Wittgenstein ’ s Critique of Moore in On Certainty

In this paper, I clarify Wittgenstein’s critique of Moore in On Certainty. I argue that this critique is largely misunderstood, for two reasons. Firstly, Wittgenstein partly misrepresents Moore. Secondly, Wittgenstein is wrongly taken to be an (access-) internalist regarding justification of knowledge. Once we take these two points into account we can understand Wittgenstein’s critique properly as a grammatical argument according to which Moore fails to see how the concepts of knowledge and certainty relate to those of justification and evidence. This reading shows that Moore and Wittgenstein were in closer agreement than many people have thought, even though Moore was not able to exploit and express the philosophical insights (which he shared with Wittgenstein) properly. 1. Moorean Propositions In “A Defence of Common Sense” (DCS) G. E. Moore claimed that there are certain “obvious truisms”, such as “The earth has existed for many years” and “I am a human being”, which everybody knows for certain (DCS: 33, 55). These ‘Moorean propositions’ are definitely true, and even though we do not know what their supporting evidence is, Moore thinks it would be the height of absurdity to talk with contempt of them (DCS: 44–45). In “Proof of an External World” (PEW) Moore describes how he holds up his hands and says, “Here is one hand, and here is another” in order to prove that “Two human hands exist”. He claims that he also knows Erlend W.F. Owesen CC-BY 72 these propositions to be true, and that it would be absurd to suggest that he did not know but only believed that they were true, or that they were not true (PEW: 146). According to Wittgenstein, the propositions Moore brought to light were interesting, but he thought that Moore was misusing the expression “I know”. This criticism has notoriously puzzled both commentators and critics ever since the publication of On Certainty (OC). The problem with understanding the nature of Wittgenstein’s criticism has, I argue, two main sources. First, Wittgenstein partly misrepresents Moore. Second, Wittgenstein is wrongly taken to be an (access-) internalist regarding justification of knowledge. In the following I will try to clarify why Wittgenstein thought that Moore was mistaken. I will also argue that there is more agreement between Wittgenstein and Moore than many interpreters have suggested. 2. Wittgenstein’s First Objection: Context and Sense The first argument Wittgenstein directs at Moore’s use of “I know” has to do with context of utterance and determinacy of sense. Considering the multiplicity of variously connected uses of “I know” Wittgenstein thinks that when Moore enumerates what he knows, out of the blue (so ohne weiteres), he is misusing the expression (OC: § 6). Presumably, the point is that the particular occurrences of “I know” are transmitted from the many distinct and peculiar contexts to which they belong, to one single context (Moore’s philosophical discussion) to which they do not belong. Wittgenstein’s point seems to be that this ostensibly innocent move affects the sense of these propositions so that we no longer understand what they mean. For Wittgenstein thinks that only the accustomed context allows what is meant to come through clearly (OC: § 237). And we can easily imagine circumstances in which a Moorean proposition could 1 Paul and Anscombe’s translation of the German idiom so ohne weiteres as “straight off like that” is misleading in this context, because the charge is not that Moore makes his claims straightaway without reflecting, but that he makes them ‘out of the blue’ (which is a better translation of the idiom, and which I use accordingly). That Wittgenstein takes Moore in this way is also evident form his examples of someone who in the middle of a conversation says “Good morning!” or “Down with him!” or “I am here!”, without any intelligible background or circumstances (OC: § 348, 350, 464). Nordic Wittgenstein Review 6 (2) 2017 | pp. 71-84 | DOI 10.15845/nwr.v6i2.3440 73 function as a move in one of our language-games (OC: § 622). For instance, someone who wanted to conduct a dangerous experiment in order to attempt to fly in the face of excessive warnings and criticisms, such as “Remember you’re only a human being!”, might say: “I know that I’m a human being! The experiment is safe, trust me.” However, uses like these are not the philosophically interesting ones with which Moore is concerned. Wittgenstein’s argument runs roughly along the following lines: If the context of utterance is not specific or specifiable, the sense of the utterance is undetermined. Moore does not provide a specific context, and hence, the sense of his utterances is undetermined. The point is not that such an utterance would be simply superfluous, but that the sense is not determined by the situation in a way that is needed (OC: § 348). This is, as Conant, following Travis, elaborates, because the total speech situation – the unitary whole of utterance and context – serves to constitute the content of what is claimed. And in these cases, where, according to Conant, there is no intelligible relation between utterance and context, there is no determinable truth-evaluable content either. As a result, Conant concludes, it is not clear what, if anything, Moore is saying. Moore’s failure lies in his not providing an answer to the question: What are the sentences in which “know” occurs supposed to mean in this situation? (Conant 1998: 241; 2011: 411–412). Regardless of whether (on Conant’s interpretation) Wittgenstein is right in that sense is interconnected with context of utterance, this criticism does not seem devastating for Moore. For he can easily answer Conant’s question and specify the context