In examining the relationship of the Didot-Perceval to the Conte du Graal, the sequence of the incidents is of importance. This is shown in the subjoined table (where the numbers given are those of the incidents as summarized, Chapter II), in which the Didot-Perceval sequence is taken as the standard.

Didot-Perceval.		Chrestien.		Gautier de Doulens.
Inc.		Inc.		Inc.
2.	Perceval sets forth in quest of the rich fisher.	11.	Only after the reproaches of the loathly damsel does Perceval first set forth in quest of the Grail.	.........
3.	Finds a damsel weeping over a knight. Adventure with dwarf and the Orgellos Delande.	8.	In so far as finding a damsel weeping over a dead knight, and (9) for overcoming the Orgellous de la Lande.	9.	In so far as a damsel is found lamenting over a knight.
4.	Arrival at the Chessboard Castle. Adventure of the stag hunt and loss of the hound.		.........	7 and 8.
5.	Meeting with sister; instruction concerning the Grail; vow to seek it.		.........	12.
6.	Meeting with, confession to, and exhortation from hermit uncle.	15.	After the Good Friday incident.	12.
7.	Disregard of uncle’s exhortations (slaying a knight), through thinking of damsel of the Chessboard.		.........	12.	In so far as a knight is slain, but before the meeting with the hermit.
8.	Meeting with Rosette and Le Beau Mauvais (the loathly damsel).		.........	11.
9.	Adventure at the Ford with Urbains.		.........	9.	Ford Amorous; entirely different adventure.
10.	The two children in the tree.		.........	20.	One child.
11.	First arrival at Grail Castle.	7.	.........	.........
12.	Reproaches of the wayside damsel.	8.	In so far as in both the hero is reproached by a wayside damsel.	.........
13.	Meeting with the damsel who had carried off the stag’s head and hound, and second visit to Castle of the Chessboard.		.........	13 and 18. Many adventures being intercalated.
14.	Period (7 years) of despair ended by the Good Friday incident.	15.	.........	.........
15.	Tournament at Melianz de Lis. Merlin’s reproaches.	13.	But told of Gawain not of Perceval.	.........
16.	Second arrival at Grail Castle. Achievement of Quest.		.........	22.

The different sequence in the Didot-Perceval and Chrestien may be explained, as Birch-Hirschfeld explains it, by the freedom which Chrestien allowed himself in re-casting the work; but why should Gautier, who, ex hypothesi, simply took up from Chrestien’s model such adventures as his predecessor had omitted, have acted in precisely the same way? If the theory were correct we should expect to find the non-Chrestien incidents of the Didot-Perceval brought together in at least fairly the same order in Gautier. A glance at the table shows that this is not the case. In one incident, moreover, the Didot-Perceval is obviously right and Gautier obviously wrong, namely, in his incident 12, where the slaying of the knight before the hero’s meeting the hermit takes away all point from the incident. An absolutely decisive proof that that portion of the Conte du Graal which goes under Gautier’s name (though it is by no means clear that all of it is of the same age or due to one man), cannot be based upon the Didot-Perceval as we now possess it, is afforded by the adventure of the Ford Amorous or Perillous, which in the two versions is quite dissimilar. This incident stands out pre-eminent in the Didot-Perceval for its wild and fantastic character. It is a genuine Celtic mÄrchen, with much of the weird charm still clinging to it that is the birthright of the Celtic folk-tale. It is inadmissible that Gautier could have substituted for this fine incident the commonplace one which he gives.

If, then, it is out of the question that Gautier borrowed directly from the Didot-Perceval, how are the strong resemblances which exist in part between the two versions to be accounted for? Some of these resemblances have already been quoted (supra, p. 75), the remainder may be usefully brought together here.[70]

First arrival at the Castle of the Chessboard—

Didot-Perceval.		Gautier.
Li plus biaux chasteaux del monde et vit le pont abeissiÉ et la porte deffermÉ (p. 439).		Le bel castiel que je vos dis ..... Et vit si biÈles les entrÉes Et les grans portes desfremÉes (22,395, etc.);

The damsel exhorts him not to throw the chessman into the water—

Votre cors est esmeuz À grant vilainie faire (p. 440).

Car Çou serait grans vilonie (22,503).

Perceval having slain the stag, sees its head carried off—

Si vint une veille sor un palestoi grant alÉure et prist le brachet et s’en ala or tot (p. 442).

Une puciÈle de malaire Vint cevauÇant parmi la lande Voit le braket, plus ne demande Par le coler d’orfrois le prist ...... Si s’en aloit grant alÉure (22,604, etc.).

On Perceval threatening to take it away from her by force she answers—

Sire Chevalier, force n’est mie droit et force me poez bien faire (p. 443).

Force À faire n’est mie drois Et force me poÉs vos faire (22,640).

In the subsequent fight with the Knight of the Tomb, he, overcome—

Se torna vers le tonbel grant alÉure et Que fuiant vait grant alÉure li tombeaux s’enleva contre moult et chevalier s’en feri enz (p. 444).

Vers l’arket et la sepouture Si est entrÉs plus tost qu’il pot (22,723, etc.).

In the description of Rosette (the loathly damsel)—

Ele avoit le col et les mains plus noires et le vier, que fer... (p. 453).

Le col avoit plus noir que fer (25,409).

When the loathly damsel and her knight come to Arthur’s court, Kay jests as follows:—

Lors pria (i.e., Kay) le chevalier par la foi que il devoit, le roi, qui li dÉist oÙ il l’avoit prise et si en porroit une autre tele avoir, si il l’aloit querre (p. 457).

