Caste and Race in India c9: Chapter V. Race and Caste
Caste and Race in India c9
- Language
- English
Contents
Chapter V. Race and Caste
c9.2
¹ This chapter is an adaptation of the author's paper "The Ethnic Theory of Caste", published in *Man in India*, 1924.
c9.3
CASTE is such an obtrusive factor of Indian social organization that since the time of Megasthenes it has never failed to attract the attention of the foreigner—be he traveller, administrator, or student of Sanskrit literature. The application of the sociological method to the explanation of caste, i.e. a systematic attempt to elucidate the genesis and growth of the institution of caste, by a comprehensive study of the contemporary castes, however, may be said to begin with the works of Denzil Ibbetson and J. C. Nesfield. Both of them, in general, endorse the view² that caste is mainly occupational in origin, i.e. occupations which were organized into guilds slowly became exclusive and stratified into castes. Nesfield went further and, affirming the essential unity of the Indian race, emphatically denied that racial distinction was the basis of caste.³ This extraordinary statement of Nesfield led Herbert Risley to use anthropometry for a solution of the riddle of caste, for that alone could decide questions of racial affinity.⁴ He carried on extensive measurements on many of the castes of Northern India and published the results of his splendid work in two volumes entitled *Anthropometric Data from Bengal* (1891). The bearing of these data on the theory of caste he discussed in the introduction to his *Tribes and Castes of Bengal*, and later on in *The People of India*. His work further led to a succession of monographs on the anthropometric data from other Indian Provinces as a part of the ethnographic survey of India.
c9.4
² Risley, (2), pp. 208-5.
c9.5
³ Ibid., p. 205.
c9.6
⁴ Ibid., p. 20.
c9.7
In any anthropometric work the student is, at the very outset, faced by two problems. The first is the question of the unit. What shall we take as a unit on which measurements may be taken and compared? Shall we take a geographical or political area as our unit and compare one with another? Or shall we take a whole people, and take our measurements on them and then analyse them? The solution of this problem is fundamental to anthropometric work. But as our main purpose is to institute comparisons between castes this problem need not engage us here. We may take one caste as a unit and compare it with another. This procedure does not involve the acceptance of Risley's view about the isolation of castes which, as pointed out by his learned annotator, W. Crooke,¹ requires much modification. We may thus take our measurements on Brahmins and compare them with those on Chamars. But the data show us that neither Brahmins nor Chamars among themselves have a uniform physical type. We ask ourselves the reason of this, and we are led to the question of the origins of the different physical types. Thus we have to deal with the general ethnology of India as, indeed, Herbert Risley did. We have to analyse racial mixture as well as compare different castes.
c9.8
¹ Ibid., p. xvii.
c9.9
This procedure opens up the second problem, viz. what is the method by which we can detect racial affinities and compare groups? For such comparisons the average has long been in use. Risley principally used the average for his work, rarely bringing in the seriations and still more rarely the absolute measurements.² Seriations, like the frequency curve, only serve the purpose of showing the actual distribution and cannot be conveniently used for purposes of comparison, a fact which explains the rare use of these by Risley when dealing with the types, though they are given in the appendices. The average without the standard deviation is an abstraction which tells us almost nothing. The standard deviation is very useful both as serving the basis of the formula that we have utilized for comparison, as well as giving us an idea of the actual range of variation. "A range of six times the standard deviation usually includes 99 per cent or more of all the observations in the case of distributions of the symmetrical or moderately asymmetrical type."³ We have therefore given in the appendix the means and their standard deviations for selected castes.
c9.10
² Op. cit., p. 87.
c9.11
³ Yule, An Introduction to the Theory of Statistics, 1916, p. 40.
c9.12
It will be observed that for the Punjab and Gujarat we have very meagre data: for Sindh proper we have none. For Madras, Mr. E. Thurston¹ has given us a long list of averages but very few individual measurements. Hence in the case of Dravidian India we can compare only two castes, data for which are taken from the paper of Sir Thomas Holland. Though we accept Thurston's averages for the general ethnology of India, yet we must point out that they are very crude in so far as the absolute measurements seem to have been recorded in centimetres.² For more scientific conclusions about Dravidian India we require fresh measurements taken on typical Dravidian castes.
c9.13
¹ Vol. I, p. xi.
c9.14
When we compare two groups, what we want to know is not merely the differences between the two groups as regards the cephalic index or nasal index separately, but we want an expression which conveniently sums up all the differences in the various attributes that we may like to compare. Such an expression was used by Mr. T. A. Joyce in 1912. In order to compare two groups he starts with one character, say, the cephalic index, and works out the fraction (M₁ - M₂)/(√(σ₁² + σ₂²)) where M₁ and M₂ are the means for the cephalic indices of the two groups, M₁ - M₂ is their difference without regard to sign, and σ₁, σ₂ are the standard deviations. Similar fractions are found for as many characters as one likes, and all of them are added together. This sum is called the "Differential Index".³ The expression is a convenient quantity showing the actual differentiation of two peoples and can be fruitfully used for comparisons. Another great advantage of this expression is that by its help we can include for comparison absolutes like the cephalic length, which have, unfortunately, been too much neglected in anthropological work. Further this item is of special importance in the field of Indian ethnology; for here we have sometimes to deal with two dolichocephalic peoples whose absolute measurements of the head are essentially different. Thus the Chamar of the United Provinces has a cephalic index of 72·90, while his cephalic length and cephalic breadth are 185·18 and 134·9 respectively. The corresponding figures for the Chhatri of the same province are: 78·12, 188·35, 137·68. We have given in the appendix the differential indices for selected castes. In this chapter we have attempted to attack the problem of caste with the help of this index.
c9.15
² Ibid., pp. lv, lxi.
c9.16
³ JRAI., 1912, p. 451.
