Framing

In this article, drawing inspiration from James C. Scott's work on state avoidance, and from Brubaker's reconceptualization of groups, ‘groupness’, and ‘groupism’, I argue that the constitution of the communities in the Nilgiri hills needs to be examined as a process of developing groupism (otherwise known as ethnogenesis) over time, in relation to each other and to the plains states that surrounded them.Footnote ⁶ To briefly define Brubaker's concept of groupism (to which I return later), it is the tendency of ethnic groups to define themselves emically as fixed and bounded entities, and the tendency of scholars and outsiders to treat those boundaries and groups as ontologically real. He further argues that scholars should view the reification of groups as the result of historical processes and not as timeless, even though this timelessness is often part of the emic narrative.Footnote ⁷

Scott's arguments about the nature of hill peoples and hill societies in Southeast Asia provide an important theorization of the relationships between hill peoples and lowland agrarian states. He argues that:

Indeed, one of the largest communities of the Nilgiris, the Badaga, have an oral tradition that suggests that they arrived in the hills as refugees from central Karnataka around the time of the fall of the Vijayanagara empire in the sixteenth century.Footnote ⁹ And, though I introduce, by way of example, this oral history of Badaga self-avowed state avoidance, various lines of evidence discussed later in the article suggest that this is not the whole story of the creation of Badaga groupness (the process of the formation of Badaga as a self-defined group).

As I will discuss later, active state avoidance in no way necessitates the total isolation of hill communities. If and when the goal is to avoid state power, domination, exploitation, or taxation, hill communities nonetheless could and did continue to interact, trade, and communicate with the communities of the regions surrounding the hills. State avoidance does not at all imply total avoidance of all contact with surrounding regions. Even though I argue that Scott's theory of the state-averse has utility in the context of the Nilgiri hills, not all aspects of Scott's analysis are perfectly or directly applicable to the South Indian case.Footnote ¹⁰

Notes on Subaltern Studies

As this article explicitly frames itself as subaltern, before continuing, I must define what I mean by the term. I use the term ‘subaltern’ to emphasize that the focus of this enquiry is on peoples who have no written history of their own, and whose history has been written thus far in a framework that was profoundly biased.Footnote ¹¹ By focusing on the subaltern subject, especially in colonial history, the Subaltern Studies movement has paved the way for a great many new interpretations that were not possible before. However, I am not concerned only with the history of the Nilgiris peoples as subjects and subalterns in relation to European colonial rulers. They were in many ways subaltern to, if not always effectively subject to, many South Asian states and empires before Europeans arrived.

The subaltern is mutually constituted ‘in a binary relationship of which the other [side] is dominance’.Footnote ¹² Just as Ramchandra Guha and colleagues have argued that subalternity cannot be conceived without its counterpart—state domination, I argue as an extension of Scott that, in the study of hill peoples, the stateless, and state-averse, we must also remain aware of the mutual co-constitution of such refugee populations and the state.Footnote ¹³ As such, it becomes clear that the subjects of this investigation, the inhabitants of the Nilgiri hills, whether permanent or transient, were more than subaltern; they adopted a position of alterity, of otherness, outside the normative structures of hierarchy in state societies, though still in a social position for which the counter-position is state domination.

Subalterns, in the classic Gramscian sense, are the peasants and the lowest members the social hierarchy that constitutes an (agrarian) state as a both social and political entity.Footnote ¹⁴ Here I argue that there is a fruitful line of inquiry on the relationship between the subalterns (in the Gramscian sense) and the state-averse (sensu Scott). It may be said that one is more ‘sub-’ and the other is more ‘-altern’, the latter perhaps having been willing to take up alterity in the form of small-scale social organization as an alternative to subordination. While I am broadly framing this work as a form of subaltern inquiry, it may be more fitting to come up with some other term for those people who, we might suspect, disliked the position of the subaltern peasant in the social hierarchy of the state enough to leave behind that social position for a life in the hills.

Much of the early literature on the Nilgiri hills demonstrates an assumption about the isolation of hill peoples (similar to the pattern observed by Scott), which was derived in part from the idea of their having been ‘left behind’ in the ‘march of progress’. It is perhaps obvious, but nevertheless worth pointing out, that it is faulty logic to assume that because the hill peoples are so divergent from the valleys and plains in terms of language, cultural practices, subsistence, and social organization that they must have become so only as the result of prolonged, and more or less absolute, isolation.

Strongly divergent languages and cultural practices need not be the result of literal isolation, but rather can result from the strategic positioning of people who are well aware of the social structures and economic systems of the lower elevations, but who are or were actively resisting participation in those structures and systems.Footnote ¹⁵

Landscapes

The Nilgiri hills are located at the intersection of the boundaries between the contemporary states of Kerala, Karnataka, and Tamil Nadu. These modern state boundaries have been drawn on the basis of linguistic lines, as well as on what was believed to be the historic boundaries of the various kingdoms that preceded British conquest.

