1 The reader should be warned of two opinions which I hold concerning Wittgenstein scholarship. First, I should like to signal agreement with a view expressed by Ambrose, Alice to the effect that Wittgenstein's writings on family resemblance can, and should, be treated in isolation from the rest of his later writings (implicitly in “Wittgenstein on Universals,” in Essays in Analysis (London: Allen & Unwin, 1966), p. 118,Google Scholar and explicitly in unpublished writings). The reason for this view is that the theory of family resemblance offers at best a solution to the classical problem of universals rather than the desired dissolution of the problem, which is to be found in the rest of his late work. In solving the problem of universals, Wittgenstein is simply ringing changes on traditional nominalistic views (“Wittgenstein on Universals,” p. 118), whereas the dissolving of the problem is a consequence of his theory of meaning. I intend to discuss the theory of family resemblance only along the traditional lines of the problem of universals and simply suspend judgment regarding the dissolution of the problem.

Second, there has been some confusion as to what constitutes the very problem family resemblance is supposed to be solving. Renford Bambrough has hailed Wittgenstein's theory of family resemblance as having solved the problem of universals by resolving some and circumventing other differences between nominalism and realism (“Universals and Family Resemblance,“ Proceedings of the Aristotelian Society 60 (1960-61 ), reprinted in Universals and Particulars: Readings in Ontology, ed. Loux, Michael J. (Garden City, N.Y.: Doubleday, 1970), pp. 109,Google Scholar 117-127; hereafter cited as UP). But the nominalist /realist distinction can be understood in at least two quite different senses. One regards the universal/particular distinction and one regards the essence/accident distinction. The traditional problem of universals is whether general terms or universals are understood as being equivalent in some sense to the particulars that make up their extensions (nominalism) or whether beyond the many particulars there is some one thing, in some sense of “thing,” in relation to which particulars are what they are and in relation to which general terms have their meanings (realism). In setting forth the theory of family resemblance in the Blue Book as a cure for “our craving for generality” and for our “contemptuous attitude towards the particular case” (The Blue and Brown Books (New York: Harper & Row, 1958), pp. 17, 18), it is clear that Wittgenstein is advancing his theory of family resemblance as a resolution of the traditional one over many problem and that family resemblance is relevant to this first sense of the realist/nominalist distinction. It is with this sense, too, that I will be concerned. Bambrough, however, bypasses this sense of the nominalist/realist distinction. Instead, he speaks of realism and nominalism in a second sense, that which regards essences. A realist with regard to essences believes in the objectivity of classificatory schemata, whereas a nominalist with regard to essences believes that classificatory schemata are matters of convention (cf. Bambrough, pp. 117, 126, and on the whole issue of real versus nominal essences see Copi, Irving M.'s “Essence and Accident,” journal of Philosophy 51 (1954),CrossRefGoogle Scholar reprinted in UP, pp. 285–300; Bambrough's conclusions on the issue are substantially the same as Copi's). It seems to me that the doctrine of family resemblance per se has nothing to say on this issue, though Wittgenstein's general theory of meaning certainly does. Further, Herbert Schwyzer has shown that Bambrough's analysis, which assimilates “meaning” to “denotation,” does not even succeed as an interpretation of Wittgenstein in general and that on the issue of classificatory objectivity, Bambrough finds realist touches which are not present in Wittgenstein, 's thought (“Essence Without Universals,” Canadian journal of Philosophy 4 (1974), 69-78).Google Scholar I will not then be concerned with Bambrough's approach to family resemblance and universals, but will be concerned with family resemblance only within the traditional context of the problem of universals.

2 My analysis here is similar to Ayer, A.J.'s (The Problem of Knowledge (New York: St. Martin's Press, 1955), pp. 10–12).Google Scholar Individuals exhibiting various gradations of a sliding scale, such as shades of a single color (Ayer's example) constitute a minimal case of individuals forming a family group. Bambrough's critique of Ayer (UP, p. 116), therefore, collapses.

3 A particular a belongs to family F by virtue of a's having a property that is a variation on the Platonic prototypic property that demarks family F.

4 By a “perfect” instance I mean a particular that corresponds exactly to a standard. If I have an extremely elastic rubber band and proceed to stretch it from its relaxed length to the length of the room, at some one point in the stretching process it will correspond in length to the standard meter in Paris. At that instant it is a perfect particular; at all others it is a degenerate particular.

5 I think it could also be successfully argued, though only at some length, that the Platonic theory I have sketched is indeed Plato's theory. Such an argument would need to establish three points. First, it would need to be shown that Plato did view Ideas as standards. Such passages as Euthyphro 6e and Laws 965b-d would seem to establish this as a view which Plato held from his earliest to his latest works. Second, it would need to be shown that Ideas, though standards, are not in addition treated by Plato as common properties. To establish this would require showing that some passages (notably Republic 596a, Theaetetus 185c-d, Euthyphro 5d and Timaeus 49b-50b), which have been taken even by some of Plato's most sympathetic critics as committing him to a theory of abstractive universals, indeed do not entail such a commitment: Third, it would need to be established that Plato specifically entertained phenomenal particulars which fall under a single Idea as forming a variation-group. Such positive evidence is to be found in the discussions in the Statesman (283b-285a) and Philebus (16a-1Bd, 26a-b, 64d-65d) of the relation of phenomena to measures, which I take to be Ideas (see my “Statesman 284c-d: An’ Argument from the Sciences',” forthcoming in Phronesis).

6 My illustrative analogy of the police artist's sketch of the Hapsburgs does not capture these qualifications ((1) and (2)). To do so, the sketch would have to be equivalent to the Form of Hapsburg. But as a human artifact the sketch cannot be a Form. Further most critics would wish to deny that Plato supposed there to be Forms of individual persons or their families. There is no Form of John Paul Jones, and no Form of the Joneses.

7 Geach, P.T. has preserved two statements on standards made in conversation by Wittgenstein (“The Third Man Again,” Philosophical Reivew 65 (1956),Google Scholar reprinted in Studies in Plato's Metaphysics, ed. Allen, R.E. (London: Routledge, 1965), pp. 267, 269).Google Scholar One is consistent with Plato's analysis of standards. It reads: “The bed in my bedroom is to the Bed [i.e. the Idea of Bed], not as a thing is to an attribute or characteristic, but rather as a pound weight or yard measure in a shop is to the standard pound or yard” (p. 67).

8 Wittgenstein goes on to use this analysis of standards as an analogue to his analysis of the utterances of words: “And just this goes for an element in [a] language-game when we name it by uttering the word ‘R': this gives this object a role in our language-game; it is now a means of representation” (section 50, Wittgenstein's emphasis).

9 The relation of original to copy is minimally a highly illustrative species of the relation of standards to their instances. Plato, of course, goes further and identifies his paradigmatic standards with originals and the instances of standards with copies or imitations of originals (e.g. Timaeus 48e5-6).

10 Wittgenstein's comparison of the usage of standards to the utterance of words (see note 8), therefore, collapses.