c9.17
We have said above that it was Risley who, under special circumstances, applied anthropometry to the solution of the problem of caste. But owing to the method which he followed, the light he was able to throw on the subject was not as decisive as was expected. Again he devoted more attention to the classification and origins of the various physical types of India than to the problem of caste proper.
c9.18
It may be taken to be an historical fact that peoples calling themselves "Arya" poured into India through the north-west somewhere about 2000 B.C. It is equally clear from our discussion that an institution closely akin to caste has been very often described in Sanskrit books, which are the works of either the Aryans or the Aryan-inspired aborigines. Can we trace a close connection between the immigration of the Aryans and the rise of the institution of caste? We have seen that the Brahmins, who were the moral guides and legislators of the immigrant Aryans, tried to keep their blood free from any intermixture with the lower classes, though they had no objection to the members of their own class having progeny from females of lower classes, provided these were not admitted into the Brahmin class. Can anthropometry shed any light on this aspect of caste? It appears to us that anthropometry will furnish us with a good guide, provided certain assumptions about the physical types of the Indo-Aryans and of the aborigines of Hindustan can be plausibly made.
c9.19
These assumptions are two: First, that the Aryan type may be described as long-headed and fine-nosed. The average cephalic index may be said to vary between 70 and 75 and the nasal index from 65 to 75. The ground for this assumption is that almost all the averages of the cephalic index and the nasal index given by Risley¹ for the castes of the Punjab and Rajputana fall within these ranges; and these are the regions which, from their geographical position, must have been the home of Aryan settlements. We may reasonably hold, therefore, that the predominant physical type of the Punjab and Rajputana represents the Aryan type. The second assumption is that the aboriginal type may be deduced from such peoples as the Musahar, who, not being within the pale of Hinduism, are like the jungle-tribes of Southern India. Their chief characteristic is the broad nose, the nasal index being above 80. Very often the head is long. The broad nose seems to have been noticed even by the Vedic poets as a characteristic of the aborigines, whom they sometimes describe as "noseless".¹
c9.20
¹ Op. cit., Appendix, iii, p. 896.
c9.21
¹ Vedic Index, vide Dasyu.
c9.22
We can now proceed to set out the conclusions we get from our anthropometrical inquiry. Taking the Brahmin of the United Provinces as the typical representative of the ancient Aryans we shall start comparisons with him. If we turn to the table of differential indices we find that he shows a smaller differential index² as compared with the Chuhra and the Khatri of the Punjab than with any caste from the United Provinces except the Chhatri. The differential index between the Khatri and the Chuhra is only slightly less than that between the Brahmin of the United Provinces and the Chuhra. This means that the Brahmin of the United Provinces has closer physical affinities with the Chuhra and the Khatri of the Punjab than with any caste from his own province except the very high caste of the Chhatri. The Brahmin is as much akin to the Chuhra as the latter is to the Khatri of the Punjab. The only valid conclusion is that the United Provinces Brahmin does not materially differ from the physical type of the Punjab, i.e. on the assumption, previously explained, that he fairly represents the physical type of the Aryans. The reality of this close affinity between the United Provinces Brahmin and the Punjab Chuhra is more clearly brought out if we look at the table of differential indices between the United Provinces Brahmin and the Brahmins of other regions. They are very high as compared with that between the Chuhra and the United Provinces Brahmin. Even the differential index between the United Provinces Brahmin and the Bihar Brahmin, who from what we know about the history of the spread of Aryan culture, is expected to be very nearly allied to the former, is just twice as high as that between the United Provinces Brahmin and the Chuhra. This ought to serve as a conclusive proof of the fact that the United Provinces Brahmin has essentially retained the same physical type as that of the Punjabis and of the ancient Aryans.
c9.23
² We have used only those differential indices which are based on six characters.
c9.24
We must now examine in what relation the United Province Brahmin stands to the other castes of his province. From the table of differential indices for the United Provinces we see that in physical affinity with the Brahmin, the Kayasth and the Kurmi stand next to the Chuhra, while at the other end of the scale come the Pasi and the Chamar. In the scheme of social precedence, the Kayasth belongs to the fourth rank, the Kurmi stands at the head of the eighth, while the Pasi and the Chamar take the eleventh and the twelfth rank.¹
c9.25
¹ United Provinces Census Report, 1901, p. 248.
c9.26
The true significance of this gradation in physical affinity with the Brahmin can be demonstrated by the study of the other provinces. On historical grounds we expect Bihar to approximate to the United Provinces. On referring to the table we find that the Kurmi comes near to the Brahmin, and the Chamar and the Dom stand much differentiated from him. But the Chamar in this case is not as much distinct from the Brahmin as the United Provinces Chamar is from the United Provinces Brahmin. The social status of the Kurmi is defined by stating that he ranks third while the Chamar and the Dom rank fifth and sixth.² Thus in Bihar the state of affairs in some way corresponds to that which we have tried to demonstrate for the United Provinces. The table for Bengal shows that identical conditions do not prevail there. The Chandal, who stands sixth in the scheme of social precedence and whose touch pollutes,³ is not much differentiated from the Brahmin, from whom the Kayasth, second in rank, can hardly be said to be distinguished. The gradation observed in the United Provinces is thus absent in Bengal. Still more is this the case in Bombay. Here the Deshasth Brahmin bears as close an affinity to the Son Koli, a fisherman caste, as to his own compeer, the Chitpavan Brahmin. The Mahar, the untouchable of the Maratha region, comes next together with the Kunbi, the peasant. Then follow in order the Shenvi Brahmin, the Nagar Brahmin and the high-caste Maratha. These results are rather odd.
c9.27
² Bengal Census Report, 1901, pp. 873.
c9.28
³ Ibid., p. 869.
c9.29
Stated in a generalized form they mean that there is no correspondence between social gradation and physical differentiation in Bombay. We venture to think that the results from Bengal would have been equally striking if we had data for the various sub-castes of the Brahmins. This contention is rendered plausible by some measurements published by Rai Bahadur Ramaprasad Chanda.¹ The following are the means and their standard deviations :-
c9.30
¹ pp. 255-7.