The hills are distinctive in their topography, vegetation, temperature, and rainfall. They are marked by steep and forested slopes, on which grow some important non-timber forest products, especially pepper (Piper nigrum).Footnote ¹⁶ The distribution of forest, grassland, and cultivated land on the upper plateau has changed significantly over the last two millennia, and continues to change.Footnote ¹⁷

The hills are themselves not internally environmentally homogeneous, though they are distinct from the very lush plains of Kerala to the west, the somewhat drier plains of the Erode and Coimbatore Districts of Tamil Nadu to the east, and the Chamrajnagar and Mysore Districts of Karnataka and Wayanad District of Kerala to the north. The slopes of the hills are steep and for the most part densely forested, except where there are sheer rock faces. These steep slopes rise to a plateau that is undulating, which was, as described by early British visitors and surveyors, mostly grassland, with pockets of forest, and some cultivated land.Footnote ¹⁸ The western part of the plateau is more forested and has less dense settlement, now and in the nineteenth century, than the eastern part of the plateau. The extent of cultivated land at the time of the first British visitors is debated, as they often described a significantly less cultivated landscape than they actually represented on maps. For example, there is an 1822 map, based on a survey by Benjamin Swain Ward, that appears to show extensive cultivated land, but there is a contrasting written account by Ward himself that downplays the extent and importance of agriculture on the Nilgiri plateau.Footnote ¹⁹

The physiographic differences of South India have also been marked culturally and in the Sangam literature, in the form of the five ‘tinais’, or landscapes, of Tamil Sangam literature. The hills are called ‘kurinji’ after the flower that blooms once every 12 years, and are associated with hill tribes and forest products like honey; the ‘pālai’, or wasteland, is dry and mostly uninhabited; ‘mullai’ is both jungle or forest and pasture, associated with pastoralists; ‘marutam’ is the lowland plains and agricultural fields, associated with rice and rural life; ‘neytal’ is the seashore, associated with fisherfolk and salt production.Footnote ²⁰ This cultural division of landscape is idealized, but also reflects the reality of numerous ecotones and ecozones, with distinct flora and fauna, and to some extent inhabited and/or utilized by different communities of people with different subsistence and economic practices. The landscape is not characterized by a simple dichotomy of hills and plains.

Even in the Sangam literature, it is clear that the hills were not part of the lowland kingdoms; the extent of territorial control of those early states is spotty across numerous types of landscapes. Even so, the Sangam literature as a corpus does seem to draw a political distinction between hills and everything else. As observed by S. Darsana, the Tamil title that translates as ‘chieftain’ was used to refer to political leaders of the hills and ‘king’ to their counterparts in the plains.Footnote ²¹ Such a semantic distinction suggests that the authors of the Sangam poems perceived a difference in the political organizations of the hills and various landscapes of lower elevation. The Sangam literature, and therefore these perceptions and representations of the hills and the hill peoples (by the Sangam bard-poets of the plains), could be as old as the third century bce but most likely dates to around the first to third centuries ce.Footnote ²²

History

To explore the history of the hill peoples, both in contact with and resistance against the lowland states and empires, requires a subaltern methodology. Some information can be gleaned from limited texts and epigraphic sources, but the fugitive aspects of hill peoples’ lives make such a project even more challenging than exploring the subaltern histories of peasants who lived in much more intimate contact with their rulers.Footnote ²³

The relationship of Nilgiris people to written history is not necessarily the same as the classic subaltern subject. They are not just another group of people with no written records. These are people, many of whom—I propose—purposefully evaded written records, and thereby any form of legibility to the state.Footnote ²⁴

The earliest and most comprehensive historical account of the Nilgiri hills, and their relations with the lowland states around them, is found in Francis's The Nilgiris Gazetteer. Ironically, he begins his historical account with the passage used as an epigraph to this article:

In prefacing his history of the Nilgiris by saying that it may ‘almost’ be said to have no history, Francis reinforces the impression that, somehow, in spite of all the forts built, and attempts at conquest and rule, the Nilgiris remained isolated from outside influence. But, though the repetition of the trope of their isolation abounds, Francis's actual accounts of the historical and epigraphic data demonstrate the repeated attempts to conquer and incorporate the Nilgiris into state territories beginning in the tenth century.

The list of dynasties and would-be rulers of the Nilgiris is rather long over the period from 900 ce to 1799, when the British took possession of the territory. Those dynasties that at least laid claim to (parts of) the Nilgiris territory include: the Ganga, Chola, Hoysala, Dannayaka, Vijayanagara, Wodeyar/Udaiyar (Rajas of Ummatur), and the Sultans of Mysore, Hyder Ali, and Tipu Sultan. Forts in the hills, especially with clear views towards the plains and approaches, were constructed possibly as early as the Hoysala dynasty in the fourteenth century, though the largest and most solid evidence for fort construction in and on the hills comes fairly late with the Rajas of Ummatur in the seventeenth and eighteenth centuries.Footnote ²⁶ All together, there are the remnants of three dozen forts in and on the Nilgiris, especially at the main approaches to the hills from the plains.Footnote ²⁷ Given that the region as a whole is a fairly small 500 square kilometres, that is a lot of forts, especially for a place so supposedly isolated and left out of the history of conquests by Francis's ‘rapacious rulers’.Footnote ²⁸

Metz, an early German missionary and ethnographer, collected an oral text that recounts a memory of a Toda king and queen, and how what might be interpreted as a Toda state was conquered by one of the Rajas of Umattur. He says:

The mention of the incorporation of the Wayanad (formerly known as the Wainad), a somewhat lower-elevation plateau attached to the north-western part of the Nilgiris, into the territories of the Ganga kingdom of Mysore in 903 ce seems to be the earliest written evidence of attempts to control the Nilgiri territories, though the claim seems not to include the upper elevations.

