SOURCE: Trindade, Ann. “The Enemies of Tristan.” Medium Aevum 63, no. 1 (1974): 6-21.

[In the following essay, Trindade discusses the narrative structure of the Tristan legend as it exists in poetic versions by Eilhart and others, placing particular emphasis on the function of antagonists in the story.]

The enemies of Tristan are many and varied in what are usually called the ‘primary’ versions of the legend. Their number and degree of individualization varies, and while the principal editions and studies of Tristan texts have included comments on individual variants as they occur, there has been, as far as I am aware, no study devoted entirely to this group of characters alone. I propose to show that there are important advantages to be gained from studying them as a group and in terms of their narrative function.

First, while a number of influential Arthurian scholars have maintained that the study of origins is less important than the application of traditional literary criteria to the texts themselves,¹ nevertheless the fact that we refer so frequently to the ‘legend’, or ‘story’, or Tristan independently of the Tristan versions indicates that the processes of transmission and development will always be a legitimate object of study. The Arthurian field is not unique in possessing a number of traditional themes and characters which persist throughout several centuries in the oral and written literature of widely divergent cultural groups, but these characteristics make it a particularly appropriate choice of field for the comparative and the historical approach. Scholars will continue to try to trace the evolution of the Tristan legend, or to assess the contribution of the Continental adaptors to the Celtic deposit. They will ask whether the enemies of Tristan, singly or collectively, figured in the earliest forms of the story, whether the variation in names and distribution of rôles is derived from the independent cognate versions. Students of comparative literature will ask whether the main narrative devices employed in the creation and distribution of rôles can be seen to relate to any universal or at least widely distributed forms with which they may be familiar, and perhaps use this material as the basis for further comparative study.

Secondly, a survey of the enemies as a group enables us to confine our attention to a limited but important area. Logically, the category of enemies might be very widely drawn, to include any character who fulfils a hostile function from the point of view of Tristan and Isolt. This would include, for instance, the kidnapping merchants and the Morholt of the first section of the story as well as the adversaries encountered by Tristan and Kaherdin in the final part, the jealous husband whom Tristan helps Kaherdin to outwit, Riole von Nantis and the other foes against whom Tristan takes up arms on his friend's behalf and even Isolt of the Whitehands, his wife. Though I shall have more to say about some of these later, it is primarily with the more restricted group of Tristan's enemies at court that I am here concerned. These are the opponents whose function is part of the intrigue surrounding the triangular relationship between Tristan, Mark and Isolt. I shall first list them briefly for convenience, with some reference to their activities, before attempting to consider them as a group.

Perhaps the most prominent among these and the one whose portrayal comes closest to modern ideas of characterization is Andret, who appears in Eilhart and Béroul (and in the Prose version, with which I am not here concerned, as Sandret), though not in Thomas. He appears chiefly in the following episodes in Eilhart: The Chips on the Stream (3250-3702); the Flour between the Beds (3765-3975); the subsequent banishment of Tristan (4203-4242); Tristan's Second and Third Visits to Cornwall (7445-7714 and 8028-8548); and finally, in a preliminary skirmish with the disguised Tristan just before Tristan enters Mark's court as the Fool (8695-8786). He also appears in Béroul, though in a considerably reduced rôle. In Eilhart's version he has certain distinguishing characteristics—he is usually represented as acting in connection with other enemies, the dwarf or the hostile counts and dukes, he is the guardian who attends the Queen and he is, like Tristan, a nephew of Mark by his sister. Ewert has argued, convincingly, that Béroul's version displays knowledge of these traditional attributes, since Andret is represented there as accompanying the Queen; though in one reference he is made to speak on Tristan's behalf and in another he is killed by Tristan, this could be the residue of an account in which he was at once a close companion and a rival of Tristan.² There is no suggestion in Béroul's text that Andret is a nephew of the king, unless we accept that the ‘nevo’ in line 2869 refers forward to ‘Andrez, qui fu nez de Nicole’, and not to Tristan, as is generally held.³

It is interesting in passing to note that all the references to Andret—or Audret—occur in the part of the poem which Raynaud de Lage,⁴ Rita Lejeune⁵ and T. B. W. Reid⁶ would assign to ‘Béroul II’, yet it is also the first part in which we find the closest correspondence between Béroul and Eilhart. If, as some advocates of the theory of dual authorship maintain, the author of ‘Béroul II’ was working from different sources from those represented in the ‘version commune’, then the appearance in this part of Béroul of Andret—with all the inconsistencies alluded to by Ewert—may lend weight to the idea that he might be in some sense the ‘original’ enemy. Most commentators have been quick to recognize the similar pattern which links Mark, Tristan and Andret with Arthur, Gauvain and Modret.⁷ Attention has also been drawn to the widespread distribution of nephews in the Chansons de geste and in the courtly romances.⁸ The Arthur-Gauvain-Modred pattern is of course connected with the persistent motif of a Celtic abduction story seen in the abductions of Guenievre, and echoed here in the Harp and Rote episode.⁹

Closely connected with Andret in Eilhart's version in the early incidents at Mark's court is the dwarf, who is named as ‘Frocine’ in Béroul and perhaps Melot in Thomas, if Gottfried is to be followed.¹⁰ In all these versions the dwarf acts as the agent of the hostile faction in plotting to trap the lovers in two consecutive episodes, the Chips on the Stream and the Flour between the Beds. Béroul alone attributes occult powers to him, but Ewert considers that this is an invention. In Béroul alone too, we hear of the dwarf's death at the hands of the king in the curious little tale of the Horse's Ears,¹¹ while in the other poems we find his banishment and reinstatement at the court. In the latter part of both the Eilhart and the Thomas versions there is no trace of the dwarf, although various minor figures appear in the same intermediate role in the episodic returns of Tristan. There is another ‘nain’ who appears in the final section of Eilhart, the ‘nain Bedenis’, for whose story, which was certainly known to him, Thomas substitutes the watered-down story of Tristan le Nain. Dwarfs of course, like nephews, abound in medieval epics and romances, and there is no reason to imagine any elaborate symbolism in such a choice of name, even though Thomas presents him in such a way as to enhance by contrast the reputation of ‘Tristan l'amerus’.

