G. E. Moore’s Proof of an External World (1993) is commonly viewed as an interesting but ultimately unconvincing response to radical scepticism (Baldwin 2010). This essay will argue that this common evaluation of Moore has been based on a misinterpretation of his argument and its aims. I will begin by presenting Moore’s ‘proof’, before advancing a less hostile interpretation, namely that Moore is not trying to answer the radical sceptic’s challenge, but rather to expose the ambiguities contained within the challenge itself. I will then deal with some of the common objections to Moore’s response, and expose their irrelevance in light of this interpretation. This will lead me to the conclusion that Moore’s ‘proof’ is a satisfactory response to radical scepticism, even if it is not a refutation.

Moore’s ‘proof’

Moore’s ‘proof’ can be represented thus:

(M1): If I know that I have two hands, there is an external world.

(M2): I know that I have two hands.

(MC): There is an external world.

At first glance, Moore’s argument, though valid, appears strikingly naïve. As Duncan Pritchard points out, his response is pre-theoretical and question-begging1, insofar as Moore helps himself to the denial of the radical sceptic’s motivating claim as an unjustified premise (Pritchard 2012: 114). However, it is also of note that Moore’s conclusion is not in conflict with the first premise of the radical sceptic’s argument i.e. he does not claim to know that there is an external world. This is because Moore claims to prove the existence of an external world, and not his knowledge of it – and it is only the latter that the ontological version of radical scepticism denies (Baldwin 2010). Moore is also personally aware that his ‘proof’ would not satisfy his opponents (Moore 1939: 168). Why, then, does Moore refuse to address directly the challenge of radical scepticism?

Re-evaluating Moore’s ‘proof’

The most plausible explanation, I believe, is that Moore was not really trying to prove the existence of an external world at all. His ‘proof’, it would seem, is not an answer to the sceptic’s question, but instead a tool to expose its futility i.e. the misguided attempt to conflate lack of knowledge of a proposition with the plausibility of its negation. Moore achieves this by holding a mirror to the radical sceptic’s argument. Where the sceptic employs modus tollens2 to question his knowledge, Moore employs modus ponens3 to affirm it. The ‘naivety’ of his ‘proof’ is its strength; it unmasks the absurdity of applying knowledge tests to empirical claims which we naturally affirm without justification. The question of why this method is unsatisfactory is left unanswered, and as a result the Moorean response to radical scepticism is not a rebuttal. But to dismiss his argument because it does not do this would be to judge it by standards irrelevant to its aim. Moore’s whole point is to change the terms of the debate. If he achieves this, then his ‘proof’ should be regarded as a satisfactory response to radical scepticism.

The question, therefore, is whether the Moorean response to radical scepticism changes the terms of the sceptical debate. The common objections to Moore’s ‘proof’ would indicate that he does not succeed in doing this. I will now demonstrate that they are no longer pertinent in light of this interpretation of Moore’s ‘proof’.

Common objections to Moore’s ‘proof’

The first objection I shall address is that Moore’s ‘proof’ commits a dialectical impropriety by begging the question of radical scepticism (Pritchard 2012: 114). While rejecting the accusation of dialectical impropriety, I do not deny that Moore’s ‘proof’ is question-begging. His argument assumes as a premise the same claim his opponent argues against, and he does not, as Pritchard believes a Moorean should, provide any ‘diagnostic theory […] to explain away the intuitive appeal of radical scepticism’ (Pritchard 2012: 114). According to my interpretation, however, Moore does not set out to refute radical scepticism, and a diagnostic theory is beyond the scope of his argument. Furthermore, while it is true that question-begging is a fallacious form of reasoning, Moore’s ‘proof’ is a special case. As I have argued, the success of the argument, in terms of answering the sceptical challenge, is irrelevant to Moore’s aim. It is intended to be unconvincing, in order to reveal that the sceptic’s argument is also unconvincing. In other words, Moore begs the question to make it clear that the question is unanswerable. Moreover, as Moore’s argument is only dialectic in appearance, and not in nature, it cannot be said to have committed a dialectic impropriety.

The second objection that Pritchard discusses is that Moore’s argument leads us to an ‘epistemological impasse’ (Pritchard 2012: 115). As both arguments are valid and both follow from intuitive premises, it is almost impossible to decide between the two. Subsequently, even if we were to accept the ‘Moorean’ argument, there is no reason to prefer it to radical scepticism. According to Pritchard, this prevents it from being ‘intellectually satisfying’ (Pritchard 2012: 115). Again, it is only the latter claim I reject. This, I believe, is Moore ‘holding the mirror’ to radical scepticism. To emphasise their similarities, I shall represent the modus tollens argument thus:

(RS1): If I know that I have two hands, then I know that there is an external world.