of his utterance – namely, that he is responding to philosophers who have doubted and found uncertain what he takes to be truisms (cf. DCS: 32, 38, 53). Moore is making epistemic claims in the face of philosophers’ doubts and denials. “Proof of an External World” was even delivered as a British Academy lecture in 1939. It is difficult to see why that is not a specific enough context for the utterance of the sentences in question. Presumably, in response to Moore’s specification of context, Conant and Travis would claim that the philosophical context or use is not a “genuine context” (Conant 2011: 412) or a “proper use” Erlend W.F. Owesen CC-BY 74 (Travis 1989: 151). But to show that would require a different and more compelling argument than merely insisting that Moore is “speaking outside of natural enough circumstances” (Travis 1989: 155). Travis and Conant do not provide any other arguments to support their view. Now, speaking in new and unnatural circumstances is surely something we do from time to time, and we do so without thinking that our utterances are devoid of sense. Besides, if “natural circumstances” is merely supposed to mean nonphilosophical, the distinction that is currently at issue and in need of argument is presupposed, and so the charge of speaking without a “genuine context” or “proper use” (cashed out in terms of speaking outside natural circumstances) begs the question. Moreover, if the point is to abandon philosophical contexts altogether, what makes Conant’s and Travis’ claims any less indeterminate? To be sure, it might be possible to argue that the sceptic is talking without an intelligible context, and so argue that Moore’s response is just as unintelligible as the sceptical claims are. But this is not how Wittgenstein, Conant, and Travis explicitly argue when complaining about a lack of specific context. For when Wittgenstein complains about indeterminacy of sense, he seems to misrepresent Moore as talking “out of the blue” (OC: § 6), rather than giving compelling arguments for the unintelligibility of Moore’s claims. Similarly, and perhaps misled by this, Travis says that “Moore thought he could catalogue the things he knew, or many of them, any time he liked” (Travis 1989: 162). However, Moore is not making these claims “any time he liked”, or “out of the blue”. He is making them in response to opposing claims made by philosophers in a specific context of a philosophical polemic. Hence, the first objection Wittgenstein directs at Moore, which Travis and Conant advocate and elaborate on, seems, as it stands, unsatisfactory, or at least not properly developed. Nevertheless, to say that the objection, as here formulated, is unsatisfactory due to misrepresentation is not to say that Moore succeeds in saying something meaningful. For Wittgenstein has another argument, to which I now turn. Nordic Wittgenstein Review 6 (2) 2017 | pp. 71-84 | DOI 10.15845/nwr.v6i2.3440 75 3. Wittgenstein’s Second Objection: Justification Wittgenstein’s second objection to Moore’s use of “I know” grants him the context and use of “I know” with which Moore is concerned – namely, his responding to philosophers’ doubts. Yet, Wittgenstein believes that Moore has overlooked some implications of this use. First he notes that merely giving the assurance “I know” does not suffice for Moore’s objective – answering those who express doubt – for that he knows remains to be shown (OC: § 14–15). In other words, Wittgenstein is complaining that in so far as Moore’s knowledge claims are supposed to answer sceptical doubts they are question-begging and presuppose what they are supposed to defend. That is the first problem. With this criticism, Wittgenstein is exploiting the connection between knowing and giving grounds, that is, justification. The second, and deeper problem with Moorean propositions is not only that Moore has in fact not given any evidence or reasons for claiming to know that they are true, it is that (under normal circumstances) nothing would count as evidence and justification for the assertions in question. Therefore, the quest for justification makes no sense. Let me explain. The reason why nothing counts as evidence or justification for Moorean propositions i


Moorean Propositions
In "A Defence of Common Sense" (DCS) G. E. Moore claimed that there are certain "obvious truisms", such as "The earth has existed for many years" and "I am a human being", which everybody knows for certain (DCS: 33, 55).These 'Moorean propositions' are definitely true, and even though we do not know what their supporting evidence is, Moore thinks it would be the height of absurdity to talk with contempt of them (DCS: 44-45).In "Proof of an External World" (PEW) Moore describes how he holds up his hands and says, "Here is one hand, and here is another" in order to prove that "Two human hands exist".He claims that he also knows these propositions to be true, and that it would be absurd to suggest that he did not know but only believed that they were true, or that they were not true (PEW: 146).According to Wittgenstein, the propositions Moore brought to light were interesting, but he thought that Moore was misusing the expression "I know".This criticism has notoriously puzzled both commentators and critics ever since the publication of On Certainty (OC).The problem with understanding the nature of Wittgenstein's criticism has, I argue, two main sources.First, Wittgenstein partly misrepresents Moore.Second, Wittgenstein is wrongly taken to be an (access-) internalist regarding justification of knowledge.In the following I will try to clarify why Wittgenstein thought that Moore was mistaken.I will also argue that there is more agreement between Wittgenstein and Moore than many interpreters have suggested.