c9.31
| | *Cephalic Index.* | | *Nasal Index.* | |
| | *m.* | *St. Dev.* | *m.* | *St. Dev.* |
| Brahmin (31) | 78.84 | 3.64 | 75.61 | 10.31 |
| Paschatya Vaidik Brahmin (50) | 79.83 | 3.56 | 71.97 | 8.62 |
c9.32
With such differences between the two sub-castes of the Brahmins of Bengal we have reason to expect more startling results than we have, though these are quite insignificant.
c9.33
Our survey of the regions of India other than Hindustan has made it abundantly clear that Hindustan is unique in this respect that here we have the Brahmin at the head of the physical hierarchy; then follows a high caste or two, hardly differing from him in physical type; then comes a group of castes, slightly differentiated from the Brahmin; low castes like the Chamar and the Pasi, whom we may look upon as the Hinduized representatives of the aborigines, stand far removed from him. This is just the situation that should have resulted from the regulations, which, as we have shown above, were being promulgated by the Brahmins to avoid their class being contaminated by the infusion of the aboriginal blood. The evidence from literature is thus amply corroborated by the physical facts. Restrictions on marriage of a fundamentally endogamous nature were thus racial in origin.
c9.34
We shall now turn to the conclusions arrived at by Risley. Believing in the "marked divergence of type that distinguishes the people of the Eastern Punjab from the people of Western Hindustan", to account for the people of Hindustan he brings in a second wave of the Aryans with few or no women. They married aboriginal women and thus modified their original type; "but a certain pride of blood remained to them, and when they had bred females enough to serve their purposes and to establish a distinct *jus connubii* they closed their ranks to all further intermixture of blood. When they did this, they became a caste like the castes of the present day".¹ We have tried to indicate that as far as the published data go the Hindustani Brahmin does not materially differ from the physical type of the Punjabis. An acute observer of Upper India, W. Crooke remarked that a traveller from the Punjab glides into Hindustan without marking any change in the physical type.² It is, therefore, not necessary to postulate a second invasion of the Aryans, who could not bring their women-folk with them.³ Nor need we propose a theory entirely contradictory to the literary records of the Brahmins.
c9.35
¹ (2), pp. 56, 274-5.
c9.36
² *Northern India*, 1907, p. 107.
c9.37
³ The hypothesis of two or three Aryan inroads has been examined by the author in his paper "The Ethnic Theory of Caste", published in *Man in India*, 1924.
c9.38
There is another proposition of Risley's which we must examine more minutely, for it is one which, if true to facts, affects our conclusion. He observes, "If we take a series of castes in Bengal, Bihar, the United Provinces of Agra and Oudh, or Madras, and arrange them in the order of the average nasal index so that the caste with the finest nose shall be at the top, and that with the coarsest at the bottom of the list, it will be found that this order substantially corresponds with the accepted order of social precedence." ⁴ In the argument elaborated above it is evident that we depend upon the uniqueness of Hindustan amongst the provinces of India as regards the gradation of the physical type for our explanation of the origin of endogamy. The statement of Risley in a way challenges that basis. For Bengal and Bombay we think we have made a good case, and our figures will speak for themselves; but as regards Madras, for lack of material, we have not been able to apply the formula for differential index. It may be urged against us that as far as nasal index, without the absolute measurements of the nose, can be an indication of racial mixture, Madras seriously damages our argument. We shall therefore examine the association between nasal index and social status province by province.
c9.39
⁴ *Op. cit.*, p. 20.
c9.40.1
Risley, after comparing the nasal indices of some of the castes of Hindustan, says that the order of gradation established by means of the nasal index is " substantially " the same as that of social precedence.¹ This mild statement has been reproduced in an exaggerated form in an authoritative book on ethnology. It is said in Keane's *Man Past and Present*² that the Hindustani type of the United Provinces has " a nose index exactly corresponding to social station ". We shall presently show that Risley's statement is only partially true, and that Keane's generalization has no ground. The following is the order in which some of the typical castes of the United Provinces stand according to their nasal index as given in our table for the province : Babhan, Brahmin, Chhatri, Kayasth, Kurmi, Tharu, Bania, Bhar, Kol, Lohar, Dom, Pasi, Musahar and Chamar. The order of social precedence as given in the *Census Report of 1901*³ is as follows : Brahmin, Babhan, Rajput, Kayasth, Bania, Kurmi and Lohar, Mallah and Bind, etc., Bhar and Tharu, etc., Kol and Muchi, etc., Pasi and Dosadh, etc., Chamar, Musahar and Dom, etc. To these we may add Kachhi, Lodha, and Koiri. Risley's⁴ averages for these are: 82·9, 83·4, and 83·6. In our list the first two will rank after Kol and the third after Lohar. But in the scheme of social precedence they are grouped together with Kurmi. These comparisons prove that the statement of Keane is baseless and that of Risley is true only in a broad sense.
c9.40.2
For Bengal Risley's⁵ averages give us the following order : Brahmin of Eastern Bengal and Kayasth, Brahmin of Western Bengal, Chandal and Sadgop, Goala, Muchi, Pod, Kaibart, and Rajbansi Kochh. These castes may be arranged according to social status as : Brahmin, Kayasth and others, Sadgop and others, Kaibarta and Goala, Bhuiya and others, Kochh, and Pod, Chandal, Muchi and others.⁶ It is evident that these lists disprove Risley's statement. From Risley's⁷ averages for Orissa some of the castes rank as below : Shashan Brahmin, Panda Brahmin and Teli, Khandait, Mastan Brahmin and Guria, and Chasa, Karan, Pan, Kewat, and Bauri. The social order of precedence is : Brahmin, Karan, Khandait and others, Chasa, Guria and others, Gola and others, Teli and Kewat, etc., Chamar, Bauri and others, Pan and others.¹ We leave out Bombay, as it is not included in Risley's observation. By looking at the table the reader can satisfy himself that the figures for Bombay are more telling than in the case of the provinces so far dealt with. Finally we come to Madras. Here we must treat of the different linguistic areas separately for the schemes of social precedence in the various areas are different. We shall begin with the Telugu region. According to the averages given by Risley (p. 398) and by E. Thurston (vol. i, pp. lxviii-lxxiii) the order of castes is as follows: Kapu, Sale, Golla, Mala, Madiga and Togata, and Komati. According to their social status they are ranked as below: Brahmin . . .
