SOURCE: Blakeslee, Merritt R. “The Authorship of Thomas's Tristan.” Philological Quarterly 64, no. 4 (fall 1985): 555-72.

[In the following essay, Blakeslee argues that the imitation of Wace's Roman de Brut evident throughout the twelfth-century Tristan attributed to Thomas supports the hypothesis of Thomas's sole authorship of that poem.]

The question of dual authorship, so long a subject of controversy among scholars of Beroul's Tristran, has latterly been raised in regard to Thomas's Tristan. A recent article by Constance B. Bouchard has suggested that the manuscript fragments traditionally ascribed to Thomas of Britain and assumed to have been composed in the second half of the twelfth century are instead the work of an early thirteenth-century author who is not the Thomas or Tumas referred to in those fragments.¹ The study hypothesizes that the original Tristan, composed by Thomas (or perhaps by a pseudo-Thomas), was left unfinished and that a continuer was responsible for the composition of the extant fragments of the romance (excepting the Cambridge fragment, which is probably the work of Thomas or pseudo-Thomas). This provocative article shows a refreshing tendency to challenge many of the traditional but unprovable assumptions on which modern Tristan scholarship rests, and it is instructive to measure our received ideas about Thomas's romance against its iconoclastic thesis. Ultimately, however, I believe that its conclusions are ill-founded and must, however stimulating, be rejected. Thus, the present study will argue that there exists no compelling reason to set aside the theory that the Thomas whose name appears in the fragments ascribed to him was in fact their author. It will contend further that there exist persuasive reasons for concluding that the entire Roman de Tristan attributed to Thomas is the work of a single poet of that name.

Before detailing Bouchard's theory, it is well to rehearse the traditional thesis advanced to account for the known facts concerning the composition of the fragments attributed to Thomas. A complete version of the Tristan narrative is presumed to have been composed in Anglo-Norman sometime in the second half of the twelfth century by a certain Thomas. This individual is assumed to have been a clerk and is usually referred to as Thomas of Britain, the appellation assigned him by Gottfried von Strassburg.² A terminus a quo of 1155 for Thomas's poem is furnished by Wace's Roman de Brut, from which it borrows. A terminus ad quem of ca. 1210 is furnished by Gottfried's Tristan und Isolde, an adaptation of the first half of Thomas's poem. The latter survives in nine fragments conserved in five manuscripts dating from the twelfth and thirteenth centuries.³ The Douce and Sneyd² fragments mention by name the poet, whom they designate in the third person as Thomas (D 862) or Tumas (Sn²820). In the early years of the thirteenth century, Gottfried von Strassburg composed a version of the Tristan legend, basing his narrative on that of Thomas, whom he cites as his source.⁴ He embellished his model profusely, creating a poetic masterpiece that breaks off at the point at which Tristan arrives in Brittany after his exile from Cornwall. Gottfried's continuers, Ulrich von Türheim and Heinrich von Freiburg, who both refer to Gottfried's untimely death, based their continuations on the narrative given by the Tristrant of Eilhart von Oberge, perhaps because, for material or linguistic reasons, Thomas's poem was inaccessible to them. In 1226 a certain Brother Robert produced an Old Norse adaptation of the entire narrative of Thomas's poem, the Tristrams Saga ok Isöndar [S]. Robert, whose text does not mention Thomas by name, followed faithfully the narrative given by his model but condensed radically the non-narrative material found in his source, notably monologue and passages of courtly casuistry. The Middle English metrical romance Sir Tristrem, an early fourteenth-century adaptation of Thomas's poem, cites four times as the authority for its narrative a certain Tomas, said to reside in Erceldoune and identified by the early editors of Sir Tristrem with the historical Thomas of Erceldoune, a Scottish poet of the late thirteenth century.⁵

Bouchard proposes a radically different set of assumptions to explain the composition of the fragments ascribed to Thomas. According to her thesis, a poet who was himself Thomas (71), who claimed to be Thomas (71), or who claimed to repeat the true version of Thomas (70), composed a French version of the Tristan matter [T1] “in the later twelfth century” (67). His version, which was “simple and direct” (70), recounted the events of the legend as far as Tristan's departure from Cornwall for Brittany. T1 was left uncompleted at this point, possibly because the author considered his story essentially complete (70) or possibly because he had “written himself into a corner” (71). T1, no part of which (save perhaps the Cambridge fragment) is extant, was the only model available to Gottfried, who adapted and embellished it, abandoning his poem at the point at which his source broke off. Ulrich and Heinrich necessarily used Eilhart's poem as the model for their continuations, for no other source for their works existed. “[A]t the very beginning of the thirteenth century” (71), a second author, “imbued with thirteenth-century courtly concepts” (71), composed for T1 a conclusion in Anglo-Norman [T2]. This second, anonymous author identified Thomas, who, he said, “told the story correctly” (68), as his source. He “wrote in a much more convoluted style” (71) than his predecessor and introduced at least one factual error into his continuation. In the scene in the salle aux images, he identified Brengvein as the person who gave the lovers the potion. However, in T1 (as conserved in the Saga), that person is said to have been a servant lad. The Folie Tristan d'Oxford [Fo], which begins at the point at which T1 breaks off and which follows immediately the text of the Douce fragment in Bodleian Library MS. Douce d.6, was composed, perhaps by the T1 poet himself, as an independent conclusion to T1 (68-69). The Tristrams Saga was composed by an adaptor who had at his disposal both T1, which he followed “slavishly,” and T2, which he abridged drastically. The Sir Tristrem poet's identification of Thomas of Britain with the late thirteenth-century Thomas of Erceldoune casts further doubt on the historicity of the former.

