Printer Friendly

Versification and translation in Sir Beves of Hampton.

Sir Beves of Hampton is one of a handful of Middle English romances in which unexpected and not always obviously motivated changes in verse form occur at some point in the text. The Auchinleck text of Guy of Warwick offers the best-known and least problematic example of the phenomenon. There the shift from couplets to tail-rhyme stanzas appears to mark a deliberate narrative caesura: a new section of the poem begins and the sense of a fresh start is underscored by the insertion of exordial material unparalleled in the Anglo-Norman original of the poem and its other Middle English redactions. The two tail-rhyme stanzas that introduce the Auchinleck version of Richard Coer de Lion remain a puzzle, while the change from septenary couplets to tail-rhyme stanzas halfway through the Ashmole Sir Ferumbras has been convincingly attributed to a change in the source text used by the Ashmole author/ translator? In all these narratives the sudden shift is limited to a single manuscript: the only manuscript of this redaction of Ferumbras, but one of several in the cases of both Guy and Richard. Beves is unique in that five of its six manuscripts, representing four different redactions of the poem, switch from couplets to tail-rhyme stanzas more or less simultaneously. (2) Such uniformity has made it harder to account for the change by reference to scribal whimsy, and there has been some debate over the years about what might have prompted it. In what follows I shall suggest that this odd feature of the versification of Beves is a direct consequence of the constraints under which the translator of the Anglo-Norman Boeve de Haumtone had to work. (3)

A translator's work is subject to a complex web of hierarchically ordered constraints. The most important, and the one over which the translator has the least control, is obviously language. Versification is less constraining in that prosodic form, unlike language, can be chosen with a certain degree of freedom; once chosen, however, it too exerts a powerful restrictive influence on a translator's choices at all textual levels. The Middle English translator of Boeve exercised his freedom of choice in a way that has puzzled the critics of the text. Unlike later translators, he did not have to worry about whether to use verse or prose for his rendering of the Anglo-Norman narrative. At this early date prose was not a real option, as the great flowering of prose romances in thirteenth-century France had no parallel in England. (4) Verse was the norm, and the main question was, what kind of verse? The answer, it seems, was not obvious. A salient feature of the versification of Beves is the abrupt change from six-line tail-rhyme stanzas to rhymed couplets at line 475 in manuscripts A, E, and C. (In SN the tail-rhyme pattern continues until line 528.) Eugen Kolbing, the romance's first editor, rejected the possibility that the couplet section of the poem was 'the work of a continuer [sic], who found the stanza too difficult or not according to his taste'. As there was nothing 'to correspond with this change in the original French versions', Kobing felt compelled to accept the shift in verse form as vet another example of what is, after all, not an isolated phenomenon in Middle English literature. (5) Derek Pearsall, who made no secret of his low opinion of Beves's artistic value, conjectured that 'a different hack' might have taken over or that 'the English adaptor recognized the unsuitability of the more "poetic" measure to this vulgar thriller') Dieter Mehl suggested 'that the first part of the poem was also originally written in rhyming couplets which by slight alteration, in many cases merely by the insertion of caudae, were turned into tail-rhyme stanzas'] This is very likely true of the continuation of the stanzaic pattern in SN (lines 475-528), which a comparison with AEC shows to be by and large the result of the mechanical addition of tail lines to existing couplets. For the earlier section of the narrative, however, this is not the case. There is certainly a fair amount of padding, though no more than in the couplet section of the poem, but many of the tail lines are syntactically important and, as Kolbing pointed out, 'absolutely necessary for the context'. (8)

More compelling was the theory proposed by A. C. Baugh in one of his examinations of Beves: that the change in the Middle English romance in fact parallels a similar change, unnoticed by Kolbing, in the Anglo-Norman original. Quoting the editor of the Anglo-Norman text, Baugh remarked that the length of the laisses in Boeve
 undergoes a very noticeable change after laisse 66 (which ends at
 line 415). Up to this point they have all been short, generally six
 lines, with an occasional divergence to five or seven. The effect
 is that of a poem written in six-line monorimed stanzas. From this
 point on the laisses are generally much longer ... Now the change
 from monorimed stanzas averaging six lines occurs at very nearly
 the same point in the story as that at which the English poem
 changes from six-line stanzas to couplets. (9)


Baugh's main concern in this article was to dissociate romances such as Beves from the oral tradition by strengthening the case for their written composition or, with regard to Beves, translation. It was therefore in his interest to connect as many features of the English romance as possible to corresponding features of the Anglo-Norman original. His purpose, however, does not weaken his argument, which is based on compelling evidence.

Like Baugh, Brian Levy explicitly connected the change in verse form with the fact of translation, noting that 'le poete-traducteur a meme reproduit, sous forme de courtes stances, les "laisses embryonnaires" du debut du poeme'. (10) And yet Jennifer Fellows, the poem's most recent editor, was disinclined to accept Baugh's claims as conclusive, on the grounds that, as Baugh himself noted, there was no 'one-to-one relation between the English stanzas and the [Anglo-Norman] laisses', and that, moreover, 'comparable metrical anomalies occur in other romances', for which translation could not be adduced as a possible influence. (11) Having examined the evidence, Fellows was more willing to align herself to a certain extent with Pearsall. Her conclusion was that the Middle English author, though reasonably smiled in the use of the tail rhyme stanza, was not skilled enough 'to compensate for the restrictions that [the stanza] imposes upon the narrator'. She therefore attributed 'its abandonment--and perhaps also the initial choice of the simpler, six-line form--to a sense of unease and a desire for greater metrical freedom on the part of the translator, who perhaps chose it to begin with because of its popularity in the fourteenth century'. (12)

Fellows backed her conclusions with an impressive array of textual evidence, and her estimate of the Middle English adapter's skill in handling the tail-rhyme stanza is by and large accurate. Nevertheless, Baugh's hypothesis cannot be dismissed so easily. The difficulty of the tail-rhyme stanza may have been a contributing factor in the translator's decision to abandon it eventually for the couplet, which is obviously easier to handle, but the reason he chose it in the first place must have been the versification of the Anglo-Norman original. (13)

I should explain that 'the Anglo-Norman original' is not an extant text that we can point to. The textual history of Boeve is in fact full of gaps. Scholars agree that the tale probably came into being in the last decade of the twelfth century, and the version in which it has come down to us was assigned by its editor to the first half of the thirteenth century, a dating on linguistic grounds which remains unchallenged. (14) The two fragmentary but complementary manuscripts in which this version was preserved postdated its presumed date of composition: Albert Stimming placed the no longer extant D, which contained lines 913-3850 of his edition of Boeve, in the second half of the thirteenth century; our only source for the beginning of the poem (lines 1--1268), manuscript B, belongs to the early fourteenth century or perhaps the very end of the thirteenth. (15) B is thus probably too late to have been used by the author/translator to whom we owe Beves and who is assumed to have produced the Middle English poem in the later thirteenth century. Besides, there are clear indications that B was copied by a scribe who, while generally competent, did not understand the versification of the Anglo-Norman poem. (16)

