Michael Roberts, responding in 1985 to a Review Essay by Susan Bayly of Cambridge University on his book on Caste Conflcist and Elite Formation, CUP 1982
Susan Bayly** has done me the honour of reviewing the book on Caste Conflict and Elite Formation: The Rise of a Karava Elite in Sri Lanka, 1500-1931 at considerable length.’ Her essay is appropriately entitled ‘The History of Caste in South Asia’. This title provides a clue to the interpretative pathways which have led her systematically to misunderstand the arguments within the book. No less problematical is her implicit belief in the possibility of constructing a composite picture of the caste system qua system on the basis of empirical data drawn from different regions, regions as widely different as Sri Lanka, southern India and western India. Let me elaborate this charge, and in doing so reiterate the arguments which I presented.
Bayly’s essay concentrates on a subsidiary theme in my book, its ‘treatment of the caste system’; and does so in the light of studies by Frank Conlon, Michael Moffatt, Arjun Appadurai and Carol Breckenridge.’ She takes issue with the use of dramatic language which, in effect, presents ‘the essence of the caste system [as] conflict between competing interest groups’ (p. 522). She points to the findings of Appadurai and Breckenridge: how a complex system of honours is operative in south Indian localities, wherein low caste groups ‘jostle for a higher place in local ranking schemes but at the same time . . . have a stake in the utsavam’ or temple festivals (p. 522). These findings show that the caste system could ‘accommodate continual reshuffling’ and was marked by common values which sustained interdependency and solidarity.
She also objects to the inferences and assertions upon which the idea of a Karava community and its collective resentments has been constructed (pp. 521, 525). In marked contrast with Conlon’s study which is attentive to the changing form of the Chitrapur Saraswat Brahman identity in the period 1700 to the 1930s, she (p. 521) finds that ‘Roberts’ approach to the question of the Karavas’ origins is quite different He insists that the Karava possessed a “solidarity” and “primordial identity . . . inherited from the past”‘ (p. 212).
It is remarkable that in presenting such a review Susan Bayly was able to dispense with the concept of ‘structural marginality’ which is central to the argument of the book and has reference to the caste system in Sri Lanka. The concept is borrowed from Bruce Kapferer;3 in minimal terms it refers to a section of a population that is ‘routinely and systematically exposed to contradictory processes’ (pp. 15-16, 233-8, quoting Kapferer). As applied to the specificities of the Sri Lankan material, it has been renamed ‘wedge marginality’. Using empirical indications from the early colonial period it was suggested that the Karava, Salagama and Durava castes (KSD in shorthand) were placed in a structurally marginal position within the Kingdom of Kotte in pre-colonial times as well as the early colonial period. In making this claim, not only was attention devoted to the role of the Sinhalese state in the organization of caste, but an explicit contrast was drawn between the relationship of the KSD castes with the numerically-preponderant and ritually superior Goyi-gama caste in comparison with that of the other non-Goyigama castes (pp. 1–2, 15-16, 138-40, 233-4 I , 285–93). On this basis it was suggested that some KSD caste families not only had the incentive and the leeway to make use of the new opportunities that arose in the colonial era, but that each caste was also subject to contradictory tendencies: on the one hand, an inclination to set themselves up as a culturally distinct and relatively autonomous body of people and, on the other, involved in a process of integration into the heterogeneous corpus of Sinhalese culture.
By the last decades of the eighteenth century the latter, centripetal movement had gained the upper hand, but the centrifugal tendency continued to show its hand (pp. 38-14o, 163-4, 285, 291-3). The most significant expression of this ambivalence in goal was the establishment of distinct monastic fraternities by separate parties of KSD noviciates and lay supporters who visited Burma between 1799 and 1812 in order to receive higher ordination. This constituted a challenge to the Goyigama monopoly of higher ordination and set in train a number of changes in the organization of the Buddhist religion in the low-country districts. These developments provided these regions and these fraternities with the capacity to mount opposition to Christianization and Westernization from the mid-nineteenth century. One unintended consequence of this Buddhist revitalization movement was the further integration of elements of the KSD population into the Sinhalese body politic (pp. 138-40, 285-93).
This was but one illustration of the transformative potential of forces embedded in structurally marginal locations (see pp. 165-71, 291 for other illustrations). Implicit to this argument, too, is the indication that the structural marginality of the KSD became less pronounced by the twentieth century, if not earlier. In short, the emphasis throughout is on a changing set of relations—an emphasis that is underlined by descriptions of the changing patterns of colonial rule under the Portuguese, Dutch and the British.
