Editorial Assumptions and the Manuscripts of The Canterbury Tales

15 Editorial Assumptions and the Manuscripts of The Canterbury Tales Simon Horobin The dissenting voices of modern textual scholarship agree that the crucial issues in determining the authorial text of The Canterbury Tales are the nature of Chaucer’s foul papers, his attitude to the poem and his working methods. This debate focuses on the possibility of authorial revision, the circulation of individual tales during the poet’s lifetime, and ultimately the relationship between the Hg and El manuscripts. Yet modern scholarship has relied too much on the order of tales, the prejudices and preferences of previous generations, and editorial subjectivity to create complex theories about the early textual history of the poem. These hypotheses have subsequently largely determined the way in which any analysis of the text must begin. Editorial assumptions are employed as the basis for an interpretation of the textual evidence, which is subsequently used to reinforce the preliminary assumptions. The authority of the text, content and arrangement of the El manuscript, justified by an initial hypothesis of multiple copies of parts of the poem before Chaucer’s death, is not a priori proof of prior circulation of Canterbury Tales material. Since the work of Manly and Rickert (1940), which demonstrated the priority of the Hg text, much work on the manuscripts has attempted to reconcile their findings with the apparent authority of the El content and arrangement, with the result that research into the textual tradition has become almost exclusively focused on these two witnesses at the expense of the vast number of other potentially important manuscripts. In order to gain a more balanced view of the textual tradition, we must give greater weight to the texts of these many neglected witnesses, allowing their evidence to inform our interpretation of the early textual history of the poem. It would be unfair to claim that previous scholarship has ignored the text completely, especially given the unique contribution of the work of ManlyRickert in their 1940 edition of The Text of the Canterbury Tales. However the exhaustive corpus of variants and classifications provided by this edition, and the conclusions reached, have provoked many contradictory reactions. While scholars have generally accepted the prominence of the Hg text shown by this work, this has had little effect on subsequent editions, and Robinson’s edition of 1933 remained largely unchanged in his second edition of 1957, and much of this appeared again in the Riverside text of 1987. Manly-Rickert’s classifications were based upon collations made against Skeat’s Student Edition, relying heavily on editorial subjectivity to determine unoriginal readings. The manuscripts were then classified into groups according to agreement in erroneous readings, although the constant shifting of textual affiliation within groups defies any simple or significant analysis, and their constant groups regularly join to comprise groups composed of up to forty manuscripts, with a THE CANTERBURY TALES PROJECT OCCASIONAL PAPERS II, 1997, 15-21 16 Canterbury Tales Project Occasional Papers II highly complex system of interrelationships. Much of this is well-known and the classifications and the method of recension were criticised by contemporary reviewers, and more recently in articles by Blake and Kane (Root 1941; Blake 1983; Kane 1984.) However despite these widespread criticisms, the manuscript groups they assembled have been retained throughout the past 50 years and form the basis for much of the textual analysis in the most recent complete study of the manuscript tradition (Owen 1991.) Despite having rejected nineteenth-century attempts to reconstruct the framework of the poem according to geographical and temporal allusions, modern scholarship has drawn heavily on the evidence of tale-order in approaching these fundamental problems. This may be an important method of understanding the development of production of copies of the poem, yet we must not place complete faith in its ability to reveal the mysteries of the Chaucerian holograph. Questions of textual affiliation are not necessarily related to those of arrangement. Reliance on the order of tales is also problematic as the evidence may be used to explain diametrically opposed theories of the development of the tradition, depending on certain initial assumptions. This is exemplified in a comparison of the interpretations