¹ Among several others, the two important reviews by social anthropologists that recently appeared on the French version of this book are: Marriott, McKim, “Homo hierarchicus: Essai sur le système des castes,” American Anthropologist, Vol. 71 (December 1969), pp. 1166–1175CrossRefGoogle Scholar. Yalman, Nur, “De Tocqueville in India,” Man (N.s.), Vol. 4 (March 1969), pp. 123–131CrossRefGoogle Scholar. The two reviews viewers, as discussed later on, arrived at different assessments of the Dumont's book, but in the process they generated several useful points about approach and methodology.

² The point is important, I think, for disentangling the substantive issues of the debate from possible mal entendu attending it in some quarters. Since I do not regard the discussion on caste hierarchy having reached a definitive stage either in theoretical constructions or in empirical testing, my remarks are made in a spirit of exploration, rather than a simple criticism directed against the author of Homo Hierarchicus. It is in this spirit that even some most recent works are referred to in the following discussion.

³ An important sociological area of study is left unexplored in the Indian context, it seems, because we have dwelt heavily on the “collective” aspects of caste and ritual behavior, denying any systematic place—conceptual or empirical—to the Indian schemes of the “individual.” The latter is often considered uncouthly, and hence relegated to philosophical conjectures. Is the concept of the “individual,” like modernity, patently and exclusively Western? Or is it that the concepts of social unit can vary according to an intellectual tradition? Dumont's Introduction requires fuller explanations in this area, because the Western “individual” is contrasted with the Indian “collective,” without telling us what Indian “individual” may be.

⁴ Without going into the details of the debate “For a Sociology of India,” one may specifically refer to: Madan, T. N., “For a Sociology of India,” Contributions to Indian Sociology, Vol. 9 (1966), p. 9–16Google Scholar; “For a Sociology of India: Some Clarifications,” Contributions to Indian Sociology (New series), Vol. 1 (December 1967), pp. 90–93CrossRefGoogle Scholar. Dumont, Louis, “The Modern Conception of the Individual,” Contributions to Indian Sociology, Vol. 8 (1965), pp. 13–61Google Scholar; and “A Fundamental Problem in the Sociology of Caste,” Contributions to Indian Sociology, Vol. 9 (1966), pp. 17–32Google Scholar. An interested reader may do well to read this discursive journal's old numbers, particularly up to 1966, to acquaint himself with Dumont's “positions.”

⁵ However, Dumont's success or failure in achieving these aims is already a matter of some discussion. Marriott, op. cit, pp. 1166–1175, found his “theories very hard to test against facts, generally,” his deductive models suffering from the fallacy of “misplaced concreteness,” and his data on practice and fact drawn from uneven and selective sources. Yalman, op. cit., pp. 123–131, definitely more agreeable than Marriott, found Homo Hierarchicus exploring on a generally reliable deductive model, whose “principles can be discovered” (p. 126); though for different (than Marriott's) specific reasons, he found the caste system of India “treated in somewhat general terms” and die derivation of Hindu ideology from the Hindu past less precise than it could have been (pp. 130–131).

I have already remarked above on some aspects of this problem, while more comments follow.

⁶ It may be noted that under the zoological rules of dimorphic nomenclature—which if followed here—h, rather than H, should be used to spell “hierarchicus.” See the French version of the book. Thus, unless referring to as a type, I shall use the book title as published in English.

⁷ Kaplan, Abraham, The Conduct of Inquiry (Scranton: Chandler Publishing Company, 1964) p. 298Google Scholar; see also Brown, Robert, Explanation in Social Science (Chicago: Aldine Publishing Company, 1963) pp. 168–171, 178ff.Google Scholar, for some general issues raised by Dumont's “idealizations” in relation to hierarchy.

⁸ A related, formalized, but intensive matrix analysis of interactional ranking by Marriott, McKim, in Structure and Change in Indian Society (Chicago: Aldine Publishing Company, 1968) pp. 133–171Google Scholar, tends to push this dimension more than any other description so far. However, given the fact that hierarchies yield a calculus, a set of “co-ordinating definitions” and a set of interpretive, interlinked models, alternative routes of analysis may be not only possible but mutually corroborative, An elaborate and intensive analysis of ranked food transactions under northern Indian conditions (see Dumont's Sections 51–65) is discussed in my forthcoming book on food, rank, and ritual, trying to evolve a “calculus” underlying varied ranked food transactions, whether normal, normative, or special.

⁹ See Kaplan, Abraham, The Conduct of Inquiry, inquiry. p. 304, pp. 311ffGoogle Scholar. However, when an author forwards a theory it is expected to replace a “poor” one by a “better” one. Dumont's scheme should also withstand the same test in terms of the “inclusiveness” it propounds for the social phenomenon of caste and ideology.

¹⁰ Emile Senart, Les Castes dans I'lnde. Les fails et le système (Paris: E. Leroux, 1864)Google Scholar; Bougié, Ceélestin, Essais sur le régime de castes (Paris: Alcan [1st ed., 1908], 1927)Google Scholar.

¹¹ Dumont is correct in arguing for the basic separation. He is equally correct in noting that “the idea of purity has hygienic functions” (p. 60; e.g., see Khare, R. S. “Ritual Purity and Pollution in relation to Domestic Sanitation,” Eastern Anthropologist, Vol. 15 (May-August 1962) pp. 125–139)Google Scholar. However, it would be wrong to conclude that the idea of physical cleanliness does not exist for the Hindu. If such an erroneous conclusion was reached, it would be because of the premise that physical cleanliness could be only either hygiene (the example of Western “socio-centricity”) or nothing. See Douglas, Mary, Purity and Danger (New York: Frederick Praeger 1966), pp. 2–6; 29–40CrossRefGoogle Scholar. The concepts of sśuca and ūsśuca (p. 49) have both individual and as well as collective (what Dumont calls culture or civilization; see p. 61) dimensions, where the former further classifies along the physical and the spiritual. One misses such a discussion in Homo Hierarchicus also because these concepts might have helped in producing an integration of the Hindu view of “individualism and holism” (cf. Sections 2 and 3). See also footnote 3 above.

