Chaucer's Metre and Scribal Editing in the Early Manuscripts of The Canterbury Tales

143 Chaucer’s Metre and Scribal Editing in the Early Manuscripts of The Canterbury Tales Elizabeth Solopova This study presents a metrical comparison of the text of The Wife of Bath’s Prologue as preserved in Aberystwyth, National Library of Wales, Peniarth 392 (the ‘Hengwrt MS’ henceforth referred to as ‘Hg’), San Marino, Huntington Library, MS 26. C. 9 (the ‘Ellesmere MS’ henceforth referred to as ‘El’) and, when necessary for the argument, in four other early manuscripts of The Canterbury Tales: Cambridge University Library Dd. 4. 24 (Dd), Cambridge University Library Gg. 4. 27 (Gg), Oxford, Corpus Christi College, MS 198 (Cp) and London, British Library, MS Harley 7334 (Ha4.) The aim of this comparison is to investigate the processes of scribal and editorial intervention at the earliest stages of the textual history of The Canterbury Tales. The particular value of metrical analysis is that the different editorial policies and scribal processes underlying the earliest manuscripts are likely to reveal themselves in their handling of metre. Metrical analysis may help to identify non-authorial intervention and provide insight into the shaping of the earliest forms of the text of The Canterbury Tales.1 Before we attempt the metrical comparison of manuscript readings it will be helpful to explain the general assumptions about Chaucer’s metre that underlie this study. It was assumed that The Wife of Bath’s Prologue is written in fivebeat syllabic verse. In its ideal form a line of such verse is a succession of alternating stressed and unstressed syllables (x for unstressed, / for stressed): x / x / x / x / x / (x.) However, in actual practice such lines are rare. Poetry where every line is perfect in the sense that it follows exactly the requirements of a single pattern and is therefore identical with every other line could not exist: its regularity would make it uninteresting. Because of this in syllabic poetry of any period weak positions are occasionally filled with stressed syllables and strong positions with unstressed syllables. It is widely accepted that, in addition, Chaucer’s verse allowed such syllabic variations as a missing unstressed syllable at the start of the line, or occasionally inside the line, or sometimes two unstressed syllables instead of one. We do not know for certain how much rhythmical freedom Chaucer allowed in verse. Rhythmically free lines in his poetry might result from scribal carelessness. However, rhythmically smooth lines can also be due to scribal intervention, as some scribes corrected the metre. In this study the lines from the manuscripts were compared from the point of view of which is more regular, that is which is closer to the ideal metrical pattern described above. This comparison has shown that some of the manuscripts are consistently metrically more regular than others, and that beyond their metrical regularity or its lack, lie certain peculiarities of language and style. More regular lines were not automatically regarded as more THE CANTERBURY TALES PROJECT OCCASIONAL PAPERS II, 1997, 143-164 144 Canterbury Tales Project Occasional Papers II acceptable or more likely to be original. No assumptions on what is Chaucerian were made out of purely metrical considerations. I also attempted to avoid as far as possible the use of examples where scansion presents problems due to final -e or similar difficulties. When the texts of Hg and El are compared it is often asserted that El is an edited manuscript and that one of the aims of its editor was to regularise Chaucer’s metre. Thus, according to Pearsall (1985, 10-11) ‘the Ellesmere manuscript itself is quite extensively edited. This editing was carried out in a highly intelligent and responsible manner, and was designed to “improve” grammar and syntax, to clear up apparent irregularities and inconsistencies, to eliminate what were thought to be infelicities, and to regularise Chaucer’s metre according to a ten-syllable pattern.’ Pearsall argues that whereas the editor of El corrects Chaucer’s metre in his preoccupation with ‘regularity and consistency,’ Hg gives a ‘more accurate representation of a flexible, idiomatic and successful metrical practice’ characteristic of Chaucer as a poet. In the opinion of Manly and Rickert (1940, I: 150) El is a manuscript edited by ‘an intelligent person, who was certainly not Chaucer.’ Manly and Rickert believed that some of the deliberate changes in El were made for the purpose of regularising the metre (e. g. II: 149-150.) According to Hanna (1987, 93) ‘Hengwrt is marginally rougher than Ellesmere, more prone not to offer hendecasyllables but headless and Lydgatian lines (either clashed stresses or preserved -e at the caesura.)’ According to Fisher (1988, 787-788) stylistic revision in El produced a change to a milder tone and greater regularity of metre in comparison to Hg. Fisher suggests that this revision could have had an authorial origin. It is common ground among commentators that El, in comparison with Hg, shows evidence of consistent stylistic editing. However, this analysis of The Wife of Bath’s Prologue provides evidence against the opinion that El contains a metrically more regular version of the text than Hg. Stylistic editing in El, as outlined by Pearsall and other commentators, seems to have been conducted on the whole without concern for metrical regularity. In some forty-four out of fifty-three cases the readings in which El differs from Hg and which conform to typical ‘El stylistic revision’ are metrically less regular than the corresponding readings in Hg. Metrically regular readings in El compared to Hg Comparison of the El and Hg texts of The Wife of Bath’s Prologue has shown that El gives a reading that is metrically smoother than the reading of Hg only about nine times through the entire prologue.2 Of these in four cases the metrically irregular readings in Hg are unsupported by the other five early manuscripts analysed in this study, and most of them are likely to be accidental copying mistakes of the Hg scribe. In these unique readings Hg omits a word in three cases (ll. 215, 397 and 792) and adds in one (l. 4.) Line 792, where Hg, probably accidentally, omits him present in all other manuscripts, and because of this has a metrically irregular reading, can serve as an illustration: 145 I with my fest· so took hym on the cheke (El) I with my fist· so took on the cheke (Hg) This line, syllabically regular in El, has in Hg a missing unstressed syllable after the third metrical beat (represented by took) which distorts the regular iambic pentameter pattern. Other cases of omission are also likely to be accidental: I sette hem so a werk· by my fey (El, 215) I sette hem awerk· by my fey (Hg) I swoor , al my walkynge out by nygœt (El, 397) I swoor , that my walkyng· out by nyghte (Hg) Line 4, where Hg adds a word relative to El, is syllabically regular in El, though its stress pattern is slightly unusual: ffor lordynges , sith I . xij. yeer was of age 3 In Hg it is syllabically irregular due to the presence of : ffor lordynges , sith I twelf yeer was of age There is a strong possibility that this reading of Hg is original because of the evidence of Chaucer’s preference for complex conjunctions with ‘that’ and because of the support of other important manuscripts.4 In five other cases where the reading of Hg is metrically less regular than the reading of El, Hg is supported by other early manuscripts, whereas El is in the minority: once it shares its reading with Gg (209), twice with Ha4 and Gg (176, 212), once with Dd and Ha4 (306) and once its reading is unique (396.) Line 306 where El Dd and Ha4 have a more regular rhythm than Hg can serve as an illustration: Yet hastow caught· a fals suspecioun (El) Yet hastow caught· fals suspecioun (Hg)5 The metrical regularity of these readings in El does not prove their authorial origin. Lines 176 and 212 involve the omission of ‘that’ in complex conjunctions: El Gg and Ha4 have a variant wheither against wheither in Hg and Dd in 176, and sith against sith that in 212. As noted above, there is evidence for Chaucer’s preference for complex conjunctions which, as will be shown later, El and Gg often simplify by omitting ‘that.’ The fact that El shares these readings with Gg and Ha4 also undermines their authority. It will be argued later that the evidence that so many of the readings shared by El Gg and Ha4 against Hg are clearly not authorial increases the likelihood that all such readings in these manuscripts are the result of editorial or scribal intervention. Metrically irregular readings in El compared to Hg Hg is metrically more regular than El in about forty-four of the analysed readings, against ten cases (discussed in the last section) where El is more regular than Hg. In the majority of these cases the readings of El are shared by Gg and sometimes Ha4. The most common textual change in El as against Hg is Metre and Scribal Editing in The Wife of Bath’s Prologue 146 Canterbury Tales Project Occasional Papers II a missing or an added word. A missing word makes a line rhythmically irregular in El in about seventeen cases; of these in six cases the readings of El are unique, nine times they are shared by Gg, once by Gg and Ha4 and once by Ha4.6 An example of a reading where El is independent of Gg is l. 113 where al present in most early manuscripts is omitted in El and Ha4: I wol bistowe , the flour of al myn age (Hg) I wol bistowe , the flour of myn age (El) In l. 391 the reading of El is shared by Gg: They were ful glad , to excusen hem ful blyue (Hg) They were ful glad , to excuse hem blyue (El)7 The missing word in both cases is an intensifying particle, not absolutely necessary for the sense (though stylistically important as it gives colloquial colouring to the text) but essential for metrical regularity. In seven cases of metrically inferior readings in El the omitted word is ‘that.’ Four of these omissions El shares with Gg. Another two—49, 51—occur before l. 77 where the text is missing in Gg and we do not know whether Gg also omitted ‘that’ in these cases or not.8 In three of these lines the omitted word is explicative ‘that,’ introducing a subordinate clause: 49 (El), 51 (El), 251 (El Gg Ha4.) In four lines it is a pleonastic ‘that’ in the phrases if that (253 El and 85 El Gg), sith that (140 El Gg) and whil that (157 El Gg.) Thus l. 85 is metrically regular in Hg, but lacks a metrically necessary syllable in El and Gg: To wedde me , if that my make dye (Hg) To wedde me , if my make dye (El) According to Partridge (1992, 32) when Chaucer uses the word ‘which’ it is often in the phrase ‘which that.’ This phrase is more common in Chaucer’s verse than in his prose, but it appears in all his prose works. Thus, in the Parson’s Tale there are seventeen instances of ‘which that,’ out of a total of 167 occurrences of ‘which’; in Melibee, thirteen out of sixty-two. In the Equatorie of the Planetis there are twenty-two uses of ‘which,’ but none of ‘which that.’ Partridge argues that this construction is characteristic of Chaucer’s linguistic usage and its absence testifies against the attribution of the Equatorie to Chaucer. This may be true about other similar phrases with pleonastic ‘that.’ Frequent omissions of pleonastic ‘that’ in El and Gg may signify the presence of stylistic revision in their texts and testify against the authorial origin of their readings. The use of complex conjunctions in Chaucer’s prose demonstrates that ‘that’ is not just a ‘metrical word,’ but a part of an idiom that Chaucer used independently of metrical considerations.9 In eleven cases, the rhythm in El is less regular due to an added word as compared to Hg. Of these eleven, just two occur only in El among the manuscripts here analysed: these are lines 35 and 110. In l. 110 the iambic pentameter rhythm is completely destroyed due to addition of the word steppes, absent from other early manuscripts: And in swich wise , folwe hym and his foore (Hg) And in swich wise , folwe hym and his steppes foore (El) 147 This word is interlined in El and was probably added to clarify the meaning, as this line was obviously difficult for some of the scribes: thus, Cp has a variant lore instead of foore. Additions in El shared by Gg which make the rhythm irregular often seem to be a result of a conscious editorial policy designed to achieve a more formal style, and a simpler and more logical syntax. Thus, in l. 183 El and Gg insert it: this adds consistency to the syntactic construction, but also gives a metrically redundant syllable: Thise same wordes , writeth Protholome Rede in his Almageste , and take it there (Hg, 182-183) Rede it in his Almageste , and take it there (El, 183) In l. 214 El Dd and Gg insert if, producing a more formal syntax and less regular rhythm: What sholde I take kepe , hem for to plese But it were , for my ∏fit· and myn ese (Hg 213-214) What sholde I taken heede , hem for to plese But if it were , for my profit· and myn ese (El 213-214) Other additions also appear to introduce minor stylistic and grammatical changes. Thus, in ll. 130 and 326 El and Gg have extra-metrical syllables due to added articles not present in Hg: That man shal yelde , to his wyf hir dette (Hg 130) That a man shal yelde . to his wyf hir dette (El 130) Of alle men , his wisdom is hyeste (Hg 326) Of alle men , his wysdom is the hyeste (El 326) An explicative that resulting in a more formal syntax, is added in El and Gg in 282, and in El Gg and Ha4 in 257: Thow seyst· som folk , desiren vs for richesse (Hg 257) Thow seyst· that som folk· desiren vs , for richesse (El 257) There are also about sixteen cases where the less regular rhythm in El as compared to Hg is due to the use of different words or forms of words in these manuscripts, or is caused by more significant textual changes.10 Twelve of these sixteen readings El shares with Gg. Another three occur in the beginning of the prologue, missing from Gg.11 Some of the readings in this category which El shares with Gg are due to minor changes which do not affect the meaning significantly: for example, that used in El and Gg instead of thilke in Hg and other manuscripts in l. 177. Others are more evidently the result of stylistic revision. The variants of El and Gg usually differ from Hg by being stylistically more formal. Thus, l. 64 is syllabically regular in Hg: Thapostle , whan he speketh of maydenhede 12 El removes the conversational pleonasm from this line and makes it metrically irregular: Whan thapostel , speketh of maydenhede Metre and Scribal Editing in The Wife of Bath’s Prologue 148 Canterbury Tales Project Occasional Papers II In l. 260 El and Gg add som for the sake of syntactic parallelism: And so~me , for gentillesse , and dalyaunce (Hg) And som for gentillesse , and som for daliaunce (El)13 In l. 250 El and Gg remove that and add and. These changes result in a more formal syntax and a less regular rhythm: And if that she be ryche , of heigh parage (Hg) And if she be ryche , and of heigœ parage (El) A similar change (that removed, the word order changed to achieve a more logical syntax) occurs in El and Gg in l. 389: Who so that first to Mille comth , first grynt· (Hg) Who so comth first to Mille, first grynt (El) In two more lines in this category El and Gg remove syntactic inversion, preferring a prosaic word order together with a less regular accentual pattern: I sey this , they maked been for bothe (Hg 126) I sey yis , that they beth maked for bothe (El 126) That ech of hem , ful blisful was and fawe (Hg 220) That ech of hem , was ful blisful and fawe (El 220) There are a number of rhythmically free lines in Hg and other manuscripts. When the editor of El and Gg introduces stylistic corrections to these lines he is unconcerned with regularising metre. Thus, l. 45 is rhythmically slightly irregular in Hg, though in a manner not at all unusual in Chaucer’s verse. El removes at as it often does, presumably for stylistic considerations, but makes no attempt to regularise the metre: Wel come the sixte , whan eu e he shal (Hg) Welcome the sixte , whan eu e he shal (El) Line 108 is headless in Hg: Bad nat euery wight , he sholde go selle In El Gg and Ha4 he is missing, which makes the rhythm yet more irregular: Bad nat euery wigœt , sholde go selle Lines which remain rhythmically smooth in El and Gg in spite of a textual change are not very common. When such lines do occur it is often possible to see in them the same policy of stylistic revision. Thus, in l. 457 El and Gg have a less emotional version of a phrase, with wel instead of how in other manuscripts: How koude I daunce , to an harpe smale (Hg) Wel koude I daunce , to an harpe smale (El) In l. 7 El has a simpler syntax and a plainly affirmative verbal construction as compared to the complex