high status local land owning groups that marry their own kind and form alliances with other high status families to expand their horizons as they retain ties to the land. Gentry families are privileged as mediators with state authorities; and because of their land holding, education, and urban connections, they are active in commercial networks. This agrarian status elite is always open to new recruits. It is rural and urban, economic and cultural, social and political. A gentry first arose in the context of Gupta state rituals, which produced dominant caste alliances that came to control agrarian assets of all sorts, including the labour of subordinate jatis. The idea of a locally dominant caste cluster maps with rough equivalence onto my sense of what a gentry is; though a gentry does not need caste ideology. Other status formations and technologies can serve the same purpose.

Dharma

Inscriptions indicate that royal Gupta lineages had settled down in all the regions of the Gupta realm and may have been settling the frontiers of Aryavarta in the sixth century. Ambitious lineage leaders may have loosened their ties to the capital as they moved farther afield, carrying with them the apparatus of Gupta power. In frontier regions, they would have needed local allies, who may have undermined their attachment to the Gupta dynasty. Gifts of land by kings and their officers to temples and Brahmans -- to sustain classical learning, the rule of dharma, and the worship of Puranic deities -- became a hallmark of new dynasties at the end of Gupta hegemony, from the sixth century onward. The Maukharis appear in the western reaches of the Gupta heartland, around Kanyakubja (Kanauj), in what would later become Awadh, and Pusyabutis emerged in the western Yamuna basin and Haryana. Dynastic core sites thus moved still farther west from ancient heartland of Bharat and so did land grants to Brahmans, which multiplied with the founding of new regimes and capital cities. In the seventh century, it is said, Harsha moved the Pusyabuti capital to Kanyakubja to better defend the plains against the Hunas, but his move also signalled the rise of the western parts of the Ganga basin as a new agrarian core for his dynasty. This event was marked by a land grant to two Brahmans. The grant was made by a soldier serving Harsha and protected by the janapadas in Harsha's realm, to represent the support of local community leaders. New dynasties and donative inscriptions also multiplied in territories very far from Gupta lands. On the northern Tamil coast, the Pallavas of Kanchipuram stepped up their donations during the Gupta decline. Many new dynasties marked territories in sites of intensive cultivation: in Kashmir, the Karkotas of Srinagar; in Bengal, Later Guptas, Sasankas of Karnasuvarana, and Palas of Gauda (at the top of the delta); in Malwa, the Paramaras of Ujjain; in Malwa and adjacent Deccan and Gujarat, the Kalacuris of Mahismati; in Berar (Vidarbha), the Rashtrakutas of Acalapura; in the Krishna-Tungabadra doab in the south Deccan, the Calukyas of Vatapi; and on the south Tamil coast, the Pandyas of Madurai and Cholas of Tanjavur.


At least forty new dynastic lineages were proclaimed during and soon after the sixth century, and from the seventh century on, they typically construct elaborate genealogies for themselves to trace their origins to mythical progenitors. Migrations of Brahmans, Gupta princes, and Gupta generals may have influenced these early-medieval trends, but most new dynasties sprang up outside Aryavarta, and even peoples who had repulsed the Guptas later adapted technologies of power which the Guptas had developed. Between 550 and 1250, the interactive expansion of agricultural and dynastic territories produced the basis for all the major agrarian regions of modern South Asia. This is the crucial formative period for agrarian history in the subcontinent. Though agricultural conditions, techniques and social relations varied across regions, and though trends in the high mountains, western plains, and central mountains are not well documented, some basic elements which pertain to many if not most agrarian territories in this period appear in data from places where major dynasties were firmly established. We know most about elements that form the explicit subject matter of the inscriptions: kingship, Brahman settlements, and temples.


The ritual and architectural complex now called "the Hindu temple" emerged in full form in the later Gupta period and its elaboration and spread from the sixth to the fourteenth century provide us with dramatic medieval remains, from Mahabalipuram to Khajuraho. Medieval inscriptional records appear predominantly in temple precincts, which were central nodes for the accumulation of power in early medieval kingdoms. By the tenth century, old theories and practices of kingship had been widely adapted in many new medieval territories. As in Rama's mythical realm in the epic Ramayana, protection and prosperity were signs of a good king; and piety, chastity, and wealth all came together under kings who nurturedfs dharma. This theme forms a continuity with very old ideas about kingship, as Kasyapa's advice that "the king should take keen interest in agriculture," resonates with a Tamil poet's advice to a Pandya king, probably in the first centuries CE. Many medieval kings followed this advice, most spectacularly in Sri Lanka.

Water and ritual were critical for medieval kingship. So were innovation and adaptation. The kings who built medieval temples nurtured forms of dharma with distinctively medieval substance. In contrast to ancient prescriptions, medieval texts do not insist that a king be a kshatriya, and in much of the subcontinent, medieval caste ( jati) ranking developed without the presence of all four varnas. Rajadharma still meant protecting dharma, but sastras now prescribed that kings protect local customs, so that kings could enshrine as dharma virtually any form of social power and style of social ranking. Land grants to temples and Brahmans confirm the adaptation of cosmic law for local purposes, by bringing Brahman powers of ritual sanction to bear in new agrarian territories to sanctify patterns of social power in Puranic temple worship. Temples were ritual and also political institutions. They incorporated many different groups who were clearing and planting the land, building towns, and contracting transactional alliances, which are recorded in the inscriptions.


The Manusmriti says that dharma includes a sacred right of first possession for the people who clear the land, even if they have taken use of the land away from others -- for example, hunters and pastoralists -- which would often have been the case. Protecting this right was a bedrock of royal authority. But beyond this, a king's ambition extended over his neighbours (samantas), as the Pandyan poet says. Homage, tribute, and services from subordinate rulers (rajas) were prizes for kings who took titles like Maharajadhiraja (Great King of Kings) and Paramesvara (Supreme Lord). In the methodology of medieval kingship, gifts to Brahmans and temples measured and dramatised royal power. In Sri Lanka, the same method channelled massive patronage to Buddhist monks and monasteries from the second century BCE. Irrigation, paddy cultivation, and monastic property expanded together until the fourteenth century. Like monasteries, temples managed by Brahmans often owned tracts of irrigated land, and individual Brahmans and Brahman settlements (brahmadeyas) often received grants that comprised royal investments in irrigation. Like monks, Brahmans meant prosperity. They attracted people to agrarian sites who had skills and assets needed for expansion. A proliferation of texts on agriculture, astrology, medicine, and related sciences, and on temples and irrigation in the lowland areas that were most favoured for Brahman settlement indicate that a Brahman intelligentsia was busily working in many fields other than ritual and Vedic studies. Even esoteric learning could be useful in constructing an agrarian order. Shankaracharya's philosophy, for instance, concerns the disputation of sacred authority. Intellectual innovations in doctrine and ritual enabled local cults to be woven together into expanding Puranic traditions. Great temples multiplied with royal patronage. So did their poetic publicity. The greatness of the gods enhanced the glamour of royal patrons. Rich centres of temple worship combined in their precincts many of the technical skills -- controlled by Brahmans -- that were needed to develop agrarian territories, from architecture and engineering to law and financial management. Building a great temple or monastery attracted Brahmans and monks and provided a theatre of royal grandeur; here a king could make alliances and enjoy dramatic displays of submission. Great kings built great temples and commanded the services of the most learned Brahmans.


A donation to Brahmans, monasteries, monks, or temples represented an investment in agrarian territoriality. Inscriptions, typically carved on a temple wall, served as contracts and advertisements. The more popular a place of worship became -- the more praised in song, and the more attractive for pilgrims -- the greater became the value of its patronage. The rise of bhakti devotional movements enhanced the virtue of pilgrimage, increased the number of worshippers, and raised the value of temple donations. Making donations became increasingly popular among aspiring groups as a means of social mobility, as temples became commercial centres, landowners, employers, and investors. The rising value of temple assets increased the value of membership in communities of worship. Increasing participation in temple rituals made them more effective sites for social ranking: temple honours were distributed according to rank, and all worshippers were positioned in ranked proximity to the deity. Rulers came first. Popular bhakti devotional movements generated more popular religious participation, more ritual power for dominant local groups, and more glory for kings (even when temple were not centres for royal cults, which they sometimes were). Devotionalism produced a populist ideology for alliances among dominant agricultural lineages and warrior kings, and formed communities of sentiment among disparate groups involved in agricultural expansion.


