So relying upon such a connection by virtue of seeing it as relevant alone construes language and world as something which is given and self interpreting; it is some way or the other, and if this is not seen in the case being analysed, that is a problem of someone's understanding rather than a problem of the methodology. In broad terms, the applicability of connections like Wittgenstein's knowledge-doubt one rests upon a privileged domain in which terms are imbued their meaning, and a connection to this domain is only ensured if you follow the pattern of argument in the knowledge-doubt (or like) analysis. It stops being a methodology using language, and reifies a particular interpretation of language as language, without the mechanisms of contextualisation that it espouses. Linguistic analysis comes (or can come) to police sense rather than clarify it. — fdrake
I don't think this is quite right, but I think this partly down to how to phrased things with the dichotomy language/world. I need to modify what I said above: it is in fact the case that language and world can 'come apart', but the key thing is to
recognise instances when they do. 'Linguistic analysis' ('LA'), as I understand it, is the attempt to track when language and world depart from one another, despite the
impression that they have not (what Witty calls 'being held captive by a picture' or somesuch). There's a passage from Cavell that I really like that brings out the critical import of LA here, where he uses a really interesting turn of phrase, on making words 'nothing but their meaning':
"Wittgenstein's notion of "speaking outside language games"... suggests that what happens to the philosopher's concepts is that they are deprived of their ordinary criteria of employment (which does not mean that his words are deprived of meaning - one could say that such words have nothing
but their meanings) and, collecting no new ones, leave his concepts without relation to the world (which does not mean that what he says is false), or in terms I used earlier, remove them from their position among our system of concepts".
I like Cavell's way of putting things because he does not say that such uses of words are meaning-less
per se, but that they are nothing 'but' their meanings: that their significance does not reach, as it were, where we would want it to reach. The words have meaning, but this meaning does not have the significance one takes it to have. I think the Wittgensteinian treatment of knowledge and doubt is exemplary in this regard: against those who ask: 'but how do you really
know the Thing with Certainty?', the Wittgensteinian counter-question is simply: 'do you understand what you're asking? Are you aware of how singular your question is, and the equivocations one risks in construing your question as though simply one more in a long line of questions about knowledge?'.
To be 'mislead by grammar' is here to think that this use of language still 'has its position among our system of concepts': to not recognize that world and language have come apart. That all said, against Wittgenstein, I'm all too happy to maintain that philosophers have long known that this is exactly what what happens in their discourse, and that Witty was simply making explicit what every competent philosopher has known implicitly since time immemorial (Wittgenstein projected, as it were, his own naivety onto the philosophers whom he never read).