Bayly’s failure to incorporate this argument into her review of the book suggests that she has totally missed the significance of the state in the organization of caste in both the Sinhalese kingdoms and the colonial states (despite considerable emphasis on this point on my part, pp. 4-5, 39, 49, I 4off, 162-4, 207). In this failure she has perpetuated a tendency that has been so prevalent among village-level anthropological studies in India.4
When she does refer to my general argument about the partial integration of the Karava within Sinhalese Buddhist society at the beginning of the sixteenth century, Bayly relegates the reference to a footnote; and adds:
“One problem here is that if the Karava were an exceptional group, outsiders who were not fully absorbed into the Sinhalese social order even by the nineteenth century, then it is hard to see them as a general case which can be applied to a discussion of caste structure and ideology among mainstream Hindu castes in India” (p. 523, emphasis added).
The last sentence is both revealing and partially valid. It reveals Bayly as a worshipper in the cult of British empiricism, a seeker after an overarching generalization constructed out of a number of empirically-rooted case studies. It is also valid in that my story of the Karava cannot be used as a building-blocks in the clarification of the caste structure among mainstream Hindus. It was never intended that the story should be employed in this manner.
When I ventured into a discussion of caste in British India,’ therefore, it was with a view towards understanding ‘the nature of caste consciousness and communal solidarity’ among mobilized caste formations;‘ and the nature of the links between elites and masses (p. 18o). My criticisms of David Washbrook were not based on the deployment of Sri Lankan case material. In part the criticism was based on Washbrook’s own data—and it is heartening that Bayly accepts my complaint that Washbrook underestimates the role of religious experience in the promotion of cohesion among certain castes (p. 526). In the main, however, I was critical of the Namierite transactionalism and the materialist determinism which infused his writings. In opposition I argued for greater attention to the cultural order, including the force of ‘cultural typifications’, a phenomenological concept derived from Kapferer and Tiryakian (see p. 197 and Bayly, p. 526).
It was in this light that I proceeded, in the rest of chapter seven, to review the question of caste consciousness and caste solidarity among the Karava. The chapter as a whole, and the latter half in particular, must be understood within the broader framework of the book. In chapter five one finds detailed illustrations of the economic advances of a Karava elite in the British period (I 796-1931). At its conclusion the reader is warned (p. 31) that one is located at a junction which forks out in two directions. Chapters six (pp. r 3 r-79) and seven (pp. 18o-224) constitute one of these branches. Chapter six treats the socio-political activities of these emergent Karava elites in British times, while chapter seven surveys the degree to which one could consider the Karava to `have possessed a sense of community before the Karava elite developed into an articulate force’ that is, before the pamphleteering and other thrusts in the late nineteenth century which are described in chapter six. As I noted (p. 132), these discussions in chapters six and seven were (and remain) also of instrumental value to our understanding of the factors that were responsible for the emergence of a Karava elite—my principal concern and the second of the branches reached by page 131, a branch to which readers are returned in chapters eight and nine.
In brief, chapter seven, the focus of Bayly’s essay, is largely centred upon British times, effectively the nineteenth and twentieth centuries. This did not prevent selective incursions into previous centuries. Nor should it prevent readers from making cautious extrapolations backwards. But it is productive of misunderstanding for Bayly to link my reference to ‘the primordial identity which the Karava had inherited from the past’ on page 212 with the question of the origins of the Karava in pre-colonial times. Here, on page 212, the statement has to be read in terms of the immediate past, just before the development of class polarization within the Karava caste during British times. It has also to be understood in terms of the special sense in which I use the word `primordial’, viz with an explicit disclaimer which denies the immutability that is normally attached to the term (p.181n.). In parenthesis, it might also be noted that the concept of a ‘cultural typification’ was also used in this manner—with the understanding that it could be altered and reconstituted (see pp. 197, 200, 219-20).
