UNIVERSITY OF FLORIDA LIBRARIES COLLEGE LIBRARY Digitized by the Internet Archive in 2011 with funding from LYRASIS Members and Sloan Foundation http://www.archive.org/details/evolutionofarthuOObruc The Evolution of Arthurian Romance From the Beginnings Down to the Year 1300 by James Douglas Bruce, Ph. D. Professor of the English Language and Literature in the University ol Tennessee. Second Edition with a supplement by Alfons Hilka (Goettingen) Volume I GLOUCESTER, MASS. PETER SMITH 1958 SECOND EDITION, I928 REPRINTED I958 BY PERMISSION OF THE JOHNS HOPKINS PRESS ALL RIGHTS RESERVED Preface. The present volumes are the outgrowth of a series of six lectures on the Arthurian Romances which I delivered before the graduate students of the Modern Language Departments of the Johns Hopkins University in December, 1912, and which I repeat- ed in the Summer School of the University of Pennsylvania in 1915 as part of a general course on the Mediaeval Romances. It was only after the repetition, however, just mentioned that I decided to work these lectures out in a fuller form for publication. The task of putting this purpose into effect grew constantly upon my hands, so that the final result is a far bulkier treatment of the sub- ject than I originally had in contemplation. On the other hand, the need of a guide through the mazes of mediaeval Arthurian ro- mance and of the vast body of modern critical writings pertaining there-to has been long felt by students of mediaeval literature, and the book, in its present form, is offered as an attempt to meet that need. I regret the necessity of fixing upon 1300 as the downward limit of the treatise. This regret applies less, to be sure, to the continental romances of the cycle — since, notwithstanding the pop- ularity in its own day of such a work as Le Petit Artus de Bre- taigne, for example, all really notable productions in this genre on the continent antedate that year — than to those written in English. For, except Sir Tristrem and Arthour and Merlin, neither of which rise above mediocrity, the extant English ro- mances of the cycle were composed after the year 1300. Never- theless, even as regards the English romances, the expressed regret was mitigated by the reflection that, after all, mediaeval English literature can boast of only four contributions of sub- stantial importance to Arthurian romance, viz. Sir Gawain and the Green Knight, the alliterative Morte Arthure, the stanzaic IV Preface Morte Arthur, and Malory's great compilation in prose. In any event, the determining factor in the limitation of time which has here been adopted was that to have attempted to carry on the work any further would have meant an indefinite postponement of its completion. In a domain where almost everything is the subject of con- troversy, as is the case with Arthurian romance, it is difficult for the author of a general treatise, like the present one, who has strong convictions in regard to the questions under debate, not to seem frequently over-dogmatic or even unfair to his opponents in the expression of these convictions. In the notes and in the division of the book entitled "Discussions", I have endeavored to justify at length my own position in many of these matters, but, owing to considerations of space, it has been obviously impossible to do this in all instances, and so I have often been compelled to content myself with inserting into my presentation of the subject such qualifying clauses as "in my opinion" and the like. But since such clauses are apt to become tiresome by repetition, I have on many occasions ventured to omit them, without intending, of course, to claim infallibility in leaving the expression of my views thus unqualified. As an example of such controversial questions tho age-long debate concerning the debt of the Arthurian romances to Celtic sources may be cited. Like Professors Foerster and Golther before me, I am convinced that this debt has been, generally speak- ing, greatly exaggerated and that personal invention was the most important factor in the creation of these romances — that we have, therefore, in this species of literature the products of a literary fashion steadily developed by successive generations of writers in an age when the sense of literary property was virtually non- existent and men did not hesitate to use their predecessors' com- positions for any purpose they chose — not, in any essential degree, the reflection of a great body of oral tradition.1 My presentation 1 This is undeniably true of the prose romances. I believe, however, that if we deduct most of the lays and a few episodes (indicated passim below) in the earlier romances, the statement will also hold good of the metrical romances. After all, most of the extant Preface V of the matter proceeds, accordingly, from this point of view, al- though I have, of course, tried to be fair to the scholars who espouse the opposite theory. Indeed, my position with regard to the whole general question concerning the folk-tale sources of the Arthurian romances will be found to be somewhat similar to that just indi- cated in the case of the Celtic problem. For, even apart from the question of specifically Celtic sources, in my opinion, the ele- ment of individual invention and of purely literary origins in the romances has been unduly minimized. The authors of these ro- mances were primarily poets, not transcribers of folk-tales, and it seems strange that scholars should so often have imputed to them the strictest accuracy in following imaginary folk-tale sources when we know that even the professional collectors of folk-tales in modern times have rarely taken down such stories from oral recitation without introducing into their texts numerous unauthorized alter- ations. A further word of explanation is, perhaps, due with regard to the inclusion of many romances that are mediocre or even worse in the division of Vol. II which is devoted to analyses. This has been done with an especial view to the students of stories, who are, of course, particularly numerous in the field of mediaeval studies. The more tedious the romance, the more grateful for such con- densed synopses, doubtless, will be the specialist who is endeavor- ing to run down the history of some motif in fiction. In any event, as I hardly need observe, these analyses have been grouped to- gether in a section by themselves, so that they do not interrupt the general narrative of the development of Arthurian romance, and, consequently, readers who may be interested in the remainder of the book will be able to omit without inconvenience such analyses in this section as do not attract them, or, indeed, the whole section. In conclusion, I wish to express my gratitude to Professor James W. Bright for having offered to include this book in the romances in verse were written after the composition of the prose ro- mances had begun, and, obviously, there was no reason why the authors of the former and of the latter, respectively, should have drawn upon sources of a different character. VI Preface series, Hesperia, — also, to the authorities of the British Museum and Widener Library (Harvard University), to whose well-known liberality and courtesy I have been deeply indebted whilst prosecut- ing my Arthurian studies in those institutions. University of Tennessee, Knoxville, Tennessee, J. D. BRUCE. November 16, 1922. Contents of vol. I. Part I: Traditions, Chronicles, Lays, and Romances Chapters Page I. Early Traditions concerning- Arthur 3 II. Origin of the Lays and Romances 37 1. The Lays 52 2. The Romances 67 3. The Origin of the Round Table and Some Other Questions of Celtic Origins 82 III. Chretien de Troyes and his Successors 100 IV. Merlin 129 V. Tristan 152 VI. Lancelot 192 Part II: The Holy Grail I. Chretien, Robert de Boron, and the Theory of Christian Origin 219 II. The Theory of Celtic Origin 2G9 III. The Ritual Theory 277 IV. Continuations of Chretien 290 V. Sir Perceval of Galles 309 VI. Wolfram's Parzival- 313 VII. Peredur, Diu Crone, and Sone de Nausay 342 VIII. Other Theories Concerning the Origin of the Grail . . .354 Part III: The Prose Romances I. Beginnings of the Prose Romances 365 II. The Prose Cycles 368 III. The Vulgate Cycle 374 1. L'Estoire del Saint Graal 374 2. L'Estoire de Merlin 395 3. Li Livres de Lancelot 397 4. La Queste del Saint Graal 419 5. La Mort Artu 426 IV. Variant Versions of Parts of the Vulgate Cycle . . . .443 V. Date of the Vulgate Cycle 450 VI. Development of the Vulgate Cycle 455 VII. The Pseudo-Robert de Boron Cycle of the Prose Romances 458 VIII. Li Contes del Brait Merlin 480 IX. The Prose Tristan 483 Abbreviations employed in the present treatise. (Herrig's) Archiv f. d. St. der n. Spr. = Archiv fur das Studium der neueren Sprachen. LB = Literaturblatt fur Germanische und Romanische Philologie. LZ = Literarisches Zentralblatt fiii Deutschland. MLN = Modern Language Notes. MLR = Modern Language Review. MPh = Modern Philology. PMLA = Publications of the Modern Language Association of America. RR — Romanic Review. RTR = Paulin Paris's Romans de la Table Ronde (5 vols., Paris, 1868-1877). Sommer = H. 0. Sommer's The Vulgate Version of the Arthurian Romances (7 vols, plus Index, Washington, D. C, 1908—1916). Vollmoller, JB = K. Vollmollers Jahresbericht tiber die Fortschritte der Romanischen Philologie. Zs. f. celt. Ph. = Zeitschrift fur celtische Philologie. Zs. f. d. A. = Zeitschrift fur deutsches Altertum. Zs. f. d. Ph. = Zeitschrift fur deutsche Philologie. Zs. f. frz. Spr. u. Litt. = Zeit&chrift fur f ranzosische Sprache und Litteratur. Zs. f. rom. Ph. = Zeitschrift lur romanische Philologie. PART I. TRADITIONS, CHRONICLES, LAYS, AND ROMANCES. Chapter I. Early Traditions Concerning Arthur in the Chro- nicles and Elsewhere. In the earliest recorded traditions concerning Arthur, he is represented as a victorious leader of the Britons — early in the sixth century, it would appear — in their desperate struggle against the Anglo-Saxon invaders. But did any such person ever actually exist? The candid historian must admit that the evi- dence on the subject (which we shall review in this chapter) is meagre, relatively late, and almost wholly fantastic, and that, consequently, an affirmative answer to this question is, by no means, inevitable. Nevertheless, we may say that under the con- ditions that prevailed in the wild years just referred to there is nothing unlikely in the supposition that some born leader of men may have emerged from the mass of the Britons and by his energy and valor have enabled them to win such signal, though temporary, successes over their enemies as to have secured for him a legendary fame with posterity. Moreover, strong confirmation of Arthur's historical character seems afforded by the fact that his name is, in its origin, not Celtic, but Roman, being derived from the name, Artorius,1 which occurs in Tacitus2 and Juvenal3 and which is, 1 Cp. Heinrich Zimmer, Gottingische Gelehrte Anzeigen for Oct. 10, 1890, p. 818. The German scholar cites ibid., note 1, various Celts with Roman names, e. g. Constantinus and Aurelius (Conan) among the Celtic princes denounced by Gildas (first half of the sixth century). Even the names of some of the best-known characters that figure in Celtic tales were derived from the Romans, e. g. Yvain-Owein from Eugenius, Geraint from Gerontius, Kei (perhaps) from Cams. Simil- arly, R. Thurneysen, Zs. f. deutsche Philologie, XXVIII, 91, note 1, with reference to Ywein, and ibid., 97, note 1, with reference to Urien (Urbgen) = Urbigenus. It was a curious fancy of A. Holtzmanns, 4 Evolution of Arthurian Romance indeed, the name of a Roman gens. During the long centuries of Roman occupation of Britain (from the first century to the fifth), many Romans, from one circumstance or another, had settled in that country, and a still larger number of the native Celtic in- habitants had become Romanized. When the Anglo-Saxon on- slaught came, such elements in the population shared, of course, the fortunes of the rest, and, owing to the superiority in education and wealth which they generally enjoyed, were, perhaps, even more likely than the pure Britons to produce a national hero. Certainly we know from Gildas (ch. 5) that one of the principal leaders of the Britons in their wars against the invaders was Ambrosius Aure- lianus, "a man of Roman race.'' When the chroniclers, then, report that Arthur won the battle of Mons Badonis (Badon Hill) — an historical battle fought, undoubtedly, as we know, in the first part of the sixth century — it seems wiser to accept this statement as authentic and to recognize in him a man of Roman descent or a Romanized Celt, who, in these times of stress, attained the leader- ship of the British hosts. There is no need, therefore, of regarding him simply as a creature of the popular imagination or of vapori- zing him into a hypothetical culture-divinity, as various scholars have done.4 Pfeiffer's Germania, XII, 279 (1867) that not only were Arthur and Vortimer, son of Vortigern and enemy of Hengist and Horsa, the same person, in British and Anglo-Saxon tradition, respectively, but that their names were, also, identical, Vortimer having become Arthur, as he sur- mises, in Welsh pronunciation. 2 Annals ; Book XV, ch. 71. 8 Satires, III, 29 — here in the feminine form, as a woman's name, Artoria Flaccilla. 4 So John Rhys, Studies in the Arthurian Legend, pp. 39 ff. (Oxford 189i). He acknowledges that the derivation of Arthur from Artor, Artorius is, phonetically, unobjectionable; nevertheless, with his customary weakness for the fanciful mythological interpretations of Max Miiller and his school — once so popular, but now generally dis- credited — he endeavors to connect the name with the Aryan root, ar — (= to plough), and hence conjectures that Arthur was by origin a culture-divinity. Among the interpretations of Arthur's name, I note, also, the follo- wing: In his article, "King Arthur and Gildas", The Academy, Oct. Early Traditions Concerning Arthur 5 Let us examine now in due order the testimony of the earliest records with respect to Arthur and his achievements. It appears surprising, at first sight, that Gildas, the British historian, who is our earliest authority on the Anglo-Saxon con- quest of Britain, in his De Excidio et Conquestu Britanniae — 5 written about 540, A. D. — should not mention Arthur, although he refers (ch. 26) to the battle of Badon Hill with which later chroniclers connected that hero's fame.6 The explanation, however, may well lie in the fact that Gildas's "Epistle", as he calls it (ch. 1), is not a regular historical narrative — it was not composed as such — but rather a homily on the misfortunes of the Britons as due to their sins — a "tract for the times", as a modern editor has aptly called it — 7 and that it gives us, therefore, simply glimpses of persons and events rather than an orderly account of them. In- deed, a large part of the work consists of a mere cento of passages 12, 1895, E. W. B. Nicholson identifies with Arthur the personage referred to by Gildas, ch. 32, as "Ursus", and regards Arthur as made up of two old Celtic words, artos (= bear) and vivos (= man). Hence, the name would mean Bear-male or He-bear, J. Pinkerton, I may remark, in his edition of Barbour's Bruce, I, 26 (London, 1790), had anticipated Nicholson's etymology, but had interpreted the combination as meaning simply "the great man". As a matter of fact, Celtic proper names with avtos (= beav) in composition are numerous. Cp. A. Holder, Alt-Celtischev Spvachschatz, 226 f. (Leipzig 1891), although Holder, himself, suggests a probable connection oiAvthur with Irish art (=== stone). 6 The standard text of this work is Theodor Mommsen's, in the Monumenta Gevmaniae Histovica, Auctorum Antiquissimorum, vol. XIII, (Berlin, 1898). Hugh Williams has reprinted this text (with a few changes) in his edition, Cymmrodorion Record Series, No. 3 (London, 1899 — 1901). He gives, also, a translation and valuable notes. Among the older editions, San Marte's — Nennius und Gildas, (Berlin, 1844) — is still useful. For the critical literature on Gildas, cp. R. H. Fletcher, The Arthurian Material of the Chronicles, espe- cially those of Great Bvitain and France, pp. 2f., note, (Boston, 1906): [Harvard] Studies and Notes in Philology and Literature, X. 0 According to the most acceptable construction of this same passage, indeed, he tells us in it, also, that he was born in the year of that battle. He wrote the De Excidio before 547. Cp. Mommsen's edition, p. 5. 7 Hugh Williams, p. V. 6 Evolution of Arthurian Romance from the Scriptures — especially, thunders of the Old Testament prophets against the sins of the Jews — both people and rulers — which Gildas here applies to his own fellow-countrymen.8 From the silence of such a work no inference can be drawn as to the matter in hand. The first record9 of the name, Arthur, occurs in Adamnan's life of Saint Columba,10 the famous Irish missionary, who founded the monastery at Iona in the Hebrides. In a Latinized form, Arturius — for Adamnan's work is in Latin - - it is there found as the name of a young Irish prince who perished in a battle at Tigernach (Ireland) in the year, 596. The father of this young man, Aed mac Gabrain, was the ruler of an Irish state, Dalraida, on the southwest coast of Scotland. Still further, in the seventh century we find a Welsh prince,11 a British prince12 and an Irish- man whose rank is unknown,13 all bearing this name, and since Arthur was not a common name among the Celts, perhaps, we are justified in interpreting these instances of its occurrence as testi- mony to the existence of legends in the sixth and seventh centuries concerning the person who was destined to render it so illustrious.14 The first direct mention, however, of the great sovereign of ro- mance occurs in the account of his exploits in the Historia Brit- h Virtually only Part I (ch. 1 — 26) gives any historical infor- mation. 9 For the first three Arthurs here noted, cp. H. Zimmer. Nennius Vindicatus, pp. 283 ff. (Berlin 1893). 10 Book I, ch. 9. Cp. J. T. Fowler's second edition (Oxford, 1920). Adamnan died in 704. His Vita Columbae was written before 697 and is preserved in a MS. not later than 714. 11 Arthur map Petr, who was ruling in Dyfed (Southwest Wales) in the first decade of the seventh century. 12 Arthur, son of Bicuir, apparently a prince. He slew a famous Ulster chieftain in Cantire (Southwest Scotland) in 625. 13 Cp. Kuno Meyer in his article "Eine verschollene Artursage", pp. 63ff. of the Festschrift fur Ernst Windisch (Leipzig, 1914). The name occurs here in a list of ecclesiastics who stood security for the execution of the law known as Cain Adamnain. The list was compiled in 697 and contains the name, Faradach ho a Artur — i. e. "Faradach, grand-son of Arthur". This Arthur, then, must have lived in the early part of the seventh century. So Zimmer, loc. cit. Early Traditions Concerning Arthur 7 tonum (ch. 56), which goes under the name of Nennius.15 Here it is said that, after the death of Hengest, the Saxon chieftain, Arthur fought against the Saxons cum regibus Brittonum, sed ipse dux erat bellorum — that is to say, that he was not himself one of the British kings, although he commanded their forces in battle.16 Then follows an enumeration of his twelve great vic- tories,17 concluding with that of Badon Hill, which we know from Gildas to be historical — only here in Nennius the narrative is already colored with legend, for it is said that Arthur alone slew 960 of the enemy in a single attack.18 16 The standard edition of Nennius is T. Mommsen's. It is in the same volume as his edition of Gildas, mentioned above. San Marte, also, edited Nennius, in the same volume as his Gildas. Cp. p. 5, note 5, above. His notes are still useful. For bibliography of Nennius, cp. Fletcher, op. cit., pp. 8f., note. 18 Richard Thurneysen, Zs. f. deutsche Philologie, XXVIII, 98, note 1 (1895), suggests that the term dux was applied to rulers who did not spring from the hochadel (higher nobility). It is evident, how- ever, here that Arthur is not conceived of as a ruler. 17 The variants in Mommsen's edition (p. 200) show great con- fusion in regard to the eleventh in the list of Arthur's battles, "bellum in monte qui dicitur Agned." We shall see below, Part IV, that Anscombe has argued that "Mons Badonis" is simply a corruption for "Mons Hagonis" ("Hill of Agon"). Except for one MS., however, all MSS. seem to have "Mons Badonis" for the twelfth. A. Wade- Evans, Y Cyrnmrodor, XXII, 155, contends that the name of this battle got into the list from "pseudo-Gildas" (i. e. Gildas, whose work he regards as spurious and late) and Bede; but the analyses of Zimmer and Thurneysen seem to establish the antiquity of the whole list. The entry in the Annates Cambriae (under 516), however, concerning Mons Badonis apparently possesses no value. The annalist, it seems, took it from Nennius and attached to it details which he found in Nennius con- nected with the eighth battle ("in castello Guinion"). Cp. on the sub- ject A. de la Borderie, Revue Celtique, VI, 2, and W. H. Stevenson, The English Historical Review, XVII, 633, note 34. The most detailed effort to identify the sites of the twelve battles is Anscombe's Zs. f. celt. Ph., V, 103ff. (1905), but in no case can he be said to have attained assured results. Anscombe, in the article just cited (p. 116), interprets this 960 as due to some scribal blunder. The figure was originally 470, he says, and indicated, really, the date of the battle of Mons Hagonis (the 8 Evolution of Arthurian Romance In the section of this work (ch. 73) called De mirabilibus Britanniae, we have still further mention of Arthur in connection with a marvel localized in the region called Buelt — the present Builth in South Wales. Here it is stated that there was in this place a heap of stones, which was crowned by one stone that bore the impress of the foot of Arthur's dog, Cabal. The stone had received this impress during Arthur's hunting of the boar Troynt — an adventure which is related fully in the mediaeval Welsh tale of Kulhwch and Olwen. Nennius says that, although one might remove this stone from the heap one day, it would always be found there the next. Immediately after this we find still another mar- vel connected with Arthur — the tomb of his son, Anir, who had fallen at the hands of his father. This tomb varied in length at every new measurement. At one time it might measure six feet, at another fifteen — "and I alone have tested it," says the vera- cious chronicler. The work which bears the name of Nennius is composite in character, and the history of its growth seems to be as follows:19 name he gives to the great British victory over the Saxons which in our text of Gildas is called, in the genitive, "Badonici montis"). "Geof- frey of Monmouth, Book DC, ch. 15, as Anscombe contends, preserves the correct figure, viz. 470, only Geoffrey, too, uses it of the number of men that Arthur killed in the battle of Mount Badon. Wade-Evans, also, adopts these views of Anscombe's. Cp. Y Cymmrodor, XXII, 128. The speculation, however, is anything but convincing. 18 My account of this growth is based on the investigations of Heinrich Zimmer, in his Nennius Vindicates (Berlin, 1893) and of Thurneyseu, in his reviews of this work of Zimmer's in the Zs. f. deutsche Philologie, XXVIII, 80ff. (1895) and of Mommsen's edition of Gildas and Nennius in Zs. /. celt. Ph., I, 157ff. (1896), respec- tively. Before the publication of Zimmer's work the value of the Historia Brittorum was underestimated and the connection of any one named Nennius with it doubted. Moreover, no part of it was regarded as of so considerable an antiquity as now seems to have been actually the case. The best representative of this depreciatory tendency is Arthur de la Borderie: L' Historia Britonum attribute a Nennius et V Hi- storia Britannica avant Geoffroi de Monmouth (Paris and London, 1883). Similarly, G. Heeger, Uber die Trojanersagen der Britten Munich, 1886), and even later, W. W. Newell, "Doubts concerning Early Traditions Concerning Arthur 9 In the third decade of the ninth century a writer named Nen- nius, who lived in South Wales, re-edited and expanded an older Historia Brittonum that dated back to 679, or, perhaps, in part, to an even earlier period in the seventh century. Now, the passage cited above concerning Arthur and his twelve victories belongs to the oldest portion of the work and shows the vigorous develop- ment of legend about the hero's name within something like one hundred and fifty years of the time when he must have lived. The section which relates the marvels of Arthur's dog and of Anir's the British History attributed to Nennius", PMLA, XX, 622ff. (1905). Both Zimmer and Thurneysen accept the earlier Historia Brittonum (mentioned in the text above) as dating from 679 and as containing still older materials. Thurneysen, Zs. f. d. Ph., XXVIII, 83 f., is inclined, still further, to ascribe to a Run (Rum) mab Urbgen (mentioned in Nennius, ch. 63) the compilation of a part of these older materials. Urbgen (a historical character who died between 572 and 579) seems to have been the Urien of the romances; Run would, therefore, be the brother of the Yvain (Ivain) of these romances. Zimmer neglected the important Chartres MS. (ninth or tenth cen- tury) of the Historia Brittonum, which preserves us the work in a pre-Nennian form. L. Duchesne published the text of this MS. in the Revue Celtique, XV, 174ff. (1894), "Nennius Retractatus ", together with a discussion of Nennius problems. Thurneysen had the advantage over Zimmer of the use of the Chartres text, and, taken altogether, his conclusions concerning the origin and development of the Historia Brittonum, as stated Zs. f. celt. Ph., I, 166f. (1896), are to be pre- ferred, viz. that in the neighborhood of Builth (in South Wales), in the year 826, Nennius (Nemnius), compiled the work in the full form, known as the Harleian recension. Immediately thereafter, on the advice of his teacher, Beulan (cp. ch. 10), he prepared an abbreviated edition of this Harleian recension. Later on, he entered additions and corrections in this abbreviated recension. Excerpts from the life of St. Germanus by Map Urbgen constitute the oldest elements of the work. For the MSS. of the different recensions, see Mommsen, pp. 119ff. and Thur- neysen, Zs.f. celt. Ph., I, 158f. — Windisch, Das keltische Britannien, etc., p. 41, speaks of the Harleian recension as "vornennianisch", which seems, however, an error. In Romania, XXIII, 432 ff. (1894) Cesar Boser, in his article, "Apropos de Nennius", had already (before Thurneysen) attacked the argument by which Zimmer had attempted to fix the date of Nennius. He suggests, however, no alternative date. 10 Evolution of Arthurian Romance tomb is one of the later additions, as the manuscript tradition proves, but it testifies also to the continued growth of wonderful legends about the British chieftain. It does not surprise us then to find that in the year, 1113, stories concerning Arthur were firmly established both in Brittany and in Cornwall. In that year certain monks of Laon in Brittany were sent to England to beg mo- ney for the rebuilding of their cathedral which had recently been de- stroyed by fire. We have an account of their experiences preser- ved in a treatise by Hermann of Tournai.20 From this account it appears that a servant of the monks got into a dispute with a Cornishman as to whether Arthur was still alive — "exactly in the same way" remarks the chronicler significantly, "as the inhabi- tants of Brittany dispute with the French over Arthur." The affair drew together a mob and there would have been bloodshed, if a local ecclesiastic had not intervened. The chronicler con- cludes naively that the man who started this brawl was punished for doing so — for he had a withered hand and had come there to be cured by the relics which the visiting monks had brought with them. The Holy Virgin, however, was evidently displeased with him, for the relics would not work a miracle on him that day. Still further, in a life of the Cornish saint, St. Carantoc, also dating from about the beginning of the twelfth century, we find Arthur reigning in Cornwall and hunting a dragon which had devastated his dominions.21 20 Cp. Migne's Patrologia Latina, Vol. 156, col. 973. 21 Cp. F. Lot, Romania, XXX, pp. 2ff. (1901). St. Carantoc, however, did not allow any one to kill the serpent. He led the monster about like a lamb. Lot identifies Dindraitliov, where, according to the legend, Arthur was reigning, with the modern Castle an Dinas (about ten miles east of Crantock). He points out, moreover, that Arthur is represented in Kulhwch and Olwen (Loth's Mabinogion2, I, 344) as chasing the marvellous boar, Twrch Trwyth, in Cornwall (Kernyw) and as having (Loth, op. cit., I, 38, 331, 334 et passim) a royal residence at Kelliwic (Bodmin) in Cornwall. He commands troops from Cornwall and Devon in the Vita Gildae, ch. 10, and has among his retinue Gwynnhyvar (Loth, op. cit., I. 277), who is a high officer in those kingdoms — a man who was, also, among the persons responsible for the battle of Camlan. Cornwall figures, furthermore, in the tale of Arthur's fight with the Cath Paluc; for the mother (the sow, Henwen) Early Traditions Concerning Arthur 11 We need not linger over the meagre borrowings from Nen- nius which are incorporated in certain chronicles of the two cen- turies or slightly upwards that followed the compilation of that work in its final form,22 for they have no independent value. In these chronicles of the succeeding centuries, only one entry is of importance, namely, that which in the Annales Cambriae, under the year 537, records for the first time in the barest possible of this monster was, according to a triad (Loth, op. cit., II, 271), in a drove of swine kept at Glynn Dallwyr, in Cornwall. On Arthur's combat with the Cath Paluc, cp. E. Freymond's monograph, Artus' Kampf mit dem Katzenungestum in the Festgabe fur Gustav Grbber (Halle, 1899). — The discussion of Arthurian localities in Cornwall by W. H. Dickinson in his Arthur in Cornwall (London, 1900) has been superseded by that of J. Loth, in his Contributions a Vetude des Romans de la Table Bonde (Paris, 1912). It should be remembered, too, that according to the Welsh tra- dition, Arthur's last battle took place at Camlan — doubtless, Camel- ford in Cornwall, with which it is still locally identified. Cp. F. Lot, Romania, XXX, pp. 16ff. (1901) and my edition of the Mort Artu} pp. 291 ff. This Welsh tradition would naturally be of Cornish origin. On the improbable identification of Camelon in Scotland as the scene of Arthur's final battle, cp. the works named, p. 73, note 72, below. 2 There are four such chronicles: 1. Annales Cambriae, compiled in Wales in the second half of the tenth century. 2. A brief Chronicle of Mount St. Michael, apparently, of Breton origin, and probably of the eleventh century. 3. Aethelweard's Chronicle, written by an Eng- lishman, probably early in the eleventh century. 4. Henry of Hun- tingdon's Historia Anglorum, English, of course, and probably not long before 1133. For these chroniclers and their relations to Nennius, the various editions of their works, etc., cp. Fletcher, op. cit., pp. 3 Iff. The best editions of the four chronicles, respectively, are by, 1. E. Philli- more, Y Cymmrodor, IX, 141 ff. (1888), reprinted by J. Loth, Mabi- nogion", H, 370ff. This gives the oldest (uninterpolated) version, which is the only one we need consider. 2. Migne, Patrologia Latina, vol. 202, cols. 1323ff. Cp. first entry. 3. H. Petrie and J. Sharpe's Mo- numenta Historica Britannica, I, 499 ff. Aethelward uses Nennius, it seems, but does not mention Arthur. 4. Thomas Arnold in Rolls Series, vol. 74 (London, 1879). Cp. p. 48. There is no need of conjecturing with Fletcher that nos. 3 and 4 may owe something, also, to popular tradition. 12 Evolution of Arthurian Romance way23 the final battle (here called of Camlann) between Arthur and Mordred (Medraut). The first indisputable indication that the fame which Arthur enjoyed among the Celts had spread to other peoples is supplied by the Gesta Regum Anglorum of William of Malmesbury, which was completed in 1125. Besides the information concerning the British chieftain which he derived from the above-mentioned pas- sages in Nennius, this writer condemns (I — 8) the idle tales which the Britons circulate about their hero as detracting from, rather than adding to, his real glory, and explains (III, 287) the belief that this hero will return from the fact that no one has seen his tomb, adding at the same time an account of the death and sup- posed tomb of his nephew, Gawain, in Wales. The same writer's work on the antiquities of Glastonbury — De Antiquitate Glasto- niensis Ecclesiae — contains more Arthurian material, but the passages are, no doubt, late interpolations and will be dealt with elsewhere.24 If we could accept the contention of some scholars, much the most interesting evidence in regard to the diffusion of Arthurian stories before Geoffrey of Monmouth, would be that which was brought to light in the year, 1888, by the Italian scholar, Pio Rajna.25 In the articles referred to Rajna collected numerous ex- amples of the names, Artusius and Galvanus (Walwanus, Wal- quanus, etc.) — that is, Arthur and Gawain, as he interpreted them26 — from the historical records of Northern Italy, including Tuscany and the Marches in the twelfth century. Artusius is found as early as 1114 and Galvanus as early as 1136 — and, in both cases, as the names of grown men — so that these names must have been given many years earlier to the persons that bore 5 Gueith Camlann (battle of Camlann) in qua Arthur et Medraut corruerunt; et mortalitas in Britania et in Hibernia fuit. Cp. Loth's Mabinogion*, II, 372. The Annales, in their earliest form, belong to the latter part of the tenth century. Cp. op. cit., p. 370. 24 Cp. Part II. 26 See his articles entitled "Gli eroi Brettoni nell' onomastica Ita- liana", Romania, XVII, 161 ff., 355ff. (1888). 3 Even Foerster accepted Rajna's interpretations without question. Cp. Zs. f. rom. Ph., XX, 247. Early Traditions Concerning Arthur 13 them — in the case of Artusius not later than 1090. Here, then, in the closing years of the eleventh century the tales concerning Arthur, if we accept Rajna's interpretation, were so widespread and popular as to affect the nomenclature of Northern Italy. Moreover, the forms of these names would appear to make it evident that the stories in question reached Italy in French ver- sions, for Artusius seems, at first sight, plainly a Latinization of Artus, the Old French nominative of Arthur, and Galvanus si- milarly the Latinization of Old French Gauvain.27 It is to be ob- served, however, that even if Rajna's identifications of those names are correct, it does not follow necessarily that his inferences are equally so, for Arthur was a Breton name and may have been brought to Italy by actual Bretons or inhabitants of French di- stricts contiguous to Brittany, independently of any specifically Arthurian traditions.28 The correctness of the identifications in question, however, is not entirely assured, and the forme, Artusius and Galvanus, respectively, may be variants of other names than those proposed by the Italian scholar.29 97 In Geoffrey of Monmouth, Arthur's name is Latinized as Ar- thurus, Gawain's as Walgannus, Walgainus — probably from French Gauvain. Cf. Lot, Romania, XXVff. (1896). 28 This objection was already, raised by Suchier in Suchier and Birch-Hirschf eld's, Geschichte der Franzosischen Literatur1, p. 141 (Leipzig, 1900). H. Zimmer, Gottingische Gelehrte Anzeigen for Oct. 1, 1890, p. 831, note, suggests that the names may have been brought to Northern Italy by Norman mercenaries, who are known to have been in the ser- vice of states in that part of the peninsula as early as the eleventh century. He recalls, too, the Norman conquest of Sicily by Robert Guiscard and his followers in the same century. 29 In his Recherches sur les sources latines des contes et romans courtois du moyen age, pp. 396 f. (Paris, 1913), Edmond Faral has suggested that Artusius may really be identical with Germanic Hartewic and Galvanus (Galgano, Gualguano, Walwanus, Walquano, Vol- vanus) with either Galganus (name of an Italian saint) or Galbanus (the Roman cognomen). Faral refers, p. 397, note 1, to studies of the question of these names undertaken by a student of F. Lot's and by Professor Bedier. As far as I am aware, however, these studies have not yet got into print. Faral's criticism (pp. 393 ff.) of Rajna's articles strikes me as the 14 Evolution of Arthurian Romance Less open to dispute are the Arthurian names which are at- tached to certain bas-reliefs over the northeast portal (Porta della Pescheria) of the Cathedral of Modena,30 representing the siege of best that has appeared. Obviously, as he observes, Arthurian names in Italian documents of later date than 1200 have no significance, since by that time the French romances had made these names familiar all over Western Europe. In the above-mentioned articles Rajna cites many such names — Galasso (== Galaad), Yvannus (= Yvain), etc., — for this later period. Besides, Artusius and Galvanus (with variants), he cites only Seldina (= Iselt, Iseut), from the year, 1180, for the period before 1200. As Faral remarks, the paucity of the names from the earlier period which Rajna conjectures to be Arthurian is in striking contrast to the abundance of indisputably Arthurian names for the later period. In the Archie fur das Studium der neueren Sprachen, CXLI, 235 (1921), F. Liebermann has noted the occurrence of Arthur as the name of a man born, at the latest, about 1140 — apparently, in Sussex. But he may have had Welsh or Breton connections of some sort or other. 30 Wendelin Foerster was the first person to bring these figures into Arthurian discussion, viz. in his article, "Ein neues Artusdokument", Zs. f. rom. Ph., XXII, 243—8 (1898). See, too, his supplementary article, "Das neue Artusdokument", ibid., pp. 526 — 529 (1898). The archivolt and its figures, however, had been the subject of repeated dis- cussion before this, in works on Italian architecture. Cp. ibid., p. 526. Foerster accepted the usual dating of the bas-reliefs (first decades of the twelfth century). He thought the situation represented on the archi- volt was not derived from any extant Arthurian romance and observes that it is nearer to the episode of Carados and the Dolerose Tor in the prose Lancelot (cp. Sommer's Vulgate Version of the Arthurian Romances, IV, 90 ff.) than to anything else in the surviving romances and was possibly derived from some lost romance, which was the source of that episode. Foerster makes no effort to reconcile the evidence of these bas-reliefs with his customary theory that Chretien invented the genre of the Arthurian romance. Foerster's reading of the names attached to the respective figures and his identifications of the same, where offered, are probably right. They are as follows (p. 244): 1. Winlogee = Guinloie or Guenloie (which occurs in two romances, viz. Li Chevaliers as deus espees and Yder), 2. Mar doc (= Madoc, or, as I believe, more likely, the common Arthurian Mariadok, etc.). 3. Burmaitus, or Burmaltus (cp. p. 526), not known elsewhere. 4. Isdernus, probably = Yder (name of four different Arthurian knights). 5. Car r ado = Caradoc (the giant, rather than Briebraz). 6. Galvaginus, probably = Gawain. 7. Galvarium Early Traditions Concerning Arthur 15 a castle by Arthur and bis followers, with knights issuing there- from to attack the besiegers. These bas-reliefs have been gene- rally dated early in the twelfth century and they include labelled images of Arthur, Kay, Caradoc, etc., which prove, if the above dating is correct, that stories concerning Arthur and the cha- racters associated with him were current in Northern Italy in that period. The evidence of these figures, however, stands in such absolute isolation for so early a period that one is disposed to await further investigation of their date, before accepting finally the dating which has prevailed among authorities on the archi- tectural history of Lombardy up to the present time.31 For the (unknown). Foerster observes that this name seems Norman, whereas G. Paris, Romania, XX VIII, 145 (1899), was inclined to connect it with Galuron in the Chronique de Nantes. 8. Che = Kai (Kay). — The castle is at the top of the archivolt, in the centre, and nos. 1 and 2 are inside its walls. On the left hand, no. 3 comes forth to attack Arthur, no. 4, and a third (unnamed) knight. On the right, no. 5 rides out against nos. 6, 7 and 8. Pictures of the figures which we have been discussing will be found in many publications that deal with the history of Italian archi- tecture or the Modena cathedral, more specifically, e. g. M. G. Zimmer- mann, Oberitalische Plastik imfruhen und hohen Mittelalter (Leipzig, 1897), Plate 18; A. Venturi, Storia dell' Arte Italiana, III (Milan, 1904), Plate 138: Giulio Bertoni, Atlante Storico-Paleografico del Duomo di Modena, Plate X (Modena, 1909); A. K. Porter, Lombard Architecture, Plate 144, (3 vols. New Haven and London, 1917). Among discussions of the figures, cp. B. Colfi, "Di una recente interpretazione [i. e. Foerster's] data alle sculture dell' archivolto nella porta setten- trionale del duomo di Modena", Atti e Memorie delta R. Deputazione di Storia Patria per le Provincie Modenesi, Serie IV, vol. IX (1899) — also, Zimmermann, p. 44, Venturi, III; 160ff., and Porter, I, 436 f., IH, 44 f. 81 Faral, op. cit., p. 395, dates these bas-reliefs about 1200. Pos- sibly, he may have authority for this dating, but the writers whose works I have cited above place them in the first half of the twelfth century, and Foerster (pp. 244 f.), on the basis of the costumes, gives the same date as the latter. Venturi, III, 160, remarks: "Si e molto discusso sull' eta della porta della Pescheria, che pure mostra ad evi- denza la mano di Niccolo, cooperatore di Wiligelmo, di lui piu giovane e meno arcaico etc." Wiligelmo, it should be remarked, worked on the cathedral at the beginning of the twelfth century. Porter, III, 48, says : 16 Evolution of Arthurian Romance names attached to them suggest an Arthurian romance from the period of the fullest development of the genre, viz., the late twelfth or early thirteenth century,32 and it accords with this dating that the architecture of the same portal contains figures from the "Between 1099 and 1106 were erected, of the existing cathedral, the crypt and also part of the facade with its sculptures and the Porta della Pescheria, but the crypt and the Porta della Pescheria were sub- sequently very materially altered." Later on, he again speaks of there being "abundant contemporary evidence that alterations and additions to the church were made constantly during the XIII and XTV centuries", but he evidently does not include the Arthurian figures among these later additions, for in his interesting section, I, 434 — 438, on "Secular Subjects" (including romances) in Lombard church architecture, he dates (p. 436) these Arthurian figures, "1099 — 1106". On the other hand, quite recently, G. Bertoni, who has made a special study of these figures, has expressed the opinion, Zs. f. rom. Ph., XL, 363 (1920) that they cannot be later than the second half of the twelfth century and more probably belong to the middle of the century. Both Venturi, III, 167 and Porter, I, 436 f., emphasize the resem- blance of these figures to those which are found on the southwestern side-portal of S. Nicolo at Bari (cp. Venturi, Plate 140, at m, 163): "From a central structure issue on foot four men, two on either side; against them come a series of men on horseback" (Porter, I, 436). Ana- logous, too, are figures at Bobbio, and Porter, loc. cit., thinks that ex- cavation of the pavement there will throw light on the Modena archivolt. 82 Guenloie, with whom Foerster identifies Winlogee, first occurs in Yder, among the extant romances, and that is dated by its editor in the second decade of the thirteenth century. Cp. H. Gelzer, Der altfranzosische Yderroman, p. LXXIX (Dresden, 1913;. To be sure, it is probably based on an older romance, but the character of the story shows that even this last source did not go beyond the latter part of the thirteenth century. If Mardoc is the same as Meriadoc, the same applies here. Cp. the Chevaliers as deus espees, whose hero is named Meriadeus (= Meriadoc), and the Latin romance Historia Meriadoci, edited by J. D. Bruce, Historia Meriadoci and De Ortu Waluuanii, Gottingen and Baltimore, (1913). Both, probably, belong to the second quarter of the thirteenth century. Mariadok is the name of Marc's seneschal in Thomas' Tristan (third quarter of the twelfth century), but the character is less prominent than would seem to be required by the Modena Mardoc. Besides, the latter evidently has nothing to do with the Tristan legend. Early Traditions Concerning Arthur 17 French beast-epic which flourished in that period.33 Even if the early dating, however, is correct, here, as in the previous case, there is always the possibility to be reckoned with that these Arthurian names may have been derived from oral tradition through Bretons or Normans.34 Leaving now these scanty indications of what may possibly be merely precipitates, so to speak, of tales transmitted orally, we come to the later written records of Arthur that have been pre- served. Whatever interpretation we may put upon the above- mentioned indications, unquestionably the matiere de Bretagne™ entered upor a new phase of influence with the publication of 83 Lucien Foulet, Le Roman de Renard, p. 60 (Paris, 1914) assigns to 1152 Nivard's Latin poem Ysengrhnus, which he thinks started the whole French beast-epic. He takes (p. 115) 1165 as the earliest possible date for any part of the Roman de Renard, but he assigns all branches but the second to the period after 1170. The sculptors of the figures may themselves have been from Normandy or Brittany, for such workmen in the Middle Ages often found employment in foreign countries. In any event, the immediate source was French. This is proved by the French forms of some of the names, viz., 1. Artus (Artus de Bretania in the inscription), not Artur or Arturus. 2. Galvaginus (formed on Galvagin). 3. Che (= French Ke), not Caius (Geoffrey of Monmouth). 4. Winlogee, not Winlogea, as the Latin form would be. Cp. Foerster, op. cit, p. 248. 5. Mar doc, without Latin -us. It must be confessed, however, that Isdernus is closer to the Welsh Edeyrn (Edern) than to the French Ider (Idler, Yder) in our MSS. of Erec, Yder, etc. Cp. too, Geoffrey's Hiderus. (The first s is, no doubt, due to the sculptor's blunder). Foerster, ibid., plausibly surmises that Carrado stands for Carradocus — only limitations of space, as an in- spection of the archivolt shows, prevented the sculptor from completing the name. 85 The term, now in general use for Arthurian traditions, is taken from the famous lines of Jean Bodel, La Chanson des Saxons, 11. 6ff.: "Ne sont que. iii. matieres a nul home antandanti De France et de Bretaigne et de Rome la grant; Et de ces. iij. matieres n'i a nule samblant. Li conte de Bretaigne sont si vain et plaisant; Cil de Rome sont sage et de san aprenant; Cil de France de voir chascun jor apparant." The poem was written in the middle of the thirteenth century. 