Lunacy and Scepticism: Notes on the Logic of Doubt Concerning the Existence of an External World

This article develops a logical (or semantic) response to scepticism about the existence of an external world. Specifically, it is argued that any doubt about the existence of an external world can be proved to be false, but whatever appears to be doubt about the existence of an external world that cannot be proved to be false is nonsense, insofar as it must rely on the assertion of something that is logically impossible. The article further suggests that both G. E. Moore and Ludwig Wittgenstein worked towards the same solution but left their work unfinished.

Regarding Moore's proofs, then, it is first of all important to note that he did not intend them to constitute a response to external world scepticism. 2 He made this quite clear, for example in the following passage that he wrote shortly after the one quoted above: I have sometimes distinguished between two different propositions, each of which has been made by some philosophers, namely (1) the proposition 'There are no material things' and (2) the proposition 'Nobody [p.1024] knows for certain that there are any material things'.And in my latest published writing, my British Academy lecture called 'Proof of an External World'-I implied with regard to the first of these propositions that it could be proved to be false in such a way as this; namely, by holding up one of your hands and saying 'This hand is a material thing; therefore there is at least one material thing'.But with regard to the second of those two propositions, which has, I think, been far more commonly asserted than the first, I do not think I have ever implied that it could be proved to be false in any such simple way.('A Reply to My Critics ', 1942, 668) Thus, what Moore is primarily interested in doing with his two proofs is to disprove idealism, which entails the thesis that no external world exists.But he takes scepticism about the existence of an external world to be of a different calibre.This distinction of Moore's can be surprising.After all, having a correct proof of an external world-or at any rate, understanding such a proof-entails knowing that an external world exists.But knowing that an external world exists may well be taken as the negation of external world scepticism.Evidently, Moore does not take it to be that.He expresses the supposedly relevant difference in terms of certainty, as when he specifies the second proposition above as being 'the proposition "Nobody knows for certain that there are any material things" '. 3   It follows from this that Moore's own later-developed response to external world scepticism does not, at least not obviously, count amongst what have 2 Even some of the best commentators are still getting this wrong.See, for instance, Schroeder 2023, 16.
come to be called Moorean responses to scepticism. 4 It will be useful to argue this point in some detail.To begin with, here is a standard formulation of an argument for scepticism: (1) If I know that p, then I know that sceptical scenario s regarding p is not the case.
(2) I do not know that sceptical scenario s regarding p is not the case.
(3) Hence, I do not know that p.
For example, s could be Descartes' evil demon who deceives me or the hypothesis that I am dreaming or that I am in a computer simulation.The socalled Moorean response is to turn an argument of this sort on its head; in this case, to turn it from a modus tollens (if A then B; not-B; hence, not-A) into a modus ponens (if A then B; A; hence, B).Thus: (1) If I know that p, then I know that sceptical scenario s regarding p is not the case.
(4) I know that p.
(5) Hence, I know that sceptical scenario s regarding p is not the case.
So, for instance, using 'Here is a hand' for p and the dream hypothesis for the sceptical scenario: (1′) If I know that here is a hand, then I know that I am not dreaming.
(4′) I know that here is a hand.
(5′) Hence, I know that I am not dreaming.
Although he explicitly mentions the above dialectic (see esp. 1959c, 247), Moore does not think it is sufficient for a response to scepticism nor, indeed, that it is necessary.The following four points should suffice to make this quite obvious.
