In the nouveau roman of Robbe-Grillet there is an attempt at a more or less Copernican change in the relation between the paradigm and the text. In Camus the counter-pointing is less doctrinaire; in Dostoevsky there is no evidence of any theoretical stand at all, simply rich originality within or without, as it chances, normal expectations. All these are novels which most of us would agree (and it is by a consensus of this kind only that these matters, quite rightly, are determined) to be at least very good. They represent in varying degrees that falsification of simple expectations as to the structure of a future which constitutes peripeteia. We cannot, of course, be denied an end; it is one of the great charms of books that they have to end. But unless we are extremely naive, as some apocalyptic sects still are, we do not ask that they progress towards that end precisely as we have been given to believe. In fact we should expect only the most trivial work to conform to pre-existent types. It is essential to the drift of all these talks that what I call the scepticism of the clerisy operates in the person of the reader as a demand for constantly changing, constantly more subtle, relationships between a fiction and the paradigms, and that this expectation enables a writer much inventive scope as he works to meet and transcend it. The presence of such paradigms in fictions may be necessary-that is a point I shall be discussing later--but if the fictions satisfy the clerisy, the paradigms will be to a varying but always great extent attenuated or obscured. The pressure of reality on us is always varying, as Stevens might have said: the fictions must change, or if they are fixed, the interpretations must change. Since we continue to 'prescribe laws to nature'--Kant's phrase, and we do--we shall continue to have a relation with the paradigms, but we shall change them to make them go on working. If we cannot break free of them, we must make sense of them.