¹² Comparatively, Marriott (op. cit., 1968; op. cit. 1969, p. 1172) seems to operate from a very simple transactional principle of “high” and “low,” with numerous gradations recognizable in between, where all such cultural specifics as the pure and impure, and inclusion and exclusion (encompassed and encompassing) get translated in the same transactional “language of hierarchy.” Thus he complains about Dumont's usage because “he does not press his uses of it [i.e., the above principle, which Dumont recognizes for showing the importance of attributional ‘theory’, see pp. 89–91] far enough to realize that it can yield a consistent cultural model of ranks in the whole hierarchy from Brahmans to untouchables …”

As this debate on the “attributional or interactional” is developing, only differences are getting better stated. But it is only a beginning, leaving too many conceptual postulates unstated or implied. Both viewpoints espouse different levels, scopes, logic, and purposes, although operate on the assumption that all else in a context is either transformable (reducible) or inconsequential. But none of the two has specifically stated either the assumptions or the steps or the rules which enable him to transform such meanings as differentiation, purity, and inclusion into interactional rank (cf. Marriott's scheme), or interactional rank into attributional “structure” (cf. Dumont's scheme). What both seem to have done is to start from the hypothesis that such a transformation is not only feasible but desirable for certain individually favored, but different, analytical aims—one micro-sociological (inductive) and the other macrosociological (deductive). What “rewrite rules” (with what logical and conceptual implications) are being followed by each of them remains to be elaborated. Even if we were to grant that either Dumont's or Marriott's view of hierarchy will be more consistently explanatory for the caste system, we still need to be further informed and persuaded in either case.

Obviously, then, there also will be those who would want to handle “meaning” and “transaction” together in an appropriate measure to see if that would bless a better union between the thought and action, underscoring hierarchies.

¹³ Let us only pose in passing: a generic definition of “priest” is where Brahman and priest are not kept ideologically inseparable, but that priesthood and status are. If Brahmans tend to be priests, it is only circumstantial (historical) rather than a causal relationship. For some empirical evidence for such observations, see Khare, R. S.The Changing Brahmans (Chicago: University of Chicago Press, 1970)Google Scholar; and “One Hundred Years of Occupational Modernization among the Kanya-Kubja Brahman,”paper presented in the Conference on Occupational Cultures in Changing South Asia, at the University of Chicago,May 1970.Google Scholar Further, do all priesthood roles go to Brahmans? Do all kinds of priestly roles carry same or similar (higher) statuses? Ideologically, what are the shifting conceptions of priesthood (separately from, and not in relation to, kingship) ? Answers should have far reaching effects on Dumont's basic premise (see pp. 90–91).

¹⁴ On the other hand, Dumont is significantly responsible for some of the finest recent studies on the subject, besides the debates that he has genneed erated in the process (for a listing of his relevant works see pp. 352–353). Still, his concern has hardly received the critical scholarly response it deserves from Indian, American, and British sociologists and anthropologists, who largely continue to emphasize “hard data” rather than “attribute” studies.

¹⁵ Dumont makes a very brief reference to the Kanya-Kubja Brahmans (see his footnote 55f, p. 300). For details on the group's hypergamy, see Khare, R. S., “On Hypergamy and Progeny Rank Determination in Northern India,” Man in India, Vol. 50 (October-December 1970), pp. 350–378Google Scholar; and “Hierarchy and Hypergamy: Same Interrelated Aspects among the Kanya-Kubja Brahmans,”paper presented at American Anthropological Association meeting,San Diego,November 1970.Google Scholar Briefly, the Kanya-Kubja Brahmans present a case of institutionalized, elaborately hierarchized, and quantified hypergamy within their caste group. They exhibit pervasive hierarchization in both kinship and marital relations.

¹⁶ For recent studies of occupational changes in India, see the proceedings of the “Conference on Occupational Cultures in Changing South Asia,” (May 1970); also Cohn, Bernard S. “Recruitments of Elites in India Under British Rule” in Essays in Comparative Social Stratification, eds. Plotnicov, Leonard and Tuden, Arthur (Pittsburgh: University of Pittsburgh Press, 1970) pp. 121–147Google Scholar. Leach, Edmund and Mukherjee, S. N. (eds.) Elites in South Asia (Cambridge: Cambridge University Press, 1970), particularly Chapters 2, 3, 5, and 9Google Scholar.

¹⁷ See Khare, R. S. “A Case of Anomalous Values in Indian Civilization,” Journal of Asian Studies, Vol. 25, pp. 229–240CrossRefGoogle Scholar. The case described is only for one—Kātyāyan—gotra, although the point to emphasize is that meat eating, even when institutionalized, does not influence a Kātyāyan's internal or external rank. They rank higher and freely marry their daughters to those vegetarians who rank higher or highest.

¹⁸ Some further comments on these aspects may be found in another appraisal of Homo Hierarchicus recently written for Contributions to Indian Sociology (New Series).

¹⁹ As already alluded in footnote 8 above, our information on food transactions in Indian society is highly selective. Intercaste interactions have received (and continue to receive) the attention they deserve, but our knowledge about domestic, kinship, and affinal food behavior is almost nonexistent, except either at the level of textual ideal or earlier ethnology, including Blunt's study Dumont briefly refers to (p. 137).