Temples were divine sites for enacting social rank among devotees who protected dharma and sustained ritual; and like kingship, rituals changed as people brought gods into changing agrarian contexts. A wide diversity of rituals brought rain, secured crops, drove away disease, delivered healthy babies and bolstered dynasties; but among all the rituals -- by all kinds of spiritualists and officiates, from all kinds of social backgrounds, in all manner of locations -- those by Brahman priests for Shiva, Vishnu, and their relatives produced most surviving documentation, because of their lasting, widespread influence. Impressive temples came to mark agricultural territory, towering over the land as sacred landmarks. For many centuries, Brahmanic rituals had evolved as a potent force in social ranking and alliance building, and specifically for ranking dominant groups in relation to royalty; and it appears that with the expansion of temple worship and popular devotionalism, principles and practices of ritual inclusion and participation provided a template and methodology for the construction of social power in agrarian territories where the most powerful people traced their sacred genealogies to the gods of the Vedas.


Ritual powers which had been confined to Vedic ritual space were generalised within farming households, communities, and kingdoms. Brahman communities spread and Brahman models of social order spread with the influence of temples and temple patrons. In communities of temple worship, the roles and terms of Vedic ritual assumed new, mundane meanings. It seems, for instance, that the Vedic jajman was transformed into a person who controls the resources needed for temple ritual, which came to include not only food and animals but also the services of workers. Hence a jajman would be the pivot of power in circuits of redistribution, and in what would be called "the jajmani system," in which village land-owning elite families receive labour and assets from subordinates throughout the agricultural year and distribute produce from their land to workers at harvest time. Inscriptions also support the inference that agrarian territory became bounded by dharma as ritually ranked circles of marriage and kinship evolved into ranked caste groups (jatis). Coercion was certainly involved in the creation of caste societies, but the practice of ranking jati groups according to the principles of varna would also have been attractive for many groups. The adoption and enforcement of caste norms consolidated and expanded caste social space as it organised agriculture and sustained agrarian states.


Religious rituals of social ranking enabled families to form political alliances by providing measures of their respective status within agrarian territory. The labour, land, and assets of low-ranking jatis were organised for production by being subordinated to the power of dominant caste families who carved out territories in strategic transactions with Brahmans, gods, and kings. Dominant caste alliances thus formed medieval agrarian territories at the intersection of kingship and local communities. The expansion of caste society appears to have been be a top-down process, which included but did not necessarily depend upon everyday coercion. It might be best characterised as an evolving caste hegemony, in which the coercive features of social power were hidden by an ideology of dharma that became widely accepted because it provided everyone a place in the ranks of agrarian entitlement. Inscriptions further support the proposition that jati ranking was propelled by strategies of alliance among rising powers in agrarian society. New dynastic realms and agrarian territories were places for social mobility where the building of ranking systems made good sense. Dynastic lineage leaders and Brahmans were critical actors in creating these systems of social difference, status, rank, and power, which enabled powerful non-Brahman families to become gentry.


Temples and Brahman settlements were sites of honour around which to form ranks of privilege. Like kings, Brahmans acquired their rank in society historically. Building dynastic territory was a complicated process that involved innovation, risk, improvisation, and experiment; and the wide open spaces of medieval times enabled people to become Brahmans as well as Rajputs, kshatriyas, Marathas, and sat-sudra gentry. To become Brahman meant to be accepted as Brahman, patronised as Brahman, respected as Brahman; thus one came to command the skills, social status, and kinship of Brahmans. Controls over access to Brahman status were strict, so that entering Brahman ranks would have been difficult. But the widespread establishment of new Brahman settlements -- duly recorded in the inscriptions -- provided many opportunities for new Brahman lineages and clans to be formed. Founders and protectors of Brahman settlements, builders of temples, and donors who financed temple rituals were the moving force behind the Brahmanisation of agrarian territories. Land grants to temples and Brahmans are therefore less an indication of traditional Brahman power or peasant subordination than a reflection of alliance-building by aspiring agrarian elites, who used ritual ranking to lift themselves over competitors, and institutionalised their status by patronising gods and Brahmans.


Giving land increased the status of a donor and allied "protectors" of the grant, who are also often named in inscriptions; and by extension, these donations elevated all their kin. As kinship circles formed around the lineages that fed gods and Brahmans, whole sets of kin groups, forming as high-status, non-Brahman jatis, elevated themselves above others in temple ritual and agrarian society. This may explain why leaders of janapadas in Harsha's realm protected a gift of land to Brahmans, and why one of his generals made the gift under Harsha's authority. In the open spaces of Rashtrakuta power, one inscription records a gift of 8,000 measures of land to 1,000 Brahmans, and 4,000 measures to a single Brahman. Similar generosity is evident in many places in the medieval period. In each specific context, an inscription of this kind appears to mark an effort by a non-Brahman power block to enhance its own status and that of its local allies. Such gambits were not without risk and opposition. Raising the status of some groups lowered others, and Brahman settlements created ranked entitlements in everyday social transactions. Brahmans did not usually farm land themselves and they were entitled by grants to receive the produce of farms, including taxes; so they became a significant social force, protected by state power, and also a landed elite whose well-being depended upon the control of other people's land and labour. Brahman lords of land established a model for other elites. Brahmanisation sustained the rise of landed elites and aspiring royalty, whose superior claims to land and labour were legitimated by their patronage and protection of Brahmans.


A small but significant set of inscriptions records opposition to Brahman settlements, to their collection of taxes, and to their claims on local resources like pastures, often contained in grants. Most opposition seems to come from leading members of local farming communities. The authority of kings who patronised Brahmans was clearly not accepted by everyone in medieval societies; and the authority of Brahmanical kingship spread slowly -- often violently -- into the vast spaces that lay outside its reach in early medieval centuries. In many instances, land grants appear to mark frontiers of royal power, and here the most resistance might be expected. Even where local society did accept the ritual and social status of Brahmans, fierce competitive struggles might flare up over land grants. Some opposition to Brahman settlements certainly came from local competitors who were fighting against the families who sought to elevate themselves by patronising Brahmans with land grants. In the ninth century, local conflicts of this kind accompanied new Brahman settlements on the Tamil coast, where they were an old and widely accepted feature of agrarian territoriality. The open spaces of Rashtrakuta ambition were another matter. Inscriptions from the northern peninsula warn that violence and curses will be heaped upon opponents of Brahman land grants, and texts proclaim that people who murder Brahmans will be punished harshly, which implies that such murders did in fact occur. But striving lineages also had options other than revolt, for as Bhisma says in the Mahabharata, "If the king disregards agriculturists, they become lost to him, and abandoning his dominions, [they] betake themselves to the woods."7


Territories of permanent cultivation, irrigation, worship, pilgrimage, dynastic authority, temple wealth, jati ranking, caste dominance, and Brahman influence grew together. Inscriptions depict idioms of territoriality as they pinpoint the location of a royal donor, the ranks of officials involved in a grant, the status of local protectors, and the names of religious personnel and institutions. Power relations in agricultural expansion were thus more complex than we can see using a simple division of the agrarian world between the state and society. There was more at work in the medieval political economy than interaction between kings and peasants or dynasties and villages. The most important social forces within medieval states worked the middle ground between rulers and farmers, where leaders of locally prominent families made strategic alliances that constituted dynastic territory. An agrarian gentry thus emerged as a constituent of royal authority. The constant rhetorical and ritual elevation of the king above all others mirrored and mobilised social ranking; it served the cause of gentry mobility. The superiority of rulers served all subordinates by elevating them above lower orders in their relative proximity to the king. When an aspiring warrior family elevated itself by declaring a new dynasty, it benefited the whole clan and their home locality. Inviting Brahmans to live in its territory, generating for itself a cosmic lineage, building temples, and adopting royal titles and rituals, the new king would pursue allies. In this pursuit, recognition by an established, superior king could be a boon. Kings would thus extend their domains by forming unequal alliances with samantas whose subordination would raise their local status. Subordinate rulers could then support or protect a grant in the great king's name, to further enhance their status. A rising dynasty would then accumulate its own subordinate samantas on the periphery of its core territory, while subordinate rulers on the frontier would improve their position at the same time. Core regions of agrarian expansion would expand as emerging leaders allied in regional dynasties, and leaders on frontiers of several royal territories shifted loyalties or combined them. A successful samanta might seek to overturn his master, so the advice of King Lalitadiya of Kashmir (in the eighth century) makes good sense: "Do not allow the villagers to accumulate more than they need for bare subsistence, lest they revolt."