First, Moore explains at the end of his 'Proof of an External World' why he has 'neither given nor attempted to give' a response to scepticism in that work, apart from the narrow sort of scepticism in the form of idealism (as noted above).For in order to do this in the form of a proof, as he correctly points out, he would have had to prove the propositions he used as premises.However, he says: Of course, in some cases what might be called a proof of propositions which seem like these can be got.… But I do not believe that any proof is possible in nearly all cases.How am I to prove now that 'Here's one hand, and here's another'?I do not believe I can do it.In order to do it, I should need to prove for one thing, as Descartes pointed out, that I am not now dreaming.But how can I prove that I am not? (1939,149) Second, in concluding his paper Moore explains why he thinks this does not mean that scepticism cannot be refuted, and what more would have to be done in his view-more, that is, than his two proofs of an external world-in order to achieve this goal: [p.1025] Kant [thought] that if I cannot prove that there is a hand here, I must accept it merely as a matter of faith-I cannot know it.Such a view, though it has been very common among philosophers, can, I think, be shown to be wrongthough shown only by the use of premisses which are not known to be true, unless we do know of the existence of external things.I can know things, which I cannot prove; and among things which I certainly did know, even if (as I think) I could not prove them, were the premisses of my two proofs.I should say, therefore, that those, if any, who are dissatisfied with these proofs merely on the ground that I did not know their premisses, have no good reason for their dissatisfaction.(1939,150) The thing Moore says here, again, that he believes he cannot prove is precisely what external world scepticism wishes to doubt.And, again, he makes it clear that this kind of doubt is not his present concern.In addition, he says that although he thinks he cannot prove the relevant propositions, that is, the premises upon which his two proofs against idealism rely, proponents of external world scepticism-that is, 'those … who are dissatisfied with these proofs merely on the ground that I did not know their premisses'-have, he says, 'no good reason for their dissatisfaction'.Thus, Moore thinks that it is only this latter kind of claim, i.e. that external world scepticism lacks good reason and (in the previous sentence) that 'I can know things, which I cannot prove', which constitutes the beginnings of a response to external world scepticism.And he is naturally aware that the alleged truth of these claims yet remains to be shown Third, in the final versions of the papers in which he eventually develops his response to scepticism, Moore still takes the correct formulation of scepticism to be unlike (1) or  (1959a).In 'Four Forms of Scepticism', he defines all four forms in terms of the clause 'that no human being ever knows with complete certainty anything whatever ' (1959b, 196).In 'Certainty', he offers a detailed discussion, on the basis of which he specifies his target thus: 'I have been and shall be only concerned with uses of "know" … such that "I know that p" does entail "I know with absolute certainty that p" ' (1959c, 236-7).
Fourth, Moore's strategy in responding to external world scepticism in these two papers constitutes an elaboration of his earlier claim, in 1939, that such scepticism would be absurd beyond repair: I knew that there was one hand in the place indicated by combining a certain gesture with my first utterance of 'here' and that there was another in the different place indicated by combining a certain gesture with my second utterance of 'here'.How absurd it would be to suggest that I did not know, but only believed it, and that perhaps it was not the case!You might as well suggest that I do not know that I am now standing up and talking-that perhaps after all I'm not, and that it's not quite certain that I am! (1939, Specifically, in 'Four Forms of Scepticism' and 'Certainty' Moore sets out to argue that external world scepticism, although it cannot be proved to be false, is in fact logically incoherent; more specifically, that any (real) doubt about the existence of an external world can be proved to be false, but whatever appears to be doubt about the existence of an external world that cannot be proved to be false is nonsense, insofar as it must rely on the assertion of something that (Moore believes) is logically impossible.
Thus, unlike so-called Moorean responses to scepticism, Moore's own response to scepticism turns not on switching around the onus of proof (as described above) but on showing that the sceptical challenge may only appear to make sense, whereas in reality it necessarily amounts to nonsense and is therefore self-undermining. 5 As is well known, Moore did not succeed in this endeavour.But his failure, as far as I can see, was due not to a mistake or thinking that he had committed one (see also his 1959a, preface). 6Rather, it was due mainly-or so I shall 5 Pryor 2004 independently recognises that this is the sort of complementary move that Moore's 1939 proof of an external world lacked in order to amount to a respectable philosophical response to external world scepticism, but he seems to have missed the fact that Moore actually made such a move in his later writings.I do not think it is rational to be as certain of any one of these [assumptions], as of the proposition that I do know that this is a pencil.And how on earth is it to be decided which of the two things it is rational to be most certain of?(1959b, 226) In 'Certainty', Moore continues this line of inquiry.However-as mentioned previously-he only gets so far.Thus, he ends the paper as follows: One final point should be made clear.It is certainly logically possible that I should have been dreaming now; I might have been dreaming now; and therefore the proposition that I am dreaming now is not self-contradictory.
But what I am in doubt of is whether it is logically possible that I should both be having all the sensory experiences and the memories that I have and yet be dreaming.The conjunction of the proposition that I have these sense experiences and memories with the proposition that I am dreaming does seem to me to be very likely self-contradictory.(1959c, 251) The remaining task for the present paper will be to explain why and how Moore was correct on this point.Specifically, I shall try to explain that Moore was completely right as far as he got, and that it was Wittgenstein who supplied at least the materials necessary to explain the reason-which Moore still owedfor its being logically impossible that someone should both be having all the sensory experiences and the memories that they have and yet be dreaming.
