University of Virginia Library

Search this document 


expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
collapse section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 

expand section 

In a recent article in Studies in Bibliography, D. C. Greetham demonstrated how a medieval translation can be stemmatically reconstructed by consultation of the translation's source.[1] Greetham's argument is a persuasive one, and it is doubly important for it advances editorial theorizing about a type of text which, judging from the number of surviving examples and manuscripts, was enormously popular in the Middle Ages but which has not attracted the attention of most modern textual critics. Indeed, the tradition of editorial discussion of medieval English texts has in large part been a reflection of the so-called canon of English literature: certain works, authors and genres have been central to editorial theorizing, while many other works and genres have been all but ignored. Moreover, for the canon of editorial discussion, as for the canon of English literature, literary preference, not textual significance, has typically determined which texts merit inclusion. But whatever the aesthetic value of romances and narrative poems, to which most textual critics have addressed themselves, if the unique character of the medieval conception of a text is to be determined, the hundreds of translations, commentaries, and religious and scientific treatises need also to be considered. Editorial discussion of the textual complexity of these works


Page 189
will provide a complementary insight into the textual identity of medieval literature and thus may in turn indicate refinements in current editorial procedures. Accordingly, in this paper I will examine the textual complexity which confronts the editor of still another type of translation. Greetham's editorial method is not applicable to this type, represented by Chaucer's Boece, because of the problems attendant upon the translation's sources, the translator's techniques, and the scribal transmission of the text.

In the "Retraction" to the Canterbury Tales Chaucer labels "Boece de Consolacione" a "translacion," and the fact that the Boece is a "translacion" has informed both textual and critical judgment. Since Boethius wrote the Consolatio in Latin, it was reasoned, Latin—specifically the text preserved in ninth-century manuscripts—was what Chaucer translated. From this perspective, H. F. Stewart's assessment that the "inaccuracy and infelicity" of the Boece "is not that of an inexperienced Latin scholar, but rather of one who is no Latin scholar at all"[2] is not surprising. While an attitude like Stewart's may have been a factor in Skeat's editorial procedure,[3] their contemporary Mark Liddell, in his edition for the Globe Chaucer, acknowledged that Chaucer used Jean de Meung's French translation as well as the Latin original.[4] The textual situation of Chaucer's sources is still more complex, however, for it is recognized today that Chaucer's sources included the "Vulgate" Consolatio (a distinct redaction that developed in the twelfth and thirteenth centuries), a French manuscript which contained many idiosyncratic readings found in Besançon MS 434, the commentary of Nicholas Trevet, and perhaps the gloss tradition associated with Remigius of Auxerre.[5]

With these identified sources it might seem that editing the Boece would be patently easy: simply compare the Boece readings with their sources and choose the best reading. The troublesome words here, however, are in fact "sources" and "best," but it is with the first word that I am immediately concerned. While we have identified Chaucer's source texts, we have not identified Chaucer's source manuscripts, and the distinction between text and manuscript is of paramount importance when assessing medieval translations. Cambridge University Library MS Ii.3.21 has long been known to have a Vulgate Consolatio which contains a number of unique readings reflected in the Boece. But many other readings indicate that this manuscript is not identical to the one Chaucer used. At 1p4.114, for instance, most Boece manuscripts read "decrets and . . . Iugements"; the corresponding Vulgate reading in C. U. L. MS Ii.3.21 is "iudiciis," while the French reading is "juigemens."[6] The older Latin Consolatio, however, reads "decretis," and this reading must underlie Chaucer's "decrets." Similarly, Besançon MS 434 cannot be identical, or even very similar to, Chaucer's French manuscript, for at times, when Chaucer is clearly following the French, its readings are so confused as to be unintelligible. Moreover, the French manuscripts have been classified in two distinct branches, and though Chaucer's French source almost always follows the α branch of the French tradition—and Besançon MS 434 is an α branch manuscript—there are a handful of occasions when


Page 190
a b reading seems to underlie Chaucer's translation. Thus, at 1p5.47 all Boece authorities read "tribulaciouns"; the corresponding Latin is "tumultus" and the corresponding French of the α branch "turbacions." The b branch, however, reads "tribulacions."