Part of Bayly’s problem is that she is looking for empirical data on the self-perceptions of the Karava at various points of time between 1500 and 1931; and for chronologically organized details on the persons who actually populated the term `Karava’. This sort of data I simply have not been able to locate. To the best of my knowledge, the source material on pre-British times does not match the type of data which Conlon ferreted out. Partly for this reason, but largely because it is a superior procedure, I tackled the issue in a different manner: by pinpointing the structural forces which contributed towards the constitution and reproduction of caste identities in the Low-Country districts of Sri Lanka in British times. These forces were so numerous that even the summary takes up a full page (pp. 219-20). By way of illustration the factors include the kin-focused, centripetal marriage principles favoured by all the Sinhalese as well as the institutionalization of caste by the Sinhalese and colonial states. In several instances the argument is inferential and explicitly couched in language which makes this clear.’ In sum, however, the arguments amount to a powerful case for the contention that by the early nineteenth century the Karava elite could not escape their Karavaness.
Since several of these factors held true for earlier centuries, especially the eighteenth, it is possible to extend the claim backwards and speculate that caste cohesion also existed then. It is not entirely insignificant, after all, that several conversions to Christianity during the sixteenth century were mass conversions encompassing large bodies of Karava under their headmen (pp. 29-30); nor it is insignificant that the kings of Kotte, the Portuguese and the Dutch used caste as an organizing principle in the extraction of labour services (rejakariya). The centrality of labour appropriation in the relations between the people and the state in Sri Lanka needs to be appreciated by Indianists. The situation appears to have differed markedly from that in many parts of India where surplus appropriation involved a land tax and permitted the intercession of intermediary revenue farmers and, thereby, tended to reproduce the power of dominant castes in each locality.’ This is no small matter: we are in need of a comparative study which examines the different modes of surplus appropriation in South Asia with an eye on their implications for both the patterns of local dominance as well as caste relations.
This, then, is one of the major points of difference between us. Bayly insists on having a profile of the caste over time its membership, its occupations, its institutions and its self-perceptions. While not denying the usefulness of such ethnographic information if it is available, I attach greater weight to an analysis of the structural relationships and cultural conceptions which sustain categorical boundaries and self-perceptions. Let me highlight this difference by a brief consideration of a work which attracts Bayly’s approval, Conlon’s study of the Chitrapur Saraswat Brahmans (CSBs), themselves originally a segment of the Gaud Saraswat Brahmans (GSBs).
Conlon provides a lucid description of the adaptations within the CSB caste from the seventeenth century to 1935. He subtly delineates the interaction between the socio-economic advancement of CSB families and the development of a corporate identity centred upon a matha (monastery) headed by a line of swamis an interaction which was not without tensions and disputes. In sum, a reader finds that ‘a caste is what a caste does’. But, as Ashin Das Gupta continues, do we know why a caste does what it does?” For this we need analytic attention to the political economy and the cultural order. While Conlon presents some information on the political and economic developments in western India, he tends to restrict his focus to the manner in which British administrative practice influenced the structure of opportunity for the CSBs (e.g. pp. 49-62, 213-14). Hence Das Gupta’s question. There is only fragmentary information on the relationship between the CSB and other castes, and on the relationship between the GSB and the indigenous states in the seventeenth—eighteenth centuries.
Yet, as Fredrik Barth has stressed, a critical aspect in the moulding of a collective identity among a set of people is their interface with other social clusters or groups.” It is significant that the establishment of a matha by some proto-CSBs (that is, GSBs) in the early eighteenth century occurred in a context in which all Brahmans were expected to have a matha.12 What is more, though only passing significance is attached to this fact, the indigenous Brahmans of western India kept the immigrant GSBs at a distance and expressed doubts about their brahmanical status (pp. 16, 214). Their relations with other Brahmans clearly influenced those GSB segments who hived out a clear, sectarian niche for themselves in the eighteenth century. But we are provided with limited information on this subject.
It is also noted that allegiance to a matha and its swami was a precondition for recognition as an endogamous group by the ruler of the territory (Conlon, p. 9). Such a political code, and the manner in which the location of the CSBs’ principal matha appears to have been altered in line with the changes in political boundaries,13 raise a host of questions regarding the relationship between the GSBs/CSBs and the fluctuating state system in pre-British times.
Because my book concentrates on attitudes to structural relationships as they altered over time, readers are, I think, better equipped to decipher the implications of one of the outstanding events in Sri Lanka’s history: the establishment of five new lines of pupillary succession within the Sangha at the beginning of the nineteenth century (as the ‘culmination’ of late eighteenth-century developments).” Readers can ponder over the reasons why this movement appeared at this point of time, why the parties were organized on caste lines, and why the castes in question were Salagama, Karava and Durava, not Goyigama or Wahumpura or Bathgam or Navandanna, etcetera.
