These later versions have been localised, and I think it is clear that the localisation took place on English soil, i.e. it is an insular and not a continental variant.
Now, from the very nature of the story it is clear that in its earliest forms it would not be attributed to any special locality, and therein the Lanzelet version again appears to be the elder; further, the variants must have arisen at a time when it was clearly understood that, however they might apparently differ, Valerîn's thorn-girt dwelling and Meleagant's water-circled castle meant one and the same thing, i.e. that both were recognised methods of describing the 'other-world.' In this connection it is instructive to recall the versions of Brynhild's wooing by Siegfried; her residence is universally admitted to be an 'other-world' dwelling, and we find it depicted under forms closely corresponding with the variants of the Guinevere story; e.g. Waberlohe (Volsunga saga)=Valerîn's hedged magic slumber; Castle surrounded by water (Thidrek saga)=Meleagant's stronghold; Glasberg (Folk-songs)=Glastonbury. The parallelism is significant.46
It is quite clear, I think, that such a story can be in no way ascribed to the invention of a poet living towards the end of the twelfth century, but must be of very much earlier date. Chrétien was dealing with a late variant of a primitive and very widely known theme. But could this variant, which, as seems probable, only reached him through the medium of a tale related by the Countess Marie of Champagne, have come from England, to which country the localisation of Glastonbury, Somersetshire, and Bath point? It is quite possible. We must remember who Marie de Champagne was: she was a princess of France, the daughter of King Louis VII. and Eleanor of Aquitaine, who, on her divorce from the French king, married Henry of Normandy, afterwards Henry II. of England. That is, at the time Chrétien wrote, the mother of his protectress was Queen of England and wedded to a sovereign who took a keen and personal interest in all that concerned King Arthur. The possibility of transmission is as clear as daylight; the question of course is, Would Marie be inclined to take advantage of it? The relations between her father and his divorced wife were certainly curious, as Louis made no objection to the marriage of the eldest son of Henry and Eleanor with his daughter by his second marriage, but whether there was intercourse between mother and daughter I have not been able to discover. But the question ought to be easily solved by some historical specialist who has made a study of that period. The point is interesting and important, and it is to be hoped some one will clear it up for us.
A question of secondary interest is whether Chrétien's poem is the source of contemporary and later allusions to the story. Of such allusions, or rather versions, we have two of special importance, that contained in Malory's compilation, and that given by Hartmann von Aue, in his Iwein. With regard to the former, I can only say that though I am in a position to offer new and important evidence with regard to the manuscript Malory used, and his method of composition, yet that evidence leaves the Charrette question unsolved. Of direct evidence there is none; the indirect and inferential evidence tends to show that Malory's source was not the poem of Chrétien de Troyes. The two points on which we can be certain are, (a) that Malory did not know the earlier part of the prose Lancelot at all, that his manuscript began at a point subsequent to the Charrette adventure; and (b) that he does not invent adventures, and but rarely details. Dr. Sommer's conclusions, as set forth in his Study on the Sources of Malory, are founded on very insufficient premises, and will need to be thoroughly revised to bring them into accordance with our present knowledge. This question I shall discuss fully in a later section. The Iwein version is of great importance, and though I have previously referred to it,47 yet in the light of Professor Foerster's strongly repeated assertion that Hartmann knew no other version of the story than that given by Chrétien, I think it is worth while going over the evidence again.
It must be remembered that Hartmann's Iwein is a translation of Chrétien's Chevalier au Lion, and though rather more diffuse, follows its source closely. In the French poem which, as we have noted above, immediately succeeded the Charrette, Chrétien deftly introduces more than one allusion to Guinevere's abduction. He says that Guinevere has been carried off by a knight d'estrange terre, who went to the court to demand her; but he would not have succeeded in carrying her off had it not been for Kay, who deceived or deluded (anbricona) the king into putting the queen in his charge (ll. 3916-39). In another place, he says that the king, 'Fist que fors del san Quant aprés lui l'an anvoia. Je cuit que keus la convoia Jusqu'au chevalier qui l'an mainne' (ll. 3706-11). Now, let us suppose that, as Professor Foerster insists, Hartmann had not read the Charrette and knew no other version of the story, what would he, who knew French well, and translates without blunders and confusion, understand by this? We must note particularly what Chrétien tells and what he omits. He distinctly says that the knight came to the court and demanded the queen (the real version of the poem is less blunt, as we have seen); that Arthur, deluded, put the queen in Kay's charge to lead her to the knight, and that they followed him. He does not say that the whole catastrophe came about through Arthur's granting a boon before he knew in what it consisted; he implies that the folly lay in Arthur's sending the queen after the knight, not in the circumstances which forced him to do so.
Now what does Hartmann say? In his version a knight appeared before Arthur and demanded a boon, the nature of which he refused to specify beforehand. Arthur granted it. It was that he should carry off the queen. This he did. The knights armed and followed. Kay was the first to overtake him, and was struck from his horse with such violence that his helmet caught in a tree and he hung suspended. He was not carried off captive. One after another all the knights are vanquished, and the queen carried off. Gawain is not at court; he returns the next day, and goes in search of the queen. Lancelot is not mentioned throughout; and the inference is that Gawain frees her.
What is specially noticeable in this account is that Hartmann agrees with Chrétien in the very feature which the French poet does not specify, i.e. the cause of the queen's abduction—a boon rashly granted, though he transfers the asking from Kay to the knight; while he differs from Chrétien in the feature which he does specify, i.e. that Kay takes Guinevere after the knight. Further, he adds details which would clear up some of the inconsistencies in Chrétien's own account: i.e. if Gawain were not present at the time, and all the knights followed one after the other and were defeated by Meleagant, we can quite understand that when Gawain returned the next day and followed on the trail, he would find traces of the severe and bloody conflict for which Chrétien's version leaves no room. On the face of it, Hartmann's version is much the more logical and coherent of the two. I have remarked above on the extreme awkwardness of the action at the outset of the story; that Meleagant should carry off Guinevere by a ruse similar to that employed by Gandîn in the Tristan poems is far more in accordance with mediæval tradition. If Hartmann's divergence is a mere 'invention,' he not only deserves praise for his sagacious skill in constructing a story,48 but excites admiration for the acuteness which enabled him to detect the leading motif of the adventure to which his source afforded absolutely no clue.
Wolfram von Eschenbach's references to the Charrette adventure are curious; at first sight it seems certain that he is referring to Chrétien's poem, but on closer examination the matter is not so clear. Thus he says that Lancelot crossed the sword-bridge, fought with Meljakanz (Meleagant), and freed Guinevere—all of which agree with Chrétien.49 But, on the other hand, he mentions Kay's suspension on the tree (Hartmann's version), and does not know that Meleagant was slain by Lancelot, or that the captive Bretons were freed by his coming—both Meleagant and the Breton knights are fighting at the tournament of Beaurösch.50 Indeed, Wolfram appears to know far more of these latter than can be gathered from Chrétien's poem. Of course, we cannot here say whether these references are due to Wolfram or to his source, which, as recent research has clearly shown, was certainly the work of a man of varied and extensive learning.51 Nor is it at all clear that Wolfram knew Lancelot as Guinevere's lover; he simply says that her imprisonment grieved him 'im was gevancnisse leit, die frou Ginòvêr dolte,'