Toda oral history, as collected by M. B. Emeneau (and W. H. R. Rivers, though Emeneau's accounts are clearer), recounts several occasions when armies invaded from the plains, though these lack any specific date. These accounts give the overall impression of ongoing antagonism and dislike of the plains people. One particular story recounts why men from particular septs cannot become priests in the ti· (sacred) dairy of the Noṣ clan, but incidentally recalls an army or raiding party from the plains who threatened to steal the sacred buffaloes, and simultaneously were extorting, or attempting to extort, cash money from the Todas:

The Toda and Kota oral texts collected by Emeneau also provide numerous examples of interaction with communities from outside the Nilgiris, though with little specific historical context or dating. Many of these were not positive interactions. There is one oral text that instructs the listener not to make friends with Tamilians, or form economic partnerships with Kurumbas.Footnote ³¹ There are instances of raids carried out by an army of men, attempts to steal buffaloes by just a few, betrayal of friendship, cheating, and so on. Several stories recall armies or raiding parties coming to steal buffaloes, which is not surprising. However, these stories have a parallel in which the Tamilians come to steal Badagas:

While most stories involving people from the plains portray them as cheating or stealing from the Todas, there is at least one story in which a Toda woman cheated a Chetti out of two bells. She promised him 30 buffalo heifers, which he took and started to depart, but she asked a priest to pray for the return of the buffaloes, and the buffaloes returned, leaving the Chetti with neither his bells nor his buffaloes.Footnote ³³ Overall, these stories suggest a history of antagonism, especially with the people of the plains who came up to the hills. The only mention of someone coming to collect a tax is the one in which the Tamilians were taking away the Badaga people, apparently because they owed taxes, perhaps to take them away and make them labour to pay off their taxes.

While tax collectors do not feature prominently in these stories, it seems stories of conquests are also mostly absent. The only example is the story recounted by Metz, suggesting that the Toda once had a king and queen whose throne was usurped by a chieftain or king from Ummattur, in present-day southern Karnataka. On the other hand, there are several stories and cases of Todas going down from the hills to interact with outsiders, and those stories are generally more positive. One example is of a Toda man who fled the hills after a family feud with his brother and his brother's wife. He went down to the Wayanad and asked for help. He was given a house, a wife, and buffaloes (on loan) and, with this assistance, he became rich and prosperous living in the Wayanad. The story essentially tells how it came to be that there was a Toda village in the Wayanad.Footnote ³⁴

It seems that the Toda perspective on relations with the world outside the hills was that those who came up to the hills usually meant trouble, but business could be conducted, help obtained, and so on, if a Toda person went down to the plains. The Toda, through these oral texts, tell of their aversion to unsolicited interference in their hill-top affairs. On the other hand, they were more than happy to obtain goods and interact with the people of the Wayanad and Tamils, on their own terms.

Archaeology: study of prehistory and para-historyFootnote ³⁵

Archaeological evidence, much of it also collected in the nineteenth century as a part of antiquarian interests and British colonial projects, presents an interesting and self-contradictory picture. On the one hand, the Nilgiris have long been considered as fully integrated into the regional South Indian ‘culture area’ from the Neolithic, or at least the Iron Age onwards.Footnote ³⁶ This is because the region is (or was at least through the nineteenth century) full of megalithic stone monuments that appeared, in most respects, to belong to the widespread tradition of megalithic construction across South India that began around 1200 bce. However, the only systematic excavation of a megalith on the Nilgiri hills produced a radiocarbon date of 910 ± 90 ce (uncalibrated), and the style of pottery and material culture found in the megaliths at Pykara excavated by Hockings is the same as that found in all of the many megaliths dug up by nineteenth-century antiquarians.Footnote ³⁷ Cross-dating the unsystematically excavated material, in particular the ceramics and terracotta figurines, with this single dated megalith and archaeological material outside the Nilgiris suggests that the entire phenomenon of megalithic burial in the Nilgiris likely belongs entirely to the first millennium ce and later, and possibly the first few centuries bce, but not the Iron Age (1200–400 bce), as previously believed.