modality in Hg and other manuscripts: If I so ofte , myghte han wedded be (Hg) ffor I so ofte , haue ywedded be (El)14 149 El significantly outnumbers Hg in metrically irregular readings. However, only a few of them are unique. In fact, there are about the same number of unique readings in El relative to the other manuscripts studied, as there are in Hg: that is, about five. These are mostly trivial and accidental in character. The present study has not revealed any process of regular editing in El independent of Gg and resulting in an increased number of metrically irregular readings. At the same time, it has not also revealed any consistent revision aimed at improving the metre, as is often claimed to be present in El. The large majority of metrically irregular readings in El are shared with Gg. The textual variants shared by El and Gg against Hg do not look like accidental changes, but like an intentional policy aimed at introducing stylistic corrections. This stylistic revision was intended to give more formality and neatness to the text, to ‘improve’ grammar, to make the style less conversational and more prose-like. In a large number of cases this revision is unconcerned with preserving metrical regularity. It seems that the editor wished to meet the requirements of a neutral and balanced prosaic style, and in an eagerness to remove syntactic inversions, ‘metrical words’ and colloquialisms often damaged the metre. Not all the changes shared by El and Gg can be so characterised. Some of the readings seem inadvertent and could be shared by coincidental variation. However, the evidence for the presence of this revision is considerable. In this context even trivial coincidences between El and Gg can be seen as the result of intentional interference. Gg in the second part of the prologue El Gg and Ha4 are very close to one another and distinctly separate from Hg at the beginning of The Wife of Bath’s Prologue, but in the second part El separates from Gg and Ha4 and joins Hg. According to Manly and Rickert (1940, II: 192) the affiliation of El and Gg ends around l. 387. The analysis of readings of Gg after that point seems to indicate that the text of Gg continues to bear the imprint of the same editorial policy that before was characteristic of both El and Gg. The changes in Gg as against Hg usually do not damage the meaning as accidental textual changes due to scribal carelessness may do, but rather seek to introduce a clearer prosaic style, less conversational, less emotional, more formal. In revising the style, the editor often removes ‘metrical words,’ producing metrically less regular readings than those of Hg. Gg also retains its association with Ha4.15 In l. 557 Gg has a unique reading with omitted thise: Ther fore , I made my visitacions To vigilies , and to processions To p chyng· eek , and to thise pilgrimages To pleyes of myracles , and to mariages (Hg, 555-558) To prechy~gis ek & to pylgry~agis (Gg, 557) This omission is likely to be due to stylistic considerations, as the presence of thise in the text of Hg and other manuscripts is justified by situation but not by Metre and Scribal Editing in The Wife of Bath’s Prologue 150 Canterbury Tales Project Occasional Papers II immediate context. In l. 606 Gg omits the metrically significant and, and makes the syntax formally more logical, while the rhythm is changed to trochaic instead of iambic: As help me god , I was a lusty oon And fayr· and ryche , and yong· and wel bigoon (Hg 605-606) As help me god I was a lustyon ffayr & riche and frosch & wel be gon (Gg 605-606) In l. 637 Gg omits is, giving the second half of the line a shape it would have had in prose: Stibourne I was , as is a leonesse (Hg) Styborne I was as a leonesse (Gg) As was characteristic of both El and Gg in the first part of the prologue, there are several instances when Gg has an explicative that absent from Hg and other manuscripts. This results in a less regular rhythm, but a more formal and logical syntax. Thus, l. 775: Bet is quod he , thyn habitaciou~ (Hg) Bet is ˇ he thyn habitaciou~ (Gg) In the following two lines the introduction of , the change from to reden to but reden, and from this to his are also likely to reflect stylistic revision: And whan I say , he wolde neu e fyne To reden , on this cursed book al nyght· (Hg, 788-789) And I saw he wolde neuere fyne But reden , on his cursede bok al ny't (Gg, 788-789) There are also examples where Gg omits ‘that’ in complex conjunctions, in the manner characteristic of El and Gg in the first part of the prologue: Who so buyldeth his hous , al of salwes (Hg, 655) Who so byldyth hise hous al of salwys (Gg, 655) And whan he say , how stille I lay (Hg, 797) And wha~ he saw how stylle I lay (Gg, 797) In l. 821 Gg removes a repetition: Keep~ thyn honour , and keep~ eek myn estaat (Hg) Keep thyn honour & thyn estaat (Gg) In l. 819 the change produces a more regular syntax: And whan that I hadde , geten vn to me By maistrye , al the soueraynetee And he seyde , myn owene trewe wyf (Hg 817-819) And wha~ne I hadde I getyn on to me By maysterye al e souereyntee Tha~ne he seyde myn owene trewe wyf (Gg 817-819) 151 In l. 801 the editor of Gg improves the sense in a rather unimaginative way: since the Wife of Bath was not actually murdered, he prefers her not to say that she was: And for my land , thus hastow mordred me (Hg) And º for my~ lond wilt Ł mordere me (Gg) These examples demonstrate that in the second part of The Wife of Bath’s Prologue Gg remains affected by the same editorial policy that it shared with El in the first part of the prologue: it is aimed at stylistic revision