Wittgenstein's First Objection: Context and Sense
The first argument Wittgenstein directs at Moore's use of "I know" has to do with context of utterance and determinacy of sense.Considering the multiplicity of variously connected uses of "I know" Wittgenstein thinks that when Moore enumerates what he knows, out of the blue (so ohne weiteres), he is misusing the expression (OC: § 6). 1 Presumably, the point is that the particular occurrences of "I know" are transmitted from the many distinct and peculiar contexts to which they belong, to one single context (Moore's philosophical discussion) to which they do not belong.Wittgenstein's point seems to be that this ostensibly innocent move affects the sense of these propositions so that we no longer understand what they mean.For Wittgenstein thinks that only the accustomed context allows what is meant to come through clearly (OC: § 237).And we can easily imagine circumstances in which a Moorean proposition could function as a move in one of our language-games (OC: § 622).For instance, someone who wanted to conduct a dangerous experiment in order to attempt to fly in the face of excessive warnings and criticisms, such as "Remember you're only a human being!", might say: "I know that I'm a human being!The experiment is safe, trust me."However, uses like these are not the philosophically interesting ones with which Moore is concerned.
Wittgenstein's argument runs roughly along the following lines: If the context of utterance is not specific or specifiable, the sense of the utterance is undetermined.Moore does not provide a specific context, and hence, the sense of his utterances is undetermined.The point is not that such an utterance would be simply superfluous, but that the sense is not determined by the situation in a way that is needed (OC: § 348).This is, as Conant, following Travis, elaborates, because the total speech situation -the unitary whole of utterance and context -serves to constitute the content of what is claimed.And in these cases, where, according to Conant, there is no intelligible relation between utterance and context, there is no determinable truth-evaluable content either.As a result, Conant concludes, it is not clear what, if anything, Moore is saying.Moore's failure lies in his not providing an answer to the question: What are the sentences in which "know" occurs supposed to mean in this situation?(Conant  1998: 241; 2011: 411-412).
Regardless of whether (on Conant's interpretation) Wittgenstein is right in that sense is interconnected with context of utterance, this criticism does not seem devastating for Moore.For he can easily answer Conant's question and specify the context of his utterancenamely, that he is responding to philosophers who have doubted and found uncertain what he takes to be truisms (cf.DCS: 32, 38, 53).Moore is making epistemic claims in the face of philosophers' doubts and denials."Proof of an External World" was even delivered as a British Academy lecture in 1939.It is difficult to see why that is not a specific enough context for the utterance of the sentences in question.
Presumably, in response to Moore's specification of context, Conant and Travis would claim that the philosophical context or use is not a "genuine context" (Conant 2011: 412) or a "proper use" (Travis 1989: 151).But to show that would require a different and more compelling argument than merely insisting that Moore is "speaking outside of natural enough circumstances" (Travis 1989:  155).Travis and Conant do not provide any other arguments to support their view.Now, speaking in new and unnatural circumstances is surely something we do from time to time, and we do so without thinking that our utterances are devoid of sense.Besides, if "natural circumstances" is merely supposed to mean nonphilosophical, the distinction that is currently at issue and in need of argument is presupposed, and so the charge of speaking without a "genuine context" or "proper use" (cashed out in terms of speaking outside natural circumstances) begs the question.Moreover, if the point is to abandon philosophical contexts altogether, what makes Conant's and Travis' claims any less indeterminate?