The relevant remarks by Wittgenstein-published in On Certainty (1969)-are all first-draft notes which were posthumously edited by Anscombe and von Wright, in full as they appeared in some of his final manuscripts, with additional numbering for ease of reference.These remarks are therefore quite unlike the text that went into his Philosophical Investigations (1953), which is another posthumously edited publication but also one that was the product of many years of his own writing, rewriting, selecting, editing and revising ('the precipitate of philosophical investigations which have occupied me for the last sixteen years', as he says in the printed preface).This does not mean that On obvious mistake.'To begin with, the following remark shows that Wittgenstein thought it possible to agree with Moore: 53.So one might grant that Moore was right, if he is interpreted like this: a proposition saying that here is a physical object may have the same logical status as one saying that here is a red patch.
A proposition saying that here is a red patch, supposing (like Wittgenstein) that this purely concerns what Moore calls 'sensory experience', is of course the kind of thing that even Russell would agree can be known in the face of external world scepticism.In fact, Wittgenstein gets Moore exactly right in this remark.
For, as we have seen, Moore ultimately arrives at precisely the view that it is possible that the relevant propositions have the same logical status, when he finally suggests that it may not be logically possible to doubt the knowledge he expresses when he says, 'Here is a hand' (etc.). 7 For an opposing view of the relation between On Certainty and the Investigations, according to which it would be as appropriate to speak of two Wittgensteins here as it is in the case of the Tractatus and the Investigations, see Moyal-Sharrock 2004 (esp. 163-5).In other words, any meaningful expression of an attempt to doubt everything without exception must, at the same time, be necessarily false, because the assumption that the expression is meaningful-hence, intelligible to someoneimplies that someone knows something (hence, that it is not the case that nobody knows anything whatsoever) insofar as whoever understands the expression thereby knows its meaning.Thus, it is not logically possible to doubt that anybody knows anything whatsoever (which may also be expressed by saying that universal scepticism amounts to nonsense).
The question of the present inquiry then becomes whether the same strategy can be applied to the narrower thesis of scepticism about the existence of an external world.Moore did think that this narrower form of scepticism also involved a logical impossibility, but he found himself unable to justify his view.
Wittgenstein made further progress in this direction.
32. It's not a matter of Moore's knowing that there's a hand there, but rather we should not understand him if he were to say 'Of course I may be wrong about this'.We should ask 'What is it like to make such a mistake as that?'e.g.what's it like to discover that it was a mistake?
In the above remark, Wittgenstein not only seems (again) to accept that Moore is right to say he knows that there is a hand-which Wittgenstein also sometimes (wrongly) questioned-but, at the same time, suggests that external world scepticism (the thesis, as Moore expresses it, that nobody knows for certain that there are any material things) might be unintelligible. 8He goes on Wittgenstein's remarks in On Certainty have often been interpreted as advancing a kind of parallelism according to which the meaningfulness of expressions of doubt and expressions of knowledge stand and fall together.There are passages in Wittgenstein's work that suggest such an interpretation, in the Investigations (e.g.section 288) as well as in On Certainty, but as a general principle both this kind of parallelism and the interpretation of Wittgenstein along such lines seem wrong to me.I cannot argue the general point here in detail-and, to be sure, the question is not essential to the anti-sceptical strategy I am developing in this paper (see note 13 below)-but in my view it should be clear to anyone who has made a careful study of the later Wittgenstein that wherever a genuine attempt to express doubt results in nonsense it does not follow, at least not without to say that what should be done to investigate this question of intelligibility is to ask about the nature of the mistake that this scepticism would appear to entail.
The first sixty-five remarks in On Certainty, which the editors believed to be the earliest of the material, were found on twenty loose sheets without dates; the remaining text stems from three small notebooks that were continuously dated and which, thus, were added in chronological order beginning with sections 66, 193 and 300 respectively.The first of these notebooks begins with the above-mentioned investigation into the nature of the mistake that external world scepticism would appear to entail.
66.I make assertions about reality, assertions which have different degrees of assurance.How does the degree of assurance come out?What consequences has it?
We may be dealing, for example, with the certainty of memory, or again of perception.I may be sure of something, but still know what test might convince me of error.I am e.g.quite sure of the date of a battle, but if I should find a different date in a recognized work of history, I should alter my opinion, and this would not mean I lost all faith in judging.
