are thus distinctly mythical in character; and though the English versions, as we now possess them, are of comparatively late date, in neither case can the Irish version be later than the eleventh century, while the internal evidence points a period anterior to the introduction of Christianity.
Let us take another instance, the story of Guinevere's abduction and rescue. Of purely mythical origin, the story was at first unlocalised, but when localised it is on insular and not on continental ground. To say, as Professor Foerster does,55 that the mention of Bath is no proof of an insular source simply shows that the writer has not grasped the real facts of the case. The mention of Bath does not cover the whole ground, it must be taken in connection with Æstiva Regis (Somerset) and Glastonbury. The latter is, if I mistake not, the real point of identification. A confusion between Glastonbury, Avalon, and the abode of the departed had taken place previous to William of Malmesbury: the exact date cannot be ascertained, but M. Ferd. Lot considers the author of the identification to have been an Irish monk writing in the tenth century. In a subsequent note M. Lot further identified Melwas=Meleagant, whom all scholars admit to be a king of the other-world, with the Irish 'king of the dead,' Tigern-Mas, of which name he considers Mael-was to be the Welsh translation.56
Now it seems to me quite obvious that the connection of the king of the other-world with the place looked upon as the special dwelling of the departed must have preceded his being considered as lord also of the surrounding lands, i.e. Tigern-Mas=Maelwas must have been connected with Glastonbury=Avalon before he was thought of in connection with Bath and Æstiva Regis. It is most probable that such a connection would take place on insular not on continental ground, and as a matter of fact the only text which connects Melwas with Glastonbury, the Vita Gildæ, is an insular text, as is that which connects Glastonbury with Avalon. Here, too, again, if M. Lot be right, we find Irish influence at work.
It is probable that we may be able to add to this list the story of Arthur's fight with the demon cat. The story is certainly told in a continental text (Merlin), and located on continental ground, but the identification of the monster with the Cath Palug of Welsh tradition and that again with the mysterious Chapalu of French romance depends on insular evidence.57
In his notice of Herr Freymond's monograph58 M. Gaston Paris suggests that the source will be found to be 'un trait sans doute fort ancien, de mythologie celtique, que Gaufrei de Monmouth n'a pas accueilli'; while M. Loth, in a note appended to this critique, remarks that the original vanquisher of the cat was certainly not Arthur but Kay. The localisation of the story in Savoy, Herr Freymond considers to have been due to the narration of pilgrims, and discusses the relations of the houses of Savoy and Flanders with our Anglo-Norman kings.
Here then we have a group of stories, possessing a distinctive (Celto-mythic) character, all of which are either better preserved, solely retained, or originally localised in these islands; i.e. the evidence of facts is here in favour of an insular rather than a continental origin. Nor do I think we shall be wrong if we ascribe a decided importance to the fact that the tales told in these islands appear to have been of a mythic rather than of a romantic character.
Granting then, that at Chrétien's time, and long previous, there was current a body of tradition, historic, mythic, romantic, dealing with the British king, how was it handed down, and in what shape did he find it? Of course it will generally be admitted that for a long time the transmission of such stories would be entirely—in Chrétien's days it would still be partially—oral.59 But in saying this we must have a clear idea of what, in the case of traditional stories, oral transmission implies. It does not mean a game of 'Russian scandal,' where the point is to see how much a story told from mouth to mouth can be made to vary from its original form in the process; professional story-tellers were, and are, more conservative than story-writers. The tales crystallise into certain formulæ of incident and expression which survive often after the real signification has been forgotten.60
In the words of a recognised authority on folk-lore: 'Among many peoples the ipsissima verba of traditional tales are insisted upon; the form, and even the details of the form, are often as much a part of the tradition as the substance of the tale.'61 Therefore when we find two stories of marked traditional and folk-lore character agreeing with each other in sequence of incident, detail, and even words, we do not necessarily conclude that the versions are connected by borrowing: they may be, but it is at least equally possible that they represent independent versions of the same oral original.
This is, of course, well understood by the folk-lore student; but unfortunately it is too often ignored by the literary critic, who is too prone to devote attention to the literary form, while he ignores the essential character of the story. Yet in solving the problem of sources it is this latter which is the determining factor.
In examining into the sources of Chrétien de Troyes it is well to remember that it is easy to exaggerate the necessity for a literary source; it is difficult to exaggerate the conservative tendencies of a professional story-teller of that date.
But besides the Arthurian legend proper, there was also current in Chrétien's time a great mass of popular folk-lore, which, certainly on the Continent, probably also on our island,62 was told, or rather sung, in the form of mythical tales or lais. These lais, in the first instance in the Breton tongue, and independent of the Arthurian cycle, were later translated into French eight-syllabic verse, and largely Arthurised—if I may use the word.
The process in vogue appears to have consisted of two stages: in the first, the king at whose court the events took place (himself generally anonymous) was identified with Arthur; in the second stage, the original hero was replaced by one of Arthur's knights. Among the specimens which have been preserved we have examples of all the stages: lais entirely independent of Arthur; lais, the scene of which is laid at Arthur's court; lais in which the hero is one of Arthur's knights; but one and all are in the same metre, that of Chrétien's poems. Of an intermediate French form we have no trace.
The lai of Tyolet, to which we have previously referred, is an excellent example of this gradual 'Arthurisation.' As we have it, the court at which the events take place is that of Arthur, the loyal friend of the hero is Gawain, but nowhere else do we meet with Tyolet as one of Arthur's knights: the inference is that we have here a lai in the first stage of assimilation. The lai consists of two parts; the latter half, the stag adventure, is found in a separate form, but here the hero is one of Arthur's most famous knights, Lancelot—the process of assimilation is complete.
The first part of the lai has many features which recall the more famous 'Perceval' Enfances. That Tyolet is anterior to the evolution of the Lancelot story we have shown above63; the probability is that it is also anterior to the great popularity of the Perceval story. When Perceval was once universally recognised as the son of the widowed lady of the forest, there would be little probability of the tale being told of a hero practically unknown to Arthurian story. His adventures taken over by more famous knights, Tyolet disappeared from the roll of heroes.
Again, among the lais we have an important group dealing with the main idea of a knight beloved by the wife of his lord, rejecting her advances, incurring her displeasure, and finally departing to fairyland with a fairy bride. Of this story we have three important variants, agreeing in their main features but differing in detail: the lais of Graalent, Guingamor, and Lanval. Of these three, the scene of the two first is laid at the court of an anonymous king; the action of the third, translated by a contemporary of Chrétien, passes at the court of Arthur. But, though the lai of Guingamor has only reached us in its earlier and independent form, Chrétien himself refers to it in an Arthurised version. He brings Guingamor to Arthur's court, and says of him,
'de