Editing the Canterbury Tales: An Overview

5 Editing the Canterbury Tales: An Overview Norman Blake Despite intense scholarly attention over many years and innumerable editions, we still know very little about the genesis of the Canterbury Tales and it seems unlikely that any unanimity will be reached soon about how to edit the poem. Such questions as did Chaucer issue all or parts of the poem in his lifetime?, Did the work exist in one or more versions?, Is it complete or incomplete? and, What were Chaucer’s working methods? all impinge on how one might edit the poem and yet we seem no nearer to answering these questions than previous scholars were. All we can say with certainty is that there is uncertainty how the poem should be edited. There are no manuscripts extant from Chaucer’s own lifetime and the scribes who copied the work after his death clearly differed as to how it should be presented. There are over eighty complete or fragmentary manuscripts of it and yet the scribes never arrived at any degree of agreement as to how the poem should be presented. This is already evident in two of the earliest manuscripts, Hengwrt and Ellesmere, which differ profoundly in organization of the text and overall presentation.1 The same applies to printed editions which have been produced regularly since Caxton issued his editio princeps in c. 1476. The existence of such a large number of manuscripts often led to an alteration in an editor’s perception about the state of the text whenever a new manuscript was brought to light. This can be noted in Caxton’s attitude to the text, for when he discovered (or, as he says, had brought to his attention by an unnamed gentleman) a different manuscript of the Canterbury Tales, he altered the text in his second edition to bring it more into line with this new manuscript. This decision was not based on any serious comparison of the two manuscripts or understanding of their differences; it was more a case of prejudice or wishful thinking.2 A new manuscript was superimposed upon the previous editorial decision. Yet in many ways the same thing could be said to have happened in more recent editions. The discovery of El led to its being used by most editors as their base text without necessarily altering the way in which the text as a whole should be presented and what should be included. Skeat (1894) who first used this manuscript did not follow its evidence regularly.3 The more recent widespread acceptance that Hg is the earliest manuscript and contains the best text has not led to its adoption as the base text in many editions; its evidence has been superimposed on that of El (Tajari 1991). As a concept it has always been accepted that the Canterbury Tales is a single work, which is normally called a poem though it contains prose. The evidence of the General Prologue when the Host proposes the tale-telling on the way out and back indicates a single work. The evidence of the Retraction is less clear. The reference to ‘this litel tretys’ (12: 1081) with which it opens is probably to the Parson’s Tale which it follows.4 But the later reference to ‘the tales of Caunterbury thilke that sownen into synne’ (12: 1086) indicates a whole work which consists of various parts and can be related to the other THE C ANTERBURY T ALES PROJECT OCCASIONAL PAPERS I, 1993, 5-18 6 Canterbury Tales Project Occasional Papers I whole works mentioned. There seems no reason to reject this conclusion, but it naturally raises its own problems. The first is whether the poem as we have it is complete. The implication of the early editions is that it is, despite the fact that anyone who read the General Prologue would realize that the final work did not reflect the plan outlined in it. No early editor calls attention to the work’s unfinished state, though Caxton could have done so in his second edition where he discusses problems of the text. Since the early editions have no lineation and since the tales proceed in a continuous sequence (whatever the sequence is), there is nothing to indicate to a reader that there is any gap. When Tyrwhitt edited the text in the eighteenth century, he introduced a continuous lineation which naturally promoted the idea that the text was complete. It is only in the nineteenth century with the spread of the idea of fragments and a lineation system which reflected this that the poem was presented to its readers as incomplete. A poem which consists of fragments will appear to its readers as unfinished, though this is not always the impression that undergraduate readers today come away with, because as there is a beginning and an end, it is possible to assume that somehow the whole poem is there. More recently the concept that the poem is complete in its incompleteness has been argued by some, though this has not had any impact on the way editions are presented or on the lineation adopted.5 A complete text should have a single lineation. Few editions agree as to what they include in the text, but most adopt the lineation of the ten-fragment system found in Skeat’s edition. This means an imperfect lineation in many editions, either with line numbers missing or with line numbers with added a, b, c, etc. A satisfactory system of lineation is now a desideratum, but the need to have a single system which would allow reference to all manuscript texts