The early medieval period -- from the sixth to thirteenth century -- lay foundations for later agrarian history in many respects. New forms of social life emerged in many places at the same time. What M.N.Srinivas called "Sanskritisation" evokes part of the process, because social groups and institutions were being formed around models of behaviour, identity, aesthetics, and patronage codified in Sanskrit texts. Brahmans were key people because they sanctified social rank and political alliance. Rising families wanted to hire Brahman genealogists and court poets, patronise Brahman and temples, endow feeding places for mendicants and pilgrimages, stage festivals, feed saints, and join the activities that united gods, priests, kings, and farmers. All this occurred as farm land was expanding and as peasant farmers, nomads, pastoralists, hunters, and forest tribes were slowly changing the substance of their social identity, over many generations, as people became high caste land owners, kings, protectors of dharma, kshatriyas, vaisyas, superior sudras, inferior sudras, untouchables, and aliens beyond the pale. Such transformations obscure the ancient identity of the people who propelled medieval agrarian history, but the result was that gentry castes filled the ranks of landowners and ruling lineages. Many of these groups remain powerful in their regions in modern times. They became Kunbi, Vellala, Velama, Reddy, Kapu, Kamma, Nayar, and other landed castes. The ancient social background of some dynasties can also be dimly perceived. Hoysalas came from Melapas, hill chiefs in the Soseyur forests. Udaiyar and Yadava dynasties descended from herders. Tevar kings descended from Marava and Kallar hunters. Many Marathas had ancestors in the hills. Gurjaras certainly had a pastoral past. Rajputs did not have one original identity but emerged from histories of warrior ranking and mythology, and many had ancient pastoral and tribal roots. All these transformations are entangled in the politics of religious leadership, devotion, and loyalty; and every state in the history of South Asia has afforded special privileges -- including tax exemptions -- to religious institutions and religious leaders. Many social movements that moderns might call "religious" might be better understood perhaps in the context of agrarian territorialism, as we will see in Chapter Four.


Geographically, early medieval territories seem to concentrate in riverine lowlands. Here the influence of Brahmans and medieval Sanskritisation was most compelling. In scattered rice-growing and irrigated lowlands, aspiring elite families patronised Brahmans and Puranic deities as they fought to control prime farm land and to creating sacred rights of first possession. Mapping agrarian space with inscriptional data is not easy at present because each transaction occurs within a moving constellation of dynastic donations and inscriptional sites themselves have not been mapped comprehensively. Territory was not defined by fixed boundaries but rather by the reach of individual transactions. Transactional territoriality remains in place only as long as the social relations formed in transactions, and inscriptional territory was defined by transactions among dominant social or ethnic groups and families. This kind of territoriality did not disappear. In the nineteenth century, a local official reported from the old Gupta core region in Saran District, Bihar, that "Brahmans, Rajputs, and Bhumihars were the only castes that figured in the `actual life' of the district."8 In early-medieval times, such groups were coming into existence, their social identities were being produced transactionally, and carved inscriptions were clearly intended to perpetuate their reputations.


Medieval kingdoms were also transactional. They did not contain bureaucratic institutions of a sort that would define a fixed territory of revenue or judicial administration. In the 1950s, K.A.Nilakanta Shastri argued that the Cholas had built an "almost Byzantine royalty," and since the 1960s, R.S.Sharma has argued that post-Gupta states in general represent a form of feudalism. Historians have developed alternatives to these models, but they have not replaced them, and today there is no consensus concerning the nature of medieval states. But inside dynastic domains, inscriptions indicate that agrarian territories were small, consisting of settlements linked together by locally dominant caste power. Inscriptional terms mark transactional space by using titles for individuals and groups and by using place names attached to the people in transactions -- terms like nadu or padi. The nadu in modern "Tamil Nadu" is a territorial marker from medieval inscriptions which designated a tiny region that was defined by its ritual, land, water, kinship, and royal transactions. The term appears most often in references to leaders of the nadu who engaged in donative transactions with temples and Brahmans. In later centuries, it acquired more expansive meanings; but medieval nadus along the Tamil coast were composed of core settlements along routes of drainage, where Brahmans and temples received land grants, and they were connected to one another by donative transactions and surrounded by vast tracts of land which was not controlled by dominant groups in core nadu sites.


Where inscriptional sites of agrarian power have been mapped, they cluster along rivers, so it appears that the nexus of power that they reflect concentrated on the control of riverine farm land. Over time, Brahman influence spread widely, and Brahmans and allied gentry and service castes became mobile state elites. Locally, gentry caste power dug in, and expanded steadily. Based in prime locations along rivers and trade routes with clusters of temple towns and old dynastic capitals, caste communities incorporated tribal, hunting, pastoral, and nomadic groups into the lower echelons of society, where the new entrants retained much of their character, redefined in caste terms. Tribal deities entered the Puranic pantheon, adding cultural complexity and expressing the richness of agrarian territories. The social and cultural character of agrarian regions emerged in later centuries, but medieval territoriality left its traces and imparted distinctive qualities to localities by giving special importance to Brahmans, temples, and high caste gentry. Where we do not find medieval donative inscriptions -- as in Punjab, in Jat territories in the western Ganga plains, and in the mountains -- the population and cultural importance of Brahmans remains comparatively small in agrarian societies today. Inscriptional territories concentrate in eastern and central Uttar Pradesh, West Bengal, Gujarat, western Maharashtra, and along the coastal plains. Brahmanism seems is less deeply rooted in other areas, where other types of social power were more prominent in the later formation of agrarian territories.



Conquest

One Candella inscription announces that Anand, brother of Trailokyavarman, reduced to submission "wild tribes of Bhillas, Sabaras and Pulindas."9 Conquering tribes and expelling them from the land is a major theme in genealogy and epigraphy. Warrior lineages expanded in number and influence after the Guptas, creating and conquering agricultural territories; and preserving their victories in inscriptions, epic poetry, folklore, and hagiography. From the eighth to the eleventh century, Gurjara-Pratiharas conquered along western plains and northern basins, moving into the central mountains and high mountain valleys. Along the coast of the southern peninsula, Pandya, Chera, and Chola lineages repeatedly conquered each others' territory. Many warriors planted settlements of warrior-farmers along routes of conquest. From the eighth to tenth century, Rashtrakutas migrated and conquered territories all across the north Deccan, Gujarat, and Orissa. The Vakatakas emulated earlier Satavahanas and expanded outward from Vidarbha (Berar); they split into four branches with shifting capitals spread from Chhattisgarh across the Deccan to the upper reaches of the Bhima river basin. In the sixth century, Kalacuris appeared in the north Deccan territories of the Vakatakas; in the eighth, inscriptions show them in Tripura, near head of the Narmada River; and as late as the thirteenth century, they appear in Bengal. Chola armies conquered northern Sri Lanka, leaving a population of colonists behind. Calukya lineages had bases in Vatapi along the upper Tungabhadra River in the seventh and eighth century, and in Kalyani, well to the north, in the eleventh and twelfth. From the fourteenth to the sixteenth century, Vijayanagar dynasties expanded from the vicinity of old Vatapi across the southern Deccan and eastern plains; and they shifted their capitals into coastal lowlands as they broke up into smaller dynasties in sixteenth and seventeenth century. Under Vijayanagar, Telugu warrior clans conquered territory all along the south-eastern coastal plain.


This is only the start of a list of medieval conquests and barely begins to indicate how widely they influenced agrarian territoriality. Conquest colonisation exerted its influence quite separately from territorial dharma, but together they produced agrarian territories that expanded to become agrarian regions. Conquering colonists knitted together many small-scale domains of localised, dominant caste power; they connected new frontiers and old dynastic sites. Activities depicted in Harsha's seventh century inscription are typical of many events in this process. Many donative inscriptions reflect similar activity by subordinate officers and local leaders, for whom gifts to Brahmans and temples represent incorporation into conquest territories. Conquest created ranks of warriors above the locally dominant castes; and superior entitlements created by conquest moved agricultural wealth along trade routes of warrior power in the form of tribute and temple endowment. Warriors used agrarian wealth from one location to support conquest in others; and traders, pastoralists, and warrior-farmers moved agrarian wealth among sites. These transfers of wealth employed monetary instruments and commercial intermediation, which are most visible in the inscriptional evidence as they pertain to temple endowments. Conquest and trade went hand-in-hand with religious endowments and investments in farming.


Warriors created territories of authority in which their officers could establish local roots within communities, where they distributed local entitlements to kin, allies, and subordinates. Distant localities were connected to one another by the social networks that formed along routes of conquest and routes for the extension of conquering clans. Local caste elites were assimilated into extensive realms of clan power as subordinates, dependants, and rising stars. Some warrior clans created a non-farming nobility living high in fortress-towns, looking down on farm communities; these became landmarks on agricultural landscapes, as prominent as temples and other sacred places. Warrior competition among siblings in each successive generation would send another wave of fighters out to conquer new territory, and in this enterprise, clans faced one another on the battlefield, so that battles among warriors became a dominant motif in hagiographies, genealogies, and local lore about the land. The exploits of great men became material for epics, rumours, gossip, and popular songs, landmarks in local history.


