AMERICAN

JOURNAL OF PHILOLOGY

VOL. XXXVII, 4.

WHOLE NO. 148.

I. THE STAG-MESSENGER EPISODE.

It seems to be generally agreed by those who are best qualified to express an opinion about mediaeval Celtic literature that that form of the so-called fairy induction motif in which a stag or hind serves as a messenger to lead a mortal hero to a fée is of Celtic origin, and the presence of this episode in a mediaeval poem is cited as evidence that the poem is derived from Celtic sources or has undergone Celtic influence. This conclusion seems to be based upon the fact that the episode in question occurs most frequently in poetry dealing with the "matière Bretagne", but the evidence is not conclusive enough, as I shall try to make clear, to warrant such definite statements as are made concerning its Celtic provenience. It should not be forgotten that before we can definitely assign a given episode, occurring in mediaeval French poetry or prose, to Celtic tradition or to a Celtic literary source, there are at least two conditions which must be met: 1. The episode must be proved beyond reasonable doubt, on evidence independent of the French work in which it is found, to have been a part of Celtic tradition, oral or written, before its earliest appearance on French soil. 2. It must not occur in any form, oral or written, which is demonstrably free from Celtic influence, and at the same time equally current, and also equally accessible and well-known to a French writer as any possible Celtic source. The first of these conditions is accepted1 even by the most enthusiastic

1 Cf. e. g. the remarks of Professor Cross, Mod. Phil. 12, 1915, p. 590; Kittredge, A Study of Gawain and the Green Knight, pp. 9 sq.

advocates of the Celtic theory, but the second, which is just as important as the first, just as necessary for convincing proof, seems to be entirely disregarded. In the following study I have tried to keep both these conditions in mind, and my endeavor has been to present, unbiased by preconceived theories, certain facts which seem to have been overlooked and yet are certainly not without importance.

As a fair example of the usual form of the episode may be cited the Lay of Graelent,' 193 sq. Graelent, having been wrongfully treated by his king and sought as her lover by the queen, mounts his horse during the former's absence and leaves the court unattended. He is riding sadly through the forest when suddenly a hind "tute blance plus n'est nois nul sor brance", starts up before him. He pursues her hotly but cannot overtake her, and finally she leads him to a fountain of clear, sparkling water in which a maiden is bathing; her clothes are hanging on a tree near by, and two other maidens are serving her. Graelent is at once smitten with her beauty, forgets all about the hind, and to keep the maiden from escaping, takes possession of her clothes, and in the end works his will upon her; she grants him her love, promises him bountiful treasure and declares that she will be with him whenever he desires her, but that if at any time he reveals their relations he shall lose her. In this story, as Professor Schofield pointed out long ago,2 there is a confusion of a fairy mistress and a swan-maiden, and he also called attention to the fact that the same situation exists in a version of the story of Wayland and the swan-maidens contained in the Middle High German poem of Friedrich von Schwaben (14th cen.). Here the details are practically the same as in Graelent. It is clear, however, that originally the hind-messenger episode was not connected with this swanmaiden story, for in the earliest version of the latter, that in the

'Ed. Roquefort, Poésies de Marie de France, I, pp. 202 sq. The Lay has been studied by Schofield, Pub. Mod. Lang. Ass. 15, 1900, pp. 121 sq. He recognizes that the so-called Breton lays contain much matter that is not Celtic, but he regards the Celtic character of the hind-messenger episode as established. This is also the conclusion of Cross in the article cited, and of Kittredge, 1. c., pp. 231 sq.

'Harvard Stud. and Notes, 5, 1896, pp. 236-7, and also his article referred to above. Cross, 1. c., pp. 616 sq., thinks that the entire lay, hind-messenger and "the other world woman of the swan-maiden type", was part of Celtic tradition before the 12th cen.

Eddic Völundarkuiða1 (9th cen.), no hind appears; Wayland and his brothers have arrived at Ulfdal, and there one morning on the border of a lake, they come upon the swan-maidens with their swan-plumage beside them.

66

In another swan-maiden story, however, a story which is very similar in many of its details to Graelent, and doubtless prior to it in date, the same method is employed to bring the hero to the maiden. This is the story of the Cygni, told by the seventh Wise Man in the Dolopathos of Johannes de Alta Silva. Here the hero goes hunting with his dogs and sees cervam nive candidiorem, decem in quolibet cornu habentem ramos ". 3 He pursues the hind down into a well-wooded valley, loses sight of both the hind and his dogs, and finally "fontem repperit nimphamque in eo cathenam auream tenentem manu nudaque menbra lavantem conspicit ". He is overcome by her beauty and, like Graelent, forgets the hind and his dogs, and approaching her by stealth seizes the chain "in qua virtus et operatio virginis constabat". He then promises to make her his wife and on the next morning takes her to his home, where in due time she bears him six sons and a daughter. These the jealous mother of the youth orders to be exposed, and they are found by an old hermit who feeds them on hind's milk and raises them as his own children. The French version of Herbert, vs. 9188 sq., shows no important variation from this

account.

This story is the earliest extant version of the Chevalier au Cygne, and this introductory episode of the hind-messenger occurs in most of the later versions; the feeding of the chil

4

'Cf. Saemundar Edda, ed. Bugge, p. 163. For further references on this subject, cf. Cross, l. c., p. 621, n. 4.

'I quote from the edition of Hilka, Heidelberg, 1913, pp. 80 sq.