is unlikely to be realized. No one has given it the attention it deserves.6 When early editors produced what was the complete text of the Canterbury Tales, they probably never considered that the work might have existed in more than one authorial version. Although most editors have been aware that the poem survives in more than one manuscript, that did not of itself make it any different from other works composed in the manuscript age. They recognized that manuscripts could differ, but the reaction was more often to think that the text had not been properly preserved and copied by various scribes rather than that Chaucer had left a number of different versions. This is made quite clear by Caxton in the prologue to his second edition: ‘Whyche book I have dylygently oversen and duly examyned to th’ende that it be made acordyng unto his owen makyng, for I fynde many of the sayd bookes whyche wryters have abrydgyd it and many thynges left out, and in somme place have sette certayn versys that he never made ne sette in hys booke’ (p. 62). This expresses the view of most editors that there is a Chaucerian text which it is the duty of editors to issue in a definitive edition containing only what Chaucer wrote and all that he wrote. It is hardly surprising that editors have not found the tools to accomplish this aim. From the nineteenth century onwards, the idea gained ground that some of the variations among the manuscripts could be the result of authorial insertion or deletion rather than scribal meddling. In a long work such as the Canterbury Tales written over a long period of time, this idea is 7 perfectly reasonable.7 As the poem consists of a series of links and tales, it need not follow that there are two completely variant versions of the whole poem, though that seems to be considered possible by Manly and Rickert.8 It is more likely that individual tales or links could have been rewritten by Chaucer. But granted that this is so in some tales and links, it could suggest a process of revision which amounts to more or less an earlier and a later version. The reaction of editors to possible authorial changes in the text has been muddled to say the least. It is not infrequent to find claims that certain passages have been added later or deleted or even that they were marked for deletion, and yet these passages are still included in modern editions. In other words, the text of a given tale which may exist in more than one authorial state is often presented in a conflated state which includes both authorial and additions (Blake 1992). Anything which an editor regards as Chaucerian is likely to be included in the edition whatever stage of the text it represents. In this respect modern editors have hardly advanced beyond earlier ones who often went in for conflation, except that modern editors usually indicate through some sign such as square brackets that the passage in question is of a different status from the rest of the text. If the poem is incomplete and exists in a fragmentary state, what does this tell us about Chaucer’s method of composition and how might that affect the way in which the work is edited? As already implied, the Canterbury Tales is unfinished in an unusual way in that it appears to have a genuine beginning and end. It is incomplete because pieces are missing from the middle. This indicates that Chaucer did not start his poem at the beginning and gradually work through and suddenly stop. Rather, he accumulated pieces, i.e. tales and links, which he was joining together to form the whole poem. Although he outlined in the General Prologue what his plan was, all authors change their minds and we cannot guarantee that the finished work would have realized this plan. It does mean that at least individual tales could have had a period of independent existence until they were linked together, if ever, with another tale or tales to form the core of a fragment. It is not unreasonable, therefore, to suppose, if Chaucer left the poem unfinished, that when he died some tales may not yet have been incorporated into the framework of the poem so that they were still freestanding and without any link. The problem is that, because the tales are often written in different styles and represent different genres, it is difficult to prove on stylistic, metrical, or linguistic grounds that a tale should be part of the poem if it is not found in the earliest manuscripts, but is included in later ones. This applies particularly to the Canon’s Yeoman’s Tale and the Tale of Gamelyn. Neither is in Hg; both are in Ha4 and Cp; and only the former is in El. Older editions tended to include both tales as part of the poem; modern ones tend to include only the Canon’s Yeoman’s Tale. The reasons for the inclusion of the Canon’s Yeoman’s Tale in modern editions are rarely stated, but are probably as follows. It is in El, which is the base manuscript for most modern editions. It is clearly integrated into the framework of the pilgrimage because its prologue shows that it must follow the Nun’s Tale. Finally, its style and narrative quality are such that critics cannot imagine that anyone other than Chaucer would have written it.9 The entry of a new pilgrim is regarded as a touch of narrative invention worthy Editing the Canterbury Tales 8 Canterbury Tales Project Occasional Papers I of Chaucer