The reliability of our shared theories does not depend upon our living in a shared world. Instead, our conviction that we live in a shared world is a consequence of the reliability of our shared theories. It is because we can make ourselves intelligible to one another that we believe we have shared experiences. It is not because we have shared experiences that we can make ourselves intelligible to one another.
It is only within our linguistic interactions that any questions at all can arise - including questions about 'sameness' and 'experience'. These tokens only have meanings within the 'game' - meanings which cannot be discovered except through the same experiments with intelligibility that form the boundaries for all our reasoning - about how the world must be, about how it actually is, or about that reasoning itself.
Before we can become engineers, before we know whether our miscalculations have led to the collapse of a bridge, we must find ourselves in relevant agreement about the identification of bridges, about what counts as a collapse, and about the way the rules of our computations work.
We can say about the world that it must be the kind of place where this game can be played, but that is not the same as saying that the game has some otherwise specifiable relationship with it. We know this because to specify such a relationship we would have to be able to specify some characteristics of its members - something that can, also, only be done within the game.
We have been confused by the vividness of our visceral empirical intuitions, and (if we are lucky) our immediate experience that these feed our capacity to render our individual worlds more predictable.
We have puzzled over the privacy of these processes but ultimately ignored its relevance. We have assumed that language is a transparent window into our internal worlds. We have presumed that our capacity to make ourselves intelligible and the reliability of our empirical intuitions were, somehow, the same thing - even though we could not have any independent access to the empirical intuitions of our interlocutors. We ignored the fact that when we tried to talk about empirical intuitions - our own and those of others - we found them to be irreducibly intentional, that we had to describe them as intuitions of something that was publicly describable. They could not be otherwise specified, either physically or formally. The features of them that burned themselves upon our sensoria could not be represented.
We did not see that this presumption of transparency was, itself, an empirical hypothesis - a hypothesis which we depended on when we projected confused internal worlds onto those we could not understand, calling them 'mad'. A hypothesis which still comforts the powerful when they project their own madness into the public realm.
If I am asked whether I am a realist, I would answer 'of course'. Not on the grounds of metaphysical conviction, but because I think that we can say things which are true about the world. The tests of truth we apply will ultimately depend upon what we can render intelligible within our conversations with one another. One of the things that is undeniably true about the world is that we have these conversations.
I am enough of a realist to reject the desperate illusions of sensory empiricist epistemology because it cannot pass that test of intelligibility.