My point here is that the certainty and objectivity of the one-to-one correspondence implied by "source" may sometimes be only apparent. Quite obviously the Boece had sources, and equally obviously we have a fairly good idea of what those sources were. But a "fairly good idea" is not the same as "exact knowledge." Chaucer's Latin text evidently had an unattested combination of Vulgate and traditional readings, while his French text presumably contained an unrecorded mixture of a, and perhaps b, readings. If we do not know the exact Latin and French manuscripts which Chaucer used, then there must be a degree of doubt about the source of every individual reading in the Boece, even if the Latin of C. U. L. MS Ii.3.21, for instance, would seem unequivocally to be the source of a given reading. And this degree of doubt is magnified by Chaucer's method of composition.

In composing the Boece, as I noted above, Chaucer used the Vulgate Consolatio, Jean de Meung's translation, Nicholas Trevet's commentary, and a few Remigian glosses. If one sits down with the manuscripts of these works—and for this purpose any of several manuscripts will do—and compares them to Robinson's or Skeat's editions of the Boece, one discovers that the Boece is a close translation in the sense that almost every phrase in the Boece is a translation from one of the four sources. But two additional points need to be made here. First, though Chaucer referred to his composition as a "translacion of Boece de Consolacione," his actual method involved much more than what Boethius wrote. In fact, Chaucer was translating from what might be called the Consolation tradition which accrued to Boethius's original text. To the Middle Ages, the Consolation was a work of moral philosophy, and then as now any intelligent and knowledgeable individual had a personal stake in moral philosophy. It is unremarkable, if not inevitable, that medieval readers grew less interested in the form of the Consolatio than in its content and that the content of "Boece de Consolacione" came to include a variety of reworkings of and commentaries and glosses on Boethius's thought.

The second point which needs to be made about Chaucer's use of his sources is that though the Boece is a close translation, it is a close translation of a "source" which exists only by implication. That is, Chaucer himself, as he was translating, in effect created his source by selectively combining portions of the tradition he called "Boece de Consolacione" in order to represent the content of Boethius as he understood it. While Chaucer's procedure has general trends, his movement from source to source is not inevitably predictable. In many cases, thus, along with the uncertainty of what Chaucer's sources read must go the uncertainty of which source text he was translating.[7]

Chaucer himself can add little certainty to the textual complexity of the Boece and its sources, for he does not state why or for whom he composed the translation. Moreover, there is good reason to believe that the Boece is


Page 191
only a penultimate draft (Techniques, pp. 111-124), and if this is the case, then the textual picture is clouded even more. Chaucer clearly was interested in understanding Boethius's ideas and in exploring language; his primary concern was not stylistic artistry (Techniques, pp. 85-110). Thus, although in the "Retraction" he does acknowledge "oure Lord Jhesu Crist and his blisful Mooder, and alle the seintes of hevene," it would have been impossible for him to have incorporated into the Boece the consciously authorial posture expressed in the well-known lines at the end of the Troilus:
And for ther is so gret diversite
In Englissh and in writyng of oure tonge,
So prey I God that non myswrite the,
Ne the mysmetre for defaute of tonge. (5.1793-96)[8]
One might well ask: Since to Chaucer "Boece de Consolacione" meant not the form of Boethius's Consolatio but the content of the Consolation tradition, would Chaucer have demanded the accurate preservation of his text? Or would he have expected, or at least not been surprised by, a similar emphasis on content among his readers, who had as much right as Chaucer did to attempt to express Boethius's ideas?[9]