The hypothesis advanced by Bouchard rests on six points that represent what she terms “anomalies and curious coincidences” (68) and for which she offers explanations at variance with the received wisdom of traditional Tristan scholarship:

• (1) The disproportion in style and length between the first and second halves of Thomas's poem, as established through a comparison of the extant fragments with the Saga and Gottfried's poem.
• (2) The identification of the Folie d'Oxford as a conscious attempt to provide an alternate conclusion to T1.
• (3) The reliance of Ulrich and Heinrich on Eilhart's poem rather than on Thomas's, “though one would have expected them to use the same source that Gottfried had used” (68).
• (4) The inconsistency between T1 and T2 concerning the cupbearer.
• (5) The contention that the references to Thomas in the Douce and Sneyd² fragments do not identify him as the author of those texts but rather, like the allusions in Gottfried's poem and in Sir Tristrem, represent the conventional evocation of a possibly apocryphal auctor.
• (6) The “remarkable coincidence” (68) that Gottfried's poem breaks off precisely at the point at which T2 begins.

Each of these points will be addressed in turn.

A. THE DISPROPORTION BETWEEN T1 AND T2

It is a well-established principle, and one with which Bouchard concurs (69), that while the Saga poet reproduced accurately the narrative detail that he retained from his model, he abridged heavily those portions of Thomas's poem containing authorial commentary.⁶ Thus, one is not surprised to discover that the Saga translates virtually verbatim the text of the Cambridge fragment, which contains only narrative material, but condenses severely the text of the other fragments, which contain lengthy authorial interventions. For this very reason, a comparison of the treatment by the Saga of the Cambridge fragment on one hand and of the other fragments on the other is unreliable in demonstrating differences in style and length between the two parts of Thomas's poem. Bouchard's conclusions concerning these differences, which she offers in support of her thesis of the dual authorship of Thomas's poem, are tautological; for they repose on the a priori assumption that the Cambridge fragment belongs to T1 and the other fragments to T2.

However, a comparison of Thomas's poem with the other primary versions of the Tristan matter (Eilhart, Beroul, and the French prose Tristan) reveals an important difference between the first and second halves of Thomas's poem. In the first half of his work (as preserved in the Saga), Thomas follows closely the traditional narrative as told by Eilhart and the French prose Tristan, although it is impossible to know how much authorial commentary he may have added. In the second part of his poem, he diverges from the narrative of Eilhart and the French prose Tristan, adding to his text not only extensive authorial commentary but much narrative material not found in the other primary versions. In an earlier article, I proposed a different explanation for the disparate character of the first and second parts of Thomas's romance.⁷ I noted that every complete primary version of the Tristan narrative (i.e., Eilhart, the reconstructed Thomas, and the French prose Tristan) contains four major parts: (1) the account of the birth and childhood of the hero (N¹), (2) the account of the hero's deeds until the marriage of Iseut and Mark (N²), (3) the account of the lovers' adulterous relationship before Tristan's definitive banishment from Mark's court (N³), and (4) the account of the lovers' relationship from the moment of Tristan's exile until their death (N⁴). These four parts are framed by a genealogical forematter and by an account of a miracle of the intertwining plants.⁸ However, while the structure of the narrative is symmetrical, its parts are not proportional in length.⁹ The first portion of the romance (N¹ and N²) incorporates three traditional narrative patterns that are widely distributed in world literature: (1) that portion of the heroic biography detailing the birth and childhood of the hero, (2) the expulsion and return formula, and, most notably, (3) the bride-quest motif. Significantly, in the primary versions of the Tristan matter, these portions of the legend demonstrate the greatest narrative stability. However, the non-canonical conclusion to the bride-quest narrative (i.e., the non-marriage of hero and heroine) serves to initiate the Mark-Iseut-Tristan love triangle and to set in motion the events of the second, non-canonical portion of the romance (N³ and N⁴).

In this second portion of the romance, the narratives given by each of the primary versions of the legend (including Beroul's poem, which gives a divergent narrative of N³) differ markedly. As Thomas, Beroul, Eilhart, and Gottfried all remind us, competing versions of the legend were in circulation.¹⁰ In composing his version of the second portion of the romance, each poet utilized to his own ends the narrative freedom afforded him by this rich and diverse tradition and by the non-canonical character of the second half of the Tristan matter. As Tony Hunt has recently shown, the author of the Thomas fragments is obsessed with the problem of Tristan's marriage to Iseut aux Blanches Mains and with the uncourtly dilemma posed by that transgression.¹¹ Thus, his version of N⁴, which is filled with authorial commentary on this question, is developed disproportionately, using all of the resources of clerkly rhetoric.