18 Evolution of Arthurian Romance Geoffrey of Monmouth's Historia Regum Britanniae about the year 1137.36 In the dedication to this work, the author remarks 39 The date of Geoffrey's work, as was pointed out by Sir Frederick Madden in his article, "The Historia Britonum of Geoffrey of Mon- mouth", The ArchaeologicalJournal, XV, 299ff. (1858), is determined by the double dedication which is preserved in the Bern MS. of the Historia. This MS., which dates from about 1160, still remains un- published. Unlike the other MSS., it contains a double dedication to King- Stephen and Robert of Gloucester. King Stephen is here extolled in almost the same words that we find applied to Robert of Gloucester in the other MSS. Robert, himself, is addressed as altera regni nostri columna and is eulogized at greater length in terms of the most exalted flattery. The text of this important preface is given in full in W. L. Jones' above-mentioned paper. Cp. ibid., p. 65, note. Now, as Madden remarks, this dedication, if genuine — and there is no reason to doubt that it is so — must have been written between April, 1136 and May 1138, for only during this period were Stephen and Robert of Gloucester on friendly terms. From the last-named date down to his death in 1147 Robert was the principal supporter of the cause of Queen Matilda in her strife with Stephen, and it is, of course, out of the question that under these circumstances Geoffrey should have coupled his name with that of the king in a joint dedication. The fact that Henry of Huntingdon, writing in January, 1139, was able to give a summary of Geoffrey's Historia in his letter to Warinus gives us that year — or better, 1138 — as the terminus ad quern for the work. Con- firmatory of Madden's dating of the Historia is A. Leitzmann's obser- vation, that Geoffrey, at the end of his dedication, refers to Henry I as dead, the "alteram Henricum*' of that passage being really that monarch (who died on Dec. 1, 1135), and not Henry H, as has often been thought. See Leitzmann's "Bemerkungen zu Galfrid von Mon- mouth", Archiv filr das Studium der neueren Spraehen, vol. 134, pp. 373ff. (1916). The Latinity of the Bern MS. is less polished than that of the other MSS., but, according to Professor Jones, who has collated it with San Marte's edition (Halle, 1854), it does not differ in any essential from the current text. Owing to differences in Henry of Huntingdon's summary, the idea formerly prevailed that there was a (lost) earlier recension of Geoffrey's Historia. R. H. Fletcher, however, in his "Two notes on the Historia Begum Britanniae of Geoffrey of Monmouth," PMLA (Publications of the Modern Language Association of America), XVI, 46 Iff. (1901), has shown that this is incorrect. For the text of Henry's Letter to Warinus see the edition of Robert de Torigni's chronicle by Leopold Delisle, I, 97ff. (2 vols. Rouen, 1872—8). Early Traditions Concerning Arthur 19 that he had often marvelled at the little that had been recorded of the kings that had dwelt in Britain before the Incarnation of Christ and even concerning Arthur and his successors who lived after the Incarnation. "Now whilst I was thus thinking upon such matters, Walter, Archdeacon of Oxford, a man learned not only in the art of eloquence, but in the histories of foreign lands, offered me a certain most ancient book in the British language (quendam Britannici sermonis librum vetustissimum) that did set forth the doings of them all in due succession and order from Brutus, the first king of the Britons, onward to Cadwallader, the son of Cad- wallo, all told in stories of exceeding beauty." Geoffrey's own work purports to be his translation of this ancient book. It is divided into twelve books, of which Book VII is really an earlier work of the same author, — viz. the pretended prophecies of Merlin and it introduces into European literature a host of new characters such as Lear, Cymbeline, Locrine, etc., whom the poets of later ages were destined to render famous. Here in the Historia (Book VIII, ch. 14 to Book XI, ch. 2), then, for the first time, we have related in pseudo-historical form, the glories of Uther Pendragon and Arthur, his son, the latter's birth and fic- titious conquests extending as far as Rome, on the one hand, and the Baltic, on the other, until his downfall in his last battle with Mordred and his translation to Avalon. Merlin's marvels of enchantment and prophecy, Gawain's valorous deeds, and Guine- vere's (Guanhamara's) marital disloyalty all form, likewise, a part of the story. The author of this book, though born in Wales, was probably the son of a Breton — a member of the Norman garri- son stationed at Monmouth.37 Between the years 1129 and 1151 he was living in the neighborhood of Oxford — probably as a canon of St. George's in the Castle of Oxford.38 Some time sub- A man named Geoffrey at this time would hardly be a Welshman. Cp. J. E. Lloyd's History of Wales down to the Edwardian Con- quest, II, 524 (2 vols. Oxford, 1911). Geoffrey signed himself Gal- fridus Artur — so his father was named Arthur. For Geoffrey's connection with Oxford, cp. the documents in which he figures as a witness (some of them new) discussed by H. Salter, "Geoffrey of Monmouth and Oxford," English Historical Review, XXXIV, 382 ff. (1919). The best summary of all that was previously known 20 Evolution of Arthurian Romance sequent to 1140 he was appointed an archdeaconry in the dio- cese of Llandaff, and on Feb. 24, 1152, after having been or- dained priest only eight days before, he was made bishop of St. Asaph in his native land, but died two years later, before he had actually assumed the duties of his office. To few works in the history of literature can the much-abused term, " epoch-making, " be so justly applied as to Geoffrey's Historia. Under any suppo- sition, it was indubitably the most notable production in the Ar- thurian field that had appeared up to that date, and, in all pro- bability, it was owing to the influence of this book, direct and indirect, that the Arthurian stories leapt into general literary popularity just at this time. The conception of Arthur as a great mediaeval monarch, the ideal representative of chivalry — not a merely fairy-tale king — ■ originated, we may say, entirely with Geoffrey of Monmouth . He succeeded in embodying this idea in his work in a truly imposing literary form, and the pretended historical character of the Historia gave a dignity to the theme which it had not hitherto possessed. We need not take very seri- ously the author's declaration in his dedication to Robert, Earl of Gloucester, that he was translating into Latin an old book in the British language furnished him by his friend Walter, Arch- deacon of Oxford. There is no evidence that he had a more than superficial acquaintance with Welsh39 and probably he knew Bre- ton no better. As a matter of fact, in addition to some borrowings from oral traditions, not exclusively of the Celts,40 he drew ma- concerning Geoffrey of Monmouth is W. Lewis Jones' paper on him in the Transactions of the Honourable Society of Cymmrodorion for the session, 1898—99, pp. 52 ff. Cp. too, H. L. D. Ward's Catalogue of Romances in the Department of MSS. in the British Museum, I, 203 ff. (1883) and E. Windisch, Das Keltische Britannien bis zu Kaiser Artur, pp. 123ff. (Leipzig, 1912). 39 Cp. Ward's Catalogue of Romances in the Department of MSS. in the British Museum, I, 205. 40 In his "Bretonische Elemente in der Arthursage des Gottfried von Monmouth", Zs. f frz. Spr. u. Lit, XII1, 231ff. (1890), H. Zimmer endeavors to prove from the forms of certain Arthurian names that Geoffrey used Armorican sources as well as Welsh — more specifically, that he derived such materials from men who were Bretons by race, hut who spoke French exclusively, or, at least, were bilingual. The Early Traditions Concerning Arthur 21 terials from Gildas, Nennius, Bede, Livy and others — in fine, from any quarter he chose, added to these materials from his own most important names are Eventus (= Yvain), Walgainus, Walguainus (= Gawain, wha in Welsh is called Gwalchmei), Caliburnus (= Arthur's sword, Excalibur, which in Old French is called Calibor(e), Escalibor(e), etc.). In the same journal, XIII1, 41, note 1, (1891), he adds Ritho to the list of such names. F. Lot, however, has pointed out, Romania, XXV, Iff. (1896), that Eweint, of which Eventus is the Latinized form, occurs as early as the tenth century in Welsh genealogies and that Caliburnus is as difficult to explain on the hypothesis of Armorican origin as on that of Welsh origin. On the other hand, he acknowledges that Walgainus is based rather on French Gauvain (Gaugain) than on Welsh Gwalchmei or Breton Walchmoe. In that case, however, I may remark, Geoffrey could have derived it as well from a Frenchman or from a French book as from a Breton. Lot, moreover, cites the form Walven for Gawain in William of Malmesbury's description (De gestis Begum Angliae, III, under the date 1086) of Gawain's tomb in the Welsh province Ros. Here it is stated that Gawain was ruler over Waluuithia (= Galloway), but was expelled from his kingdom by the brother and nephew of Hengist. Walven, however, probably rests on French Walwains, and Lot, loc. cit., conjectures that the tomb in question was really that of Maelgwn, and that the new identification was suggested by the name of a neighboring district, viz., Castell Gwalchmai (= Gawain's Castle). In any event, as Brugger, Zs. f. frz. Spr. u. Lit, XXXHI2, 59 f. (including p. 60, note 3) (1908), remarks, the connection of Gawain with Walweitha (Walwithia) is, no doubt, "eine gelehrte Erfindung", and, although the Welsh called Geoffrey's Walgainus (French Gauvain) by the name of Gwalchmei, it still remains uncertain whether the former is of Welsh origin, although this is a priori probable. In "Arthurian Notes", MLN, XVII, cols. 277 f. (1902) W. W. Newell suggests that Gawain's name (cp. Walwen) is formed from Walweitha, just as Geoffrey formed Locrin from Loegria and that Geoffrey represented him as the son of Lot (Loth), because of the usual association of Lothian and Galloway. This theory, however, seems hardly admissible. There is, of course, nothing surprising in Geoffrey's use of French sources. Apart from the Arthurian materials under discussion, Lot, loc. cit, points out his use of such sources in the episodes of Gormond (XI, 8—10, XII, 2) and Mont St. Michel (X, 3). It is evident, too, from the same scholar's discussion, Romania, XXX, llff. (1901) that among Geoffrey's Celtic sources some were Cornish. The names of the Cornish dukes, Cador and Gorlois, prove this. So, too, with the name Modredus; Cp. Lot, Romania, XXV, 2 and Loth's Mabinogion*, II, 22 Evolution of Arthurian Romance imagination — the most important element of all — and moulded the whole into the pseudo-historical wort which we know.41 In- deed, it is hardly open to question that Geoffrey's narrative re- flects, in many particulars, even actual historical events of the eleventh and twelfth centuries42 — more especially, events of Nor- 238, note 1. Zimmer argues, Zs. f.frz. Spr. u. Lit., XII1, 254 f. that it might be also a Breton form of the twelfth or thirteenth century, but that is too late for Geoffrey's source. San Marte (edition of Geoffrey's Historia, pp. LXXIIff.), like some other students of his time and of earlier times, believed that the Welsh Brut Tysilio was one of the sources of Geoffrey's Historia. But the reverse is undoubtedly true, as has been shown by F. Zarncke and B. Ten Brink in Ebert's Jahrbuch fur englische und romanische Philologie. Cp. the former's article, "Uber das Verhaltnis des Brut y Tysilio zu Galfrid's Hist. Reg. Brit.", V, 249ff. (1863), and the latter's "Wace und Galfrid von Monmouth", IX, 241 ff. (1868). For some additional points see, too, G. Heeger: Uber die Trojanersage der Britten, 79f. (Munich, 1886). Altogether, I agree with Brugger, Zs. f. frz. Spr. u. Lit, XXXII2, 127 (1908), that Geoffrey borrowed names from Welsh tradition, but little else. There are, doubtless, some exceptions, like the story of King Leir (Lear) and his ungrateful daughters (II, 11 — 15). This story, too, has analogues, it is true, outside of Celtic territory. Cp. J. Bolte and G. Polivka, Anmerkungen zu den Kinder- und Hausmarchen der Briider Grimm, II, 47 ff. (3 vols. Leipzig, 1913—1915), and Miss C. A. Harper, "A King Lear Analogue"', The Nation, (New York) for Feb. 10, 1916 (Correspondence Supplement, p. 17). The first is a German, the second an Indian folktale. The German is closer to Geof- frey. In the Indian tale — entitled The Hireling Husband and printed by Shovona Devi in her collection, The Orient Pearls (London, 1815) — the choleric old king has seven sons and the youngest fills the role of Cordeilla (Cordelia). E. G. Cox's "King Lear in Celtic Tradition", MLN, XXIV, Iff. (1909), gives no parallels to this immortal story. On the legend in general, cp. W. Perrett, The Story of King Lear: Palaestra, no. 35 (Berlin, 1904). On Geoffrey's sources, in general, see R. H. Fletcher, Arthurian Material in the Chronicles, ch. 3, and PMLA, XVI, 469ff. (1901). His debt to Virgil has been investigated by H. Tausendfreund, Vergil und Gottfried von Monmouth, Halle, 1913, and his debt to the Old Testament by Paul Feuerherd, Geoffrey of Monmouth und das Alte Testament, ibid., 1915. So Zimmer, Gott. G. A. for Oct. 1890, pp. 824ff. According 42 Early Traditions Concerning Arthur 23 man history — of the reigns of William the Conqueror and his successors. Thus he gives Arthur in the beginning of the latter's career an imaginary King Hoel of Brittany with 15,000 Bretons as allies, no doubt, because a large number of Breton troops aided William the Conqueror in his invasion of England — these troops being led by the nephews of Count Hoel, who was then ruler of Brittany. Similarly, the way in which Arthur parcels out Gaul among his followers was doubtless suggested by the Conqueror's distribution of English lands among his companions in arms. The prominence that Winchester and Carlisle have in the Arthurian narratives of Geoffrey is also plainly due to the importance which these cities possessed in Geoffrey's own time. The former, though no longer the capital of England, as in Anglo-Saxon days, was often the residence of the Norman kings, and Carlisle under Wil- liam Rufus had become one of the strong places of the kingdom. In writing his history, Geoffrey's aim was primarily to exalt his own race. Nothing is more striking about the Normans than the interest which they took in their past and a whole crop of chronicles was the result of this interest. Geoffrey opposes now to these Norman chronicles a chronicle of the Celts with Arthur as the great hero. In an age when even such a man as Lanfranc, the great Norman Archbishop of Canterbury, could connive at forgery it is not surprising that a Celtic ecclesiastic, with the lively fancy of his race, should palm off a hoax like the Historia on the public of his time. That public was, in the main, uncriti- cal and swallowed this pretended history as it swallowed the Con- stitutions of Cons tan tine, which were invented to justify the tem- poral claims of the papacy.43 Moreover, it touched the vanity of the to Fletcher (second citation in the previous note), the kernel of the Belinus and Brennius episode of the Historia, III, 1 — 10, was supplied by the feud between Harold, the last Anglo-Saxon king, and his brother, Tostig. Although most mediaeval readers accepted Geoffrey's Historia as authentic, there were writers already in the latter part of the twelfth century who voiced their scepticism — so William of Newburgh and Giraldus Cambrensis, especially, in passages that have been often quoted. See the editions of their respective works in the Rolls Series, Vol. I, of the Chronicles of Stephen, etc. (1884), William's Prooemium, and 24 Evolution of Arthurian Romance Anglo-Norman nobility, who were now identified with Great Bri- tain, that they could claim a hero who was the equal, if not the su- perior, of Charlemagne, the great hero of their Continental kins- men. The consequence of all this was that Geoffrey's Historia had an instantaneous and prodigious success and it stimulated im- mensely whatever interest there may already have been in Arthur and his companions. The first sign of Geoffrey's influence on the vernacular lite- rature appears in the metrical chronicles. Between 1147 and 1151 Geffrei Gaimar had used Geoffrey of Monmouth in compiling a chronicle of Great Britain in French verse.44 His own statement is that he followed "the good book of Oxford which belonged to Walter the Archdeacon", but it is safe to say that this really means Geoffrey's Historia. The work was undertaken at the in- stance of Constance, wife of Ralph Fitz-Gilbert, an Anglo-Norman noble to whom Gaimar was chaplain. Women rarely understood Latin in the Middle Ages, and so just as vernacular versions of Biblical books were often composed for their benefit in that period, so too with the chronicles. Giraldus' Opera, VI, 57 f. (1868) — also, the summaries of these passages in Fletcher, op. cit, pp. 101 f. and 180, respectively. Among modern scholars who have expressed a belief in the reality of Geoffrey's liber are H. L. D. Ward, "Postscript to the article upon Geoffrey in the Catalogue of Romances, Vol. I (1883)," Anglia, XXIV, 383ff. (1901), E. Windisch, Das Keltische Brittanien bis zu Kaiser Arthur, pp. 126ff., (1912), and A. Leitzmann, op. cit, pp. 3751 (1916). Ward conjectures that the liber consisted of an Old-Welsh MS. containing many British genealogies and several historical glosses, brought home from Brittany by Archdeacon Walter. Windisch and Leitzmann argue that Geoffrey would not have dared to pass off on Robert of Gloucester — the most powerful man in the kingdom — so audacious an invention. But the examples cited above show to what length mediaeval audacity might go in such matters, and, besides, Robert of Gloucester, though cognizant of the fraud, may not have taken the matter seriously, or may have even accepted the Historia gladly as a sort of national epos. As such it has been interpreted with some plausibility by Sebastian Evans in his translation of the Historia, pp. 357 ff. (London, 1904). 44 UEstorie oles Engles, edited by T. D. Hardy and C. T. Martin for the Rolls Series, 2 vols., (London, 1889), is all that survives of it. Early Traditions Concerning Arthur 25 The MSS. of Gaimar's chronicle preserve only the later por- tion of the work, — L'Estorie des Engles, as it is called. To be sure, the so-called Munich Brut,*5 which appears to be of the same date, is by some eminent scholars supposed to be a fragment of the earlier portion. Even this fragment, however, ends before it reaches the reign of Arthur. The only part of the extant por- tion of Gaimar's Chronicle, then, that contains any mention of Arthur is the story of Havelock which he has inserted in his poem. The Arthurian allusions here accord substantially with Geoffrey of Monmouth. Of far greater importance, however, was the paraphrase com- pleted by Wace in 1155 — the poem which is commonly called the Roman de Brut, or simply, Brut** although the author's own name for it was Geste des Bretons. There is no dedication of it preserved in the MSS., but the English poet, Layamon (11. 42 ff.), tells us that Wace "gave it to the noble Eleanor, who was the 46 Edited by K. Hoffmann and K. Vollmbller, Halle, 1877. There is still another fragmentary French version of Geoffrey of Monmouth in monorhymed laisses preserved in M. S. Harley, 1605 (British Museum). Cp. Ward, Catalogue of Romances, I, 272 ff. and Otto Wendeburg, Ueber die Bearbeitung von Gottfried von Monmouths Historia Begum Britanniae in der HS. Brit. Mus. Harl. 1605. Braunschweig, 1881 (Erlanger Diss.). H. Suchier, LB, HI, 107f. (1882), pronounces the dialect Picard and the date not much later than Wace. The fragments have no especial interest. The condition of this version and of the Munich Brut (both preserved in fragmentary and unique MSS.) shows, however, how easily the French original of Layamon could be lost. 46 Edited by Le Roux de Lincy, 2 vols., Rouen, 1836—1838. F. M. Warden, MLN, XTV, 95, points out that the numbering in this edition is wrong — 130 lines too high. A new edition of this work based on all the MSS. is a great desideratum. Brut is the French form of Brutus, name of the first king of Britain, according to Geof- frey. For Wace's life, cp. especially, G. Paris, Romania, IX, 592 ff. (1880). He was born in the island of Jersey about 1100 — was a canon of Bayeux in the latter part of his life. He is also well cha- racterized by G. Paris in his MUanges, I, 85 f. For a full bibliography of Wace, cp. Annette B. Hopkins, The Influence of Wace on the Arthurian Romances of Crestien de Troies, University of Chicago dissertation, 1913, p. 10, note 24 a. 26 Evolution of Arthurian Romance high King Henry's queen" — that is to say, the wife of Henry II of England, who ascended the throne in 1154. A detailed study of Wace's relation to Geoffrey47 proves that he followed his original closely, as far as the facts are concerned.48 Allusions which we shall consider later on in another connection show that he knew of an oral tradition of Arthur outside of Geoffrey, but he adds practically no new material to that which his original offered and so the advantage which his work has over Geoffrey's is merely in its superior vivacity and vividness. A French poem in octosyllabic couplets was likely to be livelier than a pseudo- chronicle in the artificial Latin prose of the Middle Ages, and Wace's style, as it happens, is particularly lively, so that Geoffrey's legends and fictions are now cast in the form which was really appropriate to them — namely, that of a metrical romance.49 He substitutes direct for indirect discourse, is fond of rhetorical questions and exclamations, and amplifies, especially, the des- criptions of battles and festivities. Even in respect to style, then, Wace is an important factor in the development of the Arthurian romances. He accordingly contributed largely to the spread of interest in the Arthurian stories, for in the Middle Ages, as well as now, more people could enjoy a tale in the vernacular than in Latin. But he influenced Arthurian literature in another way, also — that is to say, the genre of the Brut — 50 the chronicle of British history 47 Cp. Fletcher, Arthurian Material in the Chronicles, pp. 127ff., Alfred Ulbrich's, "Ueber das Verhaltnis von Wace's Roman de Brut zu seiner Quelle, der Historia regum Britanniae des Gottfried von Mon- mouth", Romanische Forschungen, XXVI, 181 ff. (Erlangen, 1908), and Leo Waldner, Wace's Brut und seine Quellen. Jena diss. (Karls- ruhe, 1914). 48 For the main additions made by Wace, cp. Waldner, pp. 120ff. 49 Substantially true is the remark of the Abbe de la Rue, who calls Wace's Brut "incontestablement le premier Roman de la Table Ronde." See his Essais historiques sur les bardes, les jongleurs et les trouveres normands et anglo-normands, etc., I, 50 (3 vols., Caen, 1834). M. Wilmotte, Le Moyen Age, XXVII, lOOf. (1914), makes the same observation. 60 The term, Brut, in the sense of chronicle, was also adopted from the French by the Welsh, and so we have in Welsh the Brut Tysilio, Gwentian Brut, etc. Early Traditions Concerning Arthur 27 which follows the Geoff reyan tradition and makes a fabulous Bru- tus the founder of the kingdom — did not end with him. Curi- ously enough, the most important representatives of this genre actually preserved are not in French, but in English; neverthe- less, the English poems undoubtedly are based on (lost) French originals. The poems in question are, 1. Layamon's paraphrase of Wace, which it is common to call, after Sir Frederick Madden 's example, Layamon's Brut — 51 a poem which was finished not long before 1205. 2. The alliterative Morte Arthure52 of the four- teenth century, which has often been ascribed — 53 wrongly, no doubt — to the Scotch poet, Huchown of the Awle Ryale. In both of these poems the principal interest is pseudo-historical. Indeed, the narrative of the second is almost wholly taken up with the continental wars of Arthur and his final conflict with Mord- red — that is, selected episodes of some lost Brut. The spirit is, therefore, Considerably nearer that of the chansons de geste or poems of the Charlemagne cycle than is the case with the Arthu- rian romances proper. Indeed, the encounter of Gawain and the Saracen prince, Priamus, together with the latter's conversion to Christianity in the Morte Arthure, is actually drawn from Fiera- bras,5i one of the most famous of the chansons de geste. More- 61 The only edition is Madden's, in 3 vols. (London, 1847), for the Society of Antiquaries of London. A complete bibliography of Layamon down to 1906 was published by B. S. Monroe, Journal oj English and Germanic Philology, VII, no. 1, pp. 139ff. (1908). For the critical literature since then, cp. J. E. Wells, A Manual of the Writings in Middle English, 1050 — 1400 (Yale University Press, 1916), and the supplements thereto. Layamon's terminus a quo is 1189, since he refers I, 3, to Henry II as dead; his terminus ad quern is 1205, date of the crisis over the collection of Peter's pence, which, contrary to Madden, seems in the allusion to this tax, III, 286, unknown to him. 62 The latest and best edition is Erik Bjorkman's (Heidelberg and New York, 1915), in Morsbach and Holthausen's Alt- und mittel- englische Texte, no. 9. The Introduction contains a full bibliography. 63 Cp., for example, G. Neilson, Huchown of the Aivle Ilyale, (Glasgow, 1902). 64 Cp. P. H. Griffith, "Malory, Morte Arthure, and Fierabras"', Anglia, XXXII, 389 ff. (1909). 28 Evolution of Arthurian Romance over, the parts played in the story by Lancelot, Gawain, and others show plainly the influence of the fully developed Arthurian ro- mances; and there is nothing surprising in this, seeing that the alliterative poem belongs to the fourteenth century. But in its essentials the Morte Arthure is based on a lost Brut.55 Of far greater importance, however, in this connection is Layamon's Brut, which is a specimen of a metrical chronicle pre- served in its entirety. At the beginning of his poem Layamon gives as his sources "the English book that Saint Bede made, another, in Latin, that Saint Albin made and the fair Austin who brought baptism in hither" (i. e. St. Augustine, the missionary), and, lastly, Wace. It is of little importance that he has cited here other books besides Wace. His poem does not show any use of such books and there can be little doubt that he took over me- chanically from his French original this citation of authorities.56 With regard to the question of his relations to Wace, Layamon's poem is about double the length of Wace's: the one contains so- mething upwards of 15000 lines, the other something upwards of 32000.57 Until quite recently it has been the common assump- 55 On its sources see P. Branscheid, "Ueber die Quellen des stab- reimenden Morte Arthure", Angjia, Anzeiger, VIII, 179ff. (1885) and, especially, R. Imelmann's treatise, named in the next note. Branscheid's idea that Layamon is a source of the poem is certainly wrong. fie In his treatise, Layamon, Versuch ilber seine Quellen (Berlin, 1906), p. 17, R. Imelmann has pointed out that this coupling of Wace and Bede as authorities is found in other chronicles which relate the history of Great Britain under Anglo-Saxon as well as Celtic rule. The addition of Albin the abbot, who died at Canterbury in 732, and St. Augustine to the list of authorities was also, no doubt, already in Layamon's original. And this addition is not hard to explain — for Bede incorporates in his work the long interrogatories which St. Augustine addressed to Pope Gregory asking for instructions as to how he should proceed in his mission of converting the Anglo-Saxons, and he still further speaks of Abbot Albin in his Preface as auctor ante omnes atque adjutor opusculi hujus. So Albin and Augustine were to a certain degree part-authors of his Ecclesiastical History. It was an easy blunder, of which the Middle Ages furnishes many examples, to make them now the authors of a separate work, although Augustine died, of course, before Albin was born. Wace's poem is written in octosyllabic couplets — which became Early Traditions Concerning Arthur 29 tion that the expansions of the English poem were due to Laya- mon himself and that its author, who was a priest at Arley-Regis (modern Ernley) on the Severn in Worcestershire, not far from the Welsh border, had derived the most remarkable of these ex- pansions from Welsh oral tradition.58 In the light of recent re- search, however, it is no longer open to doubt that this assumption was mistaken and that Layamon was merely following an ex- panded (French) version of Wace, now lost.59 As long as the the regular form of the French Arthurian romances; Layamon's in allite- rative verse, but in a somewhat disintegrated form and with frequent admixture of rhyme. ft8 Such was the view enunciated by Sir Frederick Madden in editing Layamon's poem and it has found especial advocates, also, in R. Wuelcker, "Ueber die Quellen Layamon's", Paul and Braune's Bei- trage, III, 524ff. (1876) and in A. C. L. Brown, "Welsh Traditions in Layamon's Brut", MPh., I, 95 ff. (1903). 69 R. Imelmann, in the above-mentioned treatise, was the first scholar to propose this view. See, too, J. D. Bruce, "Some proper names in Layamon's Brut, not represented in Wace or Geoffrey of Monmouth", MLN, XXVI, 65 ff. (1911). According to Imelmann, Laya- mon's original was an enlargement of Wace by combination with material drawn from Gaimar. He imagines, moreover, that the text which resulted from this combination still further underwent the in- fluence of the Vulgate Mort Artu. But the latter was certainly later than Layamon's source — later, indeed, probably than Layamon. It is to be observed that other poems of Wace's underwent expansion in the way that we are here assuming for his Brut, viz. his Conception and Assumption. For a version of the former, expanded by the inter- polation of passages from another French poem, Uhistoire de Marie et de Jesus, cp. P. Meyer, Romania, XVI, 232ff. (1887). According to MS. 749 (fol. 132) of the Bibl. Nat., Robert de Boron in his Merlin, refers to a Brut translated from the Latin by a certain Martin of Rochester, who is otherwise unknown. The passage, which is quoted from this MS. by P. Paris, RTR, II, 36, note, is not found in most MSS. and may be merely a scribal invention. Besides, in the MSS. where it does occur, we do not always find rouecestre (Rochester). Rouain and other names occur, instead. Cp. W. E. Mead's Merlin treatise, E. E. T. S., Original Series, no. 112, (1899), pp. CLI, CLIX, CLXI, CLXXI. Some scholars have accepted this Martin as Robert's chronicle-source, instead of Wace, e. g. Wechssler, Sage vom Heiligen Gral, p. 124 (1898), Brugger, Zs. f. frz. Spr. u. Lit, XXIX1, 60, 30 Evolution of Arthurian Romance forms of certain names in the French poet's Brut were known only through our sole printed edition of that work, which is based on an indifferent MS., it was possible to maintain that the greater similarity of these same names in Layamon to the corresponding Welsh names revealed a direct contact with Welsh Arthurian tra- ditions on the part of the latter. But investigation has proved that the better MSS. of Wace's Brut, which its editor left unused, show no such divergence between the two poets in respect to the forms of these names, and the argument from such alleged differences has consequently collapsed. As a matter of fact, the whole nomen- clature of Layamon's poem bears throughout the stamp of French origin and he was so ignorant of Welsh that he did not recognize what are often originally Welsh names in the French forms with which they were clothed in the manuscript that lay before him.60 Now the conclusion to which the unmistakably French cha- racter of the nomenclature points is borne out by other evidence. Above all, the circumstance that certain features which the Eng- lish poem has in common with the Old French prose romance, the Mort Artu, are entirely absent from the Wace of our MSS., puts beyond dispute the fact that these two works, which were nearly contemporary61 and quite independent of each other, had a common source, not identical with Wace's Brut in its original form.62 The features here in question are: 1. Arthur learns from a messenger of Mordred's treason in usurping the throne, whilst the former is engaged in a war against Lancelot on the continent, and sees in including note 8 (1905), XXX1, 182ff. (1906), XLIV2, 13ff. (1916), and elsewhere. If Martin's chronicle really existed, it may very well have been the expanded Wace. Sommer, Introduction to his Vulgate Version of the Arthurian Romances, pp. XX ff., cites a number of passages in the prose Lancelot (to be found III, 3, 46, V, 130, 117, 144, respectively, in his series) as allusions to a lost Brut, which Brugger, op. cit. XLIV3, 14, identifies with Martin's. None of these supposed allusions, however, are con- vincing. They are much more likely to be mere inventions for the nonce. 60 Cp. Bruce, op. cit., pp. 67 ff. 61 For the date of the Mort Artu, cp. Part III. 42 Bruce, RR, IV, 451 ff. (1913). Early Traditions Concerning Arthur 31 this intelligence the realization of a prophetic dream.63 2. At the end of the final battle between Arthur and Mordred only the for- mer and two of his knights are said to be left alive.64 3. Arthur is translated to Avalon by Morgan and her fairy ladies.65. 4. One of Mordred's sons who seized the kingdom after Arthur's death- is given a definite name.66 Many other names and details which distinguish Layamon from Wace are found scattered through different French works or works of French origin.67 From these various circumstances the conclusion, then, is unescapable that the version of Wace which constitutes the basis of the English poem had received here and there additions of considerable significance. Some of these additions, perhaps, should be accredited ultimately to Celtic — most likely Armorican — sources, but, in any event, Layamon was not re- sponsible for them. The matters which have just been cited prove beyond reaso- nable doubt that Layamon in many instances drew from an ex- panded Wace. In the light of these instances, it is a pretty safe inference that all other additions to the story of the English poem of any considerable extent come from this same source. There are, for example, long passages in the former where, over against the concise and matter-of-fact account of Wace, we have a full and circumstantial narrative with free introduction of speeches 63 Bruce's edition of MA (the Mori Artu), p. 202 and Madden's edition of Layamon, III, 117ff. There are some differences between the two, for a discussion of which cp. Bruce, RR, IV, 451 f. 64 MA, p. 244, and Layamon, III, 143. 65 MA, pp. 250 ff., and Layamon, HI, 144. 66 MA, p. 254, and Layamon, UI, 150. 67 Cp. Imelmann, pp. 24ff. The chief works which he cites are, 1. The prose Brut d'Angleterre (unpublished), apparently of the fifteenth century. According to Imelmann, it is the prose rendering of a lost Norman verse-chronicle, closely connected with Wace. 2. The Middle English alliterative Morte Arthure. 3. An imprinted French verse- chronicle preserved in the British Museum MS., Royal 13, A. XXI. This MS. is of the early fourteenth century. 4. The Munich Brut, which may well be a fragment of Gaimar's lost chronicle of the Britons. It was from such works, no doubt, that Layamon really derived the details noted by Fletcher in his article "Did Layamon make any use of Geoffrey's Historia?", PMLA, XVIII, 91 ff. (1903). 32 Evolution of Arthurian Romance and dialogues — in fine, a leisurely breadth of narrative art that approaches that of the French Arthurian romances. To say that an English writer of about the year 1200 was capable of this genial amplification is contrary to all the evidence. Not only is there no other example in this early time of such amplification of his originals on the part of any English author, but one may extend the assertion to the entire Middle English period. There is nothing like this, for instance, in the relations of the whole body of the Middle English romances to their French sources, for here compression, not expansion, is the rule. The amplified passages in question occur in all parts of the English poem,68 but those pertaining to Arthur possess naturally the greatest interest. Thus it is the English poem which first tells us69 of how the elves (fairies) conferred on the king their best gifts at his birth: "They enchanted the child with magic most strong; they gave him might to be the best of all knights; they gave him another thing, that he should be a rich king; they gave him the third that he should live long; they gave to him — that royal child — virtues very good, so that he was the most generous of all men alive." Still further, we have here among other novel- ties, detailed descriptions of Arthur's armor,70 an account which we shall return to later on, in another connection, of how the Round Table was founded,71 and an allegorical dream in which the mon- arch is warned of his impending destruction through Mordred. In this dream Arthur finds himself on top of a great hall from which he looks out over his dominions. Mordred comes and hews down the posts that hold up the hall and Guinevere aids the traitor in pulling down the building, so that Arthur and Gawain 68 For a list of them cp. Sir F. Madden's Preface, pp. XlVff. 69 Madden, II, 384 f. Layamon here shows a serious discrepancy as compared with Wace which seems to prove that the MS. of the expanded version of that author which he was using was defective in this place; for when Arthur is next mentioned (after Uther's death), he is — in some unexplained manner — in Britanny. See on the passage, Bruce, MLN, XXVI, 69 f., note. Other examples of discre- pancies like this in Arthurian literature, due to defective MSS., are there given. 70 Ibid., 463, 576. 71 Ibid., 531 ff. Early Traditions Concerning Arthur 33 had their arms broken in its fall. He rises, however, and slays Mordred and Guinevere. His followers had all fled and he wan- ders alone until a lion takes him into the sea, where he is rescued by a fish — but at this point he awakes. Most justly celebrated of all the amplified passages, however, is the one which gives us the first vernacular description of Ar- thur's translation to Avalon (the Celtic Elysium).72 Wace merely remarks (11. 13683ff.) of this translation that Arthur caused himself to be carried to Avalon to have his wounds cured. He is still there and the Britons expect his return, as they say, — with a few more words in the same vein. Instead of these bald state- ments, however, we have in the English poem the following highly poetical lines, which open with Arthur's speech to his successor, Constantine, after he (Arthur) has been grievously wounded in his final battle with Mordred: "I give thee here my kingdom, and defend thou my Britons ever in thy life, and maintain them all the laws that have stood in my days, and all the good laws that in Uther's days stood. And I will fare to Avalun, to the fairest of all maidens, to Argante the queeL,73 an elf most fair, and she shall make my wounds all sound, make me all whole with healing draughts. And afterwards I will come again to my king- dom and dwell among the Britons with great joy.' Even with these words there approached from the sea a little boat, floating with the waves; and two women therein wondrously formed; and they HI, 144 f. The one in Geoffrey's Latin poem, Vita Merlini, discussed below, is about 50 years earlier than Layamon. 73 Miss Paton, Fairy Mythology, pp. 26 1. has tried to show that Argante was a Celtic divinity, but we have here indisputably merely a corruption of Morgan (Morgant), name of the famous fairy queen of Arthurian romance — Arthur's sister — whom Malory calls Morgan le Fay. Cp. on the subject, Bruce, MLN, XXVI, 65 ff. and ER, III, 190 f. In these places examples are given of corruptions of the character's name that involve the loss of initial M and change of o to a — especially, in Roman de Troie, 1. 8024 (most MSS.), and in the Spanish romances, where the name becomes Urganda. This passage may have been suggested to the author of the expanded Wace by Geoffrey's poem Vita Merlini — for there, too, Morgan (Morgen) figures as the healer of Arthur's wounds — but, in my opinion, it was more probably drawn directly from Celtic tradition. 34 Evolution of Arthurian Romance took Arthur anon and bare him quickly to the boat and laid him softly down and forth they gan depart. Then was it accomplished that Merlin formerly said that there should be great care on ac- count of Arthur's departure. The Britons believe yet that he is alive and dwelleth in Avalun with the fairest of all elves; and the Britons yet expect when Arthur shall return."74 Although translating a French poem which constitutes a sort of spacious antechamber to the vast new palace of Arthurian ro- mance, Layamon is himself thoroughly under the dominion of the old Germanic epic tradition. He inherits its metrical form — although admitting into his verse many laxities — its stock of epic formulae, and, above all, its ethical emphasis and its spirit of staunch courage and devotion. His patriotism is so sturdy that Arthur, the scourge of the English according to Celtic tradition, becomes in his work, himself, an Englishman. Despite his dif- fuseness, which, by its insistent elaboration of detail, made a poem, already long, intolerably longer, Layamon had a larger share of 74 This account of Arthur's end accords, in general, with Geoffrey's Historia and Wace — also, with the former's Vita Merlini. There were other legends, however, about his death: 1. He is still, asleep with his warriors in a cave, whence he will issue some day to restore the glory of his people. The earliest recorded forms of this legend represent him as in Mount Aetna. Cp. Gervase of Tilbury in his Otia Imperialia (composed about 1211), Secunda Decisio, cap. 12 (ed. F. Liebrecht, Hannover, 1856) and Caesarius of Heisterbach, Dialogus Miraculorumy Distinctio XH, cap. 12, ed. Joseph Strange (Cologne, 1851). Crusaders were, doubtless, responsible for this loca- lization. There are similar legends connected with hills in England, Scotland and Wales. 2. He is the Wild Huntsman of the storm-myth. Cp. Index, below. 3. He was turned into a bird: (a) a chough. Cp. M. A. Courtney's Cornish Feasts and Folk-Lore, p. 58 (Penzance, 1890) and Notes and Queries (7th series), IV, 247. (b) a raven. This superstition is first mentioned by Cervantes, Don Quixote, Part I, ch. 13. See, too, for Cornwall, Notes and Queries (First Series) VH, 618, and for Brittany, Felix Bellamy's Foret de BrechSliant, I, 129 (Rennes, 1896). 4. He was killed by the Cath Paluc, a catlike monster. Cp. p. 41, note 9, below. Nos. 1, 2, 3 are mere adaptations to Arthur of widespread motifs. On everything relating to Arthur's end cp. Bruce, Mort Artu, pp. 298 ff. (Halle, 1910). There, inter alia, fuller details concerning the legends just enumerated are given. Early Traditions Concerning Arthur 35 the genuine poetic spirit than any other Englishman during the two centuries and a half that followed the Norman Conquest. He was connected, however, with a perishing tradition, and it was this circumstance rather than the defects of his work that caused him to pass away without leaving a trace of influence on English poetry. The author of his French original may seem, at first sight, to have been even more unfortunate, since not a single copy of his work has survived.75 He bequeathed, however, to a man of greater genius than his own — the author of the Vulgate Mort Artu — the materials for one scene, at least, through which he has merited a reflected immortality — namely, the scene of Ar- thur's translation to Avalon near the end of that romance. There is no profit in following the stream of the Geoff reyan tradition concerning Arthur — his achievements and his tragic end — further down than Layamon. As late as the sixteenth cen- tury and occasionally even beyond the pseudo-historical fables which had their fountain-head in Geoffrey found a place in all histories of Great Britain. It was through such channels that tho story of King Lear, for example, reached Holinshed, and then Shakespeare. The later chroniclers, however, it is evident, contented themselves with copying and combining the accounts of their predecessors. If they made any additions to their sources, it was not from oral tradition that they did so, but from their own Of Layamon's poem only two copies have survived, one dating from the first part of the thirteenth century, the other from the latter part, or, perhaps, even from the first part of the fourteenth. Both are in the Cottonian collection of the British Museum. The second, which is much mutilated by fire is inferior to the first and represents a new recension of Layamon's work. Cp. Rudolf Seyger, Beitrage zu Layamon's Brut (Halle diss., 1912). — Layamon's im- mediate original, after all, was substantially identical with Wace. The ratio of the added lines to those in the original Wace could not have been large. Nevertheless, one has to make allowance for the fact that all apparent expansions in Layamon of any extent were already in his French original. This is neglected by Miss Frances Lytle Gillespy in her „ Layamon's Brut: A. Comparative Study in Narrative Art", Univer- sity of California Publications in Modern Philology, III, 361 ff. (1916). 36 Evolution of Arthurian Romance fancies, mainly in the way of interpreting what seemed to them difficulties of their originals.76^ 78 In his Arthurian Material in the Chronicles, to which I have already often referred, R. H. Fletcher has sifted this enormous mass of writings — in Latin, French and English, in verse and in prose — with an industry and judgment that leaves nothing to be desired. If the result for the period subsequent to Layamon is almost wholly negative, he has none the less earned the gratitude of all students of these subjects. Chapter II. Origin of the Lays and Romances. The first French chronicles in verse, based on Geoffrey, which developed the Arthurian theme were somewhat earlier in date than any of the Arthurian romances that have come down to us, the earliest of the extant romances being the Erec of Chretien de Troyes, which was composed, as was stated above, about 1168. The influence of the chronicles on the new genre is important, but it was not in supplying the latter with specific narrative motifs for development, but rather in giving eclat to Arthur and his court and turning the attention of the literary world of the time in the direction of the stories already connected with his name,1 and in 1 For the specific influence of Wace on Chretien, the most im- portant of the romancers and the one whose works are the earliest that have survived, cp. Annette B. Hopkins, The Influence of Wace on the Arthurian Romances of Crestien de Troies (Chicago diss., 1913). Even the rather modest claims which Miss Hopkins makes on behalf of Wace are somewhat exaggerated. Brugger, moreover, in his review of her study, Zs. /. frz. Spr. u. Litt. XLIV2, 13ff. (1916), criticizes her for attributing to Wace influences that may have been really due to the versions of Gaimar or Martin (of Rochester), to the Munich Brut or the version in laisses monorimes. The versions just named, as stated above, have all perished, except for fragments that are not pertinent to our inquiry, so that one cannot control the matter with certainty. Nevertheless, we know that Wace — thongh only named once in subsequent mediaeval literature, viz. in the prologue to Guiron le Courtois — was much read and we have seen that his Bruty in an expanded form, was used in two of the most impor- tant of Arthurian texts, whilst there is no indication at all that the other paraphrases of Geoffrey ever exercised the slightest influence. I cannot, therefore, regard Brugger's point as possessing any real importance. 88 Evolution of Arthurian Romance stimulating the poets to still other inventions of the same nature. That is to say, their influence was the same as Geoffrey of Mon- mouth's — only more powerful, for, as has been already remarked, poems in the vernacular commanded a more extensive audience than pseudo-histories in Latin. Since the metrical romances have only this general relation to the chronicles, we have to look elsewhere for their origin, and here we are brought back again to the question of those traditions of the Celtic races which gave the starting-point for the narratives of Geoffrey and his followers. This problem of the relations of the Arthurian romances to Celtic tradition has been hotly debated and some extremists have gone so far as to deny that the French writers owed practically anything to this tradition.2 2 This is, virtually, the position of Adolf Holzmann in his very able article, Artus, Pfeiffer's Germania, XII, 257 ff. (1867). In most of his contentions with respect to fundamental questions of Arthurian discussion he agrees with the Anti-Celticists of later decades. Thus, pp. 262 ff., he regards the three Welsh tales, Geraint, Owen and Peredur as derived from the French; pp. 276 ff. he argues that the Arthurian passages in William of Malmesbury's De Antiquitate Glas- toniensis Ecclesiae are late interpolations, and that the same thing is true of the passage concerning Gawain's tomb in that writer's Gesta Anglorum, III, 287, else, the author would have told of Arthur in his First Book; p. 282 he declares that the Brut Tysilio is based on Geoffrey, etc., etc. Foerster and Golther do not go to such ex- tremes as Holtzmann, but the whole tendency of their work, as we shall see, throughout the present treatise, is, likewise, to minimize Celtic influences in the Arthurian romances. Perhaps, the best general statement of Golther's position in the matter is to be found in two articles which he contributed to the Zs. fur vergleichende Literatur- geschichte, Neue Folge, III (1890), viz., pp. 211 ff., „Zur Frage nach der Entstehung der bretonischen oder Artus-Epenu, and pp. 409 ff., „Beziehungen zwischen franz5sischer und keltischer Literatur im Mittel- alteru. Golther distinguishes (pp. 212 ff.) three different successive forms in which the matiere de Bretagne was embodied. 