I think I understand, though; I cannot claim his own authority for my supposition, the way in which Mallarmé wrote verse, and the reason why it became more and more abstruse, more and more unintelligible. Remember his principle: that to name is to destroy, to suggest is to create. Note, further, that he condemns the inclusion in verse of anything but, "for example, the horror of the forest, or the silent thunder afloat in the leaves; not the intrinsic, dense wood of the trees." He has received, then, a mental sensation: let it be the horror of the forest. This sensation begins to form in his brain, at first probably no more than a rhythm, absolutely without words. Gradually thought begins to concentrate itself (but with an extreme care, lest it should break the tension on which all depends) upon the sensation, already struggling to find its own consciousness. Delicately, stealthily, with infinitely timid precaution, words present themselves, at first in silence. Every word seems like a desecration, seems, the clearer it is, to throw back the original sensation farther and farther into the darkness. But, guided always by the rhythm, which is the executive soul (as, in Aristotle's definition, the soul is the form of the body), words come slowly, one by one, shaping the message. Imagine the poem already written down, at least composed. In its very imperfection, it is clear, it shows the links by which it has been riveted together; the whole process of its construction can be studied. Now most writers would be content; but with Mallarmé the work has only begun. In the final result there must be no sign of the making, there must be only the thing made. He works over it, word by word, changing a word here, for its colour, which is not precisely the colour required, a word there, for the break it makes in the music. A new image occurs to him, rarer, subtler, than the one he has used; the image is transferred. By the time the poem has reached, as it seems to him, a flawless unity, the steps of the progress have been only too effectually effaced; and while the poet, who has seen the thing from the beginning, still sees the relation of point to point, the reader, who comes to it only in its final stage, finds himself in a not unnatural bewilderment. Pursue this manner of writing to its ultimate development; start with an enigma, and then withdraw the key of the enigma; and you arrive, easily at the frozen impenetrability of those latest sonnets, in which the absence of all punctuation is scarcely a recognisable hindrance.