That this hind should have horns is noteworthy. The hind of Keryneia which Hercules hunted had golden horns (Pind. O. 3, 29: Χρυσόκερων ἔλαφον θήλειαν), and hinds are sometimes given horns i ancient writings, cf. Ael. de nat. an. 7, 39. On a gem which depicts the suckling of Hercules' son Telephus by a hind the hind is given horns; the gem is in Vienna; cf. the catalogue of Sacken and Kenner, no. 663. On this matter, cf. a paper by Ridgeway, summarized in A. J. A. IX, 1894, p. 571. He suggests that the "horned hind of Keryneia” was a reindeer, a species of deer found in northern Europe and Asia.

'Cf. Todd, La Naissance du Chevalier au Cygne, Pub. Mod. Lang. Ass. 4, 1889, Intro. p. II sq. He makes no special comment on the intro

dren is generally due directly to a hind. In the French romance, Chevalier au Cygne, however, the hind-messenger episode does not have a place, nor in the English metrical romance, Cheuelere Assigne, based upon it.1 In the English prose romance, The Knight of the Swanne, on the other hand, there is a rationalized account of the episode. The young king Oriant goes hunting with his dogs, raises a hart, and in his pursuit of it is led to a river across which the hart swims and escapes. The king then comes upon a fountain "which was so clere and pleasaunt" that he dismounts and sits down under a tree. Presently a young damsel, Beatrice, appears, accompanied by a noble knight and two squires. Oriant falls in love with her, pays his court in true knightly fashion, and later makes her his wife. The children are born, exposed, and found by a devout hermit who, not knowing how to feed them, prays to God for aid, and miraculously there appeared in his house a fair white goat which gave them suck.

Very close to the form of the episode found in these stories is that occurring in the shorter version of the romance of Parténopeus of Blois, a version based upon a lost French original and now existing in a fragmentary form in English, and in a complete but altered form in a Danish, Icelandic, and SpanishCatalan translation. In view of the undoubted non-Celtic

ductory episode, nor does Paris, Rom. 19, 1890, pp. 314 sq., or Huet, Rom. 34, 1905, pp. 206 sq. According to Paris "Jean a bien probablement recueilli ce conte dans la tradition orale du pays où il l'écrivait"; Huet notes the composite character of the story in Dolopathos and suggests a written source, a poem already attached to the Crusades and in the form of a chanson de geste. It may be noted that some claim an oriental origin for the Knight of the Swan; on this whole matter, cf. Jaffray, The Two Knights of the Swan, London, 1910, pp. 2 sq., 23 sq.

1

1 Ed. Gibbs, E. E. T. Extra Ser. VI. For the prose romance, cf. Thoms, Early Prose Romances, v. 3.

2 Professor Tupper reminds me of the opening episode in Scott's Lady of the Lake which resembles this account. In view of Scott's knowledge of these old romances it cannot be doubted that he had some such story in mind when he wrote his poem.

The English version is printed by Bödtker in his edition of Partonope of Blois, E. E. T. Extr. Ser. CIX, pp. 481 sq. The Spanish-Catalan version (in prose) which was accessible to me is an old volume printed at Gerona without date. I have not compared the Danish and Icelandic versions.

origin of the main feature of this romance, the appearance in it of the stag-messenger episode is, in view of what I shall show below, of great importance. Melior, the young queen of Byzantium, is ripe for a husband and seeks all over the world for one worthy of her. She finds him in the young prince Partonope, and in order to get him to her she raises by enchantment, for she possesses wondrous powers, a white hart with wide horns as he was hunting one day with his uncle and attendants in the forest of Ardennes. Partonope follows the hart in a vain pursuit and is led to the sea-shore, where appears a marvelous and beautiful ship. He goes on board, finds that he is the only passenger, and is carried to another land, to an uninhabited castle.1 Here unseen hands wait upon him and at night the lady comes. In the Spanish-Catalan version the inducting animal is a "porch salvatge "," which, we are told at the outset, Melior has raised by her magic power. This boar occurs also in the longer version.3 Here Parténopeus kills one boar and a second appears which, when he pursues it, leads him to the sea, into which it plunges and swims across to safety. P. wanders in the forest unable to find his way home and does not come upon the ship until the following night. Not until the queen visits him and discloses her identity do we learn that she by her witchcraft brought about the hunt, made him follow the boar, sent the enchanted ship, etc.

These details in the Parténopeus story are paralleled by several in the Lay of Guigemar of Marie de France. Here also it must be noted that there are, as in the Parténopeus story,

'This romance, as is well known, is related to the Cupid and Psyche story of Apuleius. In view of the tendency to ascribe such magic ships and voyages to fairy castles to Celtic tradition, I would call attention not only to the situation and description of the castle in Apuleius, 5, 1, "Psyche-videt lucum proceris et vastis arboribus consitum, videt fontem vitreo latice perlucidum", but also to the fact that a voyage to it by ship is implied in 5, 15, "iugum sororium-recta de navibus scopulum petunt ".

'I shall cite below a neglected passage from Ovid which is strikingly similar.

'French version ed. Crapelet, Paris, 1834; English version ed. Bödtker, 1. c. There is also a German version by Conrad v. Würtzburg, ed. Bartsch, Wien, 1871.

'Ed. Warnke, Die Lais de M. de France, 2d ed., Halle, 1900, pp. 5 sq.