¹ Locke's Indian philosopher appears on pp. 230 and 392 of his Essay Concerning Human Understanding (ed. Campbell Fraser), Vol. I, Dover, 1959.The two papers by G. E. Moore that are essential reading in connection with Wittgenstein's epistemological views are ‘A Defence of Common Sense’, Contemporary British Philosophy (ed. Muirhead), Allen and Unwin, 1925, and ‘Proof of the External World’, Proceedings of the British Academy, 1939–both reprinted in Philosophical Papers, Allen & Unwin, 1959.The present paper is largely concerned with On Certainty, Blackwell, 1969. But there is also considerable discussion of relevant parts of the Philosophical Investigations, Blackwell, 1953, and, occasionally, something comes up from the Remarks on the Foundations of Mathematics, M.I.T., 1978. I quote fairly extensively though often no more than an individual word or phrase and give a specific location only when it seems, for one reason or another, important.Google Scholar

² See Ayer's Language, Truth and Logic, Gollancz, 1936, for a characteristically straightforward statement of the methods and goals of Positivism–and if The Critique of Pure Reason (trans. Meiklejohn), Dent, 1934, isn't tight, what is?

³ The remark about our questions is from #120 of the Investigations– and the contrast between the ordinary and the logically complete and precise is quite explicit in #98 and #105–#108. #107, incidentally, concludes with the admonition ‘Back to the rough ground’.

⁴ The above sort of two-fold distinction was finally given its proper place in the tradition by the Verification Principle. Language, Truth and Logic, in particular, succeeded in summing up a more than two thousand year old tradition in a form in which its rigidity and its poverty were right on the surface. So it sounded, for some of us at any rate and in retrospect at least, the death knell of the tradition.See Kripke's Naming and Necessity (Oxford: Blackwell, 1980), for an attempt to insulate the necessary from the a priori.Google Scholar

⁵ Wisdom discusses various parallels between philosophy and abnormal psychology in ‘Philosophy and Psychoanalysis’, Polemic, 1946, and ‘Philosophy, Metaphysics and Psychoanalysis’–both contained in Philosophy and Psychoanalysis, Blackwell, 1953.

⁶ Hume says in his definition of ‘matters of fact’, a notion that for him includes scientific generalizations, that they ‘depend on a hundred different accidents’–Treatise of Human Nature (ed. Selby-Bigge), Oxford, 1978, 69.The fire, the dressing gown and the piece of wax are, of course, Descartes's version of Moore's hand. See Meditations I and II in The Philosophical Works of Descartes (trans. Haldane and Ross), Vol. I, Cambridge, 1931.Google Scholar

⁷ Descartes does, it should be said, sometimes maintain–as he does at the beginning of Meditation III–that God is the ultimate foundation.

I promised not to talk about language-games. But a word about a view that is, in the Investigations, inextricably tied to the notion, the view that a community is required for language. I think that this is a pernicious view and that Wittgenstein's arguments for it are fatally flawed. It gets very little press in On Certainty, surfacing in #298 and one or two other places–although Wittgenstein often puts his views, quite naturally, in terms of ‘us’ or ‘human beings’. There are two questions. Did even Wittgenstein believe that On Certainty needs propping up with this creaky and problematic view from the Investigations? Does it in fact? I think the answers are ‘Apparently not or not obviously’ and ‘No’ But that doesn't matter. What does is that many accounts of On Certainty are compromised from the outset by the unreasoned assumption that one or both of these questions should be answered affirmatively. See, for example, Strawson's Skepticism and Naturalism: Some Varieties, Columbia, 1985. And see Marie McGinn's Sense and Certainty, Blackwell, 1989,

