In modern anthropology, the case of the Nilgiris has been used to construct very different sociological models . It has been equally easy to prove that the inhabitants were isolated tribes or that they were part of a jajmânî-like system of interdependence, with either the Todas or the Badagas as the dominant caste. Although the weakness of the data from remote times has indeed something to do with this, these same data were still sufficient to lend themselves to divergent elaborations, which clearly points to theoretical and ideological bias in the construction of those models.
Nevertheless, I do not want to enter this debate which has, to my mind, been skewed from the start. The bases of the British distinction between ‘caste’ and ‘tribe’ were never clearly defined. These terms have a history in which scientific and political considerations have always been intertwined. When the Indian National Census endorsed the distinction and created exclusive categories in 1871-72, it placed endless difficulties before the administrators and soon compelled them to look to the transformation of tribes into castes (Sinha 1980 : 2, 7). By adopting these same sociological labels, the Constitution of 1949 established them as political dogma, inseparable from Indian Independence and unity, and turned them into political stakes. Since then, as Paul Hockings (1993 : 351) reminds us, “the academic question has been overtaken by the political ramifications of being identified with one or the other”. Challenging these labels is therefore interpreted, ipso facto, as a political issue.
For my part, I would rather question the perception and the definition of the Nilgiri peoples during the 19th and at the beginning of the 20th centuries. First of all, I shall examine what was initially written about the Nilgiris.  To a certain extent, one can follow Hockings when he writes, concerning the words ‘caste’ and ‘tribe’ : “there is no point in our seeking for consistency in the early use of these various terms. It was only after the anthropological writings of E. B. Tylor and Herbert Spencer had had some impact on the public, and the Census of India was getting itself organised (1871 and later), that these terms began to lead separate lives” (Hockings 1993 : 352). But leaving aside their intrinsic meaning or sociological relevance, a formal analysis of the frequency and the evolution of these words can throw some light on their authors and their time. The choice of what today we call sociological concepts was never discussed. But that does not mean it was wholly arbitrary.
I shall consider only eyewitness accounts, leaving out all the compilations, despite their interest. The analysis of these texts allows us to identify three main periods, or rather trends, which emerge one after the other and go on to develop without becoming entrapped by specific temporal limits.
Variability of the Terms Used with a Preference for ‘Caste’
The first reports we are able to consult were written by civil servants on assignment, and more precisely by surveyors. After the cession of the Coimbatore District, in 1799, the first task of the British was to abolish the ancient revenue system and establish a new one. Hence it was felt necessary to make a careful survey of the new territories. Three of these reports, written between 1812 and 1821 , have come down to us :
(a) William Keys, Assistant Revenue Surveyor, sent, in June 1812, “A Topographical Description of the Neelaghery Mountains” to the Collector of the Coimbatore District, W. Garrows.
(b) Lieutenant Evans Macpherson, 1820, Superintendent, Neelgherry Road, transmitted a brief report to John Sullivan, the next Collector of Coimbatore, who had entrusted him with the construction of the first road and the survey of the lands.
(c) Lieutenant B. S. Ward, Deputy Surveyor-General, was the author of a more substantial Geographical and Statistical Memoir of a Survey of the Neelgherry Mountains..., dated 1821. 
As Evans Macpherson wrote to John Sullivan, who pressed him to state his opinion, they are “In humble situation in this country”. Being technicians who spent from one to three months in the Nilgiris, they are “little in the habit of writing on any subject”, and write “with pleasure but with much diffidence”.
These first three reports, written independently, described physical, commercial and agricultural conditions more carefully than human aspects. They presented the inhabitants with their respective names, but the generic terms keep changing. In addition to ‘people’, ‘population’, and ‘inhabitants’, the term ‘caste’ was used in the majority of cases. In these three reports, I have found seventeen instances of ‘caste’, six of ‘race’, five of ‘class’, and one of ‘tribe’.
This last report referred to the Todas,  in the space of a few pages, as ‘a migratory race’ and ‘a migratory tribe’. ‘Class’ always referred to a sub-group in a broader whole.
‘Race’ had, at this early date in the nineteenth century, a very general meaning close to the old acceptation connected with European feudalism, and one understands that it could have been equivalent to caste. Indeed it emphasized appearance, a difference either from the people of the plains or within the Nilgiri people themselves, but it did not stress physical features alone.