Thomas's version is interesting in that the loosely coordinated groups of enemies in the versions of Eilhart and Béroul are replaced by one figure, the steward Meriadoc.¹² It is he who assumes Andret's rôle in the episodes of the Chips on the Stream and the Flour between the Beds, and appears in a number of loosely constructed incidents preceding these, the purpose of which seems to be to explain and motivate the growth of hostility between Meriadoc and Tristan, his former friend and confidant. As Bédier and more recent scholars such as Frappier¹³ have shown, this procedure is characteristic of Thomas, and it is an interesting reminder of some of the factors involved in any attempt to formulate a theory of mediæval poetics. On the one hand, we see the difference between modern notions of individuality and originality and the mediæval respect for tradition, oral and written, on the other, the relative sophistication of poets like Thomas, Marie de France and Chrétien in their interest in motivation and ideology. It is therefore difficult to say whether the reduction of all the enemies to one (presupposing that Thomas was working from sources, written or in part oral, which included groups of enemies) is another instance of editorial procedure, or whether it represents a tradition inherited by Thomas. As we know from studies of oral narrative and mythological material and its interaction with written forms, fission of characters and its opposite, fusion, occur with the same frequency. Both are manifestations of the transference of function rather than individual identity in traditional literature.¹⁴

The choice of name is not particularly revealing. It is intended to be Breton and characters with this name or one of its variants occur in other Arthurian texts.¹⁵ Like the dwarf, Meriadoc plays no part in the closing section of the story. Another knight, described as ‘uns riches cuns de grant alo’,¹⁶ brings the news of the ill-fated marriage to the Queen. He subsequently provokes the anger of Brengain by taunting her for having slept with a coward. Thomas has him killed by Kaherdin in revenge during Tristan's second return to Cornwall. This knight is described as a long-time admirer of Queen Isolt, and again, appears only in Thomas. In keeping with the ‘fin’ amors' ideology¹⁷ which seems to have influenced Thomas's presentation of the famous love-story, he is cast in the mould of the Provencal ‘losengier’, whose concern is to denigrate Tristan in the eyes of the Queen, rather than to reveal their association to the jealous husband. As one of the versions derived from Thomas gives Mariadokk here, it might be tempting to assume that the Meriadoc of the first part is the same as the Cariado of the Fragments; Bédier maintains that both may be inventions of the poet and this may well be so.

I turn now to a quite different group of enemies, the three barons of Béroul's version, named as Guenelon, Godoine and Denoalen.¹⁸ They are introduced abruptly in the extant part of Béroul's narrative and there is some confusion about their final destiny. In the early part of the story, where the versions of Eilhart, Béroul and Thomas follow a similar pattern (roughly from the Tryst under the tree to the banishment of the lovers to the forest) the rôle of the three barons corresponds to that of Andret and Meriadoc respectively. After the return from the forest, their rôle is less clearly defined, further confusion being engendered by the killing of one of them by Governal¹⁹ and the reappearance of all three later. Two minor points have attracted some discussion, why the barons should be three in number and whether their names have any significance. The fact is that not until line 1711 when one of them is killed, and 3138-19, when they are named, are they distinguished one from the other. Ewert's explanation of this seems an unnecessary rationalization.²⁰ Even less impressive is Vàrvaro's²¹ attempt to detect something approaching characterization here—in a way reminiscent of Bédier's remarks about characterization.²² In fact the weakness of Bédier's ground here, owing no doubt to his anxious desire to underscore the unique literary nature of his ‘archetype’, is surely emphasized by the fact that even the most ardent Bédierists have never attempted to defend his position by using this particular argument. On the contrary, the existence of three closely-related characters who seem to share an identical function is no strange phenomenon to be explained away by editorial confusion or ‘una certa misteriosa ossessione, che avrà certamente un senso che a noi sfugge della triplicità dei baroni’.²³ Irish tradition as reflected in the major tale cycles abounds with such instances; two immediate examples are the three identical foster brothers of Conaire in Togail Bruidne da Derga,²⁴ and the three sons of Uisliu in Longes mac n-Uislenn,²⁵ and there are a host of others, Irish and Welsh.²⁶ This preference for threes is also reflected in certain formal and stylistic arrangements, the Irish and Welsh Triads, for instance, and groups of stories such as the Trí Truaighe na Sgéalaigheachta²⁷—though in this case the grouping in three may be the work of a late, i.e. fifteenth or sixteenth century, author. Mrs. Bromwich links these literary conventions with the mythological systems of the early Celts,²⁸ and Georges Dumézil has devoted a life-time of study to this fundamental question.²⁹

The Romance field offers, in the Continental matière de Bretagne, an unmistakable echo of this predilection for tripling, as indeed for animal transformations, multiple personality patterns and so on. It would not then be unreasonable to imagine that Béroul's three barons may derive from a Celtic original, just as do the three brothers of Gauvain.³⁰

The names of the three are discussed by Bédier,³¹ Lot³² and Ewert.³³ Godoine and Guenelon are typical names of traitors, virtually symbolic in their context no doubt, and Ewert accepts Lot's conclusion that Denoalen may also have been a Breton traitor of note, though unfortunately there is no obvious candidate at hand.³⁴ The choice of three such names reminds us immediately of Marie de France's habit of using bilingual titles from time to time and may be said to be some indication of a multilingual milieu for which the poems were destined.