67. Could we imagine a man who keeps on making mistakes where we regard a mistake as ruled out, and in fact never encounter one?E.g. he says he lives in such and such a place, is so and so old, comes from such and such a city, and he speaks with the same certainty (giving all the tokens of it) as I do, but he is wrong.
But what is his relation to this error?What am I to suppose? 68.The question is: what is the logician to say here?It may or may not be a coincidence that the cognitive modalities that Wittgenstein introduces as examples in the second paragraph-namely, considerable further ado, that one cannot make a meaningful claim to knowledge regarding the same thing; in other words, that the logic of doubt is never simply the negation of knowledge.For useful, detailed commentary in this connection, both on relevant passages in On Certainty and the Investigations, see Conant 1998 (e.g. 241-3) andSchroeder 2023 (esp. 16-20 and28-30).By taking it that Wittgenstein accepts Moore's claim to knowledge, I am further siding with Baldwin 2011and Glock 2004, but siding against, for instance, Comesaña and Klein 2001/19, Hamilton 2014, McGinn 1989, Schönbaumsfeld 2016, and Stroll 1994 as well as the influential hingeepistemological reading of On Certainty as advanced, perhaps most prominently, in Coliva 2010, 2022, Coliva and Pritchard 2022, Moyal-Sharrock 2004, 2017, and Pritchard 2015, 2017.But in particular this is to side against Malcolm 1949, whose central claim was that Moore had misused 'know', 'know with certainty', and so on.It was arguably the perceived imperfection of this criticism by Malcolm which motivated Wittgenstein's review of Moore's anti-sceptical strategy in On Certainty (see also Kober 1996Kober /2018, 441-3), 441-3).See also, purely on the question of the linguistic legitimacy of Moore's uses of 'know', etc., Clarke 1972 (756-9) andStroud 1984 (ch. 3).In the second of the above remarks, then, Wittgenstein proposes (and invites us to consider the conceivability of) the case of someone whose memory or perception would seem to be consistently incorrect given that they are consistently wrong in what they say when they speak ('with the same certainty … as I do') about things 'where we regard a mistake as ruled out, and in fact never encounter one'.
Lastly, in the third remark, Wittgenstein notes that his consideration of this example is intended to be logical in nature; in other words, he takes it to be primarily a consideration of the logical possibility of a supposed kind of error.
His reflections on this particular case culminate in the following negative verdict regarding the possibility of such an error: 71.If my friend were to imagine one day that he had been living for a long time past in such and such a place, etc. etc., I should not call this a mistake, but rather a mental disturbance, perhaps a transient one.
72.Not every false belief of this sort is a mistake.

But what is the difference between mistake and mental disturbance?
Or what is the difference between my treating it as a mistake and my treating it as mental disturbance?
This last question he answers correctly, it seems to me, when he asks himself: 75.Would this be correct: If I merely believed wrongly that there is a table here in front of me, this might still be a mistake; but if I believe wrongly that I have seen this table, or one like it, every day for several months past, and have regularly used it, that isn't a mistake?
It is worth noting the structural parallel between the scenario Wittgenstein is considering here and what Moore is considering when he says, 'The conjunction of the proposition that I have these sense experiences and memories with the proposition that I am dreaming does seem to me to be very likely selfcontradictory ' (1959c, 251).Wittgenstein's and Moore's investigations are clearly closely associated at this point.Both want to say that the proposition that I have the sense experiences and (extensive) memories on the basis of which I make a judgement entailing the existence of an external world and expressed in a statement such as 'Here is a hand' or 'There is a table here' can be inconsistent with the proposition that I am possibly mistaken in this judgement.And it seems to me that the special nature of the supposed error, whose possibility Moore and Wittgenstein wish to deny, is what led both of them to prefer descriptions of the corresponding (hence, equally special) cognitive state that, in their view, actually occurs in this type of case in terms of certainty, or knowledge with complete certainty, rather than simply knowledge.
Supposing that Moore and Wittgenstein are right, at least thus far, obviously raises the question of how the relevant difference might be explained.
What are the conditions under which these two propositions will become inconsistent?
We have already seen that in this respect, too, the two philosophers appear to be in significant agreement with each other.Both believe an explanation is to be found in logic.Moore stops at this point, due (in my view) to a lack of time.