1. Sagen- bestandteile, i. e. the episodes which constitute the materials of the Breton epics and which, also, occur independently of these, e. g. the chastity ordeal of Iseult in the Tristan poems. 2. Sagendichtung, i. e. oral tales which might combine and elaborate earlier independent incidents. 3. Die vorhandenen altfranzosischen Epen alter Art, i. e. the Old French romances. The Celts, he thinks, (p. 213) are impor- Origin of the Lays and Romances 39 The principal circumstance which renders a satisfactory judg- ment in the matter so difficult is that we have so little preser- ved of the early literature of the two Celtic peoples who are most directly concerned in the debate — namely, the Welsh and the Bretons. Indeed, nothing but names and glosses in the Breton language have survived from the Middle Ages.3 The folk-songs which Th. H. de la Villemarque published in 1839 under the title of Barzaz-Breiz and which related in some instances to the prin- cipal characters of the Arthurian legend — for example, the king himself and Merlin — being so late, would not have had very much value for our purpose, even if they had been genuine. As a matter of fact, however, these have been proved to be substantially forgeries and not songs taken down from oral recitation, as they pretended to be.4 Indeed, they are, in a large measure, free in- ventions of their nineteenth century author — sometimes owing a mere suggestion to genuine folk-songs, but more frequently based on books. It is a case of the Macpherson forgeries over again — the fraud of a perf ervid patriot — only the connection with ac- tual tradition in this instance is even slighter than could be claim- ed for the Ossian and Fingal pieces. The conditions are somewhat better with Welsh literature, for here we have in verse the poems of the Four Ancient Books tant only for no. 1, and with regard even to no. 1, not all of the materials under this heading, by any means, are of Celtic origin. No. 2 is French (developed by itinerant and often bilingual conteurs), and we find combined here not only Celtic materials, but others drawn from the general fund of mediaeval stories. The Celtic names in the French romances of the Arthurian cycle are due, in a large measure, to fashion and caprice (p. 215), after the conquest of England had made the matter popular, and they no more imply necessarily Celtic origin for the stories in which they appear than Germanic names in the chansons de geste imply Germanic origin for the stories of those poems. Golther ascribes to the combinations and inventions of the authors a large share in no. 3. 3 Cp. H. Zimmer, Die romanischen Literaturen und Sprachen, mit Einschlufi des Keltischen, p. 133 (1909) in Paul Henneberg's series, Kultur der Gegenwart. 4 Cp., especially, F. M. Luzel: De V authenticite des chants du Barzaz-Breiz de M. de la Villemamup fParis, 1872). 40 Evolution of Arthurian Romance of Wales — 5 The Black Book of Caermarthen, The Book of Aneu- rin, The Book of Taliessin, The Red Book of Hergest. These poems, although they purport to be the work of Welsh bards of the sixth century, are dated by modern scholars in a much later time — from the tenth to the twelfth centuries.6 They relate in part to battles which fall in the period to which Arthur belonged — that is to say, the sixth century — and a few of them mention Arthur7 and certain characters connected with him who play a 6 Edited, with translation by W. F. Skene, under this title, in 2 vols. (Edinburgh, 1868.) 6 Zimmer, op. cit., p. 116. 7 English translations of the poems relating to Arthur in the four ancient books of Wales are given by Skene, I, 259 — 268. There are only four of such poems and of these four the first, entitled The Chair of the Sovereign (Book of Taliessin, XV), seems to mention Arthur only incidentally. The poem, it is true, is obscure to the last degree. In the second (Black Book of Caermarthen, XXXI) Arthur and Cai (Kay) are seeking entrance to a castle, but the porter. Glewlwyd Gavaelvawr (represented in the Welsh prose tales as one of Arthur's own porters), refuses until he knows more about the per- sons who wish to be admitted. Arthur eulogizes, then, his followers — particularly, Cai, to a description of whose exploits most of the poem is devoted. Thus, although Arthur is the speaker, the piece is really a glorification of Cai. In the fourth poem, Geraint, son of Erbin (Black Book of Caermarthen, XXII), Arthur is barely mentioned (Skene, I, 267) as "emperor, and conductor of the toil" in the battle of Llongborth (wherever that may be), in which Geraint, the hero of the poem, was slain. Only the third poem, Preiddeu Annwvn (^= The Harryings of Hades), Book of Taliessin, XXX (Skene, I, 264 ff.) celebrates Arthur as the central character. I have quoted an extract from this poem in the text above. The piece, as will be seen, refers uj an obscure manner to expeditions which ArthuY and his men — in his ship, Prydwen — made against certain places, named in the poem, but not yet identified. Sir John Rhys, pp. XXIV f. of his pre- face to the edition (London, 1906) of Malory's Morte Darthur in Everyman's Library, says: "The poem evidently deals with expeditions conducted by Arthur by sea to the realms of twilight and darkness". He also compares the quest of the cauldron of the Head of Hades in the Welsh poem to a similar quest in Kulhwch and Olwen, only the cauldron in the latter is in Ireland, not Hades. The preface, just mentioned, contains the best discussion of Arthurian allusions in these Welsh poems that we have. Rhys, however, omits to comment on Origin of the Lays and Romances 41 large part in the French Arthurian romances — for instance, Gwalchmei8 (Gawain) and Kay (Kei),9 Arthur's seneschal the passages (Skene I, 263, 295) about Llachan, son of Arthur. This son reappears in several of the French romances under the name of "Lohot". Cp. my discussion of the character in RR. Ill, 179ff. (1912). *) For Gwalchmei (Gawain of the French romances) in Welsh literature cp. Loth's Mabinogion2, I, 288, note 1. Ibid, in Kulhwch and Olwen he is called son of Gwyar, and it is said of him: "He never returned from a mission without having achieved it; he was the best of footmen and the best of horsemen. He was Arthur's nephew, son of his sister and cousin". In a triad, Loth, II 2, 89, he is one of the "three goldentongued knights at Arthur's court"; but this last may have been written under French influence. Loth observes, I2, 288, note 1, that gwalch means "male falcon" and gwyar, "blood". He does not explain mei. According to Rhys, Arthurian Legend, p. 13, Gwalchmei probably means "hawk or falcon of May". As has just been stated, Gwalchmeis father in Kulhwch and Olwen is named Gwyar; in Geoffrey (Book VIII, ch. 21, et. seq.) and the French romances Gawain's father is named Lot (Loth). In his article, "Le roi Loth des romans de la Table Ronde", Revue Cel- tique, XVI, 67 (1895), J. Loth is inclined to identify this Loth with the Lloch Llawwynnyawc who figures among Arthur's warriors in Kulhwch and Olwen (Loth's Mabinogion2, I, 264) and in the Black Book of Caermarthen (Skene, I, 262). It is to be observed, however, that further on in Kulhwch and Olwen (op. cit. p. 276) this Lloch (or Llwch, as the name is there spelt) is represented as the great - uncle of Arthur (on his mother's side), not his brother-in-law. Since Geoffrey's Lot (Loth), Book VIII, ch. 21, et. seq. is from Lothian (Londonesia), may not Geoffrey have fabricated this name from the name of the country with which he connects him, just as he fabri- cated (XII, 19) the name of his fictitious prince of Wales, Gualo, from Guallia (Wales) or Sabrina, (II, 5), name of the goldess of the river Severn, from Sabren? Loth, in the above-mentioned article, does not credit Geoffrey's invention, I believe, with its proper share in what that writer says in regard to Loth. 9) In these poems, as in Welsh literature, generally — excepting the pieces that betray French influence (cp. Loth's Mabinogion2), I, 256, note 1) — Cai (Kay) plays a heroic role — not that of a butt, as in the French romances. For instance, in the Black Book of Caermarthen, XXXI (Skene, I, 264) he slays the monster, Cath Palug or Cath Paluc (Capalu or Chapalu of the Old French), which, accor- ding to one line of Arthurian tradition, killed Arthur. Cp. on this subject E. Freymond's masterly monograph, mentioned above, viz., 42 Evolution of Arthurian Romance in the French romances. Indeed, Urien, a prince of the North, who figures in the French romances as the husband of Morgan le Fay and the father of Ivain, is even more frequently mentioned than Arthur himself. The style of these poems, however, as of the Bardic poetry generally, is obscure and oracular. They are, be- sides, lyrical, not epical, so that they could not in themselves con- tribute to the transmission of legends concerning the heroes that are mentioned in them. The following, for instance, is from one of the poems concerning Arthur in the Book of Taliessin.10 It Artus' Kampf mit demKatzenungestum: Sonderabzug aus: Beitrdge zur romanischen Philologie, Festgabefur Gustav Grober (Halle, 1899). I have summarized the main points of this monograph in my edition of the Mort Artu, pp. 304 f. The idea of Arthur's death through this monstrous cat (water demon, according to Freymond) may be one of the debts of French romance to Wales, although it is not found in the extant Welsh texts. A similar tradition may have prevailed in Brittany. Cp. Galeran de Bretagne, 5068 ff., where it is ascribed to the Bretons. In any event, early in the thirteenth centnry we find it localized at Lake Bourget in Savoy and J. Loth, Romania, XXIX, 125 f. (1900) suggests that the story reached Savoy by way of Champagne and Flanders- Alsace, whose princes (patrons of Chretien de Troyes) we know were interested in the materials of romance, Freymond, to be sure, thinks that travellers crossing the Alps brought the tale to Savoy. The earliest allusions to the killing of Arthur by a monster cat occur in texts of the latter part of the twelfth cen- tury — not all of them romances. A lengthy account of this fight will be found in the Merlin of the Vulgate cycle. Cp. Sommer's Vulgate Version of the Arthurian Romances, II, 441 ff. Here, ho- wever, Arthur is victor. The episode is plainly connected with the localization of the legend in Savoy, only in the Merlin the combat is represented as taking place on Lake Lausanne (i. e. Geneva), and not on Lake Bourget, where the Mont du Chat still preserves a trace of the early localization. For some corrections to Freymond's monograph, cp. Loth, loc. cit. In "Arthurian Notes", MLN, XVII, 277, W. W. NeweU identifies the monster of no. 30 in Child's English and Scottish Ballads with the Chapalu. According to Kulhwch and Olwen (Loth, I2, 275) Cai was slain by Gwyddawc, son of Menestyr, who, in turn, was slain — and his brothers, too — by Arthur in revenge for Cai. There is nothing to correspond to this in the French romances. 10 Skene, II, 264. Origin of the Lays and Romances 43 celebrates apparently some disastrous expedition by sea of tbe great hero: "Am I not a candidate for fame with the listened song In Caer Pedryvan, in the isle of the strong door? The twilight and pitchy darkness were mixed together. Bright wine their liquor before their retinue. Thrice enough to fill Prydwen (i. e. Britain) we went on the sea; Except seven none returned from Caer Rigor. I shall not deserve much from the ruler of literature. Beyond Caer Wydyr they saw not the prowess of Arthur. Three score Canhwr (i. e. centuries of men or a legion) stood on the wall; Difficult was a conversation with its sentinel. Thrice enough to fill Prydwen there went with Arthur. Except seven, none returned from Caer Golud. I shall not deserve much from those with long shields. They know not what day, who the causer, What hour in the serene day Cwy was born, Who caused that he should not go to the dales of Devwy. They know not the brindled ox, thick his head-band. Seven score knobs in his collar. And when we went with Arthur of anxious memory, Except seven, none returned from Caer Vandwy." To the uninitiated such a piece as this would not convey much information concerning the events to which it relates, and the rest are no better. The same is true of the dialogue between Arthur and Guinevere,11 the only Welsh poem in which these two cha- racters appear together. Besides the poems from which extracts have been given above, we have the highly characteristic brief poems called Triads.12 The nature of these pieces will be sufficiently manifest from the follo- wing specimens,13 which have reference to incidents and personages in Arthurian saga: " Three furious blows in the isle of Prydein [i. e. Britain]. "One was given by Matholwch the Gael to Branwen daughter 11 Printed, with translation, by J. Rhys, Studies in the Ar- thurian Legend, pp. 57 f. (Oxford 1891). He gives it no date. 12 The best translations are in French by J. Loth, Les Ma- binogion2, H, 223 ff. (Paris, 1913). 18 Ibid., p. 246, and p. 247, respectively. 44 Evolution of Arthurian Romance of Llyr, the second by Gwenhwy vach [ i. e. a sister of Guinevere's ] to Gwenhwyvar14 fi. e. Guinevere], which led to the battle of Kamlan [i. e. Arthur's final battle with Mordred]; the third by Golyddan,' the bard, to Kotwaladyr Vendigeit." Or again: "Three costly pillaging expeditions in the isle of Prydein [i. e. Britain]. The first took place when Mordred went to Ar- thur's court at Kelliwic in Cornwall: He left neither food nor drink in the court; he consumed everything; he drew Guinevere from her royal chair and buffeted her. The second was when Arthur went to Mordred's court; he left neither food nor drink in the court nor in the district. The third was when Aeddan Vra- dawc (the traitor) went to Alclut to the court of Kydderch Hael; after him there was left neither drink nor beast alive."' The triad-form seems to have originated in Ireland, but it received its especial development in Wales, and Loth's collection contains 153 of these pieces — only an insignificant portion, no doubt, of those that once existed. There are others besides the ones just quoted that contain allusions to Arthurian matters and we shall have occasion later on to return to them, but, after all, the information we glean from them is scanty and it is obvious, that these little poems played virtually no part in the transmission of Arthurian material to the French poets.15 14 Sir John Rhys, Studies in the Arthurian Legend, pp. 35 ff., cites one of the Welsh triads as declaring that Arthur had three wives of this name. But, according to the translation of J. Loth, Mabinogion2, II, 250, the three Gwenhwyvars there named are merely called the "principal ladies of Arthur's court". To be sure, the third of these Gwenhwyvars (the one whose father was Ocvran Gawr, the giant) is Arthur's consort. Rhys's comparison, loc. cit., with the Irish Echaid Airem and the three Etains, therefore, falls to the ground. For Gwenhwyvar (Guinevere) in Welsh literature cp. Loth, op. cit, I, 259, note 3. She is always called the daughter of Gogrvan or Gogrvan Gawr in Welsh ("Ocrvan" in the triad cited above being merely a variant of that name). Loth cites from Myvyrian Archaeology, p. 863, col. 1, a Welsh proverb, which says of her that "she was bad when she was small, but worse, when she grew up." The name means "white phantom" or „ white fairy". One of the most interesting evidences in regard to the cur- rency of Arthurian traditions in Wales occurs in the Code of Gwent Origin of the Lays and Romances 45 A much more important branch of Welsh literature for our present purpose are the prose tales called the Mabinogion, so well- known to the present generation through the reprints of Lady Charlotte Guest's translation, which appeared originally in 1838 and which is now almost an English classic.16 The Welsh word mabinogion in the manuscripts of the collection is the plural of mabinogi, 17 which in turn seems to be a term applied to the tales (South Wales). Cp. Aneurin Owen's Ancient Laws and Institutes of Wales, I, 678, (London, 1841). There it is said: "When the queen shall wish a song in her chamber, the bard will sing the song about Camlan [i. e. Arthur's last battle], bnt in a low voice, for fear lest it may cause trouble in the hall" [i. e. for fear lest the song may provoke quarrels in the hall]. The earliest MS. of the code dates from shortly before 1200, but Howel Dda, who is supposed to have compiled it in its original form, ruled in South Wales in the first half of the tenth century. 18 For scientific study, however, the French translation by J. Loth, Les Mabinogion, 2 vols, (second edition, Paris, 1913) is the only one to be recommended. 17 This interpretation of the word, however, seems to have ori- ginated in a scribal error. The proper meaning of mabinogion (plural of mabinog, and not of mabinogi) is really disciples, and this is the meaning given it in the Iolo MSS., p. 211. So, according to the most careful discussion of the sense of mabinogi and mabinogion, viz., J. Loth's Contributions a V Etude des Romans de la Table Ronde, pp. 30ff. (Paris, 1912) and Les Mabinogion2, I, 11 ff. (Paris, 1913). Elsewhere in the Red Book, the four branches, Pwyll, Branwen, Manaivyddan, Math, are spoken of as forming one Mabinogi, and this is, no doubt, correct. J. Rhys, The Text of the Red Book Mabinogion, p. VIII (1887), interpreted mabinog (mebinog) as "a literary apprentice", one who aspired to be a bard, and the still obscurer word, mabinogi, as the literary baggage of such a bardic apprentice. So, too, H. Zimmer, Gottingische Gelehrte Anzeigen for June 10, 1890, pp. 513 f. and J. Loth in the passages cited above. These scholars reject speci- fically the old interpretation of the word mabinogi, to which Lady Charlotte Guest gave currency, viz. "a tale for children". On the other hand, in The White Book Mabinogion, Welsh Tales & Ro- mances reproduced from the Peniarth Manuscripts, p. XXVI (Pwll- *heli, 1907 — published really in 1909, it would seem, since the preface is dated in that year), J. G. Evans takes mabinogion as the plural of mabinogi and denies that mabinog meant "literary apprentice" 46 Evolution of Arthurian Romance that a mabinog or apprentice in the art of storytelling practised reciting. This term applies properly only to four tales in the collection that goes under the name of Mabinogion — 18 namely, "Pwyll, Prince of Dyfed", "Branwen, Daughter of Llyr", "Ma- nawyddan, son of Llyr" and "Math, son of Mathonwy" — none of which contain Arthurian material. Nevertheless, the collection thus miscalled does contain some tales in which Arthur and other characters of his cycle appear. Three of these tales are, undoub- tedly, derived from French romances — either the extant romances of Chretien (as the present writer believes) or lost French ver- sions of the same stories.19 The tales in question are Owen or The Lady of the Fountain, corresponding to Chretien's Yvain (Chevalier au Lion), Peredur ab Evrawe, corresponding to the same writer's Perceval (Conte del Graal), Geraint or Geraint and Enid, corresponding to his Erec (Erec et Enide). The French origin of the three Welsh tales just named is indicated unmistakably by their coherent structure and by the character of the life which they reflect and which in social spirit, or "was a technical term belonging' to the bardic system". In his view, "any narrative that treats of early life is a mabinogi". He cites in proof the title, Mabinogi Jesu Crist (as Rhys, Arthurian Legends, p. 2, note 2, indeed, had already done), which is a Welsh version of the Gospel of pseudo -Matthew, called Liber de ortu beatae Mariae et Infantia Salvatoris. (This Welsh version was edited by Miss Mary Williams, Revue Celtique, XXXIII, 184 ff., in 1912). Evans's interpretation of mabinogi is adopted by Miss Williams in her Peredur essay, p. IV, note 3 (1909) and by W. J. Gruffydd in his paper, "The Mabinogion", Transactions of the Honourable Society of Cymm- rodorion, Session 1912 — 1913, pp. 39 f. J. Loth, Contributions, pp. 30 f., has argued convincingly, however, that the sense which Evans ascribes to mabinogi did not belong to the word in its earlier history, but was a later (fourteenth century) development. For a review of the publication and discussion of the Mabinogion down to 1902, see Alfred Nutt's note, pp. 323 ff. of the edition of Lady Charlotte Guest's translation of these tales published by his firm in that year. 58 This was pointed out by Zimmer, Gottingische Gelehrte An- zeigen for June 10, 1890, pp. 51 Iff. Even J. Loth had failed to make the distinction in the first edition (1887) of his above-cited translation. 19 For a full treatment of this question, cp. Part IV, below. Origin of the Lays and Romances 47 domestic custom, the usages of war — indeed, in every respect — was that of twelfth century France, as depicted by Chretien and the other French poets. Here, moreover, instead of the pre- cise localization of the genuine Welsh stories, which makes it possible even at the present day to follow with substantial exact- ness the movements of their heroes in Wales, everything is vague and shadowy.20 On the other hand, there are two tales in the col- lection that represent genuine Welsh tradition about Arthur. One of these is without suspicion of contamination from French lite- rature — namely, Kulhwch and Olwen; and the other, the Dream of Rhonabwy, although from allusions which it contains to Geof- frey's Historia it could not have been composed before the se- cond half of the twelfth century, does not seem contaminated in any important respect. These stories are touched with much of the natural magic which ever since Matthew Arnold's famous essay we have associated with the Celtic genius, but, despite great beauty of detail, which illustrates the wild fancy and vivid sense of color that are features of the Celtic endowment, in reading them one cannot but be reminded of the proverbial saying that "A. Kymro [ i. e. Welshman ] has imagination enough for fifty poets without judgment enough for one." As in most Celtic stories, from the early Irish sagas down to the tales which J. F. Campbell collected in the nineteenth century in the Western Highlands of Scotland, the main impression left on the reader is one of dif- fuseness and incoherence. Let us glance rapidly at the narrative of Kulhwch and Olwen21 for the light which it may throw on the subject of the relation of the Arthurian romances to the Welsh tales. Of all the ancient Welsh stories that have survived it is best suited to that purpose. Here we have a genuine, uncontaminated Welsh tale which we may compare with the French romances and thus gain a measure of the possible influence of the Welsh stories 20 In The Mabinogion, pp. 347 ff. (London, 1902) A. Nutt has commented on some of these features of the three tales. He remarks that, as in the French biographical romances, in these tales" the centre of interest is less the incident than the hero whose personality is the chief, if not the sole, connecting link" 41 Loth, Mabinogion9, I, 243 ff. 48 Evolution of Arthurian Romance on the latter. Like the other tales of the Red Book of Hergest, the fourteenth century manuscript which has preserved to us the so-called Mabinogion, Kulhwch and Olwen was probably written down about the end of the twelfth century — so shortly after the period in which the Arthurian romances in French appear to have begun — but it was manifestly older than this and had in all like- lihood been current in substantially the same form for at least a century before.22 It is possible that Foerster is right in con- tending that the interest in Arthur inspired by the French ro- mances was the cause of the writing down of this tale, but it shows no trace of their influence, so, even if this is true, the circumstance is immaterial. The hero, Kulhwch, is a cousin of Arthur's. He has fallen in love with Olwen, daughter of Yspaddaden Penkawr, merely from report, for he had never seen her. At his father's instigation he sets out for Arthur's court to ask him to cut his hair — that is, according to Welsh custom, to adopt him as his godson — and to assist him in winning the hand of his unseen lady-love. The journey is described with the most brilliant play of fancy. Take, for instance, the description of the dogs that accompanied the young prince: "Before him were two brindled, white-breasted greyhounds, having strong collars of rubies about their necks, reaching from the shoulder to the ear. And the one that was on the left side bounded across to the right side and the one on the right to the left, and like two sea-swallows they sported around him." And similarly it is said of the steed that "he cast up four sods with his four hoofs, like four swallows in the air, about his head, now above, now below. About him was a four-cornered cloth of purple, and an apple of gold was at each corner, and every one of the apples was of the value of an hundred kine upon his shoes, and upon his stirrups from his knee to the tip of his toe. And the blade of grass bent not beneath him, so light was his courser's tread as he journeyed towards the gate of Arthur's Palace." " Loth, ibid., p. 40, dates the first redaction of the tale in the second half of the eleventh century or beginning of the twelfth — so, too, Revue Celtique, XXXII, 436 (1911), after full discussion of the evidence. Origin of the Lays and Romances 49 He demands an interview with the king, but Glewlwyd Ga- vaelvawr, the porter who guards Arthur's gate on New Year's Day — the day of the prince's arrival — will not admit him within the precincts of the palace, where the king is engaged in revelry — but he promises him suitable entertainment out- side. "A lady shall smooth thy couch, and shall lull thee with songs; and early to-morrow morning, when the gate is open for the multitude that came hither today, for thee shall it be opened first, and thou mayest sit in the place that thou shalt choose in Arthur's hall." The young man, however, insists on admission and threatens, if his request is not granted, that he will set up three deadly shouts which will be heard throughout Great Britain and Ireland and which will render the women of the palace barren from that day on. The porter, then, informs Arthur of his arri- val and lauds in extravagant terms his appearance, so that con- trary to Kay's advice the king admits the stranger, who rides into the hall on horseback in true mediaeval style. The young prince then asks for a boon, accompanying his petition with threats. Just as in the French romances, Arthur pledges himself to make this don, before he has even learned what the stranger is going to demand, provided only that he does not ask for his ship, his mantle, his sword, Caledvwlch (i. e. Excalibur), his lance, his shield, his dagger and his wife. Kulhwch then demands the assis- tance of Arthur and his warriors in obtaining Olwen as his bride. He enumerates these warriors in a list six or seven pages long, which doubtless in the course of tradition underwent expansion from time to time, like the catalogue of the ships in The Iliad or the lists of peoples and princes in the Anglo-Saxon Widsith. The most familiar characters among them are Kay (Kai) and Bedivere. But it is a different Kay from the butt of the French romances, who is always the first to undertake every adventure announced at Arthur's court and is always ignom iniously over- thrown. Not only is he a valiant warrior here as throughout Welsh tradition, but, like everyone else in this fantastic tale, he has mar- vellous qualities: He had "this peculiarity that his breath lasted nine nights and nine days under water, and he could exist nine nights and nine days without sleep. A wound from Kay's sword 50 Evolution of Arthurian Romance no physician could heal. Very subtle was Kay. When it pleased him, he could render himself as tall as the highest tree in the forest. And he had another peculiarity — so great was the heat of his nature, that, when it rained hardest, whatever he carried remained dry for a handbreadth above and a handbreadth below his hand; and when his companions were coldest, it was to them as fuel with which to light their fires." Arthur is bound by his promise and sends forth messengers to every land within his dominions to seek for Olwen. For a year the search is vain and Kulhwch begins to reproach the king with the failure. Arthur now goes forth in person to remove the dis- honor of an unfulfilled promise and he is accompanied by the young prince and a band of his followers, including a guide who was "as good a guide in a land he had never seen as he was in his own." They come to a beautiful castle, which proves to be that of Olwen's father. Near it is a flock of sheep which seemed endless and which was guarded by a mastiff, whose fiery breath was as devastating as that of a dragon. They go to the herdsman's house, and the herdsman's daughter runs out with joy to meet them and tries to throw her arms about Kay's neck. The uncanny mastiff, however, had probably made Kay suspicious, so that he had provided himself with a billet, and when she endeavored to embrace him, he thrust forward the log instead. It was lucky that he did so, for she squeezed it so hard that it became a twis- ted coil. "Oh woman," exclaimed Kay reproachfully, "if thou hadst squeezed me thus, none could ever again have set their affec- tions on me." By the assistance of this maiden they have an interview with Olwen, whose beauty is charmingly described, and Kulhwch ob- tains from her a promise of marriage, if he will fulfill the con- ditions which her father imposes. The father cannot see his fu- ture son-in-law until the latter has raised his eyebrows which have fallen over his eyes — these eyebrows being usually supported by forks. On the fourth day the father, after the manner of the fairy- tales, announces all sorts of extravagant conditions for the achievement of his daughter's hand. The lover is, of course, not deterred and insists that they will all prove easy. The crowning Origin of the Lays and Romances 51 condition, however, is stated as follows: "Throughout the world there is not a comb or scissors with which I can arrange my hair, on account of its rankness, except the comb and scissors that are between the two ears of Twrch Trwyth, the son of Prince Tared. He will not give them of his own free will, and thou wilt not be able to compel him. This Twrch Trwyth turns out to be a mar- vellous wild boar — doubtless a parallel to the classical Mino- taur — 23 with one parent a man and the other a beast — the same that we have heard of already in Nennius — the Tortain of the Conte del Graal.2* In the tale of Kulhwch and Olwen, as in Nennius, Arthur's dog, Caval or Cabal, also takes part in the chase of the boar. In order to catch Twrch Trwyth, still other extravagant conditions are stated. Suffice it to say, however, that in the end, by the help of Arthur and his hosts, the young prince overcomes all difficulties and wins the hand of his bride, although at the expense of her father's life, whose fate it was that he should die a violent death, as soon as all the above-mentioned conditions had been fulfilled.25 28 34 Cp. G. Paris, Romania, XXVIII, 217, note. J. Rhys, Celtic Folklore, II, 521 (Oxford, 1901), identifies Twrch Trwyth with "Ore treith" in Cormac's Irish Glossary. The latter means "King's Boar". The hunting of Twrch Trwyth is des- cribed so exactly in Kulhwch and Olwen that it is possible to trace its course (in essentials) on a modern map of Wales. Cp. Rhys, ibid.f pp. 509 ff. On the local coloring of the Mabinogion and the local legends which they embody, see too, E. Anwyl, "The Four Branches of the Mabinogi", Zs. f. celt. Ph., I, 277 ff. (1897). His "Notes on Kulhwch and Olwen" in the Revue Celtique, XXXIV, 132ff. (1913) are purely textual in character. " As regards the relations of the French romances to Welsh literature it is desirable to note that F. Lot has cited some names in the French romances as of specifically Welsh origin: Thus, in Ro- mania, XXIV, 322, he points out that Lis in the name Chateau de Lis (first continuation to Chretien's Perceval and elsewhere.) is really Welsh Llys = castle, and ibid., p. 326, he explains the personal name Gorvain Cadrut (in Meraugis and other romances) as a combination of Welsh Gwrvan and Cadrod (Cadrawd). For Welsh din, dinas (= fortified place), in the names Dinas, Dinan (Beroul's Tristan) cp. Lot, Romania, XXIV, 337, but cp., also, the glossary to Bedier's edition of Beroul under Lidan. So, too, with reference to Dinas- 52 Evolution of Arthurian Romance We have taken now a hasty survey of mediaeval Welsh lite- rature, as far as it bears on our subject, and we have only to add that the literature of the Bretons of which, as was said above, nothing has come down to us from this period did not probably differ from it in essentials. Especially in the matter that most concerns us, the form of epic narrative, there is no reason to doubt that with the Bretons as with the other branches of the Celtic race this was the prose tale and not narrative in verse, as was the case with the Germanic tribes or the Greeks and the Italian races. So, for instance, the Irish epic consists of the famous prose sagas concerning Conchobar, Cuchullin, and the rest; and the only branch of the Germanic race which shows, by way of exception, a highly developed prose epic is that which at a comparatively early period underwent Irish influence — namely, the Norsemen.26 It will be necessary in the following to remember this peculiarity of the Celtic literatures.27 %. The Lays* We have seen, then, what is the form of the narrative lite- rature of the Celtic races, as it is actually preserved. Let us turn now to the consideration of a more or less hypothetical genre of this literature, about which Arthurian controversy has waged most bitterly, since it has been regarded as most closely connect- ed with the origin of the Arthurian romances. I refer to the Celtic lais. The difficulty is that we have French lais from the twelfth century down, like those of Marie de France — lais that are often said to be based on lais bretons — but we have no such pieces in extant Celtic literature, and leaving aside diron (Dinatiron) of the Roman de Troie (son of Priam), as against Lot, Romania, XXX, 195 (1901), G. Baist, Romanische Forschungen, XXIX, 319 L (1910), sees here the influence of Dinas, supposed name of the penitent thief, who was crucified with Christ. But this is not probable. 28 Cp. Andreas Heusler, "Die Anfange der islandischen Saga" in the Abhandlungen of the Berlin Academy, Philos.-Historische Klasse, Jahrgang 1913, no. 9. He concludes that the Norse prose saga-form was derived from the Irish, although we cannot point to specific borrowings. 27 Zimmer, reviewing G. Paris's treatise on the Arthurian ro- mances in verse (Histoire litteraire de la France, vol. 30) in the Gottingische G. A. for Oct. 1, 1890, was the first scholar to lay stress on this in Arthurian discussion. Origin of the Lays and Romances 53 for the moment the word, Breton, the very meaning of the word, lai, is the subject of debate.28 Suffice it here to say that the most probable etymology so far suggested29 identifies it with the Irish loid or laid — which seems to mean "song" — and so, no doubt, lai meant "song." Before discussing, however, this question of the lais, it will be well to indicate briefly the character at least of the French poems so designated by giving an outline of one of them. We may remark, in passing, that they are all in octosyllabic couplets, but vary greatly in length. I will select for the purpose the first in the collection of Marie de France, and one of the longest, na- mely, Guigemar, a poem of 886 lines. Whatever theory we may hold as to the origin of the story, it is evident that in Maries version it has been adapted to the conditions of twelfth century feudal society: Guigemar30 is the son of the lord of Liun (Leon) in Brit- tany and he serves his apprenticeship in knighthood at the court of Hoilas (Hoel), king of that country. After he is dubbed knight, he takes part in a military expedition in Flanders, and distingui- shes himself in the highest degree. He has one fault, however: he is wholly insensible to love. After a time he goes home to his parents and whilst there engages in the chase. On one occasion, being separated from his companions, he sees a beautiful white doe which he shoots at. The arrow wounds the animal mortally, 38 The history of the discussion of the word is given in Karl Warnke's second edition of the Lais of Marie de France, p. vi. 29 Proposed by H. d'Arbois de Jubainville. Cp. Romania, VIII, 422 ff. (1879). 80 The text stands first in Warnke's edition. The fullest dis- cussion of the name is H. Zimmer's Zs. f. frz. Spr. u. LitL, XIII1, 7ff. (1891). Zimmer identifies it — correctly, no doubt — with the Breton, Guilhomar. On the other hand, though granting that the tale is Breton, J. Loth maintains, Annales de Bretagne, XI, 479 (1895 — 6), that the name, Guigemar, can just as well be Cornish. It occurs, indeed, as Wihumar in Domesday Book. Variant forms of this name, viz. Guigamor (Guiagamar), Guiomar (Guionmar), appear elsewhere in Arthurian texts. On this subject, cp., besides Zimmer, op. cit., E. Freymond, in the same journal, XVII *, pp. 17 ff. (1895), and W. Hertz, Spielmannsbuch*, p. 382. 54 Evolution of Arthurian Romance but, strange to say, rebounds and wounds Guigemar also severely in the thigh, so that he has to descend from his horse. To his astonishment, the dying doe speaks to him and predicts that he will never recover, save through a lady whom he will love and who will love him in return, but this mutual passion will cause them unheard-of sorrow. Guigemar gets rid of his servant by sending him after the other huntsmen and he, himself, strikes through the forest until he arrives at the seashore. He sees there a beauti- fully appointed ship with no human being about it. He lays himself down in a magnificent bed in the ship, with the intention of resting, but before he is aware, the ship is out on the high sea, and it moves on until about evening he finds himself before an ancient city, which was the capital of the land. The lord of the city was an old man with a young wife whom he kept shut up in a castle, accessible by only one entrance and that from the sea. Her only companions were a niece and an old priest who was a eunuch. Now, the day of Guigemar's arrival the lady and her damsel were walking in the enclosed garden which was attached to the castle and they saw the strange boat approaching. When the boat has reached the castle, the damsel enters it, observes the wounded knight and returns to tell her mistress. They both go down now, Guigemar awakes and tells his story, and they take him to the damsel's room and keep him there. Soon his wound ceases to pain him, but he is pierced to the quick by the arrows of love. The damsel perceives that her lady and the stranger love each other and she exhorts the latter to disclose his love. This he does and the two live in happiness for a year and a half. But the lady has all along the feeling that their amour will be detected and that they will be doomed to separation. So she makes a knot in the shirt of her lover and compels him to promise that, if they are ever separated, he will love no woman who cannot un- tie this knot. He, on his part, gives her a girdle and requires of her a similar promise: She is never to love any man who cannot undo it. The fears of the lady at last come true. That very day a chamberlain discovers her intrigue and informs his master. Gui- gemar, when caught, tells his story to the husband and puts on so Origin of the Lays and Romances 55 bold a front that he is finally allowed to embark again in his boat. He reaches home again, but he remains melancholy and, among other things, refuses to entertain the idea of marriage, vowing always that he will wed no woman, unless she is able to untie the knot. Ladies come from all sides, but none are able to fulfill the condition. In the meanwhile the lady of the castle is as full of sorrow as Guigemar, and this continues for two years. She resolves at last to drown herself in the sea at the point from which her lover had set forth. She goes to the spot and finds the boat there, but the thought comes over her that Guigemar perhaps is drowned and she falls into the boat in a swoon. The boat takes her to Brittany, to the castle of a lord, named Meriaduc, who soon falls in love with her. She rejects him, however, alleging as an excuse her vow about the girdle. This leads him to tell her of the strange knight in this land who has made a similar vow about the knot in his shirt. The news that her lover is so near causes her to swoon. After this one knight after another tries to unloose the girdle, but without success. Finally Meriaduc proclaims a tournament and summons Guigemar to fight on his side. He comes — he and his lady recognize each other, but hold back — then Meriaduc exhorts her to try the knot in Guigemar's shirt. She unties it at once. Her lover now learns that no one as yet has undone the girdle and he demands her of Meriaduc, who refuses. Thereupon Guigemar and the assembled knights declare themselves against Meriaduc. They besiege and capture his castle — they put to death Meriaduc and the garrison — and finally the two lovers are united for good and all. This, then, is a specimen of the French lais. Even in such an abstract as this not all of the wild interest of the story is lost, but one must read the original to catch the charm of its naivete, which effaces completely the boundaries between fairyland and actual life, and of the simple, direct story of passion, which, neither hasting nor lingering, goes straight to the mark. If one were asked to name the books which represent best the romantic charm of the Middle Ages, the wisest answer, perhaps, would be the lais31 of Marie de France and Gottfried of Strassburg's Tristan. 1' 56 Evolution of Arthurian Romance Of this admirable poetess, whose naive grace cast its spell over Goethe in his old age, we know little more than her name, her nationality, and her works. We may safely affirm, however, that the unnamed king to whom she dedicates her lays in her pro- logue was Henry II of England (1154 — 1189). The designation, "de France," seems to show, moreover, that she was of royal blood.32 Indeed, she has been plausibly identified33 with an ille- gitimate half-sister of Henry II's, on the paternal side, who was abbess of Shaftesbury from as early as 1181 and was still alive as late as 1215. Let us return now, however, to the question of the origin of the French lais. Apart from other considerations, convincing testimony to the fact that stories concerning Arthur and his followers were current among the Normans both in England and on the Continent about the middle of the twelfth century is supplied by various passages 81 The little that we can gather about Marie de France from her poems is best summarized by Miss Edith Rickert: Marie de France: Seven of her Lays done into English, pp. 137 ff., (New York, 1901). Emil Winkler in his Marie de France, Sitzungsberichte of the Vienna Academy, Philosoph.-Hist. Klasse, 188 Band, 3 Abhandlung (1918), endeavors to identify her with Marie de Champagne, patroness of Chretien de Troyes. On this improbable theory, however, see Giulio Bertoni, "Maria di Francia", Nuova Antologia for Sept. 1, 1920, pp. 18 ff. Scholars differ as to whether her dialect is pure French or Franco -Norman. It seems most probable that she was from Nor- mandy or from the part of the Isle de France that borders on Nor- mandy. Cp. T. A. Jenkins's edition of her L'Espurgatoire Seint Patriz, pp. 21 ff. (Philadelphia, 1894) for a discussion of her language and for previous literature on the subject — also, Bertoni, loc. cit. 32 Cp. Marguerite de France, daughter of Louis VII of France and wife of Prince Henry of England, Henry H's oldest sod — Blanche de France, daughter of Louis IX, king of France and wife of the In- fanta of Castile — and many other royal personages in later history. Elise Richter, Zs. f. rom. Ph. XL 728 ff. (in her review of Winkler), disputes, unconvincingly, that the designation was limited to royal personages. 88 By John Charles Fox, "Marie de France" English Histo- rical Review, XXV, 303 ff. (1910) and "Mary, Abbess of Shaftes- bury", ibid., XXVI, 317 ff. (1911). Origin of the Lays and Romances 57 in Wace's Brut, a work which, as we have seen, was finished in 1155. First the famous passage (11. 9994 ff.) which contains the earliest recorded mention of the Round Table: Por les nobles barons qu'il ot Dont cascuns mieldre estre quidot . . . Fist Artus la Roonde Table, Dont Breton dient mainte fable. Illoc seoient li vassal Tuit chievalment et tot ingal . . . Nus d'als ne se pooient vanter Qu'il seist plus halt de son per. And then again (11. 10032ff.): En cele grant pais que jo di, Ne sai se vos l'aves oi, Furent les merveilles provees Et les aventures trovees Qui d'Artu sont tant racontees Que a fable sunt atornees: Ne tot menconge ne tot voir Ne tot folie ne tot savoir; Tant ont li conteor conte Et li fableor tant fable Pour lor contes ambeleter Que tout ont feit fables sanbler. A mong the followers of Arthur who perished in his last battle Wace includes (11. 13 675 f.): cil de la Table Roonde Dont tex los fu par tot le monde84). What was the form now of these Celtic stories — the term, Breton, we shall see, is somewhat ambiguous — to which Wace alludes?35 In his treatise on the metrical romances,36 mentioned 3* "Those of the Round Table, of whom there was such praise throughout the world." 86 Besides the two passages just quoted, cp. too, Wace, 11. 10 401ff. and 11. 10 555 f. (reference to the Round Table). G. Huet, Moyen Age, XIX, 234 f. (1916), "Notes d'histoire litteraire. I. Le temoignage de Wace sur les fables arthuriennes", remarks that the terms, "merveiller" and "aventures", correspond exactly to the con- tents of the French Arthurian romances and prove the existence of such tales among the Celts before Chretien and independently of Geoffrey. 58 Evolution of Arthurian Romance above, Gaston Paris interprets them as embracing both lais of the French kind (short narrative poems) and informal prose tales told by professional storytellers who haunted castle and inn. The existence of the prose tales is beyond dispute, as is made clear enough by the passages which Paris himself cites;37 it is only a question of the lais. That these, whatever their character, were set to music is manifest from passages in the French lais — as when it is said at the end of Guigemar: De cest cunte qu'oi avez Fu Guigemar li lais trovez Que hum dit en harpe e en rote Bone en est a oir la note. It is evident still further that the instruments to which they were usually sung were the Celtic rote — a rude kind of violin — and the harp, but especially the former. As we have already seen, however, there is no epic narrative in verse in extant Welsh or Irish literature and there is no reason to believe that the Bretons differed from their Celtic kinsmen of the British Isles in this matter; so what do the writers of the French lais mean by their citation of lais Bretons in connection with their own poems ?38 It would seem most probable that the Celtic lays referred to were simply lyrical pieces set to music concerning the characters who are told of in the lais of Marie de France and others. In singing 37 Histoire litteraire de la France. XXX, lit (1888). Op. cit.