(RS2): I don’t know that there is an external world.

(RSC): I don’t know that I have two hands.

As Pritchard points out, both arguments draw from intuitive (although contradictory) first premises. This similarity is intentional. By using the claim “I know I have two hands” as a premise, Moore effectively pits common sense reasoning against sceptical argumentation. The fact that the two opposing arguments appear equally (un)persuasive reveals the futility of using intuitive claims as a basis for deductive proofs. Contrary to what Pritchard says, the impasse is the reason that Moore’s ‘proof’ is intellectually satisfying – it lays bare the weakness of the method, as applied to either a realist proof or a sceptical hypothesis.

The final rebuttal hinges around Moore’s apparently contentless use of the word ‘know’ (Pritchard 2012: 115). Wittgenstein believes Moore’s ‘proof’ fails not because Moore’s claim is false, but rather because it is senseless (Kenny 2006: 161). Although Wittgenstein agrees that it is not possible to doubt the claim ‘I have a hand’, he argues that this is not the same as knowledge. According to Wittgenstein, the claim ‘I know’ (at least in the anti-sceptical sense that Moore uses4) only makes sense when one can also say ‘I do not know’ or ‘I doubt’. This is because the expression ‘I know’ implies the plausibility of the question ‘how do you know?’ (Kenny 2006: 168-9). However, as the proposition ‘I have two hands’ cannot be given any surer grounds, this question has no answer, and thus Moore’s claim to knowledge is senseless. He further argues that Moore has wrongly interpreted knowledge as an infallible mental state that he can report on, similar to belief. Wittgenstein points out that knowledge relies on not just a mental state, but also the truth of a proposition external to the mental state. For example, while I could say ‘she believes it’s true, but really it’s a lie’, I could not say ‘she knows it’s true, but in fact it’s not’. (Kenny 2006: 169-170). However, this does not mean that there is necessarily any difference between the mental state of knowing and believing.

In response to Wittgenstein’s accusation of senselessness, I believe that the two philosophers are arguing at cross-purposes. Moore may not have been aware that his premise was senseless, but he was aware that merely affirming his knowledge of everyday propositions would not counter the sceptic’s challenge. That Wittgenstein notices this, and is able to uncover the reason behind it, is entirely in keeping with the intended effect of Moore’s ‘proof’. However, the accusation that Moore views knowledge as infallible seems unfair. Moore does believe that knowledge requires justification (and is thus fallible) – in the case of his first premise, this justification consists of holding up his hands and gesturing (Moore 1939: 166). The strength of this justification is dubious– but it is also of no great import to my argument; the ‘proof’ is question-begging either way. Nonetheless, Wittgenstein’s criticism remains pertinent to the radical sceptic, who presents his affirmation of doubt as proof that doubt is possible.


To conclude, Moore’s argument is deceptive in both purpose and structure, insofar as its avowed aim (to prove the existence of an external world), is not its author’s intention, and its form (that of a deductive proof) is not connected to its substance. As I have argued, Moore’s ‘proof’ is not a rebuttal of radical scepticism, but rather a ‘mirror’ used to expose the fallibility of its method. Moreover, I have demonstrated that the key objections fail to address this aspect of his ‘proof’. This failure to strike their target would indicate that Moore succeeds in changing the terms of the sceptical debate, and therefore his ‘proof’ does indeed provide a satisfactory response to radical scepticism.


  1. Baldwin, T. 2010. George Edward Moore The Stanford Encyclopedia of Philosophy. Zalta, E. (ed.). Available at [Last accessed: May 2020].
  2. Kenny, A. 2006. Wittgenstein. Oxford: Blackwell Publishing Ltd.
  3. Moore, G.E. 1993. Proof of an External World. In: G. E. Moore’s Selected Writings. Baldwin, T. (ed.). Abingdon: Routledge. pp.147-170.
  4. Pritchard, D. 2012. Mooreanism. In: Epistemological Disjunctivism. Oxford: Oxford University Press. pp.113-115.


Question begging is a fallacious (although valid) form of argument which assumes the truth of its own conclusion as an unjustified premise. In Moore’s case, his conclusion that ‘there is an external world’ is contained within the premise ‘I know that I have two hands’.

A valid form of argument: (1) If P is true then Q is true. (2) Q is false. (3) Therefore, P is also false.

Another valid form of argument, and the ‘reverse’ of Modus Tollens: (1) If P is true then Q is true. (2) P is true. (3) Therefore, Q is also true.

Wittgenstein allows for some special exceptions to this rule where ‘I know that p’ is used in a self-ascriptive sense. However, as he makes clear, none of these could be used to respond to radical scepticism (Kenny 2006: 166-7).