To be sure, it might be possible to argue that the sceptic is talking without an intelligible context, and so argue that Moore's response is just as unintelligible as the sceptical claims are.But this is not how Wittgenstein, Conant, and Travis explicitly argue when complaining about a lack of specific context.For when Wittgenstein complains about indeterminacy of sense, he seems to misrepresent Moore as talking "out of the blue" (OC: § 6), rather than giving compelling arguments for the unintelligibility of Moore's claims.Similarly, and perhaps misled by this, Travis says that "Moore thought he could catalogue the things he knew, or many of them, any time he liked" (Travis 1989: 162).However, Moore is not making these claims "any time he liked", or "out of the blue".He is making them in response to opposing claims made by philosophers in a specific context of a philosophical polemic.Hence, the first objection Wittgenstein directs at Moore, which Travis and Conant advocate and elaborate on, seems, as it stands, unsatisfactory, or at least not properly developed.Nevertheless, to say that the objection, as here formulated, is unsatisfactory due to misrepresentation is not to say that Moore succeeds in saying something meaningful.For Wittgenstein has another argument, to which I now turn.

Wittgenstein's Second Objection: Justification
Wittgenstein's second objection to Moore's use of "I know" grants him the context and use of "I know" with which Moore is concerned -namely, his responding to philosophers' doubts.Yet, Wittgenstein believes that Moore has overlooked some implications of this use.First he notes that merely giving the assurance "I know" does not suffice for Moore's objective -answering those who express doubt -for that he knows remains to be shown (OC: § 14-15).In other words, Wittgenstein is complaining that in so far as Moore's knowledge claims are supposed to answer sceptical doubts they are question-begging and presuppose what they are supposed to defend.That is the first problem.With this criticism, Wittgenstein is exploiting the connection between knowing and giving grounds, that is, justification. 2 The second, and deeper problem with Moorean propositions is not only that Moore has in fact not given any evidence or reasons for claiming to know that they are true, it is that (under normal circumstances) nothing would count as evidence and justification for the assertions in question.Therefore, the quest for justification makes no sense.Let me explain.
The reason why nothing counts as evidence or justification for Moorean propositions is that there is nothing more certain than these propositions (cf.OC § 245).When we give grounds and appeal to evidence in the face of doubt -that is, when we are justifying our knowledge claims -we appeal to something more certain to support what is doubted.Otherwise, we will have just as much reason to doubt the ground or evidence, and then the ground or evidence cannot be coherently thought of as playing the supporting role required.Thus, if we question that the earth has existed for many years, it is unlikely that we will be convinced by checking the historical evidence.Nor would my seeing myself in the mirror constitute a ground for claiming to know that I am a human being.If we doubt the truth of Moorean propositions, we have no reason not to doubt this alleged evidence either (cf.OC: § 125), and it is no more certain than the propositions it is supposed to support.In fact, it is not clear what anyone with such doubts would let count as evidence and what not (OC: § 231).Since nothing (under normal circumstances) is more certain than these propositions, we cannot understand what it would be for something to serve as supporting grounds.Consequently, Moorean propositions are not grounded or justified by reference to something independent and more certain.Indeed, the very idea of justification here is misplaced.
Why do these points about justification matter for Moore's use of "I know"?In the relevant use of "I know" (defending a knowledge claim in the face of doubt and denial) it always makes sense to ask for justification and the source of one's knowledge.In other words, the logical possibility (that is, whether it makes sense) of asking for justification and how one knows (sources of knowledge) is part of the grammar of "I know".However, it does not make sense here to ask for evidence or justification, for nothing counts as a ground for these claims, and thus Moorean propositions fail to meet this grammatical criterion.Hence, whatever our relationship to these propositions may be, it is not a state of believing based on available evidence.Moreover, in claiming to use the expression "I know" in the relevant way (implicitly at least, since he wants to combat sceptical doubts), Moore is thus misusing the expression "I know".