Occupation in the hills probably does go back at least as far as what is known in the rest of South India as the Neolithic (circa 3000 bce), though very few data exist that can be shown to relate directly to this early settlement. Zagarell uses rock-art styles as a primary line of evidence.Footnote ³⁸ Morrison has argued that the human use of the hills in this early period was more likely temporary and transient, rather than permanent.Footnote ³⁹ Tamil Sangam literature provides another source, perhaps pertaining to the period from about 300 bce to 300 ce.Footnote ⁴⁰ This literature mentions the hills, or rather the slopes of the hills, as a source for spices, especially pepper that was brought down and loaded on to yavana (foreign) ships, bound ultimately for the Greco-Roman Mediterranean.Footnote ⁴¹

What at first appeared to the British antiquarians, archaeologists, and anthropologists such as Harkness, Breeks, and Rivers as a deep continuity between the megalithic burials associated with extreme antiquity and the living communities is more likely a product of only the most recent 1000 or so years (the medieval period and later) of human activity in the region.Footnote ⁴² In other words, this seemingly ‘primitive’ archaeological culture does not belong to the hoary past, but instead appears to be contemporary with the rise of the Chola empire in the plains of Tamil Nadu to the south and east of the Nilgiri hills.

The archaeology of the region is known mostly through unsystematic excavation of megalithic constructions. Most of these excavations took place in the nineteenth century, often by men who were in the British military or Civil Service, who had an avocational interest in antiquities, who were in some cases also members of the Royal Asiatic Society or other scholarly organizations. Francis in 1908 wrote:

Systematic archaeological enquiry did not get under way in South Asia until the early to mid-twentieth century.

Ethnography

There are, or at least there were as of the early nineteenth century, five communities in the Nilgiris identified as ‘hill tribes’: Toda, Kota, Badaga, Irula, and Kurumba (also known as Nayaka). Their classification as ‘tribes’ even in the classical anthropological sense is problematic, but this article is not the place in which to dwell on this issue.Footnote ⁴⁴

Here is the anthropological narrative (in a problematically ambiguous ‘ethnographic present’Footnote ⁴⁵ ), as originated by early anthropologists and colonial observers, most prominently W. H. R. Rivers, and later expanded and systematized by William Noble and others.Footnote ⁴⁶ David G. Mandelbaum sums it up well:

Mandelbaum's statements are representative of the view that anthropology held for a long time, namely that Nilgiris communities co-exist(ed) peacefully, and with a high degree of economic interdependence, while simultaneously maintaining ethnic and linguistic boundaries between them.

It is interesting that, while these ethnographies strongly emphasize the economic interdependence of the Nilgiris communities with each other, they rarely mention economic relations or interdependence with the plains or regions around the Nilgiris massif. This may well be a result of the bias regarding the more or less total isolation of the hills. Rivers, for example, recounts a Toda man walking down to a market in the Wayanad to buy rice, but suggests that such contacts are of much more recent origin, and that the ‘true’ or ‘original’ state of being was isolation.Footnote ⁴⁸

It does appear that goods were exchanged between regions, including the top of the Nilgiris plateau, the slopes, the Wayanad, and the plains, though the mechanisms probably varied in time and space. The slopes of the hills are known to have exchanged pepper, honey, and other non-timber forest products with the plains societies, while the Toda may have exchanged butter and ghee.Footnote ⁴⁹ The Nilgiris communities (probably all of them) were consumers of rice, and perhaps also cotton (whether raw or spun or finished textile is not clear). However, there was not a pressing or urgent need for food grains like rice, since there were other grains and products that were grown in the hills—amaranth and millets. According to Hockings: ‘Large groups of Badagas and Kotas used to go together to the market town of Gundlupet in southern Mysore, or to Karamadai, near Coimbatore. They carried hay, honey, opium, and clarified butter with them, and bartered these for salt, crude sugar, rice, coconut, tobacco, and trinkets.’Footnote ⁵⁰ While there is clear evidence, therefore, for continuous and ongoing contact and exchange of goods between people of the hills and the plains, it does not suggest a real dependence on goods from the plains to survive—with salt perhaps being a necessity and therefore an exception.

The Toda, for anthropologists, are primary in many ways. They claim primacy in the hills as the region's first inhabitants. They were also the most fascinating and enthralling for early European writers. They seem(ed) the most strongly divergent in cultural practices, language, and identity. Second, at least in order of their supposed presence on the hills, were the Kota people—part-time agriculturalists and part-time craft producers, their language is the most closely related to Toda, though still significantly distinct. Third are the Badaga, meaning ‘northerner’—a group whose oral history suggested to many authors that their origins did in fact lie to the north, and that they had migrated to the hills in the sixteenth century. Various forms of linguistic dating and historical convergences seem(ed) to suggest that this group migrated around the fall of the Vijayanagara polity, perhaps fleeing the city of Vijayanagara, which was sacked in 1565.Footnote ⁵¹ The other two groups, Irula and Kurumba, are occupants of the steep slopes of the hills and primarily engaged in hunting, gathering forest products such as honey, and perhaps some occasionally shifting horticulture. Other ‘castes’ are sometimes mentioned, identifiable as more recent and obvious transplants from the plains, though, as a result of their recent settlement in the hills, the British tended to treat them as not really ‘counting’ as part of hills society.Footnote ⁵²

One of the less agreed-upon features of Nilgiri society is whether or not, or to what extent, these communities are or were organized hierarchically in relation to one another. Some studies emphasize a hierarchy in which the Toda are the highest, based in part on concepts of purity and pollution, while others emphasize that the Badagas were given authority to arbitrate disputes between Todas and Badagas, and others. While this has been called a ‘caste-like’ system, I agree with Zagarell, who argued that the most reasonable interpretation is heterarchy. He suggests that the communities were ranked differently in relation to one another, depending on the social/economic/political context, and that additionally these hierarchies may have changed over time.Footnote ⁵³ Even as we might suggest that Toda and Badaga were neither fully equal to one another, nor did one ever have complete power or domination over the other, it is clear that both viewed Kota, Kurumba, and Irula as lower, and potentially polluting. On the other hand, they were both afraid of Kurumba sorcery, and respected Kurumba magical and healing powers.