In conclusion, there exist fundamental methodological difficulties in establishing the existence of differences in style and length between the first and second halves of Thomas's work, whether one divides the romance between N³ and N⁴, as Bouchard proposes, or between N² and N³, as I have suggested. On the other hand, on the basis of a comparison of the primary versions of the Tristan legend, it is possible to draw certain conclusions concerning the differing character of the narratives of N¹-N² and N³-N⁴ and the differing treatment of these two parts by the various poets. However, in the case of Thomas's poem, the latter differences need not be explained by the postulation of dual authorship but rather by the heterogeneous character of the narrative material present in Thomas's model or models.

B. THE FOLIE D'OXFORD AS A CONCLUSION TO T1

The juxtaposition of Thomas's poem and the Folie d'Oxford in the Douce manuscript is open to another construction than that placed upon it by Bouchard, who would see in the Folie a conscious attempt, perhaps by the very poet who composed T2, to provide an alternate conclusion to T1. First, it must be asserted that the Folie d'Oxford provides no conclusion at all to the Tristan narrative, for every full primary and secondary version of the legend composed in the twelfth and thirteenth centuries concludes with the death of the lovers.¹² Second, not only is the Folie d'Oxford episodic but it is composed within a tradition of episodic Tristan narratives recounting the clandestine returns of the hero to Cornwall to tryst with Iseut. The existence of a large corpus of return episodes, some independent (e.g., Marie's “Chevrefoil,” the Folie Tristan de Berne, and Tristan als Mönch), some inserted in non-Tristanian texts (e.g., “Tristan rossignol” in the Donnei des amanz and “Tristan ménestrel” in Gerbert de Montreuil's Continuation de Perceval), some incorporated in full versions of the Tristan narrative (e.g., those recounted in Eilhart, Thomas, Beroul, and the French prose Tristan), testifies to the existence of an established tradition of composing examples of the sub-genre that is the return episode. That this is a very old tradition in the Tristan matter is demonstrated by the existence of a Welsh return episode entitled “Three Mighty Swineherds” that is perhaps as old as the eleventh century.¹³

In D 835-84, the poet specifically disputes the authenticity and versimilitude of those versions of the legend that recount Kaherdin's adulterous love for the wife of a dwarf, Tristan's wounding by the cuckolded husband, and the protagonist's return to Cornwall in the guise of a madman, a development conserved in Eilhart's poem and the MS. B.N.fr. 103 version of the French prose Tristan.¹⁴ In replacing this version with his own (which he perhaps found in his source, Breri), Thomas rendered the sequence leading up to the tragic conclusion of the romance “more moving, more credible, and better arranged and more concentrated than in Eilhart or the prose Tristan.”¹⁵ Yet the metaphor of Tristan's madness is one of the most resonant and disturbing in the legend and, hence, a subject that lent itself to independent episodic composition. Given the existence of a tradition of return episodes, the compelling character of the narrative of Tristan's madness, the demonstrable familiarity of the poet of the Folie d'Oxford with the narrative of Thomas's poem, and the omission by Thomas of a version of Tristan's madness, it would have been natural for a poet to compose and independent version of that episode, especially if, as seems probable, he modeled his poem on the existing Folie de Berne.¹⁶ It is more plausible to argue that, in keeping with the practice of medieval scribes of collecting thematically related texts in a single manuscript, the Folie d'Oxford was grouped with Thomas's poem in the Douce manuscript because it was understood as a remaniement of the return episode motif found three times in Thomas's poem rather than because it was understood to furnish a conclusion to T1.

C. ULRICH, HEINRICH AND EILHART

The reliance of the versions of Ulrich and Heinrich on the narrative of Eilhart's poem rather than on that of Thomas's poem has traditionally been explained by the assumption of material or linguistic obstacles to their utilization of Thomas's poem and need not be taken as proof that T2 had not yet been composed. There is no reason to suppose that Ulrich or Heinrich understood French or that either had at his disposal a manuscript of Thomas's poem. It is more probable that the choice of each was determined by the fact that Eilhart's poem was the only model readily available in their language.¹⁷ Both continuers lament the untimely death of Gottfried, and while Heinrich is probably imitating Ulrich in the passage, the latter, writing in 1230-35, may have had first-hand knowledge of the circumstances surrounding the composition of his poem.¹⁸ Although their remarks are not conclusive, they suggest that Gottfried's death interrupted the composition of Gottfried's poem. If this is indeed the case, as seems likely, then there is no reason to postulate an unfinished model to account for the fact the Gottfried's poem remained uncompleted.

However, there exists a more serious objection to Bouchard's hypothesis. She proposes that the compostion of T2 took place “at the very beginning of the thirteenth century” (71). In any event, the composition of T2 can not have occured later than 1226, the date of the redaction of the Saga. Hence, the composition of T2 pre-dates the continuation of Ulrich, composed between 1230 and 1235, and a fortiori that of Heinrich, composed in the last two decades of the thirteenth century.