proposed by Blake and Benson (Benson 1981; Blake 1985.) Both scholars draw heavily on the evidence of tale-order, yet Blake concludes by arguing for the priority of Hg and Benson for El. It is important to realise that this is not simply the priority of their arrangements, but also of crucial questions of text and content. The cause of these great conflicts in interpretation stems from totally different assumptions about the state of the poem at Chaucer’s death. Blake rejects prior circulation of individual parts of the poem arguing for a single unfinished copytext in Chaucer’s possession, while Benson argues that Chaucer had finished work on the poem and had subsequently revised and released it by the time of his death. These examples reinforce the case for trying to establish objective, textual grounds for any interpretation of the pre-1400 situation. Thus in my opinion we must not be guilty of presupposing a chain of events leading to the production of our extant witnesses, but reverse the process: analysing the text first and using this analysis to understand the early history. This analysis must take account of a greater and more diverse quantity of witnesses, irrespective of external evidence and preconceptions of the development of the text. When we turn to the manuscripts themselves we confront a further problem, as much previous work has determined which manuscripts are worthy of our attention. These ingrained attitudes have constructed a canon of ‘important’ manuscripts based largely upon the preferences of Victorian scholars, allowing editors to ignore many of the non-canonical witnesses. The canon was initially created by Furnivall who published diplomatic transcripts of six of the early manuscripts for the Chaucer Society ‘Six-Text’ series1, which he subsequently extended by the inclusion of separate editions of Ha4 and Dd. These eight manuscripts also formed the core of the ten manuscripts collated by the editors of the Variorum series for their individual editions, to which were added Ad3 and He. Despite their inclusion in the Furnivall ‘Six-Text’ and the Variorum ‘Top 10,’ many of these manuscripts have received little detailed 17 study beyond the occasional lip-service of a textual footnote. While making textual study of the poem more manageable, the canon restricts research further as it allows editors to ignore many of the other, non-canonical extant witnesses. However close study of at least some of these neglected manuscripts may reveal shortcomings in our traditional methodological assumptions, and may provide vital information for recreating the authorial text. The purpose of this article is to introduce a manuscript that exemplifies both these claims, and thereby demonstrates the need for a much freer and wider view of the textual tradition of The Canterbury Tales. British Library MS Additional 35286 is dated to the period 1430-50 and therefore falls outside the period of manuscript production that is generally considered by editors and commentators as significant for the genesis of the text. It is a neat, professional copy in a cursive anglicana hand, containing a once-complete Canterbury Tales, with some textual losses due to missing folios and quires.2 Its date of copying, rather plain appearance and missing leaves have certainly influenced its exclusion from the textual canon, particularly in the work of Furnivall and Skeat.3 Since the cavalier work of these Chaucerians, editors have been more sensitive in their appraisal, but frequently equally dismissive. Most criticism of the manuscript has been directed at its unique tale-order and its profusion of scribal errors.4 I have emphasised above how unreliable tale-order can be as evidence for textual affiliation, and Ad3 presents us with further support for this claim. The order of tales in Ad3 is principally that of the El-A manuscripts, although there are three major adjustments to that familiar arrangement. Criticism of the Ad3 alterations has been directed at the inconsistencies in certain references that such an arrangement creates, and has been further prejudiced by the modern belief in the priority of the El arrangement. The attitude to the text of this manuscript reveals similar limitations in our current approach. Ad3 contains an eclectic text with a blend of affiliations, with some parts showing a close relationship with the Hg text. Any overall analysis of the text of this manuscript must therefore take account of this wealth of affiliations by considering the entire