This narrative, and in many ways this article, is attempting to side-step a debate on the construction of the categories of both caste and tribe, and the distinction between the two, hence my preference for the term ‘communities’ over both the terms ‘caste’ and ‘tribe’, and, for reasons stated above, also avoiding the term ‘ethnic group’. There is not the space here for a full treatment of the myriad issues of classification and nomenclature. Dirks dealt with the historically contingent and constructed nature of the concept of caste and of a caste ‘system’. Béteille also deals with the problematic distinction between ‘caste’ and ‘tribe’, and the history of these terms in Indian sociology and anthropology.Footnote ⁵⁴

Ethnographic work in the last 50 years has done much to add depth and complexity to these questions, and to provide valuable data. Hockings has written extensively on the Badaga community, Walker on the contemporary Toda, and Bird-David and Noble on all five communities.Footnote ⁵⁵ And, though I disagree with Hockings about some interpretations of his data, his publications have been invaluable to my discussions in this article. On the other hand, much of their work, and of most historians, has focused on recent changes and on confronting colonialism and ‘modernity’. I argue that our historical understanding of the colonial period, and later, must be informed by an understanding of the histories of interaction, avoidance, domination, and attempted domination that preceded the arrival of Europeans. This article, while focusing on the ‘pre-modern’, is therefore addressed in large part to my colleagues of the ‘modern’.

Reinterpreting social history: isolationism and groupism as process and event

Analytically in anthropology, these communities have been treated as real, observable entities. This is the problem identified by Brubaker as groupism:

Groups appear, at many levels, to be very real things. Brubaker asks:

Brubaker warns specifically that ‘We should not uncritically adopt categories of ethnopolitical practice as our categories of social analysis’.Footnote ⁵⁸

The production of groups as defined entities is both a social and a political process. Brubaker emphasizes the process of self-identification, and the relations between outsiders such as anthropologists and sociologists and the ‘group’ as defined according to those emic definitions. The process of groupness can also be the result of definitions, criteria, and boundaries imposed and constructed by outsiders with goals less noble than academic study. Politics, and relations of power, within and between states (as in imperialism) are also key to the production of groupness through the sometimes asymmetrical, yet mutual, co-construction of group identity as produced in relations of domination, in which processes of defining a group come from both inside and outside. Though he did not conceive of it in these terms, Dirks described essentially this process with reference to the caste system, and castes as groups, as a product of relations between colonizer and colonized.Footnote ⁵⁹

Here, I present two histories of groupness as process and event. I examine the processes and events of groupness of the Toda and Badaga, since they are in many ways opposites. I begin with the Toda.

The groupness, that is to say the historical reality of an internally homogenous, externally bounded group, is much more manifest for the Toda than the Badaga and, though neither is entirely internally homogeneous, the Toda are more so. Toda groupness, and the ethnogenesis of the Toda as a community of shared practices, a linguistic community, and an endogamous community, seems, from all available evidence, to have occurred very early in history.

The linguistic evidence points to a long period of isolation, not necessarily from other members of the Nilgiris branch of the South Dravidian language family, but from speakers of the plains languages. To the extent that the Toda have been reified as a group and studied by anthropologists, sociologists, linguists, etc., this reification seems to correspond fairly well to the historical reality. That is to say that it appears that Toda identity and groupness coalesced more than 1000 years ago. Evidence is obviously scanty for such a claim, but there is mention of the ethnonym ‘Toda’ as one amongst a list of other groups in an account of an 1117 ce conquest of the region by the Hoysalas.Footnote ⁶⁰ Later, in 1604 ce, the Portuguese Jesuit missionary Giacomo Finicio visited the ‘Todamala’ (Toda mountain) and describes his interactions with Toda and Badaga people. He records that he communicated with them through an interpreter who spoke Canara (Kannada), ‘the language of the Badegas [sic], neighbours of the Todares [sic]’.Footnote ⁶¹ From his account, the traditional half-barrel-shaped house style and pastoral subsistence pattern seem also to have been features of a group called Toda.

Genetic evidence also supports a very early divergence of the Toda as an endogamous pastoralist community distinct from other non-pastoralist communities around them.Footnote ⁶² While many studies of the genetics of ‘tribal’ peoples in India have been done, this study includes Toda and other pastoralist communities within a larger sample of South Asians representative of the wider population. Romero and colleagues investigate the selection for, and distribution of, alleles that result in lactase persistence, or the ability to continue to consume milk and other dairy products into adulthood.Footnote ⁶³ They conclude that Indian populations, north and south, Indo-European, and Dravidian speakers alike, share the same (-13910*T) allele that appears to have originated in Europe in the Neolithic, associated with the development of dairying as a primary subsistence strategy. However, the frequencies of this allele among different social/cultural divisions is variable and there are additional mutations that appear to have been derived since the arrival of this genetic input from Europe, several of which are concentrated in the Toda population.Footnote ⁶⁴ The genetic differences of the Toda when compared with nearby populations and the fact that some of those differences specifically relate to a pastoralist subsistence economy tend towards the interpretation that the Toda were distinctive and maintained that distinctiveness without much gene flow over a long period.