How then, in view of these textual complexities, can we speak of 'the Anglo-Norman original'? I have dealt with these problems at greater length elsewhere, arguing that the study of 'presumed originals' and 'presumed translations' is justified in dealing with medieval texts. (17) For the purposes of the present study I would single out my conclusion that the precise identification of 'original' and 'translation', which poses a major problem in attempts to disentangle macro-textual relations, is not too much of a problem in the micro-textual analysis of translational procedures of the kind that I am interested in here. When a Middle English manuscript reading matches very closely a passage in the only Anglo-Norman text we happen to have, it can safely be assumed to derive more or less directly from the ur-translation at that particular point. If several Middle English redactions agree with the Anglo-Norman, as in the opening sections of Boeve and Beves, a translational relationship between them and the presumed original is even more likely. (18)

Why did the Middle English translator choose the six-line tail-rhyme stanza for the opening of Beves? Was it really, as Fellows suggested, because of the tail-rhyme stanza's popularity? I think it unlikely. It is, of course, perfectly natural for a translator to render a verse form that has not been domesticated in his own language by turning it into whatever happens to be the most popular verse form for similar texts in the receiving literary system. But this is not what the six-line tail-rhyme stanza was. Although probably the form with which the general public is now most familiar because Chaucer used it in 'Sir Thopas', it is quite rare, the standard form being a stanza of twelve lines. (19) The six-line stanza is so rare, in fact, that it appears in none of the romances with which Beves shares its manuscript contexts. (20) Besides, it is hard to speak of 'popularity' as an established fact when A is in fact the earliest manuscript containing romances in tail-rhyme stanzas, and for all we know Beves may have been the earliest tail-rhyme romance in Middle English. (21)

The translator's other option is to remain as close as is feasible to the versification of his source and to select a verse form current in English, not for its popularity but for the ease of translation that it allows. To choose six-line stanzas as a rough approximation--in length if not in structure--of laisses averaging six lines is to solve one significant problem in any translation process: that of segmentation, i.e. of deciding what you are going to treat as the basic unit of translation. (22) To be able to match such units in the original and the translation makes a translator's life much easier. The matching need not be perfectly accurate, (23) however, and Baugh was right to dismiss the absence of a 'one-to-one relation' between the English and Anglo-Norman verse forms as of limited relevance.

The choice of basic unit of translation is a largely automatic decision, but the analysis of any process of translation requires that the automatism be reconstructed and brought out into the open. From the translator's perspective, the unit of translation may he 'the stretch of source text on which the translator focuses attention in order to represent it as a whole in the target language', but what the scholar looks for, working backwards, is 'the target-text unit that can be mapped onto a source-text unit'. (24) The tail-rhyme section of Beves can be mapped onto the short-laisse section of Boeve with remarkable consistency. This does not mean, however, that sixty-six stanzas of Beves correspond to the first sixty-six laisses of Boeve. As Baugh noted, the versification change in Beves occurs 'very nearly' at the same point as in Boeve, but not exactly at the same point. AEC switch to couplets at line 475, corresponding to the beginning of laisse 52 (line 327) of Boeve, while the tail-rhyme pattern continues in SN until line 528, corresponding to the end of laisse 60 (line 378). (25) Furthermore, the difference in length between the Anglo-Norman alexandrines and the short lines of the Beves tail-rhyme stanza would in itself make a perfect line-for-line fit impossible. (26) In fact, the 415 lines of the short-laisse section of Boeve are translated by 580 lines in A (seventy-nine tail-rhyme stanzas and fifty-three couplets) and 600 lines in S (ninety-two stanzas and twenty-eight couplets). (27) The transfer of significant information is carried out using the laisse as the basic unit, which becomes apparent when we look closely at the principal modes of transfer. (28)

More often than might have been expected, in view of the difference in metre, the translator was able to maintain a one-to-one correspondence between the laisses of the original and the tail-rhyme stanzas of the translation. Indeed, the first few Middle English stanzas closely match the first four Anglo-Norman laisses, and the translator may well have intended this to be the pattern throughout:
 Seingnurs barons, ore entendez a mei, Lordinges, herknep to me tale
 si ws dirrai gestes, que jeo Is merier pan pe
 diverses sai, niztingale,
 de Boefs de Haumtone, li chevaler Pat y schel singe.
 curtays,
 ke par coup de espeie conquist tant Of a knizt ich wile zow
 bons reys. roune--
 Si vus volez oyer, jeo vus en dirrai; Beues a hizte of Hamtoune,
 unkes ne oistes meyllur, si com jeo Wip-outen lesing.
 crai.

 Seignurs, si de lui oyer Ich wile zow tellen al
 desirez, to-gadre
 jeo vus en dirrai, kar jeo sai O pat knizt and of is
 asez; fadre,
 primes vus en dirrai de soun Sire Gii:
 parentez.
 A Haumtone fu li quens plein de Of Hamtoun he was sire,
 bontez,
 il out a noun Guioun, chevaler And of al pat ilche schire,
 fu prisez;
 meilour de lui ne fust en son tens To wardi. (A lines 1-12)
 trovez. (lines 1-12)


In a poem's introductory section, whose main purpose is to capture the audience's attention, significant information (italicized above) will be relatively scant, lost amid the generalizations and the attention-grabbing conventions proper to genre and language. If we look at the two texts side by side we can see quite clearly what the translator considered to be significant: the name of the hero, his knightly status, the initial focus on the hero's father, the father's name and position. (29) After the generalizations of the introductory laisses--which, for the most part, could easily be jettisoned without causing confusion--the plot got under way and it became much harder to pack the content of the leisurely Anglo-Norman lines into the brisk Middle English triplets. (30) Therefore the procedure had to be modified accordingly. The translator will still attempt to achieve a one-to-one fit wherever possible, and will be able to do so for a little less than one-quarter of all laisses: fifteen laisses of Boeve are translated by a single tail-rhyme stanza in both A and SN. All these laisses are short (five or six lines) and, in narrative terms, fairly simple; this allows the translator to squeeze their content into the less commodious Middle English stanza without too much cutting. Laisse 14 is very slightly longer (seven lines), but the translator's job was here made easy by the redundancy of the information in it. (31) In other cases we find different factors facilitating the transfer of information: conventionality, for example in laisses 1-3, in which important information is framed by exordial stereotypes; repetition of lines at the end of one laisse and the beginning of the next one, in a linking pattern common in chansons degeste and residually present in Boeve as well (laisses 37f.); (32) and laisses containing one significant action, speech, or state and inessential details that can easily be left out: 30, 37, 41, 48, 51.

It is not always possible to maintain this kind of correspondence even by focusing on the essential information contained in a laisse, and the translator often finds himself compelled to expand one/aisse of Boeve to two. This is, in fact, the predominant mode of transfer: eighteen laisses are turned into two stanzas in A, nineteen in SN. In some cases, the expansion of one stanza to two or three is a simple matter of accommodating the content of the laisse to a less capacious verse form without having to leave anything out. Perhaps surprisingly, the length of the original laisse does not seem to be decisive: while some of the laisses that get expanded to two stanzas are longer than the average (seven, eight, or even ten lines) others are quite short (six lines, at times five). What seems to be more important, especially in these latter cases, is the narrative complexity of the laisse and the significance of its content both for the plot in general and for the portrayal of the central characters, the hero in particular. This is where more substantial amplification may occur, especially when the original laisse is built around a conversation or monologue of particular significance for the characterization of one of the protagonists. Here is laisse 16, in which Doun, having just heard the countess's message, rejoices at the prospect of killing her husband and stepping into his shoes: (33)
 Le emperur oi que dist lui messager; 'Sai', a seide, 'icham at
 hire heste:
 seil se enloie, ne let a demaunder. zif me lif hit wile leste,
 Hit schel be do!
 Gladder icham for pat sawe
 Dan pe fouel whan hit
 ginnep dawe--
 And sai hire so.

