One of the central tasks in epistemology is to offer a definition of knowledge. The problem of the criterion, however, shows us that this task is in fact very difficult, if not impossible. We consider a very influential theory of knowledge known as the classical (or tripartite) account of knowledge, whereby knowledge is to be understood as justified true belief. Despite the surface plausibility of the classical account of knowledge, it is untenable as it cannot deal with a kind of counter-example known as Gettier cases. There is no easy answer to the Gettier cases; no simple way of supplementing the tripartite account of knowledge so that it can deal with these cases. Instead, a radically new way of understanding knowledge is required, one that demands greater co-operation on the part of the world than simply that the belief in question be true.