Biaus sire, Dites moi, si Dex le vos mire, Si plus en a en vostre terre, Une autÈle en iroie querre Si jou le quidoie trover (25,691 etc.).

These similarities are too great to be accidental. It will be noticed, however, that they bear chiefly upon two adventures: that of the chessboard and stag hunt, and that of the loathly maiden. As to the latter, it is only necessary to allude to Birch-Hirschfeld’s idea that Rosette is the original of the damsel who reproaches Perceval before the court with his conduct at the Grail Castle, a theory to state which is to refute it. The former adventure will be closely examined in the following section. There is no need to suppose direct borrowing on the part of one or the other versions to account for the parallel in these two incidents; a common original closely followed at times by both would meet the requirements of the case. It is difficult to admit that the author of the Didot-Perceval used Gautier’s continuation and not Chrestien’s original, especially when the following fact, strangely overlooked by both Birch-Hirschfeld and Hucher, is taken into account: Perceval on his first arrival at the Grail Castle keeps silence (as will be seen by a reference to the summary, supra, p. 31), because, “li souvenoit du prodome qui li avoit deffandu que ne fust trop pallier,” etc. As a matter of fact, the “prodome” had forbidden nothing of the sort, and this casual sentence is the first allusion to the motive upon which Chrestien lays so much stress as explaining his hero’s mysterious conduct at the Grail Castle. Evidently the Didot-Perceval, which, to whoever considers it impartially, is an obvious abridgment and piecing together of material from different sources, found in one of its sources an episode corresponding to that of Gonemans in Chrestien. But its author, influenced probably by the Galahad version of the Quest, substituted for the “childhood” opening of this hypothetical source the one now found in his version, and the Gonemans episode went with the remainder of that part of the story. When the hero comes to the Grail Castle, the author is puzzled; his hero knows beforehand what he has to do, sets out with the distinct purpose of doing it, and yet remains silent. To account for this silence the author uses the motive belonging to a discarded episode, but applies the words to his hermit, forgetting that he had put no such words into his mouth, and that, attributed to him, the injunction to keep silence became simply meaningless. Is the model treated in this way by the Didot-Perceval Chrestien’s poem? Hardly, for this reason. After the Good Friday incident occurs the remarkable passage, quoted (supra, p. 31), as to the silence of the trouvÈres respecting it. Chrestien gives the incident in full, and the author of the Perceval could have had no reason for his stricture, or could not have ventured it had he been using Chrestien’s work. Two hypotheses then remain; the unknown source may have been a version akin to that used by Chrestien and Gautier, or it may have been a summary abridgment of the Conte du Graal, in which, inter alia, the Good Friday incident was left out. In either case the presence of the passage in the Perceval is equally hard of explanation; but the first hypothesis is favoured by the primitive character of the incident of the Ford Perillous, and several other features which will be touched upon in their place. The Didot-Perceval would thus be an attempt to provide an ending for Borron’s poem by adapting to its central donnÉe a version of the Perceval sage akin to that which forms the groundwork of the Conte du Graal, its author being largely influenced by the Galahad form of the Quest as found in the Queste. If this view be correct, the testimony of Perceval (wherever not influenced by Borron’s poem or the Queste) is of value in determining the original form of the story, the more so from the author’s evident want of skill in piecing together his materials. It will, therefore, be used in the following section, which deals with the relationship of the Conte du Graal and the Mabinogi of Peredur ab Evrawc.

Relationship of the Conte du Graal and the Mabinogi.—As was seen in Chapter IV, opinion began with Monsieur de VillemarquÉ by accepting the Mabinogi as the direct source of the Conte du Graal, and has ended with Zarncke and Birch-Hirschfeld in looking upon it as a more or less direct copy. The most competent of living scholars in this matter, M. Gaston Paris, has expressed himself in favour of this opinion in his recent article on the Lancelot story (Romania, 1886).[71] Before dealing with the question as presented in this form, Simrock’s view, differing as it does from that of all other investigators, deserves notice. He, too, looks upon the Mabinogi as derived from Chrestien, and yet bases his interpretation of the myth underlying the romance upon a feature, the bleeding head in the dish, found only in it. But if the Mabinogi have really preserved here the genuine form of the myth, it must represent an older version than Chrestien’s, and if, on the other hand, Chrestien be its only source, the feature in question cannot belong to the earliest form of the story. Simrock’s theory stands then or falls in this respect by the view taken of the relationship between the two versions, and need not be discussed until that view has been stated.

To facilitate comparison, the incidents common to the two stories are tabulated as under, those of the Mabinogi being taken as the standard:—