It is possible that Eilhart also knew of a version in which the three barons played a larger rôle than that implied in his vague references to groups of four or seven counts or dukes. The section which occurs at the end of Béroul's text in its extant form, which describes Tristan's revenge and the death of two of the barons, has been described by Ewert as standing ‘entirely apart from all other versions’. There are, however, some slight resemblances between this section and the much shorter section in Eilhart, lines 8931-9032. In this episode, which follows immediately upon the Folie Tristan episode, Tristan and Isolt are meeting secretly at night, after the king has agreed that the Fool shall remain at court. Some of Tristan's enemies, alerted by spies—cf. Béroul 4273—plot to surprise the lovers but are prevented by a show of force from Tristan. The difference in length, which might well have been even greater had Béroul's poem not been interrupted, is not important, as we know from the development of folktale research that ‘the length of a variant is a completely irrelevant aspect as to the structure’.³⁵ The number of the enemies here is also interesting, five in all, two spies who report to a group of three.

These resemblances are rather general but there are also clues in the context of both extracts. In Eilhart the discomfiture of the evil chamberlains follows, as we have seen, the episode of Tristan disguised as a fool at Mark's court, and provides the necessary release mechanism for Tristan to leave Isolt and return to Brittany. In Béroul, the episode in question follows upon a sequence of incidents, the Tryst at the Mal Pas, when the disguised Tristan answers questions in much the same way as he does in the versions of the Folie Tristan episode, the ambiguous oath and the vindication of Isolt, which ends with the establishment of a temporary equilibrium (4262-6). Ewert thinks, with Schoepperle, and apparently also with Mrs. Bromwich,³⁶ that the original poem, and probably too Béroul's version, ended after this with Tristan's death at the hands of King Mark, or one of his agents. In Eilhart, despite the fact that the location changes to Brittany, the new sequence which opens from the state of temporary equilibrium established by the preceding section involves the wounding and death of Tristan at the hands of another foe. On the other hand, there are in both texts indications that these sections may have belonged to an autonomous episodic unit of a type which we shall see is very familiar. Ewert has drawn attention to the fact that in Béroul's version this section contains several inconsistencies and obscurities, though the poet has attempted in places to link it with what went before. In Eilhart, the appropriate section is tacked on to the end of an earlier and complete section, the Folie Tristan episode. Logically it is superfluous except that it provides the occasion for the lovers to part. It is, however, by a more detailed examination of the narrative structure of this whole section that this becomes clearer. As this examination will provide the basis for subsequent discussion, a short digression is in order.

G. Schoepperle divided Eilhart's narrative into roughly equal sections for the purpose of summary, a division based partly on content and partly on the manuscript rubrics in Lichtenstein's edition. More recently C. A. Robson³⁷ has subjected these divisions to a closer examination and has demonstrated how each segment of narrative was expanded symmetrically from its original nucleus of oral narrative substance. Applying these methods to the final set of segments in Schoepperle's list, we see that the recurrent unit is a section of roughly 68-74 lines in the Lichtenstein text, each of which corresponds to a single movement of the tale in the chain which constitutes the narrative sequence. Thus stylistic and linguistic criteria, manuscript indications, sometimes metrical arrangement, confirm at the textual level what the mental perception of the reader or hearer apprehends as a discrete segment of content, and perhaps goes on to identify, through the existence of paradigmatic inventories like the Aarne-Thompson³⁸ and Thompson³⁹ indices, as part of a universal repertoire of story telling. Lines 6106-804 in Eilhart are a good instance of this. Section ‘p’, on the other hand, where this section in question occurs, is far more chaotic in its order. The opening segments seem to follow a pattern established previously, but the pattern changes with the presentation of the Folie Tristan episode, which Eilhart seems to have arranged in a rather different way,⁴⁰ and finally there is this abortive attempt to trap the lovers, which does not follow the preceding section very logically, as we have seen. This short section falls into three clearly defined parts, each of which is roughly half the ‘average’ length alluded to above. It looks very much as if they belong to a context from which they have been displaced, or if some kind of conflation has taken place. The conclusion then towards which this points is that the three barons, named or originally anonymous, may belong to a tradition known to Eilhart as much as to Béroul.

I have discussed the principal—named—enemies of Tristan, but there are also various minor characters or intermediaries of a hostile kind; the forester in Béroul, the leper band and their leader, as well as other characters with whom Tristan comes into conflict in other parts of the story, the cowardly seneschal of Ireland, the ancestral enemies of this homeland and the enemies of Kaherdin and his father.

