But in the poem which followed the Cligés, Le Chevalier de la Charrette, Lancelot suddenly appears in both these characters, Gawain's superior and the lover of Guinevere: no explanation of the changed position is offered, but Chrétien takes for granted the familiarity of his audience with the relations between the knight and the queen. To add to the confusion, in the succeeding poem Le Chevalier au Lion, Lancelot is only once referred to, in connection with the Charrette adventure, and is never mentioned as one of the knights of Arthurs household; while in Chrétien's last poem, the Perceval, he is altogether ignored.6
It is very difficult, indeed impossible, to date Chrétien's poems with exactness. The only two which afford clear internal evidence on the point, Le Chevalier de la Charrette and Le Chevalier au Lion, fall within the years 1164-1173. Erec was the first of his Arthurian poems, and between Erec and the Charrette, certainly one work, Cligés, and it may be several, intervened.7
Very probably the Erec was written early in the decade, 1150-60, and taken in conjunction with the negative evidence afforded by the Brut and the Italian bas-relief, it goes to prove that whereas the name of Gawain, as connected with Arthur, was known by the end of the eleventh century,8 Arthurian tradition knew nothing of Lancelot till the latter half of the twelfth; and that no mention of his relations with Guinevere is found till between 1160-1170, that is, a decade after the first mention of his name. It is, of course, a well-recognised fact in the study of romance, that the date of a manuscript does not fix the date of the story contained in it; a younger manuscript may contain the same story under an older form. As a rule, the versions contained in Chrétien's poems appear to present a fairly old form of the stories they relate, saving in the case of Lancelot. About this knight, Chrétien either knows nothing or he knows too much. The earlier stages of his story he leaves unrecorded; yet an allusion in the Charrette poem9 shows that he was not unacquainted with the legend concerning his youth and upbringing. Two versions of this legend have been preserved to us, one in verse and one in prose. In the following chapter we will examine the older of these versions, and inquire into the origin of our hero's name.
CHAPTER II
The origin of the name Lancelot has been a subject of considerable debate among scholars, and has given rise to the most widely differing explanations. M. de la Villemarqué, who was a warm advocate of the Welsh origin of the Arthurian stories, derived the name from the French l'ancelot, a youth or servant, which he held to be a translation of the Welsh Melwas, or Maelwas. This solution was rejected by M. Gaston Paris, in his study on the Lancelot poems,10 in which he showed that ancelot was not a French common name, and that Maelwas did not bear the signification attributed to it. Professor Rhys,11 adopting the theory of the Welsh origin of the name, which in its present form he admitted only exists in Welsh literature as borrowed from French or English sources, decided that it represented a Welsh variant of Peredur, the root of this latter name being Pâr=a spear or lance. 'The characters,' says Professor Rhys, 'were originally the same, though their respective developments eventually differed very widely.' I doubt if this solution ever found any adherents except its author: it is sufficient to remark that the derivation of Peredur, on which it rests, is by no means universally accepted, and that Lancelot is in no special way connected with a spear or lance.12 It is certainly true that the Lancelot story shows signs of having been affected by the Perceval legend, but as we shall see the borrowings are restricted to one special and purely continental form of the story.
M. Gaston Paris, in the study referred to above, suggested that Lancelot might be either a Celtic name altered, or, more probably, the substitution, by French poets, of a name of Germanic origin for one of Breton form strange to the ears of their French audiences, e.g. it might be a diminutive form of Lanzo. This is also the conclusion of Professor Zimmer.13 The prefix Lant is often found in names of Frankish origin transferred to Breton ground: such names are Lando, Landolin; Lanzo, Lanzolin, etc.
In the introduction to his edition of the Charrette, recently published,14 Professor Foerster announces his complete adhesion to this view.
It certainly seems that the evidence points strongly to this conclusion. The fact that Lancelot's name does not appear in the earliest obtainable Arthurian documents shows that he did not belong to the original 'stoff' of the cycle; the entire silence of Welsh literature, and the practical silence of English vernacular romances,15 seem to show that he formed no part of the insular Arthurian tradition. For my own part I unhesitatingly accept Professor Foerster's dictum, 'Lancelot ist den Kymren gänzlich unbekannt, und ist unter allen Umständen Kontinentaler16 Herkunft.'17
A weak point in the proposed Celtic solutions appears to me to be that both entirely ignore the qualifying title du Lac, by which Lancelot is invariably known. Neither M. de la Villemarqué nor Professor Rhys appear to consider it of any special importance, yet if I mistake not this is just the significant point of the Lancelot story, and that which from the very outset differentiates it from the legends connected with Peredur or Maelwas. From the moment of his appearance in Chrétien's list of Arthur's knights to that in which the prose Lancelot records his death in the odour of sanctity, Lancelot is Lancelot du Lac, and the earliest version of his story which we possess amply justifies his claim to the title.
The poem of Ulrich von Zatzikhoven18 is certainly later than either the Erec or the Charrette of Chrétien, but the tradition it embodies is anterior to the poem itself. Written in the opening years of the thirteenth century, it is, as explicitly stated in the text, the translation of 'daz welsche buoch von Lanzelete,' brought to Germany by Hugo de Morville, one of the hostages who in 1194 replaced Richard of England in the prison of Leopold of Austria.19 The date of the original French version cannot, of course, be fixed. In any case it must have preceded its introduction into Germany; judging from internal evidence it represented an early and immature version of the Lancelot legend. The story as related in the Lanzelet is as follows: Lanzelet was son to King Pant of Genewîs and his wife Clarine. By a revolt of his people Pant was driven from his kingdom with his wife and child. In his flight he came to a stream, and there, overcome by his wounds, sank down and died. The queen had laid her child under a tree while she tended her husband, and before she could reach it again a water-fairy (mer-feine) came in a cloud of mist and carried off the infant. The fairy was a queen, ruling over ten thousand maidens, who knew no man. Her kingdom was called Meide-lant; there it was ever May-tide, and her palace