In this chapter I'm going to trace out an argument about what it means to claim some forms of knowledge for literature and not others, and about what happens when we try to situate literature between the worlds of work and play. The journey is both long and accelerated, and it may help if I say something about its chief stages. I'm going to suggest, with the help of a number of critical and philosophical friends, most of them personally unknown to me:

• that literature may know a lot about us that we would rather it did not know;

• that it often, perhaps mostly, lets us off the hook by telling us that what it has on us is not real knowledge;

• that literature likes to say it is only playing at knowing things but that we shall do well not to be entirely taken in by this assertion, because although it often does play at knowing things, sometimes it's not playing;

• and finally that if literature is not always playing, we may need to take the thought one stage further and recognize situations where the play has become reality without our knowing it.

‘Literature doesn't say it knows something,’ Roland Barthes once announced, ‘but that it knows of something.’ I don't think the kinds of knowledge Barthes is evoking – the knowledge that literature says it doesn't have and the knowledge it says it does – are mysterious or unusual.