For instance, the region is “inhabited by a race of people differing in language, appearance, and doubtless in origin” (Macpherson). Although they are a caste, the Todas “are evidently a distinct race”, and the Mullukurumbas “a primeval race” ; with regard to the Kotas : “This caste of people are the most industrious race” ; and to the Irulas : “These are a distinct race of people from the other highlanders” (Ward).
But, as I said, the term ‘caste’ was most frequently used. The first visitors to discover and explore the region were prepared to see castes with occupational specialisations, interrelations and even differences of status.
Of course, the context contributed to their preference for the use of ‘caste’. The region was discovered rather late in the history of British colonisation. Therefore these civil servants had been in India long enough to be familiar with the caste society and its features (such as classification of activities, sharing of food, etc.). Moreover, they were accompanied by Indians – Tamils, servants and followers, who translated what they encountered according to their usual way of thinking. Lastly, as Rivers (1906 : 15) and Emeneau (1963 : 191) have shown, “outsiders have usually communicated with the Todas through the Badagas”, and through the Badaga language, if not Tamil or Kannada. We know today that the Badagas were the most recent newcomers in the hills, the closest to the world of the plains, and the most interested in differentiating themselves from other hill peoples. The first accounts were definitely biased in favour of castes.
What was observed confirmed this opinion.
1. All these surveyors discerned a subdivided human whole : “These (nauds [areas]) are inhabited by a race of people (...), and divided into twelve castes “ (Macpherson).
2. The occupational specialization was immediately apparent for them. The Todas were exclusively herders, and migratory. The Kotas were cultivators, and very industrious craftsmen. The Irulas cultivated the slopes, gathered forest-products, “fell[ed] large trees and conveyed the timber down to the plains” (Keys). The Badagas were the principal cultivators.
3. This specialization entailed a certain interdependence, and involved economic as well as ritual relations.
On the one hand, the Todas received a part of the crop from the Badagas “and others”, who could only be the Kotas (Keys, Macpherson), which made up their only source of cereals. The same held for the Mullukurumbas, because “the produce of their fields (...) seldom affords them a sufficiency for consumption” (Ward). The Kotas got unserviceable old animals from the Badagas for a trifle. The Irulas “often come down to the plains to dispose or barter plantains and other hill productions at the market villages” (Ward).
On the other hand, the Mullukurumbas (nowadays identified as Âlu or Pâlu Kurumbas) served as priests to the Badagas, especially at ploughing and harvesting time. The Mullukurumba began the ploughing in each field, for which he received from each ryot a few measures of the ripe grain. And at harvest time he was required to reap the first handful of the grain, and “is then permitted to take as many sheaves as he can bear away on his shoulders” (Ward). As for the Todas, they received “from quarter to half a Rupee as a present” whenever wealthy Badagas celebrated a marriage (Ward).
4. Stratification  was supposedly based on criteria such as hereditary land-rights, gift of a certain portion of each crop yearly, commensality, free entrance into the house, and on comparisons with the low country. The Todas “appear to be the aborigines of the hills ; they are acknowledged lords of the soil by other castes” (Macpherson) ; they are “the hereditary claimants of the soil” (Ward), as attested by their providing a portion of the harvest. The Kotas form “the lowest class” among the three who live in the upper hills (Keys). “They are of a very inferior caste, and by their neighbours considered in the same light as the chuckler in the low country ; and none of the other castes will eat with them or even enter their houses” (Macpherson).
In short, the first discoverers encountered distinct, named groups which they identified as ‘castes’. They recognized features they had already observed in the plains and which would later be systematised as characteristic of the caste society. For them, South Indian human diversity could easily put up with a few additional groups with their own distinctive features. This trend did not suddenly vanish, and traces would be found in later works.
‘Tribes’ Take the Place of ‘castes’
The next group of authors still used such various terms as ‘race’, ‘class’, ‘tribe’, ‘caste’, but their frequency as well as their usage changed. Except for Harkness, all used ‘tribe’ in the title of their publication. I shall follow the chronological order in presenting five of them, who seem representative and who have shaped this trend.
1) In May 1819, Louis-Théodore Leschenault de la Tour, a French naturalist, spent a short period in the hills along with John Sullivan. He was very busy gathering 2,000 plant and 156 animal samples which he was to send back to France, and probably did not actually meet many aboriginal peoples. On his return, he wrote an account in a Ceylonese paper ; then on 9 September 1822, he read a Relation abrégée d’un voyage aux Indes orientales before the Académie des Sciences, in which he explained :
“Les habitans sont peu nombreux, ils paraissent fort doux, ils mènent une vie heureuse et indépendante ; ils sont divisés en trois tribus : les boggers, les cotters et les totters ; les derniers qui habitent les régions les plus élevées, sont regardés comme les habitans primitifs, ils ne sont que pasteurs ; ils possèdent de nombreux troupeaux de buffles. Les deux autres tribus cultivent la terre ou exercent des métiers utiles” (Leschenault de la Tour 1822 : 16).
