Next to the questions of authorship and of origin in point of difficulty come two others—"Which are the older: the prose or the verse romances?" and, "Was there a Latin original of the Graal story?"
Prose or verse first?
With regard to the first, it has long been laid down as a general axiom, and it is no doubt as a rule true, that prose is always later than verse, and that in mediæval times especially the order is almost invariable. Verse; unrhymed and half-disrhythmed prose; prose pure and simple: that is what we find. For many reasons, however, drawn partly from the presumed age of the MSS. and partly from internal evidence, the earlier scholars who considered the Arthurian matter, especially M. Paulin Paris, came to the conclusion that here the prose romances were, if not universally, yet for the most part, the earlier. And this, though it is denied by M. Paris's equally learned son, still seems the more probable opinion. For, in the first place, by this time prose, though not in a very advanced condition, was advanced enough not to make it absolutely necessary for it to lag behind verse, as had been the case with the chansons de geste. And in the second place, while the prose romances are far more comprehensive than the verse, the age of the former seems to be beyond question such that there could be no need, time, or likelihood for the reduction to a general prose summary of separate verse originals, while the separate verse episodes are very easily intelligible as developed from parts of the prose original.[49]
A Latin Graal-book.
With regard to the Latin Graal-book, the testimony of the romances themselves is formal enough as to its existence. But no trace of it has been found, and its loss, if it existed, is contrary to all probability. For ex hypothesi (and if we take one part of the statement we must take the rest) it was not a recent composition, but a document, whether of miraculous origin or not, of considerable age. Why it should only at this time have come to light, why it should have immediately perished, and why none of the persons who took interest enough in it to turn it into the vernacular should have transmitted his copy to posterity, are questions difficult, or rather impossible, to answer. But here, again, the wise critic will not peremptorily deny. He will say that there may be a Latin Graal-book, and that when that book is produced, and stands the test of examination, he will believe in it; but that until it appears he will be contented with the French originals of the end of the twelfth century. Of the characteristic and probable origins of the Graal story itself, as of those of the larger Legend of which it forms a part, it will be time enough to speak when we have first given an account of the general history as it took shape, probably before the twelfth century had closed, certainly very soon after the thirteenth had opened. For the whole Legend—even excluding the numerous ramifications into independent or semi-independent romans d'aventures—is not found in any single book or compilation. The most extensive, and by far the best, that of our own Malory, is very late, extremely though far from unwisely eclectic, and adjusted to the presumed demands of readers, and to the certain existence in the writer of a fine literary sense of fitness. It would be trespassing on the rights of a future contributor to say much directly of Malory; but it must be said here that in what he omits, as well as in his treatment of what he inserts, he shows nothing short of genius. Those who call him a mere, or even a bad, compiler, either have not duly considered the matter or speak unhappily.
But before we go further it may be well also to say a word on the Welsh stories, which, though now admitted to be in their present form later than the Romances, are still regarded as possible originals by some.
The Mabinogion.
It would hardly be rash to rest the question of the Celtic origin, in any but the most remote and partial sense, of the Arthurian Romances on the Mabinogion[50] alone. The posteriority of these as we have them need not be too much dwelt upon. We need not even lay great stress on what I believe to be a fact not likely to be disputed by good critics, that the reading of the French and the Welsh-English versions one after the other, no matter in what order they be taken, will leave something more than an impression that the French is the direct original of the Welsh, and that the Welsh, in anything at all like its present form, could not by any possibility be the original of the French. The test to which I refer is this. Let any one read, with as open a mind as he can procure, the three Welsh-French or French-Welsh romances of Yvain-Owain, Erec-Geraint, and Percivale-Peredur, and then turn to those that are certainly and purely Celtic, Kilhwch and Olwen, the Dream of Rhiabwy (both of these Arthurian after a fashion, though quite apart from our Arthurian Legend), and the fourfold Mabinogi, which tells the adventures of Rhiannon and those of Math ap Matholwy. I cannot conceive this being done by any one without his feeling that he has passed from one world into another entirely different,—that the two classes of story simply cannot by any possibility be, in any more than the remotest suggestion, the work of the same people, or have been produced under the same literary covenant.
The Legend itself.
Let us now turn to the Legend itself. The story which ends in Avalon begins in Jerusalem. For though the Graal-legends are undoubtedly later additions to whatever may have been the original Arthurian saga—seeing that we find nothing of them in the early Welsh traditions, nothing in Nennius, nothing in Geoffrey, nothing even in Wace or Layamon—yet such is the skill with which the unknown or uncertain authors have worked them into the legend that the whole makes one indivisible romance. Yet (as the untaught genius of Malory instinctively perceived) when the Graal-story on the one hand, and the loves of Lancelot and Guinevere with which it is connected on the other, came in, they made comparatively otiose and uninteresting the wars with Saxons and Romans, which in the earlier Legend had occupied almost the whole room. And accordingly these wars, which still hold a very large part of the field in the Merlin, drop out to some extent later. The whole cycle consists practically of five parts, each of which in almost all cases exists in divers forms, and more than one of which overlaps and is overlapped by one or more of the others. These five are Merlin, the Saint-Graal, Lancelot, the Quest of the Saint-Graal, and the Death of Arthur. Each of the first two pairs intertwines with the other: the last, Mort Artus, completes them all, and thus its title was not improperly used in later times to designate the whole Legend.