This grammatical argument has misled commentators and critics.White thinks Wittgenstein believes that nothing can be known which is not actually arrived at by considering its grounds (1986: 323-324), and accordingly criticises Wittgenstein on that basis.Clearly, such a view would be objectionable, but it is not Wittgenstein's.Coliva, Moyal-Sharrock, and McGinn think the point is not that one actually needs to provide justification, but that it must be available to the knower.Thus, Coliva claims that "Wittgenstein is deeply rooted in an internalist conception of knowledge, according to which, in order to know that p, it must be possible, at least in principle, to make explicit one's warrants for -supporting grounds […] for p." (2010:  208-209).Moyal-Sharrock also thinks that "Wittgenstein adheres to the standard view of knowledge as justified true belief" (2007: 15), and that in the case of knowledge, "a commitment is made to justification" (2007: 23).Similarly, McGinn seems to adhere to this idea when she, commenting on Moore, says, "Possessing an adequate justification for believing a given proposition, p, is part of our normal ground for asserting a claim to know p" (1989: 50).So although they would agree that White's criticism misfires, Coliva, Moyal-Sharrock and McGinn all take Wittgenstein to hold that the subject's possibility of providing justification, that is, having access to such, is a necessary condition for possessing knowledge.In this way, they take Wittgenstein to be an access internalist.
This interpretation is questionable since the connection I explained above, which Wittgenstein thinks holds between knowledge and justification, does not seem to be of the form envisaged in access internalism.However, since he thinks there is a logical or conceptual, and thus an "internal" connection (in a different sense) between knowledge and justification -in that questions about justification must make sense if one claims to know something -he may be called, in this specific sense, an internalist. 3evertheless, we must be careful to distinguish internalism as a logical point about requirements for sense from the epistemological doctrine about requirements for the epistemic subject as envisaged by these commentators. 4f the access internalist understanding of Wittgenstein's account of the relationship between justification and knowledge were correct, then Wittgenstein's criticism of Moore, which clearly draws on that relationship, would be susceptible to a number of familiar objections.Counterexamples to access internalism are easy to construct.For one does not always remember the source of one's knowledge.For instance, if one does not remember how one came to know where some famous building is located, this does not preclude one from knowing where it is.In addition, there are other arguments, such as the chicken-sexer case where one does not even realize what the source of one's knowledge is, which pose serious 3 Hamilton neglects this sense of "internalism" when he suggests that Wittgenstein should be located outside the internalism-externalism divide (2014: 301).I agree, however, that it is difficult to argue that Wittgenstein belongs to either category understood in the forms they take in contemporary debates. 4Schönbaumsfeld (2016a) has presented a reading that is close to mine, for she also takes Wittgenstein's point to be logical (about sense), rather than epistemic.However, this is sometimes misleadingly conveyed when she uses formulations such as "if knowledge is to be possible…" (2016a: 168), where it might be more appropriate to say, e.g., "if knowledgeclaims are to make sense…".
difficulties for an account holding that the subject must have access to justification.
Although counterexamples like these pose problems for the view the mentioned commentators ascribe to Wittgenstein, I do not think he is committed to any such view.To be sure, he does say things like "Whether I know something depends on whether the evidence backs me up or contradicts me" (OC: § 504).Passages like these may induce one to take Wittgenstein in the way these commentators do.But here, as in similar passages, he is not advancing a general thesis about requirements for knowledge, but talks about specific examples of knowledge claims.The context of the quoted passage is a discussion of Moorean propositions, and so Wittgenstein criticizes Moore's "lack" of justification, thereby his use of "I know".Another example that has misled commentators is § 438, which White cites  (1986: 322) in his criticism of Wittgenstein.It says, "It would not be enough to assure someone that I know what is going on at a certain place -without giving him grounds that satisfy him that I am in a position to know" (OC: § 438).Despite an appearance to the contrary, what Wittgenstein says here is not opposed to the understanding I recommend.For in this particular case it is surely true that I must give grounds, at least that I must be able (as Moyal-Sharrock, Coliva, and McGinn would insist) to provide such grounds.If I say that I do not remember how I came to know, but I do know that the neighbors are having sheppard's pie for supper, we would certainly not recognize that as knowing.To repeat, this talk about particular cases is not indicative of a general thesis about knowledge.
Moreover, there are passages that affirm my interpretation, contrary to the mentioned commentators' interpretations.Wittgenstein says: "'I know' often means: I have the proper grounds for my statement" (emphasis added, OC: § 18).The qualifier "often" is important, since one can, as mentioned, come to know something, but not remember or even realize how one came to know it, but nevertheless know it.For even though the subject is not able to provide any reasons for her knowledge claim, or to say how she knows it, the question of how she came to know and what her reason(s) is does make sense.The question has an answer, although it may not be available to the knower or anybody else.Furthermore, Wittgenstein says, "[I]f he knows something, then the question 'how does he know?' must be capable of being answered" (OC: § 550).This remark supports my reading rather than the access internalist reading, for here Wittgenstein talks about the question -which must be capable of being answered, not the knowing subject's readiness to answer (or not) such a question.He makes a logical point about sense, he is not advancing a thesis about epistemic requirements for knowing subjects.The latter quote can be rendered as follows: The question "how does he know?"must make sense, and if it does not, it is "not capable of being answered".But if so, the expression "I know" is also inappropriate, as in the case of Moore.