So it was a French savant, “naturaliste du Roi”, endowed with the prestige of his office, who deliberately used the word tribe in this region for the first time. He knew nothing about the Indian population and society, and was also ignorant of the connotations this term carried in the context of imperial India. One should know that, from the seventeenth century onwards, the word ‘tribe’ had been in common use (in English as in French) in the natural sciences, and particularly in zoology, to designate the intermediate level in the systematic classification, between the family and the genus.
2) In 1832, Captain Henry Harkness, of the Madras army and Secretary to the Royal Asiatic Society, published the first book devoted entirely to the Nilgiris and their inhabitants. He spent at least nine months travelling throughout the area, and apparently knew something of the Toda language (cf. p. 43).
The word ‘race’ appears, first in the title of his book, A Description of a Singular Aboriginal Race Inhabiting the Summit of the Neilgherry Hills, which refers to the Todas. It is then used in the book to refer to each group. It indicates a particular appearance, made up of a set of morphological and behavioural features. So for the Todas, their height, eyes and teeth, but also their bold bearing and expressive countenances “lead to the conclusion that they must be of a different race to their neighbours”. The Kotas are “a strange race and differ as much from the other tribes of the mountains as they do from all other natives of India” ; the Kurumbas “another race”, the Irulas “a race of people”, and the Badagas “a timid race”.
Yet these same groups were above all tribes. At the same time, and even though he classified three of them (the Kurumbas, Irulas, and Badagas) as Hindus, Harkness denied the existence of caste. For him, the Badagas were divided into eight different classes, and the Irulas into two ; the Kotas have “no distinction of caste” (p. 30) ; the Todas do not know of “any difference of caste among themselves” (p. 31), even though he distinguished between two branches which, “till within the last few generations, kept themselves quite distinct, and never intermarried” (p. 32).
Still this author did not isolate these new tribes, and in his opinion comparison was required if for no other reason than to further knowledge : “In order to come to any definite or satisfactory conclusion concerning the Todas, or their language, it is necessary, however, to know the prominent features, and history of the other tribes who dwell on these mountains” (p. 27).
3) The Reverend J. Friedrich Metz published several articles in English in the Madras Christian Herald in 1856, then in German in 1857 (Die Volkstämme der Nilgiri’s), before a friend later edited his notes into a book in 1864 : The Tribes Inhabiting the Neilgherry Hills. Perhaps more than by his writings, this missionary of the Basel Evangelical Missionary Society played an important part as an intermediary (and, one may well imagine, informant) for all visitors over a long period in the mid-nineteenth century.
There is again much confusion between the notions of race, class, tribe, and caste in his book, and no specific meaning or use can be distinguished.
« There are five different races. The Kotas are the only one of all the hill tribes who practice the industrial arts, and they are therefore essential almost to the very existence of the other classes. They are, however, a squalid race (...) and are on this account a byeword among the other castes » (Metz 1864 : 127)
Yet a new phrase appears – ‘hill tribe’ – in which hill becomes a qualifier : that is, one must be from the hills to be tribal. That is why the Irulas cannot be called a hill tribe, “partly because they reside so low down, and partly because their language (...) is a dialect of Tamil” (p. 11). The same held for the Badagas, as their history and relatively recent migration were better known. It should be pointed out however that Metz worked almost exclusively among the Badagas, whom he encouraged to break off relations, especially ritual, with the other groups.
4) Among those who helped diffuse the notion of tribe, special mention must be made of James Wilkinson Breeks, of the Madras Civil Service, who was the first Commissioner of the Nilagiris,  from 1868 till he died in 1872, and who made a study on orders.