Two broad geographical zones of warrior influence can be roughly discerned. One was formed by clans that became Rajputs -- Gurjaras, Cahamanas (Chauhans), Paramaras (Pawars), Guhilas (Sisodias), and Caulukyas (Solankis) -- who conquered from the eighth century onward across the western plains, northern basins, adjacent high mountain valleys, and central mountains. Local leaders rose up to ally with and to join the ranks of conquering clans, by imitation, alliance, genealogical invention, inter-marriage, and combinations of these strategies. Rajput rulers protected dharma and became ideal kshatriyas. Brahman settlements and temple rituals were not as important for Rajput royalty as warfare, genealogies, hagiographies, and court ceremonies. Their lineages measured their status in victories, alliances, marriages, and accumulations of tribute in their palaces, forts, and market towns -- including Brahmans and temples. Nobles endowed temples and employed Brahmans as ritualists and servants, but it was more their devotion to battle, to their own clan, and to the rules of martial kingship that measured their devotion to dharma. In warrior territory, the ritual ranking of agrarian society followed the logic of warrior lineages, stressing military alliance, victory, service, and publicly demonstrated powers of physical command and subjugation. Agrarian social power concentrated much less in a local landed gentry and much more in warrior lineages. Genealogies became records of rank and they proliferated across the entire span of the medieval epoch to mark the expanding area of Rajput influence. Surajit Sinha has shown that "state formation in the tribal belt of Central India is very largely a story of the Rajputisation of the tribes."10 The interaction of expansive Rajput lineages with locally powerful Jat clans produced a militarist pattern of agrarian development in the western frontiers of old Bharat, where agrarian power focused on fortified villages and strategic hill towns.


A second warrior zone lay south of the Vindhyas in Khandesh, Berar, Maharashtra, central mountain valleys, and the interior peninsula, where warriors were attached to agricultural communities and concentrated power in their own hands but followed no single dominant model like that of the Rajputs. Instead -- from the time of the Satavahanas, Vakatakas, and Rashtrakutas, to that of the Yadavas, Chalukyas, Hoysalas, Kakatiyas of Warangal, Sultans of Ahmadnagar and Bijapur, Udaiyars and Sultans of Mysore, and the Marathas -- dominant social powers in agriculture arose from and mingled with the evolution of peoples living in and drawn from pastoral, hunting, and mountain populations. Standing to fighting became part of farming. Running off to war became part of the agricultural routine. In the interior peninsula, extensive tracts of dry land were open for cultivation for the people who could fight to keep it; and throughout medieval centuries, cattle, manpower, centres of trade, strategic positions, and places with good water and access to valuable forests were more valuable assets for aspiring warrior lineages than most farmland, except in the stretches deep black soil along the big rivers that were perpetual sites of warrior-peasant competition. The dominant agrarian castes were both warriors and cultivators. In the Maratha Deccan, for instance, expanses of open land that were available in the sixteenth and seventeenth century allowed many warrior-farming lineages to carve out local territory for themselves. A lineage leader would become patel (headman); and by combining his military power with alliances with regional rulers, he could become the deshmukh for a circle of villages, and thus a player in regional politics. In such contexts, there was little scope for the rise of an agrarian gentry except when warrior-farmers could control an area long enough to build up an agrarian elite, which eventually did occur later in the medieval period in Maratha core areas, as well as among Vakkaligas and Boyas in Karnataka, and among Reddy and other locally dominant Telugu castes in Telangana.


In the eleventh century, warriors came on horseback from Afghanistan and Central Asia to engage warriors in the lowlands and they swept across the subcontinent to connect the zones of warrior colonisation north and south of the Vindhyas, which had been until then quite separate. They pushed into historical spaces of conquest colonisation that were many centuries old. The military competition that ensued increased the influence of conquest colonisation on agriculture as a whole by increasing the number and force of warriors. Because temples, cities, and irrigation represented authority and prosperity, they were natural targets in war.11 Warriors often dislocated farming communities when they attacked their enemies. The great eleventh-century irrigation builder, the Paramara Raja Bhoj, built a wall to form a huge irrigation reservoir at Bhojpur, near Bhopal, and armies of the Sultan of Mandi, Hoshang Shah, cut a dam to destroy the lake, killing the irrigation. This kind of warfare discouraged heavy investments by farmers in fixed agricultural assets like irrigation works, unless they could be protected by local arms, and wells could be, which helps to explain their popularity in all the regions of conquest colonisation. A militarisation of farming occurred in all the dry regions where wells are most important. By the fifteenth century, professional armies had established their military superiority over the local hordes of warrior-peasants, but no state could destroy the independent warrior lineages which had fought successfully to control local territories. In wide fields of warrior-farming, lineages expanded their cultivation and coercive power at the same time, under medieval dynasties. Warrior Jats colonised Punjab and the western Ganga basin and formed agrarian mini-polities which became regional states in the eighteenth century. Stable centres of professional military power emerged in many territories in the interior peninsula, some, as on the Mysore plateau, with old formations of caste dominance, endowed with all the institutional traditions of royalty and Brahman patronage.


Conquest colonisation made much of agrarian space into a battleground, and the careers of the Vijayanagar and Maratha empires reflect important features of agarian history. In a region of Telugu conquest, Vijayanagar, "the city of victory," was built in the fourteenth century, in old agricultural territory near the old Chalukya centre of Vatapi, along the upper Tungabhadra River. Endowed with magnificent irrigation, the city accumulated so much warrior wealth that Portuguese visitors took it be richer than Paris, with its great temples, royal cult, and vassals arriving at festival time with mountains of tribute. Fighters called nayakas spread Vijayanagar's dominion outward in waves of conquest colonisation, creating the first empire to embrace all the land south of the Krishna. Telugu warrior-peasants opened up new land for dry farming along the eastern coast -- in dry stretches filled with deep black cotton soil between old riverine clusters -- and these colonies would sustain a vast expansion of cloth manufacture for world markets in the seventeenth and eighteenth century, when Nayakas became kings. Vijayanagar itself was destroyed, however, by sultans at Bijapur, another site in the old Chalukya realm. The city of victory had disappeared completely by the end of the seventeenth century.


In 1640, Shivaji was married in Bangalore at the court of the Bijapur sultan, whom his father, Shahji, served as a general. Shahji's jagir of revenue villages, near Pune, became Shivaji's patrimony and base for later military expansion. In his day, good rainy seasons and peaceful times saw farmers expanding cultivation, though even in good years, they would go off to war during dry months. Building in one place accompanied destruction elsewhere. Droughts would routinely drive more farmers out to fight, loot, and colonise. The great fortresses that became great cities -- like Vijayanagar -- could disappear without a trace, because their ruling lineages and urban populations would move overnight to other sites. New cities and towns rose on ruins everyday in the world of warrior-peasants and military entrepreneurs. One story goes that in the seventeenth century, on the fringes of Maratha power, in the southern Deccan,

Doda Mastee and Chikka Mastee, two brothers with their families and cattle, came from the north and built two houses. They cleared the jungle around and maintained themselves by cultivating the soil. They invited inhabitants from other parts, advancing them money to increase the cultivation. The brothers next built a fort and gave their village the name Halla Goudennahally .... In time the brothers quarrelled because the younger brother Chikka Mastee made a tank which endangered the village. Doda Mastee being displeased with his brother's folly removed and built north of the first village another village called Goudennahally and all the inhabitants moved to the new village ... Halla Goudennahally was in the meantime inundated by the water of the tank.

Building and breaking, moving and settling -- old themes -- but the story elides a suspicion that Doda Mastee broke the tank of his younger brother to force the villagers into his new settlement. Building a new settlement often entailed violence, much like planting new fields meant cutting down the jungle. In 1630, when villages in his revenue territory (jagir) lay waste from famine, Shivaji's guardian, Dadaji Kondev, "set about repopulating and developing the jagir," for which purpose, "Deshpandes were seized and taken in hand, [and] the refractory among them were put to death."12


From the eighth century onward, conquest colonisation is well enough documented to allow us to distinguish different types which can be identified by association with specific groups. (1) Some professional warrior lineages emerged from ancient roots at the margins of old agricultural societies, from pastoral and hunter peoples for whom extensive mobility and killing were ancient skills. These warrior specialists conquered agricultural communities and formed a nobility. For simplicity, we can call this the Rajput pattern, with its various spin-offs and variants that arose with emulation and alliance-building among other warrior groups. (2) Some warrior lineages came from the ranks of dominant caste farmers who had formed a military force and allied with professional warriors, even serving under them. Conquering widely, these fighting peasants would retain their local agrarian base. If successful, they could spread their power by alliance and conquest across territories dominated by allies in similar warrior-peasant castes. This is the outline of the Maratha pattern, which had a long history in the Deccan before the rise of Shivaji. (3) A warrior lineage could split off from a dynastic authority, using symbols and alliances derived from that dynasty to lead fighting colonists into new territories, to conquer, displace, and/or assimilate local tribes and others, and thus to form new agricultural communities and new agrarian territories. This pattern, exemplified by the Mastee brothers, also typified Telugu colonists in the black soil tracts of the Tamil country. (4) Warrior-farmers could simply conquer and settle new territory, dividing the land among the lineages. The conquering group itself would form the bulk of the agrarian population in this territory. Kings would rise within it and clans would differentiate over time into many ranks. Groups would splinter off for new colonisation. This pattern may be the most common of all and probably dominated the hills and valleys of the central mountains, for instance, in Gondwana, Chhattisgarh, and other areas of tribal Rajputisation. It also typifies lowland warrior-farming groups like the Maravars and Kallars in southern Tamil country, including the Piramalai Kallar studied by Louis Dumont.