text and not by extracting pieces for sample collation. The difference between the affiliations of the taleorder and the textual relationships which this manuscript shows further demonstrates the unreliability of tale-order as evidence of genetic correspondence. The remainder of this article will consider ways in which this manuscript can aid our understanding of the development of the text. As stated above the order of tales in Ad3 is that of El with some unusual variations. The Cook’s Prologue and Tale have been placed after The Manciple’s Tale, while the two tales of Group G appear independently: The Second Nun’s Tale between the Summoner and the Clerk, and The Canon’s Yeoman’s Prologue and Tale between the relocated Cook and The Parson’s Tale. While this may appear to be simply a disarranged version of the El order, a closer look at the evidence is more revealing. The movement of The Cook’s Tale is understandable as a response to the content of The Manciple’s Prologue, although the scribe was not apparently concerned by the conflicting references to the Reeve in the Cook’s Prologue that such an adjustment creates. The positioning of the two tales of Group G is also unusual, particularly when Editorial Assumptions 18 Canterbury Tales Project Occasional Papers II compared to the arrangement common to the El-A manuscripts. Those manuscripts present these tales as a stable group situated between The Nun’s Priest’s Tale and The Manciple’s Prologue, and later manuscripts following this order have merely attempted to secure this placement further.5 The separation of The Second Nun’s Tale and The Canon’s Yeoman’s Tale as found in Ad3 is a feature more reminiscent of Hg, where The Second Nun’s Tale stands alone as The Canon’s Yeoman’s Prologue and Tale do not appear in this manuscript. Indeed, other features of the Second Nun’s Tale suggest a relationship between these two manuscripts. Both Hg and Ad3 refer to the ‘Nonne’ while later manuscripts, particularly those of the El-A group use our modern title ‘Second Nun.’ Similarly the lack of prologue and tale division found in Hg is mirrored in the Ad3 copy, while the later manuscripts include such a division at line 120 of the tale. Therefore the failure to reproduce the El-A positioning of these two tales in Ad3 seems to be due to a relationship it shares with certain features of the ordering and ordinatio of Hg. The text of The Canon’s Yeoman’s Prologue in Ad3 reveals a further unique difference, as it does not contain the opening couplet found in all other extant witnesses and thus removes any reference to the tale of Saint Cecilia, which precedes the Prologue in other witnesses. While such an omission would seem logical considering the earlier placement of this tale in Ad3, it seems unlikely that the scribe would have been troubled by this single inconsistency in the light of his retention of the references to the Reeve in The Cook’s Prologue and to the Manciple in the opening of line of The Parson’s Prologue. We have seen that the movement of the Cook’s Tale was guided by the dramatic content of The Manciple’s Prologue, while in comparison there is no clear motivation for the excision of the opening couplet of The Canon’s Yeoman’s Prologue and the unusual placement of these two tales. It seems most likely that the form of this prologue as found in Ad3 is a very early copy that preserves a stage in the composition before The Canon’s Yeoman’s Tale was linked to The Second Nun’s Tale, and ultimately moved as a group to its El-A placement. Therefore in these two tales Ad3 provides us with evidence of features closely associated with both El and Hg, and possibly with a very early stage in the compositional process unrecorded in any of our extant manuscripts. Having considered an example of an early, possibly unrevised piece of text, I will now turn to an example which suggests the preservation of a later, revised piece of text in the Ad3 copy of The Franklin’s Tale. There are two passages in this tale, lines 1455-6 and 1493-8, which appear in only two manuscripts: Ad3 and El. These lines have been included in every modern edition of the poem, with the exception of Blake’s edition of 1980, although the unusual attestation has never been explained. There has been mixed critical response to their content, although most critics seem to have considered them genuine Chaucer. Despite the editors’ opinion that lines 1455-6 are ‘so unpoetical that one would gladly believe them an editorial addition’ the couplet appeared in the ManlyRickert text, and most recently in the Riverside edition with the unequivocal note: ‘In El Ad only but genuine.’ The second passage has met with a completely contrasting critical reception. While Manly-Rickert felt that ‘there are few pages in CT from the authorship of which one would more gladly absolve Chaucer than the latter part of this complaint of Dorigen’s,’ they lauded 19 the later addition as ‘thoroughly Chaucerian in thought and style.’ In an attempt to grant Chaucer editorial absolution Manly-Rickert suggested that the later addition of lines 1493-8 may have encouraged the composition of lines 1455-6 by the scribe of the El ancestor. However such a hypothesis is evidently based on a highly subjective approach to the passage and any explanation must treat both sets of lines as one combined act of revision whether Chaucerian or otherwise. There has been much debate over the possibility of authorial revision in the Franklin’s Tale, although this has centered upon the string of exempla which constitute Dorigen’s complaint to Fortune, lines 1355-1456. This complaint has met with a volley of adverse criticism due to its length, monotony and lack of apparent structure, and has often been regarded with some embarrassment by modern Chaucerians.6 Others have defended the piece, finding a coherence in the ordering of exempla that reflects the differing virtues of their protagonists (Baker 1961; Morgan 1977.) The argument for revision of the complaint was raised by Dempster in an article which examined the structure of the piece by comparison with Chaucer’s source, chapters 41-46 of Jerome’s Adversus Jovinianum (Dempster 1937.) As Chaucer’s ordering of exempla does not follow that of Jerome, Dempster argued that Chaucer’s return to earlier portions of his source while compiling the complaint revealed several stages of composition. In a further article Dempster used the evidence of the content and placement of a marginal gloss to reinforce her argument for a separate stage of composition at line 1395 (Dempster 1939.) This gloss records the existence of more potential material in the Jerome source: ‘Singulas has historias & plures hanc materiam concernentes recitat beatus Ieronimus contra Iouinianum in primo suo libro capitulo 39o.’ In addition to its unusual content, this gloss is significant as its placement varies across the nine manuscripts that contain it. Three of these manuscripts, Hg, Bo2 and En3, place the gloss at line 1395, a point at which Dempster had previously charted a move from chapter 41 to chapter 43 of Jerome’s text, and a difference in the treatment of the source material. Two other manuscripts containing this gloss, El and Ad3, place it alongside the end of the complaint, thus seeming to indicate a subsequent revision. However the evidence of the source material, and the changes outlined by Dempster are not sufficient evidence to support a theory of authorial revision. While the placement of the gloss at line 1395 in certain manuscripts, including Hg, may be indicative of an earlier version, there is no extant text of the complaint containing this version and therefore such a hypothesis remains slight. When we turn from this discussion to the evidence of the two extra passages found only in El and Ad3 we are on firmer ground. Here we have two passages, totally unrelated in content and function, which seem to be indicative of revision of the text of the tale as a whole. This possibility is strengthened by the appearance in these two manuscripts alone of a mass of marginal source quotation on a scale unparalleled elsewhere in The Canterbury Tales. This act of textual revision and marginal annotation may further include the movement of the gloss considered by Dempster, although such a revision remains less certain. Most discussions of revision are dogged by subjectivity and a stubborn belief in an editor’s ability to recognise genuine Chaucer from mere scribal activity, as Editorial Assumptions 20 Canterbury Tales Project Occasional Papers II exemplified by the Manly-Rickert comments above. These two passages are too short to make any conclusions for authorship, and there is not enough linguistic or literary evidence to support such an analysis. However the presence of these passages in these two manuscripts does raise the provocative possibility that Chaucer revised the text of this tale and that this revised version is now extant in only El and Ad3. The question of authorial revision of tales and linking passages and the sole