In May 1871, the Indian Museum of Calcutta decided to open a section to illustrate “the state of the arts among the aboriginal and other jungle races in India”. To this end, the curator sought to obtain “collections of their arms, ornaments, dresses, household utensils, tools, agricultural implements, musical instruments (...), that will serve to illustrate the habits and modes of life of those indigenous races that have remained but little affected by foreign civilization”. Breeks replied promptly (in July 1871) and enthusiastically. It was obvious that he was already well acquainted with the question. He would be glad to procure a collection :
“to illustrate the habits and modes of life of the jungle tribes of the Nilagiris (...) for year by year the Nilagiri tribes at any rate are abandoning their distinctive customs. For example, among the Todas infanticide has been put a stop to ; polyandry is on the decline ; their buffalo sacrifices at the dry funeral are curtailed, and create comparatively little interest (...) Amongst a people in so low a state of civilization, a very few years serve to efface all trace of a custom that has been given up ; whilst a careful examination into their existing condition and habits, affords the only hope of arriving at their origin and past history.”
It was a matter of collecting and buying objects in order to document peoples that had remained “little affected by foreign civilization”, and to acquire “materials for a comparative study of the arts practised by races in an early stage of social development” — in other words, primitives, in the original sense of the word, with a view to an evolutionistic history of humankind. The long series of quotations that precedes has no other purpose than to highlight the gap between local and imperial knowledge and the fact that the latter is bound to prevail. In its search for peoples left behind by civilisation, the Museum of Calcutta was not listening to the man in the field, despite his title of Commissioner, who said that the traditional customs had been seriously transformed by colonisation and were doomed to vanish within a short time.
J. W. Breeks proposed an inquiry “as exhaustive as possible”, bearing on “the four Nilagiri, jungle tribes, viz. : The Todas, the Kotas, the Kurumbas and the Irulas”. It should be noted that the Badagas were no longer an aboriginal or jungle race, but the Irulas had been restored to their status of tribal, because they lived in the jungle. His Primitive Tribes and Monuments of the Nilagiris was published in 1873, one year after his death.
5) The same year, Colonel W. E. Marshall published A Phrenologist amongst the Toda, or the Study of a Primitive Tribe in South India. For him, the Todas were a case for studying how “savage tribes melt away when forced into prolonged contact with a superior civilisation”. This book is highly significant, since it falls unequivocally into the field of anthropology, which was just being created, and shows just how influential ideas discussed in Europe were. It attracted immediate review from one of the prominent French physical anthropologists of the time, Jean-Louis-Armand de Quatrefages de Bréau.
To summarize, this pro-tribe trend was impelled by authors with a sound educational background (naturalists, medical doctors, military men, administrators and museologists), who were also members of various scientific societies. To this list must be added : the philologist B. H. Hodgson, member of the Asiatic Society, who, on the basis of lexical information provided by Metz and other missionaries from Kaity, argued (in opposition to Harkness) against the Todas’ singularity and affirmed the close relationship between and the common origin of hill and plains languages ; J. Shortt, physician and surgeon, member of medical, anthropological, ethnological and zoological societies in Great Britain and India ; and Lieutenant later also Colonel W. Ross King, member of the anthropological society of London. The first tendency was continued by the missionary Metz, by the probably less educated James Hough, chaplain in the Madras establishment, and by occasional authors publishing in wide-circulation newspapers (Rifle 1873).
For writers of the second trend, each group was a tribe, which had (Hough 1829 : 109) or, more often, did not have caste (Harkness 1832 : 30, 31 ; Shortt 1868 : 57). Notice that caste has changed levels. For the preceding “humble men”, caste referred to each group ; now it is located within each group. What then did they mean by the term ?
At first these tribes were never considered in isolation. Their mutual relations, even though not studied in depth, were always pointed out. The same tribes were then qualified as hill-tribes, which excluded the Irulas and the Badagas, and later as jungle, primitive, and even savage tribes, which led to the reintegration of the Irulas, but not the Badagas. At this point their economic or ritual relations were no longer noted. In any case, the qualifier chosen for the word ‘tribe’ became determining. No thought was given to the notion of tribe in itself, rather it was as a function of its hill or jungle character that a group fell into the category ‘tribe’. 
‘Tribes’ redefined in terms of ‘castes’
A third trend became apparent at the beginning of the twentieth century and continued until well after independence : it was the turn of the professional scientists. There were few of them : W. H. R. Rivers (1906), M. B. Emeneau (from 1935), and D. G. Mandelbaum (from 1937). All nearly always used the word ‘tribe’, but carefully avoided the image of jungle or primitive tribe.
W. H. R. Rivers’ monograph, recalling the Todas as “picturesque and, in many ways, so unique”, does not forget that they had never been isolated, and that they had particular relations with the other tribes. In addition to an entire chapter devoted to “Relations with Other Tribes”, many notations relativized even the isolation of the plateau. Rivers went on to suggest that a more exact knowledge of Indian customs and ceremonies could show their difference to be slighter than they seem to be at present (Rivers 1906 : 4, note 2).