c9.40.3
Komati, Golla and Kapu and others, and Sale, Togata and others. Mala and Madiga occupy the lowest rank, being the Pariahs of the Telugu country.² In the Canarese region the nasal index gives the following order: Karnatak Smarta Brahmin, Bant, Billava, Mandya Brahmin, Vakkaliga, Ganiga, Linga-Banajiga, Panchala, Kuruba, Holeya, Deshasth Brahmin, Toreya, and Bedar. In the scheme of social precedence the castes take the ranks as under: Brahmin . . . Bant and Vakkaliga, Toreya, etc., Kuruba and Ganiga, Badaga and Kurumba and Solaga, Billava, Bedar and Holeya.³ The significance of the comparison is enhanced when we remember that the nasal index of the Holeya, the untouchable of the Canarese region, is 75·1, that of the highest sub-caste of the Brahmin being 71·5, while those of the jungle Kurumba and the Solaga, who when Hinduized, occupy the rank allotted to them in the list, are 86·1 and 85·1 respectively. The ranks of the Linga-Banajiga and the Panchala cannot be determined, as they do not recognize the authority of the Brahmin. The Tamil castes may be arranged according to their nasal index as follows: Ambattan, Vellal, Idaiyan, Agamudaiyan, Tamil Brahmin, Palli, Malaiyali, Shanan, Paraiyan, Irula, and Malasar. The social status of these castes is: Brahmin, Idaiyan and Vellala, Agamudaiyan and Palli, Ambattan and others, Irula and Malasar and Malaiyali, Shanan and others, Paraiyan.⁴ The nasal indices of four typical Malayalam castes are: Tiyan, 75; Nambudiri Brahmin, 75·5¹;
c9.40.4
Nayar, 76·7; Cheruman, 77·2. The order of social precedence among these is: Nambudiri, Nayar, Tiyan, and Cheruman. The nasal index of the Kanikar, a jungle tribe of Travancore, is 84·6. Thus the Cheruman belongs to the same race as the Brahmin rather than to that of the Kanikar. We hope that these comparisons will have made it abundantly clear that the proposition of Risley has almost no basis in fact outside Hindustan.
c9.41
¹ Op. cit., p. 40.
c9.42
² p. 546.
c9.43
³ p. 248.
c9.44
⁴ Op. cit., p. 400.
c9.45
⁵ p. 401.
c9.46
⁶ *Bengal Census Report*, 1901, p. 389.
c9.47
⁷ Op. cit., p. 401.
c9.48
¹ Bengal Census Report, 1901, p. 375.
c9.49
² Madras Census Report, 1901, pp. 136-7.
c9.50
³ Ibid., pp. 138-9.
c9.51
⁴ Ibid., p. 81.
c9.52
¹ Rivers, *The Todas*, p. 708.
c9.53
Outside Hindustan in each of the linguistic areas we find that the physical type of the population is mixed, and does not conform in its gradation to the scale of social precedence of the various castes. The Brahmin is not so far distinguished in his physical features from other castes as to stand out apart from them and at the head of the physical hierarchy. Some of the castes, very low in their social status, actually stand higher in physical features than some of the higher castes. In Hindustan, on the other hand, the gradation of physical types from the Brahmin downwards to Musahar corresponds very closely to the scheme of social precedence prevailing among the Hindustani castes. This state of things can be the result only of such regulations that prevented the possibility of Brahmin blood being mixed with aboriginal blood but allowed the mixture of blood of the other groups in varying proportions. As we have seen, this was just what the Brahmins attempted to do by their rules of conduct. The Brahmin kept himself pure by decreeing that only those persons both of whose parents were Brahmins were to be regarded as Brahmins. Thus the Brahmins were the first to be endogamous with reference to their class. We have also seen that the Brahmins of each of the linguistic areas show greater physical affinity with other castes of their region than with the Brahmins of other areas and of Hindustan. It is clear from this that the Brahmanic practice of endogamy must have been developed in Hindustan and thence conveyed as a cultural trait to the other areas without a large influx of the physical type of the Hindustani Brahmins.
c9.54
The treatment of the subject cannot be complete without an investigation of the physical types of India and their origins.¹ In this field also a lead has been given by Risley. It is best to start with a statement of his view. He distinguishes seven types: (1) The Turko-Iranian of Baluchistan and North-West Frontier Provinces. We do not discuss this type, because we think that it essentially lies outside cultural India. (2) The Indo-Aryan type of Kashmir, Punjab, and Rajputana. It is dolichocephalic with a narrow nose. We leave out other physical characteristics as they are far too vague. (3) The Mongoloid type of the Himalayas, Nepal, Assam, and Burma. (4) The Dravidian type extending from Ceylon to the valley of the Ganges and pervading the whole of Madras, Hyderabad, and the Central Provinces, most of Central India and Chhota-Nagpur. It was " probably the original type of the population of India ". It is dolichocephalic with a " very broad nose ". (5) The Aryo-Dravidian type of the United Provinces, Bihar and Ceylon. It is also found in parts of Rajputana. The head-form is long with a tendency to medium. (6) The Mongolo-Dravidian type of Lower Bengal and Orissa. " The head is broad; nose medium with a tendency to broad." (7) The Scytho-Dravidian type of Western India. " The head is broad; nose moderately fine."²
c9.55
¹ In our discussion of the physical types we have confined our remarks to the Hindu population.