These, then, are the principal enemies of Tristan. Despite the fact that most Tristan studies since the turn of the century have been more or less exclusively concerned with origins, scholars have been unable to say which of the enemies belong to which stages in the evolution of the legend, or whether individual choices throw any light on the earliest forms of the story. Schoepperle concluded that the different groups of enemies were introduced into the story by different redactors and remanieurs and that the poet of the estoire ‘made no effort to connect them with each other’.⁴¹ These and similar conclusions have been drawn from a study of each individual enemy taken separately. What happens when we consider them as a group, and in terms of their narrative function? One of the most profitable aspects of the modern structuralist approach to literature⁴² has been the renewed emphasis on what we might call the immanent aspect of literature. This concept of a literary work as a system of signs, a structured set of codes, which is clearly influenced by modern linguistics, is by no means revolutionary, but is in many ways a restatement of Aristotle and ‘the old methods of classical rhetoric, poetics or metrics’.⁴³ In the mediæval field, no doubt because of the numerous problems of transmission, oral and written, the different concept of authorship, the physical condition of texts and all the consequences of what Paul Zumthor in a challenging new book⁴⁴ calls ‘l'éloignement du moyen âge, la distance irrécupérable qui nous en sèpare’, this approach has not been over-favoured. None the less, it is possible to apply to the study of Arthurian texts some of the methods used by scholars who are not in the first instance mediævalists, because of the close connection between the mediæval Arthurian romances and Celtic narrative tradition and because of the enormous influence of oral tradition on both Celtic and ‘neo-Celtic’ literature.

Apart from the researches of Celticists and scientific folklorists, the most valuable development has been the productive effect of Propp's Morphology of the Folktale.⁴⁵ Commentaries, discussions, further contributions extending the range of interest from the original (Aarne-Thompson 300-749) fairy stories discussed by Propp to the whole field of narrative in general have been carried on by scholars like Bremond,⁴⁶ Greimas,⁴⁷ Pop,⁴⁸ Meletinski,⁴⁹ Dundes⁵⁰ and Köngäs-Maranda⁵¹ and the reviews Poétique, Poetica, Semiotica, Communications and Fabula, as well as the orthodox folklore journals, and occasional articles in literary or linguistic periodicals. Some of the terminology, which is still largely experimental, will be borrowed for the purposes of this discussion.

It is immediately obvious that the Tristan story, as represented in its primary versions, displays two patterns of overall narrative arrangement, one a cyclic, repetitive pattern, the other a kind of overarching pattern which approximates more closely to a complete Tale Type. Both patterns are visible in all primary versions despite the fragmentary nature of the texts. The second pattern coincides with what has long been felt by scholars to be a natural division of the story into three parts; the boyhood deeds of Tristan, the love story, and the exile, marriage and death. Mrs. Bromwich⁵² has shown how these divisions are the product of the different stages through which the story passed and the different sources from which it developed. The existence of these independent sections is confirmed by an attempt to analyse the whole story in terms of Propp's list of functions. The first section, birth and boyhood deeds, down to the vindication of the triumphant dragon-slayer, follows Propp's outline with almost no variation. This is no doubt because the type known internationally as the Dragon-Slayer is very well-defined and widely distributed.⁵³ The other two divisions are less easy to analyse along the same lines, but their independence is confirmed in the first case by the existence of a formal system of classification for Irish tales,⁵⁴ and in the last instance by the existence of parallel versions of the Man with Two Wives, which corresponds in part to a type or subtype in the Aarne-Thompson inventory.⁵⁵

The other pattern, with which we are more immediately concerned, appears to be at first one of narrative technique, since it reappears in all the versions and in the narration of the three major sections alike. It is seen most clearly in the episodic returns of Tristan to Cornwall after his exile in Brittany, but it is also illustrated in similar sequences in the middle section where the separation of the lovers is more psychological than physical, but where an identical procedure of ruse, counterruse and solution is employed. This pattern is closely linked with the arrangement of those recurrent units noted earlier in Eilhart. We notice if we analyse each example that an enemy or enemies appears in a more or less fixed place in each instance. The sequences in question, the Chips on the Stream, Flour between the Beds, and the various returns of Tristan, two in Thomas, four in Eilhart, show a basic succession of moves (and doubtless there are others, Thomas' Harp and Rote, Blades at the Bed (Eilhart), Life in the Forest, which follow a similar pattern but are excluded as the enemies do not figure significantly in them). These could be summarized as follows:

Tristan and Isolt are separated, either by physical or mental obstacles.
They attempt to reunite. An enemy plans to frustrate them.
Tristan by counter-trickery foils the adversary.
Isolt either delays the pursuer or allays the King's suspicions.
Tristan escapes or is restored to court, at any rate establishes a temporary advantage.

It is almost as if the poet were operating a simple formula. But is it the result of a personal predilection on the part of one of the many remanieurs through whose hands the story passed?

Some of Propp's interpreters, notably Bremond and Dundes, have pointed out that his remarks about the order and linking of functions⁵⁶ ought to be modified. Some of them, for example, as Propp himself was aware, seem to be in complementary distribution and others seem to be binary pairs, or again, within the succession of functions there appear groupings of two or three in a closer or more dependent relationship. Dundes has shown that North American Indian tales in particular (which are purely oral tales with a relatively simple structure) do not possess anything like all the thirty-one functions noted by Propp, but consist usually of the short sequence ‘Lack’ to ‘Lack liquidated’.⁵⁷ By contrast the structure of most European folktales is far more complicated. In short, though we cannot expect the schemes worked out by Propp on the basis of one particular corpus to apply exactly, nevertheless, as he himself foresaw, since the origins and development of many mediæval romans de chevalerie are closely linked to folktales we should expect vestiges of the same structure to appear.