Whatever Chaucer's expectations, the Boece authorities—ten manuscripts and two early printed editions which evidently derive from no longer extant manuscripts—embody a remarkable amount of scribal alteration. Assuming for the moment that the editor knows what Chaucer wrote, one is able to say that from the time it left Chaucer's pen the Boece was subject to complex transmission. In C. U. L. MSS Ii.1.38 and Ii.3.21, the two manuscripts which most consistently preserve what Chaucer probably wrote, as well as in several others which occasionally join them in readings, there is unexpectedly a cluster of Kentish dialectal forms in the middle of Book Two.[10] The language of these manuscripts is elsewhere East Midlands, and since it is improbable that Chaucer suddenly used Kentish forms in a non-poetic context, the implication would seem to be either that the archetype experienced a sudden and temporary change of scribes or that it was initially transmitted in a fragmentary fashion. In either case, one has reason to believe that Chaucer's text had undergone some scribal alteration before the earliest manuscripts.

Within these manuscripts, in any event, scribal alteration is clear. In some cases, isolating the scribe's contributions from Chaucer's can be quite simple. For example, a body of lexical glosses accrued to and was transmitted with the text of the Boece, and occasionally in the later manuscripts these glosses constitute doublets with the words they originally glossed.[11] The scribal readings are similarly easy to identify in Bodleian Bodley MS 797. The scribe of this manuscript consistently modernized syntax and lexicon—"forsaken" for "forleten" (1p6.65), for instance—and the more archaic forms found in other manuscripts are certainly Chaucer's.

But while it is often relatively easy to identify the basic lexicon of the Boece, it is sometimes quite difficult to identify by means of the source the authorial particle words in a group of variants, in part, perhaps, because


Page 192
Chaucer himself seems to have been inconsistent in his usage.[12] In a given clause, for example, it is virtually impossible to predict which preposition Chaucer used—if he did indeed use one. For instance, at 3m1.1-2 all authorities save one read "lat him first delivere it of thornes"; C. U. L. MS Ii.1.38 reads ". . . fro. . . ." The Latin is "liberat arua prius fructibus," while the French reads "il le delivre avant des espines." Elsewhere, Chaucer uses "delivere" with both "of" and "fro," and so the authorial reading here is problematic. The difficulty of determining whether Chaucer used a preposition at all is nicely indicated by the following two examples, where the existence of readings with and without prepositions as translations of the same source renders the authorial form moot. At 2m5.23 C. U. L. MS Ii.3.21 and British Library MS Add. 16165 read "mountaigne Ethna"; the remaining authorities read "mountaigne of Ethna." The corresponding Latin is "Aetnae," and the corresponding French "montaigne Ethna." At 2p6.5-6, however, all authorities save one read "mountaigne Ethna"; C. U. L. MS Ii.1.38 reads "mountaigne of Ethna." Here the Latin is again "Aetnae," but the French has an entirely different construction. These problems are multiplied when a series of prepositions is involved. For instance, at 3p2.34-35 C. U. L. MS Ii.1.38 alone reads "the entencioun of desiringes and of werkes"; the other authorities read ". . . and werkes." The Latin is "actuum uotorumque . . . intenti," and the French is "l'entencion des fais et desiriers," so that either English reading is, again, an acceptable and potentially authorial translation.

It is also nearly impossible to predict how many negatives Chaucer used in a verb phrase. At 2p4.2, for example, all authorities save one read "ne I may," which matches the Latin "nec . . . possum." C. U. L. MS Ii.3.21 reads "ne I ne may," which matches the French "ne je ne puis." But at 3p1.11-12 three authorities, including C. U. L. MS Ii.3.21, read "I am nat agrisen of hem"; two authorities, including C. U. L. MS Ii.1.38, read ". . . ne am . . ."; and four authorities read ". . . nam nat. . . ." All three of these readings would match Latin "non perhorresco" and French "je ne les redoubte . . . nient." These variations obviate any conclusions about whether Chaucer's negative constructions derive from the Latin, the French, or his own idiolect. Similar difficulties surround both pleonastic "that" after the conjunction "whan" and articles before many nouns. For instance, at 2m1.1 all authorities save one read "Whan"; the reading of C. U. L. MS Ii.3.21, which Skeat overlooks, is "Whan that." The Latin here is "cum" and the French "Quant." At 2p2.10, however, all authorities save two read "Whan that"; B. L. MS Add 16165 and Bodleian Bodley MS 797 read "Whan." The Latin and French are again, respectively, "Cum" and "Quant." The uncertainty about the use of articles is illustrated at 2m5.23, where C. U. L. MS Ii.3.21 and National Library of Wales MS Peniarth 393 read "of mountaigne"; the remaining authorities read "of the mountaigne." The Latin has no directly corresponding phrase, while the French reads "de la montaigne." At 2m8.16 all authorities save two read "knitteth sacrement"; B. L. MS Add 16165 and