It is significant that three of the five parties that went to Burma were Salagama. Considerable weight has been attached to this point. The emergence of a Karava elite was preceded by the forward thrusts of Salagama notables.’ The late eighteenth and early nineteenth centuries were marked by a Salagama-Goyigama confrontation (pp. 211, 135-6, go-I) . This was an outcome of the advantages (albeit double-edged advantages for their labouring elements) which the Salagama secured in Dutch times. One of the major developments of the nineteenth century, therefore, was the manner in which the Karava elite outpaced the Salagama and placed themselves in a position to mount the principal challenge to the Goyigama notables by the latter decades of the century. For this reason, among others, I question Bayly’s reading of my book as a study of a single caste group and an approach which does not ‘portray the evolution of relations between a variety of castes in the context of the wider field of economic, religious and political organization’ (p. 527). Indeed, one of the factors which encouraged the development of solidarity among the Karava was the localized rivalry between neighbouring Karava and Salagama villages on the coast, rivalry which occurred in a setting in which the solidarity of the Salagama was sustained by the constitution of the caste as a state department devoted to the gathering of cinnamon (from pre-Portuguese times to the 183os) .16
In retrospect, I believe that my conclusion should have dwelt more explicitly on the theme of conflict in the Sri Lankan body politic during the period under review. To this extent Bayly is correct in highlighting this thread. This emphasis, however, was not a fantasy created by twentieth-century egalitarianism (Bayly, p. 522). It was a theme that emerged out of the empirical data; and it resonated through the eighteenth-century material down to that of the twentieth century.’ This does not preclude the coexistence of commonalities associated with caste symbolism and other status claims, besides the unities enforced by the state, especially after 1815 (when the Kandyan Kingdom was vanquished). Indeed, in devoting attention to the Goyigama-ization of the KSD caste elites in British times (pp. 223-4) and in stressing the process of integration into Sinhalese Buddhist culture, I was explicitly referring to overarching unities.
The crux of the matter lies in our different units of study. Where Bayly is focusing on the caste system, my parameters are consistently those of the polity; and where Bayly looks for ‘the essence of the caste system’, that is an issue I did not address. In sequence the polities that attract my attention are: the Kingdom of Kotte, Portuguese Ceylon as constrained by the existence of independent kingdoms in the interior, Dutch Ceylon as constrained by the Kingdom of Kandy, and British Ceylon, especially after the Kandyan constraint was overcome.
In reviewing my study purely in terms of the Sinhalese caste system without attention to the role of the state and my unit of study, Bayly was placing herself on the edge of `Distortionland’. She motored her way into this territory when she sought to conflate and fuse the Sri Lankan material into an empirically-grounded and generalized picture of caste in South Asia. This is not feasible; and is likely to produce meaningless generalities of the sort typified by Gideon Sjoberg’s ‘preindustrial city’.” An overarching understanding can only be sought at a more abstract level, as attempted by Dumont and by Inden and Marriott.20
It is instructive that one of the authorities which Bayly employs in her stress on the interdependencies of the caste system is Edmund Leach.21 Leach was writing an introduction to several village-based studies by anthropologists. The emphasis, therefore, was on jajnzani-like networks.22 We would be well advised, however, to recall Dumont’s critique of village studies: the apparent sociological reality of the village is a product of the dominant caste.23 To this I would add the note that another overarching unity was constructed around the local Brahminpurohits and temple custodians who bestrode the passageways between the Indians and their Hindu gods. In such a context, one needs to extend another of Dumont’s insights; if one wishes to rid India of untouchability one needs to exterminate the Brahmans as a category.’ One could venture further and suggest an extermination of the gods.
This dramatized excursion suggests that we are in need of more Indian studies which focus on challenges to the Brahmans. We need to know more about the temple entry struggles as well as the bhakti movements. These might conceivably highlight the contradictions within the social organization of Hindu caste. Nevertheless, it is significant that several temple entry movements appear to have been relatively late developments in British India and that large parts of the sub-continent did not witness direct challenges to the Brahman priesthood. In contrast, the late eighteenth century in Sri Lanka, one found a challenge being presented to the monopoly enjoyed by the Goyigama in recruitment into the Sangha. In analysing the forces which supported the emergence of a Karava elite, my study also lays out the structural conditions and historical developments which enabled the KSD elites to oppose the Goyigama in this manner.