Mabinogi.			Conte du Graal.
Inc.			Inc.	Chrestien.
1.	Encounter with the knights.		1.
2.	Adventure with the damsel of the tent.		2.
3.	Avenging of the insult to Guinevere; incident of the dwarves; departure from Court.		3 and 4.
5.	Arrival at house of first uncle (found fishing); instruction in arms.		5.	Gonemans.
6.	Arrival at house of second uncle (Grail Castle). First sight of the talismans (head in basin and lance).		7.	Uncle found fishing; talismans, Grail and lance.
7.	Reproaches of foster-sister whom he finds lamenting over a dead knight.		8.	Reproached by his cousin; also instructed by her about the magic sword.
8.	Adventure with the damsel of the besieged castle who offers herself to hero.		6.	Blanchefleur, Gonemant’s niece.
9.	Second meeting with the lady of the tent.		9.
10.	First encounter with the sorceresses of Gloucester, who are forced to desist from assailing hero’s hostess.
11.	Adventure of the drops of blood in the snow.		10.
20.	Reproaching of Peredur before the Court by the loathly damsel.		11.
21.	Gwalchmai’s adventure with the lady whose father he had slain.		14.
22.	Peredur’s meeting the knight on Good Friday, and confession to priest.		15.	Hermit, hero’s uncle.
			Gautier.
24.	Arrival at the Castle of Wonders (Chessboard Castle); stag hunt; loss of dog; fight with the black man of the cromlech.		Inc. 7, 8, and partly 13 and 18.
25.	Second arrival at the (Grail) castle; achievement of the Quest by destruction of sorceresses of Gloucester. “Thus it is related concerning the Castle of Wonders.”		22.	In so far as Gautier ends his part of the story here with the hero’s second arrival at the Grail Castle, but no similarity in the incidents.

The sequence is thus exactly the same in the Mabinogi and in Chrestien, with the single exception of the Blanchefleur incident, which, in the French poem precedes, in the Welsh tale follows, the first visit to the Grail Castle. The similarity of order is sufficient of itself to warrant the surmise of a relation such as that of copy to original. If the Mabinogi be examined closely, much will be found to strengthen this surmise. Thus, Birch-Hirschfeld has pointed out that when Peredur first sees the knights, and on asking his mother what they may be, receives the answer, “Angels, my son”; this can only be a distorted reminiscence of Perceval’s own exclamation,

... Ha! sire Dex, Merchi!
Ce sont angle que je voi ci! (1,349-50).

as the hero’s mother would be the last person to describe thus the knights whom she has done her best to guard her son from knowledge of. Again, Simrock has criticised, and with reason, the incident of Peredur’s being acclaimed by the dwarf on his arrival at Arthur’s court as the chief of warriors and flower of knighthood. In the corresponding incident in Chrestien, the hero is told laughingly by a damsel that he should become the best knight in the world, and she had not laughed for ten years, as a fool had been wont to declare. This is an earlier form than that of the Mabinogi, and closer to the folk-tale account. Thus, to take one instance only, in Mr. Kennedy’s Giolla na Chroicean Gobhar (Fellow with the Goat-skin) [Fictions of the Irish Celts, p. 23], the hero comes to the King of Dublin, as Peredur to Arthur, clad in skins and armed with a club. “Now, the King’s daughter was so melancholy that she didn’t laugh for seven years, but when she saw Tom of the Goat-skin knock over all her father’s best champions, then she let a great sweet laugh out of her,” and of course Tom marries her, but not until he has been through all sorts of trials, aye, even to Hell itself and back. In Chrestien, the primitive form is already overlaid; we hear nothing further of the damsel moved to laughter nor of the prophetic fool; and in the Mabinogi it seems obvious that the hailing of the hero, added in Chrestien to the older laughter, has alone subsisted. Birch-Hirschfeld takes exception likewise to the way in which Peredur’s two uncles are brought upon the scene, the first one, corresponding to Gonemans in Chrestien, being found fishing instead of the real Fisher King, the lord of the Castle of the Magic Talismans, whilst at the latter’s, Peredur has to undergo trials of his strength belonging properly to his stay at the first uncle’s. Evidently, says Birch-Hirschfeld, there has been a confusion of the two personages. Again, when Peredur leaves his second uncle on the morrow of seeing the bleeding head and spear, it is said, “he rode forth with his uncle’s permission.” Can these words be a reminiscence of Chrestien’s?

Et trueve le pont abaisciÉ,
Con li avoit ensi laissiÉ
Por ce que rien nel detenist,
De quele eure qu’il venist
Que il ne passat sans arriest (4,565-69).

We shall see later on that in the most primitive form of the unsuccessful visit to the Castle of the Talismans the hero finds himself on the morrow on the bare earth, the castle itself having vanished utterly. The idea of permission being given to leave is diametrically opposed to this earliest conception, and its presence in the Mabinogi seems only capable of explanation by some misunderstanding of the story-teller’s model.

The Blanchefleur incident shows some verbal parallels, “The maiden welcomed Peredur and put her arms around his neck.”

Et la damosele le prent
Par le main dÉbonnairement (3,025-26)
Et voit celi ajenouillie
Devant son lit qui le tenoit
Par le col embraciet estroit (3,166-68).

Can, too, the “two nuns,” who bring in bread and wine, be due to the “Il AbÉies,” which Perceval sees on entering Blanchefleur’s town? It may be noticed that in this scene the Welsh story-teller is not only more chaste, but shows much greater delicacy of feeling than the French poet. Peredur’s conduct is that of a gentleman according to nineteenth century standards. Chrestien, however, is probably nearer the historical reality, and the conduct of his pair—

S’il l’a sor le covertoir mise
·····
Ensi giurent tote la nuit.

is so singularly like that of a Welsh bundling couple, that it seems admissible to refer the colouring given to this incident to Welsh sources. Another scene presenting marked similarities in the two works is that in which the hero is upbraided before the court by the loathly damsel. In the Mabinogi she enters riding upon a yellow mule with jagged thongs: in Chrestien—

Sor une fauve mule et tint
En sa main destre une escorgie (5,991-2).

“Blacker were her face and her two hands than the blackest iron covered with pitch.”

Ains ne vÉistes si noir fer
Come ele ot les mains et le cor (5,998-99).