alone. The reasons for the exclusion of the Tale of Gamelyn are the reverse of these reasons. It is not found in El, it is not worked into the pilgrimage frame because it usually follows the incomplete Cook’s Tale, and it is not regarded by critics as good enough for Chaucer from a stylistic or narrative point of view. As it happens, neither tale is known outside the Canterbury Tales. Many scholars assume that the Tale of Gamelyn existed already even though it is not attested elsewhere, but that Chaucer composed the Canon’s Yeoman’s Tale. These reasons are far from compelling. Since Ha4 contains both the Tale of Gamelyn and the Canon’s Yeoman’s Tale and may be an earlier manuscript than El, it is possible that the former entered the manuscript tradition of the Canterbury Tales at the same time as, if not before, the latter. That the Tale of Gamelyn is not worked into the framework of the pilgrimage suggests that it may have been an independent tale written by Chaucer waiting to be worked into a bigger fragment. After all, a tale like the Nun’s Tale is not worked into the pilgrimage frame either, and without the Canon’s Yeoman’s Tale it is a free-standing tale without links. As the Canon’s Yeoman’s Tale is worked into the pilgrimage frame and is linked to the Nun’s Tale, it raises the question why two tales which were linked together by Chaucer (if that is the hypothesis) should have been decoupled so that one of them is not included in the earliest manuscript. As far as we can tell, this has not occurred elsewhere in the poem. It opens up the possibility that the tale could have been linked to the Nun’s Tale after Chaucer’s death. The matter of how good each tale is and whether one is by Chaucer and the other not is a subjective matter. All one can say is that the poem contains a wide range of styles and literary excellence, which could embrace the Tale of Gamelyn. It is possible that if, for example, the Tale of Thopas or the Tale of Melibee were in the same position and of the same status as the Tale of Gamelyn, they would be dismissed as unworthy of Chaucer’s literary ability. Perhaps the main reason for the inclusion of the Canon’s Yeoman’s Tale and the exclusion of the Tale of Gamelyn in modern editions is editorial conservatism and publishers’ timidity, for as a critical tradition grows up as to what is and what is not by Chaucer it becomes increasingly difficult to challenge it. If the tales existed as independent pieces before being incorporated in the pilgrimage frame, it raises the possibility that they were circulated before Chaucer’s death to selected friends and perhaps even more widely. This possibility would in turn then mean that such tales, if circulated, could have been copied by those that had them and been subject to corruption and change. The implication of that would be that each individual tale could have its own stemma. The justification offered for prior circulation has consisted of two points. The first is that Manly and Rickert accepted that this was the only possible explanation which could account for the complicated transmission of the individual tales which they had arrived at.10 But since their editorial techniques have come under increasing attack recently, this point is hardly reliable (Blake 1983; Kane 1984). The second is that in the Envoy to Bukton the following lines occur: The Wyf of Bathe I pray yow that ye rede Of this matere that we have on honde. (Benson 1987, 655-6) 9 ‘This matere’ is that of getting married and the sorrow and suffering that marriage brings to a man. It is often urged that since this lyric was written before the Canterbury Tales was issued as a whole, it means that parts of the Canterbury Tales must have been available in Chaucer’s lifetime. The position is not quite so straightforward. The lyric is found in a single fifteenth-century manuscript (Fairfax 16) and Chaucer’s authorship of it is assumed rather than proved. In lyrics at this time, verses could be added or deleted in transmission, and as the reference to the Wife of Bath occurs in the final stanza, the possibility remains that this is an additional stanza tacked on to the poem at some date after the Canterbury Tales became available. A reference to the Wife of Bath before the poem or tale was generally available would be somewhat pointless and rather precious. The reference to the Wife must be to her prologue where she describes the sorrows she inflicted on her husbands. This prologue contains an exchange between the Friar and the Summoner, whose tales are linked to that of the Wife to form a distinct fragment. This suggests it is unlikely that the Wife of Bath’s Prologue and Tale were circulated independently, because they would almost certainly have been linked already to the Tales of the Friar and Summoner. Although there is nothing intrinsically against a fragment being circulated before the rest of the poem, one does question whether in fact it happened, particularly in this case. The last point is of some importance in that the circulation of individual tales reflects modern attitudes and editorial assumptions about the poem. We today are used to seeing editions of individual tales. Such tales are set for A level students, and in university departments of English although students are encouraged to read the whole poem, they are required to read