Late medieval warrior lineages -- from the Ghorids to the Mughals -- followed the Rajput model, which was prevalent in various forms in Central and West Asia.. As they subsumed other warriors under their authority, they increased the power of subordinates like Shivaji who could form local alliances with warrior-peasant lineages. Sultans had less interest even than Rajputs in farming themselves, living among farmers, or tinkering with production locally. They conquered warriors who already ruled over agrarian territories. They lived in fortress towns and their movements connected all the old fort-centres of warrior colonisation not only to one another but also to urban centres across inland corridors of southern Asia from Dhaka to Istanbul. Like their predecessors, the sultans brought retinues from their homelands and new technologies of power, and they encouraged migrants from their home territories to settle in their new dominions, primarily in the established urban centres, but also on agricultural frontiers, where like the Brahman settlements of the early-medieval period, ulema and Sufis provided erudition and leadership for agricultural expansion in the Indus Valley, Punjab, the western Gangetic plain, the Deccan, and eastern Bengal. Again, new skills and productive powers came into agrarian territories with new ruling elites. This was like the dispersion of productive powers that followed the end of the Gupta empire, but at a much greater scale, with more dramatic consequences, and more detailed documentation. With warriors sultans from the western reaches of southern Asia came architects, accountants, scholars, genealogists, bureaucrats, poets, scientists, merchants, bankers, musicians and the entire cultural heritage of Persia. The centre of gravity in Persian cultural history moved into South Asia, where a Persian lexicon and technologies of power organised a widespread reintegration of agrarian territories. New strata of non-farming elites were formed by grants from sultans of entitlements to revenues from the land. Again, royal patronage fed the rise of agrarian elites.


Sultanic regimes continued and reinforced long historical trends in conquest colonisation. Muslim rulers did not dramatise political alliances in temple ritual, but they sustained temple authority in agrarian territory. Old land grants to temples remained in place. New royal donations were added in the form of the tax remissions. Victorious sultans defeated old defenders of dharma, and this again brought Kali Yuga to mind for some Brahmans; some fled into mountain valleys and into rapidly expanding agrarian territories around Kathmandu. Sultans brought superior military technology into the field, which altered the competitive environment of conquest colonisation. Defeated warriors launched new waves of conquest, extending warrior power from the Deccan and the Ganga basin into deltaic and coastal regions. But imperial expansion still depended upon unequal alliances with subordinate rulers, who increased their own status by hitching their fortune to victorious sultans. Sultanic regimes developed institutions of military bureaucracy that focused authority on the emperor's family, relatives, and his highest ranking allies, who formed the imperial nobility; and from the time of Akbar's marriage alliance with Rajputs, the imperial wedding became a ritual of the highest statecraft. Rajput, Maratha, and other protectors of dharma formed a new nobility under the umbrella of sultanic regimes and increased their investments in temples; so that temple ritual and Hindu court culture flourished under the sultans. Akbar's and Dara Shuko's experiments in religion reflect a continuing effort to articulate political alliance-building with philosophical speculation; and rituals of theological disputation at Akbar's court remind us of the medieval innovations in religious thought and performance which accompanied the incorporation of new groups into agrarian regimes. Outside the court, eclectic mysticism and devotionalism expanded their reach. Brahmans were in high demand in the apparatus of imperial taxation and law, and their occupational horizons expanded. More than ever, building accompany destruction under the sultans, in the expansion of urbanism and in the expansion of agriculture. When Akbar's troops marched into Bengal, they brought in train tools and men to clear the jungles to expand cultivation. Sufis came into the eastern delta to open the jungles to farming. To protect strategic mountain routes of trade, Mughal armies conquered and settled many sites of agrarian expansion, including Kashmir. Aurangzeb began his famous 1665 farman on administration with words that echo Kasyapa and the Pandyan poet: "the entire elevated attention and desires of the Emperor are devoted to the increase in the population and cultivation of the Empire and the welfare of the whole peasantry and the entire people." Aurangzeb reiterated the Manusmriti on the sacred right of first possession when he declared that, "whoever turns (wasteland) into cultivable land should be recognised as the (owner) malik and should not be deprived (of land)."13


Patriarchy

Families passed the right of first possession from one generation to the next. At the base of medieval states and at the apex of early-modern empires, family formed a core of social power and experience. In agrarian territory, kinship formed basic entitlements to means of production. Kin groups joined together to clear land, build fields, dig wells, and cultivate. Settlements and communities formed around collections of kin. Marriage networks connected villages. Families pushed the frontiers of farming and fought for control of agrarian space within realms of ritual and conquest. Kinship underlay class and caste; and in state institutions, market networks, and community organisation, kin formed the most powerful bonds of alliance, allegiance, loyalty, and solidarity. The Ramayana and Mahabharata are fundamentally family dramas and the Manusmriti is obsessed with the implications of marriage for caste ranking. During medieval centuries, family histories, emotions, rituals, intrigues, conflicts, and loyalties permeated agrarian life and territoriality, from family farm to imperial court. Family suffused all the institutions of entitlement. On the Tamil coast, for instance, the word pangu, meaning "share," came to refer both to an individual's share in family property and to a family's share of village assets, so that pangali referred to relatives and also to share-holding landed gentry families in a farming community. The term kulam likewise referred to a household, lineage, clan, and local caste group (jati); and nadu meant an agrarian space (as opposed to kadu, forest) defined as the domain of authority of its prominent most family leaders (nattar). The inscriptional corpus is substantially the record of transactions among the heads of families who built agrarian territory, built dynasties, and travelled in search of new land to conquer and farm.


Kin followed the lead of earlier generations to create expansive domains of kinship which became localities, kingdoms, and empires. Along riverine tracts of irrigated agriculture where medieval inscriptions were most densely distributed, families sought control over expanses of farm land, grazing land, forest, and water supplies. Succeeding generations spread their power from one bit of cultivated ground to the next and prominent gentry families and an expanding set of kin folk produced small, compact, domains of dominance. Marriages formed dense links among dominant families in adjacent settlements which became related to one another in patterns that resembled the patchwork of paddy fields. In such settings, the norms and practices of kinship strongly stressed local alliances among families and they formed intricately graded ranks within gentry strata of society. Marriage also marked divisions between local elites and groups who were barred from owning land, who served the gentry as dependent servants and farm workers. These distinctions took many forms within the idiom of caste society. But in general the formation of solid traditions of local gentry and rural elite dominance entailed the reproduction of a genealogical connection to celestial and royal authorities who were the fount of their patrimony.


In this context, agriculture became a deeply patriarchal enterprise. Senior men ruled junior men and senior women lorded over their inferiors. Family ranking elevated all members of higher families, men and women, according to their lineage, clan, jati, varna, wealth, sect, office, or other mark of status rank. Patriarchy is a kind of power -- never absolute, uncontested, or unaffected by other kinds of power -- wielded by men in virtue of their rank in society, and agrarian patriarchy defined agricultural territory as a domain of ranks, entitlements, leading families, and family heads. In the ancient text, Milindapanho, Nagasena explains to King Menander that the people he calls "villagers" (gamika) are in fact the patriarchs who head the village families:

Now when the lord, oh king, is thus summoning all the heads of houses (kutipurush), he issues his order to all the villagers but it is not they who assemble in obedience to the order; it is the heads of the houses. There are many who do not come: women and men, slave girls and slaves, hired workmen, servants, peasants (gamika), sick people, oxen, buffaloes, sheep and goats and dogs -- but all those do not count.14

Many men and women did not come. Only the "heads" or the leaders of families. Such scenes have been re-enacted millions of times. Of course, the lord, the king, and the sultan are also heads of families, men of superior rank among kinsfolk and subordinates. Many medieval inscriptions depict the ranks of patriarchs being formed and reformed among men who head families at various levels of power. Family rank came to entail entitlements which became glossed as "property rights," but property in practice amounted to power over assets substantiated by and reflected in family rank. Property was also parcelled out within families to members according to rank. Agrarian territory came to be composed of proprietary units formed among families led by their senior men. From early medieval times, inscriptions indicate that property entitlements were often individualised and transferred in market and political transactions; but property was also defined and protected by social powers in communities and territories. The inscriptional authority of the local protectors of grants to Brahmans and temples indicates that powers of entitlement depended upon recognised leadership in janapada, nadu, pati, desa, grama, ur, and other named territories. Dynasties were established as patriarchies at the apex of territorial ranking systems; and states were thus composed of nested, ranked sets of family entitlements, defined by transactions among patriarchs in systems of ranking which are depicted in the inscriptions.