“And she greeted Arthur and all his household except Peredur.”

Le roi et ses barons salue
Tout ensamble comunalment
Fors ke Perceval seulement (6,020-3).

In the Mabinogi, Peredur is reproached for not having asked about the streaming spear; in Chrestien “la lance qui saine” is mentioned first although the Grail is added. Had Peredur asked the meaning and cause of the wonders, the “King would have been restored to health, and his dominions to peace.”

Li rices rois qui moult s’esmaie
Fust or tos garis de sa plaie
Et si tenist sa tiÈre en pais (6,049-51).

Whereas now “his knights will perish, and wives will be widowed, and maidens will be left portionless”—

Dames en perdront lor maris,
TiÈres en seront essilies,
Et puciÈles desconsellies;
Orfenes, veves en remanront
Et maint chevalier en morront (6,056, etc.).

In the “Stately Castle” where dwells the loathly damsel, are five hundred and sixty-six knights, and “the lady whom he loves best with each,” in “Castle Orguellos” five hundred and seventy, and not one “qui n’ait s’amie avoeques lui.” “And whoever would acquire fame in arms and encounters and conflicts, he will gain it there if he desire it.”

Que la ne faut nus ki i alle,
Qui la ne truist joste u batalle;
Qui viout faire chevalerie,
Si lÀ le quiert, n’i faura mie (6,075, etc.).

“And whoso would reach the summit of fame and honour, I know where he may find it. There is a castle on a lofty mountain, and there is a maiden therein, and she is detained a prisoner there, and whoever shall set her free will attain the summit of the fame of the world.”

Mais ki vorroit le pris avoir
De tout le mont, je quie savoir
Le liu et la piÈce de terre
U on le porroit mius conquerre;
······
A une damoisiÈle assise;
Moult grant honor aroit conquise,
Qui le siÈge en poroit oster
Et la puciÈle dÉlivrer (6,080, etc.).

In this last case certainly, in the other cases probably, a direct influence, to the extent at least of the passages quoted, must be admitted. But before concluding hastily that the Welsh story-teller is the copyist, some facts must be mentioned on the other side. Thus the incident of the blood drops in the snow, which Birch-Hirschfeld sets down as one of those taken over by the Mabinogi, with the remark that the Welsh story contains no trace of a passion as strong as Perceval’s for Blanchefleur, has been dealt with by Professor H. Zimmer in his “Keltische Studien,” vol. ii, pp. 200. He refers to the awakening of Deirdre’s love to Noisi by similar means, as found in the Irish saga of the Sons of Usnech (oldest MS. authority, Book of Leinster, copied before 1164 from older MSS.) as evidence of the early importance of this motif in Celtic tradition. The passage runs thus in English: “As her foster-father was busy in winter time skinning a calf out in the snow, she beheld a raven which drank up the blood in the snow; and she exclaimed, ‘Such a man could I love, and him only, having the three colours, his hair like the raven, his cheeks like the blood, his body like the snow.’”

Now the Mabinogi says, almost in the same words—the blackness of the raven and the whiteness of the snow, and the redness of the blood he compared to the hair and the skin and the two red spots upon the cheek of the lady that best he loved. In Chrestien there is no raven, and the whole stress is laid upon the three drops of blood on the snow, which put the hero in mind of the red and white of his lady’s face. As Zimmer justly points out, the version of the Mabinogi is decidedly the more primitive of the two; and that, moreover, as the incident does not figure at all in what Birch-Hirschfeld presumes to be Chrestien’s source, the Didot-Perceval, the following development of this incident must, ex hypothesi, have taken place. In the Didot-Perceval the hero is once upon a time lost in thought. To explain this, Chrestien invents the incident of the three drops of blood in the snow; the Mabinogi, copying Chrestien, presents the incident in almost as primitive a form as the oldest known one! Here, then, the Mabinogi has preserved an older form than Chrestien, alleged to have been its source in all those parts common to both. Nor is it certain that the fact of Peredur’s undergoing the sword-test in the Talisman Castle does show, as Birch-Hirschfeld maintains, that the Welsh story-teller confused the two personages whom he took over from Chrestien, Gonemans and the Fisher King. The sword incident will be examined later on; suffice here to say that no explanation is given in the Conte du Graal of the broken weapon; whereas the Mabinogi does give a simple and natural one. But these two instances cannot weaken the force of the parallels adduced above. In determining, however, whether these may not be due to Chrestien’s being the borrower, the differences between the two versions are of even more importance than the similarities.