D. THE CUPBEARER IN T1 AND T2

Even were one to concede the existence of an inconsistency on the question of the cupbearer, such a lapsus would hardly be determinative in establishing the dual authorship of the romance. Narrative inconsistencies are not infrequent in medieval texts, especially in longer ones. In the present case, the weight of this evidence would be further diminished by the impossibility of ascribing the error with certainty to the author of T2 rather than to Brother Robert. Moreover, the iconography of the statues is emblematic rather than realistic. Hence, they function only imperfectly and intermittently as guides to the narrative content of Thomas's poem. The narrative that the statues depict contradicts details recounted in T2. For example, the statue of Brengvain, described first as “small” (Schach, 121), has become life-sized in the scene in which Kaherdin takes it for a living person (Kölbing, Ch. 86). The detail of Iseut trampling the dwarf and that of the lion attacking Mariadoc correspond to no episode in T1 or T2. Each is instead, as Lucie Polak has remarked, “either a piece of wishful thinking on the part of Tristan the image maker, or, possibly, a magical figure destined to bring about what it portrays.”¹⁹

However, a close examination of the texts reveals that the inconsistency concerning the cupbearer is only apparent. In the first part of the romance (as given by the Saga and, presumably, by Thomas's poem), the scene in which Queen Iseut, Iseut's mother, confides the potion to Brengvain is described at length, while the page's action is accorded only passing mention (Kölbing, Ch. 46). In the second part of the romance, the figure of Brengvain, a small statue forming part of a tableau grouped around the larger, central image of Iseut, is represented offering to her mistress a cup containing the fatal potion:

On the other side of the dwarf stood a small statue in the likeness of Bringvet, the queen's attendant. The figure was well and beautifully fashioned and clad in the most lovely attire. In her hand she held a vessel with a lid on it, which she offered to Queen Isönd with a gentle smile. Around the vessel were the words that she spoke: “Queen Isönd, take this drink which was prepared in Ireland for King Markis.”

(Schach, 121)

In assuming that the statue of Brengvain represents the meschine in the act of offering the potion to the lovers during the voyage from Ireland to Cornwall, Bouchard falls into an error of which Joseph Bédier before her was also guilty. Noting that the Folie d'Oxford confirms the lesson of the Saga according to which a servant lad (un valet, Fo 647, ed. Payen) brought the potion to Tristan and Iseut, Bédier corrected the inscription on the cup (as given in the Saga) to reflect his belief that the cup was offered to Tristan and Iseut, arguing: “Ce propos est absurde: il est contraire, non seulement à toutes les formes conservées de la legende, mais à toutes ses formes imaginables” (Bédier 1:144, n. 4).

In fact, the statue of Brengvain depicts the offering of the potion not to Iseut and Tristan on board ship but rather to Iseut and Mark on their wedding night. Before her daughter's departure from Cornwall, Queen Iseut prepared a love philtre that she confided to Brengvain with the following instructions:

“Bringvet, guard this flasket well. You shall accompany my daughter. And on the first night when she and the king lie together and he demands wine, give this potion to both of them together.”

(Schach, 71)

On Iseut's wedding night, Brengvain took the place of her mistress in the nuptial bed with Mark:

When the king had fallen asleep, Bringvet went away and the queen lay down beside the king. And when he awoke, he demanded wine to drink, and Bringvet [var. MS. Reykjavík, National Library, IB 51 folio: the queen] cleverly gave him some of the wine that the queen had mixed in Ireland. But the queen did not drink any that time.

(Schach, 72)

Brengvain's words, which echo those spoken to her by Iseut's mother, are addressed to Iseut in Mark's presence. Thus, the inconsistency concerning the cupbearer is apparent only and of no validity in establishing the dual authorship of Thomas's poem.

E. THE EXISTENCE OF THOMAS AND THE COMPOSITION OF THE ROMAN DE TRISTAN

The question of the historicity of Thomas of Britain has been debated at length,²⁰ and it is not my task to join this debate. As Bartina Wind has observed of him:

Nous n'avons guère de données sur sa personne: son nom, deux fois mentionné dans les fragments, c'est tout ce qu'il nous a laissé et une oeuvre bien mutilée. L'oeuvre entière … nous donne au moins une impression de l'homme, mais pour sa vie nous ne pouvons en tirer que des conclusions trop vagues.²¹

Rather, I propose to show that there are no cogent reasons for placing in doubt the existence of Thomas nor for doubting that a single poet of that name was the author of the entire poem.

Brother Robert's failure to mention the author of his source is inconclusive in proving that Thomas or pseudo-Thomas was the author of one but not the other part of the romance. The fact that a century later the author of Sir Tristrem assimilated the Thomas of the Tristan fragments to Thomas of Erceldoune is hardly proof of the non-existence of Thomas, given the intellectual and chronological distance of the Sir Tristrem poet from his source.

On the matter of the identification of the author(s) of T1 and T2 with the Thomas mentioned in the Douce and Sneyd fragments, Bouchard contends that the references to Thomas in Gottfried and Sir Tristrem would suggest “that there was a tradition of attributing one's ideas to an imaginary ‘Thomas’ when writing of Tristan” (68). She argues further “that the fragments do not actually say they are written by Thomas, but rather that Thomas told the story correctly” (68).

In this regard, it is well to examine carefully the poet's words. In the Douce fragment, it is not Thomas but Breri who is invoked as the authority for the present version:

Entre ceus qui solent cunter
E del cunte Tristran parler,
Il en cuntent diversement:
Oi en ai de plusur gent.
Asez sai que chescun en dit
E ço que il unt mis en escrit,
Mes sulun ço que j'ai oi,
Nel dient pas sulun Breri
Ky solt les gestes e les cuntes
De tuz les reis, de tuz les cuntes
Ki orent esté en Bretaingne.