certain tales. The poem consists of separate and complete individual tales. This is reflected in their exam papers which may ask for a critical assessment of a single tale, a reflection of what one finds in the literature. To start with, the manuscript tradition presented the whole poem. If there was circulation of individual tales, one might have expected to find that reflected in the manuscript tradition with early copies of single tales. In fact this happens only later when individual tales might be excerpted from the whole poem to be included in manuscripts which contained a collection of a particular genre such as saints’ lives. Whatever the genesis of the work was, the evidence suggests that fifteenth-century readers saw the Canterbury Tales as a whole rather than as a collection of discrete parts, whereas we today tend to see it rather as a collection of tales like the Arabian Nights. Chaucer’s contemporaries almost certainly saw the Canterbury Tales as a poem like Gower’s Confessio Amantis; after all, Chaucer and Gower were closely linked in most people’s minds at that time. Although earlier editors of manuscripts and printed books saw the poem as an entity, this did not prevent them from presenting it as a series of tales, many of which were broken down into smaller sections. The manuscripts contain headings for all tales and links as well as section divisions for some tales. In the manuscript tradition this cutting up of the poem into smaller parts was carried a stage further in that most tales contained a series of glosses which tended to highlight particular passages in each tale or prologue. The division into tales and sections was thus part of an overall Editing the Canterbury Tales 10 Canterbury Tales Project Occasional Papers I presentation which emphasized within a complete poem some of the moral meaning and poetic highlights which one could get from it. This type of presentation was common in many manuscripts and probably reflected the way in which texts were read. They were storehouses of information which could be put to practical use. This is how Caxton presented Malory’s Le Morte Darthur and the growing use of indices encouraged the development of methods of breaking a text into parts to permit easy referencing (Blake 1976). This did not necessarily mean that later medieval readers did not accept the completeness of a given work, though excerpting was as common then as anthologizing today. All tales and links have headings in the manuscripts, but the manuscripts rarely agree as to what the headings should read and sometimes where they should come. The difficulty is that the poem consists of a series of tales joined by links. Precisely what constitutes the tale and what the link can be difficult to determine. More importantly the links conclude one tale and introduce another. Most people today, and probably most in the past, thought of the poem as consisting not so much of links and tales as of prologues and tales. We think of the Canon’s Yeoman’s Prologue and Tale, for example, and if these two are not genuine as has been suggested this pair illustrates how Chaucer’s immediate successors may have thought of the structure. But in order for this prologue and tale to be fitted into the poem, it follows that there must be a reference in the prologue to the previous tale and so it could be said to conclude that tale before it introduces the Canon’s Yeoman’s Tale. Although we call it the Canon’s Yeoman’s Prologue it is both the conclusion of the Nun’s Tale and the prologue of the Canon’s Yeoman’s Tale. Other tales are joined by more than one link as is true of the tales of the Physician and the Pardoner, for the prologue to the Pardoner’s Tale follows a link between the two tales. The Pardoner’s Tale itself also has an endpiece which concludes the tale but which does not introduce another tale. There is thus no regular pattern in the structure of the poem, though editorial methods try to suggest that there is or should be. More importantly, if the headings are not the same in all the manuscripts (for they can differ not only in wording but also in the language used), then one may question whether they were found in Chaucer’s original or rather, if they were found, how exactly they occurred. Once again one can look to Gower’s Confessio Amantis for a guide. In modern editions (e.g. Macaulay 1900) that poem is presented with headings and glosses in the margin so that the sense of unity is not destroyed. One might reasonably ask whether Chaucer was attempting to present his poem in a similar way. It seems fairly certain that some glosses and some types of heading were found in his original.11 All manuscripts have headings and most have glosses. It could be that Chaucer simply noted in the margin of his original some brief indication that a pilgrim began his tale at that point. This may have been in Latin to give something like ‘Uxor narrat’ for the Wife of Bath’s Tale. The headings may have been in different languages as was probably true of the glosses. But whether there would have been headings for the links is more debatable. To have a heading to indicate where some character starts his story seems more probable than that there is one to indicate when the story reverts to the frame. If the headings now found in 11 the base manuscript used by an editor were