It is alleged that after G.E. Moore gave the lecture in which his now-famous paradox was first formulated, Wittgenstein rushed around to Moore's quarters (for the two were at Cambridge together under the tutelage of Bertrand Russell) and insisted he repeat it in full. Whether or not this is true, Moore's paradox has become cemented into the philosophical literature by Wittgenstein's interest, as Philosophical Investigations invests a page or so to its consideration.
The heart of the paradox is simple, and rests on what is termed Moore's sentences, which have the logical form 'P and I do not believe P' (or 'P and I believe not-P'). The most commonly cited example is “It is raining but I do not believe it is”. At the heart of this is claimed to be a contradiction – we can say “it is raining” and “I do not believe it is raining” but not both at the same time. Yet curiously, we can make the same claim in the third person (“it is raining but she does not believe it is”) or in the past tense (“it was raining but I did not believe it was”) without any problem at all. What is going on?
This philosophical oddity has special meaning for me. When I was pursuing my Masters degree, I took a unit on introduction to philosophy led by a computer scientist with the unfortunate name of Pratt. This was the first and only formal course of philosophy I took, as thereafter I took to studying philosophy independently. At one point during a particular lecture, I believe discussing some aspect of Bertrand Russell's views of logic, Pratt remarked that it was not possible to believe things that were false. I spoke up and replied that it was perfectly possible, and indeed I did believe some things that were false, to which Pratt replied flatly “no you don't” and continued with the lecture. I was rather exasperated by this turn of events, and did not attend any more of his lectures.
Imagine my surprise when reading Wittgenstein's Philosophical Investigations earlier this year I found that Wittgenstein supported Pratt's view! He writes: “If there were a verb meaning ‘to believe falsely’, it would not have any significant first person present indicative... “I believe . . . ., and it isn’t so” would be a contradiction.” That said, I'd like to think that had Wittgenstein been my teacher he would at least have had the decency to ask me why I professed to believe in false things, since anyone who claims to be practising philosophy but is not interested in the peculiar oddities that people express therein is a meagre sort of philosopher indeed. (As to examples of false things that I believe, I will get to this in due course).
Moore's paradox rests on the sense that there is a contradiction between asserting a fact (“it is raining”) and believing the converse (“it is not raining”), and this intuition arguably hinges on the understanding of knowledge as justified true belief which originates with Plato. The contradiction in a Moore's sentence is evident if one takes knowledge to be justified true belief – that is, something you believe constitutes knowledge if it is true, and you have a legitimate justification for its truth. But I contend that Plato's model does not reliably describe how we get to a state of knowledge, nor what that state entails. I claim it is perfectly possible for us to believe in a proposition that is false (i.e. 'not P and I believe P' or the equivalent) – and when we dismiss this possibility it is because of our beliefs about what constitute both knowledge and belief.
Philosophers considering Moore's paradox routinely take knowledge as corresponding to objective states and belief to entail a striving towards those states, that is that belief is always an attempt at knowledge. Plato sets up this possibility by saying that for something to be knowledge it must be true and the means that we determine it is true must be correct. Yet it is a virtual impossibility to confirm one's beliefs against objective reality (since we never directly experience this), and equally impossible to verify one's methods of justification (and for similar reasons). In Plato's allegory of the cave, he never seems to seriously consider that after leaving the cave one might still be ill-equipped to establish truth with any confidence.
What we consider knowledge can rarely if ever be ratified by a procedure akin to Plato's; rather, we tend to grant information the status of knowledge either by virtue of its usefulness or by the (often tentative) elimination of competing claims. Thus Newtonian mechanics, while strictly false (Einstein having supplanted them with a more accurate system) are still knowledge in the sense that they are still useful (in engineering, for instance), while “Elvis is dead” is knowledge by virtue of the existence of evidence in support of the claim, and the dearth of credible evidence to the contrary. The latter could cease to be knowledge tomorrow if a 74-year old Elvis were to come out of hiding, but until such an event it comfortably qualifies as knowledge whether or not it is objectively true.
This sheds light on why Moore's sentences feel paradoxical, but not on why I can believe false things; for this, we may need some examples. Consider, as a simple case, that I believe in the fundamental goodness of human beings while knowing that this belief is objectively false, or at least, unverifiable. Since empiricists routinely equate unverifiable with false (e.g. if I cannot verify it is raining, it is not raining), the disconnect felt in Moore's paradox creeps in, but it need not. As another example, I can go to a meeting believing that I am perfectly prepared for every eventuality when pragmatically I simply could not be – I still choose to believe in my preparedness (even though I know it is false) because it helps me act confidently in the meeting to do so. This is similar to the case of the batter who believes he is going to hit a home run for certain, whilst knowing (deep down) that it is far from certain he will do so; the false belief has utility, and this outstrips logical necessity in everyday life.
Beliefs do not necessarily constitute truth claims – this is merely an assumed foundation placing objective knowledge as the goal of belief. Beliefs, in practice, may reflect personal aesthetics, moral choices or other subjective criteria. Consider, as another kind of example, that an emotivist such as Ayer or Stevenson will treat moral principles as expressions of approval or disapproval, rather than as logical propositions. An emotivist may say “rape is wrong” while denying “rape is wrong” is genuinely true. This risks inconsistency, a critic might allege, given the usual practice of speaking of truth whereby if rape is wrong then “rape is wrong” is a true proposition. But one can hardly claim that this particular practice in logic is a necessity rather than just a habit, and as such it need not be followed. Thus the emotivist can believe “rape is wrong” while knowing that “rape is wrong” is objectively a false proposition (or at least, cannot be asserted as true).
The conclusion that this leads to is that the sense of paradox in Moore's sentences comes from our unwillingness to admit of ourselves that we believe false things – our pride, you might say, conditions us into wanting to be right at all times, and although we are content to admit we were in error some time before (the past tense case), or that other people are in error (the third person case) we are not generally prepared to admit that we are wrong now. If we are wrong now, we presume we correct our mistake immediately and instantaneously and thus we must be correct at all times in the present moment. And this despite the fact that we are frequently wrong, and sometimes we even know that we are wrong, as I know I am wrong when I choose to believe in the fundamental goodness of humanity but, frankly, do not care, because it helps me (to be happy, to act positively towards others) to believe in this false thing.
As Nietzsche observes in The Gay Science, both truth and untruth have constantly proved to be useful – it is only a metaphysical commitment to the primacy of the true, a dogmatic “will to truth”, which causes us to deny the value of false things, and he tracks this stance justifiably back to Plato. Personally, I place no restriction on myself as to whether the things I believe must be true, and thus (perhaps) free myself from Moore's paradox. I frequently find it useful to believe in things which, through pragmatic eyes, I know to be false. I see no problem with this. In fact, I find it a constant source of blessings.