(D 841-51)

Thomas's name designates not an authoritative source but the poet himself who, in composing his work, exercised the taste and judgment required to select as his model from among competing versions the most authentic and versimilar narrative of the legend and more particularly of its penultimate episode:

Ensurquetut de cest' ovraingne
Plusurs de noz granter ne volent
Ço que del naim dire ci solent,
Ki femme Kaherdin dut amer:
Li naim redut Tristran navrer
E entuscher par grant engin,
Quant ot afolé Kaherdin;
Pur ceste plaie e pur cest mal
Enveiad Tristran Guvernal
En Engleterre pur Ysolt.
Thomas iço granter ne volt,
E si volt par raisun mustrer
Qu'iço ne put pas esteer.

(D 852-64)²²

The context of the passages in which the references to Thomas occur makes it abundantly clear that they were understood to designate the author of the work in which those passages are found. In the first, the poet employs the first person singular ten times (D 836-39, 844-45, 847, 877, 882-83) and the first person plural once (D 853). Thomas is evoked once in the third person (D 862). The distribution of verb tenses resolves any ambiguity, making it evident that Thomas and jo designate the same person, who is distinct from Breri. Like those which depend from jo and unlike that whose subject is Breri (solt, D 849), the verbs that depend from the subject Thomas (volt and volt, D 862-63) are in the present indicative. Thus, there can be no doubt that the poetic voice is designating itself when it speaks in the third person singular present indicative of Thomas and its source when it speaks in the third person singular preterit of Breri.

It is even clearer in the envoi of the Sneyd² fragment that the person identified as Tumas is speaking with his own voice. The envoi, which contains six first-person constructions (Sn² 826, 827, 828, 829, 830, 831), is a highly personal and self-reflective apostrophe by the poet to his readers, to whom he offers his work as a palliative against the suffering occasioned by unhappy love.²³

Tumas fine ci sun escrit:
A tuz amanz saluz i dit,
As pensis e as amerus,
As emvius, as desirus,
As enveisiez e as purvers,
A tuz cels ki orunt ces vers.

(Sn² 820-25)

In vv. 820-25, designating itself in the third person singular as Tumas, the poetic voice addresses itself to “tuz amanz.” In vv. 830-36, speaking in the first person singular, the poet avers that his work is destined “as amanz,” that they might find its matter pleasurable, memorable, and comforting:

E diz e vers i ai retrait:
Pur essample issi ai fait
Pur l'estorie embelir,
Que as amanz deive plaisir,
E que par lieus poissent troveir
Choses u se puissent recorder:
Aveir em poissent grant confort …

(Sn² 830-36)

The repetition of the apostrophe “to all lovers” by the poetic voice, speaking first in the third person, then in the first, signals the identity of these two personae and demonstrates that the Tumas of D 820 is identified with the author and not with the source of the text in which his name figures.

The evidence for an historical Thomas is minimal. The theories identifying him with Thomas of Erceldoune, with Thomas of Kent, and with “Mestre” Thomas, author of Horn et Rimel, have been discredited, while those linking him to the court of Alienor of Aquitaine are no more than speculation.²⁴ If, however, there is no more evidence for his existence than for that of, say, Beroul, there is no less and, hence, in the absence of any compelling reason for so doing, no justification for denying him the same status as that of any named but otherwise unknown medieval author.

F. T1 AND GOTTFRIED'S TRISTAN UND ISOLDE

The “remarkable coincidence” (68) that Gottfried's poem breaks off “precisely where the first of the thirteenth-century fragments begins” (68) becomes less remarkable if one notes that the Cambridge fragment coincides with a portion of Gottfried's text (vv. 18195-307). Moreover, it has been noted by more than one scholar that Gottfried transposes to his prologue material from T2, notably certain of Thomas's remarks on the principles of artistic composition (D 835-51) and on the types of lovers (Sn² 820-39).²⁵ Finally, the compilers of a recent survey of early French manuscripts have determined beyond reasonable doubt that the Sneyd manuscript belongs to the twelfth century.²⁶ Thus, Thomas's entire manuscript necessarily existed prior to the composition of Gottfried's poem.