 'Messager', dist il, 'par le cors 'And, for pow wolde hire
 Seint Richers! erande bede,
 pus ke ceo noveles vus me aportes, An hors icharged wip
 golde rede
 jeo te frai doner un bon coraunt Ich schel pe zeue;
 destrer
 e or e argent dunt tu le poez And wip-inne pis fourtene nizt
 charger'.
 (lines 95-100) Me self schel dobbe pe to
 knizt,
 zif pat ich liue'. (A
 lines 145-56)


The translator clearly feels that the cursory reference to Doun's feelings in line 95 of Boeve does not do justice to the iniquity of the character, who appears here for the first time and will be the hero's principal foe in the first half of the romance. An entire stanza is needed for the ironically lyrical outburst with which the villain greets the news. The promise of a horse charged with gold and silver can then be dealt with briefly, in as many lines as it takes up in Boeve, but this leaves the translator a triplet short. Although, as we shall see, he is not always averse to constructing a stanza out of material that belongs to two different laisses, he does try to avoid that whenever possible. Hence the addition of the promise to knight the messenger, which fills out the stanza. (34)

Another kind of amplification occurs when the translator decides to expand a brief comment in Boeve to a full stanza, as he does, for example, in translating laisse 35 :
 La dame a tresoi, ke cil va disaunt pe moder hire hap
 vnderstonde;
 hause la pume si le feert eraument, pat child zhe smot wip
 hire honde
 Vnder is ere.
 ke chaier le fist sur le pavement. pe child fel doun--& pat
 [was] scape;
 Le mestre a le enfant est sailli avaunt. His meister tok him wel
 Il out a oun Sabot, due li seit eidaunt! rape,
 pat hizte Saber.
 Chevalet fu riches, fort e caombataunt. pe knizt was trewe, & of
 (lines 220-5) [is] kinde
 Strenger man ne scholde
 men finde,
 To ride ne go.
 A was ibrouzt in tene &
 wrake Ofte, for pat
 childes sake,
 Ase wel ase po.
 (A line 319-30)


The addition of an entire stanza dedicated to a characterization of Sabot ('Saber' in the English poem) was necessitated by the impossibility of squeezing the information in Boeve, line 225, into the already packed Middle English stanza. The line could have been left out, but as Sabot is one of the principal characters, the hero's chief protector and ally throughout the story, the omission of this introductory characterization, no matter how brief, would have been awkward.

The translator obviously found this kind of technique useful, and we find him doing the same thing again, for example in laisse 22, where he remains as close to the Anglo-Norman text as he can, but again expands the final line of the laisse to an entire stanza:
 Lui quens mounta pat Erl is hors be-gan to stride;
 un destrer abrive, His scheld he heng vpon is side,
 un escu a son col, Gert wip swerd.
 en sa mein un espe;

 il ne avoit nul hauberk Moste non armur on him come.
 ne nul heaume gemme, Him-self was boute pe ferpe some
 treis companions sunt Toward pat ferd.
 ov lui muntez.

 Ore mourra lui quens a Allas pat he naddle be war
 doel e a vilte. Of is fomen pat weren par,
 (lines 138-42) Him forte schende!
 Wip tresoun worp he per islawe,
 And ibrouzt of is lif-dawe,
 Er he hom wende. (A lines
 199-210)


The five-line laisse is short enough for the significant information in it to be fitted into a single tail-rhyme stanza. This is what the translator does--and very skilfully too. However, in keeping with his general tendency to assume an active role in guiding the reactions of the audience, he is reluctant to leave out the proleptic comment in the final line of the Anglo-Norman laisse. He takes it up (line 208), but builds around it an entire stanza, in which he does his best to arouse sympathy with the earl and indignation at his impending death by treachery. (35)

A similar effort to control the audience's sympathies, though in a more discreet way than in laisse 22, is detectable in the translation of the immediately preceding laisse. In the case of this exceptionally long laisse it is perhaps misleading to speak of amplification, for the translator does not need to add any material in order to turn its twelve lines into three tail-rhyme stanzas. (36) What he does here might be described more properly as clarification, an intervention in which, by breaking down the narrative content of the lengthy laisse, he emphasizes the formal structure of the dialogue it contains and makes its implications more obvious. It is an important dialogue, in which the countess, who is feigning illness in order to send her husband to his death, asks him to bring her the flesh of a wild boar that can be caught in a nearby forest. The translator remains very close to the wording of the original, but renders some instances of direct speech as summary (which allows him to add comments on the earl's state of mind, demonstrating his genuine concern for his wife) and splits the dialogue into three sections (three stanzas), each of which contains a solicitous enquiry by the earl and a deceitful response by his wife. An effect of gradation is achieved, and the audience's indignation at the countess's perfidy grows from stanza to stanza.

The only other example of a single laisse that is amplified in translation to three stanzas is laisse 5. Though only seven lines long, it is packed with information important for both plot and characterization, as it offers us an initial glimpse of the hero's mother and sketches out the background to her marriage:
 La dame si estoit bele e afeite. pis maide ichaue of
 y-told--
 Faire maide
 zhe was & bold,
 And fre y-boren.

 Le emperur de Alemaine la out avant ame Of Almayne pat
 Emperur
 Hire hadde
 loued paramur
 Wel par-be-foren.

 e a son pere le out sovent demaunde, Ofte to hire fader a
 sente,
 And he him-selue
 peder wente,
 For hire sake.
 Ofte a ernede
 hire to wiue:

 mes lui roi de Escoce li avoit devee pe King, for no ping
 aliue,
 Nolde hire him take.

 si la dona Guioun ov la chere membre. Sipe a zaf hire to
 Sire Gii--
 A stalword
 erl and hardi,
 Of Souphamtoun.

 Pus en perdi le chef (alias, quele destine!) Man, whan he fallep
 in-to elde,
 pur la amour de la dame que il out espose. Feble a
 (lines 24-30) wexep and vnbelde,
 Pourz rizt resoun.
 (A lines 31-48)


The translator omits the narrator's comment in the final line of the laisse, having used a very similar formulation at the end of the stanza immediately preceding the quoted passage. Until then, however, he had consistently matched every line of Boeve with a half-stanza of his own. A parallel analysis of this laisse and its translation alone might misleadingly suggest that the translator tried to establish a pattern of correspondence wherein one alexandrine of the Boeve text is expanded to a triplet. The Anglo-Norman lines in the above example function, however, as both metrical units and units of meaning. As metrical units they do not carry enough weight to affect the primacy of the laisse as the determining factor in the formal organization of the Middle English translation; but as units of meaning they can disrupt the effort to match tail-rhyme stanzas to laisses. The laisse may be the main unit of translation in this section of the narrative, but it is a complex unit and its manipulation is facilitated by breaking it down into smaller components, sense units between one and three lines long. (37)

As in the above example, the translator normally tries to pack the content of an Anglo-Norman sense unit, no matter how long, into one English triplet (half-stanza). In the following example (laisse 46) he does this with sense units of different lengths in the original:
 Li emfes vint devaunt le emperur avis fer, Al aboute he gan
 hardiement comenca a parler: be-holde;
 To Pemperur he spak
 wordes bolde,
 Wip meche grame:

 'Entendez vers moi, beau duz sire cher, 'Sire, a sede, 'what
 ky vus dona conge cele dame acoler? dostow here?
 Whi colles pow aboute
 pe swire
 Pat ilche dame?