What are these? The French romances belonging to the Perceval type of the Grail quest give two versions of the search for the magic talismans, that of the Conte du Graal and that of the Didot-Perceval. The latter pre-supposes an early history which, as already shown, cannot be looked upon as the starting point of the legend without postulating such a development of the latter as is inadmissible on a priori grounds, and as runs counter to many well-ascertained facts. The former is not consistent with itself, Manessier’s finish contradicting Chrestien’s opening on such an essential point as the cause of the maimed king’s suffering. Still the following outline of a story, much overlaid by apparently disconnected adventures, may be gathered from it. A hero has to seek for magic talismans wherewith to heal an uncle wounded by his brother, and at the same time to avenge him on that brother. What, on the other hand, is the story as told in the Mabinogi? A hero is minded by talismans to avenge the death of a cousin (and the harming of an uncle); it is not stated that the talismans pass into his possession. It is difficult to admit that either of these forms can have served as direct model to the other. If the Mabinogi be a simple copy of the Conte du Graal, whence the altered significance of the talismans? whence also the machinery by means of which the hero is at last brought to his goal, and which is, briefly, as follows? The woe which has befallen Peredur’s kindred is caused by supernatural beings, the sorceresses of Gloucester; his ultimate achievement of the task is brought about by his cousin, who, to urge him on, assumes the form (1) of the black and loathly damsel; (2) of the damsel of the chessboard, who incites him to the Ysbydinongyl adventure, reproves him for not slaying the black man at once, and then urges him into the stag hunt; (3) of the lady who carries off the hound and sends him to fight against the black man of the cromlech; “and the cousin it was who came in the hall with the bloody head in the salver and the lance dripping blood.” The whole of the incidents connected with the Castle of the Chessboard, which appear at such length in both the Conte du Graal and the Didot-Perceval, but without being in any way connected with the main thread of the story, thus form in the Mabinogi an integral portion of that main thread. Would the authors of the Conte du Graal have neglected the straight-forward version of the Welsh tale had they known it, or could, on the other hand, the author of the Mabinogi have worked up the disconnected incidents of his alleged model into an organic whole? Neither hypothesis is likely. Moreover the Conte du Graal and the Didot-Perceval, if examined with care, show distinct traces of a machinery similar to that of the Welsh story. Thus in Chrestien, Perceval, on arriving at the Fisher King’s, sees a squire bringing into the room a sword of such good steel that it might break in but one peril, and this the King’s niece (i.e., Perceval’s cousin) had sent her uncle to bestow it as he pleased; and the King gives it to the hero for—

... biaus frÈre ceste espÉe
Vous fu jugie et destineÉ (4,345-6).

After Perceval’s first adventure at the Grail Castle it is his “germaine cousine” (4,776) who assails him with her reproaches; she knows all about the sword (4,835-38) and tells him, how, if it be broken he may have it mended (4,847-59). So far Chrestien, who furthermore, be it noted, makes Blanchefleur Perceval’s lady-love, likewise his cousin, she being niece to Gonemans (3,805-95). A cousin is thus beloved of him, a cousin procures for him the magic sword, a cousin, as in the Mabinogi, incites him to the fulfilment of the quest, and gives him advice which we cannot doubt would have been turned to account by Chrestien had he finished his poem. Turning now to Gautier, in whose section of the poem are to be found the various adventures growing out of the chessboard incident, this difference between the Mabinogi and himself may be noted. In the former, these adventures caused by Peredur’s cousin serve apparently as tests of the hero’s strength and courage. The loss of the chessboard is the starting-point of the task, and the cousin reappears as the black maiden. Nothing of the sort is found in Gautier. True, the damsel who reproaches Perceval is in so far supernatural, as she is a kind of water-nix, but it is love for her which induces the hero to perform the task; she it is, too, who lends him the dog, and she is not identified with the “pucelle de malaire” who carries it off (22,604, etc.). But later on Perceval meets a knight who tells him that a daughter of the Fisher King’s (thus also a cousin of Perceval) had related to him how a knight had carried off a stag’s head and hound to anger another good knight who had been at her father’s court, and had not asked as he should concerning the Grail, for which reason she had taken his hound and had refused him help to follow the robber knight (23,163, etc.). This makes the “pucelle de malaire” to be Perceval’s cousin, and she plays the same rÔle as in the Mabinogi. True, when later on (Incident 13) Perceval finds the damsel, nothing is said as to her being the Fisher King’s daughter; on the contrary, as will be seen by the summary, a long story is told about the Knight of the Tomb, brother to her knight, Garalas, and how he lived ten years with a fay. She is here quite distinct from the lady of the chessboard to whom Perceval returns later. The version found in the Didot-Perceval agrees with the Mabinogi as against Gautier in so far that the hero is in love with the mistress of the castle, and not with the damsel who reproaches him for throwing away the chessmen. This reproaching damsel is not in any way identified with the lady who carries off the hound, who is described as “une vieille,” and of whom it is afterwards told “elle estoit quand elle voloit une des plus belles damoiselles du monde. Et est cele meismes que mon frÈre (the brother of the Knight of the Tomb, who here, as in Gautier, is the lover of a fay) amena À la forest,” i.e., she is the fay herself, sister to the lady of the Chessboard Castle, who hated her and wished to diminish her and her knight’s pride (p. 469). Here, again, a connection can be pieced out between the various personages of the adventure; and it appears that the hero is driven to his fight against the Knight of the Tomb by a fair damsel transformed into a mysterious hag.[72] The Mabinogi thus gives one consistently worked-out conception—transformed hag = Peredur’s cousin—which may be recovered partly from that one of the two discordant versions found in Gautier which makes the pucelle de malaire to be the Fisher King’s daughter, hence Perceval’s cousin, and connects the stag hunt with the Grail incident, partly from the Didot-Perceval, which tells how the same pucelle de malaire is but playing a part, being when she wills one of the fairest maids of the world. Now we have seen that the stag hunt is just one of those portions of the story in which are found the closest verbal similarities between Gautier de Doulens and the Didot-Perceval. It is, therefore, perplexing to find that there is not more likeness in the details of the incident. But the similarities pointed out concern chiefly the first part of the incident, and are less prominent in the latter part (the hero’s encounter with the Knight of the Tomb). This, taken together with the difference in the details of the incident just pointed out, strengthens the opinion expressed above, that the Didot-Perceval and Gautier are not connected directly but through the medium of a common source, the influence of which can be seen distinctly in certain portions of either story, and that when this source fails they go widely asunder in their accounts. That such an hypothesis is not unreasonable is shown by the fact that Gautier has two contradictory forms of this very story, one of which, that which makes the hound-stealing damsel a daughter of the Fisher King, is on all fours with the Mabinogi, whilst the other is more akin to, though differing in important respects from, that of the Didot-Perceval. In this case, at least, Gautier must have had two sources, and if two why not more?