However, the strongest refutation of Bouchard's hypothesis lies in the demonstration that the Sneyd manuscript, Oxford, Bodleian MS. Fr. d. 16, is a mutilated fragment of an originally longer codex. The Sneyd manuscript (together with a large number of other manuscripts including the Beroul manuscript, now MS. B.N.fr. 2171) was purchased by the Rev. Walter Sneyd in Venice in 1835 and was acquired by the Bodleian Library in 1934. It consists of fourteen parchment leaves of the original manuscript, numbered in this century from 4 to 17, as well as fly-leaves and binding papers of more recent date.²⁷ The Sneyd¹ fragment, a quire of four bifolia, the first of which has lost its first leaf, begins in medias res at the top of the first column of folio 4r. (Folios 1 and 2 consist of modern preliminary material; folio 3, also modern, has been removed.) This fragment ends as abruptly as it begins, at the bottom of the second column of folio 10v; and the Sneyd² fragment, also a quire of four bifolia, the last of which has lost its second leaf, begins at the top of the first column of the next folio. Thus, an indeterminate number of leaves have also been lost between folios 10 and 11. In its present form, the manuscript contains nothing but the two Thomas fragments. However, the sewing holes and folds in folio 21, a notarial document dated 1366 that served at one time as a limp-vellum cover to the book, suggest that the codex originally contained approximately thirteen quires of eight leaves each, only two of which, Sneyd¹ and Sneyd², remain. Thus, there is every reason to suppose that the Sneyd manuscript originally contained other portions, now lost, of Thomas's poem, particularly material immediately preceding the extant Sneyd¹ fragment. Moreover, it would appear that the lost portion preceding Sneyd¹ was of substantial length. An original quire-signature at folio 10v indicates that the quire containing the Sneyd¹ fragment was the eighth of the original codex. Assuming that each quire was a regular quaternion of four bifolia and assuming that the missing leaves preceding Sneyd¹ contained only the text of Thomas's poem, this missing portion would consist of 57 pages or approximately 7296 lines (7 quires of 8 leaves + 1 leaf of quire 8 x 4 columns per leaf x 32 lines per column). Hence, the “remarkable coincidence” concerning Gottfried's poem and the Sneyd fragment is nothing more than coincidence and must not be taken as evidence of the original state of the Sneyd manuscript or of Thomas's work.

G. THOMAS AND WACE

In addition to the negative arguments adduced above against the thesis of the dual authorship of Thomas's poem, one positive argument in favor of the unitary thesis of composition may be put forward. The utilization in both T1 and T2 of narrative material imitated from Wace's Brut strongly suggests the composition of both parts of the romance by a single author.

The particularity of Thomas's poem resides in his style, found in his rhetorical embellishments, and in his personal conception of his subject matter, articulated in his authorial commentary. Yet, the adaptation of the Saga does violence to both of these, making it impossible to discern stylistic or ideological similarities between the Saga and the extant fragments. However, one signature, this attached to the narrative material—the only element of his model that Robert renders with some fidelity—may be identified. Thomas's borrowing of narrative material, personal names, and descriptions from the Roman de Brut of Wace has long been recognized.²⁸ While certain scholars have perhaps been incautious in overstating the extent of this imitation, particularly as regards descriptive material, it is demonstrable that Thomas knew Wace's work and utilized narrative material derived from it in both parts of his romance, a fact that argues strongly for unitary authorship. The correspondances in question have been demonstrated at length by others.²⁹ I will merely cite those that I believe to be incontrovertible:

• (1) The description in T1 (S Ch. 27) of the tribute demanded of the Cornish by Le Morholt imitates closely the speech of Arthur denouncing the claims of the Roman senate.³⁰
• (2) The description in T2 of the combat between Arthur and the giant Orguillos, uncle of Urgan (Sn¹ 664-728), shows a close similarity to that of Arthur and Riton in the Brut (11565-92).³¹ Both are appended to the description of another combat between the hero and a giant.
• (3) The description in T2 of the giant of Mont-Michel, the predecessor of Moldogog (S Ch. 78) corresponds closely to the description of Dinabuc in the Brut (11279-95, 11309-21, 11397-412, 11321-36, 11459-550).³²

In every case, the similarities concern narrative event, physical detail, and personal names—those elements that the Saga poet reproduces with the greatest fidelity.

In conclusion, the contention of a recent article that the extant fragments of the Roman de Tristan are not the work of Thomas and that T1 and T2 are the work of different poets is without foundation. The allusions of the poetic voice to Thomas and Tumas are self-referential and do not designate a literary source. Thus, there is no reason for abandoning the convention by virtue of which we designate the author of the romance in which his name appears as Thomas or the cautious credence that we lend to his authorial pronouncements. The authorial signature represented by the imitation of narrative material from Wace's Brut is found in both halves of Thomas's poem, a circumstance that argues strongly in favor of the presumption that they were the work of a single poet. I would propose for the case of Thomas the conclusion of Gwenneth Whitteridge concerning the authorship of Beroul's romance: “auctores (like entities) non sunt multiplicandi praeter necessitatem.”³³

Notes

1. The study in question is Constance B. Bouchard's “The Possible Nonexistence of Thomas, Author of Tristan and Isolde,” MP 79 (1981-82): 66-72. The authoritative edition of the Thomas fragments, from which I quote, is Thomas. Les Fragments du roman de Tristan, poème du XIIe siècle, ed., Bartina H. Wind, Textes Littéraires Français, No. 92 (Geneva: Droz; Paris: Minard, 1960) [Wind 1960]. Because I believe the Old Norse Tristrams Saga to be a more reliable guide than Bédier's reconstruction to the lost portions of Thomas's Tristan, I will refer to Eugen Kölbing, Die nordische und die englische Version der Tristan-Sage. I. Tristrams Saga ok Isondar mit einer literar-historischen Einleitung, deutscher Uebersetzung und Anmerkungen zum ersten Mal herausgegeben (Heilbronn: Henninger, 1878) [Kölbing] or quote from The Saga of Tristram and Isönd, trans., Paul Schach (U. of Nebraska Press, 1973) [Schach] when I seek to draw conclusions concerning the missing content of Thomas's poem. However, I will also refer to Joseph Bédier, Le Roman de Tristan par Thomas, poème du XIIe siècle, Société des Anciens Textes Français, 2 vols. (Paris: Firmin-Didot, 1902, 1905) [Bédier 1 and Bédier 2]. References to these works will be given parenthetically.

