Acknowledgements: I thank Chris Bayly, David Hardiman, Richard Fox, Ron Inden, Tapan Raychaudhuri, David Stark, Burton Stein and David Washbrook, for their comments on earlier versions of this paper. These were presented to several seminars and workshops in the course of 1983 held at the London School of Economics, University of Pennsylvania, Duke University, at the Universities of Chicago and Wisconsin, and finally at St Antony's College, Oxford.

¹ In Eric Wolf's words (1982: 148), the culturalists ‘attract all those who believe in the primacy of mind, who see humankind as spinning ever more complex webs of signification through autonomous processes of the symbolic faculty. In this perspective, signification sets up the human relation with the material universe’. On the other hand, ‘the materialists cleave to the belief that human affairs are caused by the ways human beings cope with nature’ (thus including the notion that culture forms part of “superstructure”).

² The literature on ‘tribe’ and the ‘invention of tradition’ are especially revealing; see, for example, the references in Perlin, 1983, 42, n. 42.

³ See, for example, Giddens, 1979, Introduction and ch. 1; and Abrams, 1971 and 1982, from among a large critical literature.

⁴ Among others, Williams, 1973 and 1977, chs 1/4 and II/1, and Godelier, 1978 and 1983, both of whom reject the base/superstructure model. But for a conception close to that advanced in this essay, see Cohen, 1978, ch. VIII.

⁵ Skorupski, 1976, gives a critical overview of these different hypotheses.

⁶ Sperber, 1982, 170: ‘Our capacity to form semi-propositional representations gives us the means to process information—and in particular verbal information—which exceeds our conceptual capacities’.

⁷ Dumont, 1980, 105–6; this book has now become the classic exposition of the relationship between caste, religion and material exchanges; despite much dispute around particular hypotheses composing it, the basic principles hold. See also Marriott, 1976, for the caes in exemplary miniature.

⁸ Dumont, 1970, 210 and 1980, 165.

⁹ Dirks, 1979, and Bayly, 1983, 50, basing themselves on Hocart, 1950.

¹⁰ A classical and consistent argument for this view is Meuvret, 1971, the two essays in Part III. Raychaudhuri, 1982, i, 358 (and see also ii, 22) presents it as the ‘curious paradox of the Mughal economy’. See Perlin 1978, 223, n. 13, for other references and for their logically arbitrary character.

¹¹ Dumont, 1980, ch. 4; Raychaudhuri, 1982, i, 280–1; Habib, 1963, 128–9 (retracted in 1982, 248, but with little to substitute for it but a single reference to an opinion of Baden-Powell).

¹² Dumont, 1980, 99: after first mentioning the ‘village community’ literature (which is itself highly problematical), he goes on: ‘Wiser was the first to describe it in detail…some time before 1930’ and 372, n. 42 d, ‘Blunt mentions the term prior to Wiser’ (the references are Wiser, 1936, and Blunt, 1931). For the deficiency of historical sources, see Raychaudhuri, 182, i, 279–80, and ii, 9–10.

¹³ Here, again, a reference to recent work on ‘the invention of tradition’ is relevant, especially that on Africa; see Ranger, 1982, and the reference in n. 2 above.

¹⁴ Dumont, 1980, 106, that‘…what would otherwide be economic is subjected to [the religious] reference’.

¹⁵ Harriss, 1982; Gough, 1960; Raychaudhuri, 1982, i, 279–80, 325–7, and ii, 9–10, 21; Rudra, 1981, 2144.

¹⁶ Ironically, within the narrow limits of the culturalist paradigm, there seems to be more dispute concerning the status of Dumont's particular theory than amongst marxists.

¹⁷ Giddens, 1981, 11, for the quotation, and 1979, for a general critique of the social sciences which surely contradicts it; Godelier, 1978, e.g. 767. The latter appears to set down all the elements of the culturalist perspective but with marxist terminology; thus Bloch's critical summary of his views (1978, 768).

¹⁸ For both Thompson and Medick, see Medick, 1983, 85 and 89.

¹⁹ See n. 24.

²⁰ Comparable cases are the terms ‘totemism’ and ‘hysteria’, as discussed in Levi-Strauss, 1964, and Foucault, 1965, 136ff, respectively.

²¹ The contents of this and the next paragraph are discussed at greater length in Perlin, MAS 1985. Evidence and discussion of certain of these aspects are central to Levi-Strauss, 1966 (classification of nature, naming, thus, for example, ethno-biology), and as part of an older tradition of historical writing well represented by Gray, 1915.

²² Sperber, 1982, 171ff; Bloch, 1977, is another important case.

²³ The quotation comes from Bloch, 1978, 768, characterizing Godelier's position.

²⁴ This contrast, in its more characteristic expression as the two stages of a historical transition between types of society, stands at the centre of the Enlightenment view of the old order in Europe (put particularly well by Taylor, 1979, 8–9, 15, 20–2 and 141–2; see also Marcuse, 1955, 3–4), since when it has entered mainstream social-science thought both as such and in the more static differentiation between western and nonwestern societies according to this distinction. This is less a statement of origins so much as of an epistemological development which itself requires study.

²⁵ Bourdieu, 1968.

²⁶ Briefly discussed in Perlin, 1985, MAS and JPS, Section II, 1985.

²⁷ For example, Thompson, 1971 and 1978, and Medick, 1983, and for the French literature, see Clark, 1983. Ziegler, 1978, is a unique Indianist example, although not framed in these terms.