In fact, the linguistic evidence, in particular the phonological divergence, suggests that Toda language was spoken only with in-group members and that there were rarely any non-native learners of the language.Footnote ⁶⁵ Instead, Toda used Badaga language as a medium of communication with non-Toda people. According to Rivers, ‘The Todas are a bilingual people speaking Badaga in their intercourse with other races and keeping Toda for themselves’.Footnote ⁶⁶

Many early observers made arguments about the distinctiveness of Toda phonology as evidence of the isolation of the community. I argue that we should consider the historical processes of Toda groupness, and to some extent intentional isolationism, as either directly or indirectly resulting in the distinctive features of the language that led scholars to initially interpret the Toda's language as evidence of their profound (and passive) isolation. One of the renowned linguists of Dravidian languages, Murray B. Emeneau, argued that:

While I disagree that, specifically, hierarchical difference is the only way to explain this, and prefer the heterarchical argument made above, I think Emeneau is right about the social production of difference.Footnote ⁶⁸

Indeed, most early European accounts of interaction with the Toda are described as taking place through the medium of Badaga language. Most of the place names given for Toda places are actually Badaga names for those places. This may or may not have been a conscious strategy to keep the British (and others) at arm's length. It seems likely, however, that the choice to use Toda language for communication with group members resulted from a Toda perception of their own groupness. The use of Badaga as a lingua franca by multiple communities in the hills may also have facilitated an isolationist attitude in the use of Toda language, in that it provided Toda speakers with an alternate medium with which to communicate with outsiders. This may be comparable to (if perhaps a less extreme version of) the ways in which the Roma communities in Europe went to great lengths to maintain Romani as a language exclusive to those born into the community, even going so far as to intentionally mislead outsiders who enquired about the language.Footnote ⁶⁹

Based on these diverse pieces of information, it seems as though Toda groupness was certainly formed by the early twelfth century ce, perhaps significantly earlier, and that the cohesiveness and boundedness of this identity was something that was likely socially enforced through powerful prescriptions regarding language and endogamy, among other things.

In contrast to the restricted use of Toda language, it appears that Badaga language was used as a lingua franca during the early colonial period, and likely earlier. Christiane Pilot-Raichoor suggests that the origin of what is known now as Badaga language is a conglomeration of vocabulary and grammatical structures from many Dravidian languages, building on a base that is most similar to Alu Kurumba language (the language of one of the communities inhabiting the steeper slopes).Footnote ⁷⁰ In other words, Badaga is at its core a Nilgiri language—a language of the hills—not a dialect of Kannada. Instead, we can say that it developed through koineization, a process of agglomeration of different linguistic features.Footnote ⁷¹

Pilot-Raichoor argues that, while the Badaga language is heavily influenced by Kannada, it is not a direct descendent or dialect of it. This is an important point. The concept of koinization is based on an assumption that language is not biological, does not function as a biological species, and therefore should not be thought of only in terms of ancestry and descent. Instead, koinization is a process that must be understood as one of interaction between speakers of different languages, and the mutual co-constitution of a koiné as a shared medium of communication. As comparative and historical linguistics has waned in popularity, especially in American- and English-language research, there has been limited progress in our understanding of the relationships between Nilgiris languages. But, whether or not one agrees about the origins of Badaga as a Nilgiris language, or even whether such an endeavour as searching for linguistic origins is worthwhile, the evidence that Badaga was used as a lingua franca within the hills remains well documented, at least in the colonial period.Footnote ⁷²

The formation of Badaga identity as an ethnicity, as a group, likely did not take place until around the time of initial European colonial contact—around the late sixteenth or early seventeenth century. This may be concluded in part based on the mention of the Badagas as the ‘neighbours’ of the Toda in a letter by Fr. Finicio in 1604, though it would appear that other communities were integrated into the Badaga community as late as the eighteenth century.Footnote ⁷³

Hockings's recent book, So Long a Saga, adds a great deal of data to this story, detailing the many and diverse origins of the different phratries within the Badaga community.Footnote ⁷⁴ Hockings recounts: ‘There are two very small phratries living on the northern edge of the plateau near Si:gu:ru. These are the Kurumba:ras (“potters”) and the Be:das (“hunters”), the latter from Nagatahalli, and both groups report that they originally came to this locality while in the service of the chieftain of Ummattūr.’Footnote ⁷⁵ These two phratries or sub-groups of the Badaga community thus probably arrived in the seventeenth century—a tumultuous time, after the earlier migrants who came around the time of the fall of the Vijayanagar empire in 1565. Yet, despite different chronological and geographical origins, these communities were considered fully a part of the Badaga community by the early nineteenth century.Footnote ⁷⁶

To return to the point made by Scott for Southeast Asia, the hills were places of refuge for many people fleeing the state-making enterprises of the plains, fleeing conscription, corvée labour, slavery, taxation, and so on.Footnote ⁷⁷ The actual inhabitants of the hills were not necessarily fixed populations or monolithic ethnic groups, but rather fragments, or what Scott called ‘shards’ of populations who had fled from the plains.