appearances of revised versions in later manuscripts is highly problematic, but the possible textual evidence for the presence of revision in The Franklin’s Tale must force us to look at similar examples across the whole of The Canterbury Tales and across all manuscripts, irrespective of the received canon of important witnesses. The question of revision has been hotly debated for some time, although such discussions have become ossified by a concentration on favourite examples, such as the ‘Adam Stanza,’ and the Nun’s Priest’s Prologue and Endlink. While these examples are important, future research must turn to the wealth of similar problematic passages found across the many manuscript witnesses. This brief study of Ad3 shows that by devoting more attention to individual manuscripts beyond the Hg-El deadlock we may learn more about the development of the text, and about the place of Hg and El within the tradition as a whole. The issue of dating may be of less importance than has traditionally been assumed, and it is important that we begin to look beyond those manuscripts currently dated to the first quarter of the fifteenth century. By conducting detailed studies of individual manuscripts, paying attention to the whole variety of textual evidence they contain, we may obtain a more complete view of the overall textual history of the poem. These neglected manuscripts raise a number of uncomfortable questions which do not fit into the traditional framework developed for the textual study of The Canterbury Tales, thus challenging us to address many of the foundations upon which such a framework has been constructed. We have accepted that the Riverside edition is not The Canterbury Tales. We must now accept that neither is Hg nor El. Notes 1Furnivall’s ‘Six-Text’ edition included Hg, El, Cp, La, Pw, and Gg. 2For a full description of the manuscript, see Mosser 1996. 3Skeat’s Clarendon edition describes the manuscript simply as ‘imperfect.’ See Skeat 1894, iv. xiv. 4Dempster describes the variations to the El order as ‘defects...no doubt largely [due] to difficulties in obtaining copies of some tales,’ while Owen terms them ‘curious lapses’ (Dempster 1949; Owen 1991.) 5Seven of the A manuscripts include the Nun’s Priest’s Endlink, while four of these contain an additional six lines linking the Nun’s Priest’s Tale with the Second Nun’s Tale. 6Manly complained that Chaucer was ‘unfortunately as much seduced by this astonishing fad [the use of exempla] as was any of the French imitators of Machaut...In the Franklin’s Tale a fine story finely told is nearly spoiled by one 21 hundred lines of rhetorical exempla.’ Tatlock cited Dorigen’s complaint as evidence of ‘Chaucer’s interest in learning for its own sake’ (Manly 1926; Tatlock 1935.) Bibliography Baker, D. C. 1961. ‘A Crux in Chaucer’s Franklin’s Tale. Dorigen’s Complaint.’ JEGP 60: 56-64. Benson, L. D. 1981. ‘The Order of the Canterbury Tales.’ Studies in the Age of Chaucer 3: 77-120. Blake, N. F. 1983. ‘The Editorial Assumptions in the Manly-Rickert Edition of the Canterbury Tales.’ English Studies 64: 385-400. —. 1985. The Textual Tradition of the Canterbury Tales. London and Baltimore, Edward Arnold. Dempster, G. 1937. ‘Chaucer at Work on the Complaint in the Franklin’s Tale.’ Modern Language Notes 52: 16-23. —. 1939. ‘A Further Note on Dorigen’s Exempla.’ Modern Language Notes 54: 137-8. —. 1949. ‘The Fifteenth-century Editors of The Canterbury Tales and the Problem of Tale Order.’ PMLA 64: 1123-1142. Kane, G. 1984. ‘John M. Manly and Edith Rickert.’ In Editing Chaucer: The Great Tradition, ed. P. G. Ruggiers. Norman, Okla., Pilgrim Books: 207-29. Manly, J. M. 1926. ‘Chaucer and the Rhetoricians.’ Proceedings of the British Academy 12: 95-113. Manly, J. M. and E. Rickert, Eds. 1940. The Text of the Canterbury Tales: Studied on the Basis of All Known Manuscripts. 8 vols. Chicago, Chicago University Press. Morgan, G. 1977. ‘A Defence of Dorigen’s Complaint.’ Medium Ævum 46: 77-97. Mosser, D. W. 1996. ‘Witness Descriptions.’ In The Wife of Bath’s Prologue on CDROM, ed. P. M. W. Robinson. Cambridge, Cambridge University Press. Owen, C. 1991. The Manuscripts of The Canterbury Tales. Cambridge, D. S. Brewer. Root, R. K. 1941. ‘The Text of The Canterbury Tales.’ Studies in Philology 38: 1-13. Skeat, W. W. 1894. The Complete Works of Geoffrey Chaucer. 6. Oxford, Clarendon Press. Tatlock, J. S. P. 1935. ‘The Canterbury Tales in 1400.’ PMLA 50: 100-39. Editorial Assumptions 22 Canterbury Tales Project Occasional Papers II