Patriarchy formed a dynamic, productive power which connected intimate family life with wide historical trends. Family is most often assumed to be a cultural constant within the realm of tradition. Scholars tend to think of kinship as a durable structure that is reproduced innocuously in private life. But family is an engine of change in politics and in struggles over resource control at every level of society. Marriage decisions, rituals, and alliances became politically important in ancient times as lineage leaders began building early state institutions. Elite matrimony became a political event of the first order. Competition among sons generated expansive political domains as younger sons went out in search of new territories, and in early medieval centuries, dispersions of ranked lineages created wide nets of alliance and competition. Medieval inscriptions record transactions among patriarchs, articulate ranks among them, and thus encode episodes in family history which formed agrarian territory and dynastic genealogies. Temple rituals articulate alliance, loyalty, devotion, and competition among units of patriarchal power and kinship. Many if not most groups with collective identities in South Asia use some form of genealogical reckoning to expressfamily feelings and histories. Ancestral patriarchs and mythical progenitors populate origin stories. In the dense forest of north Bihar, in the eighteenth century, a Kayasth named Dullah Ram founded the village of Changel, and local lore preserves the story that he obeyed a dream and found a horde of gold coins near the temple to the mother goddess, which he duly dug up "and his descendants lived happily ever after"; whereas the more prosaic truth is that Dullah Ram and his kinsfolk founded the village by usurping the land and subjugating the labor of the local tribal population that worshipped the mother goddess there.15 Landless Buinhya workers in south Bihar recount the victories of their heroic progenitor, whose valiant exertions clearing and turning the jungle into farm land dignify their subservient labour today.16 Vellala gentry on the north Tamil coast trace their origins to a royal Chola ancestor who migrated north with 48,000 Vellala families, conquering Kurumba hunters. Genealogies from Mysore and Andhra begin with great patriarchs. So do family histories among Rajputs and other royal families across the northern basins. A group that calls itself "rulers of the hills" (Malaiyalis) traces its descent from ancestors who migrated up from the Tamil lowlands with their gods in hand at the sharp end of enemy spears. Countless genealogies depict patriarchs as founding fathers who begin the chain of succession and entitlement that runs down over generations. Ancestral personalities become icons of group identity. Their exploits become collective accomplishments. "Our history" for many groups became a story of family feelings forged by lore and worship, beginning with great patriarchs whose offspring populated the land. As if to replicate earth in heaven, the Puranic pantheon filled up with marriages and families of gods. Earthly kings became descendants of Vedic divinities. Cosmic and mundane genealogies together defined social identities around powers and sentiments that linked families to one another in territories of divinity and heritage. Temples embodied the cosmic power of gods in the territory of patriarchs.


Caste -- jati -- defined units and idioms of family alliance and ranking within varna ideologies, but patriarchy also transcended caste and escaped the rule of dharma. Warrior kings connected disparate, distant territories to one another, and the rule of dharma could organise only parts of these expansive territories. In the sixth century, groups outside the ranks of caste society comprised the bulk of the population and though dharma did subsequently expanded its reach by various means, people outside caste society -- whether beneath the lowest of the low or outside the pale altogether -- remained numerous. Though excluded from temples and other rituals in respectable gentry communities, low castes and non-castes lived in agricultural territory. Because the power of caste society expanded downward from the top ranks and outward from centres of ritual and conquest, groups at the lowest ranks and on the margins of dominant caste control comprised a moving borderland between caste society and its surroundings. Outsiders in and around localities of high caste control were critically important for the vitality of every agrarian locality, and many did enter into the rituals of dharma in various ways, but many also remained outsiders. Such people continued to arrive in every agrarian territory with new waves of migration and conquest colonisation throughout the first and second millennium. Idioms and practices of patriarchal alliance allowed for the loose inclusion of countless groups within transactional territories formed by systems of market exchange and political ranking. Lineage and clan leaders among tribal groups, merchant patriarchs from distant places, travelling artisan headmen, nomadic chiefs, and military commanders from virtually any background could form alliances with locally dominant caste patriarchs based not on their caste ritual rank but rather on the mutual recognition of their respective patriarchal powers. Heads of households and heads of state could negotiate as patriarchs because they could rely on one another to command the labour and allegiance, assets and loyalty, of their kin folk. This produced trust, confidence, and stability in transactions that relied upon payments in the future for promises in the present, whether loans, contracts, or agreements to pay taxes in return for entitlements to land.


Patrimonial entitlements thus defined property rights and powers over labour independently of the rituals of caste and temple worship; and transactions that formed proprietary entitlements also produced state revenues as well as profits and capital for the market economy. Medieval inscriptions depict very complex market transactions, and in many cases, temple donations represent tribute. In payments of tribute, conquers, kings, and financiers took payments from local patriarchs in transactions that constituted ranks of patriarchal entitlement. Routines of tribute collection became systems of taxation as they became routinised and acquired ideological legitimacy in agrarian cultures as instruments for transactional ranking and entitlement. Within the rituals of taxation, dharma, markets, and conquest, patrimonial property became securely established as a foundation for agrarian authority.


As disparate groups with different backgrounds settled in agrarian territory and worked with one another over generations, they developed complex etiquettes of rank, deference, and residential segregation, expressed in housing, personal habits, marriage, clothing, language, ritual, literature, and cuisine. Agrarian societies were not conceptualised by their participants as being composed of jatis working within a unified ritual order of caste ranking or dharma; so there is no "caste system" described in the records of medieval agrarian communities. Creating a system of ranks to include all participants in any local society was actually beyond the scope of dharma. The adaptive and inclusive capacities of temple and caste ritual could not keep pace with the expansive diversity of agrarian social space. Local societies came to include not only Buddhists, Jains, Parsis, Muslims, Christians, and tribals; but also many groups who did puja but were excluded from temples, who did observe ritual ranks among themselves but were excluded from the territories controlled by dominant agrarian castes. Vast open space and towns sprouting up here and there allowed many groups to establish settlements with their own local rules of ranking and alliance, within which they lived and from which they engaged in social relations of great importance to their livelihoods with groups elsewhere who followed different rules of ranking. Settlements of people who followed mutually incompatible systems of ranking could not relate to one another on the basis of either ranking system; and transactions among such groups included barter, exchange, employment, patronage, alliance, conquest and subordination. Dharma could not define this kind of transactional territory, but patriarchs could always represent their own people in relations with others. Genealogies that begin with founding patriarchs produced legitimate authority for the head men of prominent families, community leaders, village elders, and family heads who were members of the panchayat, shalish, and village councils. Thousands of little social groups occupied and partitioned agrarian space and acquired names and genealogical histories as they interacted with one another. Some became part of a caste structure in agricultural settlements, but many stood apart. Many hunters, tree cultivators, herders, fisher folk, harvesters of the forest, merchants, artisans, miners, diggers of tanks and wells, tribal groups, and peasants formed their own little ethnic communities outside territories of dominant caste authority. Their headman patriarchs represented these groups in their relations with one another and received recognition and entitlements as the natural leaders of their communities.