This is true of the narrative segments we have isolated in the Tristan versions. The rough schema preceding can be further reduced and it will be seen that the pattern conforms fairly closely to the important group of functions which Propp calls basic to all forms of the fairy tale, and which Bremond has tried to apply to all types of narrative in general.

Separation:	Lack. (Propp VIII a)
Initiative:	Lack made known, Hero accepts action, Hero departs. (Propp IX-XI)
Counter ruse:	Hero under attack. (Propp XII)
Foiling of enemy:	Reaction of (hostile) Hero to donor, (XIII 8 or 9)
Establishment of advantage:	Lack Liquidated. (Propp XIX)

It is noticeable that 6 functions are missing from Propp's list. This is not particularly important, since Propp and his successors themselves pointed out that even in those fully developed tales which conform most closely to the abstract model not all the functions will always be present. What we have here is a much simpler model than the one drawn up by Propp, one which seems to be particularly suited to oral composition, delivery and transmission and which could only be incorporated in a longer, more complex and partly written version by repetition.⁵⁸ This is the simplest type of linking and though the individual poets of the extant versions, Thomas, and to a lesser extent Eilhart and Béroul, have introduced some degree of overall motivation, the repetitive nature of the pattern is still obvious.

It seems clear that the earliest forms of the story were in fact separate ones about Drust the Dragon-Slayer, Drystan, Essyllt and March, and perhaps versions of the Dragon-Slayer in a Breton setting.⁵⁹ Does it then follow that of the two types of narrative structure we have noticed, the ‘overarching’ or more complex type is the earlier or more basic form, and the simpler arrangement, the concatenation bout à bout of identical sequences differing only in content is a feature of the poetic technique of individual poets or redactors?

This seems highly unlikely, since we possess an obvious example of this latter type in Triad 26 of the Welsh Triads,⁶⁰ which Mrs. Bromwich has said ‘belongs to the oldest stratum of Arthurian tradition in Wales’. Bédier dismissed this triad as part of a vague and amorphous body of tradition which preceded the formation of the poem by a man of genius who was certainly a Frenchman. Even now, some scholars have attempted to minimize the importance of the Celtic contribution, not only just to the substance of the story but also to its shaping. On the contrary, the relation between Triad 26 and the extant versions is a most interesting one, since it suggests that Triad 26 provides the model for the later versions and confirms very strongly Mrs. Bromwich's belief that it was in S. Wales that the celebrated love story was first composed. This early Welsh Tristan story, in versions partly or wholly oral, was almost certainly not as lengthy as our present versions. But if this narrative sequence involving separation, attempts at reunion, frustration by an enemy, counter ruse of the hero and ultimate success represents the original nucleus of the story, then it is clear that the various enemies are a product of the process of repetition and expansion.

This implies, in turn, that there is somehow one basic enemy. To examine this implication, we must look at the second aspect of Propp's study of the folktale, the distribution of dramatis personae. Propp said that the tale could be defined in terms of two axes, its functions (or predicates) and its performers (or subjects) which in the case of the Russian fairy tale he found to be equally constant in number, seven to be exact. Rather less work has been done here by Propp's followers, and in this area the work of A. J. Greimas is probably unique. The mixed reception accorded to his work on semiology should not obscure the fact that his comments on Propp are more readily appreciated despite the use of jargon to which some of his English-speaking reviewers have taken exception.⁶¹ Valuable work has also been done by Köngäs and Maranda. Such studies have broken down Propp's distinction between functions and performers, and highlighted that which sometimes appears between rôles and actors. Basic to all forms of narrative is a hero or centre of interest; Köngäs and Maranda have labelled this the ‘first term’ and advocate the startlingly simple method of statistical assessment to answer the question who is the hero of a given story. The hero is simply the one mentioned the greatest number of times. This is not so naïve as it may seem: in other words, the term has meaning only at the level of discourse, his mode of existence is purely linguistic. Having established the hero, ‘most European folktales add an opponent … plus … a number of actors who bring into relief, or make possible, the actions of the hero (and the opponent) and who can become helpers or adversaries. But they are as a rule undefined, because the motivation of their actions rests with the hero's fate, not with their own selves. Thus we are inclined to see two “fronts” in a folktale and a host of actors who can become allies of one or the other.’⁶² Here we see the difference between actors and rôles discussed by Greimas and it is interesting to consider in this light the relationship which exists in the Tristan texts between the two groups of go-betweens, the helpers and the enemies.

Assuming this process, the addition of intermediaries attaching themselves to either of two ‘fronts’, to be visible in the diachronic process—the ‘history’ of the evolution of the legend—, the obvious figure on to whom all the other enemies can be mapped is King Mark himself. The sympathy and pathos of his relationship with Tristan in the French versions and in Gottfried is irrelevant here. In the various Irish analogues, disregarding for a moment the question of derivation, the husband is depicted in active pursuit of the hero. Indeed it is this very point which leads Carney⁶³ to maintain that the Irish analogues are based on a lost original British Tristan, and not vice versa; he contrasts the remorseless, treacherous Conchobor of Longes mac n-Uislenn with the paternal figure of the Ulster cycle, and points out that Finn too, in the Tóraigheacht Dhiarmada agus Ghráinne, is at variance with his usual portrayal. The second Isolt, according to Carney, did not belong to the early British Tristan, since the third party who in most of the Irish stories misinterprets the sign and thus brings about the tragic ending is a man. ‘This points also to the part played by Isolde, the wife, having been played by Mark in the primitive story.’⁶⁴ Had Carney been using the terminology of Greimas, he might have said that the ‘acteur’ Isolt of the Whitehands was here filling the slot, or rôle, of the ‘actant’ enemy, originally lexicalized by Mark! (It should be noted, however, that the structural study of narrative form practised by Greimas and others is resolutely synchronic and avoids identifying the notion of hierarchy in the grouping of elements with any chronological stage in the development of a form. This would be one of the main differences between Dundes and the so-called ‘Finnish’ or ‘historicogeographical’ school of folklorists, of which Stith Thompson is the chief American representative.)