It may be urged in explanation of the similarities between Gautier and the Mabinogi, that the author of the latter used Gautier in the same free way that he did Chrestien, but that getting tired towards the close of his work he abridged in a much more summary fashion than at first. If the comparison of the versions of the stag hunt found in either work be not sufficient to refute this theory, the following consideration may be advanced against it: if the Mabinogi derives entirely from the Conte du Graal, how can the different form given to the Grail episode be accounted for?—if it only knew Chrestien, where did it get the chessboard adventure from, and if it knew Gautier as well as Chrestien why did it not finish the Grail adventure upon the same lines as it began, i.e., partly in conformity with its alleged model?

Is Manessier any nearer than Gautier to the Mabinogi in the later portion of the tale? The chief points of the story told by him may be recapitulated thus:—The Grail damsel is daughter of the Fisher King, the damsel of the salver, daughter of King Goon Desert, his brother (i.e., both are cousins to Perceval); Goon Desert, besieged by Espinogre, defeats him, but is treacherously slain by his nephew Partinal, the latter’s sword breaking in the blow. Goon’s body is brought to the Fisher King’s castle, whither the broken sword is likewise brought by Goon’s daughter to be kept until a knight should come, join together the pieces, and avenge Goon’s death. In receiving the sword the Fisher King wounds himself through the thighs, and may not be healed until he be avenged on Partinal. Perceval asks how he may find the murderer, the blood vengeance (faide = O.H.G. Fehde) being on him. Perceval fights with Partinal, slays him, cuts off his head as token of his victory, returns to the Fisher King’s castle, lighting upon it by chance, heals the Fisher King by the mere sight of the head, which is fixed on a pike on the highest battlements. At the death of his uncle Perceval succeeds him as King of the Grail Castle. Here, then, as in the Mabinogi, the story turns definitely upon a blood feud; the same act which brings about the death of one relative of the hero, also causes, indirectly, it is true, the laming of another, even as in the Mabinogi the same supernatural beings kill Peredur’s cousin and lame his uncle; the cousin reappears again, bringing the magic sword by whose aid alone the hero can accomplish the vengeance, and uttering the prediction the fulfilment of which will point out the destined avenger. Finally, if the Mabinogi seems to lay special stress upon the head of the murdered man, Manessier lays special stress upon the head of the murderer. Now it is quite evident that the Mabinogi cannot have copied Manessier. It has been alleged that the Welsh story-teller, adapting Chrestien to the taste of his fellow countrymen, substituted a blood feud for the Grail Quest, but what reason would he have had for thus dealing with Manessier? He had simply to leave out the Christian legendary details, which in Manessier are, one can hardly say, adapted to the older form of the story, to find in that older form a clear and straightforward account with no admixture of mystical elements. It is impossible to explain the strong general similarity of outline with the equally marked divergences of detail (Sorceresses of Gloucester instead of Partinal, etc.,) except by saying that both, though going back to a common legendary source, are unconnected one with another.

The facts thus dealt with may be recapitulated as follows:—There is marked similarity in general outline between the Mabinogi and the Conte du Graal in the adventures common to both; in that portion of the Conte du Graal due to Chrestien there occur, moreover, many and close verbal parallels, and the corresponding part of the Mabinogi is told at greater length than the remainder of the incidents common to both works. That which answers in the Mabinogi to the Grail Quest forms a clear and straightforward whole, the main features of which may be recovered from the Conte du Graal, but in varying proportions from the various sections of that work. Thus the indications of this Mabinogi talisman quest, the central intrigue, as it may be called, of the tale, are in Chrestien of the slightest nature, being confined to passing hints; in Gautier they are fuller and more precise, though pointing to a version of the central intrigue different, not only in details but in conception, from that of the Mabinogi; in Manessier alone is there agreement of conception, although the details still vary. Finally, those portions of the Mabinogi which are in closest verbal agreement with Chrestien contain statements which cannot easily be reconciled with this central intrigue.