and unconcerned with preserving metrical regularity. Metrical regularisation in Ha4 Ha4 is another heavily edited manuscript but the editorial policy present in its text is very different from that of Gg and El.16 In contrast to the case with Gg, the interventions in the text of Ha4 show an interest in producing metrically regular readings.17 Whoever was responsible for these had a good ear for rhythm: the unique readings of Ha4 are often metrically smooth. In l. 383 Ha4 corrects a rhythmical irregularity present in all the other manuscripts analysed in this study: On Iankyn , and on my Nece also (Hg) On Iankyn and vpon my nece also (Ha4) In l. 825 it has a version different from other manuscripts but metrically regular: And also trewe , and so was he to me (Hg) And also trewe was he vnto me (Ha4) The spelling practices in Ha4 also indicate concern about metre. Its spelling usually reflects scansion very exactly. Though this does not hold for every case, on the whole the scribe intended all the vowels he spelt to be pronounced. There are very many cases when metrically redundant inflexional -e occurs in spelling (though presumably not in pronunciation) in all the manuscripts, but not in Ha4. Thus, for example, cacche/cach in l. 76: Cacche who so may , who renneth best lat se (Hg) Cach who so may who rennith best let· se (Ha4) frere/frer in l. 834: A frere , wol entremette hym eu e mo (Hg) A frer wil entremet· him euermo (Ha4) housbonde/housbond, seyde/sayd in l. 19: Is nat thyn housbonde , thus he seyde c teyn (Hg) Is nou't· in housbond us he sayd certayn (Ha4) lordynges/lordyngs in l. 4: Metre and Scribal Editing in The Wife of Bath’s Prologue 158 Canterbury Tales Project Occasional Papers II Gg and El can not be attributed to Chaucer. The arguments against this attribution are as follows: 1. The stylistic corrections in El and Gg often make the verse unmetrical. There is no doubt that Chaucer’s verse allowed some degree of metrical freedom and that rhythmical strictness is not always a proof of the authenticity of a line. However, the textual changes associated with the stylistic editing in El and Gg frequently go beyond what can be called rhythmical freedom or flexibility, and look more like a corruption of metre. They are often openly unmetrical and unpoetical. In changing the syntax and removing ‘metrical words,’ the editor destroys the iambic pentameter pattern without substituting another structure that would allow the line to conform to the metrical context even if with a considerable degree of freedom. The variants of El and Gg against Hg are often not ‘headless’ or ‘Lydgatian lines,’ but a plainly unmetrical, prosaic text. 2. The introduction of a more formal style and removal of colloquialisms impedes Chaucer’s expression of the characters through their speech. 3. Many of the changes favoured by the El/Gg editor conform to what has been described as typical ‘scribal responses.’ Windeatt (1979, 134-139) lists the following among the scribal changes common in the manuscripts of Troilus and Criseyde: a tendency to introduce the implied verb, or the implied subject or object of the sentence avoided in Chaucer’s laconic poetic syntax; a preference for more self-contained individual lines; the elimination of inversions; the introduction of a more predictable and prosaic word-order; and the insertion of relative pronouns. According to Kane’s (1988, 115-165) account of the patterns of scribal behaviour in the manuscripts of Piers Plowman the scribes were likely to introduce what seemed to them more correct, more easily intelligible variants. Kane remarks that they tended to more complete grammatical representation of meaning, to more explicit reference and more precise designation, to simplified language and connotation, all resulting in a more prosaic style. 4. If it is true that complex conjunctions such as while that, or sith that are characteristic of Chaucer’s linguistic usage, Gg and El tend to eliminate such usage. Changes could occur very easily in manuscript culture due to an attitude which permitted editorial interference with a text. It has been observed that the sacred character of a text did not necessarily stop scribal revision.25 Neither did high artistic quality or poetic authority, as the state of Chaucer’s texts demonstrates. Scribal editing as a common practice meant that in the case of verse scribes had to adjust their changes to fit the rhythm. Such adjustments could range from readings perfectly fitted into the rhythmical context, to those only roughly approximating the metrical pattern. The scribes were occasionally successful in regularising the rhythm of their copies and it is not necessarily true that scribal textual changes lead to looser metre and that free verse is an equivalent of poor verse corrupted by the scribes. It would be a simplification to see metrical freedom always as a result of corruption of a metrically regular 159 authorial text. Metrically regular variants of Hg and Ha4, often corresponding to metrically inferior variants in El and Gg, have a completely different status: such variants in Hg are likely to be original, whereas in Ha4 they are the improvements of a skilful reviser.26 At the same time not all the editors of the early manuscripts, even among the intelligent and responsible, were interested in metre. This analysis of the six early manuscripts of The Wife of Bath’s Prologue has shown two types of editorial practice: regularising the metre as in Ha4, and concerned with style but fairly indifferent to metre as in Gg and El. Understanding scribal attitudes towards the style and metre of their