Wittgenstein offers the following elaboration: 80.The truth of my statements is the test of my understanding of these statements.
81.That is to say: if I make certain false statements, it becomes uncertain whether I understand them.
82.What counts as an adequate test of a statement belongs to logic.… Thus, Wittgenstein is saying that the truth of the relevant statements, such as Moore's saying 'Here is a hand' (etc.), belongs to logic-in other words, that it is a logical necessity that they (or, more precisely, a significant proportion of them) are true-because if I attempted, in all seriousness, to make statements to the contrary I would stop making sense. 10Wittgenstein reiterates this point in the following remark: [p.1029] 155.In certain circumstances a man cannot make a mistake.('Can' is here used logically, and the proposition does not mean that a man cannot say anything false in those circumstances.)If Moore were to pronounce the opposite of those propositions which he declares certain, we should not just not share his opinion: we should regard him as demented.

And finally:
219.There cannot be any doubt about it for me as a reasonable person.-That'sit.-210.The reasonable man does not have certain doubts.
According to Wittgenstein, then, external world scepticism-the thesis that nobody knows for certain that there are any material things-is necessarily false, because it supposes that there could be a mistake which, in fact, is logically impossible, due to the fact that assuming it will lead to the production of nonsense-in particular, apparently false statements of the sort 'I do not know that here is a hand' (etc.)-thesum of which, in one person, amounts not to a series of mistakes but to insanity.So, for these reasons Wittgenstein agrees with Moore that any doubt about the existence of an external world can be proved to be false, but whatever appears to be doubt about the existence of an external world that cannot be proved to be false is nonsense, insofar as it must rely on the assertion of something that (both Moore and Wittgenstein believe) is logically impossible.
Up until this point, we have considered external world scepticism merely formally, namely, as the thesis that nobody knows for certain that there are any material things.I have said that Moore and Wittgenstein agree this is false.
Wittgenstein clearly implies, and Moore would surely agree, that any reasonable person in fact, by logical necessity, knows that there are material things insofar as they know, by logical necessity, propositions such as those that Moore proved in his two proofs of an external world.But at this point it might be objected that a sceptical scenario is precisely what will motivate external world scepticism in such a way that it will not result in nonsense.This is a plausible objection, which must now be addressed.For example, the sceptical hypothesis that I am dreaming may be thought to give the kind of motivation required.When asked why one denies statements such as Moore's one may try to give a reason by saying that one does not know one is not dreaming.Moore explicitly said he believes this does not work, but he did not elaborate on his reasons for so believing.Wittgenstein did not leave a useful response to the objection either. 11So what should we say?
Although I believe that the general strategy presented in the present paper is in fact sufficient, sceptical scenarios do deserve specific treatments, if only in order to break the stronghold that they tend to exert over our intuition.But I think that the strategy presented in this paper will invariably serve as a useful basis in practical situations too, insofar as its general application in debate with our old (imaginary) friend, the external world sceptic, may perhaps be formulated as follows (where p is a variable for the propositions Moore declared certain, and pn stands for 'an appropriate number of similar propositions' 12 ). 13(i) If external world scepticism is true, then you are doubting both that p and pn.
(ii) Now, either you know that p or you appear to be doubting that p.
(iii) Of course, if you know that p, then you are not doubting that p.
(iv) But if you appear to be doubting that p, then either you are doubting that p or you are not.

(v)
And if you are doubting that p, then you appear to be doubting that p.
(vi) Yet, if you appear to be doubting both that p and pn, then you are not doubting both that p and pn.
(vii) You appear to be doubting both that p and pn.
(viii) Hence, external world scepticism is false.
Premises (i), (iii), and (iv) should be unproblematic.Premise (ii) is obviously simplifying matters, but not unreasonably so: in reality, one may be doubting that p without appearing to be doubting that p, but the present case is that of someone-our aspiring sceptic-who wishes to doubt both that p and many similar propositions; indeed, so many similar propositions that it would seem unrealistic to think this could be done without the appearance of doing so (i.e. [p. 1030] without saying things or otherwise behaving in ways that are consistent with a sceptical desire or intention of this kind).The same holds for at dealing with the dream hypothesis in sections 383 and 676 of On Certainty fails.But I believe that there is a better response available in that work than the one Baldwin briefly sketches, as it is possible to successfully apply the anti-sceptical strategy that I have argued can be extracted from On Certainty (as is demonstrated in the final paragraphs of this paper).
