¹ I am grateful to Prof. F. Willett of Northwestern University for the Yoruba example and the possible explanation of it given later; and to Messrs S. Lukes, P. M. Hacker and A. Kenny for many helpful discussions of the matters raised.

² Langer, S., Philosophy in a New Key (Harvard University Press, 1963, third edition).Google Scholar

³ Ibid., p. 96f.

⁴ Carnap, R., Philosophy and Logical Syntax (K. Paul, Trench, Trubner & Co. London, 1935, p. 28)Google Scholar, quoted by MissLanger, in Ch. IV, p. 84.Google Scholar

⁵ Evans-Pritchard, E. E., Witchcraft, Oracles and Magic Among the Azande (Oxford, 1937)Google Scholar. I am not implying that Evans-Pritchard thinks this line to be correct.

⁶ Cf. Leach, E. R., Political Systems of Highland Burma (Bell and Sons, 1954) p. 14Google Scholar: ‘In sum then, my view here is that ritual action and belief are alike to be understood as forms of symbolic statement about the social order’.

⁷ Langer, S., loc. cit., p. 97.Google Scholar

⁸ Ibid., p. 99.

⁹ It will be seen that I am much indebted to W. V. O. Quine (especially to the second chapter of Word and Object) for several ideas in this paper. But, in virtue of his doctrine that all beliefs are revisable, he is, I think, an empiricist (what C. I. Lewis calls a Conceptual Pragmatist), and I have tried to show that no empiricist can sail so close to the rationalist wind.

¹⁰ This might be denied by supposing a tribe which dealt exclusively in imperatives. Here an analogous assumption needs to be made about when an imperative has been put into effect. But, if we stick to imperatives, we shall never get to beliefs at all, much less to ritual beliefs. I shall therefore ignore this enticing alley.

¹¹ E.g. Bruhl, L. L. Levy; La Mentalité Primitive (Herbert Spencer Memorial Lecture, Oxford, 1931), p. 21Google Scholar. But see also his Les Carnets (Paris: Presses Universitaires, 1949), pp. 130f.Google Scholar

¹² Evans-Pritchard, , loc. cit., p. 338Google Scholar (Pt. 3, Ch. IV, Sec. viii).

¹³ Ibid., p. 476. This is one of 22 reasons cited to account for the failure of the Azande to perceive the futility of their magic.

¹⁴ Ibid., p. 195.

¹⁵ E.g. Winch, Peter, ‘Understanding a Primitive Society’ (American Philosophical Quarterly, 1964)Google Scholar and the references cited in that article.

¹⁶ De Sacramento Eucharistiae, P. G. 161, 496Google Scholar. The passage is quoted and discussed by Fr Leeming, Bernard S.J., in his Principles of Sacramental Theology (Longmans, Green and Co., London, 1956), p. 256.Google Scholar

¹⁷ Mr Hacker has pointed out that the argument, if sound, should hold against Intuitionist Theories of Ethics.

¹⁸ It is important also that a similar ‘civilised’ distinction between instrumental and symbolic behaviour does not always hold in primitive thought. Thus Evans-Pritchard remarks of the Azande:

‘When a man chooses a suitable tree and fells it and hollows its wood into a gong his actions are empirical, but when he abstains from sexual intercourse during his labour we speak of his abstinence as ritual, since it has no objective relation to the making of gongs and since it involves ideas of taboo. We thus classify Zande behaviour into empirical and ritual, and Zande notions into commonsense and mystical, acording to our knowledge of natural processes and not according to theirs. For we raise quite a different question when we ask whether the Zande himself distinguishes between those techniques we call empirical and those techniques we call magical.’ Loc. cit., p. 492Google Scholar. The passage does not imply relativism.

¹⁹ Evans-Pritchard, E., Mier Religion (Oxford, 1956), Ch. XII, p. 322.Google Scholar