⁸ Wittgenstein's uneasy relationship with scepticism has been noted by others. See Kripke, Wittgenstein on Rules and Private Language, Harvard, 1982, and Simon Blackburn, Spreading the Word, Oxford, 1984–and see Strawson for some more general concerns. This sort of thing can get out of hand though. So Wittgenstein is a ‘Pyrrhonian sceptic’ for Fogelin in Wittgenstein, Routledge and Kegan Paul, 1987, and both a ‘sceptic’ and a ‘phenomenalist’ at once for John Cook in ‘The Metaphysics of Wittgenstein's On Certainty,’ Philosophical Investigations, 1985. But scepticism preoccupied Wittgenstein–and it is still very much in the air. It is central to the ongoing debate between Realists and anti-Realists. As B. Martin puts it: ‘[Philosophers] will unashamedly use the skeptic to frighten off an opponent and to effect for themselves the most terrible reductions, ontic irresolutions, and debilitating relativities. They do not even stop to think that they may be running scared’, ‘Antirealism and the World's Undoing’ Pacific Philosophical Quarterly, 1984. Martin's comment is made as a preface to a discussion of Davidson and Putnam but it also applies to Dummett. Martin's own response to scepticism, however, is not much better. His only suggestion is that we would be somehow hard done by if the sceptic were right and we were wrong. Descartes thought this too–but at least he had God to turn to when it occurred to him that the world might be epistemologically unfair.Google Scholar

⁹ The remark about shades of meaning is from #254 of the Investigations. The seeds of Wittgenstein's argument, which comes up from time to time in On Certainty, can be found in the Investigations, #246 and p. 221. It contains other flaws as well as those I have mentioned but it seems not to be an argument that has impressed many philosophers. A spirited though in the end unsatisfactory defence of it is given by Cook in ‘Wittgenstein on Privacy’, Philosophical Review, 1965–reprinted in Wittgenstein (ed. Pitcher), Doubleday, 1966. Further discussion can be found in John Canfield, ‘"I know that I am in pain" is Senseless’, Analysis and Metaphysics (ed. Lehrer), Reidel, 1975–reprinted in The Philosophy of Wittgenstein, Vol. 8 (ed. Canfield), Garland, 1986–and Newton Garver, ‘Neither Knowing nor not Knowing’, Philosophical Investigations, 1984. Wittgenstein also offers other criticisms of Moore in On Certainty– that there is, for example, something inappropriately personal in Moore's saying ‘I know’ and that Moore's procedure implies both that knowledge is an introspectible state and that it guarantees its object. These criticisms strike me as desperate attempts to convict Moore of something, of anything, over and above the familiar charge of question-begging.Google Scholar

¹⁰ The sceptic does employ other weapons–especially in particular cases. But the above are, I think, the two most important general weapons. See #392 of On Certainty for Wittgenstein's remark about doubt. It reads in full: ‘What I need to show is that a doubt is not necessary even when it is possible’. And see Austin's ‘Other Minds’, Aristotelian Society, Supp. Vol., 1946, reprinted in Philosophical Papers (ed. Urmson and Warnock), Oxford, 1961, for some interesting suggestions about the above principle. Counter-possibilities, by the way, are most naturally thought of as compatible with what is not in question but conflicting with what is. The notion is, in fact, broader than that as should be clear from Gettier's ‘Is Justified True Belief Knowledge?’, Analysis, 1963, reprinted in various places. The point is emphasized by Barry Stroud in The Significance of Philosophical Scepticism, Oxford, 1984.Google Scholar

¹¹ Wittgenstein is sometimes nervous about the idea that grammar or what is philosophically significant might not be exhausted by matters of meaning or, at least, of ‘meaning’. And he tries the ‘meaning move’, not only with ‘knowledge’, but also with ‘argument’, ‘mistake’, ‘judgment’, ‘reasonability’ and so on. It is rejected–convincingly, it seems to me, and along with most other standard responses to the sceptic–by Keith Lehrer in ‘Why Not Scepticism?’, Philosophical Forum, 1971, reprinted in Essays on Knowledge and Justification (ed. Pappas and Swain), Cornell, 1978. Reasons, at least, and of a compelling kind are needed here–just as they are for the view that the sceptic does not make sense or that the language- game is somehow sacrosanct. Stroud too is very good on the implausibility of the grounds usually offered for not embracing scepticism and in the negative, critical, part of his book he proceeds, quite rightly, as though we ought to be able to tackle the extremely powerful considerations in favour of being a sceptic head-on.