c9.56
² Risley, (2), pp. 83-4.
c9.57.1
The Indo-Aryan type: The first objection against this type is that the term used to denote it in the scheme proposed by Risley is unsatisfactory. In the case of the other types the names are compound terms composed of the names for the two physical types from which the particular type is supposed to be derived. Thus what is meant by the term Aryo-Dravidian is that the type is considered to be a mixture of the Aryan type and the Dravidian type. But in the case of the Indo-Aryan type no such mixture is meant. The Indo-Aryan type is not to be understood to be a mixture of the Indian and Aryan but to be a pure type. Indo-Aryan means nothing more than the Aryan of India as opposed to the Aryan of Iran. Thus it will be seen that in Risley's scheme the term is rather misleading. Nevertheless, as in our treatment of the subject we wish to avoid all compound terms we shall retain the term Indo-Aryan to denote the Aryan type of India. Instead we cannot use the term Aryan, for in philology it is generally applied to the Iranians as well as the Indians. By Indo-Aryan, then, we mean the physical type to which the Vedic Indians largely contributed. The physical characteristics of this type are solely inferred from those of the present population of the Punjab. We have no reason to think that after the Aryan immigration into India and before Darius' dominion there was any considerable influx of foreigners.
c9.57.2
When Darius held sway over a part of the Punjab the country appears to have been already very populous and prosperous.¹ The physical type of the Punjab is so uniform as to preclude any possibility of large mixture. Hence we can reasonably assume that a large bulk of the present inhabitants of the Punjab are the descendants of the Vedic Aryans. Their number might have swollen by later immigrants whose physical type must have been similar to theirs but whose cultural affinities we may not be able to trace. Secondly, an objection has been levelled at Risley's classification of the Rajputs as Indo-Aryans.² It has been contended that some of the clan-names of the Rajputs are the same as those of the Gurjara tribe or caste and that the reigning dynasty of Kanauj during the ninth century and after was of the Gurjara stock. Hence it is argued that some of the Rajput clans are of non-Aryan descent.³ Even the Minas and Meos of Rajputana are not physically much differentiated from the Rajputs.⁴ The Rajputs are thus of the same race as the other peoples of Rajputana. We cannot hold that Rajputana was untenanted before the foreign tribes entered it in the sixth century A.D. Perhaps it will be said that the Gurjaras were of the same physical type as the population of Rajputana and hence their presence cannot be detected by physical anthropology, though they had come in large numbers and produced the present population of Rajputana. This supposition is not tenable;
c9.57.3
for the Gurjaras are believed to be allied in blood to the White Huns,⁵ who seem to have been a brachycephalic people.⁶ It is reasonable therefore to assume that a few ruling families of the Gujaras with a small band of followers succeeded in gaining ascendance and imposing their culture over the then population of Rajputana. It is also urged that some of the Rajput clans "are descended from the so-called aboriginal tribes—Gonds, Bhars, Kols and the like",¹ apparently because both the Rajputs and the aboriginal tribes have certain clan-names in common. In physical characteristics these tribes are sharply distinguished from the Rajputs, and yet we are asked to believe that they are essentially the same because some features of social organization are common to both. The common clan-names can better be explained as a borrowing by the lowly castes or tribes from the dominant Rajputs. Thus there does not appear to be any serious objection to the Rajputs being classed as Indo-Aryans. Thirdly, there does not seem to be any foundation for Risley's classification of the Kashmiris as Indo-Aryans. The Kashmiris appear to be a mesaticephalic people, perhaps a mixture of the Alpine folk of Central Asia and the Indo-Aryans of the Punjab. Lastly, we have made it clear that Risley was not right in restricting the type to the Punjab and Rajputana. The type includes one or two high castes of Hindustan as well.
c9.58
¹ Cambridge History of India, vol. i, pp. 385-7.
c9.59
² Risley, p. xx.
c9.60
³ V. A. Smith, Early History of India, 1914, pp. 411-14.
c9.61
⁴ Risley, p. 396.
c9.62
⁵ Smith, op. cit., p. 321.
c9.63
⁶ Keane, p. 320.
c9.64
¹ V. A. Smith, *op. cit.*, p. 413.
c9.65
As regards the affinities of Indo-Aryans outside India, Dr. Haddon would perhaps see in them modified Proto-Nordics.² But it is better to connect them with the tall variety of the Mediterranean Race.
c9.66
² *Wanderings of Peoples*, pp. 26-7.
c9.67
The Mongoloid type: Here we have no comment to make except to draw attention to a mistake of Risley,³ repeated in *Man Past and Present*.⁴ Both the Kanets of Kulu and Lahoul are there classed as Mongoloid. A glance at the table for the Punjab will show that whereas the Kulu Kanets are very much akin to the Khatri of the Punjab, the Lahoul Kanets are quite distinct from them. The Kanets of Kulu are Indo-Aryans, while the Kanets of Lahoul are a mixture of the Kulu Kanets with the Mongoloid folk. Indeed, it was to elucidate the process of racial mixture between the Kanets of Kulu (who show marked points of contact with the Punjabis) and the Mongoloids that Holland, our authority on the subject, undertook a study of these two peoples.⁵
c9.68
³ p. 84.
c9.69
⁴ p. 547.
c9.70
⁵ *JRAI.* 1902, pp. 114, 120.
c9.71
The Dravidian type: Here we come upon the weakest part of Risley's great work. The type is supposed to spread from Ceylon to the valley of the Ganges, the best representatives of it being Nayars, Paniyans, Santals, and others.¹ This view has been repeated in *Man Past and Present*,² where though the jungle tribes like the Paniyans are classed apart as Pre-Dravidian, the Nayars and the Santals are given as the typical representatives of the Dravidian type. Now even if we leave aside complexion, the Santal stands as far apart from the Nayar as the Paniyan. The nasal index of the Nayar is 76·7; that of the Santal, 88·8; that of the Paniyan, 95·1. The close connection of the Santal with the jungle tribes of Southern India comes out more clearly if we take into account other tribes like the Irula and the Kanikar. It may be broadly stated that the nasal index of the jungle tribes ranges from 80 to 90. We have therefore to connect the Santal with the jungle tribes and separate the Nayar and others like him from them.
c9.72
¹ Risley (2), pp. 34-5.
c9.73
² p. 347.
c9.74
We propose to break up this composite group into four distinct types:—
c9.75
(a) The Pre-Dravidian type.