If this interpretation is right, Wittgenstein's critique of Moore cannot be dismissed as based on an access internalist epistemic requirement -that one must be able to say how one knows and justify one's knowledge claim.5It seems, therefore, that Wittgenstein hits his target much more forcefully with the second objection than with the first.

Wittgenstein on Moore: Critique or Elaboration?
One thing that has puzzled commentators is how Moore's proof seems to beg the question in that he does not provide justification for the premise (expressed by holding up his hands and uttering "Here is one hand and here is another") -which is exactly what the sceptic is questioning.However, Moore is perfectly well aware that his proof begs the question against the sceptic.He has neither attempted to prove, nor proved what these philosophers are requesting.Nor does he think such a proof can be given (PEW: 148-149).Hence, we might better ask if Moore is trying to convey something more than just the blatantly question-begging answer to sceptical doubt.Wittgenstein does so too, I believe.In the remainder of this section I will argue that Wittgenstein, rightly, credits Moore with both philosophical insight and error.
How should we understand Moore's contentment with not providing a proof, and not believing any proof feasible, of the proposition "Here is one hand and here is another"?I would suggest that rather than being just a bizarre case of dogmatism, this is where Moore connects with Wittgenstein, that is, he does not think the idea of evidence applies to his propositions.To begin with, proving the premise would, presumably, involve some appeal to evidence.But there is neither any evidence for, nor evidence against this and the other (Moorean) propositions.That is why Moore calls them "obvious" (DCS: 33) and says that it would be absurd to doubt them (DCS: 45; PEW: 146).Next, Moore says that in order to prove such a thing as the premise he would also have to prove that he is not now dreaming (PEW: 149).This is reminiscent of Wittgenstein's point that nothing is more certain than these propositions, for if we doubted them, we should have no reason to trust any sources of evidence (and so we would have no reason to think that we are not, e.g., dreaming).When Moore talks about our not knowing what the evidence for these propositions is, that is, how we came to know them (DCS: 44), this could therefore be read as him expressing, in misleading terms, the observation that the notion of evidence is unintelligible.
On this reading, we can properly understand why Moore does not think it is possible to provide proof of the premise (PEW: 149)viz.appealing to evidence -since there is here no such thing.He is misled to put the peculiarity of these propositions in terms of ignorance of evidence and incapability of proof because he has already (mis)interpreted the situation epistemically, instead of logically, and brought in the concept of 'knowing'.Thus Moore, although formulating himself in a confused way, can actually agree with Wittgenstein on the point that nothing counts as evidence.Accordingly, it is questionable whether commentators (Coliva 2010:  208; Moyal-Sharrock 2007: 122; Pritchard 2011: 525; 2012: 257) are right in charging Moore with the incorrect view that the notion of evidence applies to the propositions in question.
Wittgenstein, on the other hand, does not directly ascribe the misconceived view to Moore, but points out how Moore is committed to it if he uses "I know".In fact, he never mentions Moore when talking about (inapplicability of) evidence, but frequently talks about "Moore's mistake" as being to say that he knows (OC:  § 178, 403,  521).And since the view of applicability of evidence is incorrect, Moore must be wrong in using the words "I know".Nonetheless, there is no good reason to think that Moore does not share Wittgenstein's thought that nothing is more certain than these propositions, and that they are not believed on the basis of evidence.So Moore is right in his observations about these propositions.That is the philosophical insight.