} pp. 9ff., from Thomas's Tristan, from Chretien's Ere-Cy from the continuation of his Perceval by Wauchier de Denain (Gaucher de Dourdan), and from still others. 38 It was natural that Gaston Paris, who had no firsthand know- ledge of Celtic literature, should have mistaken the force of these allusions and assumed that they referred to narrative poems in a Celtic language of the same character as the French lais themselves. But, on the whole, the assumption seems unjustified. As Warnke has pointed out, pp. XXVI ff., Marie de France, for instance, does not say that she is translating or paraphrasing a lai Bretun: she says that she is following the cuntes or tales from which the lais too, it seems, sprang: So, for example, near the beginning of her collection she uses the following words: "The tales which I know to be true, concerning which the Britons have made the lays, I will relate to you quite briefly. " Origin of the Lays and Romances 59 such a piece, the singer would find it necessary to recite in prose the story to which the poem related. This would be the case es- pecially, of course, if the Breton singer were singing before French-speaking audiences who did not know the legends of his race. There is, in all probability, then, no reason for trying to distinguish between epical and lyrical lais, as has sometimes been done. There were, we may assume, as a matter of fact, only lyri- cal lais in the Celtic language and the term was extended by Marie de France and her fellows to the short narrative poems in French which they composed; but this poetical genre had really nothing to correspond to it in the Celtic and was their own inde- pendent invention. In French literature, besides, the word, lai, continued often to designate brief lyrical pieces (chansons) as we have just assumed that it did in the Celtic. Indeed, lai in the sense of tale (conte) was never very popular in France and passed out of use altogether by the end of the thirteenth or first quarter of the fourteenth century.39 Granting that the French lays embody tales that eame to their authors from Celtic sources — and this is certainly true of some of them, at least — it remains for us to inquire from what division of the Celts did the French poets derive these tales. Was it from the Celts of Great Britain or of Brittany (Armorica)? The answer to the question turns on the meaning of the term, Breton, which qualifies the word, lai, in the references of Marie de France and others to their sources. That the term, in some cases, is equivalent to Armorican is plain. For example, the names of two of Marie's lays — Bisclavret (=Werewolf) and Laustic (= Nightingale) — are somewhat distorted Breton words and could not come from the Welsh.40 Furthermore, the name, Guigemar, ** In only five of her lays does Marie use lai as a synonym of conte. Cp. Foulet, Zs. f. rom. Ph., XXIX, 303 f. He traces there her hesitating adoption of the term in this sense. He points out, p. 311, its very limited use in the anonymous lays. In Italian lai meant lament. Cp. 0. M.Johnston, Studj Medievali, H, 554ff. (1907). 40 This was first pointed out by Zimmer in the Gottingische Gelehrte Anzeigen for Oct. 1, 1890, pp. 800f. The article is of the first importance with reference to the origin of the lays. So, too, is the same writer's "Beitrage zur Namenforschung in den altfranzbsischen Arthurepen", Zs, f. frz. Spr. u. Litt., XIII \ Iff. (1891). 60 Evolution of Arthurian Romance occurs in numerous historical documents in Brittany from the early eleventh century down and not in Wales.41 In the case of still other lays — Equitan, Lanval, Les Dous Amanz, Eliduc — geographical allusions connect them definitely with Brittany.42 These observations afford irrefragable proof that in the particular cases the word, Breton (Bretun), is synonomous with Armorican. Indeed, a comprehensive inquiry into the subject has shown that, in the general usage of the Middle Ages, the terms Bretagne, Breton, and their Latin equivalents43 were employed only of con- temporary Brittany and its inhabitants and language, or, retro- spectively, of Great Britain before and during the period of the Anglo-Saxon conquest and of the ancient Celtic population that inhabited it.44 Consequently, when Marie de France and her con- temporaries refer to lais Bretons as their sources, they have Brit- tany in mind, and their own lays are, accordingly, based on Breton stories, as far as they are of Celtic origin at all. It is true that Marie and her fellows sometimes localize the plots of their lays in Great Britain. For example, Chievrefueil (Honeysuckle), in which Tristan sends down stream to Iseult messages carved on pieces of wood, is connected with Tintagel (in Cornwall); in Yonec (and in L'Espine, under the influence of Yonec)*5 Caerleon (in 41 Cp. Zimmer's last named article, pp. Iff. 42 Cp. Zimmer, Gott. G. A., for October 1, 1890, pp. 797ff. 48 Brito, Brittanicus, Britannia, 44 These are the conclusions of Brugger's article, "Uber die Be- deutung von Bretagne, Breton in mittelalterlichen Texten", Zs. f. frz. Spr. u. Litt., XX1, 79 ff. (1898). One has to except Britannia, used as a name for Great Britain in learned writings. Still further, as Lot, Romania, XXIV, 508 ff. (1895), and in his reply to Brugger, entitled "La Patrie des lais bretons", ibid., xxviii, Iff. (1899), has shown, these archaic terms, Britones, lingua Britannica, Britannia are, also, occasionally applied to contemporary Wales, its people and its language, respectively, in the Latin writings of Welshmen, (Geoffrey of Monmouth, Giraldus Cambrensis, etc.). For a full bibliography of the controversy over the meaning of Breton, etc., see Miss Hopkins's dissertation, pp. 114ff. Brugger, like Zimmer, is an advocate of the- Armorican origin of the stories told in the French lays. 45 Cp. Foulet, Zs. /. rom. Ph., XXIX., 36. Origin of the Lays and Romances 61 Wales) and St. Aaron's church in that city are mentioned.46 But such localizations were merely innovations of the French poets, under the all-pervading influence of Geoffrey or Wace.47 Again, two of Marie's lays bear alternative titles, English and French, for the poetess says of Chievrefueil that the French call it by that name, but that the English call it Gotelef (Honeysuckle), and of Laustic that the French call it Russignol, but the English Nihtegale (Nightingale).18 G. Paris drew49 from this latter circumstance the far-reaching inference that even before the Norman Conquest Welsh musicians had spread among the Anglo-Saxons a knowledge of the stories with which the lays deal. But in this matter, also, as it turns out, we have merely another mark of Wace's influence — an imitation of his trick of giving names in different languages,50 as, for in- 46 In his articles, "De la provenance des lais dits bretons", Romania, XXIV, 513 ff. (1895) and "La patrie des lais bretons", ibid., XXVIII, 321 ff. (1899), Lot tried to establish a Welsh or Cornish origin for some of the lays, such as Yonec, Chievrefeuil, L'Espine, Le Cor. 47 This is Brugger's suggestion. Cp. his last-cited article, pp.147 ff. As he points out, this influence is patent in the case of Marie's Milun, where she says that her hero was known "in Ireland and in Norway and Guhtlande [i. e. the island of Gotland in the Baltic], in Loengres li. e. Logres, name for England in the romances] and in Albania [i. e. Scotland]". This enumeration was taken indisputably from Geoffrey's list of Arthur's conquests, Book IX, ch. 10, or, possibly, Wace's diffuse paraphrase of the same, 11. 9900 ff. Especially significant is the occurrence of the Baltic island in both lists. 48 Through a slip, G. Paris, Hist. Litt., XXX, 7, cited, also, Bisclavret as an example, for Marie says that the Normans called this lay Garulf (Garual). But this has nothing to do with English origins. A9 hoc, cit., and, before this, in Romania, XIV, 604 ff. So Lucien Foulet, "English Words in the Lais of Marie de France", MLN, XX, 109 ff. (1905). He points out the example here cited — also, another in the episode of Hengist's treachery towards Vortigern (11. 7409 ff.), where Wace explains (11. 7473 ff.) that the English call the knives which Hengist's men used sexes (= Anglo- Saxon seaxas), the French costiax. Foulet points out still other mannerisms of Wace which Marie imitates. — The trick just discussed 62 Evolution of Arthurian Romance stance, when in speaking of the famous dolmens or prehistoric temples at Stonehenge, on Salisbury Plains, in England, he says (11. 8383 ff.): "The Britons in the British language call them Karole as gaians [ i. e. Giants' Dance or Round ] ; in English they are called Senhange [i. e. Stonehenge]; in French Pierres Pandues [i. e. Scattered Stones]." We have here an exact parallel to the Lausiic of Marie de France, where the title of the poem is given first in Breton, and next in French and English. Although Marie, then, derived the materials for some of her lays from Breton sources, there is no ground for supposing that she, herself, ever came into personal contact with Bretons.51 In any event, she was certainly ignorant of the Breton language — otherwise she would not have distorted Breton bleiz lauaret (i. e. is not confined to Wace and Marie. It is found in other mediaeval writers. Cp. Foulet, Zs. f. rom. Ph., XXXII, 275f. (1908). Before Foulet, Warnke, in his second edition of Marie's lays, p. XXI (1900), and Bedier, edition of Thomas's Tristan, II, 127ff. (1905), had both offered unsatisfactory explanations of the alternative titles of the above-mentioned lays: Warnke suggests that, writing in England after the Norman Conquest, Marie provided her poems with the alternative titles to suit either a French or English audience as the case might be. Similarly, Bedier — only he ascribes the double naming to Breton jongleurs in England. For criticism of Bedier's suggestion, cp. Foulet, loc cit. 51 So L. Foulet, in his article, "Marie de France et les lais bretons", Zs. f. rom. Ph., XXIX, 319 ff. (1905). Ibid., 315 ff., he shows from Guigemar, 11. 22 ff. (compared with corresponding ex- pressions in her Fables, which was, of course, certainly based on a written source), that she used written sources — doubtless, French. He has proved, too, "Thomas and Marie in their relation, to the conteurs'% MLN, XXIII, 205 ff. (1908), that her occasional appeals to an oral source is "a meaningless mannerism". In his "Marie de France et la legende de Tristan", Zs. f. rom. Ph., XXXII, 274, note 4 (1908), the same scholar remarks pertinently : "Je ne trouve dans Marie aucun texte ou elle nous affirme avoir entendu la mtme personne, ou le meme groupe de personnes, chanter la 'note' et raconter le conte qu'elle se propose de nous repeter en vers frangais — aucun texte ou elle nous affirme avoir entendu un Breton chanter ou raconter, quoique se soit. . . * Ainsi le conteur Breton n'apparait nulle part chez Marie. Les musi- ciens qui faisaient en harpe et en rote le lai de Guigemar etaient- ils bretons? C'est possible, voila tout ce qu'on pent dire." Origin of the Lays and Romances 63 talking wolf) into Bisclavret — the name of her werewolf -hero — or have failed to see that the L of Laustic — the title of another one of her lays — was merely the French article prefixed to Aostic, the Breton word for "nightingale."52 It is evident, there- fore, that she was entirely dependent on intermediate written ver- sions in French for her knowledge of these Breton stories. And what was true of Marie was, no doubt, true of the other authors of French lays. Even in Marie's collection of lays there are some which have no real connection with Brittany. Thus, she, herself, tells us that the incident on which her lay, Les Dous Amanz (= The Two Lovers), is based occurred in Normandy. When she adds that the Bretons made it the subject of a lay, one begins to suspect that the term lai Breton is already becoming conventionalized53 and may be applied to any short narrative poem of the kind which Marie had been composing on genuine Breton themes. One may say somewhat the same thing of Le Fraisne (= The Ash) — a tale with no distinctively Breton features, which the poetess localizes in Brittany, but which is found in variant forms in every part of the world. Above all, it has been immortalized by Chaucer in his Patient Griselda (Clerk's Tale).5i What we have suspec- ted here in the case of these two lays by Marie is true, beyond question, of the lays of later authors in this genre: the term became a purely conventional one with the meaning just indicated 68 Cp. Zimmer, G. G. A. for Oct. 10, 1890, pp. 800f. G.Cohen, Zs. f. frz. Spr. u. Litt, XXIV2, 3 (1902), suggests that the fusion of the article with the noun in Laustic may have been the work of a scribe, but the equating of the word with simple russignol and nihtegale (without articles) at the beginning of the lay shows that the suggestion is groundless. 63 F. Lot, Romania, XXIV, 5271 (1895), makes this observation with regard to this lay and Le Fraisne. Ibid., p. 525, however, he cites the fact that in UEspine an Irishman is said to have sung a lay (Aelis), as proof that the Irish took part in the propagation of lays, but the Irish had long before figured in the romances — especially since the appearance of the Tristan poems — so we doubtless have in this instance, too, a purely conventional use of the term, lai. 54 With this story in Le Fraisne is combined the widespread popular notion that twins are always the offspring of different fathers. 64 Evolution of Arthurian Romance and the Celtic coloring in such poems is, if not invariably, cer- tainly more often than not, purely factitious.55 A main factor in rendering the use of the term, lai Breton, a popular literary artifice was Thomas's poem on Tristan, com- posed about 1170.56 Here, probably under the influence of Marie de France, that writer makes his British hero an accomplish- ed harper who charms audiences with his lays, and so it was the example of Tristan, the most popular figure in the literature of the time, that led subsequent poets to represent their charac- ters as singing lais bretons, or even to designate their works by that title as in the case of the French versions of Havelok and Horn, which have, of course, no connection with Brittany. Marie is the only author of genuine narrative lays57 whose name has been preserved. We have, however, besides the collec- tion which is unquestionably her work, a number of anonymous lays58 of the same character, one or two of which may be also from her pen. Some of these pieces deal with the matiere de Bretagne and have much the same charm of ,, faerie" — the charm of wild and delicate fancy — that we have noted in Marie's re- cognized works. So in the story of Guingamor (probably Marie's), where, after enjoying the love of a fay in the Otherworld for a hundred years, that passed like three days, the hero is allow- ed to return to his native land, but by disobeying the command of his mistress not to partake of mortal food, whilst there, loses both his apparent gift of immortal youth and the power to return to her — or, again, in Tydorel, where a king, the son of a water- 55 In the opinion of the present writer, this has been clearly established by L. Foulet in his "Marie de France et les lais bretons", Zs. f. rom. Ph., XXIX, 19ff., 293ff. (1905) and his subsequent ar- ticles on the lais bretons in the same journal, vols. XXX and XXXII. 88 Cp., especially, Foulet, „Marie de France et la legende de Tristan", Zs. f. rom. Ph.r XXXII, 161 ff., 257ff. (1908). B7 he Cor by Robert Biket (Biquet) in verses of six syllables, last edited by H. Dorner (Strafcburg diss., 1907), is a fabliau, and Ignaure by Renaut, also, does not belong properly to the genre. We know nothing of these authors, save their names. *8 For these lays and the critical literature relating to them see Part IV, below. Origin of the Lays and Romances 65 sprite and a mortal mother, on being taunted once with not being human, since, in consequence of his supernatural origin, he never sleeps, learns with mortification the true secret of his birth from his mother and plunges into his father's lake, never to be heard of again. There is something suspicious in the fact that out of the seven anonymous lays that deal with the matiere de Bretagne three constitute variants of themes which Marie has treated.59 The choice of themes in such cases was, doubtless, determined by her example, and, as a matter of fact, these lays, in phrasing and other matters, also, betray her influence — so distinctly, indeed, that the question has been raised as to whether the authors of the pieces in question drew at all on Celtic tradition. This skeptic- ism has included — and, justly, it would seem — even the beautiful Franklin's Tale of the Canterbury Tales in which a wo- man's honor is saved by the generosity of her lover, after she had placed herself in his power by promising him her favors, in case he performed a seemingly impossible task, which, however, he actually performs.60 Nevertheless, Tydorel, at least — to say 69 Melion is like Bisclavret; Graelent and DesirS are like Lanval. For the literature of these lays, cp. Part. V, below. Zimmer, Zs.f.frz. Spr. u. LitL, XIII1, Iff., has shown that the name Grae- lent is identical with Gradlon mur, name of a hero in early Ar- morican saga. Nevertheless, to my mind, Foulet, op. cit., XXIX, 19ff., has made out so strong a case for his thesis that Graelent is essen- tially a combination of Marie's Lanval and EUduc that I hesitate to accept it as embodying an independent Celtic tradition. 60 Chaucer describes it as a "Breton" lay, but cp. the article, "Le prologue du Franklin's Tale et les lais Bretons", Zs. f. rom. Ph., XXX, 698 ff. (1906), by Foulet, who argued that the term is here con- ventional. On the other hand, the conclusions of Pio Rajna, "Le origini della novella narrata dal Frankeleyn nei Canterbury Tales del Chaucer", Romania, XXXII, 204ff. (1903) — viz., that this tale is derived from Boccaccio's Decameron, X, 5 — are unconvincing. But one need not agree therefore with W. H. Schofield, "Chaucer's Frank- lin's Tale", PMLA, XVI, 405 ff. (1901), that Chaucer's source was a genuine Breton folk-tale. J. S. P. Tatlock has latterly, "The Scene of the Franklin's Tale Visited", Publications of the Chaucer Society (1914), even tried to determine the part of the coast of Brittany where the story is laid. 66 Evolution of Arthurian Romance nothing of Guingamor, which is probably Marie's composition — appears to the present writer to possess the natural magic of Celtic fancy. There is no convincing evidence that before Marie de France narrative lays existed at all in French literature. She, was doubt- less, the creator of the genre61 and her genius seems to have dominated it during the brief vogue which it enjoyed.62 C1 This is Foulet's conclusion, Zs. f. rom. Ph., XXIX. 56, Apart from the fact that Le Cor by Robert Biket is really a fabliau (cp. already Brugger, Zs. f. frz. Spr. u. Litt., XX1, 140), there is no valid ground for dating this poem earlier than Marie. Cp. the dis- cussion of the subject by Foulet, op. cit., p. 55, note 1. As the same scholar observes, pp. 302 f., the lais bretons which are first mentioned in the fabliau, Richeut (usually dated 1159, although it was in reality probably composed considerably later. Cp. Foulet, „Le poeme de Richeut et le roman de Renard", Romania, XLII, 321 ff.), and to which we have allusions in Chretien, the Roman de Troie, Horn, etc., are all lyrical, not narrative, and it was only by degrees that Marie, herself, came to use the word in the new sense. — For views at variance with the one here adopted, cp. Warnke's Introduction and the literature there cited — also, H. DornerV edition of Le Cor, cited above, and Ezio Levi, "I lais brettoni e la leggenda di Tristano,", Studj romanzi editi a cur a di E. Monad, XIV, 113 — 246 (Rome, 1917). The only original point which Levi tries to make is that he finds in canzoni of the Countess of Dia (who died p orb ably in 1193) and Guittone d'Arezzo echoes of the words "Isolt ma drue, Isolt ma vie" in the Tristan lax (Thomas's Tristan). He assumes that these canzoni were too early to be influenced by Marie. The dates of both, as a matter of fact, are still uncertain. Above all, however, the similarity is so general as to have no value. 62 Interesting early testimony to the popularity of the lais is supplied by Gautier d' Arras's Ille et Galeron, 11. 928 ff., composed about 1167, or perhaps somewhat later. Cp. E. S. Sheldon, MPh., XVII, 383 ff. There, with reference to the fluctuations of love, it is said. Mes s'autrement n'alast r amors, Li lais ne fust pas si en .cours, Nel prisaissent tot li baron. Equally emphatic is the testimony on this subject in Denis Pyramus's La vie St. Edmund le rei, 11. 35 ff. (ed. F. L. Ravenel, Philadelphia, 1906) — quoted by Warnke8, p. XXXVI. Denis's poem dates, it would seem, from the last decade of the twelfth century. Origin of the Lays and Romances 67 2. The Romances. We have been dealing so far with the lays, but the problem of the origin of the romances63 — more especially the question of their relation to Celtic tradition — is essentially the same, not- withstanding the position which has sometimes been taken that it is different. The fact, however, that most of the lays do not mention Arthur and his knights does not alter the case; for the most distinctive features of these romances still remain of the same general character as in the case of the lays. Take, for in- stance, the magic fountain at the beginning of Chretien's Yvain, whose waters, if dropped on the neighboring stone, raise the storm, Whereupon a strange knight rushes forth to encounter the offender, or the search of Lancelot for Guinevere in the land from which no one returns in the same writer's Lancelot, or Perceval's adven- tures at Arthur's court in the Conte del Graal. We move in the same world of romance and marvel in each genre** — only, in On the development of the meaning of the word, "romance", cp., especially, P. Voelker, "Die Bedeutungsentwickelung des Wortes Roman", Zs. f. rom. Ph., X, 485 ff. (1887). It (OFr. "romanz") was used first in the sense of "a book in the vernacular*' (earliest example noted is in Samson de Nanteuil's Dits Salamon — in the forties of the twelfth century) — but only of translations from the Latin. Wace, Roman de Ron, Part III, 11. 5331 f. (H. Andresen's edition), was the first writer, as far as we know, to drop this restriction. In the second half of the twelfth century, the term is applied to what we call "ro- mances", but is, also, employed of chronicles, and continues to be em- ployed of such works down into the fourteenth century. Nevertheless, the meaning "fictitious narrative" predominates from the thirteenth century on. W. Meyer-Llibke, Zs. f. frz. Spr. u. Litt., XLIV \ pp. 131 1., (1916) draws attention to Chretien's contrasted use of conte (short narrative) and romanz in Cliges, 11. 22 f. But the poet does not consistently maintain this distinction. Cp. the titles of his romances: Conte de la Charette, Conte del Graal. It has been maintained by certain Arthurian scholars — espe- cially by Gr. Paris, Histoire litter aire de la France, XXX, 9 (1888), E. Brugger, Zs. f. frz.. Spr. u. Litt, XX1, 151 (1898) and with unessential differences, J. L. Weston, Legend of Sir Lancelot du Lac, pp. 20, 66 ff. (1901) — that the romances were, in the first in- stance, made up of a combination of lays. The theory is similar to 68 Evolution of Arthurian Romance the case of the romances, the Arthurization is not limited to a few superficial details as in the Arthurized lays, but is thorough-going. Similarly, the adaptation of these motifs to the conditions of feudal society is even more drastic in the case of the romances. Now, in view of what has been said about the lays there can be no reasonable doubt that Brittany was one source of whatever is Celtic in the romances of Chretien and his followers. One may go further and assert that the historical conditions render this part of the Celtic territory a far more likely source of that element in the romances than Great Britain could ever be. The emigration to Armorica of the Britons who fled from Great Britain under the pressure of Anglo-Saxon invasion was in progress during the fifth and sixth centuries — so from this early period they were in constant contact with their neighbors in Gaul. —Especially close were their relations with Normandy. Indeed, from the first part of the tenth century the Bretons were vassals of the Norman duke and intermarriages between the reigning houses of the two states testify to the intimacy of the intercourse which existed between them. By the tenth century the portion of Brittany which was closest to Normandy was thoroughly assimilated to Norman-French civilization and there was a considerable zone in which, to a large extent, the population was bilingual. Under these conditions, of that which Paris maintained — e. g., in his Histoire Poetique de Charlemagne2, pp. llf., 69 ff . (Paris, 1905) — with regard to the origin of the chansons de geste. These, too, he thought, sprang from the combination of shorter pieces (cantilenae) — often, to be sure, completely transformed. We have here, of course, the familiar theory of epic origins which was first applied to the Homeric poems in the latter part of the eighteenth century (the so-called Wolfian theory) and afterwards to the mediaeval epics. This theory, as applied to the latter, however, — to say nothing of the Greek — has latterly become so generally discredited that there is no need of our discussing here its application to the romances. Cp. Bedier's Les LSgendes Epiques, with reference to the chansons de geste. For supposed influence of the Greek romances on the Arthurian romances — especially, in structural matters — through the intermediation of the Crusaders, cp. W. J. Courthope, "The Connexion between Ancient and Modern Romance", Proceedings of the British Academy for 1911 — 19 12, pp. 245ff. The theory, however, has no sound basis. Origin of the Lays and Romances 69 course, it was easy for the legends of the Bretons to penetrate into Normandy and the rest of France.65 How was it, on the other hand, with the insular Celts — the inhabitants of Wales and Cornwall and Scotland? We have seen already that the reasons advanced by Gaston Paris for supposing that even before the Norman Conquest Celtic tales were in cir- culation among the Anglo-Saxons are not sound. Moreover, in extant Anglo-Saxon literature there is no trace of such stories. We may say, in general, that the long and bitter hostility which prevailed between the two peoples was not favorable to the exchange of the materials of literature. In the main, the same thing is true of the relations that existed between the new Continental invaders and the old Celtic population of Great Britain in the decades that immediately followed the Norman Conquest. The Normans had utterly humiliated the hated Saxons, but the feeling of the Celts towards the conquerors was at first not so different from that which they had entertained towards their old enemies as might have been expected, for they found themselves engaged almost at once in a desperate struggle with the same foe. Nevertheless, during the reign of Henry I, South Wales was thoroughly subdued to Norman rule, and early in the twelfth century the conditions for peaceful intercourse between victors and vanquished were establish- ed.66 Moreover, within such Celtic areas, there was one es- pecial circumstance that doubtless facilitated the transmission of insular Celtic materials to the French-speaking invaders — na- mely, the part played by the Bretons in the conquest of England.67 Except the Normans, themselves, no other people had so large a share in this enterprise or in the spoils that resulted from it. There is abundant documentary evidence of the gifts of land which 65 In this connection it is significant that the names of the well- known Arthurian characters occur much oftener in the mediaeval do- cuments of Brittany than of Wales. On this subject, cp. Franz Putz's important article, "Zur Geschichte der Entwicklung der Artursage", Zs. f. frz. Spr. u. LitL, XIV1, 161 if. (1892). 60 Cp. J. E. Lloyd's History of Wales from the Earliest Times to the Edwardian Conquest, II, 433 ff. (London, 1911). 67 So Zimmer, Gott. G. A. for Oct. 1, 1890, pp. 78911 The sugge- stion has found general acceptance. 70 Evolution of Arthurian Romance they received in different parts of the conquered country, including Celtic Cornwall. Especially large were the holdings of the Breton prince, Alan Fergant,68 in Yorkshire. But wherever there were large Norman garrisons, as in South Wales, there can be no doubt as to the presence of bilingual Bretons among them. Now, as we know from Giraldus Cambrensis, about the year 1200, say, the language of the Bretons was still so near to that of their Welsh and Cornish kinsmen that there was no difficulty in members of these different branches of the Celtic race understanding each other. So here was a ready channel for the diffusion of Welsh stories among the Normans and thence among the French. Since we have no direct testimony on the subject, the matter must re- main more or less in the realm of conjecture, but the romances, them- selves, furnish, to some degree, reasons for supposing that such a diffusion did take place, and it is accordingly highly probable that the Bretons did play some such role as we have indicated. The problem of the relations of the romances to Celtic tra- dition — more particularly to Celtic insular tradition — would be simple, indeed, if we accepted, with some critics, the theory that the three Welsh tales, Oiven and Lunet (or The Lady of the Fountain), Peredur ab Evrawc, and Geraint and Enid — all in- cluded in the Mabinogion — which deal with the same themes as Chretien's Yvain, Perceval and Erec, respectively, were not mere adaptations of the latter, but went back independently to common Welsh sources. The differences, however, in style and construc- tion between the three Welsh tales just named and the unquestion- ably native tales in the Mabinogion, such as Kulhwch and Olwen or Branwen, Daughter of Elyr, are too plain to deny. Instead of the incoherence and ineffectual rambling of the latter, we have here organization and order. Moreover, as far as he appears, Arthur in the three tales is the roi faineant of the French romances, not the active hero of Welsh saga. These various differences, are, indeed, so manifest that it was obviously impossible to accept the above-mentioned hypothesis in an unqualified form. Consequently, Zimmer, however, was mistaken in representing this prince as taking part in the Norman Conquest of England. Cp. J. Loth, Annates de Bretagne, XI, 479 (1895 — 6). Origin of the Lays and Romances 71 nearly all scholars who have studied the problem have agreed that the three tales were influenced by French originals, whether such originals were the poems of Chretien still extant, or those of some other writer, now lost.69 There seems really, however, no need of calling into existence imaginary French sources: Chretien will suffice. For, although in the Welsh tales the narrative is more condensed than in the corresponding romances of Chretien, the two concur in their incidents throughout, and, save for one exception (in The Lady of the Fountain), in the order of these incidents. Such divergences as we observe in comparing the two sets of narra- tives are, for the most part, not difficult to explain. The differ- ences between Welsh and feudal French society at the end of the twelfth century were enormous — in respect to custom, culture, intellectual outlook — everything, indeed — and a Welsh adap- tation of a French poem would, of course, inevitably, reflect these differences. In these three Welsh ,tales pf the Mabinogion, then, we have, no doubt, the three above-mentioned poems of Chretien adapted, in part, to the conditions of contemporary Welsh society and transmuted by that riotous fancy of the Celt which has never had its equal — by some professional story-teller, we may say, then, of twelfth or thirteenth century Wales, the representative of a rude civilisation, to whose mind the marvellous was as familiar as the real. After all, however, it is questionable whether the transmutation in these cases is so great as that which was suffered by the Tale of Gamelyn, as it passed through the hands of Lodge and Shakespeare, becoming, in succession, the Euphuistic novel, Rosalynd, and the loveliest of pastoral dramas, As You Like It, or — to select an Arthurian example, — as that which befell' Malory's account of Arthur's last phase in Tennyson's Guine- vere.10 69 For recent writers who adopt the latter view, cp. Part IV, below. 70 As far as change of incident goes, there is, obviously, no comparison — the modern writers are so much more drastic. But the contrast in respect to sentiment is, in reality, equally striking. For an admirable statement of the case for Chretien as the source of the three Welsh tales, cp. P. A. Becker, Literaturblatt fur ger- manische und romanische Philologie, XXXIV, 19ff. (Jan., 1913). 72 Evolution of Arthurian Romance We have reviewed already above the scanty testimony of the chroniclers with respect to the existence of Arthurian traditions in Great Britain, and we have seen, moreover, that the sources of the extant lays were all continental. The question of these insular traditions will again arise when, in subsequent chapters, we endeavor to trace the history of the development of the legends concerning Merlin, Lancelot, Tristan, and the Holy Grail. With regard to these legends, we shall see that only for those that relate to Merlin and Tristan can it be claimed with certitude that they originated in Great Britain and that, even in the case of the former, the connection with Celtic popular tradition is of the slight- est kind. Is there, now, any other evidence which would suggest that widespread traditions concerning Arthur and his companions were in circulation in Great Britain, or, indeed, in the British Isles — for we shall include Ireland in our examination — tra- ditions which the authors of the Arthurian romances might draw upon directly or which, if not originally connected with the cha- racters in question, have been actually applied to them by the romancers? The chief evidence that bears upon our inquiry would appear to be as follows: 1. Arthurian localizations. Many localities in Great Britain bear Arthurian names — especially in those parts of the island where the Celtic element is strongest — that is to say, in Wales, Cornwall, and Somerset, and also in Scotland and Northern England.71 It is to this cir- cumstance that Tennyson refers in the well-known lines in the Epilogue to his Idylls of the King, when he speaks of That Peredur is derived from Chretien's Perceval and its continuations is now very generally conceded. Cp. Part. II, Peredur section, below. With regard to the two remaining tales, cp. Part IV, where a full history of the Mabinogion controversy is given and the evidence in the case examined. Full collections on the subject are given in J. S. Stuart- Glennie's Arthurian Localities — published as an introduction to Part III (1869) of the.E. E. T. S. edition of the Middle English prose Merlin (Original Series, no. 36). For Cornwall see, besides, W. H. Dickinson s King Arthur in Cornwall (London, 1900). Unfortunately, both writers are wholly uncritical. Origin of the Lays and Romances 73 "that gray king" whose name, a ghost, Streams like a cloud, man-shaped, from mountain peak, And cleaves to cairn and cromlech still." Thus, "Arthur's Round Table" is a term applied to enclo- sures of one sort or another in various parts of Great Britain, — for instance, to what is really the remains of a Roman amphi- theatre at Caerleon-on-Usk in Wales, — to a space surrounded by trenches on a hill at Cadbury in Somersetshire, and so on. This same Cadbury is identified locally with the Camelot of the French romances, the capital of Arthur's kingdom. In Cornwall, besides Tintagel, where the hero was born, and Camelford, where he fought his last battle with Mordred, we have King Arthur's Bed — a group of rocky hills — King Arthur's Cups and Saucers, rock basins in the slate of a promontory along the coast. But the localizations are especially numerous in Southern Scotland.72 Every visitor to Edinburgh will remember the hill called Arthur's Seat which rises to the south of the city, but the same name is met with elsewhere in the south of Scotland. Further north at Meigle is Guinevere's grave — the queen, according to local tra- dition, having been tied to wild horses and torn to pieces as a punishment for marital infidelity. These are only a few well- known examples which might be increased almost indefinitely. It is frequently taken for granted that this widespread Ar- thurian nomenclature furnishes indisputable proof of an equally widespread diffusion of early Arthurian traditions in Great Bri- tain, those traditions being embodied, as it is conjectured, in fan- tastic tales of the same general character as the Welsh Kulhwch and Olwen. But the validity of this assumption depends entirely on the antiquity of the localizations, and, in default of a critical work on the subject, no one can affirm that such localizations 13 On the basis of these frequent localizations, Skene, Four Ancient Books of Wales, I, 60, and Stuart-Glennie, p. LXI, have placed the scene of Arthur's last battle in Scotland at Camelon, on the south bank of the river, Carron, where there are remains of an old Roman town. W. H. Dickinson, pp. 82 f., also argues at length in favor of this view. All the conditions of sixth century Britain, however, point to the South as the scene of this last battle, if, indeed, it is not purely legendary. 74 Evolution of Arthurian Romance were not really suggested, in the majority of instances, by the chronicles and romances. They would occur particularly in the regions named above — partly, because it was known that the severest conflict between the Celts and the Anglo-Saxon invaders did take place in just those regions, and, partly, because people of Celtic descent would naturally feel the greatest interest in con- necting with objects in their own territory the names of these mythical characters of their race whom the chroniclers and ro- mancers had made world-famous. It is significant that the parts of Great Britain where the localizations in question are most nume- rous of all — viz., Northern England and Southern Scotland — were also just those parts where Arthurian romance as a literary genre took deepest root. It was here (in Northern England), for example, that the finest of the English romances, Sir Gaivain and the Green Knight and the alliterative Morte Arthure, were written, and the last of the species that was composed on British soil was a Scottish romance, Lancelot of the Laik. It will be observed, then, that the evidence of these place-names is inconclusive.73 2. Traditions concerning Arthur's death (journey to Avalon) and expected return. In spite of their enthusiastic patriotism, native Welsh autho- rities are agreed that Arthur was a late-comer in the traditions of Wales.74 Moreover, one of the most eminent of Celtic scholars 73 Zimuier, however, maintains, Gott. G. A. for June 10, 1$90, pp. 525 f., that the localization of Arthur's capital at Carlisle (Carduel), which occurs often in the French romances, is a genuine tradition inherited from an early period. 74 The testimony of the native Welsh authorities on this sub- ject is virtually unanimous. Cp., for instance, Thomas Stephens, The Literature of the Kyrnry, pp. 400 ff. (second ed., London, 1876). He regards (p. 406) Arthur as an Armorican creation. E. Anwyl, Zs. f celt Ph., I, 293 (1897). declares that "in the legends of Gwynedd [Northwest Wales] and Dyfed | Southwest Wales] he [Arthur] had no place whatever", and W. J. Gruffydd in his paper "The Mabinogion", Transactions of the Honorable Society of Cymmrodorion, Session of 1912—1913 (London, 1914), says (p. 31) that to Cornwall and Devon "must be given the credit of cradling the superb mythology of Arthur", and (pp. 32 f.) that "whether he appears in the well-ordered tales of Chretien de Troyes or in the chronicles of Geoffrey of Mon- Origin of the Lays and Romances 75 on the Continent, the late Professor Zimmer, even went so far as to deny that there was any popular tradition at all among the Welsh concerning Arthur's voyage to Avalon (the Celtic Elysium) and expected return.75 According to the German scholar, this was a purely Armorican (Breton) conception and became known to the Welsh only through the Bretons — after the appearance of Geoffrey's Historia — whose passionate attachment to the idea of the espoir breton (hope of the Britons), as it was called, is mouth and in others or in Malory, or, above all, in Welsh literature he is clearly no dweller in Wales or in the Gogledd" [i. e., the old Celtic territory in Northwest England and Southern Scotland]. Gruffydd avers (p. 32) that even in Kulhwch and Olwen Arthur is "no neces- sary part of the romance", and with reference to the same tale, J. Loth, Revue Celtique, XXXIV, 379 f. (1913), observes that, if Arthur plays a preponderant part in a Welsh story, this is a sign of late date. Gruffydd, op. cit., p. 35, asserts that Arthur's tables, mounds, etc., in Wales are all of late origin. The little that is found concerning Arthur in Welsh literature, he thinks, came in through Powis. The Anglo-Saxon victory at Deorham (in 577) cut off henceforth the Celts of Wales from those in Devon and Cornwall. Gruffydd suggests that this event prevented the Welsh from absorbing the Arthurian traditions which their Southern brethren had created. It seems significant that South Wales — the part of Wales nearest Cornwall — should, among all divisions of Wales, best pre- serve the tradition of Arthur and his last battle (Camlan). Cp. F. Lot's note, "La bataille de Camlan", Romania, XXX, 16ff. (1901). As Lot says, the people of Cornwall localized Camelford (twelve miles east of Bodmin) as the scene of the battle and the river, Camel, as the Cambula. In the Welsh Mabinogion and Triads we find also the Cornish localization of Arthur's residence at Kelliwic, which Lot, ibid., pp. 13 f., identifies with Bodmin (capital of Cornwall). 76 In his article, "Bretonische Elemente in der Arthursage des Gottfried von Monmouth", Zs. f. frz. Sj)r. u. LitL, XII1, 231 ff. (1891) — especially, pp. 238ff. In this same article Zimmer tried to draw a distinction between a historical Arthurian tradition in Great Britain (cp. Nennius) and a romantic tradition developed by the inhabitants of Brittany in the course of the generations which followed their separation from the actual scenes of his struggle and triumphs. But the distinction is a fanciful one and has found no favor with scholars. The popular imagination could have been just as active in full sight of Badon Hill or on the field of Arthur's last battle as in far-away Brittany. 76 Evolution of Arthurian Romance well attested by contemporaries, such as Wace7e and Alanus de Insulis. 77 The latter, for example, in his commentary on Geof- frey's Prophecies of Merlin (Book VII, of the Historia) avers that any one who in Brittany disputed the superstition concer- ning Arthur's return would be cursed and stoned. Geoffrey, him- self, indeed, has only a brief phrase concerning the matter: Mine ad sananda vulnera sua in insulam Avallonis advectus (XI, 2). But, interest in the subject having been awakened by this state- ment, according to Zimmer's view, the Welsh later on drew a fuller account from the Bretons, by way, doubtless, of the French chro- niclers and romancers. As a matter of fact, the evidence of Welsh belief in this most famous of the conceptions that relate to Arthur is surprising- ly meagre. Nennius, for instance, does not mention it, nor is it mentioned in the Mabinogion. In fact, only two texts have been cited that seem to furnish any evidence as to the currency of these conceptions in Wales: 1. a description of Wales in French octosyllabic verse which dates from the middle of the twelfth cen- tury and which is attached to Gaimar's chronicle in certain MSS. of that work. Describing the Norman invasion of Wales after the Conquest of England, the writer says that the Welsh recovered much of the ground which the invaders had taken from them and that they averred that they would recover it all in time through Arthur. 2. A Welsh poem of about the same date in the Black Book of Caermarthen. Here it is said in an enumeration of tombs: A tomb for March, a tomb for Gwythyr, A tomb for Gwgan of the red sword. A tomb for Arthur would be foolish. That is to say, of course, because he was still alive.79 76 Brut, 11. 13683ff., quoted in part, p. 57, above. Wace, with Geoffrey, VII, 3, in mind, says that, according to Merlin, Arthur's end would be doubtful. 77 For the words of Alanus in his Prophetia Anglic ana cp. San Marte, Sagen von Merlin, p. 55 (Halle, 1853), and Zimmer, op. cit., p. 240. They were written during the reign (1154 — 1189) of Henry II. 78 By F. Lot, Romania, XXVHI, 16 f. (1899). '" W. J. Gruffydd, who maintains, as we have seen above (p. 74 note), that Arthur was an alien to genuine Welsh tradition, suggests, Origin of the Lays and Romances 77 Does not Giraldus Cambrensis, however, the famous Welsh ecclesiastic of the late twelfth and early thirteenth centuries, tes- tify to a Welsh tradition of Arthur's expected return in the well- known passage of his Speculum Ecclesiae (II, 9)80 which de- p. 35 of the Transactions of the Honorable Society of the Oymm- rodorion, Session, 1912 — 1913 (London, 1914), that the real meaning' of the last line is: "His exploits were not performed in Welsh terri- tory; his grave was unknown, simply because he was not buried in Wales." But even if Gruff ydd's general theory about Arthur and Wales is correct, Lot's interpretation of the line strikes me as pre- ferable. Arthur's last battle (Camlann) is mentioned often in Welsh texts. See Index to Loth's Mabinogion under Camlann. The text above, however, is the only one in the Welsh language that even hints at the conception of Arthur's translation to Avalon, and there are none in Irish — for Irish literature is virtually entirely ignorant of Arthur. In my edition of the Mort Artu, pp. 300 f., I have summarized and discussed the suggestions that have been made as to Irish parallels to the story of Arthur: Death of Cuchullin (cp. Eleanor Hull, The Cuchullin Saga, pp. XXVHIff., London, 1898), the legend of Mongan (cp. A. Nutt, The Voyage of Bran, II, 23, London, 1897), and the legend of Finn (cp. A. Nutt, Revue Celtique, XII, 190). There is no real similarity, however, in any of the cases — least of all, in regard to the end of the respective heroes. J. F. Campbell, Popular Tales of the West Highlands, IV, 240ff. (London, 1893) was better in- spired, when he cited as a parallel the Gaelic legend of Diarmaid and Grainne, a cognate, as we shall see below, of the Tristan-Iseult story. The parallel is closer, if we take (with Geoffrey) Mordred as the adulterer, instead of Lancelot, as Campbell does. Here, as in the stories of Tristan and Iseult and Arthur and Guinevere, we have the faithless wife who commits adultery with her husband's nephew. On the other hand, the legend in question does not end with an Other- world voyage. Altogether, there is probably no historical connection between the stories. The idea that only three men survived Arthur's last battle is known to Welsh literature. Cp. the triad in Loth's Mabinogion*, II, 290, and Kulhwch and Olwen, op. cit., I, 270. For a discussion of these passages see Bruce, Mort Artu, pp. 295 f. and on Arthurs mythical end, in general, cp. pp. 298 ff: Cp., too, Miss Paton, op. cit., pp. 25 ff. 80 Giraldus, loc. cit., erroneously states that it was in Henry ITs reign, but Henry died in 1189. It occurred really in the reign of Richard I. Cp. J. S. Brewer's edition of Giraldus's works for the 78 Evolution of Arthurian Romance scribes the pretended exhumation of Arthur and Guinevere at Glas- tonbury in 1191? To increase the importance of their monastery, which, according to their impudent assertions, was founded by King Arthur, the monks concocted this solemn farce of the ex- humation, an account of which Giraldus received from an eye- witness. The bodies were found buried sixteen feet deep in the ground and when a monk seized a yellow lock of Guinevere's hair, it is said to have vanished away. The Welsh writer con- trasts with the evidence of Arthur's real death which this affair supplied the absurd tales which are propagated by fabulosi Bri- tones et eorum cantatores; for the latter were accustomed to relate, he says, how a certain fantastic goddess, Morganis — Dea quaedam phantastica — took Arthur to Avalon to be healed of his wounds and how, when they are cured, he will return to reign over the Britons. Giraldus comments still further on the likeness of this belief to that of the Jews in the coming of the Messiah. Everything here turns on the disputed meaning of Britones. Is Giraldus referring to Armoricans or Welshmen? Despite con- tentions to the contrary,81 Welsh writers in the Middle Ages do occasionally, though not often, use this term of contemporary Welshmen, as well as of the whole race of Britons combined — both continental and insular.82 Consequently, the passage is open to varying constructions and throws no decisive light on the sub- Rolls Series, IV, 48 ff . (London, 1873). A few sentences which are unintelligible in the Latin text, owing to mutilation of the unique MS. of the Speculum, have been preserved in the Welsh version of Giral- dus's account edited by Timothy Lewis and J. D. Bruce, "The pre- tended exhumation of Arthur and Guinevere," Revue Celtique, XXXIII, 432 ff. (1912). The inscription on the cross over the tomb in Giral- dus, p. 50, speaks of Guinevere as Arthur's second wife. In a Welsh triad, according to Rhys, Arthurian Legend, pp. 35 ff., he is said to have had three wives, all named Gwenhwyvar (Guinevere). Nowhere else in Arthurian literature, however, is he given more than one wife and, besides, as we have seen above, p. 44, note 14, Loth does not translate the Welsh in the same manner as Rhys. According to his translation, the triad does not speak of Arthur's three wives, but of the three principal ladies at his court. 