The politics of patriarchy also propelled a medieval transition that came with the second millennium. From the eighth to the thirteenth century, patrilineal warrior clans with backgrounds in pastoral nomadism conquered farming communities all across south, central, and south-west Asia. In the vast territories of the warrior clans, competition by junior members and collateral branches propelled expansion and conflict. Marriage formed ranks and alliances among all the warrior lineages, and when warriors did marry into agrarian communities, they formed new ranks in which the sons of kings remained superior to the sons of the soil. In the early centuries of the second millennium, Ghaznavids, Ghurids, Khaljis, and Tughluqs expanded into lowland territories of military competition among Yadavas, Calukyas, Paramaras, Sisodias and others. These specialised, warrior groups had much in common. They unified their own forces by kinship and ritual practices that formed extensive family ties. They made alliances by marriage. They conquered farming groups to rule and protect them. They lived in fortress towns and formed an elite strata ranked above the kin networks of farmers. They formed hierarchical alliances among superior and inferior families, lineages, and clans. Their family ranks within military hierarchies allowed for strategic calculations in political hypergamy. They gave "subaltern" a distinctive meaning: subalterns among warrior clans were junior patriarchs in the ranks of lineages and dynasties. A son born to a ranking lineage member inherited a family position that provided a specific set of options for the ranking of his own family, as the son became a patriarch himself. Alliances gave subaltern families leverage in their struggles to maintain and to improve their position. Becoming a subordinate ruler raised the subordinate family's rank in relation to peers and competitors. Accumulating subordinate patriarchs (samantas) under one's own authority was the very definition of a king. Among the warrior clans, daughters married up -- to express the subordination of a patriarch and to seek upward mobility for the family -- and sons married down, to express the superiority of a patriarch by the stature of his allied subalterns. Patriarchal polygamy expanded the possibilities of subordinate alliance building, as women became hostages to fortune and some became the mothers of kings. In these settings, pardah and sati became auspicious expressions of female purity, piety, devotion, and heroism. Strength and sacrifice sustained one another. In the political institutions formed by competitive alliances among warrior patriarchs, subordination was a moment of power in which all alliances were built upon measurable inequalities of rank. Dominance rested upon extensive alliances with subalterns whose movement up in the ranks often meant challenging superiors in war. War and marriage, militarism and family ties, rank and alliance, negotiation and resistance -- all together formed patriarchal power in the warrior clans.


Dharma could not produce stable alliances among these groups, because warring medieval patriarchs invoked the names of many different gods in prayer and the rituals of war gave losers the option of moving out to look for other farmers and warriors to conquer. Moving out from Central Asia, Turk and Afghan warrior clans pushed Rajputs down the Ganga basin, into high mountain valleys, and into the central mountains; and they pushed Telugus up the Tungabhadra basin toward Vijayanagar and Telugu Nayakas into the Tamil country. In all the regions of later medieval warrior competition, marital and martial techniques of social ranking provided a cultural basis for new, sultanic regimes. Rathore Rajputs married daughters to sultans before Akbar's time, and almost all the great Rajput patriarchs would marry into the Mughal nobility, strengthening Mughal power and opening wide avenues for mobility and advancement for Rajput clans, and at the same time opening a status division between rising Rajput nobility and lesser Rajput and Thakur lineages. One Rathore princess married Prince Salim. He became Jahangir and she bore a son who became Shah Jahan, as lesser Rajputs lineages declined. In the eighteenth century, Qazi Muhammad Ala said that ordinary chiefs (rausa) "are now called zamindars."17


In late medieval times, from the thirteenth to the sixteenth century, new institutions of patriarchal ranking evolved; they formed a cultural context of sultanic secularism, which contained many kinds of canon, idioms of ritual, and moral systems under the umbrella of sultanic authority. Sultans rose above all patriarchs. Their regimes compiled the small agrarian territories of the early medieval period into regional forms with distinctively early modern characteristics. By the seventeenth century -- most dramatically in Maharashtra and Punjab -- alliances among imperial, military, fiscal, and agrarian patriarchs produced regional patriarchies. Great patriarchs like Shivaji and Guru Nanak formed the basis for regional identities of a new kind. Rights of first possession expanded metaphorically to include collective rights for all the people of the dominant castes assembled under a great patriarch to rule their homeland. This early formation of territorial and ethnic nationality emerged from the expansive powers of patriarchal authority produced under sultanic regimes as it absorbed the intense attachments to the land among dominant caste farm families and local conquest regimes.


When Rajputs and Mughals married, they tied together two traditions of patriarchal power which though expressed in different spiritual idioms had basic commonalities that formed a coherent logic of ranking, competition, and alliance. Mughal sultans became apical agents and icons of ranking for all patriarchs below. In the Mughal regime, mosque, temple, or church could mark communities of sentiment; sacred genealogies could be reckoned from Rome, Palestine, Arabia, or Aryavarta; because the Mughal institutions of patriarchal power -- within which patriarchs ranked one another and held patrimonial entitlements -- superseded and encompassed the ideology of dharma. No religion constrained a sultan's power to confer rank on subordinates. A sultan's status arose from rituals of conquest and entitlement whose authority went back to the days of the Gurjara-Pratiharas. Sultanic power reached its height under the Mughal, Safavid, and Ottoman dynasties, but its logic was not contained by Islam. Hindu practitioners included not only the Rajputs but also the Rayas of Vijayanagar, who were effectively sultans of the south. The English East India Company used sultanic authority for its own Christian imperialism. Thus the expansion of Muslim dynastic power in South Asia should not be conflated with the expansion of Islam: the Mughal imperial system set itself apart from all its predecessors by making the rituals and conditions of patriarchal entitlement more agnostic than ever before.


Social ranks defined by Mughal imperial titles inflected the idioms of social rank almost everywhere in South Asia, influencing group identity subtly and pervasively. From village and caste headmen to hill chiefs, to merchants and bankers and artisans; and from Rajputs, zamindars, nayakas, chaudhuris, ray chaudhuris, jagirdars, palaiyakkarar (Poligars), and rajas to all kinds of tax farmers; positions of leadership, authority, and political mediation in state institutions became focal points for social identity formation. Mughal entitlements and modes of patriarchal ranking entered family strategies of marriage alliance and thus influenced kin-group formation at many levels of society. In the Mughal system, patrimonial entitlements depended upon personal recognition by a superior patriarch under the authority of the sultan. In families, occupational groups, sectarian organisations, and caste and tribal societies, an officially recognised headman had to attain his status -- at a price -- in rituals of state. The darbar became a centre of transactions that defined agrarian territories, locally and regionally, in the ranks of all the patriarchs. Moving down the ranks, superiors granted honours, titles, and entitlements to those below. Moving up the ranks, inferiors paid tribute, service, and allegiance to those above. State revenues were collected in return for honours and titles conferred by state authorities; and the increasing value of these revenues accumulated at the higher ranks as they fed the evolution of early modern states. At the lowest echelons, peasant patriarchs paid for titles to land.


From 1500 onward, the agrarian utility and spread of money also increased along with supplies of precious metals and sultanic currencies. Money could buy a wider variety of entitlements to resources within disparate agrarian territories connected by systems of sultanic ranking that were open to participants of all sorts. Aggressive patriarchs bought and fought their way into positions of power in agrarian territories; they became military officers and revenue intermediaries entitled to collect local revenues from local head men. A diffusion of imperial titles and ranks facilitated a broader commercialisation of the agrarian economy that was also propelled by the military integration of ecologically diverse agricultural territories and by increasing state demand for cash payments, as we will see in the next chapter. Buying titles and official positions of rank became a basic patriarchal strategy. This further accelerated a broad shift away from dharma, caste, and Brahman ritualism as the most prominent means to secure assets in agrarian territories, though technologies of temple ritual also expanded their territorial reach under sultanic regimes. The pace of temple building and temple endowment accelerated steadily after 1500, as patriarchs with state entitlements and commercial assets sought additional resources through investments in temples.


Politically and socially, any group could be defined by its representation at court. Though temple and caste rituals extended their reach, darbari dramas had wider powers of incorporation and entitlement. Transactions that defined agrarian territory came increasingly to focus on key people who provided states with revenue. Mughal revenues fattened the nobility and fuelled the war machinery, and like other acts of submission, paying revenue constituted entitlements for patriarchs who paid to secure positions of power. Because revenue payments secured patrimonial property at every rank -- right down to the lowest levels of the peasantry -- it is understandable that a myth emerged among Europeans that the sultan owned all the land in India: who would counter this claim when all patriarchs held their property rights by submission to the Emperor? In this political culture, acts of resistance and rebellion were also acts of negotiation and strategic positioning. Patriarchs faced opposition all around -- in the land, unruly nature; in the home, unruly women; in the fields, unruly workers; in the villages, unruly peasants; in the forests, unruly tribes; in the provinces, unruly zamindars and rajas -- and negotiations among patriarchs always had to take into account resistance from below and demands from above. Patriarchal expectations for obedience and loyalty often met frustration. Many new, assertive identities formed around rebellious patriarchs who like Shivaji had official entitlements to represent "their people" in transactions with higher authorities.