(b) The Munda type.
(c) The Dravida type.
(d) The Western type.
c9.76
Before dealing with the types we must make a note about the use of certain terms. The Malayalam and the Tamil regions seem to have been collectively referred to by Sanskrit writers as "Drāviḍa deśa", i.e. the Dravid country, as opposed to the "Āndhra deśa" or the Telugu country. The type that we propose to designate Dravida is predominant only in the Malayalam and the Tamil regions. Hence it should be deservedly called the Dravida type. "Dravidian" is only an anglicized form of the word "drāviḍa". But in philology it has been used to denote not only Tamil and Malayalam but Canarese, Telugu, and kindred languages. It is also employed to designate a culture different from the Aryan culture. We think that it was developed by the interaction between the peoples of the Dravida type and those of the Western type. It is convenient to retain this word to denote a culture which in its social organization was characterized at one end by such practices as cross-cousin marriage and at the other by dual organization and matrilineal descent.
c9.77
(a) The Pre-Dravida type: The characteristic representatives of this type inhabit mostly the jungles of Southern India. They are also found in Western India, in the hilly country of Central India, in Rajputana, and the United Provinces, everywhere penetrating like a wedge. Such a distribution makes it probable that these peoples were the first occupants of Southern India. Being pressed by later immigrants they seem to have taken to the hills and jungles, or again managed to become low members of the social polity of the immigrants. The immigrants who pressed upon them were the people of the Dravida type. Hence the name of the type under consideration. If we regard the Irula, the Kadir, the Kanikar, the Paniyan and the like as the best representatives of this type, its physical characteristics may be thus summed up: The head is long, the cephalic index being generally below 75; the nose broad, the nasal index being always above 80. The Bhil and the Katkari of Western India are members of this racial stock. The Musahar, the Pasi, the Chamar, and the low castes in general of the United Provinces belong to the same type. From the fact that among the Kanikars of the hills the system of inheritance is from father to son, only a portion of the personal property being given to the nephew, while among those of the plains an equal distribution of one's self-acquired property is made between the sons and the nephews,¹ it is reasonable to assume that the pre-Dravida people were essentially patrilineal. When they moved northwards they seem to have imbibed many of the items of the Dravidian culture. They carried with them the practice of raising megalithic tombs and monuments, which are found in the Central Provinces, Eastern Rajputana, and the Mirzapur District of the United Provinces. They spoke Dravidian languages; for some of the names of the villages in the Central Provinces end in a Canarese termination² and some village-names in the districts of Gaya and Mirzapur are distinctly Dravidian.³ As for social organization they seem to have carried the practice of cross-cousin marriage, which now exists among them for the most part as a survival.¹
c9.78
¹ Thurston, op. cit., iii, p. 169.
c9.79
² Central Provinces District Gazetteers: Chanda, p. 128.
c9.80
³ JASB. 1903, pt. iii, pp. 92-3.
c9.81
¹ JRAS. 1907, p. 626.
c9.82
This type has till now been known as pre-Dravidian. Some authorities connect these people with the Sakai of the Malay Peninsula and the Australians.² They might have been the autochthones of India. That there is a negrito strain in this type is rendered clear by the researches of Dr. B. S. Guha.³
c9.83
² Keane, p. 422.
c9.84
³ Nature, 1929, pp. 942-3.
c9.85
(b) The Munda type: Peoples having a very slight tendency towards mesaticephaly and a broad-nose, nasal index above 80, are massed together in Chhota-Nagpur and Western Bengal. They are also found in Bihar, but not in the United Provinces.⁴ In this region Munda culture seems to have originally flourished.⁵ Hence it is best to designate the type as the Munda type. Some of the peoples belonging to this type speak Indo-Aryan and Dravidian languages. In culture the Munda peoples have affinities with the Indonesian and the Melanesian regions. That this Munda culture had a far wider distribution than at present is certain because "Munda languages must once have been spoken over a wide area in Central India, and probably also in the Ganges Valley".⁶ Further, some of the names of villages in the districts of Gaya and Mirzapur are Munda in origin.⁷ The Munda languages are closely related to the Mon-Khmer languages spoken by the Sakais and Semangs of the Malay Peninsula and are further connected with the languages of Indonesia, Melanesia, and Polynesia in such a manner as to allow us to group them together as one family.⁸ A survival of a peculiar custom, viz. the marriage of grandfather with his granddaughter, is confined only to this part of India. The existence of it among the Oraons was first made known by Rai Bahadur S. C. Roy. A counterpart of it is known from Melanesia.⁹ Roy has also noticed certain practices from Bihar which are reminiscent of this survival.¹⁰ Another sort of survival of this curious custom is reported from Ganjam, where even now a Munda language is spoken.
c9.86
⁴ Risley, pp. 399-400.
c9.87
⁵ Haddon, p. 26.
c9.88
⁶ Linguistic Survey of India, iv, p. 9.
c9.89
⁷ JASB. 1903, pt. iii, pp. 92-3.
c9.90
⁸ Linguistic Survey of India, iv, p. 14, and f.n. p. 21.
c9.91
⁹ S. C. Roy, The Oraons, pp. 352-4.