The only thing, on this view, that Moore is wrong about is his interpretation of these observations -his drawing epistemic conclusions, rather than logical.For instead of taking the quite correct observations about his propositions as objections to claiming to know, he takes them as indicating the peculiar security of our knowledge claims, how they are "definitely true", "obvious truisms" which we know with certainty (DCS: 33,45).He takes certainty and inapplicability of evidence, rightly, as objections to queries for justification and accusations of ignorance (that is, sceptical doubts), but wrongly, infers that this affirms knowledge."I am a human being, but I do not know it" is not false (in that I do know it) -it simply makes no sense as an epistemic claim.This logical exclusion of ignorance is not the same as the epistemic presence of knowledge.Hence, Moore's mistake lies not in his certainty, which Wittgenstein agrees with (OC: § 446, 511), nor in his rejection of sceptical queries about our justification for Moorean propositions and accusations of ignorance.The mistake is only his misrepresenting these features as a case of knowledge.That is the philosophical error.
Another way in which to express this point is by invoking a distinction between 'logical', and 'epistemic' sense of "I know" (see Schönbaumsfeld 2016a: 173).The 'logical' sense of "I know" indicates that expression of doubt or ignorance is senseless, as in the grammatical proposition "Only you can know if you had that intention" from Philosophical Investigations § 247, explaining the meaning of the word 'intention'.The 'epistemic' sense of "I know" indicates a straightforward knowledge claim, where the invocation of evidence is appropriate.With this distinction at hand, we could say that Moore exploits a use of "I know", which is really a logical or grammatical point about what does (not) make sense, but he conflates this 'logical' use with an 'epistemic' use.As Wittgenstein puts it: "'I know' is here a logical insight.Only realism can't be proved by means of it" (OC: § 59).
This understanding of the relationship between Wittgenstein and Moore differs from the common view that Wittgenstein thought that Moore was totally misconceived about the status and implications of the propositions in question. 6He credits Moore with both philosophical error (misuse of "I know") and insight (inapplicability of evidence, presence of certainty, and rejection of ignorance and doubt).This diagnosis fits well with Moore's remarks that it would be absurd to doubt the truth of these propositions and say he only believed them and that they are not quite certain (DCS: 45; PEW: 146-147).According to Wittgenstein's analysis, it is quite correct that to doubt them is absurd, but Moore's conclusion about certain knowledge does not follow.Nevertheless, there are for the most part correct observations that underlie Moore's claims.The misunderstanding arises from Moore's misinterpretation of these observations in epistemic terms, and taking them to imply certain knowledge.Wittgenstein's concern with Moore is thus no less an elaboration than a critique.

Conclusion
Wittgenstein's writings are notoriously difficult to understand.The tentative and unpolished (cf.OC: Preface, p. vi) remarks that make up On Certainty are not an exception.Confusion regarding his critique of Moore is, I have argued, partly his own fault, partly due to our inability to comprehend the originality of his thinking.What is his own fault is his misrepresentation of Moore as speaking without any specific context.This has misled some commentators such as Travis and Conant to overstate the importance of Wittgenstein's first, and, because it rests on misrepresentation, unsatisfactory objection against Moore.Furthermore, our proneness to place Wittgenstein in fixed philosophical categories has made commentators unable to realize the originality of his clarification of the connection between knowledge and justification.For Wittgenstein is not talking about accessibility to supporting evidence as an epistemic requirement.He is talking about a logical connection, in that questions of validation and appeal to evidence must make sense if one claims to know.Since it does not make sense for Moore, this constitutes a much more serious worry than the accusation of talking without any specific context.Wittgenstein's objection is also interesting in its own right, for it shows how we can maintain a connection between knowledge and justification without being vulnerable to the objections directed at access internalism.This highly original and promising aspect of his epistemological reflections is misrepresented, misunderstood, or neglected by most commentators.
Of course, Wittgenstein's target in On Certainty is not only Moore.But a misunderstanding of the criticism of Moore -which provides the background of the book -will affect the understanding of other insights and arguments.For unlike Moore, Wittgenstein is not simply ignoring the tension between the fact that we cannot doubt, and the fact that we cannot justify Moorean propositions.Instead, his approach is to resolve the apparent tension, in showing how the absence of justification neither poses a threat to our certainties, nor renders them dogmatic.This obviously has consequences for our understanding of scepticism.Contrary to the sceptic, Wittgenstein does not see our inability to ground Moorean propositions as an epistemic failure that must be remedied if our acceptance of them is to be undogmatic.Rather, our "inability" is a logical, not an epistemic feature of these propositions, which both Moore and the sceptic fail to see properly. 7Wittgenstein's critique of Moore is not only important because it teaches us how to avoid the latter's mistakes, but because it is crucial in order to understand Wittgenstein's thinking about scepticism and other epistemological issues. 8