Brittany, where it was profoundly modified, Tristan being provided with a Breton parentage and home, and where in all probability it was worked into the Arthur saga.
Such is a very bare summary of this brilliant and fascinating hypothesis. I must leave the criticism of it to those who are more familiar than I am with the oldest French forms of the Tristan story. I would merely note that evidence, which Prof. Zimmer himself quotes, shows that the saga must have been current in Wales before 1081, and probably before 1072. Moreover, that no light is thrown upon the curious Welsh traditions concerning Tristan, traditions which cannot either be explained from the French romances. But let us accept his results and see what bearing they have upon his general theory of the Arthurian romance. Here is a story, originating in these islands, unknown in Brittany before the close of the eleventh century, and yet the oldest French forms are Breton in locale and characterisation of the personages. What reliance then can be placed upon Breton traits in other branches of the romances as evidence of their specific Breton and non-insular origin? What has happened once may have happened more than once—the early specific Breton lais to which Prof. Zimmer traces back the French romances may be, as he himself claims that the Tristan lais are, mere Bretonised variants of insular originals.
Thus whilst admitting in a very large measure the validity of Prof Zimmer's claims on behalf of the Breton element in the formation of the Arthurian romance, I cannot but think that he has often misinterpreted the nature of that element, that he has exaggerated the consequences to be drawn from the facts he has stated, and that he has unduly depreciated the influence of the insular element. Be this, however, as it may, the services he has rendered to the study of the cycle are of extreme value, and for years to come his investigations must form the basis of further research.