c9.92
¹⁰ Ibid.
c9.93
Among the Sollokhondia section of the Gaudo caste if a girl fails to secure a husband before puberty she has to go through a nominal marriage with an old man, " preferably the girl's grandfather."¹ A similar practice prevails amongst the Kurumo of Ganjam.² The custom of marriage between the granddaughter and the grandfather must have formed part of the Munda culture. A neolithic artefact known as the " shouldered celt " is peculiar to the highlands of Bengal and Assam. Identical tools are found in the Malayan region.³ The limited distribution of this implement in India lends support to the conclusion that the people who brought it to India must have come from the Malayan region. Further as the tool is not found west of the area characterized by the Munda culture it is reasonable to suppose that the Munda people were responsible for the introduction of this artefact. The conclusion about the origin of the Munda type is that the people of this type came into India from the Malayan region when they were in the Neolithic Age.
c9.94
¹ Thurston, op. cit., ii, p. 270.
c9.95
² Ibid., vol. iv, p. 170.
c9.96
³ Journal of the Department of Letters, Calcutta University, i, p. 165.
c9.97
The typical representative of this type are the Munda, the Santal, the Musahar, and the Chamar of Bihar.
c9.98
(c) The Dravida type: This type is characterized by a long head, the cephalic index being below 75, and by a medium nose, the nasal index being less than 77. It is restricted to the Tamil and Malayalam countries.⁴ Matrilineal descent seems to have been the chief characteristic of the social organization of the peoples of this type. A statement of Megasthenes is construed to refer to this peculiarity of the social organization of Malabar.⁵ A physical trait which we have not so far taken into consideration is the hair. It is of special importance in connection with the inquiry of the affinities of this type. Most castes of Malabar excepting the Nambudiri have little or no hair on the cheeks and the chest. They are clean-shaven, and the Nambudiri imitating them shaves all his body excepting the top-knot. " Amongst the people of good caste in Malabar, to speak of one as a hairy man is to speak of him reproachfully."⁶ This lack of hair on the cheeks definitely connects the Dravidas with the Brown Race as it is characterized by Elliot Smith.¹ When they came to India from Mesopotamia or Arabia, cannot be ascertained. They might have brought with them the matrilineal type of social organization from these countries; for there is some evidence for thinking that both the Arabs and the Babylonians practised mother-right.²
c9.99
⁴ Thurston, op. cit., p. 41.
c9.100
⁵ W. Logan, Malabar, i, p. 247.
c9.101
⁶ Madras Museum Bulletin, vol. iii, p. 38.
c9.102
¹ The Ancient Egyptians, pp. 50-62.
c9.103
² H. Schaeffer, The Social Legislation of the Primitive Semites, p. 3.
c9.104
The best representatives of this type are: The Nayar, the Tiyan, the Badaga, the Agamudaiyan, and the Vellala. We exclude the Nambudiris from this type because their cephalic index is above 75, and they are extremely hairy. The Todas are a problem. Their cephalic and nasal indices are such as to include them under this type; but their hairy system is particularly well developed. Rivers connects them with the Malabar castes and attributes their hairiness to their environment.³ Against this explanation R. Chanda has justly urged that the Badagas, another tribe of the Nilgiris, do not seem to be hairy.⁴ In view of the fact that "of all the castes or tribes of Malabar, the Nambudiris perhaps show the greatest number of resemblances to the customs of the Todas ",⁵ it is not unlikely that they might have some connection with the Nambudiris.
c9.105
³ The Todas, p. 708.
c9.106
⁴ Journal of the Department of Letters, Calcutta University, vol. viii, p. 300.
c9.107
⁵ The Todas, p. 709.
c9.108
(d) The Western type: We shall deal with this type in connection with the Scytho-Dravidian type of Risley.
c9.109
The Aryo-Dravidian type: As the discussion so far must have made it clear we cannot speak of any Aryo-Dravidian type. We have shown that the high castes of the United Provinces (and perhaps also of Bihar) must be classed as Indo-Aryan. The lower castes of the United Provinces must be referred to the pre-Dravidian type and those of Bihar to the Munda type. That we must separate the low castes of Bihar from those of the United Provinces is made clear by the differential index. The differential index between the Brahmins of the two provinces is 2.05; that between the Kurmis is 8.49; and that between the Chamars is 4.01. The intermediate castes of the United Provinces must be considered to be the result of a mixture between the Indo-Aryan and the pre-Dravidas, while those of Bihar that of the Indo-Aryans and the Mundas, with perhaps a pre-Dravida strain.
c9.110
The Mongolo-Dravidian type: This type, according to Risley, is a mixture of the pure Mongoloid and Dravidian types and comprises the population of Lower Bengal and Orissa. He describes the head of this type as broad, but we think it would be better to describe it as medium. Only the Rajbansi Maghs of the Chittagong Hills show a cephalic index of over 79; and they are essentially Mongoloid. In Bengal the nasal index, with very few exceptions, is below 78.¹ The nasal index of the Mongoloid type in India is in most cases above this figure.² Nor does the orbito-nasal index support any idea of a Mongoloid mixture. In the case of the Mongoloid people it is below 110.³ In three lower castes of Bengal—the Pod, the Bagdi, and the Chandal—the figures are: 111.5, 112.2, and 114.0. In Orissa the orbito-nasal index varies from 112 to 117. We do not wish to imply that there is no Mongoloid mixture in this part of India, but only to point out that the published data of anthropometry do not give us any clue. The Savaras of Ganjam are distinctly Mongoloid.⁴ Again some of the castes of the United Provinces show marked cultural affinities with the Mongoloid peoples. Among the Bahelias, Dhangars, Dharkars, Dombs, Dusadhs, and Nais, marriage is permitted in the line of one's mother's sister.⁵ Everywhere else in India such marriages are strictly prohibited; hence the permission for such marriages among these castes is very unusual. We can explain it only as a borrowing from the Mongoloid people, the Bhotias of Sikkim. Among them one can marry one's cousin on the maternal side, either mother's brother's child or mother's sister's child.⁶ We think that the Bengalis are only an extension of the Western type. The differential index shows that the Bengal Brahmin is more nearly related to the Shenvi and the Chitpavan Brahmins of Bombay than to the Bihar Brahmin and far more so than to the United Provinces Brahmin. There is evidence to show that the people of Bengal have affinities with the Dravidian culture. The use in marriage of the Chank-bangles even by the Brahmins is an instance in point.¹ Risley thinks that there is "a strain of Indo-Aryan blood in the higher groups ". ² In view of the fact that the differential index between the Bengal Brahmin and the Chandal is 1.11, while between him and the United Provinces Brahmin is 3.89, there does not seem any ground for this supposition. In Orissa there appears to be some mixture of the Munda type, because the nasal index is sometimes very high. The people of Bengal and Orissa, therefore, would best be regarded as a mixture of the Western and the Munda types slightly modified in some cases by the Mongoloid type.
