1 Canfield, John V. Wittgenstein Language and World (Amherst, Massachusetts: University of Massachusetts Press 1981).Google Scholar Hereinafter, WLW. Numbers in brackets in the body of the text refer to page numbers of WLW unless otherwise indicated.

2 Wittgenstein, Ludwig The Blue and Brown Books (Oxford: Basil Blackwell l958) 25Google Scholar

3 Ibid.

4 Ibid.

5 Wittgenstein, Ludwig Zettel, ed. Anscombe, G.E.M. and Wright, G.H. von trans. Anscombe, G.E.M. (Oxford: Basil Blackwell 1967) 492Google Scholar

6 Putnam, Hilary ‘Dreaming and Depth Grammar,’ in Butler, R.J. ed., Analytic Philosophy (Oxford: Basil Blackwell 1962)Google Scholar

7 Ibid. 221

8 From Canfield's preliminary discussion of the scope of criteria.

9 From The Blue Book 25.

10 Shoemaker, Sydney S. Self-Knowledge and Self-Identity (Ithaca: Cornell University Press 1963).Google Scholar Canfield notes that Shoemaker has now given up this position (80).

11 Lycan, W.G. ‘Noninductive Evidence,’ American Philosophical Quarterly, 8 (1971) 109–25Google Scholar

12 Ibid. 110

13 Albritton, Rogers ‘On Wittgenstein's Use of the Term “Criterion”,’ in Pitcher, George ed., Wittgenstein The Philosophical Investigations (Garden City, NY: Doubleday 1966) 245Google Scholar

14 Something is in principle statable if it fails to be in principle unstatable. By the use of ‘in principle unstatable’ or ‘in principle ineffable’ I intend to distinguish cases where an adequate description is unavailable owing to logical or conceptual properties inhering in the proposed candidate for description, from cases where an adequate description is unavailable owing to situation specific limitations on time, space, fluency, or the like. The proper contrast to ‘in principle unstatable’ is ‘practically unstatable.’ More specifically, in the first instance, any possible statement of what it would be for a criterion to be met would in some significant way misrepresent the situation which would actually obtain were the criterion to be met; in the second instance, there is some statement of what it would be for the criterion to be met which does not misrepresent the relevant situation - but such a statement may be too long, or too complicated to make the giving of that statement practicable; or it may be beyond the intellectual capacities of the particular speaker, or require greater interest or fluency than in fact she has, etc. I leave aside here the question of whether it is possible for anything to be in principle, as opposed to practically, unstatable. What is important is that the claim that something may be in principle unstatable is a much stronger claim than the claim that it may be practically unstatable; and that there is textual support in Wittgenstein's work only for the weaker claim.

15 Ibid. 247-8

16 This seems to be the view endorsed on p. 115, though.

17 The formulation of the objection is Carl Ginet's. In the full version, quoted by Canfield on p. 97, the word ‘these’ appears as the word ‘those.’

18 See, for example, 34, 9, and 32; but also see note 22 below.

19 Wittgenstein, Ludwig Philosophical Investigations, trans. Anscombe, G.E.M. (Oxford: Basil Blackwell 1958).Google Scholar Hereinafter PI.

20 Cavell, Stanley The Claim of Reason (Oxford: Clarendon Press 1979; New York: Oxford University Press 1979)Google Scholar

21 As it occurs, for example, on p. 207; or on p. 112: ‘We obviously can know someone was pretending … We know he was pretending in the usual way ….’

22 Cavell claims that Wittgenstein intended criteria to close the gap between language and world, and that his appeal to community (and, as Cavell sees it, our individual responsibility for maintaining community as a basis for judgement) is not intended as an answer to the skeptic – who cannot be answered – but rather as a response: an acknowledgement of the question's unanswerability, and a transcendence of this fact through acceptance of the naturalness of our community-affirming activities.

While I find the conclusion of this view – that it is possible to admit the power of the skeptic's position, but nonetheless to affirm the epistemic access which it denies – attractive, I must confess that I do not understand how the transcendence is to be achieved. How, that is, we are at once to play by rules which leave us the possibility of an appeal to community and yet by rules which also allow us to feel the genuine force of the skeptic's question. Perhaps Cavell and, if he is right, Wittgenstein, have discovered a coherent perspective from which the skeptic's position can be seen to have force, but not quite enough force to destroy the perspective in question. This vision still eludes me. It seems to me that the tension between the rational indefeasibility of skepticism and the apparent fact of its psychological unacceptability is one which is not resolved in Wittgenstein's writings.

23 The phrase is Richard Rorty's in ‘Pragmatism, Davidson, and Truth.’ typescript, 1983.

24 See, for example, PI 217; or Remarks on the Foundations of Mathematics, Wright, G.H. von Rhees, R. and Anscombe, G.E.M. eds., trans. Anscombe, G.E.M. (Oxford: Basil Blackwell 1978)Google Scholar Part VI,§ 38ff.: or Philosophical Grammar, Rhees, Rush ed., trans. Kenny, Anthony (Oxford: Basil Blackwell 1974),Google Scholar Part I: VII §§ 99, 105, VIII.

25 Canfield usually uses the terms ‘grammatical truth’ and ‘depth grammatical truth’ interchangeably.

26 What Canfield actually says here is that grammatical remarks about grammatical remarks obey the very rules they set out. This cannot in all cases be true. Consider ‘It is a grammatical truth that proper names function like singular nouns.’ The rule which this meta-grammatical remark sets out is, presumably, ‘Proper names function like singular nouns.’ The object of the rule is proper names, and thus the rule does not govern even a part of the meta-grammatical remark, let alone the whole thing. Thus, I took Canfield here to have intended the same claim he makes a few pages later on p. 207.

27 See PI 372. Canfield quotes this remark on p. 18.

28 Note that what is intended by the claim that the world is mutable cannot be achieved on the assumption that it is merely phonemes which change; rather, for the world to mutate, what must change are the rôles of phoneme place-holders vis a vis one another. For reasons too complicated to retail here it would also have to be the case that either no world-constituting language exhausted all possible phoneme rôles, or that the relations between some rôles failed to be uniquely determined by the identity of the rôles involved.

29 One way of reading Wittgenstein is to take him to be arguing that we should cease to be troubled by meta-level questions, given the phenomenal identity of solipsism and realism. It is one thing, however, to point out that phenomenal identity, and another to summon psychological complacency in the face of the conceptual distinction.

30 I would like to thank Paul Boghossian for many provocative discussions on the subject of Wittgenstein's later philosophy; and also Duncan Macintosh and Kathryn Morgan for their comments on an earlier draft of this notice.