 EIe est ma mere, ne vus enquer celer, Me moder is pat pow
 e kaunt a moi ne volez conge demaunder, hauest an honde.
 jeo vus frai sa amur mou cher achater; What dostow her vpon me
 rendez moi ma tere, ieo vus voil loer. londe,
 (lines 289-96) Wip-outen leue?
 Tak me me moder and mi
 fe:
 Boute pow pe raper
 hennes te,
 I schel pe greue!
 (A line 421-32)


The laisse can be roughly divided into four syntactically self-contained statements: (1) Boeve addresses Doun (two lines); (2) 'Who gave you the right to embrace this woman?' (two lines); (3) 'She is my mother, and if you want her you must ask for my permission' (three lines); (4) 'Give me back my land!' (one line). Each of them becomes a half-stanza in Middle English. While some of the translator's changes cannot be explained by reference to segmentation only (for example, the increased stress on the usurpation of land), others are almost certainly owed to the translator's concern not to lose step with his original. The final two lines (464f.), in which the 7-year-old hero threatens his stepfather with unspecified harm, are felicitously chosen and add a touch of comedy to the scene, but their primary function was to fill out the stanza, not to achieve a comic effect. They are inspired padding--but padding nevertheless.

The matching of sense units is a general tendency, not a strict rule, and whenever other constraints (most often metre and rhyme) prove to be more demanding, the translator will not hesitate to reorder and recombine the shorter sense units within a laisse rather than sacrifice the sense of a laisse as a self-contained unit, to be translated by one or two (or occasionally three) equally self-contained units.

Respect for laisse boundaries is considerable, but not unconditional. There are times when inspired padding is hard to come up with and mindless padding is something the translator generally tries to avoid, at least in the initial section of the poem. At such times he departs from his basic matching pattern and renders one Anglo-Norman laisse as one and a half tail rhyme stanzas. (38) But even when he is forced to do this he is anxious to return to stricter one to-one matching as soon as possible. As shown in the table below, these cases of out of-step matching are kept close together, so that three adjacent tail-rhyme stanzas normally preserve the substance of two adjacent laisses; (39) we still feel that the integrity of the laisse as a narrative unit is of great importance and that the laisse is instrumental in binding the translation to its original.

The validity of this general principle is negatively illustrated at the point where all manuscripts except S and N abandon the tail-rhyme stanza and switch to couplets. Whoever was responsible for the redaction preserved in S and N tried to maintain the original versification for a while longer. (40) In so doing, however, the redactor upset the pattern of correspondences between laisses and tail-rhyme stanzas established by the original translator. Anglo-Norman laisses are now predominantly, rather than exceptionally, translated by one and a half tail-rhyme stanzas. As noted by Mehl (sec above), the author of the SN redaction begins by adding tail lines to the couplets of A. (41) After a few stanzas, though, we see him changing the couplets themselves more and more substantially, until he diverges entirely from both AEC and Boeve, adding three and a half stanzas with new narrative material (SN lines 509-28). (42)

Whatever the reason for the delayed change of verse form in SN, it is clear, I think, that the initial choice of the six-line tail rhyme stanza in the earliest Middle English Beves was prompted by the versification of Boeve and that this choice made the segmentation of the text for the purpose of translation much easier. This kind of effort to match the verse form of the original is also the best explanation for the translator's eventual abandonment of the tail-rhyme stanza. When he switched to couplets it was not so much because he could not handle the tail-rhyme stanza. He may not be the best versifier in this particular form, but he is far from the worst. Rather, the looming replacement of short laisses with long, and even very long, ones suddenly left him unanchored, having to puzzle out the problem of segmentation anew. (43) He did the best thing under the circumstances and decided to change to a verse form that still allowed him the approximate matching of manageable units, except that the units were now shorter: from now on, in passages of close translation the tendency will be for one Middle English couplet to be roughly equivalent to one or two Anglo-Norman lines. Once the restrictions involved in matching stanza for laisse were no longer operative, however, other constraints weakened too, so that a few sections of fairly close translation are followed by more serious divergences between Boeve and Beves. Henceforth the translator also takes greater liberties in ordering his material within longer episodes and no longer feels obliged to follow the expository order of the Anglo-Norman text. The rearranging of the narrative building blocks in the first encounter between Boeve and Hermine, an early indication of the change in the translator's approach, (44) is followed almost immediately by the translator's first major interpolation, the hero's Christmas Day battle against the Saracens. (45)

According to the evidence of the extant manuscripts, Beves is one of the rare Middle English romances featuring the six-line tail-rhyme stanza. It is by far the earliest of the few romances to use this verse form, and the structure of its stanza differs from the others in having very short two-stress tail lines. Though early, it failed to start a trend: the overwhelming majority of later tail-rhyme romances use versions of the twelve-line stanza. Boeve is itself unusual in that it uses, in the initial section, short monorhymed laisses of a kind that we do not encounter elsewhere in Anglo-Norman romance). (46) This intriguing similarity between the two texts is more than mere coincidence: an analysis of the translation procedure shows conclusively that the versification of the original must have acted as a major constraint on the work of the translator. (47)

NOTES

I would like to thank Dr Jennifer Fellows for her valuable comments on a draft of this article, and the anonymous readers for Medium. AEvum for their helpful and constructive suggestions. The article was prepared with the financial support of the Social Sciences and Humanities Research Council of Canada, which I here gratefully acknowledge.

(1) There is a useful summary of critical views on the versification shift in the Auchinleck Guy in Paul Price, 'Confessions of a godless killer', in Medieval Insular Romance." Translation and innovation, ed. Judith Weiss, Jennifer Fellows, and Morgan Dickson (Cambridge, 2000), pp. 93-1 10 (pp. 97-100). For Ferumhras see H. M. Smyser, 'Charlemagne legends', in A Manual of the Writings in Middle English 1050-1500, ed. J. Burke Severs (New Haven, Conn., 1967), pp. 80-100 (p. 85), and, for a more cautious attitude to the hypothesis of a change in exemplar, Stephen H. A. Shepherd, 'The Ashmole Sir Ferumbras: translation in holograph', in The Medieval Translator: The Theory and Practice of Translation in the Middle Ages, ed. Roger Ellis (Cambridge, 1989), pp. 103-21 (pp. 107f.).