texts may help in estimating the textual value of different readings more than purely metrical analysis: there is no safety in adopting readings from editorialising manuscripts whether they are metrically regular or not.27 A study of a scribe’s tendencies in regard to wording, style and metre may allow us to determine as closely as possible the layers of editorial activity in a given manuscript and create grounds for discerning whether apparently ‘better’ readings are the result of scribal improvement, or of the preservation of the original forms. The conclusions made in this essay on the material of The Wife of Bath’s Prologue need to be further elaborated and explored across the whole range of The Canterbury Tales with reference to the early manuscripts. Notes 1For the present study I used collations of the six early manuscripts of The Canterbury Tales made available in the Lineated Collation of Hg El Dd Gg Cp Ha4 for the Wife of Bath’s Prologue, An Internal Publication of the Canterbury Tales Project #2, 6 March 1994, Office for Humanities Communication, 13 Banbury Road, Oxford OX2 6NN. I would like to express my gratitude to Norman Blake, Anne Hudson, and Peter Robinson for their comments and suggestions on drafts of this essay. 2This investigation was based on the study of lines which pose no problems of pronunciation, that is where the syllable count does not depend on vowels that could have been elided, or on inflexional endings that could have been silent, particularly on final -e. Ambiguous lines, such as the ones discussed below, were not used for metrical analysis. For example, l. 56 reads in El and Ha4: And Iacob eek· as ferforth as I kan (El) In Hg Dd Cp there is an extra-metrical syllable introduced by eu e: And Iacob eek· as fer as eu e I kan (Hg) However, this syllable consisting of -efollowed by a sonant was very likely to undergo contraction and to lose most of its syllabic value in pronunciation of verse. This is a very common metrical licence in Chaucer. Lines where it occurs were not included in the count of metrically irregularity lines. Another example of a metrically ambiguous line is l. 540, which reads: Metre and Scribal Editing in The Wife of Bath’s Prologue 160 Canterbury Tales Project Occasional Papers II That made his face , often reed and hoot (Hg and other manuscripts) That made his face , ful often reed and hoot (El Dd) The version of Hg is regular if read without elision of final -e in face. This pronunciation would be legitimate, for there is no indication that the rules for elision were absolutely strict in Chaucer’s verse. The reading of El and Dd is regular if the final -e in face was silent. In l. 400 the reading of Hg is metrically regular if the word kyndely is three-syllabic, whereas the reading of El is regular if it is disyllabic: To wommen kyndely , whil they may lyue (Hg) To wo~men kyndely , whil that they may lyue (El) Lines such as these were not used to assess the metrical superiority of one manuscript over another. They were disregarded in the count of metrically regular or irregular lines in the manuscripts and did not influence the conclusions or statistics. 3The ending -es in lordinges was probably not syllabic. This is born out by the spelling lordyngs in Ha4. 4The use of complex conjunctions with ‘that’ is discussed later in this paper. 5Cp has its own unique reading, most likely a result of corruption: Yet· hastow ought· and fals suspeciou~ 6Four of these lines occur before l. 77 where the text is missing in Gg. This leaves only two readings that are certainly unique. 7Unstressed to in this line probably behaved as a proclitic and merged in pronunciation with the first vowel of excusen. It is a common metrical practice in Middle English poetry, often testified by spelling in Chaucer and Hoccleve verse manuscripts (cf. Jefferson, 1987, 99.) 8In l. 51 the reading of El is shared by Si, a manuscript close to Gg in The Wife of Bath’s Prologue, which makes it likely that it was also shared by Gg. 9Hanna (1987, 92) includes alternation between conjunction and conjunction plus that, between parataxis with and, hypotaxis and no connective, and between wel, ful and no intensifier in a list of variant readings of Hg and El which are difficult to resolve. Understanding the direction of scribal editing in early manuscripts may assist in deciding between such variants. 10 Apart from lines discussed below these are ll. 12, 54, 59, 91, 121, 136, 173, 180 and 191. 11 These are ll. 12, 59 and 64. It is very likely that these variants in El are due to the editorial revision which El shares with Gg: readings in ll. 59 and 64 are supported by Si and Bo1, and in l. 12 by Si—manuscripts closely affiliated with Gg in The Wife of Bath’s Prologue. 12 The final syllable in speketh was probably syncopated as the spelling spek in Ha4 suggests. 13 According to Blake (1993, 13) some of the editorial ‘improvements’ present in the El text are stylistic. In particular the editor frequently altered the language to make parallelism or contrast more obvious. 161 14 This line is missing from Gg, but the reading of El is shared by Si, a manuscript close to Gg in The Wife of Bath’s Prologue. This line was not classed among the verses which are metrically less regular in El than in Hg, although it presupposes the absence of elision. As has been pointed out already there is no evidence that elision was always strictly observed in Chaucer’s verse. 15 A statistical study by Moorman (1993, 54-56) confirms that Gg is close to El in the first part of The Wife of Bath’s Prologue, and to Ha4 through the entire Wife of Bath’s Prologue. 