Murray B. Emeneau, a linguist and Sanskritist, was the great decipherer of the languages and oral literatures of the Nilgiris, and in particular those of the Todas and Kotas. It may be presumed then that he had the most intimate knowledge of these groups. Although he did not specialize in political or social organisation, he set out expressly to supplement Rivers’ account and found the existence of a system of matrilineal clans which Rivers had missed (1938, quoted in Rooksby 1971 : 114-115). For him the Kotas were “a small tribe or caste”. He entitled the first Kota myth : “How Kitu:rpayk (...) originated the three castes”.  But he still used a formula which brought sharp reactions from Marian W. Smith : “For a long time (...) these tribes formed a local but not too aberrant version of the Hindu caste system” (Emeneau 1944 : 1). The specialization and the interrelations led him to write : “The communities are symbiotic economically and even, to some degree, religiously”. He was even firmer in his “Introduction” to the Toda Songs (1971 : xxiv) : “The Toda community is often referred to as a tribe. It is in reality part of a local caste-like system, the Nilgiris system, which includes also Kotas, Kurumbas, and Badagas”.
David G. Mandelbaum arrived in India in 1936, and almost immediately went to work with Emeneau. At first he did not question the notion of geographic isolation, the idea of a social and cultural ‘enclave’, constituted of several tribes living in economic and social symbiosis. But from 1952, he described their economy as “geared to a caste-like division of labour”, then in 1954 as “a caste-like interdependence”. In 1955, he even went so far as to write : “Yet for all their differences from the usual south Indian patterns, all the Nilgiris peoples accorded, in fundamental ways, with the prevalent village tradition”. He made the comparison even clearer in 1956 (1989) : “Kotas are, and were, decidedly more akin to what Sinha calls the Hindu peasantry of the plains than they are like the tribesmen of the central hill belt”.
Apparently, it was his interest in the modern change that led him to new interpretations and expressions : “If the Kotas have not been a caste before, they are becoming more and more like the lower castes of the plains below their homeland” (1989 : 184). But one can also believe that he evolved as he came to know Indian society better, as he could compare his data and analysis with other South Asian scholars, whereas he had arrived in India after writing a dissertation on the Plains Cree, “An Indian Tribe”.
The Scientific Issues
In the Nilgiris, far from being immediate, the notion of primitive tribe was a romantic,  administrative, and scientific construction : three aspects which could be found in the same person. The notion of tribe prevailed as the British settled there, as roads were opened on every side of the plateau  and as markets were organised, making it possible to send off local products as well as to acquire goods from outside, as villages of immigrants from the plains grew up (Hockings 1989 : 337). The more the region opened up geographically, socially and economically, the more the aboriginal people were seen as being different, primitive, associated with the jungle (or savage), and consequently as the epitome of a tribe. This was aggravated as the description and organization of the region was taken out of the hands of local administration to depend on decisions made at the higher level of British India (e.g. the orders from the India Museum), or on concerns having little or nothing to do with India (as will be seen, ideas relevant to European scientists).
The needs and interests of the British authorities and administration in India have always been advanced as an explanation, and of course they must be taken into account. By the second half of the eighteenth century, the East India Company had already developed an administrative system, with its own categories and ways of dealing with them. But once the Europeans adopted the word ‘caste’, and throughout the seventeenth and eighteenth centuries, the terms ‘caste’ and ‘tribe’ were used interchangeably, ‘caste’ sometimes being taken for the vernacular term and ‘tribe’ for its translation (Yule and Burnell 1903 : 171-172). The two terms were still approximately equivalent at the turn of the nineteenth century. 
In the Nilgiris, both trends (pro-caste and pro-tribe) left their mark on, and were in turn influenced by, debates on the matter of land tenure and revenue. These discussions set different parties at odds. On the one side, people like John Sullivan, the founder of Ootacamund, who had paid the Todas for the land he took from them, “endorsed the absolute proprietary rights of the Todas to the entire Nilgiri plateau, on the presumption that they were the earliest settlers there”. On the other side, those like Stephen R. Lushington, Governor of Madras, or the Military Commandant of Ootacamund, argued that “throughout India proprietary right in land belonged to the Government” (Hockings 1989 : 340 ; Grigg 1880 : 314-343). One party insisted on their being particular and called the hill people ‘tribes’, whereas the other wanted to consider them to be just like other South Indian castes.