to be pushed into the margin in a modern edition, this might bring them more into line with the glosses and with the section titles. In some manuscripts, for example, the names of the various pilgrims as they are described in the General Prologue occur in the margin at the beginning of their description. The same applies to the brief tragedies which make up the Monk’s Tale. These are not usually reproduced in modern editions, though they are occasionally found. The division of a tale into sections is, however, more likely to be found in a modern edition so that the Knight’s Tale and the Clerk’s Tale will often be divided in a modern edition into several parts. Unfortunately, once again the manuscripts do not always agree as to where these divisions should come, though most modern editors follow El’s arrangement. But modern editors do not usually put in the glosses, except in one or two places such as at the beginning of the Knight’s Tale where a quotation from another text is considered to be not so much a gloss as an integral part of the text, though such a decision is a fine one. We have in modern editions lost the sense of a continuous text provided with a running commentary of glosses and divisions and subdivisions to direct our reading. Our response to the text is now quite different because of the way editors present it to us. The two major decisions which most editors have to make are what manuscript to use as a base text and in what order to organize the parts they intend to include. It is probably true to say that the only editors to have undertaken any comprehensive work on the manuscript tradition of the Canterbury Tales are John Manly and Edith Rickert. This is simply because the number of manuscripts is so large and the divergence among them so great that it requires a Herculean effort to organize the variant readings in a way which would allow a stemma to be constructed either for the whole text or for its individual parts. The conclusions which Manly and Rickert reached, that there were many divergent lines of descent for the various tales or even parts of the tales, and that the manuscript which probably reflected the original most closely in textual matters was Hg, created so many problems for editors that their work has not been followed in any detail by subsequent ones. Consequently, in their decision as to which manuscript to follow, editors have been more motivated by the order of the tales and what they want to include in their final version. The view that the El order and its contents reflect most closely Chaucer’s plans for his text towards the end of his life (no matter what the earlier plans may have been) has been an influential one with editors (e.g. Benson 1981). The lively debate among critics and commentators as to what Chaucer’s intentions were and what the relationship among the early manuscripts is has had much less impact on editors. Most editors tend to be more influenced by what previous editors have done rather than by what studies about the manuscripts reveal. The diversity of attitudes among those who write on the manuscripts is thus not reflected in editions, which tend to be uniform in their presentation of the text apart from one or two notable exceptions.12 The only influence such studies have had has been to increase the number of Hg readings into a text based on El, as we have already noted. In order and contents El is largely followed in modern editions, but this has led to the following problems and areas of dispute. The end of the Man Editing the Canterbury Tales 12 Canterbury Tales Project Occasional Papers I of Law’s Tale has in Ha4 , Cp, and Ne a conclusion known as the Man of Law’s Epilogue, which is included in most modern editions. In these manuscripts the reference to the next teller after the Man of Law varies among the Squire, the Summoner, and the Shipman, but this rarely influences editors as to the tale which follows, since most proceed with the Wife of Bath’s Prologue and Tale. The absence of this link in Hg and El among other manuscripts and the unwillingness of editors to follow the reading of line [B11179] as to which tale should follow illustrate the problem editors find themselves in when they decide to mix their manuscript sources by including all that they feel Chaucer wrote. In the Wife of Bath’s Prologue there are five passages not in Hg of which only four occur in El. Most editors include all five in their text, usually on the grounds that these were late insertions in the text and consequently not reflected in many manuscripts. They do, however, profoundly alter the portrayal of the Wife and her attitude towards marriage (Blake 1982; Kennedy forthcoming). The Clerk’s Tale concludes in most manuscripts with the Host’s Stanza, and it is found in all early manuscripts. Some modern editors do not include it in their text on the grounds that it was cancelled by Chaucer despite its inclusion in the early manuscripts. Its excision in some manuscripts reflects almost certainly the attempt to impose a prologue-tale pattern on the poem into which this stanza fits uncomfortably. It is at this point in most editions that the question of which manuscript is used as base text becomes most acute in that the order of the tales of the Clerk, Merchant, Franklin, Squire, and Nun and how they are linked together vary