It is unwise to equate scientific activity with what we call reason, poetic activity with what we call imagination. Without the imaginative leap from facts to generalisation, no theoretic discovery in science is made. The poet, on the other hand, must not imagine but reason--that is to say, he must exercise a great deal of consciously directed thought in the selection and rejection of his data: there is a technical logic, a poetic reasoning in his choice of the words, rhythms and images by which a poem's coherence is achieved.