It has struck me that there is a diagnosis of what is wrong with ‘p, but I believe that not p' which is at once perfectly simple and very strong. It brings out in a way which no other diagnosis which I have seen achieves, how very like a contradiction that form of assertion is. I make just two assumptions in my account, and at another time I would defend each of them to the hilt. The first is that statements of the form ‘I believe that p’ may be judged for truth without any knowledge of the truth of ‘p’. That we may judge ‘He believes that p’ for truth without any knowledge whether p, is generally accepted. But if, for instance, Urmson’s account of ‘I believe…’ sentences were correct, my assumption would be false.