Page 193
Bodleian Bodley MS 797 read "knitteth the sacrement." The Latin reads "sacrum . . . nectit," and the French "enlace le sacrement."

While in poetic texts, which have been the focus of editorial theorizing for medieval literature, editorial decisions in these matters can rely, sometimes controversially, on metrical criteria, we do not yet know enough about Chaucer's prose stylistics to say what forms he as a prose writer and translator preferred. Yet the structures just discussed are basic to the unity and character of any piece of prose. Doublets are also basic to the character of the Boece, and these present difficulties as well, for in many cases some of the manuscripts have only half the doublet. When both the single word and the doublet are contextually appropriate, Chaucer's reading is certainly not manifest, since even if we knew Chaucer's exact source, we do not know how he translated it. At 2p7.30-31, for example, C. U. L. MS Ii.3.21 and the authorities closely related to it read "thinken ye to manifesten your renoun"; C. U. L. MS Ii.1.38 and the authorities closely related to it read ". . . manifesten or publisschen. . . ." Either reading is appropriate for Latin "de peruulganda fama . . . cogitatis" and French "penéz vouz de vostre renommee monteplaier." Similarly, at 3m3.1 the first group of manuscripts reads "a river," while the second group reads "a river or a goter"; the corresponding French and Latin are, respectively, "un gort" and "gurgite." Here one might argue that "goter" is due to a perceived connection—the words are in fact etymologically unrelated—between it and French "gort." But, again, if one relies on the sources alone, one cannot know whether in this passage the perceived connection was made by Chaucer or a scribe who had access to the French text. Indeed, since some scribes clearly did have access to Latin and French texts, the variants often present several acceptable translations of both the Latin and the French with no a priori way of distinguishing among them.

Earlier I suggested that the logical way to edit the Boece would seem to be to compare it to its sources and choose the best reading. I then objected to the word "source"; now I would like to object to the word "best." "Best" in what way? The objective of an editor of the Boece is not to prepare an edition of an accurate translation of Chaucer's sources. Nor is it to prepare an edition of an artistically successful translation, for as I noted above, in the Boece, unlike in his poetry, Chaucer's primary concern was not literary artistry. The objective of the Boece editor is to recover Chaucer's particular version of "Boece de Consolacione," and this version is sometimes inaccurate and infelicitous. Many of the Boece scribes, as I implied earlier, were knowledgeable and intelligent. Like Chaucer, they were interested in the conglomeration "Boece de Consolacione"—the Consolation tradition—and they could produce readings which are aesthetically or semantically superior to Chaucer's. But the editor of the Boece, or of any similar work, cannot lose sight of the fact that the desired text is not always the same as the most pleasing one.[13]