²⁸ Stated in spite of Skorupski, 1976, whose interesting book remains again within the culturalist framework of relevance.

²⁹ Godelier, 1978, 768, rhetorically asks: ‘Ought we really to be using words such as class and state when referring to hierarchized precapitalist ancient or exotic societies?’ He suggests that Marx (that is Godelier), does not consider that ‘we ought to start hunting for classes hiding behind the [medieval] orders—classes that Marxists alone would be able to discover and that the Greeks or the Romans, history's actors, could never have seen for themselves’. This puts the culturalist case very neatly. But ought we to be using words like ‘precapitalist’, ‘exotic’ and ‘societies’?

³⁰ The most obvious example of procedural substantivism is Polanyi, 1957 (his ‘economic substantivism’), but it is clearly a much more generalized phenomenon, central to the traditions of anthropological thought and cultural relativism, and even part of the common sense of the social sciences.

³¹ See also Sperber, 1982, 164–6.

³² This latter question is discussed at length in Perlin, 1983. Barnett, Fruzzetti and Ostor in their critique of Marriott assure us that ‘Of course we do not challenge the fundamental unity of India nor the relational unity of mind and matter’ (1977, 599). Williams, 1977, 80, notes that ‘the analytic categories … almost unnoticed, become substantive descriptions, which then take habitual priority over the whole social process…’.

³³ For these arguments, see Perlin, 1983 and MAS 1985.

³⁴ See for example Barkan, 1975, James, 1983 and Kantorowicz, 1957.

³⁵ I develop this point in my essay on state-formation (1985). Giddens, 1981, also emphasizes the role of what he terms ‘information storage’ in the organization of dominance and surveillance (e.g. 5–6): ‘Following Mumford, I regard the city in non-capitalist societies as a special form of “storage container”’; I would instead argue that it facilitates storage in dispersed centres of power; conversely, Bourdieu's notion of habitus is relevant, although here I would want to refer to the habitus of an uncontained population and its component habiti (see Bourdieu, 1976, 141).

³⁶ Godelier pays lipservice to variation but provides no grounds for it (1978, 768).

³⁷ Habib, 1982, i, 363–71.

³⁸ These are problems dealt with at length in my essay on money-use (1987).

³⁹ Clear in the case of Gujarat (see Perlin, in MAS 1985, and JPS, 1985); for the Russian mir, for example Male, 1971, 6; on western and central Europe, Blum, 1971; Bloch, 1966; there is also Habib's (1982, i, 248) gnomic reference to Baden-Powell, concerning India.

⁴⁰ For example, Gellner, 1982, but even the more sensitive discussion ofTaylor, 1982 (both in the admirable volume edited by Hollis and Lukes); it is as if western societies consist only of university intellectuals, of rationalists and atheists not disposed to faith and ritual, or subscribing to ideologies (and as if university intellectuals could be characterized wholly in such terms).

⁴¹ See n. 3, and Bourdieu, 1968.

⁴² Discussed, for example, in Lukes, 1982, 162ff; Sperber, 1982, 153–7; Hollis and Lukes, 1982, 12ff; Gellner, 1982, e.g. 183, seems to bypass the question.

⁴³ Surely, the necessary condition for technical communication among those international groups of migrant workers, whose origins include a variety of backgrounds of the types treated by relativists.

⁴⁴ See n. 21.

⁴⁵ See Clark, 1983, for example 67 and 81, for various references.

⁴⁶ Gellner rejects the accusation of his being a relativist, as Godelier would of being a culturalist, but for purposes of discussion, in terms of the uses and definitions applied in this paper and conventional to the media, such characterizations are inescapable.

⁴⁷ We would say that ideologies are full of subjects (corporate and incorporate), but that the ‘objective’/instrumental sphere consists of subjects and objects, being that in which, therefore, the problem of ‘agency’ arises. It follows that a distinction needs to be made between ‘corporate’ and ‘incorporate’ individualisms, on the one hand, and an ‘instrumental’ individualism, on the other; Foucault's critique of Bentham concerns the pathology of the latter (1977, 195ff), while Rousseau's General Will presents an actualization of the former.

⁴⁸ Pearson, 1976, ch. VI, and Kulkarni, 1969, among many.

⁴⁹ Not to be confused with Dumont's use of the term. ‘Complementarity’ and the contrast between classical and quantum mechanics are discussed in a short, sympathetic and excellent essay by Holton (1970); at length, but rather simplistically by Feuer, 1982; usefully defined by Bynum et al., 1981, 73, and fiercely criticized by Popper, 1965, 100–1. A good lay treatment is given by Zukav, 1979, 93–5 and 201n.

⁵⁰ Duby, 1980.

⁵¹ Ziegler, 1978.

⁵² Giddens, 1981, 230–6 (and 281–2), who rightly emphasizes the fundamental difference between mere conflict and contradiction.

⁵³ But this is not to say ‘random’, except at the level of data collection.

⁵⁴ Quotation from Medick, 1983, 84; in the case of Godelier, for example, 1978, 765, ‘…the dominant “superstructure” functions simultaneously as a relation of production’.

⁵⁵ Cohen, 1978, ch. VIII, provides an excellent reconsideration of the base/ superstructure question in Marx, with which I largely agree given minor adjustments. Thus my substantive and instrumental categories conform with the basic distinctions made by Cohen.