(2) The five manuscripts I shall be referring to are the following: National Libra," of Scotland, MS Advocates' 19.2.1, also known as the Auchinleck MS (A); British Library, Egerton MS 2862 (S), also known as the Sutherland or Trentham MS; Naples, Biblioteca Nazionale, MS XIII.B.29 (N); Cambridge, Gonville and Caius College, MS 175 (E); and Cambridge University' Library, MS Ff.II.38 (C).

(3) The Anglo-Norman text will be quoted from Der anglonormannische Boeve de Haumtone, ed. Albert Stimming, Biblioteca Normannica 7 (Halle, t899), and the Middle English from The Romance of Sir Beues of lqamtoun, ed. Eugen Kolbing, EETS, ES 46, 48, 65 (London, 1885-94; repr. 1973). Kolbing's edition is very inconvenient to use when referring to manuscripts other than A, but remains the only one available until the publication of Jennifer Fellows's eagerly awaited edition of the N(S) and C texts.

(4) All extant Anglo-Norman romances of that period are in verse; the first prose romances in Middle English date from no earlier than the mid-fifteenth century. There is no parallel in England to the near-immediate translation of French thirteenth-century prose romances into German, partly because there was no need for such translation in the literary system of thirteenth-century England: those with a taste for French romances, verse or prose, would have been able to read them in French. It is also true, though, that the earliest French prose romances are all Arthurian, and thirteenth-century insular literature, with its distinctly anti-Arthurian bent, would have provided a hostile environment for their domestication in either French of English. For a sketch of the beginnings of prose romance in England see Helen Cooper, 'Romance after 1400', in The Cambridge History of Medieval English Literature, ed. David Wallace (Cambridge, 1999), pp. 690-719 (pp. 691f.). On the anti-Arthurianism of Anglo Norman literature see Rosalind Field, 'The Anglo-Norman background to alliterative romance', in Middle English Alliterative Poetry and its Literary Background, ed. David Lawton (Cambridge, 1982), pp. 54- 69 (pp. 64f.) or, by the same author, 'Romance in England', in The Cambridge History of Medieval English Literature, ed. Wallace, pp. 152-76 (PP. 163f.).

(5) Sir Beues of Hamtoun, ed. Kolbing, p. xi.

(6) Derek Pearsall, 'The development of the Middle English romance', Mediaeval Studies, 27 (1965), 91-116 (pp. 91f.). While Pearsall is too ready to dismiss the Middle English translator as a fifth rate hack, Carol Fewster, Traditionality and Genre in Middle English Romance (Cambridge, 1987), pp. 46-9, may be too ready to ascribe to him a remarkable degree of subtlety. Fewster's claim that 'the change of metre is at a significant point, and helps to suggest a different literary structure for the story' by establishing Beves as the hero of the romance after two potential heroes have been rejected is rather strained and takes no account of Boeve, the Anglo-Norman original.

(7) Dieter Mehl, The Middle English Romances of the Thirteenth and Fourteenth Centuries (London, 1969), p. 212.

(8) Sir Beues of Hamtoun, ed. Kolbing, p. xi.

(9) Albert C. Baugh, 'Improvisation in the Middle English romance', Proceedings of the American Philosophical Society, 103 (1959), 418-54 (p. 431). Baugh's account is not entirely accurate: although the average length of the laisses in this section of Boere is slightly over six lines (6.28, to be precise) there is more divergence than he suggests. There are five laisses of eight lines, and one each of ten and twelve lines. Nevertheless, the overall effect is indeed of 'a poem written in six-line monorimed stanzas'. Baugh reiterated this argument, in much the same terms, in 'Convention and individuality in the Middle English romance', in Medieval Literature and Folklore Studies: Essays in Honor of Francis Lee Utley, ed. Jerome Mandel and Bruce A. Rosenberg (New Brunswick, NJ, 1970), pp. 12346 (p. 126) and, in somewhat greater detail, in 'The making of Beves of Hampton', Library Chronicle, 40 (1974), 15-37. In this last he claimed, paradoxically, that the author of Beves was able 'to maintain the parallel between laisse and stanza' only because 'he is not attempting a translation' (p. 18). It is obvious, however, that Baugh had in mind a very restrictive definition of translation, one that has since been abandoned by students of medieval translation (and increasingly of post-medieval translation too).

(10) Brian Joseph Levy, 'The ancestral romance in mediaeval French with special reference to Anglo-Norman literature' (unpub. Ph.D. diss., University of Edinburgh, 1966), p. 516.

(11) Jennifer Loney Fellows, 'Sir Beves of Hampton: study and edition', 5 vols (unpub. Ph.D. diss., UniversiD, of Cambridge, 1980), I, 86.

(12) Ibid., I, 89.

(13) Linda Marie Zaerr recently related the change in the versification of Beves to oral performance in 'Meter change as a relic of performance in the Middle English romance Sir Beues', Quidditas, 21 (2000), 105-26. Zaerr's dismissal of the Anglo-Norman text's relevance to the problem of metrical variation in Beues (pp. 113f.) is surprising in what is otherwise a thoughtful and carefully argued article.

(14) Judith Weiss, 'The date of the Anglo Norman Boeve de Haumtone', MAE, 55 (1986), 237- 41 Der anglonormannishe Boeve de Haumtone, ed. Stimming, pp. lvii-lviii.

(15) MS Firmin Didot (D) perishcd in the bombing of Leuven in the Second World War. On its fate see Judith Weiss, 'The Anglo-Norman Boeve de Haumtone: a fragment of a new manuscript', Alodern Language Review, 95 (2000), 305-10 (p. 305). Stimming (Der anglonormannische Boeve de Haumtone, p. ix') dated MS Paris, Bibliotheque nationale de France, MS nouv. acq. fr. 4532 (B) to the beginning of the fourteenth century, which is probably correct, though some letter forms would appear to suggest a slightly earlier date.

(16) See below, n. 45.

(17) See Ivana Djordjevic, 'Mapping medieval translation', in Medieval Insular Romance, ed. Weiss et al., pp. 7-23, and, especially, 'Mapping medieval translation: methodological problems and a case study' (unpub. Ph.D. diss., McGill University, 2002), pp. 81-9.

(18) When slightly different forms of the same passage occur in several manuscripts, it is usually not impossible to figure out how individual readings may have evolved through scribal transmission.

(19) It has been plausibly suggested that the six-line stanza antedated the more familiar twelve-line form. See, for example, Maldwyn Mills, 'Sir Lsumbras and the styles of the tail-rhyme romance', in Readings in Middle English Romance, ed. Carol M. Meale (Cambridge, 1994), pp. 1-24 (p. 24 and n. 42). Indeed, with the sole exception of the Ashmole Sir Ferumbras (c. 1380), in which the six-line stanza is used only in the second half of the poem, the few texts in which the six-line stanza appears are either very early or very late. Mills suggests (pp. 23f.) that such late examples, dating from the late fifteenth and early sixteenth centuries (they include the short romances The Jeaste of Syr Gawayne, The Weddyng of Sir Gawen, The Grene Knight, and The Turke and Gowin), represent a brief revival of the old verse form. The earliest examples, including Beves, may have been based on Anglo Norman models (see below, n. 21).