Faced with uncertain sources and occasionally unadjudicatable readings,


Page 194
the editor of the Boece finally asks: Is there a text in this text? Most editorial problems are rooted in what Chaucer and his contemporaries understood by "Boece de Consolacione." The reference is not to the specific form and content which Boethius produced but to an undulating, living tradition which Boethius inspired. Though Chaucer did write one text, the nature of the Consolation and the latitude of Middle English prose led his scribal readers to alter the text according to their own perceptions of the Consolation tradition. Ironically, the fact that the Boece is a translation both eliminates many of the fundamental problems the editor of a narrative poem or romance confronts and creates many others he does not. In the editing of any translation the source text is not a panacea, for even if the exact manuscript source is known, we frequently still do not know how the translator would have proceeded or even if he would have translated accurately; a given Boece reading, for instance, may be due to a Chaucerian misperception—either of the form or of the meaning of a source word. In these circumstances, the modern critic can never know precisely what Chaucer's sources read or how he represented them. Moreover, the nature of Chaucer's sources restricts editorial emendation at the same time it invites it. For given the composition of the Boece, an individual reading may have arisen from any one of four sources. There may be several variants for the Boece reading and for each of the four source readings. It is also possible to conjecture a source reading to justify any one of the Boece variants; and it is possible to conjecture a Boece variant to justify any one of the source readings. Thus, an editor who employs even a mildly eclectic method in this textual situation opens up a Pandora's box of readings, whereby the Boece and its sources can be formed and reformed in any number of uncontrolled ways. Rather than a panacea, indeed, the source text can be every bit the blandishment Chaucerian metrics can be.

The textual situation of the Boece requires, then, a best-text method, and the two best texts, as was noted above, are C. U. L. MS Ii.1.38 and C. U. L. MS Ii.3.21; indeed, one or the other of these two manuscripts has served as the base manuscript for every modern critical edition of the Boece. In a sense, the former manuscript is clearly superior, for its text is often the more readable because its readings are often the easier. For example, in the reading at 2p7.30-31 (cited above), C. U. L. MS Ii.1.38 in effect glosses "manifesten," which modern dictionaries indicate was a rare word in Chaucer's day, with the more familiar "publisschen." Similarly, at 4p4.19 C. U. L. MS Ii.1.38 reads "power" for the unusual and rare verbal noun "mowinge," found in C. U. L. MS Ii.3.21 and all the other authorities. A thorough comparison of the textual qualities of the two Cambridge manuscripts is not possible here, but it can be said that it is the very textual refinement of C. U. L. MS Ii.1.38 which, given the way medieval readers tried continually to adapt and improve the Consolation in general and the Boece in particular, suggests that the text of C. U. L. MS Ii.1.38 embodies scribal editorialization of what Chaucer wrote. C. U. L. MS Ii.3.21 is the more authoritative manuscript, then, but it is


Page 195
certainly not without error, and in editing it an editor must of course draw on the evidence of Chaucer's sources. In a passage largely indebted to the French, for instance, the authorial reading in a given group of variants is likely to be that which most closely resembles the French. And the sources can aid in the correction of the order of a group of sentences which were jumbled in transmission or explain why a clause present in the Latin but not the French, for example, is also absent in the Middle English; thus, the clause which Skeat supplies from Caxton's edition at 4p6.188-189 and which does not occur in any of the manuscript authorities is almost certainly not Chaucer's but Caxton's translation of the Latin, for the corresponding French, which Chaucer follows throughout this passage, also lacks the clause. But in all doubtful cases priority must be given to the reading of the base manuscript as it stands; otherwise, if an editor of a text like the Boece employs Greetham's method, he runs the risk of producing a text possibly better than, and certainly different from, the one Chaucer wrote.[14]

By way of a conclusion I would like to stress two general observations which emerge from a discussion of the textual complexity of the Boece. The first observation is that the textual identity of a work of medieval literature is dependent both on what the author intended and on how the original audience approached the work. The Boece manuscripts are what they are because Chaucer's concern in the translation was primarily with meaning and language, not with literary artistry, and because for both Chaucer and his readers the Consolation was a living text which invited reader involvement. It is this textual identity which finally must determine the method an editor employs. The second observation which emerges from a discussion of the Boece is that a coherent picture of medieval textual attitudes can develop only from consideration of the many types of literature which existed in and were peculiar to the Middle Ages. If textual discussion is motivated by literary preference, many of the works which characterize the Middle Ages will remain only textual oddities. Indeed, the aesthetic quality of individual works is in fact irrelevant in the construction of significant textual generalizations about medieval literature. We need to explore and explain translations, commentaries and treatises as well as poems of love.