Susan Bayly’s review article is as revealing as it is earnest and scholarly. It places her in the British empiricist tradition. This empiricism is taken to extreme, positivist lengths. Thus, in criticising the metaphors of conflict which I employ, she blithely notes that ‘this dramatic language . . . offers value judgements in place of objective analysis’ (p. 522); and then proceeds to offer readers her own interpretation of the caste system (of South Asia!) as a system of interdependencies which accommodated jostling and the ‘continual reshuf-fling’ of local schemes of caste ranking an interpretation which is presumably objective. Such a combination of naiveté and blinkered arrogance would be amusing if not for its disastrous implications; unfortunately, it is also impregnable because so incorrigible.
In sum, then, Susan Bayly reveals herself to be a seeker after chronologically-organised historical detail who, at the same time, wishes to construct a general model of caste in South Asia. The latter effort, as attempted in an empiricist manner, leads her into the dangerous waters of a static, structural-functionalist and classificatory view of caste.25 The two tendencies are probably not compatible, but the limits imposed on Bayly by a review essay do not take her far enough to test this. However, the end-product is a misrepresentation of Caste Conflict and Elite Formation. While this is disquieting, it is not without its pleasures. She has exposed her flanks.
University of Adelaide MICHAEL ROBERTS
1 Modern Asian Studies 17, 3 (1983) 519-27.
2 Appadurai and Breckenridge, ‘The South Indian Temple: Authority, Honour, and Redistribution’, Contributions to Indian Sociology lo ( I 976), 187-211; Arjun Appadurai, Worship and Conflict under Colonial Rule: a South Indian Caste (Cambridge University Press, i981); Michael Moffatt, An Untouchable Community in South India (Princeton University Press, 1979); Frank C. Conlon, A Caste in a Changing World. The Chitrapur Saraswat Brahmans, 1700-1935 (Berkeley, University of California Press), 1977.
3 ‘Structural Marginality and the Urban Social Order’, Urban Anthropology (1978), 287-32o.
4 One of the strengths of the studies by Washbrook and Baker is their attention to the colonial state and the manner in which the extension of state agencies influenced local politics and caste relations.
5 Cf. ‘If we are to construct the social history of India on the basis of dozens of local studies, whose informing theory is questionable and whose design makes difficult the development of more adequate theories, we may never build it very high’, David Washbrook in reviewing Karen Leonard’s book on The Kayasths of Hyderabad in Pacific Affairs 52 (Winter 1979-80), 735-6.
6 The chapter on caste in British India was not part of the original plan and was inserted because the referees for the Cambridge University Press desired my study to be related to the literature on caste mobility in India. In responding to this ‘suggestion’, I warned them that such a venture could open a Pandora’s box. Bayly’s article is a proof of this prediction. In retrospect, I believe that my study would not suffer greatly from the elision of the first part of chapter seven, though the survey of the factors sustaining the solidarity of the Karava in the latter half of the chapter is a distinct gain.
7 The term ‘caste formation’ is used in my study as an equivalent for the word jati and as a generic device to encompass all levels of a segmentary structure of caste—thereby leaving open the question whether it possessed a sense of community which justified its description as a ‘caste’ (see p. 18 onwards).
8 Conlon is also circumspect in some of his conclusions. I think his study gains in value from this imprecision and cautiousness. As Lawrence Stone remarks, the study of history forces scholars back ‘upon the principle of indeterminacy’, in part because the variables are numerous and in part, one might add, because of deficiencies in the source material (see his ‘The Revival of Narrative: Reflections on a New Old History’, Past and Present 85, 1979. p. 13).
9 A general land tax was not found in Sri Lanka from pre-colonial times. Thus, in the British period, there was only a tax on paddy; and the taxation rate was one-tenth, not one-half as in much of India. When I glanced at some of the paddy-tax registers of the mid-nineteenth century, it was my impression that the revenue-farmers (renters) were mostly Goyigama.
10 Review of Conlon’s book in The Indian Economic and Social History Review XVI (1979) 999- 1101.
11 ‘Introduction’ in Frederik Barth (ed.), Ethnic Groups and Boundaries (Bergen: Universitetsforlaget, 1969).