81 Zimmer, Zs. f. frz. Spr. u. LitL, XII \ 241ff. (1890). S3 Op. Lot's articles cited p. 60, note, 44, above. Origin of the Lays and Romances 79 ject of our inquiry. The form of Morgan's name in G-iraldus, Morganis (instead of Morgana) seems determined, it is true, by the French nominative form, Mor gains,83 but even this detail cannot be taken as settling the whole question; for, in forming the Latin name, Giraldus might well have been influenced by the most popular works in which the character had appeared (the French romances) even though he knew, also, from Welsh sources of Arthur's exitus dubius.8* Morgan85 and Avalon86 being, of course, inseparable from 83 Cp. Brugger, op. tit., XX \ 100 (1898). Ibid., pp. 97 ff., lie presents the strongest argument that has been offered for the in- terpretation of Britones as Armoricans in this passage. 84 "Doubtful exit" — i. e. from life. 86 For ttye name which, according to J. Rhys, Arthurian Legends, 348 1., must have been Morgen (as actually in Geoffrey's Vita Mer- lini) in mediaeval Welsh, cp. L. A. Paton, Studies in the Fairy Mythology of Arthurian Romance, pp. 9 ff . : Radcliffe College Mono- graphs, no. 13 (Boston, 1903), and J. D. Bruce, MLN, XXVI, 67' note 16, where the opinions of the leading Celticists on the subject are assembled — in the latter, also, similar observations on the common Welsh man's name, Morgan. Miss Paton, loc. tit., proposed to iden- tify Morgan with Morrigan, an Irish battle-goddess, but this unlikely suggestion has won no adherents. — Rhys, op. tit., p. 22, interprets Welsh Morgen as etymologically identical with Muirgen, meaning "sea-born", which actually occurs as a woman's name in Irish. Morgan (Morgen) is first mentioned in Geoffrey's Vita Merlini, 11. 918 ff. (composed probably in 1149) — and here as the person who is to heal Arthur in the Isle of Apples (Celtic Elysium). She is next mentioned in the Roman de Troie, 1. 8024 (Constans' edition, I, 434), — composed about 1160 — as a fairy who hated Hector, because he did not return her love. The name here, it is true, is in# most MSS. disguised by mutilation, as we have seen above, p. 33, note 73, that it was also in Layamon. Chretien, in his Erec, 1. 1957, com- posed about 1165), mentions her as the mistress of Guigomar, lord of Avalon, and, 11. 4216ff., as Arthur's sister and as highly skilled in healing. The two passages combined betray a knowledge of the part which she played in the story of Arthur's end. The poet, him- self, however, doubtless, invented the idea that she was Arthur's sister, which was afterwards adopted, among others, by Robert de Boron in his Merlin. To be sure, it must be confessed that as early as about 1170, in the Draco Normannicus (a Latin metrical para- phrase of Geoffrey's Historia), the author (Etienne, a monk of Bee, 80 Evolution of Arthurian Romance the myth of Arthur's end, the problem of their origin is simply a part of the general problem of the origin of that myth. it seems), also, makes Morgan (Morgana) Arthur's sister. Cp. Fletcher's Arthurian Material, pp. 145 f. This is more likely, however, to be due to Chretien's influence, direct or indirect, than to a misunder- standing of the bearing of the word, sorores (Morgan and her fairy companions), in the Vita Merlini, 1. 919, as Fletcher surmises. Miss Paton's hypothesis, pp. 29, 33, that the conception of Ar- thur's translation to Avalon (in company with Morgan) is merely a fairy-mistress story transformed has no support whatever in the in- numerable Arthurian texts. Cp. Bruce, Mort Artuf p. 300. Miss Paton, op. cit., has followed Morgan through the romances and her collection of material on this character and on the fairy mythology of the Arthurian romances in general is invaluable, but her work suffers from the fundamental error of assuming virtually always that the romancers in what they say concerning the great fairy are drawing on a body of popular tradition concerning her. As a matter of fact, their draughts on tradition are infinitesimal — they drew on their imaginations, instead. In the course (p. 326) of his article "Morgue la Fee et Morgan Tud", Romania, XXVIII, 321 ff. (1899), F. Lot suggests that Geoffrey based his famous description of Arthur's end on an Irish tale, which probably reached him in a Welsh version. The grounds of the sug- gestion, however, are insufficient. The name, "Morgan Tud", occurs only in the Welsh "G-eraint and Enid" (Loth, Mabinogion2, II 143, 174) — as the name of a wonderful male physician. Zimmer, Foerster's edition of Chretien's Erec, pp. XXVII ff., argued that this otherwise unknown person was a blundering creation of the Welsh writer, which was due to his ignorance of the fairy- queen (Morgue, Morgain) of his French original (Chretien's Erec). For a refutation of this hypothesis, cp. Miss Paton, op. cit, pp. 259 ff., and J. Loth, Contributions a V etude des romans de la Table Ronde (section entitled, Morgan Tut), pp. 5 Iff. (Paris, 1912). For a re- futation of the opposite theory, according to which the change of Morgan's sex was due to a misunderstanding on the part of Chretien of what was originally a Welsh source, see Miss Paton, loc. cit. This latter theory was proposed by J. Loth, Revue Celtique, XIII, 496 f. (1892) and defended by F. Lot in his article just named. Miss Paton, herself, pp. 263 ff., urged that Morgan Tud of the Welsh text was merely a corruption of the Old Welsh proper name, Morgetiud (and variants) — modern Meredith. This seems to me the most probable explanation of all, but subsequently, Loth, Contributions, pp. 56 ff., has connected the Welsh Tut (Tud) with Old Irish tuath, which Origin of the Lays and Romances 81 In the face of the scanty and, for the most part, ambiguous evidence, one is compelled to fall back on general considerations in order to reach a final decision in the case. Looking at the question from this point of view, we are forced to conclude that the above-mentioned conception must have already prevailed among the Bretons before their exodus from Great Britain; for there would seem to be no reason why such a conception should have sprung up among them, after they had left the land with which Arthur's, glory was associated, and, besides, we know that their kinsmen of Cornwall, who remained behind, were, at the begin- ning of the twelfth century, cherishing the faith in the hero's sometimes means "magician". He supposes that this was translated by an Anglo-Norman "Morgan le Fe or le Fed" (fee being masculine in the Norman dialect) and that Chretien misconstrued this as feminine. Cp. too Loth's recent article, Revue Celtique, (1920). To the present writer, however, all of this is unconvincing. Geoffrey of Monmouth, Benoit de St. Maure and the author of Layamon's original, not to look for others, all knew of Morgan as a supernatural female and there is no need of hazardous speculations, therefore, to explain how Chretien conceived of her similarly. *e For a summary of discussions concerning Avalon (Avallon) or Avallach and the literature of the subject, see W. E. Meads Se- lections from Morte Darthur, pp. 31 6 f. (1897) and Bruce, Mort Artu, p. 273. Zimmer, Zs. f. frz. Spr. u. Litt., XII1, 238ff. (1890) andBrugger, ibid., XXVII1, 98ff. (1898) contended that Avalon, like everything relating to Arthur's translation, was unknown to the Welsh. On this subject, see still further pp. 74 ff., above. Lot in his "Glastonbury et Avalon", Romania, XXVII, 553ff. (1897) attacked Zimmer's thesis, but, himself, advanced the improbable view that Avalon was originally the name of a man (or god) and not of a place and that the latter conception is due to French misunderstandings of the original Celtic form of the name, Avalloc. — F. M. Warren points out in "The Island of Avalon", MLN, XIV, 93f. (1899) that Avalon is already known to the author of Le Couronnement de Louis, who, 11. 1796, 1827, uses the expression, "tot Tor d'Avalon", as a hyper- bole for riches. This poem was written in the Isle de France about the middle of the twelfth century. Warren regards it as earlier than Wace's Brut. This cannot be determined definitely, but the phrase at least shows no influence from the Brut. For the identification of Avalon with Glastonbury, cp. Part II, below. 82 Evolution of Arthurian Romance return with equal fervor — a faith which, of course, they must have inherited from their ancestors of something like six cen- turies before. On the other hand, Arthur was certainly, also, a character in Welsh saga in the first part of the ninth century (when Nennius compiled his chronicle), near the end of the tenth (when the Annates Cambriae were compiled), and, again, round about 1100 (when Kulhwch and Olwen was composed). There are even numerous allusions, as we have seen,87 in Welsh texts to his last battle, although none to his voyage to Avalon and expected return, save in the one obscure line which has been given above. It seems strange, indeed, that, having been adopted, as we know from the sources just indicated, as one of their national heroes by this branch of the Celts, also — although, perhaps, with less ardor and universality than was the case with his nearest kinsmen (the Celts of Devon and Cornwall and Brittany) — the conception of his apotheosis should have made no impression on them. Never- theless, in view of the evidence that confronts us, we can only conclude that the two passages cited above (one French and one Welsh) do not express a general Welsh faith. The two writers concerned doubtless knew of the "hope of the Britons" that pre- vailed in Brittany and in the old Celtic territory of Southwest. England and adopted it for the Welsh, although in Wales it was no part of general popular tradition-88 The source, then, of the fine description of Arthur's trans- lation to Avalon in Geoffrey of Monmouth's Vita Merlini, of which we shall hear more anon, was, no doubt, also Armorican.89 3. The Origin of the Round Table and Some Other Questions of Celtic Origins. 1. The Round Table: (a) According to Wace's Brut (11. 9994 ff.)90 the Round Table was instituted to prevent quarrels 87 Cp. p. 77, note 79 above, 88 As far as the French writer is concerned, he may have simply invented the incident, or, if this is not so, the Welshmen, referred to, may not have spoken in all seriousness. 89 It is to be remembered that, apart from general knowledge of Breton matters, Geoffrey had special sources of information on this subject, if his father was a Breton, as seems most likely. 80 Quoted below, in the chapter on Merlin. Origin of the Lays and Romances 83 in regard to precedence among Arthur's barons. Now we hear of similar brawls at the royal board in the Irish sagas91 — for example, at Conchobar's table in Emain Macha, according to the Tale of MacDatho's Pig,92 and at Bricriu's, according to Bri- criu's Feast.93 The following is an incident of this character which took place at Bricriu's table. He prepared a great feast for Conchobar, the renowned king of Irish saga, and the nobles of Ulster. Knowing that he would not be allowed to share in the feast because of his evil tongue, he caused to be built for him- self a little chamber or balcony from which he could see all that went on. True to his character, he determines to stir up strife among his guests, and when they begin to arrive, he urges each secretly to demand at the banquet "the hero's portion", as it is called in the Irish sagas — that is, the biggest portion, which was supposed to go to the worthiest warrior at the table. In this way he approaches successively Loegaire, Conall, and Cuchulinn, the last-named being, so to speak, the Achilles of the Old Irish epic. The feast commences and Bricriu retires. Soon, however, the cha- rioteers of the three champions get up and each demands the hero's portion for his master. A dispute arises over this, and arms begin to clash, but it is agreed after a while to divide the portion equally among the three claimants and to refer to King Ailill of Connaught the decision as to who is the greatest champion. Quiet is restored for the time being, but the dispute again breaks out and Ailill is urged to render his decision. To test the warriors' abi- lities he puts them through a series of adventures against strange 91 Zimmer first called attention to these Irish parallels to the Round Table, in the Gottingische G. A. for June 10, 1890r pp. 518 ff. He says, too, p. 518, note 1, of the Pentecost festivals with which the Arthurian romances so often begin: "Die Maiversammlungen der Ar- thursage sind durchaus im keltischen Altertum begriindet". He cites the May festivals held at Tara, according to Irish saga, to which the overlord of Ireland summoned all his men. Cp. on the subject of this note, also, Brown, "The Round Table before Wace", pp. 193ff., in- cluding notes. 92 D'Arbois de Jubainville, Cours de UttSrature celtique, V, 66 ff. 93 Ibid., pp. 120ff. The Irish title is Fled Bricrend. Bricriu means "Poison Tongue". 84 Evolution of Arthurian Romance monsters. Cuchullin fares best in these trials, but Ailill, who is apparently afraid of offending the rest, gives each one privately a goblet as a token of superiority. Cuehulinn's, however, is of gold, whereas the others are of inferior metals. So when subsequently at a feast the old dispute over precedence begins anew and each presents his goblet, much bitterness is produced. The claimants are again sent forth now on a fresh series of adventures to test their worthiness, but at this point the MS. breaks off. (b) Like these episodes of the Irish sagas is the account which Layamon gives (11. 22736ff.) of how the Round Table came to be instituted.94 At a great feast on Yule-day which Arthur gave, says Layamon, a sanguinary quarrel sprang up among the guests, "because each, on account of his high lineage, wished to be with- in'' (whatever that may mean). Several had lost their lives be- fore the king succeeded in quelling it. Shortly after, when the king was in Cornwall, a smith there offered to make him a table at which 1600 and more people might sit, "all around about so that none be left out without and within, man against man". More- over, the king could carry it about with him anywhere. In four weeks' time the work was completed and thereafter all was peace and fraternity at Arthur's feasts. This passage has nothing to correspond to it in the extant text of Wace's Brut, but, in view of the Irish parallels, must be accepted as undoubtedly derived ultimately from Celtic tradition. It has been taken95 as a proof that Layamon drew directly from 94 The similarity was pointed out by A. C. L. Brown, "The Round Table before Wace", Harvard Studies and Notes in Philology and Literature, VII, 183ff. (1900). 96 By A. C. L. Brown, op. cit, and "Welsh Traditions in Laya- mons Brut", MPh., I, 95 ff. (1903). In his review of Brown's "Round Table before Wace", in Romania, XXlX, 634 (1900), G. Paris, however, differs from Brown, inasmuch as he supposes that the Welsh traditions in question reached Layamon not directly, but through the English. This is, of course, a corollary of Paris's very questionable theory con- cerning the Anglo-Norman origin of the Arthurian romances. In the Zs. f. frz. Spr. u. Litt, XXIX3, 247, note 11 (1905), Brugger has attacked, in particular, Brown's assumption that Layamon derived his story of the Round Table from Wales. He points out very aptly that in no writings of Welsh authorship (Mabinogion, Giraldus Origin of the Lays and Romances 85 the oral traditions of his neighbors, the Welsh. We have seen, however, that the English writer was, in all probability, wholly dependent on a French source (an expansion of Wace) for his so-called additions, so that the inference is unwarranted and we are left in the same state of doubt as to the Welsh or Armorican provenance of this incident as of the incidents of the Arthurian romances generally — with the usual balance of probabilities, how- ever, in favour of the latter. (c) It has also been proposed96 to derive the Round Table from some Celtic feast, like the Beltane or May-day feasts of the Highlands of Scotland — a spring festival which descends no doubt from pagan times. According to an account of such a feast at Callender, recorded in the nineteenth century, the boys in that neighborhood on May-day cut a table in the green sod by digging around it a trench of sufficient circumference to hold the entire Cambrensis etc.) is the Round Table mentioned. He suggests (pp. 245 ff.) that the name "Round Table" arose in Armorica and that Wace pro- bably invented the idea that it was made round in order to forestall quarrels as to precedence. On the other hand, according to La}'amon (or his source), who was modifying Wace, the purpose was to put an end to such quarrels. Like Ten Brink and Mott, Brugger believes, as against Brown, that the story about the Round Table which is told in Layamon originally was not connected with Arthur at all. This may well be so, bnt I see no reason to imagine ihat this connection was first made by the author of the expanded Wace (Layamon's French original) rather than in the sources (probably Armorican by origin) on which he drew. Similarly with the idea, which we find in Wace, that quarrels over precedence caused a round table to be constructed. In this same article, p. 246, note 9, Brngger speaks of a simi- larity between the Round Table and the table of the Last Supper in the Gospels, which, he says, is not purely accidental, for the latter, too, is "ein Uberrest altheidnischen Opferbrauches". But wherever any snch similarity exists, it is under the influence of the Grail ro- mances. The Grail Table was, of course, modeled by Robert de Boron directly after the table of the Last Supper. 90 By Lewis F. Mott, "The Round Table". PMLA, XX, 231 ff. (1905). Brugger, Zs. /. frz. Spr. u. Litt, XXIX2, 238ff. (1906), in reviewing this article, expresses substantial approval of its results. I agree, however, with F. Lot, Etude sur le Lancelot en prose, p. 245, note 5, in rejecting it. 86 Evolution of Arthurian Romance company. Then after making a fire in the circle and cooking and eating certain prescribed things they put bits of cake into a bonnet and blindfolded draw them out. Whoever draws a certain black bit is to be sacrificed to Baal to induce him to render productive the year which has just begun. The sacrifice is now a joke, but, originally an actual human sacrifice was, very likely, involved in the ceremony. There is a far cry, however, from this custom observed in a limited district of the Highlands to the Round Table of Ar- thurian romance — even in its most fantastic form, that of Laya- mon's Brut. Besides, it is a serious weakness of the theory that the part which disputes over precedence play in accounting for the form of the table is unjustifiably treated as an afterthought. (d) Miss J. L. Weston97 connects the Round Table with some hypothetical turning table of Celtic tradition of mythical significance. The only evidence for the Round Table as such a turning table — if evidence it can be called — is a single line in Beroul's Tristan, 1. 3384 (end of the twelfth century), where, in replying to a messenger of Iseult's who is inquiring about the king, a shepherd says: "Sire", fait il, "il sit an dois. Ja verroiz la Table Reonde, Qui tornoie come le monde: Sa mesnie sit anviron." From the eighth century Irish text called The Voyage of Maelduin down, we have turning castles in Celtic tradition; but no mention is made of equally marvelous turning tables in this same tradition. Beroul's words may possibly refer to the vicissi- tudes of life to which the company gathered about the board, like the rest of the world, are subject, or, if construed literally, they may express a passing fancy of this particular poet — but, whatever the meaning of this obscure line may be, it does not justify us in jumping with Miss Weston to the conclusion that the Round Table is connected with some supposed solar ritual. Whatever we may think of the theories just presented, there 97 See her article "A hitherto unconsidered aspect of the Round Table", iu the Melanges offerts a M. Maurice Wilmotte (Paris, 1910). Origin of the Lays and Romances 87 is hardly room for doubt that Arthur's Round Table belongs to the paraphernalia which attached itself to him in Celtic tradition. We cannot regard the legends relating to it as a mere development out of the old stories concerning Charlemagne and his twelve peers, as has been suggested. The Irish parallels, moreover, render it probable that the conception obtained, also, among the insular Celts,98 as we know from Wace that it did among the Bretons. Perhaps, after all, in seeking for an explanation of this most famous of tables and the customs which were connected with it, we need go no further than the account which is given of Celtic feasts by the Greek philosopher and traveller, Posidonius, who lived in the first century before Christ, his observations being made probably in Southern Gaul. He tells us" that at their feasts the Celts sit in a circle and that the bravest sits in the middle like the leader of a chorus. Moreover, primitive Celtic houses were often circular. It is quite possible that these actual customs may have determined the shape of the celebrated table of romance. 2. Excalibur (Caliburnus in Geoffrey's Historia, IX, 4), which is Arthur's sword in Geoffrey and commonly in the ro- mances,100 is certainly identical with the sword Caladbolg of the Irish prose epic Tain bo Cualnge (The Cattle-Raid of Cooley),101 which is there the property of Fergus, the fugitive ruler of Ulster. It is said to have become of the size of the rainbow, whenever any 98 99 Zimmer, Gott. G. A., for June 10, 1890, p. 525, denies this. Quoted by A. C. L. Brown, "The Round Table before Wace", p. 195, note 3, from Carl Mtiller, Fragmenta Historica Graecorum, in, 260 (Paris, 1849). 100 In the relatively late Vulgate ilfer/m-continuation Arthur presents it to young Gawain. Cp. Sommer, Vulgate Version of the Arthurian Romances, II, 253. 101 The identification was first made by Zimmer, Gott. G. A., for June 10, 1890, pp. 516 f. For Caladbolg, cp. Miss Winifred Faraday's translation of the Irish epic: The Cattle-Raid of Cualnge, London (Grimm Library, no. 16), 1904. Loth maintains, Revue Celtique, XIII, 495, that French Calibor (= Caladbolg) is drawn from a written Welsh form. See, too, F. Lot, Romania, XXV, 1 f., who contends that Geoffrey's Caliburnus does not necessarily come from the Breton. 88 Evolution of Arthurian Romance one struck with it. Fergus cuts off the tops of three hills with this sword. Moreover, like Excalibur, it was made in fairy-land. 3. Miscellaneous Celtic folk-tale motifs, such as are found, for example, in the Tristan romances.102 With regard to such motifs, it should be premised that even in instances where insular records appear to offer parallels to incidents in Arthurian ro- mance, it may be that the same stories were current in Brittany and transmitted thence to the French writers — only the com- plete absence of all Breton records from the Middle Ages leaves us without the means of control. We may select for notice here, particularly: (a) the motif of the Turning Castle mentioned above, which is found in La Mule sans Frein, 11. 440ff., the Middle High German Diu Krone, 11. 12 95 Iff., and other romances, both metrical and prose, and is, likewise, familiar to Irish saga, Voyage of Maelduin, etc.;103 (b) the motif of the Beheading Game:104 A strange visitor turns up at court and offers to submit himself to decapitation at the hands of a knight, provided that at the expiration of a given period the knight will, in turn, subject him- self to the same test of courage. The knight naturally accepts this as a very easy test, but to his surprise the stranger picks up his head after decapitation and at the appointed time is ready to return the blow. This incident is found in several romances — earliest, perhaps, in the so-called Livre de Caradoc (Livre de Karados),10[> which is an interpolation in the first continuation of Chretien's Pe?xeval. It is even better known, however, through the Middle English romance, Sir Gawain and the Green Knight, 102 Cp. especially Gertrude Schoepperle: Tristan and Isolt, II, 267ff. (Frankfort and London, 1913). 103 Cp. Gideon Huet, "Le chateau tournant dans la Suite du Merlin", Romania, XL, 235 ff. (1911). Cp., too, W. E. Sypherd, Studies in Chaucer's House of Fame, pp. 114ff. (Publications of the Chaucer Society, London, 1907), where numerous Celtic examples of Whirling Houses are given, but, also, some from other sources. 104 On this motif both in the romances and in folk-tales, cp. G. L. Kittredge, A Study of Gawain and the Green Knight, pp. 9 If., 147ff. (Cambridge, Mass., 1916). 106 Cp. Potvin's Perceval li Gallois, III, 117—221. For the episode in question, cp. ibid., pp. 125 ff. Origin of the Lays and Romances 89 which certainly goes back substantially to a lost French original. But we have this same incident also in the Irish saga, named above, Bricriu's Feast,106 of which a manuscript from circa 1100 has been preserved, but which is itself much older. One of the tests of courage which the three champions Loegaire, Conall, and Cuchullin, have to undergo in that tale is an adventure of this nature with the giant Uath Mac Imomain (Fright, Son of Great Fear). Cuchulinn, however, alone is equal to the test. When the giant has been decapitated, he clasps his head to his breast and jumps into the lake. The next day he returns and when Cuchulinn exhibits no fear even when stretched out and ready to receive the fatal blow, his enemy spares him, at the same time, declaring him to be the first of Irish warriors. 4. The individual episode in the romances the insular Celtic — apparently Scotch — origin of which appears best assured is one that also occurs in the Livre de Caradoc,101 just mentioned — viz. the episode of Carados (Caradoc, Karados) and the serpent, which occurs also as a Gaelic tale108 and was still in oral circu- 108 Cp. George Henderson, Fled Bricrend, pp. 116ff. (London, 1899). Kittredge, op. cit., pp. 10 ff., gives an analysis of the episode. 107 Potvin, op. cit., II, 191 ff. Secondary to this version is the one in the Roman de Renart le Contrefait. See Miss C. A. Harper, MLN, XIII, 422 f. 108 Cp. J. F. Campbell's Popular Tales of the West Highlands, IV, Introduction, pp. XCVf. Campbell took it down from a travelling tinker. "The Queen of Scotland", no. 301 in Child's collection of ballads, is, also, a variant of this Carados story. It was Miss C. A. Harper, "Carados and the Serpent", XIII, 417 ff. (1898), who first observed the resemblance of the episode in the Livre de Caradoc (Karados} to the Highland tale. She discusses there the relations of all the four versions which I have named. G. Paris, "Caradoc et le Serpent", Romania, XXVIII, 214ff. (1899), following up Miss Harper's study, pointed out that this story was also a part of Welsh* tradition in the Middle Ages, for a know- ledge of it is implied in a Welsh triad (Loth, Mabinogion2, II, 284 f.) on three famous chaste young women of Britain. Here the heroine is called, to be sure, not Guinor (in variant forms), as in the Livre de Karados, but Teg an Eur on (Eurvron), i. e. Tegan of the Golden Breast. Assuming that Caradoc Brechbras belonged to Armorican tradition, Paris concluded that the tale reached Wales from Brittany. 90 Evolution of Arthurian Romance lation in the nineteenth century in the Highlands of Scotland. The essential features of the story as we find it in the French are as follows: Carados (Caradoc) is the son of the enchanter Eliaures and of Queen Ysaune who has been unfaithful to her husband. After he has grown up, he learns from his true father the secret of his birth, but he is indignant when he hears the story and repeats it to the king, whom he has hitherto supposed to be his father. The king then shuts Ysaune up in a tower, but her lover still manages to visit her, until they are detected by their son, who again informs on them. Ysaune appeals to Eliaures, who offers to punish Carados in this way: He will create a horrible serpent and shut him up in her cupboard. When Carados visits her, she is to send him there immediately for her mirror. As soon as he puts in his hand, the serpent will wind about his arm. His flesh will then waste away and in two years he will die. The mother agrees to this and carries out her part in the plot. The serpent winds about the arm of Carados and his mother Ysaune then tells him that he has got what he deserves and that the best thing for him to do is to go forth and repent of his sins during the two years of life left to him. Carados takes the advice and goes into an abbey. Here he is discovered after a long search by his friend Cador of Cornwall, who wishes to kill the serpent, but cannot, since his friend's life is bound up with that of the creature. Cador, however, induces Carados to go with him to Ysaune's tower, to see if something cannot be done for him, and, owing to his repro- aches, she consults with her paramour, who prescribes the follow- ing procedure by which the unhappy Carados may be released from the serpent: Only a beautiful, well-born maiden who loves Carados loyally can effect the release. This maiden must pre- pare two caldrons and fill one with milk and the other with the sourest wine she can find. She must get into the caldron of miLk The story, however, he believed, was ultimately Irish and had nothing to do with Caradoc. In "Caradoc et Saint Paterne", Romania, XXVIII, 568 ff. (1899), however, F. Lot shows that Caradoc was really a hero of the Britons of Strathclyde and Cumberland, that the tale in question was Scotch, not Irish in origin, and passed into Wales by way of Northwest England, not Brittany. Origin of the Lays and Romances 91 and Carados must get into the caldron of wine. Then she must show her breast over the edge of the caldron and pray to God to cure Carados immediately. The serpent, disliking the wine and tempted by the sweet milk, will then leave Carados and seize her breast. On going home, Cador learns that his own sister loves Carados, and when she hears of the condition of his release, she agrees to sacrifice herself for her lover. Cador, however, is to try to kill the serpent as he passes from the one to the other. The serpent, however, actually seizes hold of the girl, and Cador, to accomplish the release, has to cut off a part of the breast to which it has attached itself. He then kills it between the caldrons and Carados marries the girl. Later on, after a number of strange adventures, Carados obtains from a knight named Alar- din del Lac a shield-buckle which has a wonderful power of heal- ing wounds. If a knight has lost half of his nose, a touch of this buckle will make a golden nose just like the one he had before. So with any other part of the body. Carados takes the buckle home and touches his wife's breast with it. She is at once cured, but henceforth has a golden breast. She is, later on, the only lady at Arthur's court who can stand the test of chastity. One may mention in conclusion that in this story Carados's arm was said to have always remained smaller after this experience with the serpent, so that he was known henceforth as Carados Brisie Braz (Carados with the broken arm) or Briebraz (short' armed). This rests on a curious misunderstanding of the Celtic epithet of the hero — namely, Breichbras (breich-&rm, bras- strong), which really means virtually just the opposite of what the French romancers (including the author of the Livre de Cara- doc), being misled by the similarity to words of their own language, imagined.109 109 It is a delicate question, however, whether F. Lot, op. ciL, pp. 222 ff., is justified in assnining that the retention of B as the initial sound of this epithet proves that the French authors used written sources, the b's in their language, as he declares, being pronounced like a v from the eighth century on, so that, if only heard, the French would have written down v, b. Confusion in catching the spoken name is too common to admit of a confident decision in this case. Cp. too, W. Meyer-Liibke, Zs. f. frz. Spr. u. Litt, XLIV1, 164, on the subject. 92 Evolution of Arthurian Romance 5. Arthur himself is virtually unknown to mediaeval Irish lite- rature of purely native origin,110 but, as will have been observed 1 10 In his article, "Eine verschollene Artursage", pp. 63 ff. of the Festschrift fur Ernst Windisch (1914), Kuno Meyer calls attention to an Aigidecht Artuir (= Entertainment of Arthur) in a list of Irish sagas in the Book of Leinster (twelfth century). Meyer conjectures on the analogy of the Middle Irish ballad, Find and the Phantoms (Book of Leinster), that the lost saga described how Arthur, whilst hunting a boar, came to a haunted house and had strange adventures with the spirits there. In the Revue Celtique, X, 185ff. (1889), Max Nettlau published from Stowe MS. 992 (Br. Museum) some excerpts from an Irish ro- mance concerning the Holy Grail — so, too, F. N. Robinson, Zs. f celt. Ph., IV, 381 ff. (1903) fragments from a MS. in the Franciscan library at Dublin. In both cases, however, the Irish is not original, but based on the Old French prose romance, Queste del Saint Graal. R. A. Stewart Macalister has published from MS. Egerton, 128 (B. M.), dated 1748, The Story of the Crop-Eared Dog and The Story of Eagle-Boy (Irish Texts Society, London, 1908). He entitles them Two Irish Arthurian Romances, and, as a matter of fact, King Arthur appears in both, and the principal character in the first bears the name of Sir Galahad. Nevertheless, Miss Weston, Folk-Lore, XX, 36 Iff. (1909), very properly questions their right to be called Ar- thurian. The stories belong to the "Wonder- Voyage" type of tale, as their editor, himself, remarks, and the mere adoption of Arthurian names does not render them Arthurian. A. C. L. Brown, MPh., IX, 120 ff. (1911) has expressed an opinion similar to Miss Weston's. The first of the two stories is probably based on a French original, the second certainly. Cp. T. P. Cross, MPh., X, 229, note 1 (1913). In his " Arthurian Motifs in Ghadelic Literature", Miscellany presented to Kuno Meyer, edited by Osborn Bergen and Carl Mar- strander, pp. 18 ff. (Halle, 1912), George Henderson calls attention to another "Wonder Voyage" tale of which Lancelot and Galahad are the heroes, preserved in Rawlinson MS. B. 512 (Bodleian Library). This, however, is as little Arthurian as the two stories published by Macalister. Henderson observes that, apart from fragments about the Grail, Arthurian motifs may be traced in 1. Eachdoaidh An Ama- dain Mhoir (Tale of the Great Fool). 2. Am Brat (The Cloak). 3. Am Br on Binn (The Melodious Sorrou). These are, however, all late and manifestly derived from the French romances. No. 3 is a Gaelic ballad which Henderson, himself, took down at Dalibrog, South Uist, in 1892, from oral recitation. The King of Britain dreams of a fair lady and falls in love with his vision of her. Gallomhai Origin of the Lays and Romances 93 from the preceding pages, this is not true of many of the motifs that occur in the Arthurian romances. We have seen above that mediaeval Breton literature is practically non-existent and that the bulk of extant works of the imagination in Welsh for the same period is not very large. On the other hand, an enormous quan- tity of Irish folk-tales and sagas has been preserved to us in MSS. of the eleventh and following centuries and of this body of narra- tive literature much was actually composed at a still earlier date. These sagas, etc, reflect, moreover, a more primitive state of society than the extant literature of any other branch of the European races, and hence they offer a store of material for the folk-lorist which is without parallel, as far as the Occident is con- cerned. Under these circumstances, it is natural that students of the romances who approach the subject from the folk-lorist's point of view should so frequently seek in Irish tradition the origin of the tales that make up the plots of the French romances. In cases where the theory of Irish origins is justified, it would seem most probable that the stories in question reached France by way of Wales and England. Nevertheless, difficult as navigation be- tween Ireland and France might seem to be under mediaeval con- ditions, direct intercourse between the two countries was not un- known in the Middle Ages, and the late Professor Zimmer has even striven to prove that the usual trade-route between Ireland and the Continent in those centuries was across the sea from the Irish coast to the mouth of the Loire.111 Inasmuch as all records (Gawain) traverses the sea for weeks to find her. A "big man" (i. e. giant) keeps her captive in a tower. In collusion with Gawain, she lulls him to sleep with her harp. Gawain then cuts off the giant's head and they escape. In general, it should be said that Henderson was ignorant of Arthurian discussions and his observations on the sub- ject are utterly confused and without value. For evidence regarding the intercourse between the Celts of Ireland and those of Great Britain by which the traditions of the former might have been diffused among the latter, cp. T. P. Cross, Revue Celtique, XXXI, 421 ff. (1910), where the earlier literature of the subject is given in the notes. For direct communication between Ireland and Gaul (France) cp. H. Zimmer, Uber direkte Handels- verbindungen Westgalliens mit Irland in Altertum und friihen Mittelalter in the Sitzungsberichte, Philos.-Hist. Klasse, of the Berlin 94 Evolution of Arthurian Romance of the sagas current in Brittany during the period under discus- sion have perished, we cannot say whether, after all, these same motifs and incidents which we find in the Irish sagas may not have been familiar also to the Bretons, and so have passed di- rectly from them into France. Besides, the ancestors of the Bre- tons were brothers of the ancestors of the Welsh and so even where such tales are, indeed, Irish by origin, they may have been brought over from Ireland to Great Britain before the emigration to Armorica began. Waiving, however, these considerations, the debt of the French romances to Irish sources, in any event, seems to the present writer to have been greatly exaggerated. But it is best to let the reader judge for himself by putting before him the materials in a typical case — viz., the outlines, respectively, of a French romance, Chretien's Yvain, and of an Irish saga, Serglige Conculaind (Cuchulinri's Sick Bed) — a version of the fairy-mistress theme — which, in some variant form, is supposed to be the source of this same romance:112 Academy of Sciences for 1905, pp. 363 ff., 430 ff., 543 ff., 582 ff. — for 1910, pp. 1031 ff. See, too, this scholar's Auf welchem Wege kamen die Goidelen vom Kontinent nach Irland? in the Abhand- lungen of the same Academy for 1912 — especially, pp. 55 ff. As. against D'Arbois de Jubainville, Rhys, etc., Zimmer (approved by Kuno Meyer) argues that the Gaels in their original settlement of Ire- land did not reach that island by way of Great Britain, but direct from West Gaul, and, moreover, that down to the English conquest of Ireland (late twelfth century) the main intellectual as well as trade relations of Ireland were with West Gaul and not with Great Britain. The argument, on the whole, seems plausible but, as J. Vendryes has remarked in his review of the paper (Rev. Celt.), Zimmer does not sufficiently consider the archaeological evidence. 112 This is the thesis of A. C. L. Brown in his Iwain, a Study in the Origins of Arthurian Romance: Harvard Studies and Notes in Philology and Literature, vol. 8 (1903). Cp., too, his articles, "The Knight of the Lion", PMLA, XX, 673 ff. (1905) and "Chretien's 'Yvain'", MPh., IX, 109ff. (1911). In these articles Brown brings together from various Irish sagas parallels — frequently far-fetched, as it seems to me — to all the principal features of the Yvain — a parallel to one detail from one saga, a parallel to another detail from another saga, and so on. He concludes that all the motley incidents of the French poem are derived from a single Celtic marchen. Origin of the Lays and Romances 95 In Yvain the hero, incited by the report of another knight, Calogrenant, tries the adventure of the fountain in the Forest of Broceliande: If one pours water from this fountain on a stone nearby, a terrible storm arises which is followed by a calm. A knight then comes forth and jousts with the offender. On the way to the fountain Yvain (Ivain, Iwain), like Calogrenant, passes a flock of fierce fighting bulls guarded by a giant herdsman — a black creature with a head larger than that of a horse. He goes through the fountain adventure, but, unlike his predecessor, overthrows the knight and wounds him mortally. Yvain pursues his wounded adversary into the latter's castle. He finds himself caught between two iron gates, but he is recognized by Lunete, a damsel who attends the lady of the castle and who had once been befriended by him at King Arthur's court. This girl gives him a magic ring which will render the holder invisible at will. By the use of this talisman Yvain witnesses in safety the funeral of the knight he has slain and falls in love with the latter's widow, Laudine. Immediately afterwards he begins the courtship of the widow and with the help of the damsel gains her consent. He now becomes the lord of the castle and defender of the fountain. I have given the two stories according to Brown's own analysis, with only slight compression here and there, so that whatever resem- blance there may exist between the two loses nothing in my outlines. hi his review, MLN, XIX, 80 ff., (1904) of Brown's Iwain, — which he describes as "more especially a contribution to the history of Celtic mythology than a monograph on the romance of Crestien de Troyes" — Nitzo dissents from the author's hypothesis that the French poet was following "a clearly defined conte" . Similarly, E. Windisch, Das keltische Britannien bis zu Kaiser Arthur, p. 181 (Leipzig, 1912). — Already Axel Ahlstrom, Melanges de philologie romane dedies a Carl Wahlund, p. 302 (Macon, 1896), expresses the opinion that Chretien's sources for the Yvain, besides contes and popular legends (such as he supposes, gave rise to Marie's lais), embraced, also, "de petits romans en prose" on the same subjects. But Chretien knew only French (no Celtic), and there was no French prose so early as this. In his Sage vom heiligen Gral, pp. 138 ff. (Halle, 1898) Wechssler argues that there were already romances of chivalry among the Celts, but the works which he cites to prove this are, in reality, French, e. g., Chretien's Perceval, the lays, Milun and Boon — and just how far they contain Celtic elements is a debated question. 