The remaining works are much better and much more important. It is true that the _De Compositione_ (as its title [Sidenote: _The_ Composition.] honestly holds forth) belongs to the lower, not the higher, division of the school-grouping of the subject—to Composition, not to Rhetoric. But proper Composition, even in the school sense, is the necessary vestibule of style; and, until attention has been paid to it, there is no hope of anything further that shall be of real use in literary criticism. And it is also not only something, but a great thing, to make an advance upon that (one had but for a sacred shame, almost said) ignorant and unintelligent contempt of words as words which we find in Aristotle himself. Dionysius indeed, as in duty bound, glances at the contempt of _lexis_ which the great Master of the Walk had made fashionable. It is true, he says, that boys are caught by the bloom of style, but it takes the experience of years to judge it rightly. And he promises a supplementary treatise _On the Choice of Words_, which we should be very glad to possess. But for the present he is busied, not with their choice, but with their arrangement after they are chosen; and he deals with this partly by positive precept, but chiefly by the use of examples, from Homer in poetry and Herodotus in prose. Dionysius was a fervent devotee of his admirable countryman, allowing his devotion, indeed, to carry him to the length of distinct injustice to that countryman’s great rival Thucydides; but it has here inspired him well enough. And Homer could not lead him wrong; though perhaps we may note here, as elsewhere with the ancients, a distinctly insufficient appreciation of the differences between poetry and prose. He begins quite at the beginning with the letters, touches on onomatopœia—that process which the great poetic languages like Greek and English admit so readily, and of which the less poetic like Latin and French are so afraid—and on the practice (of which, like a true critic, he has no fear) of reviving archaisms when desirable. Then he attacks the question how beautiful diction and composition are to be attained. Here again, and necessarily, he proceeds more by example than by precept, for indeed precept, of the _a priori_ kind, is in these matters mostly valueless. But one sentence (p. 96, Reiske) is worth quoting at length, because it puts boldly the truth which Aristotle had evaded or pooh-poohed in his excessive devotion to the philosophy of literature rather than to literature itself: “So that it is necessary that that diction should be beautiful in which there are beautiful words, and that of beautiful words beautiful syllables and letters are the cause.” Dionysius knew this, as Longinus knew it three hundred, as Dante knew it thirteen hundred, years after him: but, six hundred years after Dante, there are still persons who seem to regard the fact as somehow or other degrading.