Studying the Successful

I shall almost always take the poems that I admire, and write with pleasure about their merits; you might say that, from the scientific point of view, this is a self-indulgence, and that as much is to be learnt from saying why bad poems are bad. This would be true if the field was of a known size; if you knew the ways in which a poem might be good, there would be a chance of seeing why it had failed. But, in fact, you must rely on each particular poem to show you the way in which it is trying to be good; if it fails you cannot know its object; and it would be trivial to explain why it had failed at something it had not been trying to achieve. Of course, it may succeed in something that you understand and hate, and you may then explain your hatred; but all you can explain about the poem is its success. And even then, you can only have understood the poem by a stirring of the imagination, by something like an enjoyment of it from which you afterwards revolt in your own mind. It is more self-centred, therefore, and so less reliable, to write about the poems you have thought bad than about the poems you have thought good.

Empson, (introducing) Seven Types of Ambiguity