16 Dempster (1946, 400) records Manly’s opinion that Ha4 is ‘characterised by editing almost as bold and extensive as is found in any CT manuscript of any date.’ 17 Tatlock (1909, 8-16) gives numerous examples of metrical revision in Ha4. According to him the reviser of Ha4 ‘had a much better ear, and much more independence, than most scribes.’ Tatlock recognises that some of the readings in Ha4 presuppose a good literary taste, but denies any possibility of them reflecting Chaucer’s own revision. 18 According to Manly and Rickert (1940, I: 222) ‘Ha4 is the earliest example of the commercial type of MS picked up from many sources and edited with great freedom by some one other than Chaucer.’ There seems to be further evidence that Ha4 is a contaminated manuscript from the following observations by Ramsey (1986, 140): Cp has one uncorrected unique variant per 53.81 lines, Ha4—one per 10.97, La—one per 10.85. If in Cp and La the rate of unique variants is the same from tale to tale, in Ha4 it differs greatly. Similar statistics are found in Moorman (1993, 61-63), who also points to a great range in the amount of unique variants in Ha4 from tale to tale, and connects it with the change of affiliations in various tales. The rate of change is greater when the scribe was using an exemplar with a B-type text. 19 In spite of the great differences between the C-version of the text in Cp and the text in Ha4 there are numerous textual similarities between Cp and Ha4 (see Blake, 1985: 96-122.) According to Pearsall’s (1983, 99) description of Cp it shows ‘some marked similarities with the kind of editorial activity that lies behind Ha4 (which is, however, on the whole more intelligently edited.)’ According to Manly and Rickert (1940, I: 93-96) ‘there is evidence that Cp and Ha4 were in the same shop at the same time.’ Some corrections in Cp may have been done from Ha4. 20 According to Blake (1985, 119) there is evidence that Cp was copied before Ha4 and that many features in Ha4 are explicable only on the assumption that Cp is earlier. Blake also argues that the scribe was more experienced when he was working on Ha4 than on Cp, among other matters in copying verse; he made fewer mistakes in copying tales written in stanza form than he did in Cp (ibid., 115.) 21 According to Manly and Rickert (1940 I: 102) Dd is much the earliest representative of the subgroup Dd of group A; ‘it is also most frequently away from its associates, both by shift of exemplar, probably due to loss of leaves, and by extensive correction (usually invisible), partly independent, partly from an unknown source near the original.’ Metre and Scribal Editing in The Wife of Bath’s Prologue 162 Canterbury Tales Project Occasional Papers II 22 According to Dempster (1946, 394-399) Manly believed Hg to have remarkably few accidental errors and hardly any editorial variants. He thought its scribe to be very accurate and conservative. As for the scribe of El, Dempster remarks (ibid., 398): ‘Somewhat surprisingly, though the scribe is believed to have been that of Hengwrt, characterised as “a very mechanical copyist” (cr. n. to A 3322) Manly tends to hold him responsible for a large share both of the accidental and the editorial variation in the Ellesmere text.’ 23 ‘Since different interpretations occur in copies produced by the same scribes it seems more likely that the scribes were following different instructions or different exemplars whilst executing different commissions than that they were responsible for the different interpretations themselves. We believe that the roles of B and D in the preparation of these copies were as subordinate as their roles in the preparation of the Trinity copy of Gower, and that at most they were responsible only for the realisation of each ordinatio in terms of preparing rather than planning the layout of these copies’ (Doyle and Parkes, 1978: 194.) A view different from Doyle’s and Parkes’ is held by Ramsey (1982, 1986) who defends identification of the El editor with the scribe. Analysing the unique variants in El which Manly and Rickert believed to be editorial, he observes that some of these variants are trivial and are shared, probably by accidental coincidence, with other manuscripts; some involve violation of an immediate context. According to Ramsey such ‘editorial’ variants lacking intentional policy and sometimes pointing to inadequate attention to the context are best explained as introduced by the scribe. 24 See footnote 20 above. 25 Greetham (1987), 62 observes that ‘as is well known in Biblical textual criticism sacred texts such as the New Testament seem to encourage rather than forbid scribal licence.’ Behind this is the wish to adapt such texts to the needs of the audience and to facilitate their understanding. 26 In spite of criticism metrically ‘correct’ readings from Ha4 long preserved their attraction for the editors. Dempster (1946, 400, fn. 117) observes that ‘Many Ha4 variants adopted by Skeat were rejected by later editors. But most editors, including Robinson (not Koch), follow Ha4 when its text has extra monosyllables intended to eliminate trochaic lines; see A 686, 752, 3350, B 1502, 1623.’ See also Pearsall’s (1991, 55) critique of editors adopting variants from manuscripts regularising the metre, such as Ha4 and Cp, and Moorman’s (1989, 102-103) discussion of Skeat’s use of Ha4 for emendation in spite of his statement of the pre-eminence of El and comparative worthlessness of Ha4. 