EPISTEMOLOGY deceiving me totally, then he is deceiving me about the meaning of the word ‘deceive’ So ‘The evil genius is deceiving me totally’ does not express the total doubt that it was intended to Even within the language-game, there must be some propositions that cannot be doubted ‘Our doubts depend on the fact that some propositions are exempt from doubt, are as it were the hinges on which those turn’ (OC 341) But if there are propositions about which we cannot doubt, are these also propositions about which we cannot be mistaken? Wittgenstein distinguished between mistake and other forms of false belief If someone were to imagine that he had just been living for a long time somewhere other than where he had in fact been living, this would not be a mistake, but a mental disturbance; it was something one would try to cure him of, not to reason him out of The difference between madness and mistake is that whereas mistake involves false judgement, in madness no real judgement is made at all, true or false So too with dreaming: the argument ‘I may be dreaming’ is senseless, because if I am dreaming this remark is being dreamt as well, and indeed it is also being dreamt that these words have any meaning (OC 383) Wittgenstein’s purpose in On Certainty is not just to establish the reality of the external world against Cartesian scepticism His concern, as he acknowledged, was much closer to that of Newman in The Grammar of Assent: he wanted to inquire how it was possible to have unshakeable certainty that is not based on evidence The existence of external objects was certain, but it was not something that could be proved, or that was an object of knowledge Its location in our world-picture (Weltbild) was far deeper than that In the last months of his life Wittgenstein sought to clarify the status of a set of propositions that have a special position in the structure of our epistemology, propositions which, as he put it, ‘stand fast’ for us (OC 116) Propositions such as ‘Mont Blanc has existed for a long time’ and ‘One cannot fly to the moon by flapping one’s arms’ look like empirical propositions But they are ‘empirical’ propositions in a special way: they are not the results of inquiry, but the foundations of research; they are fossilized empirical propositions that form channels for the ordinary, fluid propositions They are propositions that make up our world-picture, and a worldpicture is not learnt by experience; it is the inherited background against 166