96 Evolution of Arthurian Romance In his new capacity, Yvain barely avoids an encounter with Arthur, but when the king departs, his wife gives Yvain permission to accompany him and to be absent for a year. He overstays his time, is reproached for it, goes mad in consequence, but is finally cured by an ointment of Morgan le Fay's. There are all sorts of disconnected adventures that follow upon this. He succors a lion who is fighting with a serpent and the lion henceforth follows him. With the beast's assistance he rescues Lunete, who is con- demned to be burned to death on the false accusation of a wicked seneschal. In the midst of this affair, however, he combats a giant who is besieging a castle and threatening to kill the sons of the owner or carry off his daughter. Next he has an indecisive encounter with Gawain, without recognizing his identity, in the affair of two sisters, who are disputing over an inheritance. But this adventure is interrupted likewise by another in which the hero does away with an evil custom that requires thirty girls of the Isle of Maidens to be annually delivered over to two mon- sters. Lastly, Yvain returns to the fountain and stirs up a storm there. Lunete discovers who he is and brings about a reconciliation with Laudine. Let us turn now to Cuchulinn' s Sick Bed: 113 Cuchulinn tries to kill two strange birds linked by a golden chain that appear at a lake in Ulster. Being unsuccessful, he sits down against an upright stone and falls asleep. In his sleep he sees two women, one clad in green, the other in red, who come up to him, laugh and keep striking him until he is almost dead. (As it turns out this is the fairy, Fand, and her sister.) He was carried into a house, where he lay till the end of a year without speaking to anyone. Then, as he lay in the bed, a man mysteriously ap- peared, who sang verses promising him health and strength, if he would accept the invitation of the daughters of Aed Abrat, one of whom, named Fand, wished to marry Cuchulinn. The man departed after that and they knew not whence he came or 118 In connection with this tale, I will call the reader's attention to an interesting account of the Irish fill (professional story-tellers) and of the way in which Irish tales were handed down, which A. C. L.Brown gives, MPh., IX, 121 ff. (1911). Origin of the Lays and Romances 97 whither he went. Cuchulinn rose up and spoke and went back to the upright stone where he saw again the woman in the green cloak. From her he learned that Fand, deserted by her husband Manannan mac Lir (a Celtic divinity), had fallen in love with him. Her own name is Liban. She is sister to Fand and wife to Labraid, who has sent her to ask Cuchulinn for one day's assistance against Labraid's enemies, promising in return to give him Fand to wife. Cuchulinn sent his charioteer, Loeg, to see the mysterious land from which she came. Liban and he went till they came to the place where Fand was waiting for them. Then, it is said, Liban took hold of Loeg by the shoulder. "O Loeg", said Fand, "thou wilt not come out alive today, unless a woman protect thee! ,] " I have not been much accustomed to woman's pro- tection" was Loeg's reply. Then they came to the water's edge, where they entered a boat of bronze and crossed over to an is- land. Loeg saw Labraid and his palace and, returning, told his story to Cuchulinn and everyone else. Again Liban came to invite Cuchulinn to Mag Mell. She sang to him in praise of the place where Labraid dwelt — evi- dently fairy-land.114 "I will not go at a woman's invitation," 114 Otherworld journeys are frequent in Celtic literature. For an account of them cp. The Voyage of Bran, 2 vols. (London, 1895 — 1897), by A Nutt and K. Meyer, A. C. L. Brown's Iwain, 27 if. (Boston, 1903), Josef Baudis, "The Mabinogion", Folk-Lore, XXVII, 31 ff. (1916) — particularly, pp. 35 ff. There seems to have been no distinction in the minds of the Celts between their "Land of the Dead" and their Elysium. So, lately, Baudis, loc. cit, who, I believe, is right. Hence we include both in the term "Otherworld". For a discussion of the whole subject, cp. J. A. MacCulloch, The Religion of the Ancient Celts, pp. 362 ff. (Edinburgh, 1911), whose conclusions are different. MacCulloch (pp. 374) disputes the view held by DArbois de Jubainville, Baudis, etc., that the two Otherworlds were not dis- tinguished. He deduces (p. 370) the idea of the Celtic Elysium from some early myth of a Golden Age. The literature concerning the descriptions of the Celtic Otherworld is given by T. P. Cross, Revue Celtique, XXXI, 461, note 3 (1910). In "Two Otherworld Stories". MLN, XXXH, 280 ff. (1917) John C. Hodges has compared with Celtic conceptions of the Otherworld two stories from the, Arabian Nights, 98 Evolution of Arthurian Romance he declares; but he again consents to send his charioteer. Fand admonishes the latter that Cuchulinn must come with speed, for the battle is appointed for that day. Loeg returns with Fand and describes at even greater length the fairy palace of Labraid and the beauty of a woman there, who is, doubtless, Fand her- self. Cuchulinn is now persuaded, mounts his chariot, and ac- companies Loeg and Fand to Mag Mell. At early dawn he trans- fixes with his spear one of Labraid's chief enemies, as he was washing himself at a well. After that he slew still another and won a victory for Labraid. In return he received Fand, with whom he lived for a month. When he departed, she said to him: "I will meet thee in whatever place thou shalt appoint for me to come." After Cuchulinn returned home, he revealed to his wife, Emer, the appointed place of meeting. The jealous queen lay in wait with knives to murder Fand. Cuchulinn rescued her, but when Manannan mac Lir heard of it, he suddenly appeared, visible to Fand alone. When she saw her husband, the sea-god, coming over the sea, she forsook her mortal lover and went away with him. When Cuchulinn perceived that she had gone, he sprang three leaps upward and three leaps to the right of Luacra, so that he was for a long time without drink and without food among the mountains and "tis there that he slept every night upon the road of Midluacra — " that is to say, he ran mad. Emer persuaded Conchobar to send " poets and people of wis- dom and druids of the Ulstermen" to heal Cuchulinn, but "he sought to murder the people of wisdom". However, they sang (viz., The Porter and the Three Ladies of Baghdad and The Man Who Never Laughed during the Rest of his Days. Burton, I, 151 ff., VI, 160 ff., respectively), which seem to parallel the Celtic more closely than any others in that collection. In the Arabian tales the mortal, not the fairy mistress, takes the initiative. For the literature of the Fairy Mistress theme in Celtic cp. (besides Brown's Twain) Cross, MPh., XII, 598, note 1, 594, note 2. For a contemporary Breton ballad on the fairy mistress theme, taken down by J. Loth, cp. "Le Comte et la Fee", Annates de Bretagne, XXVII, 199 ff. (1911) — also, a Welsh tale noted by G. Dottin, "Notes sur le Folklore Gallois, ibid., XXIII, 462 ff. (1907—8). Origin of the Lays and Romances 99 their druidical charms over him till they captured his feet and hands and till he recovered a little of his senses. He asked for a drink then. They gave him a drink of forgetfulness. As he drank the drink, there was no recollection to him of Fand nor of anything that he had done. Manannan shook his cloak bet- ween Cuchulinn and Fand, so that they should never meet again. Most readers, we believe, will agree with us, that it would be impossible for the French poet to extract from such a story the plot of Yvain, as we have recounted it above. Except that in each the lover runs mad, on losing his mistress, the two stories have virtually nothing in common. In the Yvain, besides, the loss is only temporary. Chapter III. Chretien de Troyes and his Successors. We have now come to Chretien de Troyes, who, even if he had predecessors, gave, as we may safely assert, a new value, and probably a new character, to the poems relating to Arthur and his knights. The fact that all poems devoted especially to Arthurian subjects before Chretien — granting that such existed — have disappeared, shows that, on the whole, they could not have been marked by any great literary excellence.1 We know nothing positively of the life of the poet. The designation, "de Troyes", which he uses in the Erec, seems to prove that he was a native of Troyes in the province of Cham- pagne, and that he was a herald has been conjectured2 on the 1 This is, of course, contrary to Brugger's view, Zs. f. frz. Spr. u. Litt., XXXI2, 143 f. that all the finer Arthurian romances have perished and that we possess only the productions of a period of decadence. 2 By G. Paris, Journal des Savants for June, 1902, pp. 295 f. = pp. 251 f. of the reprint in the G. Pans-Melanges (Paris, 1910). Owing to the logic-chopping debates between Fenice and Thessala in Cliges (11. 3085 ff. and 4409 ff.) and Laudine and Lunete in Yvain (11. 1589 ff.), M. Wilmotte has suggested, p. 167 of the Bulletin de la Classe des Lettres, etc., for 1890 of the Academic Royale de Belgique that Chretien was a jurist; but Ins scholastic training would sufficiently account for the character of the passages in question. Cp. the same scholar on Chretien, Le Moyen Age, for 1914, pp. 102 ff. For various speculations concerning Chretien's life and a very thorough sifting of such speculations see especially, G. Paris, loc, cit., pp. 293ff. and Foerster, Chretien Worterbuch, 22 ff., 39 ff. Paris, Melanges, I, 260, conjectures — and W. Meyer-Liibke, Zs. f. frz. Spr. u. Litt., XLIV l, 162, expresses his approval of the conjecture — that Chretien visited England. The evidence to that effect is: 1. that in his Cliges, 11. 4579 ff., the poet makes the in- significant English town, Wallingford (Galinguefort), the scene of one Chretien de Troyes and his Successors 101 basis of the Lancelot (11. 559 Iff.), where the herald recognizes Lancelot, who has come incognito to a tournament, and cries out: Or est venuz qui aunera (i. e., the conqueror has come). Chretien adds that this was the first time the expression had been used: Nostre maistre en fu li hira Qui a dire le nos aprist, Car il premier ement le dist. In any event, his translations from Ovid prove that he had the usual scholastic education of the time and even render it prob- able that he shared the renewed ardor with which the study of the classics, and, particularly, Ovid, was pursued in learned circles toward the middle of the twelfth century. In a well-known pas- sage at the beginning of Cliges the poet enumerates his works up to that date (about 1168): Erec and Enide, a poem on King Marc and Iseult, and translations of Ovid's Ars Amandi (L' Art a" Amors) and Remedia Amor is (Les Comandemenz Ovide), with two other pieces which appear to be derived from the same writer's Metamorphoses.3 of Arthur's tourneys, as it had actually been the meeting-place of Henry II and his barons. 2. that he gives, Cliges, 11. 276 ff., a rather vivid description of sea-sickness. These matters are supposed to have been drawn from his personal experience. Obviously, however, the inference is not imperative. Cp., contra, Foerster, Chretien Wbrter- buck, Introduction, pp. 40 f. 8 The Cliges begins as follows: Cil qui fist d'Erec et d'Enide Et les Comandemanz Ovide Et l'Art d* Amors an romanz mist Et le Mors de l'Espaule fist, Del roi Marc et d'Iseut la blonde, Et de la Hupe et de l'Aronde Et del Rossignol la Muance, Un novel conte recomance D'un vaslet qui an Grece fu Del lignage le roi Artu. It is generally agreed that considerations of rhyme, not fact, dictated the order in which these works are enumerated. Only Meyer-Lubke, Zs. f. frz. Spr. u. Litt, XLIV \ 129 ff. objects. Le Mors d'Espaule has been taken to be a version of the classical legend concerning Pelops's shoulder wich was devoured by 102 Evolution of Arthurian Romance If we accept the Philomena episode of the Ovide Moralise as one of these pieces,4 our author would be identical with the enigmatical Crestiiens li Gois, the author of that work, but this is more or less doubtful. From the omission of the hero's name in the allusion, G. Paris argued5 that Chretien's poem con- cerning Marc and Iseult was episodical in its nature and not a genuine Tristan romance. The inference seems hardly justi- fied, but, in any event, the work has completely disappeared.. Of the extant romances, Erec (about 1168) is the first, and Cliges (about 1170) the second. Next come his three latest romances, with the following approximate dates:6 Lancelot, 1172, Yvain, Ceres. This leg-end receives only a passing mention in Ovid — though accessible, doubtless, to Chretien in Hyginus, Fable 83 — and be- sides, Mors (bite) does not necessarily imply devouring. On these grounds, G. Paris, Melanges, I, 248, objected to the above-mentioned identification of the story and suggested a Celtic origin. Nevertheless, the identification is, most likely, correct. The poem dealing with the muance (transformation) of the three birds was, of course, the story of Tereus, Procne and Philomela in the Metamorphoses. It is possible that both UArt d' Amors and the Comandemenz Ovide refer to the Ars Amandi. — For Ovid's influence on Chretien see F. E. Guyer's exhaustive study, "The Influence of Ovid on Crestien de Troyes", RR, XII, 97ff., 216ff. (1921). I regret that the article appeared too late for me to cite it in detail. It throws new light on many con- ceptions in Chretien's poems — especially on those that relate to love. In his "Zum Text des Erec", Zs. f. frz. Spr. u. Litt., XXXVHI1, 97 f. (1911), G. Cohen has tried to show that the above lines of the Erec are spurious, but his argument is futile. As far as I know, no one has expressed approval of this suggestion, save W. Meyer-Liibke, "Crestien von Troyes Erec und Enide", ibid., XLIV \ 136. On the subject of the above lines, cp., too, Foerster, Cliges3, pp. VI ff. (Halle, 1910). 4 So G. Paris, loc. cit, and C. De Boer in his edition of the Philomena (Paris, 1909); but for an argument contra, cp. Foerster, op. city p. VII, note 2. 5 Op. cit., pp. 254 ff. e So G. Paris, op. cit., pp. 262 ff. He dates the Erec about 1168. For discussions of the respective dates of Chretien's romances, except Perceval, see Foerster 's introductions to these romances in his editions. His maturest opinions on the dates of all (including the Perceval) will be found pp. 33 ff. of the Introduction to his Chretien Wbrterbuch. Ibid., p. 39, he expresses the opinion that Chretien's Chretien de Troyes and his Successors 103 1173, Perceval, 1175. The Perceval remains a fragment. This was the last of the poet's works, its composition being broken off, it seems, by his death. We leave aside Guillaume d1 Angleterre, since, even if it is Chretien's,7 the poem does not belong to the genre of Arthurian romance. It is significant that at the very beginning of his career Chretien should have been occupying himself with Ovid and with a subject which lay so close to the Arthurian cycle — namely, the story of Iseult. With the exception of the Cliges, whose con- nection with Arthur is wholly artificial, all the writings of our poet from the Erec down relate to the matiere de Bretagne and they all, without exception, betray the influence of Ovid. It would appear that the course of his development was as follows: He first becomes interested in the love-poetry of Ovid (it may be in his school-days), then he is attracted by the theme of Tristan, which he knew through oral tradition, or, possibly, some earlier French poem on the subject. With a poem of some kind on this hero he makes his first essay in the matiere de Bretagne. If we had this Tristan poem, we should probably find that in the ana- lysis of love-moods the influence of Ovid and of the French ro- mances of antiquity was already traceable in it. The charm of the Celtic stories has laid its hold on the author and in all of his succeeding works, with the single exception of the Cliges, he draws incidents from this source, though by no means exclusively. Since Geoffrey and Wace, however, the authority of Arthur is great, production began slightly before 1160 and ended not so very long after 1170. He dates Cliges about 1164 (p. 38) and accepts the Erec as earlier, though he suggests no definite date for the latter. The Lancelot and Yvain he places (p. 33) between 1164 and 1173, the latter being the later. Perceval he dates not much later than Yvain. Paris and Foerster agree as to the order in which Chretien's Arthurian works were written and we may indeed regard the question as settled. See further on the date 1170 as the upper limit of the date of the Yvain M. Morner: "Le 'terminus a quo' du Chevalier au Lion", Archivium Romanicum, III, 95 f. (Geneva, 1919). 7 Cp. Introduction to Foerster's edition (Halle, 1911) for dis- cussion of the question of its authorship. Foerster upholds the as- cription to Chretien. 104 Evolution of Arthurian Romance and under the influence of these works a new standard of refine- ment is already, to a certain degree, associated with his court. Chretien avails himself of this prestige, which had more or less historical validity in the eyes of his contemporaries, since per- haps the majority of them accepted as historical truth Geoffrey's pictures of Arthur's glories, and makes Arthur's court the star- ting-point of the adventures of all his heroes. This conception, it is safe to say, was his own invention — only the setting and, in part, the incidents are Celtic. The ideas which his romances embody are those of his own time and country, and even the in- cidents he has adapted in thorough-going fashion to the con- ditions of French feudal society in the second half of the twelfth century. A comparison has often been made between Tennyson and Chretien, as regards their attitude towards their originals, and it is, in many respects, just. The French poet probably ad- hered less closely to a single original for each poem than Tenny- son did — although, on the other hand, in the matter of incident, he, no doubt, drew less from his own invention than the English poet. Both, however, are alike in making the stories of an earlier time the vehicle of the ideals and sentiment of their own age. Chretien's principal importance, then, lies in the fact that he is_ the chief literary exponent of the feudal society of his age. Just as the Song of Roland and the cycle of William of Orange reflect the feudal life of the early twelfth century, so Chretien's ro- mances reflect more perfectly than any other works the life of the aristocratic classes in the later twelfth century. Now, before illustrating somewhat more in detail by reference to the specific works the observations we have just made with regard to Chretien's development and his shaping of the Arthu- rian theme, we should recall briefly what were the distinctive ideals of this society: Chretien himself during the period of his literary activity was connected mainly with the court of Marie, Countess of Champagne, whose father was Louis VII of France and whose mother was Eleanor of Poitiers, the latter (the Eleanor of Shakespeare's King John) being successively the wife of Louis VII of France and Henry II of England. At a later period he found a patron in Philip of Alsace, Count of Flanders. The dif- Chretien de Troyes and his Successors 105 ference between the ideals of the society of such a court as Marie's in the second half of the twelfth century and those of the earlier time represented by the chansons de geste is connected, above all, with the improved position of women. This accords with the gene- ral advance of civilization from the unrestricted reign of physical force in the earlier Middle Ages and is manifest even in the political arrangements of the time, since women were now permitted to inherit the great fiefs. The centre from which this move- ment radiated was the south of France — the land of Eleanor of Poitiers and the land where all ideas of social and intellectual liberty had so far received their highest development. The love lyrics of Provence show the growing idealization of woman, and the conception of the amour courtois is the product of this ideali- zation. The starting-point of this idealization, to be sure, was earthy enough in its nature — for it was, first of all, the motive of self- advancement that prompted these Provencal singers — the trou- badours, as they are called — to confer a poetic exaltation on the wives and daughters of the great barons of the South, and where panegyric passed, apparently, into love-poetry, the love was, in the vast majority of instances, a pure fiction,8 just as in the case of the amorous effusions with which the Elizabethan poets were accustomed to tickle the vanity of their Virgin Queen. But the very unreality of the theme of the troubadours, who, as a rule, were separated by a social gulf from the objects of their pretended adoration, drove these poets within themselves to an idealization of the persons addressed.9 For the poet who found himself in 8 On the origins of the amour courtois (Middle High German Minne) cp. E. Wechssler's "Frauendienst und Vassalitat", Zs. f. frz. Spr. u. Liu., XXIV1, 158ff., and, above all, his Das Kulturproblem des Minnesangs: Band I, Minnesang und Christentum, Halle, 1909 — also K. Heyl: Die Theorie der Minne in den dltesten Minne- romanen Frankreichs: Marburger Beitrdge zur romanischen Philo- logie, IV (Marburg, 1911), and T. F. Crane, Italian Social Customs in the Sixteenth Century and their influence on the literatures of Europe, ch. 1 (New Haven, 1920). 9 Cp. Wechssler's book, pp. 219ff. As appears from this same work, there were other Christian elements that helped to shape the 106 Evolution of Arthurian Romance this attitude of inward contemplation the doors were thrown wide open to the spirit of mysticism which was at that time coloring contemporary thought in all its higher forms and which, through St. Bernard of Clairvaux and his fellows, in the same century, had already imparted a new life to the Church. Thus woman who had been, in the beginning of the Provengal lyric, merely the object of a rhetorical lip-service, became in the course of its devel- opment the object of a cult, which varied in genuineness according to the spiritual capacity of the individual singer. As all stu- dents of literature know, this half mystic and half philosophical conception of woman and love passed from Provence into Italy and after being still further developed there by Guido Guini- celli and other Italian poets, attained immortal expression in Dante's idealization of Beatrice Portinari in the Vita Nuova and 1 Divina Commedia. For the great mass of noblemen and noble ladies, as well as writers, however, these conceptions remained, no doubt, merely a convention which they adopted for the sake of fashion or amusement, but they, at least, pretended to take them seriously, and so from about the middle of the twelfth cen- tury the duties of the lover towards his mistress amount to a com- plicated code10 — a code which was perhaps not very strictly - conformed to in real life, but which at least expressed an ideal. Chretien's patroness, Marie de Champagne, as is evident from conception of the amour courtois — the worship of saints, the doctrine of Charitas, the methods of scholastic philosophy — but mysticism was much the most powerful single influence. Contrary, however, to the view which has been often expressed, Wechssler points out, pp. 434 ff., that the worship of the Virgin Mary (mariolatry) was a con- sequence, not a cause, of the cult of woman that was developed in Provence. It does not arise until about the end of the twelfth century. 10 Cp. especially, the one embodied in Andreas Capellanus's De Amore Libri Tres, edited by E. Trojel (Copenhagen, 1892). Andreas was for a time, at least, in the service of Marie de Champagne, and composed his treatise apparently before 1196. It is improbable that there were ever any serious "courts of love" (to decide points that might arise in the relations of lovers). On the whole subject, cp. W. A. Neilson, "The Origins and Sources of the Courts of Love", Harvard Studies and Notes in Philology and Lite- rature, VI (Boston, 1899) and Crane, op. cit., ch. 1. Chretien de Troyes and his Successors 107 the poet's own statement in his Lancelot, played an important part in the development of the literary tradition relating to these things and his work is the very highest expression of su.oh ideals. In contrast with the chansons de geste, pictures of courtesy and refinement of manners, still more than in the romances of anti- quity, constitute one of the main sources of interest of the narra- tive and much space is given to the functions, the feasts and festivities, where these traits are exhibited or developed. Even fighting appears in a more civilized aspect. The tourney, with its formal regulations, takes the place of the rude combats of the chansons de geste and the somewhat similar descriptions in the romances of antiquity, and the valor of the knight is now mainly employed in the succor of distressed women. The whole modern conception of the relation of the sexes, finds here its beginnings. An advance towards these conceptions, to be sure, had already been made in the romances of antiquity, and there can be little doubt that both in respect to form and sentiment these works — particularly the Thebes and the Eneas11 — were the true pre- cursors and, in a large measure, the models of Chretien's romances. 11 Both were probably composed in the fifties of the twelfth century. On the relative date of the Eneas see, last, Faral, pp. 169 ff. There are the beginnings of a spirit of courtesy in the fragmentary Latin poem, Ruodlieb, (edited by F. Seiler as Der dlteste Roman des Mittelalters, Halle, 1882), which has usually been regarded as of German origin (near Tegernsee) and dated about 1030, but which M. Wilmotte, "Rodlieb, notre premier roman courtois", Romania, XLIV, 373 ff. (1916 — 1917), now assigns to Northeastern Gaul (valley of the Meuse, between Liege and Namur) and the early twelfth century. The author is not so much interested in combats as in the manners, diver- sions, decorations, etc., of aristocratic society. Wilmotte sees in all this an early independent development in the North in the direction of the romances of antiquity and Chretien. Wechssler, Sage vom heiligen Gral, p. 139 (Halle, 1898) and Problem des Minnesangs, p. 30 (Halle, 1911) had already mentioned Ruodlieb in this connection, but had distinguished between the pictures of courtly life here and the special ideal of courtesy which we find in the romances and which he derives, I believe, justly from Southern France. In any event, there is no evidence that the Ruodlieb exercised any influence on subsequent poetry. 108 Evolution of Arthurian Romance It is important to recollect these distinctive traits of French feudal society in the latter part of the twelfth century when we come to consider the element of originality in Chretien's works — when we come to determine what is Celtic in incident or manners in his poems and what is possibly of his own invention or at any rate of non-Celtic origin in these works. A greater contrast with the position and character of woman as depicted in these romances than that which is furnished by the Celtic sagas could hardly be imagined.12 In the latter we have a conception of woman which is inferior not only to that of the Arthurian ro- mances, but to that of the sagas of any other branch of the Indo- European races, even the earliest in date. The Irish heroines have strength and stature, but the abundant examples which Professor Zimmer has adduced from the Cuchulinn sagas amply justify the term gemein ("low-down") with which he stigmatizes these women. Indeed, the exploits of some of the most famous of them, like Medbh, wife of King Ailill of Connaught, do not bear repetition in modern polite society.13 Let us turn now to Chretien's romances themselves. Of the romances which have come down to us the earliest is Erec. The Welsh tale of Geraint and Enid, which corresponds to Chretien's 12 Alfred Nutt, The influence of Celtic on Mediaeval Romance (London, 1904), and others, to be sure, have tried to establish the existence of chivalrous feeling towards the weaker sex in these sagas — , for instance, in that of Deirdre, the great heroine of Irish legend — , but, as it appears to me, without success. I will refer the reader for the proof of the contrary to the remarkable paper by the late Professor Zimmer, published } after his death in the Sitzungsberichte der kbnig- lichen preufiischen Akademie der Wissenschaften. Philosophisch- Historische Classe, 1911, IX, 174ff. entitled Der Kulturhistorische Hintergrund in der altirischen Heldensage. 18 J. Vendryes, Revue Celtique, XXXII, 232 ff. (1911) in his review of Zimmer's paper, has tried, to be sure, to show that the in- ferences of the German scholar are unjust. J. Loth, too, has objected, ibid., XXXIII, 260, note (1912), that we might as well judge the manners and morals of the women of early Greece from the conduct of the Homeric goddesses. But there is no such vulgar abandon in the case of these goddesses. Chretien de Troyes and his Successors 109 poem, is, doubtless, derived directly from the latter14 and not vice versa, but evidence of Celtic elements in the Erec, itself, are not lacking. To begin with, the name Erec is Celtic, being an alteration of Breton Weroc.15 The same seems true of some of the other principal names in the poem: Eurain, Mabonagrain, etc. Take, moreover, the episode (11. 5461 ff.) known as the Joie de la Com:16 There is an enchanted garden next to the palace of King Eurain. It is enclosed with a wall of air, which is, never- thelessT just as effective as a wall of iron, so that no one can enter it save by the will of the possessor. To overcome this enchantment now is the task of Erec. Stakes with human heads crowning them at the entrance of the garden confirm the warnings which the hero has received from Eurain and his companions against under- taking the adventure. In the garden under a sycamore he comes upon a beautiful girl reclining on a silver couch; but just at this moment the girl's lover appears — Mabonagrain, a giant knight — and threatens him. A combat follows in which Erec, of course, vanquishes his opponent and thereby dispels the enchantment of the garden. It turns out that Mabonagrain had promised his dam- sel that he would never leave the garden until he had been van- quished in combat. Now Celtic parallels seem to establish the Celtic origin of the episode.17 Similarly, the custom of hunting the white stag 14 Cp. Part. IV, below. 15 Cp. G. Paris, Romania, XX, 166, note 1 (1881), and J. Loth, Revue Celtique, XIII, 482 ff. (1892), who prove that Zinimer's deri- vation, Zs. f. frz, Spr. u. LitL, XIII \ 26 ff., is wrong. 16 F. Lot, Romania, XXIV, 321ff. (1895). Evrain is a mistake for Euuain (i. e. Owen). Mabonograin is a combination of Mabon and Evrain — names of two magicians in the Bel Inconnu. Cp., too, E. Philipot, "Un episode d'Erec et Enide, la Joie de la Cour", Romania, XXV, 258ff. (1896) and W. H. Scliofield, Studies in the Libeaus Desconus, pp. 124 ff. 17 Cp. the examples of "druidic mist" (walls of air from which there is no exit) cited from Celtic stories by Miss Paton, Fairy Mytho- logy, p. 84, note 3. Foerster, Erec\ p. XXIH (Halle, 1909), on the other hand, maintains that we have here merely a variant of the marchenmotiv diffused all over the continent of Europe of the deliverance of a maiden held in captivity by a giant. According to Foerster, Chretien again used this motif in the Yvain. 110 Evolution of Arthurian Romance at the beginning of the romance reminds one irresistibly of the marvellous white doe in Guigemar, which, as we have seen, is, no doubt, of Celtic origin. On the other hand, the principal motif of the poem, the Verliegen (to use a convenient German term) of the hero — that is, the way in which, through uxoriousness, Erec neglects deeds of valor and gives himself up to love, so that his decline in knightly prowess becomes a matter of general re- proach, and, at length, of sorrow, even to the young wife herself — bears on its face the marks of its French origin. Spurred by the grief of his wife, the hero, though angry at the implied reproach, arouses himself from his sloth and redeems his reputation, just as a young French knight of the twelfth century doubtless would have done, when touched in the point of honour. The wife en- dures patiently the tribulations to which his wounded vanity sub- jects her and in the end her love triumphs — a motif that recalls the Patient Griselda of Chaucer's Clerk's Tale. There is nothing Celtic in the conception, and, considering that it is intimately interwoven with another motif that is manifestly French — namely, that of Verliegen — there seems to be no reason for ascribing to it a Celtic origin. The same is true of the sparrow-hawk incident in the earlier part of the romance, in which the bird is the prize of beauty which each knight endeavors to win for his lady — also, of the kiss of the most beautiful girl at Arthur's court, which is to reward success in the chase of the white stag. These incidents exhaust the motifs of the romance, for the combats with robbers and chance knights, like Guivret le Petit, of which there are many, have no significance. If the above outline of the sources of Erec et Enide is correct, the poem owes little, after all, to Celtic models. It is merely the Arthurian setting, the nomenclature, the Joie de la Cour episode, and a detail here and there that we may ascribe to this origin. The novelty consists in the happy selection of incidents, in the elegance of style, in the idealized pictures of chivalry, with its refinement of manners and its courage enlisted in the service of beauty. But just as the direction of Shakespeare's art at the beginning of his career is determined in a considerable measure by Marlowe and Lyly, so Chretien, too, has had his precursors — Chretien de Troyes and his Successors 111 in the authors of the romances on subjects of classical antiquity, as we have already said, the Roman d' Alexandre, Roman de Thebes, Eneas, Roman de Troie. The dates of these works are not fixed precisely, but there can be no reasonable doubt that the composition of the last three fell within the ten or fifteen years preceding the composition of Erec. In this group of romances we already have the effort to depict the new ideals of courtesy and chivalry, so that Eneas and Dido, Achilles and Polyxena, are twelfth century knights and ladies. One can observe the influence of these works on Chretien even in matters of phrasing and tech- nique,18 and besides, the length of his romances was, no doubt, approximately determined by the example of the Thebes and Eneas. But since, in questions of art, sentiment, form, and style are even more important than subject-matter, it is, after all, the debt of the French poet to these predecessors rather than his meagre and un- certain borrowings from Celtic tradition that needs to be emphasi- zed. The full force of these models — the romances of antiquity — did not make itself felt, however, in the Erec. As far as the refinements of life are concerned — rich and brilliant externals, careful etiquette and polished manners — these are already given prominence in Chretien's first extant romance; but he has not yet accepted love as the all-dominating factor in human life, in the spirit of the amour courtois, which had already touched the Eneas and Troie, in particular. In his subsequent romances this fashionable conception of the relations of the sexes — doubtless through the influence of his patroness — was destined to impose 18 This has been convincingly set forth by M. Wilmotte, If evo- lution du roman frangais aux environs de 1150, Brussels, 1903, and F. M. Warren, "Some Features of Style in Early French Narrative Poetry", MPh., Ill, 179ff. (1905), 513ff. (1906). IV, 655 ff. (1907). For the general influence of the romances of antiquity on sub- sequent literature, including Chretien, see, besides Foerster's Chretien Worterbuch, pp. 33 ff., and Faral's oft-quoted book, the following Grottingen dissertations: R. Witte: Der Einfiufi von Benoits Roman de Troie auf die altfranzosische Literatur (1904), A. Dre&ler, Der Einfiufi des altfranzosischen Eneasromanes auf die altfranzosische Literatur (1907) and G. Otto, Der Einfiufi des Roman de Thebes auf die altfranzosische Literatur (1909). 112 Evolution of Arthurian Romance itself upon the different love-stories that he dealt with, until in the Perceval he set his hero a new aim, and sexual passion, now subordinate in his scheme, strips itself again of ephemeral con- ventions.19 Let us examine now the manner in which Chretien composed his romances.20 The plan and the main idea of each romance are his own. He makes up his narrative by combining motifs which he derived from earlier sources, in some instances, using episodes drawn from such sources to embellish or lengthen out the story which is the basis of his work. One may question whether a single fundamen- tal idea21 runs through the whole of the Erec or the Yvain or the Perceval, but otherwise the above is, probably, an accurate statement of the manner in which these works were composed. 19 Myrrlia Borodine (now Mme. Lot-Borodine) in her important study, La femme et V amour au Xlfe siecle d'aprhs les poemes de Chretien de Troyes, (Paris, 1909) distinguishes three stages in Chretien's development in his treatment of love: 1. Erec: the ideal of active knighthood (V ideal chevaleresque) is stronger than love. 2. Cliges, Lancelot, Yvain: love is the sovereign power. 3. Perceval: the re-, ligious ideal is stronger than love. If one remembers that Chretien's religious feeling had no great depth, these divisions may be accepted as valid. Foerster, Chretien WB, p. 55*, has observed that the Erec resembles the chansons de geste not only in the view of woman's subordination to man, but in literary technique and in its outbreak against the conteurs — moreover, that it contains a number of allusions to various heroes of these works. 80 The views on this subject which I adopt in this place are, in the main, those of W. Foerster, Yvain, pp. XXI ff. (Halle, 1913). 91 Grundidee is Foerster's word. He regards Chretien's ro- mances as Thesenromane. Cp. Chretien Wbrterbuch, pp. 43*ff. Foerster, moreover, minimizes, perhaps, even more than is due the Celtic element in the poet's works. To be sure, in my opinion, there is no reason to believe that this element was at all large. For an argument — in the main, correct — to the effect that Chretien's object in composing his romances was to interest by his narrative, not to illustrate a thesis, see W. Kiichler, „Uber den sentimentalen Ge- halt der Haupthandlung in Crestien's Erec und Ivainff, Zs. f. rom. Ph., XL, 83ff. (1919). Chretien de Troyes and his Successors 113 That this was the mode of their composition, indeed, seems clear from the next of the series — namely, the Cliges, for here the principal element in the story is a familiar one and we can observe how the poet has united it with other material in order to secure the requisite length for his work. To elucidate his method, it will be necessary to outline briefly the narrative of the ro- mance. The poem falls into two unequal divisions: 1. The story of the hero's parents. 2. The hero's own story. It contains 6784 lines, of which 2382 are given to the story of the parents, the remainder to the hero's, except about 300 lines, which serve as a connecting link between the two. As was said above, it is generally agreed that no part of this romance has any genuine connection with the matiere de Bretagne and the Arthurian affiliation which Chretien gives it is of the most artificial kind. First as to the story of the parents: The father, Alexander, is a prince of Con- stantinople, who, on account of Arthur's fame, journeys to Bri- tain to receive from the great king the order of knighthood. Ar- thur goes to Brittany and leaves Count Engres regent of his king- dom. On the voyage he is accompanied by Alexander and his niece, Soredamors, who fall in love with one another, but fear to declare their mutual passion. But Engres, like Mordred in Geoffrey of Monmouth, from whom he is evidently imitated,22 endeavors to seize the kingdom in Arthur's absence, so that Arthur is compelled to return. He besieges the faithless Engres at Windsor. Finally, it is Alexander who by a ruse gains possession of the castle for the king. In the meanwhile, he had often been thrown with his lady-love in the court-circle of Guinevere, who guesses the secret of the lovers. She brings about the long-delayed declaration, and they marry. Cliges, the hero of the story, is the fruit of their union. This ends the first division of the poem, and next comes the connecting link. During the absence of Alexander from Constantinople his 29 Cp. Bruce's edition of the Mort Artu, p. 287 (Halle, 1910). Cp. also, Annette B. Hopkins, The Influence of Wace on the Ar- thurian Romances of Crestien de Troies, p. 35 (University of Chicago dissertation, 1913). Curiously, Chretien does not mention Mordred directly in any of his works. 114 Evolution of Arthurian Romance father, the emperor, has died and a younger son, Alis, has as- cended the throne, having received false intelligence of his brother's death. The latter now comes to Constantinople and makes a com- pact with Alis by which Alis is to retain the imperial title, but he is to be the real ruler. Moreover, Alis agrees not to marry, so that Cliges may inherit the throne. Before his death Cliges's father urges his son to test his knighthood at Arthur's court by encounters with the best knights there — especially Gawain. Both father and mother die. Now follows the second and main division of the poem. Alis breaks his promise and carries off by force as his bride Fenice,23 the daughter of the German emperor, who had been engaged to a Saxon prince. But Fenice and Cliges fall in love with one another. Fenice appeals for aid to her nurse, Thessala, who saves her from the consummation of her marriage with Alis by giving the husband a magic drink that deludes him into belie- ving that he has consummated it, when he has not. On the way home Cliges confirms his hold on his lady-love's affections by defending her from the Saxon prince who tries to recover his betrothed. The hero now carries out his promise to his father by going to Arthur's court and measuring himself there in tourneys with the greatest knights.24 He distinguishes himself, but love in time drives him back to Constantinople. Here the lovers again avail themselves of Thessala's magic arts, and it is agreed that, to elude her husband, Fenice is to take a magic potion which will throw her into a death-like trance. According to their anticipations, she is carried out as dead to her grave. Certain skilled physicians from Salerno see through the deceit, but they are slain and the girl is buried. Her lover evades the watch and brings Fenice into 23 This name is, doubtless, Greek Phoinike, although, G. Cohen, Zs. f. frz. Spr. u. LitL, XXV2, 146, derives it from Macedonian Pherenike (= Berenike). The Greek names in the romance are classical, rather than Byzantine in origin. Cp. Krumbacher's com- munication to Foerster in the latter's Cliges9, p. XXXVII. 24 We have here in the tournament, lasting four days, a variant of the widespread "three days' tournament motif". Chretien de Troyes and his Successors 115 a tower, where she is restored to life by her nurse. Here they live for months tagether. A new trouble, however, awaits them, for they are detected by a young knight, named Bertram. Nevertheless, they escape to Arthur's court, and after a time, on the death of Alis, they return to Constantinople and reign there for the rest of their lives. The kernel of this romance is the love-affair of Cliges and Fenice, including the deceits which it involves to defeat the con- summation of the latter's marriage with the uncle of Cliges and to bring about the union of the lovers- All the rest is secondary and, for the most part, as is obvious, added merely to give variety to the poem. Now we have here in this main story merely an adap- tation of the oriental motif of the wife who feigns death in order to deceive her husband for the sake of her lover.25 In what is, perhaps, the earliest of its known forms, the Solomon and Marcolf tale, it is the wife of Solomon who practises the deceit. Though not of identical origin, the story is similar to that which is fa- miliar to all the world in Shakespeare's Romeo and Juliet. At the beginning of his romance Chretien tells us that he derived his story from a book which belonged to the library of the church of Saint-Pierre de Beauvais. An allusion in the poem itself shows that he was acquainted with the Solomon and Marcolf tale, but his version of the motif bears so close a resemblance to that which is preserved in the thirteenth century prose conti- nuation of the Seven Sages, called Marque de Rome, that one may reasonably conclude that the tale which he found in the above- mentioned book had virtually this form.26 The deluding of Alis by Thessala's magic drink, however, has no parallel in the oriental 35 Cp. Foerster, Introd. to small Cliges (second ed.), XXXII ff. and G. Paris, Melanges, 309 ff. 28 So Foerster, Cliges3, pp. XXXII ff. (as in his previous editions) and G. Paris Melanges, p. 312. W. Golther, Zs. f. frz. Spr. u. Litt, XIII8, 7 (1891) and ibid., XXIV2, 7ff. (1902), expresses the opinion that the passage in the Marque de Rome is derived from Cliges and not vice versa, but see on this subject Foerster, Cliges3, p. XXXIV (1910). Besides, in his Tristan und Isolde, p. 213 (Leipzig, 1907) Golther seems to have retracted his earlier opinion. 116 Evolution of Arthurian Romance tale. In this we have a motif imported from another source27 — namely, from the old chanson de geste, the Enfances Guillaume, where the delusion is the effect of sorcery. That the magic drink should be substituted for the sorcery of the old poem is plainly due to the influence of the Tristan story, which was occupying at this time so large a place in Chretien's mind. It has been pointed out by Professor Foerster and others28 27 Cp. G. Paris, Melanges, p. 293. 28 Foerster was the first to advance the theory that Cliges was composed as an Anti-Tristan, viz., in the Introduction to his small Cliges (1901). He repeated it in subsequent editions. Cp. the third (final) edition (1910), pp. XLVIff. The theory was accepted by Golther, Zs. f. frz. Spr. u. Lilt., XXIV2, 7ff. (1902) — see, also, his Tristan und Isolde, pp. 21 Iff. (Leipzig, 1907) and A. G. van Hamel, "Cliges et Tristan", Romania, XXXIII, 465 ff. (1904). G. Paris also accepted it, but with the modification indicated in the the text above. See his fourth Cliges article, Journal des Savants for August, 1902, pp. 442 ff. (in the Melanges reprint, pp. 288 ff.). For a reply to Paris, cp. Foerster, op. cit.f p. LIII, note 2. In his "Randglossen zum Athisroman", Zs. f. rom. Ph., XXXVI, 727 ff. (1912), Foerster finds, also, strong resemblances between this famous tale of friendship (especially, in Part II of the text of the Tours MS.) and Cliges. He argues that Athis et Prophilias contains' no allusion to Arthur or Tristan, and no marks of the Minne spirit — hence is older than Cliges. The reasoning, however, to my mind, is, by no means, conclusive. The omission of allusions to Arthur and Tristan is not surprising in a romance based on Oriental materials, and, as a matter of fact, many passages in the romance betray the new Minne ideal, e. g. 11. 2805 ff. (Athis and Gaite fall in love with each other), 3360 ff. (Athis's monologue on his love-sickness). Cp. A. Hilka's edition, I, 99, 118 (Dresden, 1912). — Op. cit., p. 736, Foerster thinks that Chretien's Perceval, 1. 9336 (Baist's edition, 1. 7926) shows, also, the influence of Athis et Prophilias, 11. 2936 ff. In each we have a discourse extending through several lines on the necessity of being on one's guard against catching cold after overheating. Prophilias thinks that Athis, owing to his neglect of the requisite precautions in such cases, has taken cold and is "sanmellez", and the girl in the Perceval warns Gawain, "Que maint an sane mesle an sont", from the same cause. Foerster declares that he knows of no other parallel to these passages in Old French literature and that only in the Athis and Prophilias passage is the meaning of sanmellez (sane mesle) clear. The two passages are certainly very similar, but, Chretien de Troyes and his Successors 117 that the whole character of the Cliges, indeed, was determined by the lost Tristan poem. Not only have we here the introductory ro- mance of the hero's parents told in full, but the whole poem may be regarded as an Anti-Tristan, or, in any event, as a pendant of Tristan.29 The manner in which Iseult in the older story yields herself to the embraces of two men, her husband and her lover, was shocking to the sentiment of the cultivated, though immoral, society for which the poet was writing, so that the Cliges was manifestly composed by way of polemic against the coarser views of love which underlay that romance. It is hardly correct to as- sume30 that our poet is here adopting the bourgeois ideal, which permits the union of lovers only through the marriage ceremony. Such an ideal would have had but little attraction for Chretien's aristocratic patrons of the twelfth century31 — either women or men — and, besides, the deceit which Fenice practised on Alis, her legitimate husband, and her subsequent conduct are little con- apart from the unsettled relative date of the two romances, the simi- larity may be due to the fact that both authors drew independently from contemporary notions concerning the physiological process in such attacks. In his U e" volution du roman francais aux environs de 1150, p. 345, M. Wilmotte states that Cliges was composed manifestly under the influence of Gautier d' Arras's Eracle, which, like Foerster, he dates (p. 337) 1164. He even lists pp. 351 ff. (including notes) a number of verbal coincidences in which, as he thinks, Cliges shows the in- fluence of Eracle. It is difficult to say how far their common classical sources account for the resemblances of situation and of phrasing, but Wilmotte's argument is plausible. — W. M. Stevenson assumes in his Der Einflufi des Gautier d 'Arras auf die altfranzosische Kunst- epik, insbesondere auf den Abenteuerroman, p. 102 (Gottingen diss. 1910), that Gautier and Chretien mutually influenced each other. This seems very likely. 29 This last is the opinion of G. Paris, Melanges, p. 293. 30 So Foerster, Cliges*, pp. XXIXff. 31 Andreas Capellanus ascribes to Chretien's patroness, Marie de Champagne, the sentiment "Amorem non posse suas inter duos conjugales extendere vires", (Trojel's edition, p. 133). We know, too, from the author of the Eracle that she suggested to him the story of Athenais's adultery in Part II of that romance. 