27 According to Hanna (1991, 36) understanding scribal policies may help to resolve some metrical variants: ‘That preferable return to the manuscripts suggests one possible move toward a resolution of some metrical variants—although not a “ready expedient” like Greg’s rule. All Tales manuscripts, or some appropriate selection like the nine I cite, can be entered into a data bank and surveyed, not for possible rectitude but for simple scribal habit in specific verbal-metrical contexts. If one cannot openly determine the anteriority of one reading over another, one can determine how individual scribes react to certain possible lections. One could then answer such questions as how often Hg includes the word eek when attested elsewhere in the sample, how often it reproduces “headless” lines. Such a data bank may reveal a variety of distinctive scribal profiles...’ 163 Bibliography Blake, N. F. 1979. ‘The Relationship between the Hengwrt and Ellesmere Manuscripts of the Canterbury Tales.’ Essays and Studies n.s. 32: 1-18. — 1985. The Textual Tradition of the Canterbury Tales. London and Baltimore, Edward Arnold. —. and P. M. W. Robinson, Eds. 1993. The Canterbury Tales Project Occasional Papers. Oxford, Office for Humanities Communication. –. 1993. ‘Editing the Canterbury Tales: an Overview.’ In Canterbury Tales Project Occasional Papers I, ed. N. Blake and P. M. W. Robinson, 5-18. Oxford: Office for Humanities Communication. Dempster, G. 1946. ‘Manly’s Conception of the Early History of the Canterbury Tales.’ PMLA 61: 379-415. Doyle, A. I. and M. B. Parkes. 1978. ‘The Production of Copies of the Canterbury Tales.’ In Medieval Scribes, Manuscripts and Libraries: Essays Presented to N. R. Ker, ed. M. B. Parkes and A. G. Watson. London, Scolar Press: 163-210. —. 1979. ‘Palaeographical Introduction.’ In The Canterbury Tales. A Facsimile and Transcription of the Hengwrt Manuscript with Variants from the Ellesmere Manuscript, ed. P. G. Ruggiers. Norman, Ok., University of Oklahoma Press: xix-xlix. Fisher, J. H. 1988. ‘Animadversions on the Text of Chaucer.’ Speculum 63: 77993. Greetham, D. C. 1987. ‘Challenges of Theory and Practice in the Editing of Hoccleve’s Regement of Princes.’ In Manuscripts and Texts, ed. D. Pearsall. Cambridge, D. S. Brewer: 60-86. Hanna, R. W. III. 1987. ‘Problems of ‘Best Text’ Editing and the Hengwrt Manuscript of the Canterbury Tales.’ In Manuscripts and Texts, ed. D. Pearsall. Cambridge, D. S. Brewer: 87-94. —. 1988. ‘Authorial Versions, Rolling Revision, Scribal Error? Or, the Truth about Truth.’ Studies in the Age of Chaucer 10: 23-40. —. 1991. ‘Presenting Chaucer as an Author.’ In Medieval Literature: Texts and Interpretation, ed. T. W. Machan. New York, Center for Medieval and Renaissance Studies: 17-39. Jefferson, J. A. 1987. ‘The Hoccleve Holographs and Hoccleve’s Metrical Practice.’ In Manuscripts and Texts, ed. D. Pearsall. Cambridge, D. S. Brewer: 95-109. Kane, G., Ed. 1960. Piers Plowman: The A Version. London, Athlone. —. 1983. ‘The Text of the Legend of Good Women in CUL MS Gg.4.27.’ In Middle English Studies Presented to Norman Davis in Honour of his Seventieth Birthday, ed. D. Gray and E. G. Stanley. Oxford, Oxford University Press: 39-58. —. 1986. ‘“Good” and “Bad” Manuscripts: Texts and Critics.’ Studies in the Age of Chaucer Proceedings 2: 137-145. 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. Moorman, C. 1989. ‘One Hundred Years of Editing the Canterbury Tales.’ Chaucer Review 24: 99-114. —. 1993. The Statistical Determination of Affiliation in the Landmark Manuscripts of the Canterbury Tales. Lewiston, Queenston, Lampeter, Edward Mellen Press. Metre and Scribal Editing in The Wife of Bath’s Prologue 164 Canterbury Tales Project Occasional Papers II Owen, C. 1991. The Manuscripts of The Canterbury Tales. Cambridge, D. S. Brewer. Parkes, M. B. and R. Beadle, Eds. 1979. The Poetical Works of Geoffrey Chaucer: A Facsimile of Cambridge University Library MS Gg.4.27. Norman, Ok., Pilgrim Books. Partridge, S. 1992. ‘The Vocabulary of The Equatorie of the Planetis and the Question of Authorship.’ In English Manuscript Studies, 1100-1700 3, ed. P. Beal and J. Griffiths. Oxford, New York, B. Blackwell: 29-37. Pearsall, D. J., Ed. 1984. The Nun’s Priest’s Tale. A Variorum Edition of the Works of Geoffrey Chaucer. Norman Ok. and London, University of Oklahoma Press. —, Ed. 1985. The Canterbury Tales. London, George Allen & Unwin. —. 1991. ‘Chaucer’s Metre: The Evidence of the Manuscripts.’ In Medieval Literature: Texts and Interpretation, ed. T. W. Machan. New York, Center for Medieval and Renaissance Studies: 41-57. Ramsey, R. V. 1982. ‘The Hengwrt and Ellesmere Manuscripts of the Canterbury Tales: Different Scribes.’ Studies in Bibliography 35: 133-54. —. 1986. ‘Paleography and Scribes of Shared Training.’ Studies in the Age of Chaucer 8: 107-44. Robinson, P. M. W. and R. J. O’Hara. 1993. ‘Computer-Assisted Methods of Stemmatic Analysis.’ In The Canterbury Tales Project: Occasional Papers I, ed. N. F. Blake and P. M. W. Robinson. Oxford, Office for Humanities Communication: 53-74. Ruggiers, P. G., Ed. 1979. The Canterbury Tales, Geoffrey Chaucer. A Facsimile and Transcription of the Hengwrt Manuscript, with variants from the Ellesmere Manuscript. A Variorum Edition of the Works of Geoffrey Chaucer, vol. I. Norman, Ok., University of Oklahoma Press. Samuels, M. L. 1988. ‘The Scribe of the Hengwrt and Ellesmere MSS.’ In The English of Chaucer and his Contemporaries, ed. J. J. Smith. Aberdeen, Aberdeen University Press: 51-69. Tatlock, J. S. P. 1909. The Harleian Manuscript and Revision of the Canterbury Tales. Chaucer Society 2nd. ser. 41. London, Kegan Paul, Trench, Trübner. Windeatt, B. A. 1979. ‘The Scribes as Chaucer’s Early Critics.’ Studies in the Age of Chaucer 1: 119-41.