(20) Apart from Sir Tristrem's unique eleven-line stanza, the standard verse forms in the romances collected in A are the short couplet (Sir Degare, Floris and Blauncheflur, the first part of Guy of Warwick, Of Arthour and of Merlin, Lay Le Freine, Otuel, Kyng Alisaunder, Sir Orfeo, and Richard Coer de Lion, which begins with two twelve-line tail-rhyme stanzas) and versions of the twelve line tail-rhyme stanza: aabaabccbddb (pe King of Tars, Amis and Amiloun, the second part of Guy of Warwick, Reinbrun, Horn Childe) and aabecbddbeeb (Roland and Vernaw). When the six-line stanza occurs, it is reserved for lyrics or for religious narratives (St Patrick's Purgatory', The Assumption of the Blessed Virgin, Hou our leuedi saute was ferst founde, The Thrash and the Nightingale, The Sayings of St Bernard). Of the romances in other Beves manuscripts, only Sir Eglamour, preserved in S and C, contains some six-line tail-rhyme stanzas, but their stress pattern is different.

(21) In a recent conference paper, Rhiannon Purdie drew attention to the use of tail-rhyme sizains in Anglo Norman narratives of moral or spiritual instruction, notably Beneit's La Vie de Thomas Becket ('Beues: the knight in shining white stanzas', in Romance in Medieval England, 8th Biennial Conference, Durham Castle, Durham, 18 April 2002). With its two stress tail lines, the anomalous Bares stanza ([a.sup.4] [a.sup.4] [b.sup.2] [c.sup.4] [c.sup.4] [b.sup.2]) matches the stress structure of the Anglo-Norman sizain ([a.sup.4] [a.sup.4] [b.sup.2] [a.sup.4] [a.sup.4] [b.sup.2]) in a way that other six-line tail-rhyme stanzas in Middle English, with their predominantly three-stress tail lines, do not. Two other Anglo-Norman texts--a Desputeison de l'alme et du corps and a fragmentary parable, both tentatively attributed to Nicholas Bozon--are even closer to the versification of Bares, which they match not only in stress structure but in rhyme scheme as well. The translator of Bares could have imitated the versification of texts such as Beneit's Vie or Bozon's [?] works while responding to problems posed by a different French verse form: the short monorhymed laisse. It is also possible, however, that the translator had in mind English models, the stress structure of which he modified to suit his own purposes. None of those Middle English texts was a romance. As in the Auchinleck MS, they were either lyrics or narrative poems with a strong religious bent. The only exception is Dame Sirith, one of the rare fabliaux in Middle English, written in a mixture of metres, including six line tail rhyme stanzas with a fluctuating number of stresses.

(22) For a still useful general introduction to the problem of segmentation see L. G. Kelly, The True Interpreter A History of Translation Theory and Practice in the West (Oxford, 1979), pp. 120-6. As Kelly points out, 'the act of translation begins from assumptions about the unit of translation', in an illuminating article on Guy of Warwick ('Techniques of translation in the Middle English versions of Guy of Warwick', in Medieval Translator 2, ed. Roger Ellis (London, 1991), pp. 209-29), Maldwyn Mills identifies the same need for manageable units for the translator to work with in another translation from Anglo-Norman, but one that involved turning octosyllabic couplets into twelve-line tail-rhyme stanzas: 'before proceeding to any detailed rendering of the words of the Anglo-Norman, the translator would first have broken down that source into passages, each sufficiently coherent in its subject matter to serve as the basis for an effective tail-rhyme stanza' (p. 221). See also Roger Ellis, 'The choices of the translator in the late Middle English period', in The Medieval Mystical Tradition in England, ed. Marion Glasscoe (Exeter, 1982), pp. 18-46 (p. 36).

(23) Even in modern translational practice it hardly ever is, except in highly technical texts. There may be a fair amount of overlap between segments and of blurring at the edges, hut the important thing is that the outline remains clear. As long as we can tell that a particular segment of the Middle English text corresponds to a particular segment of the Anglo Norman, the segments need not be delimited with the utmost precision.

(24) Kirsten Malkmjaer, 'Unit of translation', in Routledge Enocyclopedia of Translation Studies, ed. Mona Baker (London, 1998), p. 286.

(25) There are many differences in wording between A and C, but the way that Anglo-Norman laisses are matched by tail-rhyme stanzas is substantially the same in both. N and S can be considered identical for most practical purposes. Nothing definite can be said about E because of a sizeable lacuna in this section; what remains of the text indicates that initial kinship with the S redaction very quickly gives wav to a close alignment with A.

(26) The versification of Boeve being far from regular, not all lines are perfect alexandrines: the number of syllables varies between ten and thirteen.

(27) The matching section of C, from which four stanzas found in related redactions are missing, is 556 lines long. It is worth noting that comparison of the number of lines in an original text and its translation is a very, misleading indicator of the extent of the translator's intervention if the length of the lines is not taken into account. A single line of Boeve is often nearly as long as an entire triplet (three line half stanza) of Beves, and even when the Middle English text eventually switches to couplets they will be composed of shorter lines: octosyllabics, as opposed to the Anglo-Norman alexandrines.

(28) For a detailed presentation of the correspondences between Anglo-Norman short laisses and Middle English stanzas in this section of the text see the appended table.

(29) Scribes and redactors may have different views as to what constitutes essential information, however. In manuscripts SN and E, the opening stanza is followed by one that does not appear in either A or C:
 He was a stalworth man,
 And many kyngdomes wan
 To Goddes lawes.
 He was De best pat com in felde,
 And most wan with spew and shelde
 By his dawes. (S line 6)


This might be interpreted as a later redactor's embellishment, were it not for the obvious echoes (here italicized) of line 4 of the opening stanza of Boeve. Most likely, the stanza preserved in SE was part of the ur-Beves but was left out of A and C, for whatever reason. Alternatively, the stanza may have been suppressed at an early stage in the textual transmission and reinserted by a later scribe who had access to Boeve and felt that an important piece of information had been left out of the AC version. The manuscript tradition is too complex to allow for anything beyond mere guesswork. But if someone added the stanza he must have been familiar with Boere--and convinced that here was an omission to be rectified.

(30) The procedure was somewhat facilitated by the fact that the Anglo Norman author was not averse to using conventional line fillers for the sake of rhyme or assonance--a common feature of the chanson de geste, the metrical form of which was adopted by the author of Boeve. The presence of chevilles lightened the narrative load of individual lines; moreover, they could easily be disregarded in translation. Some of the conventional expressions in Boeve are context specific and work best in narrative situations of a particular kind (combat, journey, conversation, description, etc.), while others could be used virtually anywhere. 'Pur veritez', 'ben le sachez', 'sachez de veritez', 'he le quidez', 'com dist fi escris', 'com oyer purrez', 'ja nevus err cele', 'nent vus estut doter', 'a men escient', 's'e verite prove', and other expressions of this kind are used liberally by the Boeve author.

(31) Beves lines 133-8. This laisse, in which a messenger sent by the hero's mother delivers a message to her lover Doun (lines 82-8), merely repeats the countess's own words to the messenger (lines 53-9), although with some changes in wording. The thrifty translator simply copied verbatim his translation of the original message (lines 91-6), making only the necessary grammatical changes: from present tense to past and from third person singular to second: 'bid [him]' (line 91) thus became 'bad pe' (line 133) and 'pat he be to fizte prest, / Wip is ferde ...' (lines 94f.) became 'pat ze be to fizte prest, / Wip zour ferde ...' (lines 136f.). Three other laisses that are translated by a single stanza also consist of the delivery of messages that the audience has already heard and does not really need to hear again (laisses 11, 15, 32).

