neglect so fine a means of conveying shades of significance. But on the whole the structure of thought in language does not offer such scope for exceptional, immediate expressiveness as the diction which builds up that structure; we naturally look for poetic quality in words and idioms rather than in the planning of sentences, the quality of which has no great range of variation. The chief importance of syntax in poetry is of another order. It is not remarkably concerned with the actual symbolising of imaginative experience which we call poetry; it is rather concerned with the organisation of this business into poems. The intellectual shape of poetical matter-the coherent isolation of it into self-sufficient organism-to effect that is the affair of syntax in poetry. Without it, we could never know the real significance of poetic detail; but the detail must first be there in order to give syntax a function at all. Given poetic matter, conveyed in the vigours and delicacies of diction, it is the continuous organisation of syntax which moulds this into that intellectual unity, the achievement of which we call form.

I

V

DICTION: THE SOUND OF WORDS

HAVE insisted on the importance to poetic theory

of the spoken life of words as regards their semantic qualities. Now, words, of course, are sounds which symbolise ideas: we listen to them for the purpose of understanding them. But it is well known that words, especially combinations of words, very often produce remarkable effects by their sounds over and above the necessary and open conveying of meaning; the mere existence of such terms of linguistic criticism as euphony and cacophony witnesses to this. The art of poetry endeavours to make language not so much the vehicle of thought as the equivalent of experience itself; and it can only do so by deliberately using every quality which language is capable of exhibiting—and by using, as far as possible, all the qualities appropriate to its purpose simultaneously. Accordingly, we find that the phonetic aspect of language is in poetry very conspicuous; so much so, that the meaning of the words cannot impress itself on our minds without the sound of the words distinctly impressing itself also. It is not merely that we become aware, as we seldom do elsewhere, that it is sound to which we are attributing meaning; nor is it that the effect of sound over

and above its regular symbolism of meaning is simply an effect of euphony. The whole sound of the words and the whole of its effect-whether as the necessary sign of grammatical meaning, or whatever may be redundant to this (e.g., rhythm)—must contribute so decisively to the art of poetry, that its audible technique may well be discussed separately; not indeed because verbal sound operates on its own account in poetry, but because it is an addition to the common use of language so notable and so characteristic that its contribution cannot but be distinguished.

When, therefore, we consider the phonetic aspect of poetry, it is obvious that we assume the spoken existence of language; so obvious that it might hardly seem worth mentioning. But we are sometimes apt to forget, even in this connexion, that language has, and has had for many centuries, a double life, in poetry as elsewhere. Language lives as the spoken word, and it lives also as the written (or printed) word. The spoken word cannot be anything else than sound accepted as the symbol of an idea; and the written word was originally the symbol of this spoken sound: that is to say, the symbol of the symbol of an idea. But the written word can be, and has long since in part become, something else than the symbol of a symbol; it can be a symbol in its own right. For the human mind will always short-circuit a process when it can. As soon as the habit of reading to oneself was established, the second-hand symbolism of the written word was shortcircuited; and the written word became itself the symbol

of the idea, without having to pass through the symbolism of sound. Printing has fixed this short-circuit in our civilised mentality so deeply that many of us are scarcely aware of it.

Language, as communicable symbolism of ideas, has two modes of existence: it exists as audible signs and it exists as visible signs. Generally, the only practical relation between the two kinds of signs is that they both refer to the same thing: the English we read has almost given up pretending to have anything to do with the English we hear. As a rule, when we are reading to ourselves, the printed word refers immediately to its idea; the sound of the word very likely comes in, but it is not required, and may be due to association with the idea as much as to taking the letters as phonetic signs. The sound of the word at any rate only comes in as a faint unnecessary accompaniment, to which we only attend as a sort of mild corroboration. But, though it may be proper enough to read our newspapers in that style, it will not do for poetry. Poetry consists absolutely of the word spoken and heard: the printed word must always be frankly the symbol of articulate sound. We must hear what the poet has to say; if we are reading to ourselves, we must hear it mentally. Otherwise we shall miss half his technique; and that means, we shall miss half of what he is trying to express.

But there is still something to be said on this matter; the visible word is not to be dismissed in poetry altogether, in favour of the audible word. There is no

doubt, in fact, that the existence of language as printed words has had a profound influence on the art of poetry, though it would take too long to investigate this here exactly. We read poetry to ourselves more often than hear it read aloud; and poets, consciously or not, have taken advantage of this. Poetry will always take advantage of anything that will increase or refine its expressive power. I said, that the printed word in poetry must always be taken as a symbol of an articulate sound. I did not say, it could only be taken so. When we read poetry to ourselves, it is not, I think, the usual thing to refer the word to the sound and thence, through the sound, to the idea; I think rather we refer the printed word immediately to the idea and simultaneously to the sound as well. And this is important; for eye-language is a much subtler and nimbler affair than earlanguage. We can get, in printed language, in the appeal through the eye, a more instant and more certain apprehension of fine associations of ideas, of delicate shades of significance, than you can ever get through the ear. One of the chief differences between such an art as Homer's and such an art as Dante's or Milton's is that Homer never thinks of any appeal but through the ear; whereas Dante and Milton both know their verses will meet with eyes as well as ears. Their art is certainly not greater than Homer's, but it has finer modulations of significance. The thing is, that Dante and Milton, like every other printed or written poet, take advantage of the eye-appeal without losing the ear-appeal. However, we are not concerned with