These facts seem to warrant some such deductions as these. Bearing in mind that the Mabinogi is an obvious piecing together of all sorts of incidents relating to its hero, the only connecting link being that of his personality, its author may be supposed, when compiling his work, to have stretched out his hand in all directions for material. Now a portion of the Peredur sage consisted of adventures often found elsewhere in the folk-tale cycles of the Great Fool and the Avenging Kinsman—cycles which, in Celtic tradition, at least, cover almost the same ground as the one described by J. G. von Hahn under the title, “Die Arische Aussetzung und RÜckkehr-Formel.” In the original of the Mabinogi this portion probably comprised the childhood and forest up-bringing, the visit to Arthur with the accompanying incidents, the training by the uncle (who may have been the Fisher King), the arrival at the (bespelled) castle, where the hero is to be minded of his task by the sight of certain talismans and of his cousin’s head, the reproaches of the loathly damsel, her subsequent testing of the hero by the adventures of the chessboard, stag hunt, etc., the hero’s final accomplishment of the task, vengeance on his kindred’s enemies, and removal of the spells. There would seem to have been no such love story as that frequently found in stories of the Great Fool class, e.g., in the Irish one (supra, p. 134). This original was probably some steps removed from being a genuine popular version; the incidents were presented in a way at once over-concise and confused, and some which, as will be seen in the next chapter, the living folk-tale has preserved were left out or their significance was not recognized. What more natural than that the author of the Mabinogi in its present form, knowing Chrestien, should piece out his bare, bald narrative with shreds and patches from the Frenchman’s poem? The moment Chrestien fails him, he falls back into the hurried concision of his original. His adaptation of Chrestien is done with singularly little skill, and at times he seems to have misunderstood his model. He confines his borrowing to matters of detail, not allowing, for instance, Chrestien’s presentment of the Grail incident to supersede that of his Welsh original. In one point he may, following Chrestien, have made a vital change. It seems doubtful whether the Welsh source of the Mabinogi knew of a maimed king, an uncle to be healed through the hero’s agency; the sole task may have been the avenging the cousin’s death. True the “lame uncle” appears at the end, but this may be due to some sudden desire for consistency on the arranger’s part. But whether or no he was found in the Welsh story preserved in the Mabinogi, he certainly played no such leading part as in the Conte du Graal. The two stories deal with the same cycle of adventures, but the object of the hero is not the same in both, and, consequently, the machinery employed is not quite the same. The present Mabinogi is an unskilful fusion of these two variations upon the one theme.[73]

Light is also thrown by this investigation upon the question of Chrestien’s relationship to his continuators. Birch-Hirschfeld’s theory that the Didot-Perceval was the source of Chrestien and Gautier has already been set aside. Apart from the reasons already adduced, the fact that it does not explain from whence Manessier got his ending of the story would alone condemn it. It must now be evident that Chrestien and two of his continuators drew from one source, and this a poem of no great length probably, the main outlines of which were nearly the same as those of the Welsh proto-Mabinogi given above, with this difference, that the story turned upon the healing of the uncle and not the avenging the cousin’s death. This poem, which seems also to have served, directly or indirectly, as one of the sources of the Didot-Perceval, had probably departed from popular lines in many respects, and may, though this would be an exceedingly difficult question to determine, have begun the incorporation of the Joseph of Arimathea legend with its consequent wresting to purposes of Christian symbolisms of the objects and incidents of the old folk-tale.

Such an incorporation had almost certainly begun before Chrestien’s time, and was continued by him. There can be little doubt that he dealt with his model in a free and daring spirit, altering and adding as seemed best to him. This alone explains how Manessier, slavishly following the common original, tells differently the cause of the lame king’s wound. Gautier, who lacked Chrestien’s creative power, though he often equals him in the grace and vivacity of his narrative, seems to have had no conception of a plan; the section of Conte du Graal which goes under his name is a mere disorderly heap of disconnected adventures brought together without care for consistency. But for this very reason he is of more value in restoring the original form of the story than Chrestien, who, striving after consistency, harmony, and artistic development of his tale, alters, adds to, or retrenches from the older version. Gautier had doubtless other sources besides the one made use of by Chrestien. This does not seem to be the case with Manessier, who, for this portion of the story, confined himself to Chrestien’s original, without taking note of the differences in motif introduced by his predecessor. What is foreign to it he drew from sources familiar to us, the Queste and Grand S. Graal, from which more than two-thirds of his section are derived.

In working back to the earliest form of the Perceval-sage, Mabinogi and Conte du Graal are thus of equal value and mutually complementary. Both are second-hand sources, and their testimony is at times sadly corrupt, but it is from them chiefly that information must be sought as to the earlier stages of development of this legendary cycle. They do not by themselves give any satisfactory explanation of the more mysterious features of the full-blown legend, but they do present the facts in such a way as to put out of court the hypothesis of a solely Christian legendary origin. Before proceeding further it will be well to see if the English Sir Perceval has likewise claims to be considered one of the versions which yield trustworthy indications as to the older form of the story.

This poem, described by Halliwell as simply an abridged English version of the Conte du Graal, has, as may be seen by reference to Ch. IV, been treated with more respect by other investigators, several of whom, struck by its archaic look, have pronounced it one of the earliest versions of the Perceval sage. It has quite lately been the object of elaborate study by Paul Steinbach in his dissertation: “Uber dem Einfluss des Crestien de Troies auf die altenglische literatur,” Leipzig, 1885. The results of his researches may be stated somewhat as follows: the two works correspond incident for incident down to the death of the Red Knight, the chief differences being that Perceval is made a nephew of King Arthur, that the death of his father at the hands of the Red Knight is explained as an act of revenge on the part of the latter, that Arthur recognizes his nephew at once, and tells him concerning the Red Knight, and that the burning of the Red Knight, only hinted at in Chrestien’s lines—

Ains auroie par carbonees.
Trestout escarbelliÉ le mort, etc. (2,328-9).

is fully told in the English poem. After the Red Knight incident the parallelism is much less close. The English poem has incidents to itself: the slaying of the witch, the meeting with the uncle and nine cousins, the fight with the giant for the ring, the meeting with and restoring to health the mother. Of the remaining incidents, those connected with Lufamour are more or less parallel to what Chrestien relates of his hero’s adventure with Blanchefleur, and that of the Black Knight, with that of the Orgellous de la Lande in Chrestien. Of the 2,288 verses of the English poem the greater part may be paralleled from Chrestien, thus:—