But I would like to emphasize the influence of scientific issues on the designation of the Nilgiri groups. Bernard S. Cohn (1968 : 16-17) pointed out that the administrative-official view of caste reflected the anthropological interests and theories of the period between 1870 and 1910. I believe that this influence made itself felt even earlier, before the birth of anthropology as we know it today, in other words, basically social and cultural anthropology. But this can only be outlined within the scope of this essay.
This influence can be surmised by the swiftness with which publications circulated, and by the interest they raised in Europe. Leschenault de la Tour was read by his English contemporaries and quoted as an authority. Marshall’s book was reviewed at length (three articles for a total of forty-three pages) the next year by A. de Quatrefages, one of the founders of French physical anthropology. He had already read King’s publication (1870) and had seen photographs and cranial measurements of the hill people brought back by Mme Janssen, wife of the astronomer Pierre-Jules-César Janssen, who had gone to India to observe the great solar eclipse in 1872. De Quatrefages was also an admirer of Adolphe Pictet, Sanskritist, inventor, in 1859, of an ethnology of the Aryas, whose perfect language could only be the reflection of an ideal state of humanity (Olender 1989 : 127-134). Elisée Reclus, and his brother Elie, both libertarian French geographers and tireless travellers, showed in the Nouvelle Géographie Universelle (1883) and in Les Primitifs (1885), that they were astonishingly well documented on the Nilgiri people. At a later period, Emeneau’s Kota Texts were discussed by Marian Smith, an American folklorist and linguist, two years after they had been published. 
Conversely, learned writers who described the Nilgiri people were probably well informed and nourished the discussion in European scientific circles. Since the study of Sanskrit began, India had been an exemplary place in which to investigate various issues of the human sciences to come, all of which were connected and associated with the attempt to trace the origins of mankind, with the old dream of discovering the roots of humanity, religion and language. All depended on the notion of race as an “explanatory framework” ; races which, in Renan’s words, were “intellectual and moral moulds” (Laurens 1988 : 376). For the new discipline of anthropology,  the heir to zoology, one of the main issues was the classification of human races, which opposed monogenists and polygenists, and encouraged the search for autochthonous, unknown races. The Nilgiris constituted a true laboratory for observation and experiment. They presented salient features allowing the observers to infer that some of the Nilgiri people were either a “hardly modified sample of the first human races” (Marshall), or “representative of an exceptional type, remained in a pure condition” (Quatrefages).
The relations between observers in India and new fields of research can throw a different light on the use and evolution of words like ‘race’, ‘caste’, ‘tribe’, which were not yet sociological concepts. They may explain the omission of caste by some writers in the mid-nineteenth century. Caste did not exist outside India. It did not exist for the European anthropologists, who only knew the tandem ‘race’ and ‘tribe’.
‘Race’, an extremely common word with no precise definition was the object of a variety of interpretations coloured by tendencies and persons. Moreover, the term evolved considerably over the course of the nineteenth and twentieth centuries, and one must be wary of anachronistic interpretations. Until sometime around 1875,  a race was defined not only by physical features, but also by way of life, intellectual and social state, and above all by the language spoken. The nineteenth century witnessed the triumph of comparative philology. For scholars of this time, it was language that guaranteed a people’s identity over time, which revealed the people’s fundamental features and even authorized one to speak of ‘linguistic race’. This reminds me of an aspect of the descriptions of the Nilgiris that I overlooked in presenting the texts, which was the attention given to linguistic resemblances and differences, and the collection of vocabularies for comparative purposes. The notion of ‘race’ was applied to all mankind and, either did not carry a value judgement, or carried the idea of progress on the path to the perfection that was Christian civilisation.
The notion of ‘tribe’ was less extensive. It applied only to particular races, to which nineteenth-century evolutionism assigned a particular stage in human evolution, leading from the simple primitive to the complex civilised. Geographical and social isolation (up on a hill or out in a jungle) vouched for their primitiveness, original purity and scientific value. European scientists looked for correlations between the different criteria to constitute distinctive entities and typologies. Their method increased the isolation of each group.