considerably in the extant manuscripts. Hg contains the tales in the order Squire, Franklin, Merchant, Nun, and Clerk; but most editions follow El in the order Clerk, Merchant, Squire, and Franklin with the Nun pushed back further in the order. Links which could be late additions in Hg for some of these tales occur with different names and in different sequences in El. Most editors argue that the Hg scribe adapted these links by himself to fit them to the order he had already imposed on the tales. These links are generally assumed to be genuine, though their status has been questioned (Blake 1979). It is the decision to accept the Canon’s Yeoman’s Prologue and Tale as genuine which pushes the Nun’s Tale further back, because the former is linked with the latter and contains geographical references which indicate a later stage in the pilgrimage. The Monk’s Tale and the Nun’s Priest’s Prologue and Tale have a series of alterations in the text which may be associated. Hg does not include the Adam stanza in its tragedies, though this stanza is found in most later manuscripts. In the same tale there are four tragedies dealing with fourteenth-century people known collectively today as the Modern Instances. In Hg, El, and many other manuscripts these are placed at the end of the Monk’s Tale, but other manuscripts have them earlier in the tale between the tragedies of Zenobia and Nero. Despite the position of the Modern Instances in El, many modern editors put the Modern Instances in the middle of the tale. Doing this makes the tragedy of Croesus the last one in the tale, which some editors think is referred to in the Nun’s Priest’s Prologue which follows the Monk’s Tale. However, that prologue itself exists in two forms, a shorter and a longer one. The latter is the more frequent in 13 the manuscripts, though the shorter is found in Hg, Cp, and Ne. At the end of the Nun’s Priest’s Tale, a few manuscripts including Dd contain a link which does not indicate who the next teller is because the link appears to be incomplete. This link may have been written to accommodate the placing of the Nun’s Tale after the Nun’s Priest’s Tale, for the former has no prologue within the pilgrimage frame. Editors who include the Canon’s Yeoman’s Tale have to link it to the Nun’s Tale, which is then placed after the Nun’s Priest’s Tale. This in turn prompts them to include this endlink for the Nun’s Priest’s Tale even though it is attested in a few mainly late manuscripts and its genuineness must be questioned. As with all Middle English texts the editing of the Canterbury Tales exhibits a wide array of solutions to the questions of metre, spelling, and punctuation. Metre also involves style. For the scribes of the manuscripts style was probably the most important factor and some tried to improve on Chaucer. There is evidence that El is an edited text and some of the improvements which its scribe tried to introduce are stylistic. In particular, he frequently altered the language to make parallelism or contrast more obvious. Thus in the General Prologue where Hg has ‘As wel in cristendom as hethenesse’ (1: 49), El alters this to ‘As wel in cristendom as in hethenesse’ where the second ‘in’ makes the parallelism between the two phrases much closer. There are many similar examples from El (Pearsall 1985, 10-12). Other scribes alter the language in different ways. The northernisms in the Reeve’s Tale are expanded and regularized in some manuscripts (Blake 1979b). In his changes the scribe of El may destroy the metre or make it more regular. Some scribes seem to have been influenced by the need to regularize the metre, at least in those lines where Chaucer may have been most daring. The early scribes were little influenced by the spelling or the punctuation of their copytext in that they simply followed what was then the accepted convention of imposing their own spelling and punctuation on the text they were copying. As far as punctuation goes that attitude still prevails. Modern editions impose a modern punctuation on the text, and where they differ it is only in the amount of punctuation which is introduced and in how they interpret the text. In the nineteenth and early twentieth centuries it was common to introduce as much punctuation as was considered necessary to make the grammar of the text clear. In other words the text was punctuated as though it had been written in modern times, even though this created some problems. Recent editions have tended towards a lighter punctuation, partly because the punctuation in the manuscripts is light and partly because that is the current fashion. Even a light punctuation forces decisions to be made as to the relationship of various clauses which may have been left ambiguous in the manuscripts. Fourteenth-century English has fewer conjunctions than modern English and consequently two statements may be placed one after the other without any linking punctuation. To introduce modern punctuation may heighten or diminish the connection between the two statements and thus subtly alter the meaning of the text. For example, in the description of the knight in the General Prologue we find these two lines which are a little way apart from each other. And eueremoore he hadde a souereyn prys (1: 67) He was a verray, parfit, gentil knyght (1: 72) Editing the Canterbury Tales