P. of G.				Cr.
1-160				485-940
161-188				941-1,206
189-256				1,207-82
257-320				1,283-1,554
321-432				1,555-1,828
433-80				1,829-1,970
481-600				2,091-2,170
601-56			{ {	2,055-90
				2,135-59
				2,171-2,225
657-740				2,268-2,312
741-820				2,313-2,398
1,061-1,108				4,000-4,060
1,109-1,124				5,511-553
1,381-1,540				5,600-5,891
953-1,012	} }
1,125-1,380			{	2,900-3,960
1,541-1,760				4,088-94
1,761-1,808				4,095-4,150
1,809-1,951				4,865-5,375

the incidents comprised v. 821-952 and 1,953-2,288, being the only one entirely unconnected with Chrestien. This general agreement between the two works shows the dependence of the one on the other. But while evidently dependent, the English poem, as is shown by the differences between it and its French original, belongs at once to a less and to a more highly developed stage of the Perceval sage. The differences are thus of two kinds, those testifying to the writer’s adherence to older, probably Breton, popular traditions and those due to himself, and testifying to the skill with which he has worked up his materials and fitted portions of Chrestien’s poem into an older framework. Of the first kind are: the statement that Perceval meets with three knights instead of five as in Chrestien, the English poem agreeing here with the Mabinogi; the mention of his riding on a mare and of his being clad in goat-skins, the English poem again agreeing rather with the Mabinogi than with Chrestien, and showing likewise points of contact with the Breton ballads about Morvan lez Breiz, printed by VillemarquÉ in the Barzaz Breiz. The combat with the giant may likewise be paralleled from the Lez Breiz cycle in that hero’s fight with the Moorish giant. These points would seem to indicate knowledge on the author’s part of popular traditions concerning Perceval forming a small cycle, of which the departure from, and return to the mother were the opening and closing incidents respectively. This form of the story must have been widely spread and popular to induce the author to leave out as much as he has done of Chrestien’s poem in order to bring it within the traditional framework. He accomplished his task with much skill, removing every trace of whatever did not bear directly upon the march of the story as he told it. In view of this skill differences which tend to make the story more consequent and logical may fairly be ascribed to him. Such are: the making Perceval a nephew of Arthur, the mention of a feud between the Red Knight and Perceval’s father, the combat with the witch arising out of Perceval’s wearing the Red Knight’s armour, and the other adventures which follow eventually from the same cause, the feature that the ring taken by Perceval from the lady in the tent is a magic one, endowing its wearer with supernatural strength, the change made between this ring and his mother’s which prepares the final recognition, etc. The original poem probably ended with the reunion of mother and son, the last verse, briefly mentioning the hero’s death, being a later addition. To sum up, Sir Perceval may be looked upon as the work of a folk-singer who fitted into the old Breton framework a series of adventures taken partly from Chrestien, partly from the same Breton traditions which were Chrestien’s main source, and with remarkable skill avoided all such incidents as would not have accorded with the limits he had imposed upon himself.

Against this view of Steinbach’s it might be urged that a writer as skilful as the author of Sir Perceval is assumed to be could easily have worked Chrestien’s Grail episode into his traditional framework. A more plausible explanation, assuming the theory to be in the main correct, might be found in the great popularity in this country of the Galahad form of the Quest, and the consequent unwillingness on the author’s part to bring in what may have seemed to him like a rival version. Steinbach has not noticed one curious bit of testimony to the poem’s being an abridgment of an older work, more archaic in some respects than Chrestien. When the hero has slain the Red Knight he knows not how to rid him of his armour, but he bethinks him—

... “My moder bad me
Whenne my dart solde brokene be,
Owte of the irene brenne the tree,
Now es me fyre gnede” (749-52).

Now the mother’s counsel, given in verses xxv-vi are solely that he should be “of mesure,” and be courteous to knights; nothing is said about burning the tree out of the iron, nor does any such counsel figure either in Chrestien or in the Mabinogi, which in this passage has copied, with misunderstandings, the French poet.[74] The use of Chrestien by the author of Sir Perceval seems, however, uncontestable; and, such being the case, Steinbach’s views meet the difficulties of the case fairly well. It will be shown farther on, however, that several of the points in which the German critic detects a post-Chrestien development, are, on the contrary, remains of as old and popular a form of the story as we can work back to. Accepting, then, the hypothesis that Sir Perceval, like the Mabinogi, has been influenced by Chrestien, what is the apparent conclusion to be drawn from the fact that the former omits the Grail episode altogether, whilst the latter joins Chrestien’s version to its own, presumably older one, so clumsily as to betray the join at once? May it not be urged that Chrestien’s account is obviously at variance with the older story as he found it? may not the fact be accounted for by the introduction of a strange element into the thread of the romance? This element would, according to Birch-Hirschfeld, be the Christian holy-vessel legend, and it would thus appear that the Grail is really foreign to the Celtic tradition. Let me recapitulate briefly the reasons already urged against such a view. The early history of the Grail, that part in which the Christian element prevails, must certainly be regarded as later than the Quest, to which it could not have given rise without assuming such a development of the romance as is well nigh incredible—the Quest versions, moreover, all hang together in certain respects, and point unmistakably to Celtic traditions as their source. These traditions must then be examined further to see if they contain such traces of the mystic vessel as are wanting in the Mabinogi and the English poem, and as may have given rise to the episode as found in the French romances. As Perceval is the oldest hero of the Quest, and as the boyhood of Perceval, forming an integral part of all the oldest Quest versions presents the strongest analogies with the folk-tale of the Great Fool, it is this tale which must now be examined.

---