In the early twentieth century, social anthropology had developed in India as elsewhere. The caste society was now sufficiently recognized to give rise to attempts at synthesis and explanation. The choice between ‘tribe’ or ‘caste’, terms inherited from the past, then referred to and had to agree with the prevailing historicist view of Indian society, and particularly of South India. According to this perspective, the coincidence between the arrival of Aryans, Hinduism and the caste system allowed for the development of a great and complex civilization. The Dravidian population was originally comprised of tribes that had been either transformed into castes by integration into the population of Indo-European origin and Hindu religion, or remained unchanged.  To fit with this view and to confirm it, Nilgiri people had to be isolated, non-Hindus, that is tribals. Considering the inhabitants of the Nilgiri as castes would have upset this model. While researchers in the field tended to see castes or something resembling them, they were held back by bigger interests (bigger than the researchers or the field of Indian studies) which they were neither ready nor disposed to confront.
By its interest in contents, in the manufacture of scientific facts and not only in the framework or the socio-cultural context, the history of sciences showed long ago that it can be nothing if not social, and even anthropological. The perception and the constitution of sociological reality are inseparable from the concepts which express them, and these concepts are not given in advance. They are the outcome of an often controversial epistemological and sociological construction and have evolved over the course of time.
The elements at stake go well beyond the main forces at play and the research trends generally recognized in Indian studies. Combining and comparing these elements has given rise to various solutions shaped by local contexts and historical periods. We have no right to regard these as arbitrary or inconsistent without examining them, unless, that is, we assume that our predecessors were less well endowed with logical capacities than we. These solutions only seem arbitrary or inconsistent because we project contemporary meanings onto past periods, forgetting that the elaboration of these very meanings took several centuries of observation, reflection and evolution. The logic of these discourses cannot be assessed or appreciated using meanings crystallized at a later time. To state it baldly, as for example Surajit Sinha does (1980 : 1), without indicating the period, that the British discovered a rigid system of castes and tribes “outside the Brahmanic Varna-Jati hierarchy” is one such anachronistic interpretation and inevitably leads to the discovery of inconsistencies. These terms were being used well before they were constructed as two sociological poles structuring Indian society. Exactly when and how did the transformation come about ? This remains to be studied. The case of the Nilgiris cannot be generalized as it stands. Nevertheless it invites care in the reading and interpretation of texts from the past and the multiplication of analyses localized in space and time.
Breeks, James W. 1873 - An Account of the Primitive Tribes and Monuments of the Nilagiris. London, India Museum, 137 pp. + lxxxii plates. Reprint : Delhi, Cultural Publishing House, 1983.
Cohn, Bernard S. 1968 - ’Notes on the History of the Study of Indian Society and Culture’, pp. 4-28 in Milton Singer and Bernard S. Cohn (eds), Structure and Change in Indian Society. Chicago, Aldine Publishing Co.
Emeneau, Murray B. 1944-46 - Kota Texts. Berkeley and Los Angeles, University of California Press, 4 vol. (University of California Publications in Linguistics).
_. 1963 - “Ootacamund in the Nilgiris : some notes”, Journal of the American Oriental Society, vol. 83 : 188-193.
_. 1971 - Toda Songs. Oxford, Clarendon Press, xlvii-1004 pp., indexes.
Grigg, h. b. (edited by) 1880 - A Manual of the Nilagiri District in the Madras Presidency. Madras, printed by E. Keys, at the Government Press.
Harkness, Henry 1832 - A Description of a Singular Aboriginal Race Inhabiting the Summit of the Neilgherry Hills, or Blue Mountains of Coimbatoor, in the Southern Peninsula of India. London, Smith, Elder, and Co.
Hockings, Paul 1989 - ’British Society in the Company, Crown and Congress Eras’, pp. 334-359 in P. Hockings (ed.), Blue Mountains. The Ethnography and Biogeography of a South Indian Region. Delhi, Oxford University Press.
_. 1993 - ’Ethnic Identity in a Complex Society : the Badaga Case’, Bulletin of the National Museum of Ethnology, vol. 18, n° 2 : 347-363.
Hodgson, B. H. 1856 - ’Aborigines of the Nilgiris’, Journal of the Asiatic Society of Bengal, vol. 25, n°1 : 31-38 and n° 6 : 498-522.
Hough, J. 1829 - Letters on the Climate, Inhabitants, Productions, etc., of the Neilgherries, or Blue Mountains of Coimbatoor, South India. London, J. Hatchard and son, 166 pp.
Kennedy, Dane 1991 - ’Guardians of Edenic Sanctuaries : Paharis, Lepchas, and Todas in the British Mind’, South Asia, vol. XIV, n° 2 : 57-77.