¹² Like those suggested by Peter linger–who is, as far as I know, the only sceptic in print who makes no bones about it at all–in Ignorance, Oxford, 1965. Such accounts of absolute certainty should not be confused with the ‘absolute conception of reality’ described by Williams in Descartes: The Project of Pure Enquiry, Harvester, 1978. That has to do with the fact or reality side of the fence. It seems required for knowledge, according to Williams, but he thinks of it, in its comprehensiveness, as analogous to Nagel's ‘view from nowhere’–for which see The View from Nowhere, Oxford, 1986. It is then unattainable by anyone but God. There is, though, no reason why we shouldn't split it up into bits. And Wittgenstein's paradigms, I will argue, are indeed bits or elements of the absolute conception of reality. I mean: They're facts or they're absolutely true. What more could you ask?Google Scholar

¹³ For Wittgenstein, by contrast: ‘A totality of judgments is made plausible to us’.–On Certainty, #140; ‘What I hold fast to is not one proposition but a nest of propositions’.–#225; and ‘Experience can be said to teach us these propositions. However, it does not teach us them in isolation: rather, it teaches us a host of interdependent propositions. If they were isolated I might perhaps doubt them’.–#274.

¹⁴ Doesn't Wittgenstein's failure to insist on the paradigms or insist enough give everything back to the sceptic? Even here there are clues. Sometimes he does insist–as with the following by now anachronistic example: ‘We all believe that it isn't possible to get to the moon; but there might be people who believe that that is possible and that it sometimes happens. We say; these people do not know a lot that we know. And, let them be never so sure of their belief–they are wrong and we know it’.–On Certainty, #286. He also says: ‘Knowledge is in the end based on acknowledgement’.–#378. I don't profess to know what Wittgenstein means by that difficult word but what I say in the rest of this paper is one more or less plausible account. And then there is the fascination with Moore's obviously and flatly question-begging response to scepticism. Stroud too devotes a fair amount of space to Moore. What is needed, however, is to turn question-begging into an art.

¹⁵ Stroud discusses, in his brief positive comments, Thompson Clarke's distinction between ‘internal’ questions and the ‘external’ questions of the sceptic and the idea that the latter can be shown to somehow bypass the former or even to be incoherent. The distinction seems to mean, for Stroud and Clarke: ‘plain’ as opposed to ‘philosophical or sceptical’. But the above simple facts are just plain simple facts. There is nothing particularly philosophical about them, nothing fancy or potentially corrupt–and there is certainly nothing sceptical, not when they are considered in themselves. Otherwise, Realism would, as is sometimes maintained, imply scepticism. This distinction of Stroud's and Clarke's, it should be noticed, differs from Jonathan Dancy's in An Introduction to Conemporary Epistemology, Blackwell, 1985. For Dancy, an ‘internal’ account of knowledge requires that we know that we know whereas an ‘external’ one does not. Dancy claims, unconvincingly, that his own response to the sceptic–we know without knowing of a crucial factor in virtue of which we know–is not externalist. I think there are external cases of knowledge in Dancy's sense. The sceptic need have no worries, though, if all we can say is that we believe we know. The regress that threatens in internal cases? The human mind is sophisticated enough to allow for a handful of levels but the regress peters out as our limitations set in.

¹⁶ A difficulty Wittgenstein runs into in certain contexts. Hume puts the point about arguments against scepticism with regard to reason being question-begging in temporal terms. First, scepticism holds sway in the mind, then the anti-sceptical or ‘dogmatical’ position, then scepticism... and so on. Fogelin too notices the relevance of Hume's discussion to Wittgenstein's project–although his interpretation of the issues is very different from mine. But he, like Hume, puts the point temporally–see pp. 228–9. Hume does say, however, that ‘the sceptical and dogmatical reasons are of the same kind’. Hume has, it seems to me, no equal in his understanding of the strength of sceptical arguments. Neither should it be supposed that he was entirely happy with his arguments for scepticism– and not just because of the alleged brute psychological impossibility of endorsing them. There are a number of attempts to undermine scepticism in the Treatise although none of them, it must be admitted, is very successful. Still, Hume did appreciate– much more fully than Descartes, for example–the paradox as well as the power of scepticism.