2. Gottfried von Strassburg. Tristan und Isold. Text, ed., Frederich Ranke, 9th ed. (Zurich and Berlin: Weidmann, 1965), l. 150.

3. The Cambridge fragment [C]: Cambridge, University Library, Add.MS 2751 (D.D.15.12), “sans doute le plus jeune [des manuscrits]” (Wind 1960, 9); the Sneyd fragments [Sn¹ and Sn²]: Oxford, Bodleian Library, MS. Fr. d. 16, fols. 1-14, “probablement du XIIe siècle” (Wind 1960, 9), one of a group of manuscripts “dont l'appartenance au XIIe siècle peut être reconnue soit sans contestation soit avec beaucoup de vraisemblance” (Brian Woledge and Ian Short, “Liste provisoire de manuscrits du XIIe siècle contenant des textes en langue française,” Romania 102 [1981]: 1-17 at 2); the Douce fragment [D]: Oxford, Bodleian Library, MS. Douce d. 6, fols. 1-12, “de la fin du XIIe ou du début du XIIIe siècle” (Wind 1960, 9); the Turin fragments [T¹ and T²], “sans doute du XIIIe siècle” (Wind 1960, 10), whose present location is unknown; and the Strasbourg fragments [Str¹, Str², Str³]: Strasbourg, Bibliothèque du Séminaire Protestant, “du XIIIe siècle [?]” (Wind 1960, 10), destroyed in 1870. It is tempting to speculate that this was the very manuscript utilized by Gottfried von Strassburg. In the Douce manuscript, which contains 22 leaves, the Douce fragment (fols. 1a-12c) is followed immediately by the text of the Folie Tristan d'Oxford (folios 12d-19a). The manuscript also contains a debate in French verse between Pride and Humility (folios 19-20) and meditations in French and Latin prose on the True Cross (folios 20-22).

4. Ranke, pp. 155-71.

5. Sir Tristrem, ed. George P. McNeill, Scottish Text Society, No. 8, (Edinburgh: Blackwood, 1886; rpt. New York: Johnson Reprint Corporation, 1966), pp. 1-2, 397, 412, 2787. On the authorship question, see pp. xxxii-xlv.

6. On the question of the fidelity of the Saga to its model, see Bédier 2: 71-75 and Alfrún Gunnlaugsdóttir, Tristán en el Norte (Reykjavik: Arna Magnússonar, 1978), pp. 329-32.

7. “The Pattern(s) of the Heroic Biography: Structure and Sense in the Tristan of Thomas of Britain,” Michigan Academician 9 (1978-79): 375-90.

8. The latter episode is found only in Eilhart's poem, in the Saga, and in the variant version of the French prose Tristan contained in MS. B.N.fr. 103. In a forthcoming study, I contend that a version of the episode originally existed in Thomas's romance, from the extant manuscripts of which it was subsequently excised. See “Mouvance and Revisionism in the Transmission of Thomas of Britain's Tristan,” Arthurian Literature 6 (1986).

9. It was Eugène Vinaver who first noted the symmetrical structure of the Tristan narrative. See his Etudes sur le Tristan en prose. Les sources; les manuscrits; bibliographie critique (Paris: Champion, pp. 1925), pp. 7-10.

10. Thomas, D 835-84; Beroul (ed. Defourques), 1265-70, 1787-92; Eilhart (ed. Buschinger), 9446-57; Gottfried (ed. Ranke), 131-54.

11. Tony Hunt, “The Significance of Thomas's Tristan,” Reading Medieval Studies 7 (1981): 41-61.

12. By “full” (Fr. totalisantes), I mean those versions that recount the entire narrative of Tristan's life and death. Beroul's and Gottfried's poems, full versions that are incomplete, are the exceptions to this statement, as are, of course, the episodic versions discussed below.

13. See Rachel Bromwich, “Some Remarks on the Celtic Sources of Tristan,” Transactions of the Honourable Society of Cymmrodorion (1953): 33-35.

14. In B.N.fr. 103, the adulterous knight is Ruvalen, Kaherdin's brother.

15. Douglas Kelly, “En uni dire (Tristan Douce 839) and the Composition of Thomas's Tristan,” MP 67 (1969-70): 11.

16. Ernest Hoepffner, La Folie Tristan de Berne, publiée avec commentaire (1934; 2nd ed., Paris: Belles Lettres, 1949), pp. 6-12.

17. It is clear, however, that Heinrich, writing between fifty and seventy years after Ulrich, also consulted the latter's version.

18. Michael S. Batts, Gottfried von Strassburg (New York: Twayne, 1971), pp. 12-13.

19. Lucie Polak, “The Two Caves of Love in the Tristan by Thomas,” Journal of the Warburg and Courtauld Institutes 33 (1970): 52-69, at 68.