At the crucial Welsh stage in the formation of the legend, then, there may well have been many current versions of a simple story involving March, Essyllt and Drystan of which Triad 26 is a reminder and an example. In a manner closely reminiscent of a modern cultural equivalent, the radio, T.V. or comic strip serial, the story tellers who wished to extend the story or revive an old favourite for frequent recitation had recourse to one of the commonest of narrative procedures, repetition, doubling and tripling. The mediæval Continental poets who adapted the Welsh or Breton material may have been responsible for incorporating these concatenated sequences into the more complex structure of the biographical outline as we know it. It is at this point, one is inclined to think, that the ‘lexicalization’ of the rôle of enemy took place. Although some of the individual enemies show, as we saw earlier, characteristics which are consistent with a Celtic milieu, or at least one in which Irish and Welsh stories and traditions were not just a fossilized deposit—a rival nephew closely connected with the Queen, three hostile figures sharing virtually one identical function—the growth of a tradition concerning the enemies of Tristan is much more likely to have been a consequence of the evolution of the rôle of King Mark. At the level of structure, as the quotation from Köngäs and Maranda points out, the process of development from simple oral narrative forms in which the hero is balanced by an enemy⁶⁵ gives rise in more complex versions of the tale to the use of intermediaries who act out in a concrete way the hostile relationship between the two principals. At the level of content, the situation in which a monarch is portrayed as dependent upon the counsels of his advisers and subject to a detailed system of rules of procedure is somewhat more reminiscent of the literary presentation of feudalism familiar to us from the chansons de geste as from the romances. As Idris Foster points out,⁶⁶ the Arthur of Culhwch is far closer to the shadowy but powerful ‘dux bellorum’ hinted at in the few precious references in early Welsh tradition than to the ‘roi fainéant’ of the romances.

It is not possible to say at what precise stage in the progress of the legend these changes took place. Frappier⁶⁷ has argued with a good deal of sense that Bédier's theory of a fully developed written narrative immediately preceding our present versions must be modified, and it is likely that most Tristan scholars today would agree that such a rigid distinction should not be drawn between oral and written forms. But we are probably entitled to assume that the original enemy of Tristan was Mark himself, and that the earliest versions of the story of their hostility were Welsh and consisted already of concatenated sequences based upon the model given in Triad 26, in which first Mark, then Tristan gained the advantage. It is not unlikely, though purely conjectural, that some intermediaries appeared in the versions at that stage as indeed Triad 26 itself suggests.⁶⁸ As the story was used by different conteurs with different skills and in different cultural contexts the enemy was ‘lexicalized’ in different ways, but his basic function, his place in the narrative sequence, remained unchanged. Having introduced them or inherited them, however, the poets had to get rid of them and it is here that the uniform structure breaks down, with the clumsy attempts at harmonization by Béroul, the sudden fade-out of Andret in Eilhart and so on.

These observations are made on the basis of a study of the narrative structure of the existing versions and a comparison between them. The methods used by Propp and so on coincide here in part with some of those favoured by modern structuralist critics like Barthes and Todorov and, in this field, Zumthor. In so far as such methods fall within the sphere of well-established disciplines like stylistics and poetics, they are hardly revolutionary; their value lies in the attempt to broaden the base of such studies to include not merely the textual or surface aspects of a narrative—its ‘style’ as one might have said—but also what Köngäs and Maranda have called, using an analogy from contemporary linguistics, its ‘deep structure’. From such studies is emerging an interest in the characteristic shape and properties of narrative in general, as a legitimately autonomous area of literary discourse.⁶⁹

Notes

1. Helaine Newstead ‘Recent Perspectives on Arthurian Literature’ Mélanges Frappier (Paris 1971) II, 877-83.

2. A. Ewert ed. The Romance of Tristran by Béroul (Oxford 1970) II, 120.

3. Unfortunately the context is ambiguous. The line itself is an example of the transitional, almost formulaic lines with which the poet proceeds from one narrative movement to the next, and so could refer either to the preceding section, as a kind of summary statement or to the following one. On the other hand, the words nevo, nies are used frequently throughout the entire poem, often in direct exchanges between Tristan and Mark. It is therefore probable that Ewert is right and that this reference is to Tristan.

4. G. Raynaud de Lage ‘Faut-il attribuer à Béroul tout le Tristan?’ Le Moyen Age LXIV (1958) 249-70 and LXVII (1961) 167-8.

5. Rita Lejeune ‘Les “influences contemporaines” dans les romans français de Tristan au XIIe. siècle’ Le Moyen Age LXVI (1960) 143-62.

6. T. B. W. Reid ‘The Tristan of Béroul: One Author or Two?’ Modern Language Review LX (1965) 352-8. See also A. Vàrvaro Il Roman de Tristan di Béroul (Napoli 1966) pp. 9-22 and Ewert op. cit. pp. 1-3.