(32) The translator occasionally mimics the laisse-linking repetitions (for example, lines 170f. in Boeve, which connect laisses 26 and 27, are paralleled in lines 240f. of Beves, which likewise link two stanzas), but more often he simply elides the repeated lines. For another example of laisse-linking lines preserved in a Middle English translation, see Shepherd, 'The Ashmole Sir Ferumbras', p. 109. Shepherd suggests that this is a feature of what he calls 'the first state of translation', and it is possible that the author of Beves suppressed some initially retained linking repetitions when he revised his own translation.

(33) I have broken up the laisse to show more clearly the close correspondence between the Anglo-Norman text and its translation.

(34) On the inappropriateness of the promise see Fellows, 'Sir Beves of Hampton', V, 5. Other examples of laisses consisting mostly of speeches expanded to two stanzas include laisses 39f. (four tail-rhyme stanzas in Beves, lines 355-78) and laisse 46, which I shall examine shortly. Examples of composite laisses in which speeches are combined with action too important to be abbreviated are 23, 27, 28, 33-6 (in which the hero makes his first appearance), 49, and 50. They too are expanded to two tail-rhyme stanzas in translation.

(35) It is unfortunate that he could not find a more convincing tail line with which to conclude this little tirade. 'Er he hom wende', a tag that works well in different contexts, is quite ludicrous here and rather spoils the desired effect. Elsewhere the translator shows greater skill in the amplification of narratorial comments, for example in laisse 12. There he picks up a brief parenthetical remark, 'Lui messager s'en turne (deu lui doint mau jur!)' (Boeve, line 70) and gives it more prominence in a less obtrusive way, by turning it into a running commentary on the messenger's journey:
 Now pat masager him gop--
 pat ilche Lord him worpe wrop,
 pat him wrouzte!
 To schip pat masager him zode;
 Allas, pe wind was al to gode,
 pat him ouer brouzte! (A lines 109-14)


(36) Boeve, lines 126-37; Beves, lines 181-98.

(37) Kelly compares 'the question of unit' to a 'Chinese Box puzzle, with units of larger relevance giving way to smaller' (The True Interpreter, p. 120).

(38) Ten laisses are translated by one and a half stanzas in A, thirteen in SN. Two extreme examples are laisses 7 (eight lines) and 8 (six lines), translated respectively by three and a half and two and a half stanzas.

(39) Laisses 7-10, 17-20, 42-5. The single exception is the translation of laisse 30, which intrudes somewhat awkwardly between two one-and-a-half-stanza units.

(40) For another eleven tail-rhyme stanzas, corresponding to laisses 52-8 (lines 327-69) of Boeve.

(41) This is one of the reasons for the impression that we have of him as a much less skilled versifier than the original translator. The tail lines he adds are mostly meaningless adverbial tags of the kind that have given the tail-rhyme stanza a bad name: 'pat ilke day' (SN lines 476, 497), 'with-out delay' (line 488), 'sone a-noon' (line 492), 'ryzt hastely' (line 508). The original translator used tags too, but less often. Besides, he frequently managed to integrate them into the stanza in such a way as to actualize their adverbial meaning.

(42) As I suggested above (n. 23), the fact that some of the lines in SN are actually closer to Boeve than the corresponding lines in A indicates that the SN version may be the work of a redactor who had access both to the Anglo-Norman text and to an earlier translation, close to the A redaction. Compare, for example, line 326 of Boeve, 'Sabot le amena en une chaumbre musser', with the two Middle English versions: 'Saber Beues to his hous ladde' (A line 475) and 'Saber in to a chamber him lad' (SNC).

(43) There could be a simple explanation for the fact that the change of verse form in Beves occurs earlier than it does in Boeve. Namely, the translator may have got to the end of a right-hand page of his exemplar, turned the leaf, and found out that the familiar pattern of more or less regular six-line blocks of verse was no longer there. (We do not have his exemplar, but it is likely that in it, as in most manuscripts containing texts written in laisses, the habit of marking the beginning of each laisse with a coloured initial would have made the metrical scheme immediately visible.) Realizing that the original could no longer guide him, he gave up trying to match its versification without waiting to reach the end of the short-laisse section of Boeve. Incidentally, the Anglo-Norman text in manuscript B, the only one in which the opening of Boeve has been preserved, could not have been the source of the ur-translation. Not only was it never completed (its scribe seems to have been working from an incomplete exemplar) but it shows no understanding of the laisse as a unit of versification. Contrary to what the careful layout of Stimming's edition and his introductory comments (p. iv) might suggest, out of the 114 laisses in the section of the text it preserves B identifies the beginnings of only forty seven by an initial (more precisely, by leaving a space for an initial). On the other hand, it once provides for an initial where it is entirely out of place (line 1046, at the end of a laisse instead of at the beginning of the next one).

(44) Boeve, lines 380-410; Beves, lines 527-76.

(45) Fellows notes that 'from this point onwards, the relations of the [Middle English] texts both to [Boeve] and to each other are much less close than hitherto' ('Sir Beves of Hampton', V, 28). Although she does not link the placing of the first interpolation with the change in verse form in Boeve, it cannot be a coincidence that precisely at this point the short-laisse section of Boeve gives way to an indiscriminate mix of laisses of widely differing lengths, starting with the twenty eight-line laisse 68. It is, I think, not too fanciful to imagine the translator, suddenly faced with a major disruption in the sequence of narrative segments of roughly the same size, deciding that since he clearly could no longer rely on the original for easy guidance, he might just as well improvise by adding material of his own. Judith Martin [Weiss] had suggested ('Studies in some early English romances' (unpub. Ph.D. diss., University of Cambridge, 1967)) that the author of Beres 'was impressed by the style of the opening of the Anglo Norman poem, and by the literary quality of the enfances, and made some attempts to imitate it. For his first 555 lines, he borrows more words from Boeve than in any other section of his poem, and diverges very little from events or even details in the Anglo Norman' (p. 114). Martin added in a note that 'the manuscripts of Belles are reasonably similar throughout the enfances, and only start to diverge widdely afterwards'. She did not refer to the versification of the original as a possible explanation for the closeness of the translation at this point, but in a recent personal communication she indicated that she now believes this to have been an important factor.

(46) The predominant verse form in Anglo-Norman romance is the octosyllabic couplet. Only Thomas's Horn and the Roman de toute chevalerie by Thomas of Kent were written in alexandrine laisses. The laisses of Horn are significantly longer than those of the first section of Boeve (11-37 lines in one of the two manuscripts and 18-44 lines in the other). In the Roman de route chevalerie, a version of the story of Alexander, there are several sections in which short laisses predominate, but this is mostly due to the nature of the material they contain (the form tends to be restricted to brief pseudo-documentary descriptions of the marvels Alexander encounters on his expeditions). In Old French chansons de geste short laisses do occur, but always mixed with longer ones. In the Chanson de Roland, for example, laisses range from five to thirty lines; the average length is fourteen. In Rater texts laisses tend to grow longer. I am not aware of long sequences of short laisses such as we find in Boeve.