12 'Appropriate' here means appropriate to p on the condition of (vi).
13 Notice that the argument does not require that p, or pn, is known (see also note 8 above).The schema given in the text, (i)-(viii), includes the Moorean part of the dialectic.The Wittgensteinian part alone may be put as follows (where  abbreviates 'both that p and pn'): If external world scepticism is true, then you are doubting .(v′) If you are doubting , then you appear to be doubting .
(vi) Yet, if you appear to be doubting , then you are not doubting .(viii) Hence, external world scepticism is false. premises (v) and (vii).The reason why, in (vi), the subject will stop doubting is precisely because their words will stop being meaningful, as explained in the second half of this paper.
In the case of dreaming, then, I think Moore was exactly right to question 'whether it is logically possible that I should both be having all the sensory experiences and the memories that I have and yet be dreaming ' (1959c, 251).
But in addition to making the requisite substitutions in the argument above, I believe that one can demonstrate this to oneself in the form of the following exercise.Take some small-sized object available to you (perhaps one of your hands) and inspect it closely, focusing on this task for at least twenty seconds with the aim of getting as many different and rich sensory experiences as possible; now try to remember what you did yesterday; alternatively, try to remember why you ended up reading this paper right now.I think we can agree that neither of us have ever dreamt so vividly.'Surely, it is not impossible to dream so vividly?', someone may respond.But, yes, it is indeed impossible: if anyone's dreams ever come close to resembling the sorts of sensory experiences and memories you will have had if you followed my instructions, not to mention the coherent thinking that would normally go along with it, we will distinguish it from what we call 'dreams' and will do so easily as well as correctly.In other words, your having all the sensory experiences and the memories that you have logically excludes the possibility that you are dreaming, because neither you nor anyone else has ever dreamt so life-like; it is not what is called 'dreaming', and if it came to be (correctly) called 'dreaming', that new notion of dreaming would be essentially different from ours today.This also provides a possible answer to the common rejoinders 'What if you never wake up?' and 'What if we develop the technology to induce perfectly life-like dreams?' Another possible answer to these questions is what can be said in response to the evil demon and computer simulation scenarios.
The hypothesis that an evil demon may constantly deceive me contains some special difficulties, but for present purposes I will treat it, like many people nowadays do, as simply the stylistic predecessor of the computer simulation.Whilst Moore rightly suspected that the conjunction of the proposition that I have the sense experiences and memories as described in the previous paragraph with the proposition that I am dreaming is selfcontradictory (which indeed it is), it is not logically impossible that all of you and me and all the rest should be part of one big computer simulation.The word 'simulation' can be misleading here; it is hopefully clear that the proponent of external world scepticism cannot simply derive from the dictionary meaning of this word that anything inside the supposed simulation is not real, and so on.
On the contrary, it seems that the discovery that the computer simulation hypothesis is true would just be an interesting, possibly scary, discovery about the fundamental structure of this world. 14It would not be the first such discovery.Maybe physics will discover a physical foundation of this world that is different from what I thought; that will not show that all I took myself to know about the physical world was in fact false, or indeed that I did not know those things.The same will be true if that discovery is that my whole world has, until now, somehow been dependent on a computer program.My knowledge will not shrink, it will increase.
memory and perception-are [p.1028] also Moore's favoured examples for refuting external world scepticism. 9 Certainty does not contain brilliant philosophical work, but it does mean that it also contains raw material.At the same time, I agree with von Wright's assessment when he writes, 'Considering that the remarks constitute a first, Here, however, I will simply pick and choose what seems valuable to the line of argument that interests me.Whilst putting the philosophy first in this way is obviously not in itself a historically sensitive method, the result will ultimately have to be measured for both its philosophical and historical accuracy, and I shall of course offer textual evidence whenever and to the degree that it is at all convenient to do so.Hence, although I will refrain from entering into exegetical discussion in this paper, everything I say about Wittgenstein's text is, to the best of my knowledge, historically accurate.In particular-and this will not be entirely uncontroversial-everything I say about Wittgenstein's text is, in my view, perfectly consonant not only with the rest of On Certainty but also Philosophical Investigations (as, I think, it should be). 7 The argumentative strategy I now want to develop, following Moore and Wittgenstein, is illustrated in the following two remarks by Wittgenstein concerning not external world but universal scepticism, that is, the thesis that