The Mughal regime brought more kinship groups under one system of ranking and military alliance than any before. All its constituent groups became designated by terminologies that effectively formed an ethnic typology. Ethnic identities, based on combinations of language, religion, and region, emerged dramatically among Rajputs, Marathas, and Sikhs, but also in many other places at lower registers. Competitive alliance formation raised the most powerful agrarian patriarchs up into the status of regional leaders. Shivaji inherited a jagir that his father obtained under Ahmadnagar and he continued the project of constructing a multi-jati Maratha warrior-farming elite by acquiring titles from other sultans in the Deccan. Over several generations, in a long process of competitive alliance building, conquest, and institutional formation, Marathas built a state that became deeply involved in the enforcement of family ranking and in regulating female behaviour, as warrior patriarchs set about defining Maratha territory and identity. The subsequent preoccupation of Maratha hagiographers with Shivaji as the ideal ruler not only reflects the capacity Muslim states to nurture Hindu leadership, but more importantly, it represents the creation of a semi-deified patriarchal icon around which new collective identities were formed, combining ethnicity, language, and religion.


The long term interaction of family and statecraft produced geographical patterns in regional styles of kinship within agrarian territories. Irawati Karve once argued that more extensive and intensive kinship territories typified the northern plains and southern peninsula, respectively, with Maharashtra being a bridge between the two,18 but Bina Agarwal's more detailed analysis of kinship practices and women's land rights reveals three broad zones of kinship in South Asia. In each zone, kin groups form distinctive types of territory, and regions are characterised by the prevalence of the kinship strategies pursued by prominent land owning groups, most importantly, dominant castes. The position of women is a critical feature of these kinship and territorial regimes.



(1) The north-east high mountains, the southern peninsula, Sri Lanka, and Nepal: "In all of these, women marry either in their natal villages or in nearby ones, and close-kin marriages are preferred. There is no adherence to purdah, and the overall control of female sexuality is less than in other parts of the subcontinent. Women's labour force participation varies between medium and very high."

(2) The western plains and northern basins: In Punjab, Haryana, Uttar Pradesh, and Rajasthan, "village endogamy is typically forbidden, marriages are often at some distance from the natal village (especially among the upper-caste land owning communities), close-kin marriages are usually taboo, purdah is practised, control over female sexuality is strict, and women's labour force participation rates are low." Though in Pakistan and Bangladesh, "village endogamy and close-kin marriages are permitted, and women's inheritance rights are endorsed by Islam .... female seclusion practices negate these advantages to a significant degree...."


(3) The central mountains, Maharashtra, and West Bengal: "Village endogamy is not common but neither is it usually forbidden, and women in many communities do marry within the village or nearby villages. Some communities do allow close-kin marriages. Purdah is practised in some communities and not others." 19


Agarwal stresses that tribal kinship patterns, most prevalent today in the mountains, differ significantly from those in the agrarian lowlands, especially in the amount sexual activity allowed for women outside marriage. Historically, groups called "tribes" are by definition those which have been relatively isolated from lowland agrarian states in modern times and therefore distant from the enforcers of dharma. Tribal groups interact with caste societies and economies but they have also been kept apart, especially in the tropical high mountains in the north-east. Agarwal notes that the prevalence of tribal communities in the central mountains and their territorial admixture with non-tribal communities creates the mixed character of her third kinship zone.


The broad division portrayed by Karve and Agarwal is between extensive and intensive strategies of kinship alliance, which appear to predominate according to the respective influence of warrior colonisation and territorial dharma in pre-modern centuries. There is also an overlap between forms of kinship territoriality and the prominence of irrigated agriculture, rice cultivation, medieval inscriptions, and pastoral nomadism. In general, when we move from low lying riverine tracts of the early medieval gentry, where older inscriptions cluster, into drier areas dominated by warrior-farmers, we see a transition from more intensive to more extensive kinship practices. In this same transition, we see a shift in gendered substance of patriarchal power. Matrilineal descent was prevalent only in Agarwal's first zone, which also contains territories of intensive kinship where women live within a small circle of kin for their whole lives. By contrast, in extensive kinship regimes, women pass between distant kin groups as icons of family honour and agents in marital alliances. Maharashtra contains both kinds of kinship territory, and also transitions not only between tribal hills and caste lowlands but between coastal regions of more intensive medieval wet-farming and interior regions of more extensive dry-farming. So Maharashtra is not so much a transition zone between cultures of the north and south, described by Karve, as a mix of practices that characterise different types of agrarian territory. A separation of irrigated lowlands from dry uplands also divides coastal Andhra from its interior and the Kaveri basin in southern Karnataka from the Deccan; and more intensive kinship patterns usually pertain in the wetter regions. In general, the distribution of more intensive forms of kinship even today coincides with that of more intensive farming in pre-modern centuries. Extensive regimes, like warrior power, spread widely over time. Extensive kinship strategies concentrate today in territories which have more pastoral nomadism and warrior colonisation in their agrarian history. In Maharashtra, Bihar, Bengal, Bangladesh, and elsewhere, extensive warrior patterns of kinship were imposed upon more intensive local strategies, creating elite alliances and models of status which diffused downward like the powers of dharma. Extensive kinship patterns would have helped to extend the agricultural frontier in the east, and more intensive kinship forms perhaps developed in old pastoral areas that became characterised in modern times by more intensive irrigated farming, especially in Punjab. A mixing of kinship forms occurred everywhere with the rampant migratory resettlements of the early modern period. In the Tamil country, Telugu warriors settled tracts between river valleys and some have retained extensive kinship strategies and even observed pardah until recently.


Despite all the imperfections in the fit between old farming regimes and kinship, we can see that intensive kinship alliance-building is a good strategy for protecting family property in local communities and territories. Family alliances that formed the local gentry also produced the funds and controlled the labour which built up early medieval irrigation and paddy cultivation. Patriarchs sought to control contiguous territories for the expansion of succeeding generations. Marriages formed dense links among dominant families who became related to one another like their paddy fields, as sustenance flowed from one family to another and from one generation to the next. Families partitioned social space into contiguous kinship territories, which became more diverse by the inclusion of new groups into jatis and by the fissioning of lineages, but retained an intricately kin flavour. In riverine lowlands, kinship stressed local alliances among families and formed intricately graded ranks within gentry strata. Families maintained a genealogical sense of descent from medieval kings, but domestic patriarchy concentrated on markers of status within local communities. The marriage of sons and daughters was normatively contrived within finely graded social strata, within close proximity to the natal village, and within an existing nexus of family ties. Agricultural communities and regions were organised around webs of intensive, intersecting family alliances. On the Tamil coast, it would not be uncommon for people to be related to one another in several ways at once, as a result of cross-cousin marriages over generations. The language of family permeated all the institutions of agrarian patrimony, as the meanings of pangu, pangali, kulam, nadu, and nattar attest. Institutionalised among dominant castes, the intensive pattern of kinship and its idioms of village share holding also became typical among agricultural workers and other groups in irrigated regions of the medieval era. In this setting, appropriate female decorum is a community as well as a family concern. Husband and wife are most often reckoned as kin from birth, sharing the same agrarian home-space. A wife's devotion to her husband does not need to conflict with loyalty to her father; and though she leaves her father's house at marriage, she effectively never leaves home. Migration brought similar types of family networks into being in nested agglomerations of family ties that came to characterise the paddy-growing lowlands.

In the dry territories of nomadism and in domains of conquest colonisation, by contrast, and particularly among warrior colonisers like the Rajputs and Jats, a woman's transition from daughter to wife came to mean moving into the household of a stranger whose superiority to her father was dramatised in the marriage itself. Patriarchal strategies of marriage alliance designed for upward mobility put women in a difficult, intermediary position, as marriage helped to extend lineage power out over territory in a pattern like that of the banyan tree. Expressions of family power focused on the wife as the icon of her family's honour and rank. Multiple marriages expanded the power of a great patriarch, his wives being ranked as a representative of their fathers. A woman's devotion to her in-laws always conflicted in principle with loyalty to her parents and siblings. Personal, intimate, ritualised expressions of devotion to her husband as opposed to her father were built into the disciplinary activities particularly of her mother-in-law, who had survived this same transition. But the status of wife and mother in a superior family could also open up new opportunities for her natal kin and their offspring; so that serving her husband would most likely be her father's most fervent desire, because pleasing her husband would be the best way to improve her natal family's prospects. In these settings, pardah and sati became auspicious expressions of female purity, sacrificial devotion, sacred heroism, and divine power. Extreme controls over female sexuality enhanced family honour in a culture of heroic sacrifice and harsh discipline. Sati became divine at landmarks of heroism which marked warrior territory. Forts and palaces enshrine the valour of great men. Perched high above Ajmer on a rocky mountain ledge, Prithvi Raj Chauhan's fort has today become a icon of militant Hindu nationalism. At Mandu, it is said, hundreds of palace women became sati as the sultan marched his troops to death in battle. At Mandore, in 1459, Rao Jodha "took an extreme step to ensure that the new site proved auspicious" by burying alive one Rajiya Bambi in the foundations of the new fort, promising that "his family and descendants would be looked after by the Rathores."20