7. Ewert op. cit. p. 218.

8. See R. Bezzola ‘Les neveux’ in Mélanges Frappier II 88-111. Bezzola's article concludes with a long list of nephews found in both chansons de geste and romances. He is of course seriously mistaken when he refers to ‘les mythologies germaniques et celtiques, òu, à notre savoir, les neveux ne jouent jamais un rôle de premier plan’. One has only to cite a few examples: Cú Chulainn nephew of Conchobor (and referred to in the Táin as ‘the king's sister's son’—mac sethar ind ríg), Diarmaid nephew of Finn, Gwern nephew of Bran (Branwen verch Lyr), Gilfaethwy and Gwydion nephews of Math (Math), Lleu Llaw Gyffes nephew of Gwydion (Math), and of course we learn at the start of Culhwch that Arthur and Culhwch were fellow-nephews, evidently a relationship of some importance. See Rachel Bromwich Trioedd Ynys Prydein (Cardiff 1961) p. 371 and T. M. Charles-Edwards ‘Some Celtic Kinship Terms’ Bulletin of the Board of Celtic Studies XXIV (1971) 105-21.

9. Helaine Newstead ‘The Harp and the Rote, an episode in the Tristan legend and its Literary History’ Romance Philology XXII (1969) 463-70.

10. M. Delbouille ‘Le nom du nain Frocin(e)’ Mélanges Istvan Frank (Univ. des Saarlandes 1957) pp. 191-203.

11. J. Bédier Le roman de Tristan de Thomas (Paris 1902-5) II 250; Ewert op. cit. pp. 160-2.

12. Bédier op. cit. pp. 245, 247.

13. J. Frappier ‘Sens et structure du Tristan; version commune, version courtoise’ Cahiers de Civilisation Médiévale VI (1963) 255-80 and 441-54.

14. R. S. Loomis Arthurian Tradition and Chrétien de Troyes (New York 1949); A. Olrik ‘Epic Laws of Folk Narrative’ in A. Dundes ed. The Study of Folklore (Englewood Cliffs 1965) pp. 131-41 (originally appeared in Zeitschrift für deutsches Altertum LI (1909) 1-12). See also K. Krohn Folklore Methodology (Texas 1971) pp. 78-98, and V. J. Propp ‘Transformations des contes merveilleux’ first published 1928 translated in Morphologie du Conte (Paris 1970).

15. The name appears in Geoffrey of Monmouth, as Conanus Meriadocus. See also M. Roques ed. Erec (Paris 1955) lines 2076ff. and J. Rychner ed. Lais de Marie de France (Paris 1966) Guigemar (index). See also R. S. Loomis ed. Arthurian Literature in the Middle Ages (Oxford 1959) pp. 379ff., 472ff.

16. B. H. Wind ed. Les Fragments du Roman de Tristan (Paris-Geneva 1960) Fr. Sneyd 1, line 796.

17. Moshe Lazar Amour courtois et ‘fin amors’ dans la littérature du XIIe. siècle (Paris 1963).

18. Ewert op. cit. p. 119.

19. Ibid. lines 1656-1750.

20. Supra n. 18.

21. Varvaro op. cit. p. 150.

22. Bédier op. cit. II 175ff.

23. Vàrvaro op. cit. p. 150.

24. E. Knott ed. Togail Bruidne da Derga (Dublin 1963).

25. Vernam Hull ed. Longes mac n-Uislenn (New York 1949).

26. Three sons of Tuirill Bicred or of Nechta Scéne, and the various groups of triple deities like the Trí Dé Donand (Book of Leinster ed. Bergin and Best (Dublin 1954) I 124). See further Bromwich op. cit. p. 155.

27. Myles Dillon Early Irish Literature (Chicago 1949) p. 62.

28. Op. cit. p. lxiiiff.

29. For bibliographical information see C. Scott Littleton The New Comparative Mythology (Berkeley and L.A. 1966).

30. Loomis Arthurian Tradition p. 487.

31. Bédier op. cit. II 125.

32. Romania XXXV (1906) 605-7.

33. Op. cit. pp. 222-3.

34. The name itself is well-attested. The Cartulaire de Redon ed. Courson gives several instances of variations on the form Donoalus, Donuuallonus etc. (pp. 74, 86, 261, 243, 299, 129, 333). Later instances from the Cartulaire de Quimperlé see J. Loth Chrestomathie Bretonne p. 202. For the probable etymology see Ellis Evans Gaulish Personal Names (Oxford 1967) pp. 196-7 and Fleuriot Grammaire du vieux breton pp. 41, 1135.

35. E. Köngäs and P. Maranda Structural Models in Folklore and Transformational Essays (Hague 1971) p. 73.

36. In her review of Binchy's edition of Scéla Cano Meic Gartnáin (Dublin 1963), in Studia Celtica I (1966) 152-5.

37. C. A. Robson ‘The Technique of symmetrical composition in medieval narrative poetry’ Ewert Miscellany (Oxford 1961) pp. 26-75.

38. A. Aarne and Stith Thompson The Types of the Folktale (Helsinki 1961).

39. Stith Thompson Motif Index of Folk Literature (Copenhagen 1955-8).

40. The section in question is divided clearly into three segments, clearly indicated by the switch of attention from one actor to another and by formulaic lines—8787, 8845 and 8887. These segments are slightly shorter than the average narrative unit of 68-74 lines. A similar treatment involving an initial, pivotal and conclusive section is found in Eilhart lines 5700-6105, a description of a battle. This is slightly more complicated though as we might expect from Eilhart's taste for battle scenes, each of the three main stages is divided into two subsections.