Taking into consideration the cultural, linguistic, and religious heterogeneity within the Badaga community, we can consider the historical origins of that heterogeneity as composed of various migrant communities from the plains; this explains the linguistic picture of diverse dialects, and koinezation, as well as the heterogeneous religious, marriage, and food-consumption practices described by Hockings for different Badaga sub-communities and phratries.Footnote ⁷⁸ This revised picture does not preclude the possibility that there was a core or original refugee population in the sixteenth century, and that they probably had some influence over later arrivals. The migration from Karnataka in the sixteenth century cannot explain all aspects of what we know about the people who identified themselves as, or were identified as, Badaga by ethnographers, census-takers, and colonial observers.

It is, however, possible that we should not give sole credit to the colonial project for creating and/or reifying Badaga group identity. There may have been other reasons for ethnopolitical actors to want to crystallize Badaga groupness, and that process may have been going on for centuries before colonial contact. We should consider that, while there had been refugee populations of assorted linguistic backgrounds, potentially speaking a koiné that we call Badaga for a long time, the arrival of new refugees or migrants may have been a catalyst leading members of this heterogenous population and language community to try to establish a more formal groupness—a group identity that could be put on the same footing as Toda and perhaps Kota communities that had already established their groupness, and had established land rights/ownership and other privileges associated with their groupness.

Agglomeration of refugee populations into the heterogeneous mix of Badaga-speakers may have taken several forms, depending on the social position that those people had occupied in plains societies. If the refugee populations had belonged to low-status groups, they would have had powerful motivation to forget about their former membership in castes, jatis, or other groups/categories/labels. If one could migrate up into the hills and escape the strictures or stigma of low caste or untouchability, surely this would make many people more than willing to leave any sort of prior groupness behind.

If, on the other hand, the members of the group of refugees formerly enjoyed some benefits of groupness associated with a caste/jati or other marked group identity in their previous homeland, the sudden loss of these privileges would be problematic and difficult. The degree of influence of sixteenth-century Kannada on Badaga language is quite strong, which is, in part, what led to its misidentification as a dialect of Kannada.Footnote ⁷⁹ This may suggest that the previously Kannada-speaking sixteenth-century refugee population was powerful both numerically and perhaps also socially. From this, we might infer that the sixteenth-century refugee Badagas had previously been of at least moderate social status before immigrating to the Nilgiris.

Above, I argued that the Toda's groupism was an early process compared with the Badagas, perhaps as early as the twelfth century ce. But the Toda community has its own internal heterogeneities, as well as internal hierarchies of purity, and status. The Toda may have become externally bounded early on in history, but they are not now, nor have they likely ever been, internally homogeneous. There are several divisions of Toda society, though these divisions do not produce major power differences; they are not status-neutral. In addition to the divisions between clans, the rules of purity and pollution severely limited the status and role of women in Toda society.

To return to the question of hierarchy between the Nilgiris communities, I wish to return to the relation between the Todas and Badagas, and which of these might be viewed as higher-status. For those who argue that the Todas are at the head of the Nilgiris hierarchy, the key fact is that of gudu, translated as ‘rent in grain’ or ‘rent in kind’, which was (at least just before the British period) paid by Badagas to Todas for the lands they cultivated. This has been viewed as a kind of peasant–landlord relationship, analogous to such relations between peasants and landlords in the plains. However, I would argue that these relationships are not at all analogous. In other parts of India in which the peasant–landlord relation has been historically discussed, landlords typically have a significant amount of power, in the form of both might and wealth, which can be used to extract rents. In such a feudal system, if there ever were a landlord who did not receive the share of grain he expected, he would simply go and seize it.Footnote ⁸⁰ Nothing like this situation ever seems to have existed between the Todas and the Badagas. In fact, there were two means of recourse for resolving such conflicts. One was to call a meeting of the naim, or Toda council, which typically included the heads of the Toda clans, as well as the paramount head of the Badaga community.Footnote ⁸¹ The other alternative was to go to officials of the state, whichever one that might be. One Toda song recounts going to remedy grievances between Toda clans, to the Tamilians, the ‘golden official’, which was probably someone (whether a British official, or representative of Tipu Sultan, or the Wodeyars, or any of the other prior occupying powers) residing at the Dannayakankote fort.Footnote ⁸² In addition, it seems that, whenever a plains-based state was effectively collecting taxes (as was occasionally the case), the Toda were expected to pay in grain, which they only received from the Badagas through gudu, or possibly through the barter of ghee. Thus, the Badagas paid (or gave) gudu to the Toda, and did so without much coercion from the Toda. Toda claim a sort of superiority by ritual purity over the Badaga but, though this is a form of hierarchy, it is not accompanied by great imbalances in power. Kotas and Kurumbas are tied to the Toda and Badaga communities through historic relationships of obligation and service. These might be viewed as producing greater power differential, and perhaps even marked by a level of coercion in extracting the obligations and services of the Kota or Kurumba parties. Yet, to say that the system is caste-like is too extreme. The Kota and Kurumba might have occupied positions that multiple communities viewed as lower, yet they were not oppressed or exploited in their labour to nearly the same extent as the lowest members of caste society.