118 Evolution of Arthurian Romance sonant with such an ideal. To be sure, the heroine frequently expresses a horror of being associated in the minds of the people with Iseult. She exhibits, however, no repulsion for Iseult's sin, but merely for the evil reputation which the detection of her guilt had given the latter throughout the world.32 This main story Chretien has complicated and retarded by the episode of the hero's visit to Arthur's court, the most interesting feature of which is the well-known Three Tournament motif where the knight, incognito, wins the tournament on three successive days, appearing each day in a new colour — red, white, and black — still further, by the final episode, borrowed from the Tristan, in which Bertram detects the lovers together in an orchard — an incident which is common in tales of Eastern origin. No less important, however, for an insight into our poet's methods of composition is the first division of the Cliges — the story of the love of the hero's parents. As already observed, it was obviously suggested by the corresponding division of Chre- tien's Tristan original — the story of the love of Rivalin and Blanchefleur, the parents of Tristan; but the correspondence does not extend to details — at least, to the more important features. Paris has justly noted the inferiority of the imitation to the ori^- ginal in the essential matter of its connection with the main theme of the romance. The whole life of Tristan is determined by the tragical fate of Rivalin and Blanchefleur, whereas there is no such relation between the stories of Cliges and his parents. In this the Cliges, then, exhibits a manifest want of unity. Altogether the connection between the first and second di- vision of the narrative is full of blunders in composition.33 But the first division is, itself, also composite. For the rest, this first division contains a rather commonplace story of love between two young people of high rank. Doubtless, the fact that their love- affair begins at sea is a pale reflection of the same feature of the Tristan romance with the love potion left out; and similarly with some other features. The whole story, however, such as it So (t. Paris, loc. cit. G. Paris has enumerated them, Melanges, p. 285. Chretien de Troyes and his Successors 119 is, is interwoven with a supposed episode in Arthur's career, ac- cording to which the king goes abroad, appoints a regent in his absence, but is compelled to return, because of the treachery of this regent. There is an incident somewhat like this in the Tristan story, but, as I have noted above, there can be little doubt that the real original which Chretien here had in mind was Mordred's treachery to Arthur, under similar circumstances, as related in Geoffrey of Monmouth's Historia and Wace's Brut. Apart from the incidents, the monologues on the subject of love constitute a marked feature of this first division of the Cliges, but these, too, have their originals; they were suggested by the monologues of the Roman de Troie and Eneas — especially the latter — and are ultimately descended from Ovid and Virgil34 — for example, the monologues of Byblis and Myrrha in the Metamorphoses and Dido's speeches in the Aeneid, which, of course Chretien also knew at first hand.35 It will be observed that virtually every incident even of this first division of the poem can be traced to its source. It approaches, however, close to invention — certainly as near to invention as we can expect of a poet of the Middle Ages or of most poets of modern times — that is to say, Chretien, has taken widely sepa- rated details and brought them into a new combination, so as to produce the effect of originality. That the total impression in this case is not particularly striking is due to the fact that the 34 Cp., especially, E. Faral: Becherches sur les sources latines des contes et romans courtois du moyen age, pp. 73 ff. (Paris, 1913). 36 The Greek names of certain characters have led some scholars to imagine that Chretien used a lost Greek source for this division of his romance. Cp. E. Herzog, Liter aturblatt, XXV, 19 (1904). But these names, insofar as they are genuine, are derived from Ovid and the Old French romances of antiquity. Cp. Foerster's Cliges*, p. XXXVIII (Halle, 1910). In his "Byzantinisch-Geschichtliches im Cliges und Yvain", Zs. /. rom. Ph., XXXII, 400ff. (1908), F. Sette- gast sees in the Byzantine element of Cliges a reflex of the history of the Byzantine emperor Isaac I (1057 — 1059) and, especially, of his two nephews, Isaac and Alexios I. The latter reigned 1081 — 1118 and was well-known in the West, because of his relations to the First Crusade. On this fanciful hypothesis see Foerster, op. cit., pp. XXXVIf. 120 Evolution of Arthurian Romance elements combined do not individually possess any especial in- terest — far less, for instance, than attaches to the ancient story that underlies the second division. It cannot be too strongly emphasized that the Cliges con- stitutes the proper starting-point for any study of Chretien's rela- tion to his sources.36 We can analyze with certainty his prin- ciples of composition in the case of this poem, whose sources can be fixed, one may say, even in the minutest details. Now we may be reasonably sure that these principles were essentially the same in the poems whose sources we are unable to fix — at any rate, so exactly. And this is substantially true, whether we assume his originals to have been mainly Celtic or not. The analogy of Cliges then certainly tells against the theory37 as to a uniform Celtic tale's forming the basis of the Yvain. The narrative of that romance is also obviously a composite one and it is impos- sible to fit it into the scheme of some hypothetical Irish analogue. We have already considered the question of the sources of the Yvain in another connection and the same question in the case of the Lancelot and Perceval will be dealt with when we come to trace the history of these heroes in romance. In general, it is sufficient to say here that the sources are, in each instance, com-" posite and that the writer exercises a certain amount of invention in combining originally disparate elements — moreover, that these elements are only partly Celtic- Taken altogether, Chretien is undoubtedly the best of the French authors of metrical romances that deal with the matiere 36 This has been stressed by A. G. van Hamel, Romania, XXXIII, 486 (1904), and in his "Bydrage tot de vergelijking van Cliges en Tristan", Taal en Letter en, XIV, 193 ff. (Leiden, 1904) — especially, p. 211. On the other hand, Nitze, Romania, XLIV, 33 (1915) ob- jects that what is true of Cliges is not necessarily true of the rest — especially, since Cliges is a combination of an Oriental tale with an Arthurian theme, as the others are not. The last point, at least, however, is of no weight, for, as Faral, pp. 41 8 f., has observed, the division of the romances into four groups according to their subject- matter is largely artificial. 37 A. C. L. Brown, "The Knight of the Lion", PMLA., XX, 673 ff. (1905). Chretien de Troyes and his Successors 121 de Bretagne. In saying this, however, we are making an acknow- ledgement of the limitations of what was achieved in this genre — at least, in France; for in the works of this writer there is no question of the higher imagination, of philosophical insight into the riddles of existence, of "the dower of spanning wisdom", with regard to either character or the conduct of life, or of the magic of diction and phrase, which have distinguished the represen- tative poets of many other ages. His imagery is confined to a few similes — the majority of which are of a purely conventional kind — and to a restricted, though somewhat richer, store of metaphor. His "criticism of life" is merely that of a shrewd, alert, man of the world of his time. He was quite contented with the feudal society in which he moved and he delighted in the bustle and splendor of its festivities, its pageantry, and its tourna- ments. Within the bounds of this society, apart from the externals just mentioned, the code of chivalry and the problems of the rela- tions of the sexes — the latter especially in the new form which these problems had assumed under the system of the amour cour- tois — were the things that most attracted him. Moreover, living in a naive age, when in the elementary interests and emotions the grown man was nearer to the child than at present, he was keenly susceptible to the spell of the marvellous, as were his contempo- raries generally. Consequently, the setting which he gives to the life of chivalry and to his solution of the above-mentioned pro- blems is taken largely from the folk-tales of Celtic regions and of the Orient, where such fancies most abounded, with occasional admixture of classical motifs. In the case of the Perceval, per- haps, he did not understand the full significance of the materials that he drew from his sources, but, in general, the combination of the various elements of content and setting in his poems produces on the reader an effect of harmonious unity, and the creation of this new world in which mediaeval barons and ladies jostle fairies and even stranger Otherworld figures is no mean achievement. _. Since love is the dominant theme of Chretien's romances, it is natural that he should display most knowledge of the human heart in his characterizations of women. We are wearied occasion- ally, it is true, with the hair-splitting analyses of amorous emo- 122 Evolution of Arthurian Romance tions in the lovers' soliloquies of his romances and with the conceits which such analyses engender, but the patient loyalty of Enid, the, sovereign haughtiness of Guinevere and the piquant fickleness of Laudine in their relations with their lovers or husbands are depic- ted with much truth to nature and with no little charm, and, in the last-named case, also, with an effective touch of malice. It is particularly however, as a born conteur that Chretien can claim a notable place among the poets of the Middle Ages. In telling a story his sparkling vivacity never fails him, and, as a German critic has happily said, he makes the impression of a juggler who can shake couplets out of his sleeve as long as he pleases. He, himself, thoroughly enjoys telling his tale and his enjoyment is consequently contagious/ We have seen above that his work is marred in some instances by serious faults of construc- tion; but the same thing is true of Marmion and Rokeby — yet, in spite of these defects, Scott is, perhaps, our greatest English master of narrative poetry. Chretien, of course, paints on no such broad canvases and with no such variety of poetical resources as the nineteenth century writer. With greater vividness and agre- ment, nevertheless, than any other poet of the Middle Ages, he has preserved for us, at least, the outward lineaments of the society which Scott idealized. After all, however, his most memorable services to the great cause of poetry, perhaps, were in stimulating immeasurably the imagination of his contemporaries — for the vast forest of mediaeval Arthurian romance sprang mainly from the seeds of his sowing — and in enriching the whole poetic tradition of Europe with new and beautiful themes on which greater men than himself have exercised their genius, from the age immediately succeeding his own down to that of Tennyson and Wagner. During the life-time of Chretien there were other romance- writers like Gautier d'Arras, author of Eracle and Ille and Galeron, which, though similar in the most essential respects, cannot be classed as Arthurian romances, inasmuch as Arthur and his knights do not figure in them. No one, however, but Chretien, during these years, as far as the extant works show, made it his object to embody the ideals of contemporary feudal society in romances connected with Arthur.38 Thomas's Tristan was written about Chretien de Troyes and his Successors 123 1170 and those critics are probably right who see in it here and there the influence of Chretien rather than vice versa,39 but, apart from its dependence on the lost Tristan romance, the spirit of this work is somewhat different from Chretien's. The new ideals of chivalry occupy here a subordinate place in comparison with the passion of the lovers and the subterfuges to which they resort in the fulfilment of that passion. The Beroul Tristan fragments are, at least in the main, later than Chretien, but they belong, be- sides, to the tradition of the popular minstrels and are on a diffe- rent level from the poems we are dealing with. It may be that the supremacy of Chretien during his life- time deterred others from trying their hand in a genre which he had made his own. In any event, he was the object of admiration for the generations that immediately succeeded him. The fre- quent allusions to his works prove this, but even more decisively the variety of influences which they exercised on subsequent lite- rature. In the first place, we have from the closing years of the twelfth century and the first part of the thirteenth many bio- graphical and episodical romances which continue the genre. So, for instance, Li Biaus Descouneus (Le Bel Inconnu)*0 by Renaud de Beaujeu, on the theme which has also been treated in the middle English romance, Libeaus Desconus — itself based, of course, on some lost French original — that is to say, the theme of the un- known young knight who claims the first adventure after his arri- val at Arthur's court, astonishes the girl whose champion he has been made by his acts of prowess, visits the fairy isle, L'llle d'Or, resists its temptations and abolishes its bad customs, delivers the enchanted princess who has been turned into a serpent by venturing the fier baiser (dangerous kiss). Similar are the works of Raoul de Houdenc, at the beginning of the thirteenth century, which show, 38 On the probable mutual influence which Chretien and Gautier exerted upon each other, cp. p. 117 note 28 above. 39 Cp. M. Wilmotte, L' evolution du roman frangais, p. 67, and Foerster, Chretien WB, pp. 65 *f., as against G. Paris, Melanges, p. 282. *° For editions of this and the other metrical romances mentioned in this chapter, cp. the Bibliography at the end of the present work. 124 Evolution of Arthurian Romance to a marked degree, the influence of his great predecessor — the Meraugis de Portlesguez, where the heroine, Lidoine, exemplifies the idea that courtesy — the perfection of courtly breeding — is superior to mere beauty, and the Vengeance de Raguidel, which re- counts the vengeance that Gawain took for a knight named Ra- guidel, who was slain before the action of the romance begins. These are only a few of those that have survived. The list drawn up by Gaston Paris41 includes thirty-seven romances, and some additional ones have come to light since the publication of his treatise. Even now (1921) one of the very best of Paris's list remains unedited — namely, Gliglois. And so the Arthurian me- trical romance drags on into the fourteenth century until it ex- pires with the Meliador of Froissart in the second half of that century.42 The influence of this product of French genius outside of France is well-known to all the world. In Germany, before the end of the twelfth century, we have in the Lanzelet of Ulrich von Zatzikhoven an adaptation from a French romance, and, about the same time, the Erec and Iwein of Hartmann von der Aue — paraphrased, of course, from Chretien's romances of the same names. Then come Gottfried von Strassburg's adaptation of Thomas's Tristan and Wolfram von Eschenbach's Parzival, based, according to his own statement, on the work of an enigmatical Kyot — a French poet, of course. Both of these illustrious writers belong to the early thirteenth century and there are others less illustrious of the same nation who about the same time drew likewise from the vast store-house of the French metrical romances. Especially to be noted is Heinrich von Tiirlin (about 1220) whose Crone preserves in German dress much lost French material which we know only from this poem. The same thing is true of the vast *l At the beginning of the Histoire litteraire de la France, XXX (1888). 42 For references to Arthur in works outside of the Arthurian romances, cp. F. L. Critchlow, "Arthur in Old French Poetry not of the Breton Cycle", MPh., VI, 477 ff. (1909). The author's list, however, is by no means exhaustive. For analyses of these romances see Part V, below. Chretien de Troyes and his Successors 125 thirteenth century Dutch compilation in short-rhymed couplets called the Lancelot.*3 Its 87000 lines (and somewhat more) em- braces, besides large portions of the prose-romances of the Ar- thurian cycle, more than one metrical romance which in its original French form has perished. Owing to the special conditions which prevailed in England after the Norman Conquest, when the language of the patron-class was for many generations French, the English were slow in adap- ting in their language these most characteristic reflections of the life of the ruling classes in the Middle Ages. As long as the language of those classes was French, there was, of course, no incentive to produce English versions of the French originals, and when at last there was such an incentive in the growing decline of French in familiar use among the upper classes, these poems belonged to a past age and no longer represented, at least with such perfection, the spirit of contemporary society. Perhaps, this may account in part, at least, for the fact that the Middle English romances, on the whole, are inferior to German production in the same genre; for the latter, it will be recalled, fell at the end of the twelfth century and beginning of the thirteenth — that is to say, it was coincident with all but the first period of production in France itself and so was a vivid representation of actual con- temporary ideals. It is significant that the best English Arthu- rian romances — those which seem addressed to the more cultured classes and not to rude popular audiences — belong to the north — the most backward part of the kingdom, where the old feudal ideals and the literary tastes corresponding were least changed. To this region belongs the gem of the English romances, Sir Gawain and the Green Knight, which was composed about the middle of the fourteenth century — at a time when the metrical romance on the continent was utterly decadent. Notwithstanding the impression of uncouthness which Middle English alliteration always makes on the modern reader, this romance, by the general verdict, belongs among the best productions of the Arthurian muse 48 Edited by W. J. A. Jonckbloet, 2. vols. (The Hague, 1846 -1849). 126 Evolution of Arthurian Romance in the Middle Ages. It is safe to say, too, that although derived from a lost French original, it contains a larger share of origi- nal invention than is the case with any other English specimen of the genre. A similar vigor and a similar uncouthness mark the alliterative Morte Arthure, which was produced in the same general region. It is a pity that the large proportion of words of French and Scandinavian origin in this romance and in Sir Gawain and the Green Knight that time has rendered obsolete should have rendered these fine poems sealed books to all but pro- fessional students at the present day. The earliest Middle English metrical romance that has come down to us is the King Horn, which dates from the middle of the thirteenth century, and the earliest Middle English metrical romance connected with the Arthurian cycle is the Sir Tristrem which belongs to the close of that century. The latter is a rude poem in stanzas of eleven lines and it was composed in the North- west of Midland England. It is based on Thomas's Tristan, but, like virtually all the Middle English romances, whether Arthurian or not, it condenses the original. The English audiences of the time were evidently unequal to the task of appreciating the re- finements of the French romances and the poets themselves stood - on a lower level of culture than the authors of their originals.44 To the following century, however, — the fourteenth — be- long the great bulk of the English metrical romances, including those with Arthurian connections. One of the most important of these, the alliterative Morte Arthure, which is often assigned to the Scottish knight, Huchown of the Awle Ryale, although it more likely originated in Northern England, represents, as we have already seen, rather one of the later developments of the chronicles in verse than the metrical romance proper, as we have it in the works of Chretien and his followers. Moreover, a large number of the Middle English metrical romances are derived from 44> ) It has been contended by Bossert, Kolbing etc. that the English poet was not following a written text, but was merely working up in his own language the narrative of Thomas, as he recollected it from a previous reading or from having heard it read. For a con- trary view, cp. Bedier's edition of Thomas's Tristan, II, 87 f. Chretien de Troyes and his Successors 127 French romances in prose and not in verse. This is the case with the stanzaic Morte Arthur of the end of the fourteenth century — a poem of much naive charm, whose ultimate original is the great French prose romance of the early thirteenth century, cal- led the Mort Artu, (Mort Artus, Mort d 'Arthur).*5 It is also the case with the Middle English poems that deal with the Holy Grail theme — namely the alliterative Joseph of Arimathie (Mid- land dialect about 1350) and Lovelich's Holy Grail (South or South Midland, about 1480), both based on the thirteenth cen- tury Estoire del Saint Graal (Grand St. Graal) — one of the great French prose romances of the Vulgate cycle. It is the case, also, with the Arthour and Merlin (end of thirteenth century) and the Merlin of Lovelich, the fifteenth century poet, both of which are derived from the Old French prose Merlin in the same cycle of the prose romances as the Estoire del Saint Graal. So too with the Scotch Lancelot of the Laik of the late fifteenth century, which merely paraphrases an episode of the famous Lancelot del Lac — the prose romance of the early thirteenth century. This is in- teresting as the latest in date of all the metrical romances com- posed in Great Britain. It must be confessed that with the exception of the stanzaic Morte Arthur, none of the English romances in verse that are based on French romances in prose rise above mediocrity. Love- lich's works, especially, are monuments of tediousness. But, besides the English Arthurian romances in verse which are based on French romances in prose, we have, in addition to Sir Trislrem, a number of others which are based on French ro- mances in verse and which are markedly superior in literary qua- lity to the former. First, the Sir Perceval of Galles which is, probably, a mere adaptation of Chretien's Perceval 46) — then the Libeaus Desconus, mentioned above, whose lost French original was closely akin to the Li Biaus Descouneus of Renaud de Beaujeu — the Yvain and Gawain, an adaptation of Chretien's Yvain — Sir Gawain and the Green Knight (already mentioned), whose French original is also lost — the Au/iters of Arthur, made up, 46 Cp. Part IH, below. 46 Cp. Part II, below. 128 Evolution of Arthurian Romance it seems, from various French sources, which is also probably true of the Avowing of Arthur, preserved in the same MS. and appa- rently of the same date. All of the romances just enumerated were composed in the North Midland or Northern districts of England during the fourteenth century. The Scotch Gologrus and Gawain, the story of which is taken from the continuation of Chretien's Perceval, belongs to the fifteenth century. Still further, five brief romances, hardly more than ballads, concerning Gawain have been printed by Sir Frederick Madden in his great edition of the ro- mances47 relating to that hero published by the Bannatyne Club, London, 1839. If we add in conclusion a version of Marie de France's Lanval*8 and the brief fragment called Arthur which Dr. Furnivall published in the second issue of the Early English Text Society,49 we have completed the list of the Middle English Arthurian romances. This last piece is a very condensed narrative based on Geoffrey's Historia. 47 Sir Gawayne: a collection of Ancient Romance-Poems by Scottish and English Authors relating to that celebrated Knight of the Bound Table, with Introduction, notes and a glossary. London, 1839. For everything relating I., xr-ca poems — sources etc. — cp. G. L. Kittredge, A Study of Gawain and the Green Knight (Boston, 1914). They are entitled respectively, (1) The Green Knight (fifteenth century), which is derived from the English romance, Sir Gawain and the Green Knight. (2) The Turk and Gowin (fifteenth cen- tury). (3) The Jeaste of Syr Gawayne (fifteenth century). (4) Syre Gawene and the Carle of Carelyle (fifteenth century). (5) The Weddynge of Syr Gawen and Dame Ragnell (fifteenth century). Part IV, below, I have discussed the use which Miss Weston makes of these poems in her hypothesis of a primitive "Gawain-complex". 48 Last edited by Rudolf Zimmermann at Konigsberg in Preufeen, 1900. For previous editions and literature of the poem, see this edition or Miss A. H. Billings's A. Guide to the Middle English Metrical Romances, pp. 144 ff. (New York, 1900). The author was Thomas Chestre and he wrote in the fourteenth century — probably the first half. 49 Arthur, a short sketch of his life and history in English verse of the first half of the fifteenth century, copied and edited from the Marquis of Bath's MS. By F. J. Furnivall. E. E. T. S. Original Series, no 2. London 1864. It was composed about 1400 in the Southern dialect. The fragmentary poem (only 642 four-stressed lines) is based on one of the Bruts. Chapter IV. Merlin. It is only in the writings of Welshmen that we find any trace of a character corresponding to Merlin before the Arthurian romances. As has been said above more than once, literary re- cords from mediaeval Brittany are practically non-existent and the Breton ballads about Marzin or Merlin which in the last cen- tury La Villemarque published in his Barzaz-Breiz and Myrhddin, (having collected them, as he pretended, from oral tradition) are now universally regarded as forgeries.1 Indeed, the lady from the neighborhood of Molaix, whom he cites as an authority for these ballads, has been pronounced by one critic to be probably a sister of the imaginary Mrs. Harris in Dickens's Martin Chuz- zlewit. Furthermore, Merlin, like the other characters of Ar- thurian romance, does not, of course, appear in the literature of the Gaels. Now, with regard to the Welsh evidence: In the body of ancient Welsh poetry2 there are eight poems which purport to be the compositions of a bard named Myrddin. This is evi- dently the same name as the Merlinus of Geoffrey of Monmouth — only as Gaston Paris has remarked,3 Geoffrey probably shrank from Latinizing the Welsh name in the natural way as Merdinus, owing to the similarity to a French word of unpleasant associations which would have resulted, and so changed the d to an Z. But the value of these poems as evidence of traditions concerning Merlin among the Welsh is seriously impaired by the doubt which at- taches to their antiquity. The result of the most authoritative 1 Cp. pp. 39 ff. and notes. 2 Edited by W. F. Skene, Four Ancient Books of Wales, 2 vols. (Edinburgh, 1868). 3 Romania, XII, 376 (1883). 130 Evolution of Arthurian Romance examination of this question that has been made4 is that all the Welsh poems in question were really composed later than Geoffrey of Monmouth, with the single exception of the Dialogue of Merlin and Taliesin,5 which is either contemporary with Geoffrey of Monmouth or perhaps slightly earlier.6 Moreover, all of them, except the Dialogue, betray the use of Geoffrey's Vita Merlini.1 Nevertheless, these poems give evidence as to the existence in Welsh tradition of Merlin and the personages associated with him in the Vita Merlini — only the few concrete allusions they con- tain bearing on the history of Merlin seem derived from Geoffrey. Besides, after all, they merely show that Merlin was a bard like Taliesin, who certainly enjoyed greater fame among the ancient Welsh than Merlin ever did. The Dialogue, which, as we have seen, is the least open to suspicion of the Welsh poems gives us a conversation between Myrddin (Merlin) and Taliesin concerning 4 By F. Lot, "Etudes sur Merlin", Annates de Bretagne, XV, 324ff., 505 ff. (1900). 5 Skene, I, 368ff. (translation), II, 3 f f . (Welsh text). 6 In Y Cymmrodor, XI, 48 (1892), E. Phillimore, one of the best Welsh authorities, speaks of the Welsh Merlin-poems "following-" Geoffrey's Vita Merlini in making the hero go out of his mind, in consequence of the battle of Arderydd, so that he, too, evidently re- gards these poems as subsequent to Geoffrey. In the same article ("Additional Notes" to J. E. Lloyd's "Welsh Place-Names", ibid., pp. 15 ff.) this scholar interprets Myrddin (the Welsh name for Merlin) as a mere eponymus "derived from the place-name Caerfyrddin (= modern Carmarthen), like Efrog (Peredur's father in the Peredur) from Caer Efrog (Modern English York) or Lleon Gawr from Caerleon (Modern English Chester). And again, pp. 47 f., he remarks that, in changing the narrative of Nennius (his source) and putting "the fatherless boy" (Geoffrey's Merlin) at Carmarthen, "Geoffrey had here in his mind's eye the connection of Myrddin with Caerfyrddin'' — that, moreover, Geoffrey in the Vita Merlini makes Merlin king of Dyfed (his mother in the Historia was already daughter of a king of Dyfed), because in those days (down to 1132) Dyfed included Carmarthen. It would seem, however, from Lot's study that, even if the name Myrddin is derived from Caerfyrddin, the derivation is older than Geoffrey. May not, indeed, the latter name be derived from the former? 7 Cp. Lot, op. cit., p. 515. Merlin 131 the battle of Arderydd in which Merlin expresses sadness at the slaughter. This battle, it may be remarked, was fought at Ar- thuret, near Carlisle, about the year 575, between Rhydderch ap Tudwal, the friend of St. Columba, and another Celtic chieftain of the North, Gwenddoleu ap Ceidio, the former being victorious.8 The poem, however, is so meagre and its meaning so dark, that we merely gather from it that Merlin is a bard. At the end of the Dialogue we have the obscure words: Since I, Myrddin, am next after Taliessin, Let my prediction become common. This is the only hint in the poem that Merlin has prophetic gifts — indeed, the only thing in works not demonstrably depen- dent on Geoffrey that suggests the possession of prophetic powers on the part of Merlin in all Welsh literature. But, apart from the doubtful antiquity even of the Dialogue, this would be a slender basis for supposing that the Welsh regarded Merlin as a prophet, especially as, notwithstanding the words just quoted, there is actually no prediction in the poem — a fact which excites the suspicion9 that these words may have been a later addition. If the poems which we have been considering throw little light on the subject of Merlin's place in Welsh tradition, it is even worse with the remaining Welsh records — the Triads and the Mabinogion. Two of the former mention Merlin — the first of them as one of the three chief bards of Britain, the second in connection with his marvellous disappearance. But these triads are late10 and, hence, worthless for our purpose. The Mabinogion do not mention Merlin by name at all and the only connection whatever between the tales in this book and the traditions con- cerning Merlin is that in one of them, Lludd and Llevelys, there is a story of a combat between a white and a red dragon which figures in the narrative of Merlin and Vortigern's tower in Geof- 8 Cp. J. E. Lloyd's History of Wales, I, 166. 9 Cp. Lot, op. cit, p. 535, note 2. 10 Cp. E. Phillimore's observations on the subject in the Early English Text Society's edition of the Middle English Merlin, Part IV, p. XCVin, note 3. 132 Evolution of Arthurian Romance frey's Historia. This same story, to be sure, appears in Nennius connected with another name. The conclusion to be drawn from extant Welsh literature, then, is, evidently, that Merlin was of relatively little importance in Welsh tradition, and, if we depended merely on the mention that the above records make of him, we should say that he was known to that tradition simply as a bard — not as a magician or prophet. It will simplify subsequent discussion, if we state at once the conclusion to which we are forced by all the evidence — namely, that Merlin owes his fortune in the history of fiction and popular tradition to Geoffrey of Monmouth. He is virtually the creation of Geoffrey. Let us take up now the history of this wondrous being. In the Historia Brittonum of Nennius (ch. 40 — 42) u we have the following story concerning the tower which Vortigern tried to build on Mount Hereri or Snowdon as a refuge from the Saxons: Workmen begin to lay the foundations of the building, but on three successive nights the work done during the preceding day is destroyed. Vortigern's wise men inform him that, before the tower can be built, the ground must be sprinkled with the blood of a child born without a father. The king at once sends out messengers in search of such a child. When they arrive at the field of Elleti in Gleguissing they hear a boy jeer at a comrade, because he has never had a father, and the boy's mother confirms the truth of the taunt and asserts that the child is, indeed, the son of no mortal man. The messengers accordingly take him to Vortigern whom the boy proceeds to question shrewdly, until the king admits why he has been brought thither. Then the child orders the wise men to declare what there is beneath the spot where Vortigern wishes to build. When they say that they do not know, he bids them dig into the ground where they will find a pond in which there are two vases; in the vases is a folded tent and in the tent are two sleeping dragons, one white and the other red. The men dig and find that the boy's words are true. Suddenly the dragons 11 See Mommsen's edition, pp. 181 ff., Monumenta Germaniae Historica, 1904. Merlin 133 begin a terrible combat with each other, in which the red dragon succeeds in routing the white one. The boy proceeds to explain to the king that the pool signifies the world, the tent Britain, the red dragon the British nation, the white dragon the Saxons. Vor- tigern, he adds, must depart from this place, but he himself to whom fate has allotted it will remain there. The king asks him his name, and he replies: "I am called Ambrosius."12 The king next asks as to his origin and he replies that he is the son of a Roman consul. This is, of course, a flat contradiction of what was said at the beginning of the episode. Vortigern obediently assigns the site of the stronghold to Ambrosius with all the other provinces of Britain and himself departs elsewhere. Later Nennius tells us that, after the death of Vortigern, his son, Pascentius, received two provinces from Ambrosius, "who was the great king among the kings of Britain." Miss L. A. Paton has justly observed13 that we evidently have here a reminiscence of the barbarous " custom of offering a human being to the deity of a selected site as a foundation sacri- fice". Whether, as she maintains, the two dragons were origi- nally two supernatural shape-shifters who had assumed for the time being the form of dragons is, I believe, much less certain. The part which Ambrosius plays in this affair14 seems borrowed from one of the numerous Talmudic legends concerning Solomon in which the wise king by a ruse captures the demon Aschmedai (or Asmodeus, as we are accustomed to call him) and makes him 12 Nennius's Ambrosius is identical, of course, with the Ambrosius Aurelianus who, according to the historian Gildas, led the Britons against the Anglo-Saxon invaders after the withdrawal of the Roman legions. E. Phillimore remarks, Y Cymmrodor, XI, 48 f. (1892), that the identity is proved by the gloss which is found in all the MSS. of Nennius after the words "Ambrosius vocor", in this passage, viz., "id est Embreis Guletic". Now, Gwledig (Guletic), which is equi- valent to the Latin Princeps or Imperator, could indicate here no one save the historical chieftain named by Gildas. 13) In her article on this episode, "The Story of Vortigern' s Tower — an Analysis", Radcliffe College Monographs, no. 15, pp. 14ft. (1910). 14 Cp. Miss Paton, op. tit., p. 18, note 1. 134 Evolution of Arthurian Romance tell how he can obtain possession of a certain wonderful creature who is necessary in the building of the Temple of Jerusalem. Now this story of Nennius was used by Geoffrey of Mon- mouth (Book VI, ch. 17 — 19) with certain differences of which the most important are as follows: He calls the boy without a father Merlin as well as Ambrosius.15 Moreover, the messengers take Merlin's mother, who is the daughter of a king of Demetia, to Vortigern with the lad, and she gives the king a full account of Merlin's birth, explaining that his father was an incubus.16 Just after this story Geoffrey inserts the famous prophecies of Merlin, which, as a matter of fact, in large part, refer to recent political events of the author's own time. They make up the entire Seventh Book of his Historia, but had already been in cir- culation as a separate work.17 The next book begins with a pro- 15 Cp. Geoffrey, Book VI, ch. 19: Tunc ait Merlinus, qui et Am- brosias dicebatur". Similarly, VII, 3. In his "Merlin and Ambrosius", Kittredge Anniversary Papers, pp. 119ff., G. H. Maynadier explains (p. 125) Geoffrey's identification of Merlin with Ambrosius as follows: Nennius says, ch. 41, that Am- brosius was found in the region Gleguissing, which, even if it did not actually include the vale of the river Towy, was, in any event, near that stream. Now, Caermarthen, which is connected by name with Merlin (Myrddin) — see p. 130, note 6 above — lies on the Towy. Hence, "nothing would be more natural than for Geoffrey to combine the two men". — Apart, however, from our ignorance as to the extent of Gleguissing, the vague geographical connection here assumed seems unnecessary for the identification. The inclusion of a great magician in his work was suggested to Geoffrey by Nennius's Ambrosius and the designation, Ambrosius Merlinus, testifies to this historical connection between the two characters, but Geoffrey's enchanter was to have an altogether different scope from Nennius's — as a matter of fact, he was destined to prove one of the greatest creations of the mediaeval imagination — and so he calls him, after a brief period of hesitation, exclusively by the new name. 16 This notion of the sexual commerce of incubi with mortals was common in the Middle Ages. 17 On Geoffrey's so-called Libellus Merlini and the numerous political phrophecies which were modelled after it in the Middle Ages, see San Marte, Die Sagen von Merlin, pp. 32 ff., (Halle, 1853) and Rupert Taylor, The Political Prophecy in England, (New York, 1911), Merlin 135 phecy of Merlin concerning the death of Vortigern and, later on, tells of his exploit of moving the stones of Stonehenge — or Giants' Dance, as it is called — from Mount Killaraus in Ireland to their present site. In this affair, which was probably invented by Geoffrey, Merlin appears as hardly more than a marvellous engineer. He reappears, however, (ch. 15), as interpreting to Uther Pendragon, the father of Arthur, a portent which betokens the death of Aurelius Ambrosius — the Ambrosius Aurelianus of Gildas — whose gallant deeds constitute a large part of the narrative of this Eighth Book. Then follows (ch. 19 — 20) the last and most famous of Merlin's achievements related by Geof- frey — his transformation of Uther into the likeness of Gorlois, Duke of Cornwall, so that he might have access to Igerna, the wife of Gorlois — on which occasion he begets Arthur. It has often been maintained, and no doubt correctly, that this episode is merely an adaptation of the well-known story of the conception of Hercules — how Jupiter assumed the appearance of Amphitruo and in the latter' s absence deceived Alcmena and begot the fa- mous hero. Looking back over the Merlin material in Geoffrey's Historia, which I have just summarized, it seems in the highest degree unlikely that any of it was derived from oral tradition. Nennius, classical legend, and his own fancy seem to account for the whole. Things are not so plain — at first blush, at any rate — when we come to the next important Merlin document — namely, the poem in Latin hexameters called the Vita Merlini — which is also commonly ascribed to Geoffrey of Monmouth. To be sure, the correctness of this ascription which is found at the end of the poem itself, has at various times been disputed,18 but the fact that in Columbia University Studies in English. Despite the title of the latter, it contains also, ch. 6, a discussion of the Galfridian type of prophecy in other countries than England. There is, in addition, a thirteenth-century French (prose) work entitled Les Prophecies de Merlin, which, notwithstanding its title, is very different from the Libellus Merlini. On this later work cp. Part III, below. 18 Cp. San Marte's Die Sagen von Merlin, pp. 273 ff. and various scholars cited by H. L. D. Ward, Catalogue of Romances in the De- 136 Evolution of Arthurian Romance in the concluding lines of the poem the Historia is spoken of as celebrated throughout the world is no reason for doubting the authenticity of the ascription. Modesty was certainly not the strong point of the writer who tried to palm off the Historia as an authentic narrative. Moreover, the prophecy of Ganieda found in this poem (11. 1474 ff.) relates to actual political events during the years 1139 — 1144,19 and 1148 — 9 has been established as its date.20 These facts support the view that the work is really Geoffrey's. It is, indeed, singular that a poem by so famous an author should exist in only one complete MS. (Vespasian E. IV, of the British Museum), and that of the late thirteenth century — yet this does not justify us in rejecting it as spurious. The Vita Merlini is the most original of all the works con- cerning Merlin which the Middle Ages have left us, and parts of it are genuinely poetical — so that it will be necessary to consider for a moment this curious production, the action of which is sup- posed to fall after the translation of Arthur to Avalon. Following is a condensed analysis of its contents: Merlin was a king and a prophet who had already lived an indefinite age, when a war breaks out between the princes of Britain, Peredur, King of Venedotia (North Wales), being the leader on one side, and Guennolous, King of Scotland, on the other. Merlin and Rodarchus, King of Cumberland, were with Peredur. The Scots are routed, but Merlin loses three brothers in the battle. This misfortune plunges him into grief and madness and he flies to the desert. There he becomes a wild man (Sil- vester homo), lives on fruits, and fills the Caledonian forest with his lamentations. Rodarchus, however, who had married Ganieda, Merlin's sister, tried to bring back the fugitive to his court. At partnient of Manuscripts in the British Museum, I, 278 ff. So far as I know, however, Brugger is the only living scholar who denies the poem to Geoffrey. For his views cf. Zs. f. frz. Spr. u. Litt., XX1, 105 (1897) and ibid., XXX1, 217, note (1906). For an answer to the objections which have been raised to Geoffrey's authorship of the Vita Merlini, cp. Ward, loc. cit. 19 Cp. Ward, op. cit, pp. 282 ff. 20 Cp. Ward, pp. 285 f. and F. Lot, Annates de Bretagne, XV, 333. Merlin 137 last, this is effected by a messenger, who finds Merlin seated by a spring in the forest and watching the play of the wild ani- mals. This man sings to a "cithara" the woes of Merlin's wife, Gwendoloena, on account of her absent husband. The music heals Merlin of his madness and he allows himself to be led back to the court of Rodarchus. But there his madness soon returns on him and he is eager to return to the oaks and high mountains of the forest. To prevent his escape, Eodarchus has him put in bonds. Merlin falls into a profound melancholy and does not utter a word. One day the queen was traversing the hall, when her husband took her by the hand and made her sit down; but, as he leant over to kiss her, he observed a leaf in her hair. He took it and threw it on the ground. Merlin burst into laughter at this. Everyone was astonished, but the sage refused to explain his singu- lar merriment, unless he was set at liberty. Rodarchus had him unbound and Merlin explains that the leaf which the king had just removed from his wife's hair had fallen there shortly be- fore as she was passing through the thicket for a rendezvous with her lover. The queen tries to discredit Merlin's power of divi- nation by putting him to a test. She has the same child brought in three times in different costumes, so as to deceive the enchanter as to his identity, and Merlin each time predicts a different kind of death for him: he will fall from a high rock — he will die on a tree — he will be drowned. The queen is exultant over the success of her ruse, but, as a matter of fact, Merlin's prophecies all come true; for the child, when he grew up, whilst chasing a stag, fell from a high rock and was drowned in a stream that flowed at its base, but his feet caught in a tree and there he was suspended. Merlin returns to the forest and gives permission to Gwen- doloena to marry another man, but the bridegroom must beware of letting Merlin see him. One bright night the seer reads in the stars that his wife has taken him at his word and is about to marry. He assembles a great herd of wild deer and goats and mounts a stag. Driving the remaining animals before him, he goes to his wife's house and calls her out to see the wedding pre- sents he has brought her. She naturally smiles at the spectacle, 138 Evolution of Arthurian Romance but the bridegroom puts his head out of the window and laughs. Filled with rage, Merlin tears off the antlers of the stag he is riding, hurls them at his unlucky rival and kills him on the spot. He then flees, but is stopped by a torrent, made prisoner, and brought back before his sister in chains- He falls into a sullen dejection and refuses food. To distract him, his brother-in-law has him taken out for a walk through the town, but at the palace gate they meet a porter miserably clad, asking alms of the passers- by. Merlin bursts out laughing and acts similarly when, a little further on, in the market-place, he sees a young man buying shoes and pieces of leather to repair them with. In both cases Merlin refused to explain his laughter, unless his freedom was to be the reward. Rodarchus at last consents, and Merlin explains that the beggar had at that very time a hidden treasure under his feet and the young man was drowned immediately after making his purchases. The enchanter returns again to the wilderness. His sister wishes to follow him, but he prevents her — only, later on, he per- mits her to visit him and bring him food in the cold season. He gets her, moreover, to build him a house in the forest with 70 doors and 70 windows, so that he can sit within and predict the future by observation of the heavens. His prophecies are taken down by 140 scribes. Here he foretells most of the calamities which are to befall the Britons — including the death of Rodarchus, after which event Ganieda resolves to fly from the world and live with her brother. Taliesin joins Merlin and gives a long description (11. 737 ff.) of the universe,21 winding up with the famous account (11. 908 ff.) of the Happy Isle, where Morgan lives with her eight sisters.22 21 The immediate source of this description, as of several other similar descriptions (11. 