King, W. Ross 1866 - The Aboriginal Tribes of the Nilgiri Hills. London, Longmans, Green, and co, 52 pp. (A paper read before the Anthropological Society of London, May 3, 1870 and published by the Society in the Journal of Anthropology).
Laurens, H. 1988 - ’Le concept de race dans le Journal asiatique du XIXe siècle’, Journal Asiatique 276, 3-4 : 371-381.
Leschenault de la Tour, L.-T. 1822 - Relation abrégée d’un voyage aux Indes Orientales. Extrait des Mémoires du Museum d’histoire naturelle, t. 9, 5e année, 30 pp. Paris, Imprimerie de A. Belin.
Mandelbaum, David G. 1952 - “Technology, Credit, and Culture in an Indian Village”, Human Organization 11, n°3 : 28. (Reprinted as “Technology, Credit, and Culture in a Nilgiri Village”, pp. 103-105 in M. N. Srinivas (ed.) India’s Villages. Bombay, Asia Publishing House, 2d ed., 1960.)
_. 1954 - “Form, Variation, and Meaning of a Ceremony”, pp. 60-102 in Robert F. Spencer (ed.), Method and Perspective in Anthropology. Papers in Honor of Wilson D. Wallis. Minneapolis, University of Minnesota Press.
_. 1955 - ’The World and the World View of the Kota’, pp. 223-254 in McKim Marriott (ed.), Village India. Studies in the Little Community. Chicago, University of Chicago Press.
_. 1989 - ’The Kotas in their Social Setting’, pp. 144-185, in P. Hockings (ed.) Blue Mountains. The ethnography and biogeography of a South Indian region. Delhi, Oxford University Press. (1st ed. 1956 in R. Redfield and M. B. Singer (eds), Introduction to the Civilization of India. Chicago).
Marshall, W. E. 1873 - A Phrenologist amongst the Todas, or the Study of a Primitive Tribe in South India. History, Character, Customs, Religion, Infanticide, Polyandry, Language. London, Longmans, Green, and Co.
Metz, J. F. 1864 - The Tribes Inhabiting the Neilgherry Hills ; their Social Customs and Religious Rites. From the rough notes of a German Missionary. Edited by a friend. Second enlarged edition. Mangalore, Basel Mission Press.
Olender, Maurice 1989 - Les Langues du Paradis. Aryens et Sémites : un couple providentiel. Paris, Gallimard/Le Seuil.
Pictet, A. 1877 - Les Origines indo-européennes ou les Aryas primitifs. Essai de paléontologie linguistique. (1re ed. en 2 vol. 1859 et 1863), Paris, Sandoz et Fischbacher.
Pouchepadass, Jacques 1982 - ’The ’Criminal Tribes’ of British India : a Repressive Concept in Theory and Practice’, International Journal of Asian Studies, II, 1 : 41-59.
Quatrefages, Armand de 1873-74 - ’A Phrenologist among the Todas... By William E. Marshall, lieutenant colonel of her Majesty’s Bengal Staff Corps, London, 1873’, in Journal des Savants, décembre 1873 : 729-745 ; janvier 1874 : 5-22 ; février 1874 : 96-106.
Reclus,Élisée 1883 - Nouvelle géographie universelle. La terre et les hommes ; XIV Inde méridionale : Madras, Maïsour, Courg, Cochin, Travancore. Paris, Hachette et Cie : 511-578
Reclus, Élie 1885 - Les Primitifs. études d’ethnologie comparée. Paris. G. Chamerot.
Rifle 1873 - “The Hill Tribes of the Neilgherries” (Madras Standard, Oct. 18), The Indian Antiquary, vol. II, January : 32
Rivers, W. H. R. 1906 (1986) - The Todas. Reprint : Jaipur, Rawat Publications, 2 vol.
Rooksby, R. L. 1971 - ’W. H. R. Rivers and the Todas’, South Asia, 1 : 109-121.
Shortt, J. (ed.) 1868 - An Account of the Tribes on the Neilgherries and Geographical and Statistical Memoir of A Survey of the Neilgherry Mountains, by the late captain J. Ouchterlony (1847). Madras, Higginbotham and Co., 76 pp. + 84 pp.
Sinha, Surajit 1980 - ’Tribes and Indian Civilization : a perspective’, Man in India 60, 1-2 : 1-15.
Smith, M. W. 1948 - ’Kota Texts : a Review of the Primitive in Indic Folklore’, Journal of American Folklore : 283-297.