20. For the studies devoted to this question through 1977, see David J. Shirt, The Old French Tristan Poems: A Bibliographical Guide (London: Grant & Cutler, 1980), T.d. 18-22, and especially the excellent mise au point of Bartina Wind, “Nos Incertitudes au sujet du Tristan de Thomas,” in Mélanges de langue et de littérature du moyen âge et de la Renaissance offerts à Jean Frappier (Geneva: Droz, 1970), 2:1129-38.

21. Bartina Wind, Les Fragments du roman de Tristan, poème du XIIe siècle par Thomas (Leiden: Brill, 1950) [Wind 1950], p. 10.

22. In twelfth and thirteenth-century French romances, authorial references that employ a proper name are virtually always couched in the third person, although first-person singular subject pronouns designating the author may be found in the same passages. E.g., Le Roman de Brut (ed. Arnold), 7, 3823, 13282, 13866; Beroul (ed. Defourques), 1268, 1790; Cligés (ed. Micha), 43, 6664; Le Chevalier de la charrete (ed. Roques), 25, 7102; Yvain (ed. Reid), 6815; Le Conte du Graal (ed. Roach), 7, 62; Le Bel Inconnu (ed. Williams), 6249.

23. For analogous examples of passages containing a mixture of first and third person authorial references, see Cligés (ed. Micha), 1-13; Le Chevalier de la charrete (ed. Roach), 1-29; and Le Bel Inconnu (ed. Williams), 6247-66.

24. Wind, “Incertitudes,” 1130-34.

25. See Bédier 1:1, n.1; Jean Fourquet, “Le Prologue du Tristan de Gottfried,” Bulletin de la Faculté des Lettres de Strasbourg 31 (1952-53): 252; and Magda Heimerle, Gottfried und Thomas. Ein Vergleich (Frankfurt: Diesterweg, 1942), pp. 13-17.

26. Woledge and Short, p. 7. I owe this observation to the kindness of Susan Crane.

27. The descriptions of the Sneyd manuscript given in Bédier 2:2-4 and Wind 1950, 1-2 are mutually inconsistent and contain minor inaccuracies. My own statements concerning the manuscript are based on information generously communicated to me by Dr. Bruce Barker-Benfield, Assistant Librarian in the Department of Western Manuscripts of the Bodleian Library.

28. On the use of authorial “signatures” in the attribution of authorship, see Frédéric Deloffre, “Stylistique et critique d'attribution,” in Au Bonheur des mots: Mélanges en l'honneur de Gérald Antoine (Nancy: Presses Universitaires de Nancy, 1984), pp. 509-20.

29. See Bédier 2:99-101; Margaret M. Pelan, L'Influence du “Brut” de Wace sur les romanciers français de son temps (Paris: Droz, 1931), pp. 72-97; Anthime Fourrier, Le Courant réaliste dans le roman courtois en France au moyen-âge. I. Les Débuts (XIIe siècle) (Paris: Nizet, 1960), pp. 19-109; as well as the salutory warning voiced by Joël Grisward in “A propos du thème descriptif de la tempête chez Wace et chez Thomas d'Angleterre,” in Mélanges de langue et de littérature du moyen âge et de la Renaissance offerts à Jean Frappier (Geneva: Droz, 1970), 1:375-89. I subscribe to Grisward's conclusion that elements of the description of the storm in Thomas's poem (D 1592-1614, 1695-99) are found elsewhere in twelfth-century literature and may represent not a direct imitation of Wace but “un souvenir imprécis de l'Enéas, du Roman de Troie ou du Guillaume d'Angleterre, voire tout simplement un écho mêlé de toutes ces lectures, un reflet de sa Culture” (p. 389).

30. Le Roman de Brut de Wace, ed. Ivor Arnold (Paris: Société des Anciens Textes Français, 1940), 2:10819-20, 10825-34, 10893-94, 10903-4. See Bédier 1:81-82 and Pelan, 92-94. While the Saga has eliminated all proper names from the passage, as is its general habit, it is possible exceptionally to identify in the Saga the translation of a paraphrase by Thomas of the following passage from Wace:

    Mais force n'est mie dreiture
    Ainz est orguil e desmesure.
    L'um ne tient mie ço a dreit
    Que l'um ad a force toleit.

    (10829-32)

    En ofríki er ei séttindi, heldr opinberlig skomm ok rangindi. … Allt þat fé, er med rán fæst, hvarvitna er þat illa aflat. (Kölbing, Ch. 27, pp. 32-33)

    But tyranny is not justice, but lawlessness and an obvious shame and injustice. … All goods that are seized with robbery everywhere are ill-gotten. (Schach, p. 40)

31. See Bédier 1:289-91 and Pelan pp. 75-77.

32. See Bédier 1:307-9 and Pelan pp. 77-80. Moreover, the account in the Saga closes with a formula similar to that which closes the account of Orguillos and Arthur:

    A la matire n'afirt mie,
    Nequedent boen est quel vos die …

    (Sn¹ 729-30)

    But as far as the story of the giant whom the king slew is concerned, the only part that pertains to this saga is the fact that. … (Schach, p. 119)

33. Gweneth Whitteridge, “The Date of the Tristan of Beroul,” Medium Aevum 28 (1959): 171.