¹⁷ A difficulty Wittgenstein runs into in certain contexts. Hume puts the point about arguments against scepticism with regard to reason being question-begging in temporal terms. First, scepticism holds sway in the mind, then the anti-sceptical or ‘dogmatical’ position, then scepticism... and so on. Fogelin too notices the relevance of Hume's discussion to Wittgenstein's project–although his interpretation of the issues is very different from mine. But he, like Hume, puts the point temporally–see pp. 228–9. Hume does say, however, that ‘the sceptical and dogmatical reasons are of the same kind’. Hume has, it seems to me, no equal in his understanding of the strength of sceptical arguments. Neither should it be supposed that he was entirely happy with his arguments for scepticism– and not just because of the alleged brute psychological impossibility of endorsing them. There are a number of attempts to undermine scepticism in the Treatise although none of them, it must be admitted, is very successful. Still, Hume did appreciate– much more fully than Descartes, for example–the paradox as well as the power of scepticism.

¹⁷ We all know that the only position to adopt is to be anti- Foundational, Realist about the physical world and anti-sceptical all at once; otherwise, we sound naive, mannered or outrageous. I know of no other sort of account than the above that would reconcile these views. In particular, it should be emphasized that recent discussions of closure don't help here, or not by themselves. See Nozick's Philosophical Explanations, Harvard, 1981, for some of the details. The critical point for closure or its absence is at the limit of our abilities. And unless sense can be made of that, unless it can be justified in some way or defended, the notion of closure poses no problem for the sceptic. Besides, there is nothing to prevent us, after we have vindicated the paradigms, from saying it follows, as it certainly seems to, that we know we're not dreaming, that an evil genius isn't deceiving us, that we're not brains in vats-and so on. Closure?

¹⁸ Romanticism is given a central place in that most romantic of books, Stanley Cavell's The Claim of Reason: Wittgenstein, Skepticism, Morality and Tragedy, Oxford, 1979-and in the discussion of scepticism, in particular. It is hardly central for me. And there is not much else that Cavell's treatment of scepticism has in common with mine. What I have said is also reminiscent–in some ways and vaguely at least–of certain Phenomenologists and Existentialists. Merleau-Ponty is perhaps the best of them on the topic of scepticism. But I hope I have put things more straightforwardly and, so to speak, more analytically than he ever does.

¹⁹ Descartes's remark can be found in Rules for the Direction of the Mind, Vol. I, p. 10. Wittgenstein might seem to claim that God is bound by our knowledge. But the irony should be clear: ‘... as soon as I say this sentence ["I know that I am now sitting in a chair"] outside its context, it appears in a false light. For then it is as if I wanted to insist that there are things that I know. God himself can't say anything to me about them’.–On Certainty, #554.

²⁰ I should emphasize that the above dismissive attitude towards anti- Realism has to do with physical reality, in particular. I take Nagel to have shown–in ‘What is it Like to be a Bat?’ Philosophical Review, 1974, reprinted in Mortal Questions, Cambridge, 1979–that we should indeed be anti-Realists, species by species, about mental reality. But what Nagel offers as a solution to the sceptical problem is no solution at all. I have the uncomfortable feeling of keeping a score-card or a log here. Nagel's suggestion, in The View from Nowhere, is that ‘the correct course is not to assign a victory to either [the objective or the subjective] standpoint but to hold the opposition clearly in one's mind without suppressing either element’. This is simply Hume's predicament with the label ‘Correct’ pasted over it

²¹ Our notion of knowledge or our employment of it is therefore such that while it comes with a sliding scale so the more intelligent you are, the more demanding it is or the less you are able to get away with, human beings do qualify as knowers–but don't listen to any stories about epistemologically- gifted chickens or frogs. The notion makes a gesture towards the significance, in the grand scheme of things but for us in particular, of that distinction. The above views about argument in philosophy are from the Investigations, #126 and #129; the remark about standing before the abyss of scepticism is from On Certainty, #370.