(47) Before generalizations on the problem of segmentation in translated Middle English romances can be made with some confidence, more work remains to be done on individual texts. I have here offered some evidence of the way that Beves lends itself to this kind of analysis. For very useful contributions to the study of related problems see, for example, the articles by Maldwyn Mills and Stephen H. A. Shepherd quoted above or the informative sketch in Rhiannon Purdie's introduction to her recent edition of Ipomadon, EETS, os 316 (Oxford, 2001), pp. lxiv-lxx.

(48) The left hand column lists the laisses of Boeve, with the number of lines in each laisse given in parentheses. The remaining four columns indicate how many stanzas in the manuscripts of Beves match a particular laisse. As MSS S and N are virtually identical for the purposes of the present study, I have not treated them separately.

(49) This segment of C contains an anomalous nine-line stanza.

(50) The first of many lacunae in E begins here.

(51) The second stanza is anomalous: the tail lines have been omitted, leaving two couplets.

(52) In all manuscripts except S, material from laisses 28 and 29 is mixed up in this and the following segments. Only S preserves the expository order of Boeve. Fellows believes this to be 'probably coincidental, since the apparently accidental transposition of lines is of not infrequent occurrence in this MS ... and since ANC here reads better--because in a less clumsily repetitive manner--than does S' ('Sir Beves of Hampton', V, 14).

(53) Or, more accurately, the second half of one stanza and the first half of the next one.

(54) The discrepancy between the manuscripts at this point is hard to explain. According to Fellows, the episode of Beves's conflict with the porter 'has become unaccountably garbled in the ]Middle English]' ('Sir Beves of Hampton', V, 18).

(55) Manuscripts AEC switch to couplets at this point.

(56) The second half of one stanza and the first half of the next one.

(57) The Middle English text begins to drift away from the Anglo-Norman. In the ten lines of AEC, Boeve material (an account of the hero's passage over the sea and his arrival in Hermine's kingdom) is freely moved around and combined with some additions. Of the four stanzas in SN, three contain material that is not found in either Boeve or the other Middle English manuscripts.

(58) The author of the SN redaction must have been disoriented by the change from tail-rhyme stanzas to couplets. Otherwise he would hardly have left out a passage as important as this: the audience's first glimpse of the heroine, who is here briefly described.

(59) The substance of the Anglo-Norman laisses is preserved in translation. There are a number of close verbal resemblances between Boeve and the Middle English manuscripts, but the various segments in this episode (an exchange between Hermine and Bevis) have been shifted around. The two lines added to SNC are not based on Boeve.

IVANA DJORDJEVIC

Dalhousie University

Halifax, Nova Scotia
The short-laisse section of 'Boeve' and the
corresponding section of 'Beves' (48)

Boeve (MS B) A SN E C

1 (6) 1 2 2 1
2 (6) 1 1 1 1
3 (5) 1 1 1 1
4 (6) 1 2 2 2
5 (7) 2 3 3 3
6 (6) 0 0 0 0
7 (8) 3 1/2 3 1/2 3 1/2 3 1/2
8 (6) 2 1/2 2 1/2 2 1/2 2 1/2 (49)
9 (6) 1 1/2 1 1/2 1 1/2 1 1/2
10 (7) 1 1/2 1 1/2 1 1/2 1 1/2
11 (6) 1 1 1 0
12 (5) 2 2 -- (50) 2
13 (7) 2 2 -- 2
14 (7) 1 1 -- 1
15 (6) 1 I -- 1
16 (6) 2 2 -- 2
17 (6) 1/2 1/2 -- 1/2
18 (5) 1/2 1/2 -- 1/2
19 (6) 1 1/2 1 1/2 -- 1 1/2
20 (8) 1 1/2 1 1/2 -- 1 1/2
21 (12) 3 3 -- 3
22 (5) 2 2 -- 2
23 (5) 2 2 -- 2 (51)
24-5 (7+6) 1 1 -- 1
26 (10) 2 2 -- 2
27 (6) 2 2 -- 1
28 (6) 2 2(52) -- 3
29 (6) 1 1/2 1 1/2 -- 1 1/2
30 (6) 1(53) 1 -- 1
31 (6) 1 1/2 1 1/2 -- 1 1/2
32 (6) 1 1 -- 1
33 (6) 2 2 -- 2
34 (7) 2 2 -- 2
35 (8) 2 2 -- 2
36 (6) 2 2 -- 2
37 (5) 1 1 -- 1
38 (6) 1 1 -- 1
39 (6) 2 2 -- 2
40 (7) 2 2 -- 2
41 (6) 1 1 -- 1
42 (6) 1 1/2 1 1/2 -- 1 1/2
43 (6) 1 1/2 1 1/2 -- 1 1/2
44 (8) 1 1/2 2 1/2 -- 0 (54)
45 (7) 1 1/2 1 1/2 -- 1
46 (8) 2 2 -- 2
47 (6) 1 1 -- 1
48 (6) 1 1 -- 0
49 (6) 2 2 -- 2
50 (6) 2 2 -- 2
51 (6) 1+2 lines (55) 1 1/2 1+2 lines 1+2 lines
52 (6) 4 lines 1 (56) 4 lines 4 lines
53 (7) 6 lines 1 1/2 6 lines 6 lines
54 (6) 6 lines 1 1/2 6 lines 6 lines
55 (5) 8 lines 1 1/2 8 lines 8 lines
56 (5) 6 lines 1 6 lines 6 lines
57-8 (6+6) 10 lines (57) 4 10 lines 10 lines
59 (4) 6 lines 0 (58) 6 lines 6 lines
60 (6) 8 lines 4 lines 8 lines 8 lines
61-3 (6+7+6) 30 lines (59) 32 lines 32 lines 30 lines
64 (7) 8 lines 8 lines 8 lines 8 lines
65 (5) 8 lines 8 lines 8 lines 8 lines
66 (5) 4 lines 4 lines 4 lines 4 lines
67 (9) 4 lines, which correspond to lines 1-4 of
 the laisse. The second half of the laisse
 introduces the hero's fight with the boar,
 but the Middle English translator here
 inserts his first interpolation, the
 Christmas fight.
68 (28)
COPYRIGHT 2005 Society for the Study of Mediaeval Languages and Literature
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2005 Gale, Cengage Learning. All rights reserved.

 Reader Opinion

Title:

Comment:



 

Article Details
Printer friendly Cite/link Email Feedback
Author:Djordjecic, Ivana
Publication:Medium Aevum
Geographic Code:4EUUK
Date:Mar 22, 2005
Words:10679
Previous Article:A local source for Horn Child and Maiden Rimnild.
Next Article:Chaucer's Book of the Duchess: a proposal.
Topics:


Related Articles
The Matter of Identity in Medieval Romance.
The polyphony of Jewish culture.
Thomas Maillet (?): Les Proverbez d'Alain.
Cultural Encounters in the Romance of Medieval England.
Le Chant des Chanz.
Matthew Paris: The History of Saint Edward the King.

Terms of use | Copyright © 2014 Farlex, Inc. | Feedback | For webmasters