41. Gertrude Schoepperle Tristan and Isolt (Frankfurt 1913) I 252.

42. The literature on Structuralism is unfortunately increasing copiously and in a manner which, due to the diffuse nature of the subject, makes bibliographical reference difficult. See M. Lane ed. Structuralism, a Reader (London 1970) and T. Todorov et al. eds. Qúest-ce que le structuralisme (Paris 1968).

43. R. Wellek and A. Warren Theory of Literature (London 1963) p. 139.

44. P. Zumthor Essai de poétique médiévale (Paris 1972).

45. V. J. Propp Morfologija skazki (Leningrad 1928) translated as the Morphology of the Folktale (Indiana 1958) and new translation (Austin 1968), and Morfologia della Fiaba (Torino 1966) second Russian edition 1969 translated into French 1971, see n. 14. Of the two English editions the second is preferable, since it contains an excellent introduction by Alan Dundes. The Italian version adds a Comment by Lévi-Strauss, also available in the International Journal of Slavic Poetics and Linguistics III (1960) 122-49, plus Propp's reply which is not found there. All references are to the French version unless otherwise stated.

46. C. Bremond ‘Le message narratif’ Communications IV (1964) 4-32; ‘La logique des possibles narratifs’ ibid. VIII (1966) 60-76; ‘Postérité américaine de Propp’ ibid. XI (1968) 148-56; ‘Morphology of the French Folktale’ Semiotica II (1970) 247-76.

47. A. J. Greimas Sémantique structurale (Paris 1966) pp. 172ff. ‘La structure des actants du récit’ Word XXXIII (1967) 221-38; Du Sens (Paris 1970) pp. 185-230; ‘Narrative Grammar; Units and Levels’ Modern Language Notes LXXVI (1971) 793-806.

48. M. Pop ‘Aspects actuels des recherches sur la structure des contes’ Fabula IX (1967) 70-7; ‘La poétique du conte populaire’ Semiotica II (1970) 117-27.

49. E. M. Meletinski ‘Problèmes de la morphologie historique du conte populaire’ Semiotica II (1970) 128-34 and see the extremely useful synthesis ‘L'étude structurale et typologique du conte’ appended to the French translation of Propp, pp. 201-54.

50. A. Dundes ‘From Etic to Emic Units in the Structural Study of Folktales’ Journal of American Folklore LXXV (1962) 95-105.

51. See note 35.

52. R. Bromwich ‘Some remarks on the Celtic Sources of Tristan’ Transactions of the Hon. Society of Cymmrodorion (1953) pp. 32-60.

53. Aarne-Thompson op. cit. pp. 88ff.

54. R. Thurneysen Die Irische Helden- und Königsage bis zum 17ten Jahrhundert (Halle 1921) p. 21ff.

55. Aarne-Thompson op. cit. pp. 128ff.

56. There is a problem of terminology here. Propp's ‘function’ is defined thus: ‘function must be taken as an act of dramatis personae which is defined from the point of view of its significance for the course of action of a tale as a whole’ (first English ed. p. 20). The confusion is compounded by the fact that, for instance, Greimas' ‘syntagme narratif’, Propp's ‘sequence’, Dundes' ‘motifeme’, and others such as Dorfman's ‘narreme’ (The narreme in the Medieval Romance Epic (Toronto & Manchester 1969)) do not always correspond exactly. See Köngäs-Maranda op. cit. p. 21.

57. A. Dundes The Morphology of North American Indian Folktales (Helsinki 1964).

58. Propp op. cit. 112, attempts to distinguish between a ‘Conte qui peut se composer de plusieurs séquences’ and a text which contains more than one conte juxtaposed. By any criteria advanced by Propp what we have here is not an example of the latter case.

59. A. Witte ‘Der Aufbau der ältesten Tristandichtungen’ Zeitschrift für deutsches Altertum LXX (1933) 161-95 and R. Bromwich art. cit. p. 59.

60. Trioedd Ynys Prydein p. 45.

61. See for instance Ullmann in Lingua XVIII (1967) 296-303.

62. Op. cit. p. 33. The intrusion of Isolt does not really prove an exception here as she can legitimately be regarded as Tristan's alter ego, see on this J. Campbell The Hero with a Thousand Faces (New York 1949) p. 342.

63. J. Carney Studies in Irish Literature and History (Dublin 1955) vi ‘The Irish Affinities of Tristan’.

64. Ibid. p. 204.

65. Cf. Olrik op. cit. p. 135, the Law of two to a scene, and p. 139 Concentration on a leading character. As Dundes points out in his editorial note, p. 135, it is interesting to see how some of Lévi-Strauss's structuralist ideas have been anticipated here. In fact this little-quoted essay by an eminent Danish folklorist of the past merits more attention, since it anticipates in non-technical language many of the findings of contemporary structuralists and folklorists.

66. In Arthurian Literature in the Middle Ages pp. 31-9 at p. 33.

67. See n. 13.

68. The other two important pieces of evidence in Welsh, the late but almost certainly independent Ystoria Trystan and the problematic fragments in the Black Book of Carmarthen, do not in any way alter this. The former involves the polarization of Tristan and Mark, and the presence of intermediaries, and appears to repeat the familiar sequence of threat, counter-ruse and so on; the fragments would appear at very least to hint at enmity between the two principals, the presence of intermediaries and possibly a specific reference to one of the early incidents in the story. See R. Bromwich art. cit. p. 58.

69. See above n. 47 for the number of MLN devoted to Comparative Literature.