12 ‘Hindu society, speaking in this case through the Nayaks of Ikkeri and Bednur, required that each caste should organize itself and the Brahmans should have a recognizable spiritual descent’ (Ashin Das Gupta in review of Conlon, The Indian Economic and Social History Review XVI (1979), 99).
13 As Karen Leonard notes, complaining that Conlon leaves it to the readers to chart out these connections over time (review in the Journal of Asian Studies XXXV II (1978), 564).
14 My summary of the background to this event was unashamedly derived from Kitsiri Malalgoda’s excellent account in Buddhism in Sinhalese Society, 1760-1900 (Berkeley California: University of California Press, 1976). However, it was supplemented by two significant pieces of evidence (see Caste Conflict, p. 138) which confirmed Malalgoda’s argument.
15 In reviewing an earlier work (drafted in 1969-70) in which I had dwelt briefly on the rise of the Karava, D. A. Kotelawele rightly criticized me for not taking account of the rise of the Salagama (Nineteenth Century Elites and their Antecedents’, Ceylon Historical Journal XXV (1978) 204-12). I did not come across this essay until I visited Sri Lanka early in 1979, by which time I had come across the relevant data and incorporated this development into my analysis.
16 See Caste Conflict, pp. 207,211-12, 178, 170, 137–8, 90–1.
17 In questioning my reference to ‘caste warfare’ on the ground that the conflict ‘was a far cry from the communal violence of north India or the honours disputes and left hand-right hand caste conflicts of south India’ (p. 525), Bayly displays mind-boggling pedantry.
18 If I wished to depict the essence of the caste system, I would probably concentrate upon the social organization of the Kingdom of Kotte or that of the Kingdom of Kandy, or both. Apart from the choice of arena, however, the more problematic issue is the theoretical approach which one considers appropriate to the search for ‘the essence’ of a structured social order. Bayly seems unaware of the minefields surrounding such a venture.
19 The Preindustrial City. Past and Present (Glencoe, Illinois: The Free Press, 196o). Despite being cited often, Sjoberg’s study is shot through with most of the shortcomings of the modernization theory. His ‘pre-industrial city’ is a residuary category, with its lowest common denominator being its technologically-determined contrast from the industrial city. No wonder that Moses Finley dismisses this work in a footnote (`The Ancient City: From Fastel de Coulanges to Max Weber and Beyond’, Comparative Studies in Society and History 19 ( 977) 309n). Also see Paul Wheatley, ‘What the Greatness of a City is Said to Be’, Pacific Viewpoint 4, (1963), 163-88, espec. 163-4,171-2,187-8.
20 Louis Dumont, Homo Hierarchicus (London: Paladin, 1972); McKim Marriott and Ronald B. Inden, ‘Caste Systems’, Encyclopaedia Britannica, 15th edn, 3:982-91 and `Toward an Ethnosociology of South Asian Caste Systems’ in Kenneth David (ed.), The New Wind (World Anthropology Series, The Hague: Mouton, 1977), pp. 227-38. Whereas Dumont’s theoretical analysis places the Muslims and the Sinhalese Buddhists in South Asia outside his schema (ibid., ch. 1o), Marriott and Inden explicitly claim that their cognitive and ethno-sociological theory encompasses ‘the caste systems of Buddhists in Sri Lanka [and] those of Muslims in Pakistan and Bangladesh’, ibid., p. 227).
21 ‘Introduction. What should we mean by Caste’ in E. R. Leach (ed.) Aspects of Caste in South India, Ceylon and North-West Pakistan (Cambridge University Press, 1960), pp. 1–10.
22 That is why Leach’s work is surveyed by Pauline Kolenda, together with those by the Wisers, T. 0. Beidelman, and Harold Gould, in her ‘Toward a Model of the Hindu Jajmani System’, Human Organization 1963, 22: I 1-3o. Note too, my passing reference to `the locality-centred functionalist studies of caste networks perpetuated by . . . F. G. Bailey and Edmund Leach’ (p. 195).
23 Dumont, ‘Village Studies’, Contributions to Indian Sociology 1957, I:23-41.
24 This is a stronger statement than Dumont’s; he maintains that ‘untouchability will not truly disappear until the purity of the Brahman is itself radically devalued’ (Homo Hierarchicus, p. 92).
25 The previous reference to Sjoberg’s study was not introduced without reason. His approach is explicitly structural-functionalist (see The Preindustrial City pp. 12-13, 139, 341-2). Also see fn. 19 above