827ff., 859ff., 1179ff„ 1298ff.), in the Vita MerUni, is the Etymologiae of Isidore of Seville, who, in turn, derived his information from Pliny's Historia Naturalis. Cp. F. Lot, Ro- mania, XLV, Iff. (1918—1919). "Insula pomorum quae Fortunata vocatur, Ex re nomen habet, quia per se singula profert: Non opus est illi sulcantibus arva colonis; Omnis abest cultus nisi quern natura ministrat: Merlin 139 Thither Taliesin in a boat, steered by Barinthus,23 has taken Arthur Ultro foecundas segetes producit et uvas, Nataque poma suis praetonso germine silvis; Omnia gignit humus vice graminis ultro redundans. Anuis centenis aut ultra vivitur illic, lllic iura novem geniali lege sorores Dant his qui veniunt nostris ex partibus ad se: Quarum quae prior est fit doctior arte medendi, Exceditque suas forma praestante sorores; Morgeu ei nomen, didicitque quid utilitatis Gramina cuncta ferant, ut languida corpora curet; Ars quoque nota sibi qua scit mutare figuram, Et resecare novis quasi Daedalus aera pennis; Cum vult est Bristi, Carnoti, sive Papiae Cum vult in nostris ex aere labitur horis. Hancque mathematicam dicunt didicisse sorores, Moronoe, Mazoe, Gliten, Glitonea, Gliton, Tironoe, Thiten, cithara notissima Thiten. Illuc, post bellum Camblani, vulnere laesum Duximus Arcturum, nos conducente Barintho, Aequora cui fuerant et coeli sydera nota. Hoc rectore ratis, cum principe venimus illuc, Et nos quo decuit Morgen suscepit honore, lnque suis talamis posuit super aurea regem Strata [MS. has stulta] manuque sibi detexit vulnus honesta. Inspexitque diu; tandemque redire salutem Posse sibi dixit, si secum tempore longo Esset, et ipsius vellet medicamine fungi. Gaudentes igitur regem commisimus illi, Et dedimus ventis redeundo vela secundis." Only in this passage is Morgan represented as one of the nine sisters. For a plausible explanation of the names of these sisters see L. A. Paton's Fairy Mythology, p. 44, note 2. There is a similar description in the Gesta Regum Britanniae — a thirteenth century adaptation in verse of Geoffrey's Historia. The island, however, is unnamed and the healer is merely called regia v'xrgo. For the text of this passage and a discussion of the same see Bruce, Mort Artu, p. 300. 23 In the Revue Celtigue, XXII, 339 ff. (1901), A. C. L. Brown has pointed out that this character is really a seagod, who, like Manannan in the Welsh tales, conducted voyagers to the Celtic Other- world. We have a similar Christian adaptation of this idea in the Navigatio Brendani. 140 Evolution of Arthurian Romance aiter the battle with Mordred, in the hope that the fay might be able to bring him back to life. Merlin then laments over the misfortunes which the Saxons are destined to inflict on the Britons; but no one now alive will see that time. Conan of Brittany and Cadwallader, however, will expel the invaders. During the dialogue of the two friends an attendant announces that a fountain of pure water has sprung up in the valley. Merlin drinks of it, to quench his thirst, but it also restores him to reason. He thanks the creator, and Taliesin takes occasion (11. 1179ff.) to describe the celebrated fountains of the world. At the end of this conversation Merlin is again plunged into sadness by the sight of a madman who comes up and whom he recognizes as a friend of his youth. The madman had been a noble and worthy knight. He and some of his companions had gone hunting with Merlin. They came upon a spring with fruits lying about them. Merlin picked them up and distributed them so generously among his friends that there was not one left for himself. But the men who ate them were deprived of reason and ran away like wolves. The poisoned fruit had really been inten- ded for Merlin by a quondam mistress of his whom he had aban- doned (11. 1422 ff.). They take the poor madman now, compel him to drink of the spring, and he recovers his reason. He an- nounces his purpose of remaining in the wilderness with Merlin. Taliesin sends home the kings who had assembled to hear Merlin's prophecies and the enchanter is left with his two friends and his sister. The latter, Ganieda, is, in her turn, seized with the pro- phetic gift and declares a number of marvels. With her speech the poem ends — only there is an epilogue in which Geoffrey bids the Britons weave a crown for him. On first reading this strange narrative concerning the great enchanter of Celtic legend, as we are accustomed to call him, one is naturally inclined to take for granted that the materials on which it is based are Celtic — but a closer analysis of its elements reveals the fact that the sources are only to a very limited degree Celtic, in any proper sense, and that the Orient has supplied a far larger proportion of the motifs.2* The longest step made in the Merlin 141 analysis of the Merlin legend we owe to the late H. L. D. Ward, who published25 in 1893 the so-called Lailoken fragments preser- ved in the British Museum MS., Titus A. XIX. Here we have the most important single source for the Vita Merlini. To be sure, Ward does not seem to have been fully aware of the im- portance of his own discovery. There are two fragments in the manuscript just named. Both relate to a madman named Lailoken, but the second is later than the first, and its author seems to have been familiar with the first fragment. In the first fragment at one point Merlin is identified with Lailoken, but this sentence, as Ward has observed, is the addition of a later hand under the influence of Geoffrey. In the first of these fragments Lailoken gives the triple prediction of the death of one person (only the person is himself). In the second we have the incident of the leaf which betrays the queen's adul- tery. The fragments give these narratives in a more consistent and primitive form than the Vita Merlini, and there can be no question that Geoffrey made use of these Lailoken stories and possibly of others concerning the same personage that have not survived. The whole conception of Merlin in the Latin poem is borrowed from this earlier madman and prophet, and the one ranges the Caledonian forest, because the other had done so be- fore him. In the words of Phillimore,26 the eminent Welsh scholar, Merlin has simply stepped into the shoes of Lailoken. 24 This has been proved by Ward and Lot in the discussions which T have already cited above. 25 In "Lailoken or Merlin Silvester", Romania, XXII, 504ff. (1893). 28 Cp. Y Cymmrodor, XI, 45 f. (1892). He observes, also, (p. 48): "With Geoffrey the identification of Merlin and Lailoken was complete; but with the later Scottish writers it was still a matter of doubt." Phillimore says, loc. cit., that Lailoken (Lailocen) "is mentioned in ch. XLV of Jocelyn of Furness' Life of St. Kentigern as Laloecen or Laloicen, a fool at the court of Rhydderch Hael, who possessed the gift of prophecy; in the Welsh Merlin poems this word (there used as a name or epithet of Merlin) is made into Llalogan, which has been explained as meaning 'twin-brother'; but there is plenty of evidence that it was a personal name." He then cites (p. 48) Lalocan in the Cartulaire de Redon, 125 (already cited by Ward) and the simpler form, Lalocc, as a woman's name, in Whitley Stokes' Tripartite Life of St, Patrick, pp. 82, 104, 317. 142 Evolution of Arthurian Romance According to the first fragment of the Cottonian MS., Lai- loken had an evil disposition and, like the Irish Bricriu, was con- stantly stirring up discord among his compatriots. As he was one day watching a battle, however, (localized, it would seem, in the neighborhood of Carlisle) which was due to his malicious efforts, a voice from heaven reproaches him with the responsibi- lity and condemns him henceforth to a life among the beasts. He also saw a supernatural light and hosts of angels casting their lances at him. At this sight he loses his reason and flies to the desert. But he would often come to a rock in sight of Glasgow and utter predictions which people took down in writing. One day whilst Saint Kentigern, the famous Scottish saint, is cele- brating mass, Lailoken disturbs the ceremony by howling and de- manding communion. Kentigern sends a messenger to bid him be quiet — but without success. The messenger goes three times and each time the madman predicts that he is about to die — stating, however, in each instance a different mode of death. He adds also a prediction of the impending death of three eminent personages in Britain. After this he runs away, but the same day the triple prediction of his own death is fulfilled. In the second fragment Lailoken is captured by king Meldred. and kept in chains at Dunmeller. He will not gratify his cap- tors, however, by uttering prophecies — on the contrary, for three days he will neither eat nor speak. Then comees the incident of the leaf, when, like Merlin, he bursts out laughing and will only explain his laughter on being promised his liberty. At the same time he gives directions as to his burial, for in a few days he is to die the triple death. On being released, he discloses the queen's adultery. She tries to discredit the prophet by pointing out the impossibility of the triple death. The king, however, will not believe her. On the other hand, the queen, later on, out of revenge, compasses the death of Lailoken, and he is buried, as he had requested, at the junction of the Pausayl with the Tweed. The Lailoken legend, then, is the most important source for the Vita Merlini, but it is doubtful how far this legend itself re- presents oral tradition among the Celts. The motif of the pro- phecy of a triple death is found in the East — so, too, the reve- Merlin 143 lation of adultery by the scornful laughter of a captive gifted with supernatural knowledge.27 The same is true of the motif of the boy whose ignorance of his impending fate, when making provision for the repair of his shoes, causes the prophet to laugh. Parts of the remainder, like Ganieda's prophecy, are, no doubt, mere inventions of Geoffrey,28 whilst the apples that cause mad- ness and the wild story of Merlin's appearing at his wife's marriage riding a stag with whose antlers he slays the bridegroom are, in all probability, both drawn from oral tradition. The latter has every mark of a folk-tale, but there is no reason to suppose that it was originally told either of Merlin or Lailoken, and Geof- frey was, no doubt, the first to connect it with the former. Before leaving the Vita Merlini, it should be remarked that Giraldus Cambrensis distinguishes two Merlins29 — Merlin Am- brosius and Merlin Silvester — and this fictitious distinction has descended to our own time and confused some modern students of the Arthurian romances. But there is no reason whatever for such a distinction;30 the Merlin of the Historia and the Merlin of the Vita Merlini differ in some respects, but, as we have seen, both are substantially the creations of Geoffrey and neither have any root in popular tradition. He himself identifies them by a specific reference in the Vita Merlini (11. 68 Iff.). During the forty years or so that followed upon the com- position of Geoffrey's Latin poem the fame of Merlin was spread far and wide, not only by Geoffrey's own writings, but by deri- 27 Compare the Indian tales cited in another connection by Miss Paton, PMLA, XXII, 241 ff. (1907). 28 Miss Paton, in her article, "Merlin, and Ganieda", MLN for June, 1903, has tried to prove that in the original form of the story Gauieda must have been the enchanter's fairy-mistress, not his sister. But this theory is a pure exercise of fancy, for there is no documen- tary evidence whatever to support it. Cp. this scholar's similar theory in regard to the original relations of Arthur and Morgan le Fay, in her Fairy Mythology, pp. 29, 33. 29 Cp. Itinerarium Kambriae, Book II, ch. 8. 30 This has been made clear by F. Lot, Annates de Bretagne, XV, pp. 333 ff. 144 Evolution of Arthurian Romance vatives from those writings — especially, by extensions of Mer- lin's Prophecies and by paraphrases of the Historia in the verse and prose chronicles. It was not until the appearance of the Merlin of Robert de Boron, however, that the stories that centre about the character received any material development. This poem — which, for the most part, survives only in a prose version — falls probably in the last fifteen years of the twelfth century. As we shall see more fully later on, it constitutes the second member of an intended series of poems and follows, in the main, the ac- count of Merlin given by Geoffrey, whom Robert, however, pro- bably knew only through Wace.31 The most striking thing in the poem — the story of Merlin's conception — is based on Geof- frey, but the original story is happily combined with a motif drawn from popular notions concerning Anti-Christ in the Middle Ages. We find here again some of the motifs of the Vita Merlini — which reached the French poet, however, no doubt, through indirect channels — viz., the triple prediction, the incident of the man about to die who is so careful in regard to his future footwear. Some variations in the account of Arthur's conception, as com- pared with Geoffrey and Wace, may be due to the lost chronicle of a certain Martin of Rochester, which appears to have covered, these events.32 In the account of Arthur's birth and childhood, however, Robert makes a radical departure from his sources in order to increase the importance of his hero, Merlin. In Geoffrey (Book VIII, ch. 20), immediately after Uther Pendragon had be- 31 In his Introduction to the Huth-Merlin, pp. Xff., G. Paris has established, in essentials, the sources of Robert's Merlin. Note, besides, that the incident of the man who so carefully provides himself with foot-gear, although, as a matter of fact, in a few minutes he is destined to die, occurs in Robert (Huth-Merlin, I, p. 48 f.), as well as in the Vita Merlini, 11. 495ff. — so, doubtless, the latter should, also, be reckoned among Robert's sources. Cp. Lot, Annates de Bre- tagne, XV, 336, note 1. 32 According to MS. 749, fol. 32 (Bibl. Nat.), as quoted by P. Paris, RTR, II, 36, note, Robert used a chronicle by this person, and some scholars, as we have seen, — e. g., Brugger, Zs. f. frz. Spr. u. Lit, XXX1, 182ff. (1906) — accept the statement as authentic. With regard to this Martin cp. p. 29, note 59, above. Merlin 145 gotten Arthur upon Igerna, having assumed the form of her hus- band, Gorlois, through Merlin's magic art, Gorlois himself is killed and Uther marries his widow. Thus when Arthur is born, there is no scandal concerning his birth, only Uther and Merlin being in the secret of his conception. But Robert puts the marriage two months after the conception, so, to obviate scandal, Merlin has the boy turned over to him, as soon as it is born, and entrusts it to a good man named Antor (Auctor)33 who does not know the child's rank and whose wife gives it suck. Fifteen years after- wards Uther dies and Britain is without a king. The kingdom is to go to the man who can draw a sword from a certain mar- vellous anvil. Arthur alone proves equal to the task.34 The barons murmur at the success of a person of apparently low rank, but he soon convinces them that he possesses royal qualities, and he is first knighted and then crowned king at the Feast of Pen- tecost. Here Robert's poem ends. Finally, we should note that, in a passage35 which occurs before the account of Arthur's conception, Robert connectc the poem with an unnamed knight who is destined to achieve the ad- ventures of the Grail.36 Merlin relates here the story of the 33 This is, probably, a mere corruption of Arthur. Gawain and Helyas (the Swan-Knight) were, also, named after their foster-fathers. Cp. Bruce, Mort Artu, p. 288. This sword motif is found, also, in the leg-ends of Theseus and Sigmund. Cp. Bruce, op. cit, pp. 297 f. 36 Sommer, II, 53ff. 36 Scholars (e. g., G. Paris, Huth-Merlin, I, p. XVII) have been accustomed to assume that this unnamed knight was Perceval. This, is, however, by no means, certain, although I have no doubt that Robert knew Chretien's Perceval. The passage runs in the Huth MS. (cp. Huth-Merlin, I, 98) as follows: "Tant te puis jou [i. e., Merlin, who is speaking to Uther Pendragon] bien dire qu'il ne sera pas emplis en ton tans. Et cil qui l'emplira naistera de celui qui engendrer [Paris's emendation for MS. emplir] le doit. Et n'a point encore de feme prise ne ne set riens qu'i[l] le doie engenrer. Et couverra que cil qui emplir le doit acomplisse chelui [lieu] avant ou li vaissiaus del graal siet, car cil qui le gardent ne le virent onques acomplir; ne che ne sera jamais en ton tans, ains averra au tans le roi qui apries toi verra." In the MS followed by Sommer (II, 56) the second sentence in 146 Evolution of Arthurian Romance two holy tables of Jesus and Joseph of Arimathea — the table of our Lord's last supper and the table which Joseph instituted in memory of our Lord's table — and he induces Uther to in- stitute a third table at which there is to be a vacant seat as at the former tables. Merlin prophesies that this seat will only be filled in the reign of Uther 's son and by a knight who will put an end to the adventure of the Grail. This idea, like most of Robert's inventions, had a great success with the prose romancers. Robert de Boron's poems have no striking merit — they are far inferior to those of Chretien de Troyes — but in the genre of the prose-romances their influence is of capital importance. He is, above all, responsible for three innovations in Arthurian ro- mance: 37 he gave it both a religious and a pseudo- historical co- loring and he cast his compositions in a cyclic form. His Merlin, with which we are now more especially concerned, was turned into prose, and this prose rendering, as we have seen, forms the first section of the various prose Merlins which were composed in the first half of the thirteenth century. These romances, accor- dingly, begin with Robert's story of Merlin's conception, whose elements we have indicated above. The devils hold council as to how they shall counteract the work of Christ. They agree that the passage just quoted is represented by: "& cil qui l'acomplira nest encore mie engendres", whilst there is nothing at all corresponding to the third sentence. If the reading of the Huth MS. is correct, the allusion certainly would not fit with anything that we learn about Perceval's father in Chretien's poem. On the other hand, it seems to correspond rather with the allusion to the future heir of Alein (Hebron's son) in Robert's Joseph, 1. 3467: "Et ques oirs de li peut issir", where there is no hint that Perceval is intended. In any event, even if the unnamed knight in the passage under consideration is to be accepted as Perceval, to judge by Robert's extant works, he would have been a very different character from Chretien's and his story would have differed accordingly. On this subject, see, still further, Part IV, below. 37 This has been pointed out by Brugger, Zs. f. frz. Spr. u. Litt, XXIX1, 75 ff. (1905). Merlin 147 this can only be done by a man who is born of a virgin and one of the devils undertakes to engender such a person. The narrative of how this enterprise is executed is weird in the extreme and has a real tragic force, but, fortunately, the innocence of the victim defeats the object of the fiends and the child who is thus begotten inherits his father's supernatural powers, but not his wickedness. We have already seen how this being was destined to watch over Arthur's infancy and childhood up to the coronation of the great king with which Robert de Boron's poem ends. In the several continuations of Robert's Merlin, the enchanter appears again from time to time as a deus ex machina, but, after all, the Merlin ma- terial in these huge continuations — the Vulgate continuation, for example, is more than four times the length of the de Boron section — constitutes only a small portion of the whole romance. Take, for instance, the continuation, just named, which is the oldest of the three extant continuations. It is really, as we shall see, a pseudo-history of Arthur's reign down to the appearance of Lancelot on the scene — Arthur's wars with his rebellious barons and with the Saxons, his marriage with Guinevere, and what not. Most of this narrative is a patch- work of motifs deri- ved from Wace or from the lost expanded version of that writer's Brut or again from romances like the Meraugis de Portlesguez. Only occasionally do we have a bit of genuine Celtic tradition as in Arthur's fight with the Capalu, a monster-cat.38 As far as Merlin is concerned, there is no reason to believe that any of the new incidents connected with his name in the Vulgate con- tinuation (or Livre d'Artus, as it is also called) are of Celtic origin. The most distinctive of these are: 1. the Grisandole episode. 2. The episode of Merlin and Viviane, which ends in V ' enserrement Merlin, as the Old French romance calls it: he succumbs to Vi- 38 See Sommer, II, 442 ff. Freymond (cp. p. 41, note 9, above), in a masterly monograph, Artus' Kampf mit dem Katzenungestilm (in the Festgabe fur G. Grober, Halle, 1899), has shown that the Capalu is the Cath Panic of Welsh saga, although in Welsh the creature is not at all connected with Arthur. 148 Evolution of Arthurian Romance viane's spell and is imprisoned forever in the forest of Broce- liande.39 First, as to the Grisandole episode,40 it is briefly as follows: Julius Caesar, emperor of Rome, is troubled by an incomprehensible dream. Merlin knows of it, assumes the form of a stag, comes to Caesar's palace, followed by a multitude of people, and tells him that only the wild man of the woods can interpret his dream. The enchanter then opens the palace gates by magic and vanishes. The emperor promises his daughter's hand as a reward for the capture of the wild man or the stag. Many knights go on the quest, but all later abandon it, save Grisandole, who is really a princess disguised as a man. The stag appears to Grisandole and tells how the wild man may be caught. She is to spread food for him in the forest as a snare. She does this, and when the creature, after gorging himself, lies down to sleep, she and her men capture him and take him back to court. We have now the motif of the sudden bursts of laughter, like those of Merlin in the Vita Merlini. There is no need of relating the three incidents — all significant, however, of the wild man's powers of divination. In the end he explains his laughter in each case, as Merlin does in the Latin poem. The main thing is that, on reaching the palace, he inter- prets Caesar's dream as signifying that the twelve ladies-in-honor to the queen are really twelve youths in disguise with whom she is leading an adulterous life. He further explains that he had laughed on looking at Grisandole, because a woman by her craft had taken him prisoner when no man could capture him. Gri- sandoie's sex is thus revealed. As a result of it all, the queen and her paramours are burned, Grisandole marries the emperor and her brother the emperor's daughter. The wild man refuses to reveal who the great stag is or his own name and leaves the hall abruptly, writing an inscription in Hebrew on one of the door-posts, as he passes out. Some time afterwards a messenger from Greece interprets the inscription, which explains that the wild man and the stag are one and the 39 For this latter episode see Sommer, H, 208 h\, 280, 421 40 Cp. Sommer, II, 282 ff. Merlin 149 same being, named Merlin, Arthur's counsellor. Instantly there- after the letters vanish. The main motifs of the above episode are found combined, already in a number of stories current both in Europe and the East — that is to say, the story of the disguised girl who captures the wild man or satyr, as he is sometimes called, this wild man being endowed with the gift of divination by which he reveals his cap- tor's sex and a queen's infidelity to her husband. We even have in these tales the sudden outbursts of laughter.41 The stories of Solomon and Aschmedai (Asmodeus), as we have seen, had already furnished motifs to the Vita Merlini, and in the above episode, no doubt, we have these same stories drawn on; for in the Hebrew legend of Solomon Asmodeus plays exactly the same role as the wild man does here. As regards Merlin's assumption of the form of a stag, this seems a sort of confused recollection of the passage in the Vita Merlini, where Merlin rides to the palace on a stag. This preface of the Grisandole episode was evidently introduced merely to connect with Merlin the piquant tale of the disguised maiden, the faithless queen, and the mar- vellous wild man of the woods; but the connection is, after all, of the most awkward kind. The Grisandole episode, then, is not Celtic in origin. The same thing is true of the story of Merlin and Viviane, immortalized 41 In her paper on this episode, "The Story of Grisandole", PMLA, XXII, 234ff. (1907), Miss L. A. Paton tries to show that we have here again the hard- worked fairy-mistress theme (see p. 256) in disguise (Grisandole was originally a fairy and Merlin her lover) and that the story is drawn from oral Celtic tradition. For my own part, as I have already intimated, I am skeptical on principle as to virtually every story of any importance in mediaeval romance being a variant of the fairy-mistress theme. But, besides, as Miss Paton herself acknow- ledges, some features of the tale — to my mind the most essential — are indisputably of Oriental origin — for example, the capture of the laughing wild man who is compelled to exercise his powers of divi- nation. Why not take this wide-spread story at its face value, instead of reading into it the worn out fairy mistress motif? This explains everything in the episode, except Merlin's part in it, and that is plainly a clumsy addition, as anyone might have guessed, even if a comparison with the cognate tales did not prove it. 150 Evolution of Arthurian Romance by Tennyson, and here the case is even plainer than in the epi- sode just examined. It was pointed out by Gaston Paris,42 that the essential conception of this lamous story — the only really pretty one in the whole Merlin continuation,43 is first found in the prose Lancelots There Merlin betrays the magic spell to Viviane or Niniane (the MSS. vary as to the form of the name),45 on her promising him her love, but, after obtaining the secret, she uses her magic power to make him believe that he has enjoyed her favours, though, in reality, he has not. At last she shuts him up forever in a cave in the forest of Darnantes. Vi- viane in this passage is identified with Lancelot's foster-mother, the Lady of the Lake. This story of Merlin and Viviane is, probably, an early interpolation in the Lancelot*6 but, in any event, it is certainly the original of the more fully developed episode in the prose Merlin. Furthermore, it is unquestionable that, contrary to former views on the subject, the conception here involved was first attached to Merlin by the author of the above- mentioned passage in the prose Lancelot and has nothing Celtic about it; for, obviously, the Merlin-Viviane incident is merely a new adaptation of the old fabliau motif of the wise man deceived by a woman, which is ultimately of Oriental origin.47 Aristotle,- " Cp. Huth-Merlin, I, p. XLVI. 43 Cp. Brugger, Zs. f. frz. Spr. u. LitL, XXXIII1, 149 (1908). 44 Sonnner, III, 19ff. 46 The MSS. show still other variants. Cp. Sommer's Index to his Vulgate Version under Viviane. This is, most probably, the correct form of the name. It would be, then, the feminine, corres- ponding to the masculine Vivian (Vivien), so well-known from the chansons de geste. Names, however, are subject to such corruption in mediaeval MSS. that, after all, Diane (Diana) may be the true etymon of Viviane. In her Fairy Mythology, ch. 4, Miss Paton treats at length Viviane's character, as well as her name, and con- cludes that both are Celtic. The episodes in the Huth-Merlin, etc., however, which she discusses, in reality, are all of purely literary manufacture. We shall see in our chapters on the prose romances that this was true, in general, of these works. 46 In the Zs.f.frz. Spr. u. LitL, XXX1, 175 ff. (1906), Brugger seems to me to have proved this. 47 A. Jeanroy was the first to point this out in his review of Miss Paton's Fairy Mythology. Cp. Romania, XXXIV, 120 (1905). Merlin 151 Hippocrates and Virgil, as we have seen above, figure successively in these tales as the butt of feminine deceit. Merlin's turn was sure to come. That this is the true derivation of the Merlin-Vi- viane story is so evident that no further argument on the subject is required. Even the detail noted in the Lancelot passage of the magical illusion under which Merlin labors as to his enjoyment of his lady's favors is found in the Solomon and Marcolf legend and had already been employed by Chretien in Cliges, as was ob- served above. It is a noteworthy mediaeval feature of the Vulgate Merlin that Viviane was only twelve years old when Merlin fell in love with her. The author of the Merlin of. the so-called Robert de Boron prose cycle, best represented in the somewhat shortened form of the Huth-Merlin, took over this episode from the Vulgate romance and from his work48) it passed into Malory49 and thence into Tennyson's Idylls of the Kifig.50 This episode and a gene- ral reputation for magic and prophetic powers are all that can be still called vital in the stories concerning Merlin which are found in the books that we have been passing in review. Brugger had, independently, arrived at the same interpretation. Cp. his elaborate argument mZs.f.frz. Spr. u. Lift., XXX1, 1 1 7 ff . (1906). 43 II, 191 ff., in the edition of G. Paris and J. Ulrich. 49 Book IV, ch. 1. 80 Merlin and Vivien. Chapter V. Tristan. In one important respect the study of the story of Tristan is easier than is the case with that of Lancelot: there is sub- stantial agreement among authorities on the subject that all the mediaeval romances and shorter poems concerning this hero go back to a lost French romance1 of a considerably earlier date 1 Owing to its archaic character and, the (in many respects) rude civilization which it depicted Bedier, II, 314, dates this lost romance back "jusq' aux premiers temps de la conquete de l'Angleterre par les Normands", which must mean, at the latest, the early years of the twelfth century. Golther, p. 73, however, objects that the in- troduction of Arthur and his knights proves that it must have been written after Geoffrey's Historia, which dates from about 1136, had made these characters familiar figures — more specifically, between 1140 and 1150. Indeed, Golther, p. 34, was inclined to believe that the adultery of a nephew with his uncle's wife in this pri- mitive Tristan was imitated from the similar relations of Mordred, Arthur, and Guinevere in Geoffrey's work, as Muret, Romania, XVI, 322 (1887) had already suggested. Miss Schoepperle, p. 183, dates the lost romance "very shortly" before the extant redactions, none of which, according to her^ antedate the last decades oi the twelfth century. Bedier, II, 154f., accepts 1154 as the date of Bernart de Ven- tadour's lyric which contains the earliest allusion to Tristan and Iseult, and hence as the terminus ad quern of the primitive Tristan, but Miss Schoepperle, pp. 112ff., has shown that the true date ot com- position of this lyric is wholly .uncertain, and, consequently, that it cannot be used for dating the lost archetype of our Tristan poems. Her own late dating of this archetype (pp. 120ff.) is based on the observation that it contained motifs -that did not become current until the latter half of the twelfth century: 1. A girl who eludes an im- portunate lover (Kaherdin-Camille episode), imitated, according to Miss Schoepperle, from the pastourelles ; 2. Two lovers deceive a jealous husband, despite all his precautions (Kaherdin-Gargeolain), imitated from Tristan 153 than any on the subject that is now in existence. The close rela- tion of the incidents in the various extant versions of the story, despite individual divergencies, was explained by scholars of the last generation as due to the fact that the writers all drew from the same body of lays or prose tales which were supposed to be current orally,2 but the conviction gradually forced itself on the minds the chansons de mat mariee. 3. Notions that are rooted in the amour courtois (cp., already, Muret, Romania, XVI, 360), such as the extra- vagant tests of humiliation to which a lady puts her lover and the idea that a lover must do anything, if appealed to in the name of his lady-love. 4. The stereotyped characters of Arthur, Kay and Gawain, which show the influence of the Arthurian romances. In Eilhart 11. 5047 ff. especially, I may add, show unmistakably this influence. They describe as accurately as possible the plan of the regular Arthurian romances. F. Lot's criticism (Romania, XLIII, 128 f.) of these points has invalidated, I think, so late a dating of the primitive Tristan as Miss Schoepperle's (which, besides, would be impossible, if Thomas' poem were written about 1170), but it seems to me that the evidence, espe- cially under the above heading, is sufficient to prove that this lost archetype was not earlier than the romances of antiquity (1155 — 1165). G. Huet, Moyen Age, 2e. serie, XVIII, 380ff. (1914), has added Iseult's love-monologue (Eilhart, 2398 ff.) to Miss Schoepperle's illu- strations of the amour courtois. The matter is not susceptible of determination, but it seems most likely that ideas of the amour cour- tois were in the archetype, since they are found in all the extant versions. The same thing applies to the Arthurian connections of these versions. To be sure, in the case of Thomas, this connection is very slender, consisting entirely of Arthur's encounters with two giants (Bedier, I, 290 ff., 307), which really stand outside of the Tristan adventures and may very well have been borrowed directly from Ge- offrey's Historia, Book X, ch. 3 or Wace's Brut, 11. 11634ff. In Beroul, 11, 3706 ff. et passim, and Eilhart, 11. 5231 ff., Arthurian characters are more intimately connected with the narrative, and Lot. op. cit., pp. 131 f., argues that the episode in Eilhart was expressly introduced by the author for the purpose of imparting novelty to the story. This, of course, is possible. In his "Tristan bei Cercamon ?", Zs. f. rom. Ph., XLI, 219ff. (1921), C. Appel, detects an allusion to Tristan in a poem of this Provencal poet. He assigns the poem to the period, 1150 — 1160, and contends that this is the earliest extant allusion to Tristan. The evidence, however, is too uncertain. 2 So R. Heinzel, Zs. f. d. A. XIV, 272ff. (1869), W. Golther, Die Sage von Tristan mid Isolde, pp. 30 ff. et passim (Munich 1887). 154 Evolution of Arthurian Romance of students of the Tristan romances that the resemblances in question were really due to a common original — a definite ro- mance, now lost.3 This solution of the problem was given scientific F. Novati, Studj di filologia romanza, II, 390 (1887), G. Paris, Manuel, pp. 99 ff. (1888). For historical surveys of Tristan studies, cp. Bedier, II, 168ff. and Golther, Iff. 3) In his review of Rbttiger's Program in Romania, XXVII, 608ff. (1898), E. Muret makes Chretien's lost poem and Thomas the sources (one or the other) of all extant Tristan versions. E. Brugger, Zs. f frz. Spr. u. Litt. XX, 134, note (1898) suggested a single source (undefined). Golther, however, ibid. XXII1, 23 (1900), was the first to work out a definite scheme, based substantially on the hypothesis of a single source (Chretien). In his later work Golther gives up Chretien as the author of the Ur-Tristan. In the Journal des Savants for 1902, p. 301; note 2, G. Paris expressed the belief that all French Tristan poems go back to a lost English poem, itself incomplete. In that case, the story of Tristan and Iseult would have passed from the Cornish (or Welsh) to the English and from the English to the French. For a discussion of this hypothesis and the reasons which led Paris to adopt it see Bedier II, 314ff. In so far as these reasons are connected with Paris's general theory of the Anglo-Norman origin of the Arthurian romances, I have dealt with them elsewhere. For the rest, I will add to what Bedier says on the subject that the testimony of the unpublished French poem Waldef (eleventh or twelfth century) as to the existence of an early English Tristan (Tristram) is, on the face of it, valueless, since in the same line the author of this poem speaks of an early English original for the French Bruit (Brut). The line of works entitled Brut, however, all go back to Geoffrey of Monmouth (cf. Gaiinar. Wace etc.) and have, of course, nothing to do with any earlier English works. Cp. on this subject Golther, Zs. f. frz. Spr. u. Litt. XXIX 2, 151 ff. (1906) and Brugger, ibid. XXXII 2, 136ff. (1907), and, above all, R. Immelmann in his edition of the fifteenth century Latin Waldef romance, Johannes Bramis' Historia Regis Waldei, pp. XXX ff. (Bonn, 1912). The argument from the Waldef which Paris had al- ready cited, Romania XV, 597 (1886), XVIII, 510 (1889), in con- nection with the discussion of Tristan's origin, was adopted, also, by W. Hertz, Tristan und Isolde von Gottfried von Strassburg, pp. 477 f. — so, too, apparently by W. H. Schofield, English Literature from the Norman Conquest to Chaucer, p. 202 (New York, 1906). Schofield here goes so far as to connect the incident of Tristan's sending his message to Iseult by the chips on the stream with the Anglo-Saxon lyric which is usually entitled "The Husband's Message/' Tristan 155 demonstration by Bedier in the Second Volume (1905) of his great edition of Thomas' Tristan* and substantially the same results were reached independently by W. Golther in his Tristan und Isolde (1907), which, though published two years later, had been written before the appearance of Bedier 's work. The con- clusions of Bedier and Golther have been disputed by a few scho- lars, as will be observed from the notes below, but, in general, we may say that the existence of a single primitive Tristan ro- mance (Ur-Tristan, as German scholars call it) from which all extant versions are ultimately derived is one of the few matters of Arthurian discussion on which students are definitely agreed.5 4) Also in a popular article in the now defunct International Quarterly (March-June, 1904). 5 It would seem natural to identify this Ur-Trlstan with the lost poem "del roi Marc et d' Iseut la blonde" which Chretien, in the list of his works, Cliges, 11. Iff., tells us that he wrote, and, as a matter of fact, Foerster, Cliges*, p. LXVIII, argues that the two are identical. But its connection in this list with Chretien's tales from Ovid (cer- tainly compositions of the poet's youth) shows that the lost poem was an early work. Now, the Ur-Tristan was evidently a masterpiece and superior in construction and in poetical content to even the ma- turest romances from Chretien's pen, so that it is inconceivable that it was one of Chretien's compositions. Golther, p. 74, very properly raises this objection to the theory of the identity and adds with less force that Chretien is not likely to have had the knowledge of Eng- lish conditions that the Ur-Tristan (as reconstructed by Bedier and himself, with a high degree of probability) implies. Foerster, loc. cit., tries to meet Golther's objection, with the supposition that Chretien had before him a still earlier Tristan poem (Ururtristan) and that the merits of construction may have come to him from his original. In the Journal des Savants for 1902, pp. 299 ff., G.Paris has argued that Chretien did not write a full poem on the story of Tristan and Iseult, but merely a brief one, dealing with some episode in which Marc and Iseult figured. Novati, Studj di filologia romanza, II, 411 and Rottiger, Der heutige Stand der Tristanforschung, pp. 28 f. (1897) had already discerned a significance in the absence of Tristan's name from Chretien's allusion to his lost poem. Paris's hypothesis is based on the consideration that in all the literature of the Middle Ages there is no allusion to a Tristan poem by Chretien. We have, however, no episodic poems from Chretien's pen, and it seems to me more likely that he planned a long poem on Tristan and Iseult, but 156 Evolution of Arthurian Romance Bedier has reconstructed the narrative of this primitive Tristan by the comparative method, and so has Golther. Their respective failed to complete it, as he later failed to complete his Lancelot (to say nothing of the Perceval) — only in the case of this earlier com- position he never put his work into circulation, recognizing it - as im- mature. — In Paul's Grundriss der germanischen Philologie II, 1, 459 (1890) J. te Winkel conjectures that there was a (lost) Dutch translation of Chretien's Tristan, hut gives no reasons for this con- jecture. In the Roman de Renard, Branch II, there is an allusion to a lost Tristan poem by a certain La Chievre and in a conte devot (Foerster Festschrift, Halle, 1902) there is a similar allusion to this lost poem — only its author's name is here given in the Picard form, Li Kievres. The poem, of whose contents we know nothing, was probably composed in the twelfth century and its author should, doubt- less, be identified with the lyric poet, Robert La Chievre of Rheims. On these subjects cp. Grbber's Grundriss, Band II, AbteilungI, pp. 494, 671, and G.Paris, op. cit. p. 299. In 1. 2119 of his Tristan, Thomas appeals to "Breri" as his authority for representing that Tristan sent Kaherdin, and not Governal, as his messenger to Iseult on a certain occasion. G. Paris, Romania, VIII, 425 ff. (1879) identified this "Breri" with "famosus ille fabulator Bledhericus, qui tempora nostra paulo praevenit", of Giraldus Cam- brensis, Descriptio Kambriae, ch. XVII, in Vol. VI (London, 1868) of that writer's Works edited for the Rolls Series by J. F. Dimock. As Paris, himself, however, says, Thomas does not hereby imply that he is using a book by Breri, but is merely appealing to the authority of a person of that name who is said to know more about British history than anybody else. "Breri" and "Bledhericus" are, indeed, probably the same name and Paris's identification of the persons con- cerned is a priori admissible. The identification was favored, also, by H. Zimmer, Gottingische Gelehrte Anzeigen for Oct. 1, 1890, p. 805, note, and Zs. f. frz. Spr. u. LitL, XIII1, 84, (1891). Never- theless, Thomas's appeal to Breri is, no doubt, merely one of the in- numerable instances in mediaeval literature of a writer's bolstering up his narrative by the citation of fictitious authorities. He puts off on Breri the responsibility for innovations in the story which are really, his own. Cp. Muret, Romania, XVII, 608 f. For a full discussion of the Breri question cp. Bedier, II, 95 ff. He lists p. 95, note 1, the chief previous discussions of the subject. Add J. L. Weston, Romania, XXXIH; 334 ff. (1904), who identifies Thomas's Breri with the Bleheris (Blihos-Bleheris) of the additions made to Chre- tien's Perceval — also, W. Golther, Tristan und Isolde, pp. 139 f. Tristan 157 reconstructions do not differ in essentials, so that the task of the Tristan student at the present time consists mainly in the study of the sources of this hypothetical romance, as reconstructed by these scholars. The materials which form the basis of these reconstructions are. 1. The Tristan-poem of the Anglo-Norman poet, Thomas, composed somewhere between 1155 and 1170.6 2. The fragments (Leipzig, 1907), who points out that Thomas, in the passage referred to above, cites Breri as a great authority on British history, yet, as a matter of fact, derives all his knowledge of that history from Wace. Very important is L. Foulet, "Thomas and Marie in their relation to the conteurs," MLN, XXIII, 205ff. (1908). Foulet shows that Thomas's reference to the "conteurs", to whose varying accounts he opposes the authority of Breri, is really a "meaningless mannerism", copied from Marie de France, and does not imply any knowledge of oral traditions concerning Tristan. We shall have to return to this mysterious Breri in the discussion of the continuations to Chretien's Perceval. 6 The upward limit is fixed by the author's use of Wace's Brut, the lower by his influence on Chretien's Cliges. F. Lot, Romania, XXVII; 42 (1898), called attention to the first point, G. Paris, Jour- nal des Savants for July, 1902, pp. 354ff. to the latter. For a full discussion of the subject cp. Bedier, I, 37 ff. M. Wilmotte, U evolution du roman francais, p. 67 (Brussels, 1903) and W. Foerster, Cliges*, pp. LXVIff. (Halle, 1910; have disputed Paris's proofs as to the priority of Thomas's Tristan over the Cliges, on the ground that Chretien and Thomas may have been drawing from a common source (one of the lost Tristan poems), in the cases where Paris assumes imitation of the latter by the former. The question is difficult to decide, but it seems to me that Thomas would hardly have appropriated from a predecessor so distinctive a play on words as amer: mer. The limits of date would be still further narrowed, if we could accept S. Singer's contention, "Thomas, Tristan, und Benoit de Saint Maure," Zs. f. rom. Phil. XXXIII, 729ff. (1909), that the Roman de Troie influenced Thomas's Tristan; for the former was written about 1165. Singer compares especially the description of the loves of Rivalen and Blanchefleur, Bedier, I, 12ff., with those of Achilles and Polyxena, R. de Troie, 17554ff. But the matter is too inde- finite, and the passage in Thomas, just named, may be imitated from some other romance embodying the new spirit of the amour courtois. Singer, himself, compares, also, Gottfried, 11, 16478ff. (which goes 158 Evolution of Arthurian Romance of a French poem which is, at least in part, by a poet of Normandy named Beroul,7 coupled with the Middle High German poem back, doubtless, to Thomas) with Eneas, 9885 ff. (eye belongs to love, hand to grief). Traces of the influence of the Disciplina Clericalis in Thomas throw no light on the subject of date, since the author of that work, Petrus Alfonsi, flourished in the early part of the twelfth century. Cp., on this influence, A. Hilka, "Der Tristanroman des Thomas und die Disciplina Clericalis," Zs.f.frz. Spr. u. LitL, XLV1, 38 ff. (1917). That Thomas was not identical with the poet of the same name who wrote Horn et Rimenhild has been shown by W. Soderhjelm, "Sur l'identite du Thomas, auteur de Tristan, et du Thomas, auteur de Horn", Romania, XV, 175 ff. (1886). The identification of our Anglo-Norman Thomas with Thomas of Erceldoune is, of course, merely the individual fancy of the author of the English Sir Tristrem. Novati, Studj di filologia rornanza, II, 403, note 3, conjectures that Thomas, author of Tristan, was an ecclesiastic. There is no means of deciding the matter, but the conjecture hardly seems probable. 7 The poem has been preserved in the unique MS. 2171 of the Bibliotheque Nationale and was edited by H. von der Hagen in Vol. II of his edition of Gottfried von Strassburg (Breslau, 1823) and by F. Michel in his edition of the Tristan fragments I, Iff. (London and Paris, 1835). The authoritative editions, however, are E. Muret's 1. Le roman de Tristan par Beroul et un anonyme (Paris, 1903, for the Societe des Anciens Textes Frangais) and 2. Beroul, le roman de Tristan, poeme du XII6 si&cle (Paris, 1913, in Les Classiques fiancais du moyen age). LI. 1268 and 1790, the poet calls himself Berox (nominative form). Down to 1. 2754 the narrative accords closely with that of Eilhart. Not so with 11. 3028—4485 (end), nor with 11. 2767 — 3031, which latter connects the two principal divisions of the poem. In his first edition of Beroul (Paris, 1903) Muret, pp. LXVff., concluded with G. Paris and others that Beroul II (11. 3028 — 4485) was by a different hand from Beroul I (11. 1 — 2754) and that the two were connected by still a third hand. Beroul II, he observes, is grosser and more barbarous, is not marked by the same literary know- ledge or influences of chivalrous courtesy. In his second edition (Paris, 1913) he repeats this opinion, though with some hesitation (pp. VIII ff.), being affected, it would seem, by Bedier's view (Legendes Epiques, III, 399), that the whole of the Beroul fragment is by one person. Heinzel (Zs. f. d. Altertum, XIV) had maintained, as no one would now, that it consisted of nineteen different lays by different authors. Bedier's view seems to me, the most likely. — The Beroul Tristan was undoubtedly addressed to an audience of lower social position Tristan 159 on Tristan by Eilhart von Oberge.8 Both belong to the closing years of the twelfth century and both draw evidently from the same source — a lost derivative of the primitive Tristan? 3. Por- than Thomas's poem. It is a jongleur's version. There is nothing, however, to support the view, formerly held, that it was older than Thomas. The upward limit of date for Beroul II, is fixed by an allusion (1. 3853) to the epidemic of leprosy which raged among the crusaders at Acre, 1190 — 1. This part of the poem probably falls in the last decade of the twelfth century. Muret (p. LXIV of his first edition) assumes that Beroul I was not earlier than 1165 or 1170. There is no evidence, however, to prove that this part of the poem was not composed substantially at the same time as Beroul II, granting even that the two parts were by different authors. 8 Edited * by Franz Lichtenstein (Strassburg,