c9.111
¹ Risley, p. 401.
c9.112
² Ibid., p. 31.
c9.113
³ U.P. Census Report, 1911, p. 212.
c9.114
⁴ Ibid., p. 402.
c9.115
⁵ Thurston, op. cit., vi, p. 312.
c9.116
⁶ Bengal Census Report, 1911, p. 320.
c9.117
¹ J. Hornell, *The Sacred Chank of India*, pp. 145-6.
c9.118
² Op. cit., p. 33.
c9.119
The Scytho-Dravidian type: What Risley exactly meant by the term Scythian is not clear. He could not have meant Mongoloid, for then there would have been no point in distinguishing this type from the Mongolo-Dravidian. Perhaps he used the word as equivalent to the Śakas of Sanskrit writers. That there is no evidence of the Śakas having reached so far as this type stretches is amply proved.³ We must, therefore, give up this designation of the type. We propose to call it the Western type, because it characterizes the Western Coast from Gujarat to South Canara, and thence spreads inwards into Coorg, Mysore, the Deccan, and the Telugu country and through Orissa into Bengal. The head is mostly medium with a slight tendency towards broad; the nasal index in most communities is below 78. Representatives of this type are: Sale, Bant, Vakkaliga, Coorga, Shenvi, Prabhu, Nagar, Chitpavan, Mala, Madiga, and Holeya. Chanda seeks to connect the Gujaratis, the Marathis, and the Bengalis with a people of the Alpine Race that, he supposes, came from Chinese Turkistan. And he has the support of no less an authority than Dr. Haddon, who first postulated an immigration of the Alpine folks to account for the "strongly marked brachycephalic element in the population of Western India".⁴ Against this route of immigration it must be objected that if the Alpine people came through Kashmir and the Punjab, how is it that we have no trace of either brachycephaly or mesaticephaly in the intervening area till we come to Gujarat? Further, the highest cephalic indices are recorded from the southern part of the region characterized by this type, i.e. from Bellary. The distribution of the type sketched above suggests quite a different route of immigration, i.e. on the western coast by sea. Most of the castes are mesaticephalic and nine of the castes of the Marathi region, pooled together, give 77.84 as the mean cephalic index. It is therefore more reasonable to seek the origin of this type in a mixed stock rather than in a purely Alpine one, for the latter has a very broad head, the cephalic index being often 85 and upwards.¹ Western Asia seems to provide us with such a people; for there we notice a prolonged process of mixture between the Brown race and the Alpine race.² Dr. Schoff sees in Cutch or Kachh reminiscences of the Kassites who migrated from Elam about 2200 B.C.³ Mr. H. J. E. Peake thinks that about 2000 B.C. some of the Nordics had made their appearance in Asia Minor.⁴ If some of them accompanied the people who landed on the western coast of India, we can explain the colour of the eyes of the Chitpavans of Bombay. It is greenish grey rather than blue. Their eyes are known in Marathi as "cat-eyes".⁵
c9.120
³ Risley, (2), p. xx.
c9.121
⁴ Op. cit., p. 27, and Map 1.
c9.122
¹ Keane, p. 438.
c9.123
² The Periplus of the Erythrean Sea, p. 131.
c9.124
³ JRAI. 1916, p. 172, and Map 2.
c9.125
⁴ Bombay Gazetteer, vol. x, p. 111, f.n. 3.
c9.126
⁵ Journal of the Department of Letters, Calcutta University, v, p. 226.
c9.127
Giuffrida-Ruggeri has suggested the following classification of the ethnic elements in India: (a) Negritos. Apparently there are no representatives of these. He thinks that they survive in tribes like the Bhils, the Gonds, etc.⁶ (b) Pre-Dravidians (Australoid-Veddaic); (c) Dravidians. They are connected with the Ethiopians, with the exception of the Somalis and Gallas.⁷ Under this category he includes the mesaticephalic and the brachycephalic peoples of the Canarese-Telugu regions.⁸ Being afraid of sea migration—witness his gibe at Elliot Smith⁹—he brings the Dravidians into India from Iran, and explains the "elevation of the cephalic index among the inhabitants of the south by a mixture with the brachycephalic Negritos, the most ancient population between India and the Persian Gulf".¹ He entirely misses the significance of the distribution of the dolichocephalic and mesaticephalic types explained above. (d) Tall dolichocephalic elements (Toda),² (e) dolichocephalic Aryans. Under this class he includes Kafirs, Dards, Rajputs, and the Kanet of Lahoul.³ (f) Brachycephalic Leucoderms. He accepts the hypothesis of Chanda,⁴ about immigration of the Alpine folk. It will have been quite clear from our discussion that such a classification is untenable. The ascertained facts of Indian anthropometry are far better accommodated by the classification that emerges out of our discussion.
First 131 of 143 passages. Read on