But a social hierarchy is not a state. Zagarell argues for the more or less constant intervention of segmentary states (à la Burton Stein), saying that:

On the other hand, considering the oral account presented by Metz, it does not mean that there were not attempts to create a Toda state, ruled by a king and queen.Footnote ⁸⁴ Zagarell's argument suggests that the lack of a state (or states) can be taken as evidence of the interference of plains states, and therefore evidence of the profound non-isolation of the Nilgiris from the surrounding regions.

His argument parallels mine in many ways in suggesting that the formation of the communities of the Nilgiris, as they came to be found at the beginning of the colonial period, owes much to the histories of entanglement between plains and hills, and the intrusions of plains states into the hills. However, I differ in my understanding of the processes by which these communities specifically arrived at the groupism that they displayed at the point of colonial arrival.

I argue that we should consider the ways in which these communities had greater agency in relation to the states they were avoiding, or attempting to avoid, and that they did not simply fail to evolve into a state society of their own. Such states (or statelets, as they have sometimes been called) did emerge in other contexts, like north-eastern India.Footnote ⁸⁵ This returns us to the concept of isolationism, rather than passive isolation. What does isolationism mean in this context? You might be wondering ‘how does a community or a society of multiple communities, come to the decision to adopt an isolationist stance?’. I would suggest that it is not a ‘decision’ per se, but rather the gradual development of an ethos. That ethos is one of avoidance, of outsiders of various kinds, but also avoidance of the kinds of trappings of state power, especially bureaucracy.

There is no evidence to suggest that the Todas, Badagas, or other hill communities ever raised a revolt, armed or otherwise, against any of their successive overlords from the earliest hints of conquests in the tenth century onwards. They do not appear to have been militarized in any way. Nor did they ever adopt or invent a writing system, which would have facilitated systematizing tribute (or gudu) collection. The Toda were not the feudal overlords of the Badagas, Kotas, Kurumbas, and Irulas; nor, despite the oral history collected by Metz, does it seem that they ever really tried to be. As compared with the tribal communities of western India discussed by Guha and Skaria (among others) and north-eastern India discussed by Guite (among many others), the Nilgiris communities seem to have been uninterested in the kinds of raiding and military resistance displayed in both those cases. But, again, this need not imply a passive ‘failure’ to resist.

Rather, the ethos was one of avoidance, perhaps at least partly encapsulated in the tale enjoining the listener ‘don't make friends with a Tamilian’. Isolationism took the form of subsistence strategies illegible to states, avoiding taxation, or, at the very least, making it difficult to levy taxes. This is easily argued for the moveable herds of Toda buffalo pastoralism, as well as the Kurumba and Irula, living on the steep slopes, foraging (and trading forest products). This isolationism was evasive rather than aggressive. Toda spoke the Badaga language to outsiders and, according to some accounts, acquired many of their plains products through Badaga intermediaries in the form of long-term trade partners, who had obligations to provide some of those materials to the Toda, while the Toda had reciprocal obligations to them.Footnote ⁸⁶ What they may have done, or how they might have conducted trade, before the arrival of Badagas in the sixteenth century is a difficult question to answer. If the oral histories provide any clue, it is that the Toda may have preferred to go down to the plains to trade and obtain salt, and maybe cloth, rice, and other items. They probably preferred to preserve the isolation of their homes in the hills and not to invite outsiders in. This is arguably a fairly effective strategy of avoidance, while simultaneously allowing them to benefit from the goods produced outside the hills.

Once Badagas of various stripes arrived in the sixteenth century, their status as outsiders of a sort may have been something of a boon, allowing Todas to isolate themselves further from direct contacts with outsiders. The Todas appear to have delegated the trade with lowlands merchants to the Badagas and, as a result, individual Toda families developed specific trade partnerships with individual Badaga families, who in turn developed relationships with Chettis who were perhaps both merchants and producers of goods like cloth.Footnote ⁸⁷ Kotas also developed similar trade relationships with the Badagas, and depended on them and their trade connections for cloth as well.Footnote ⁸⁸

As a result of their isolationist attitudes, the Toda in particular developed a strong sense of the boundary of Toda and others. This ethos of isolationism, and the resulting groupism, may have driven them to find ways to consciously express their differences, through the characteristically distinctive styles of hair and dress, as such practices have been argued in many contexts for the marking of ethnic boundaries.Footnote ⁸⁹ This may also have resulted in the active resistance against adopting outside gods or religious practices and the establishment of a religious orthodoxy. And so they appeared, with distinctive hairstyles, dress, language, and religion, as a community frozen in time, utterly isolated from the plains societies around them.