Authorial or Scribal? : spelling variation in the Hengwrt and Ellesmere manuscripts of The Canterbury Tales
Caon, L.M.D.
Citation
Caon, L. M. D. (2009, January 14). Authorial or Scribal? : spelling variation in the Hengwrt and Ellesmere manuscripts of The Canterbury Tales. LOT, Utrecht. Retrieved from
https://hdl.handle.net/1887/13402
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/13402
Note: To cite this publication please use the final published version (if applicable).
Conclusion
This dissertation consists of an analysis of the language of the Hengwrt and Ellesmere manuscripts of The Canterbury Tales, and folios 9-32v of the Trinity Confessio Amantis, with references to two other manuscripts, the Hatfield House fragment of Troilus and Criseyde and the Cambridge University Library, Kk 1.3/20 fragment of the Prioress’s Prologue and Prioress’s Tale; all of these texts were copied by Scribe B. My aim has been to try to shed light on the scribe’s spelling practice in order to explain on the one hand why he often used different spelling variants in Hg and El, and on the other hand to what extent these differences were due to the preservation of Chaucer’s language in each of these two manuscripts. In this chapter I will summarise my findings, and try to relate them to previous research on the same subject, in order to show that the spelling differences between Hg and El were chiefly due to Scribe B’s attempt to normalise the spelling in El rather than to linguistic changes.
To explain the reason of spelling variation between Hg and El, I first had to establish what kind of copyist Scribe B was. I believe that his spelling in Hg and Tr in particular shows that Scribe B was neither a faithful copyist nor a translator, but that he was, in McIntosh’s (1963) terminology, a Type C scribe, who mixed faithful copying with translation (see Chapter 2). I have shown that Chaucer’s spelling variants occur alongside forms belonging to the scribe’s own practice in Hg, and similarly, both Gowerian and scribal forms are found in Tr. To some extent the same can be said about El, although I think that the spelling in this manuscript is purposely designed to reflect Chaucer’s own practice. This does not mean that the scribe was more faithful to the exemplar when he copied El; rather, it indicates that he tried to normalise the spelling in this manuscript by employing forms that he considered to be more representative of Chaucer’s language. To normalise the orthography in El, Scribe B must have made choices whenever two or more variants of the same lexical item occurred in his exemplar. Even so, there is evidence that even in El this copyist rarely adopted a single form for words that exhibited several variants in Hg. However, he generally succeeded in decreasing spelling variation in El by selecting fewer variants for the same lexical items, and also by using one variant more frequently than others. In Chapter 3, where I discussed the use of single and double graphs for the spelling of long vowels, and in Chapter 4, where I dealt
CHAPTER 7 218
with the use of -ow-, -o- and -ou- for long and short [u], I demonstrated that all of these features are found in Hg and El. However, in El there is a clear preference for the use of double graphs for long vowels, especially in rhyming position, where the syllables in which they occur must be marked with greater weight, and for variants spelled with -ou-, all of them resulting from a pattern imposed by the scribe on the spelling of this manuscript. In addition, I also argued that when scribal forms can be distinguished from authorial ones in Hg and El, there is evidence that the scribe’s practice was to some extent influenced by the spelling of the bureaucratic language.
As a result, forms that regularly occur in the documents issued by the Chancery, such as aske and atte, are also found in these manuscripts, along with Chaucer’s preferred spellings axe and at the.
A final observation needs to be made about Scribe B’s practice. When looking at the images of folios 13r and 14v of another manuscript that has been ascribed to this copyist, Trinity College, Cambridge MS B.15.17, a copy of Piers Plowman, one cannot fail to notice the regular use of two old-fashioned characters, i.e. yogh, as in no©t, and thorn, as in þe, sooþlice and þis in the first four lines of the detail from folio 13r reproduced in Figure 1:
Figure 1. Detail from Trinity College, Cambridge, MS B.15.17, fol. 13r (Duggan 2007)
The use of these characters is rather striking, given that © is found just once in El, in the word thur© in PA par. 505, which is in the section of PA that is missing from Hg, while þ is attested more frequently in both manuscripts, but only in a restricted number of words, as explained in Chapter 6. Þ is mostly used in the abbreviation þt, and only occasionally it does occur in other words, such as troweþ (TM par. 206 in Hg) and deþ (PA par. 137 in Hg), often to save space at the end of lines of prose.
Horobin and Mooney (2004:87) argue that the Trinity Piers Plowman was copied before Hg, which could explain why it contains old-fashioned forms of the language.
Yet I want to suggest that if Scribe B is the copyist of this manuscript, he faithfully preserved these two characters from the exemplar in his copy, just as he did in his stint of the Trinity Gower. In that manuscript he employed þ very frequently because of the influence of the Gowerian language in the copytext. The absence of © and the
occurrence of þ only in certain environments in the two Chaucerian manuscripts therefore prove that these characters were used very infrequently or not at all by Chaucer. All this contributes to the impression that Scribe B was a conscientious and accurate copyist, a professional who did not limit himself to producing merely a faithful copy of his exemplar, but probably dealt with the language of his copytext quite systematically. Most likely, he normalised the spelling of the exemplar, as he was used to doing in the bureaucratic documents that he copied daily as a writer of the Court Letter (see Chapter 2), but, as a literary copyist, he also tended to preserve features that were distinctive of the author’s language, as they would give authority to his copy of the text.
After having established that the spelling in Hg and El is neither entirely authorial nor entirely scribal, I tried to explain why these two texts copied by the same scribe are characterised by spelling variation. For one thing, there was of course no standard spelling, as English spelling would not be fixed until the beginning of the eighteenth century (Scragg 1974:80). More importantly for this study, however, I pointed out in Chapter 2 that Samuels suggested that the differences between Hg and El are due to scribal adaptation to the ongoing linguistic changes, and that they therefore indicate a progression in Scribe B’s orthographic practice. I have demonstrated in many instances throughout this study that this, however, is not the case. Spelling differences between Hg and El are the result of a different approach towards the two texts, Hg being a manuscript meant to collect all tales in one codex, and El being the more prestigious version and beyond doubt a high-quality production of the same work. When discussing the importance of Hg and El as representative examples of Chaucer’s spelling, Samuels argues as follows:
which of those two, Hengwrt or Ellesmere, is to be preferred? The differences between these two MSS pose a problem, for although in general the forms used overall are the same, their distribution in the two MSS varies considerably, the earlier type of variants being more frequent in Hengwrt, the later type in Ellesmere.
(Samuels 1988b:34) This is exactly what I argue against in my study, by proving that there is no progression in scribal practice from Hg to El. Old forms are also attested in El, but there they are used intentionally, in order to lend authority to this manuscript. Hg should thus be considered a more genuine example of Chaucer’s language than El, because it is very likely that the spelling variants that I identified as authorial in this manuscript derive more or less directly from Chaucer’s working papers. The language of El, by contrast, consists of a selection of forms that probably belonged to Chaucer’s repertoire, and were thus chosen to represent the author’s own spelling;
for the most part they are old-fashioned variants, but there is also evidence of forms that we now consider to be more modern. It must, however, be specified at this point that since no Chaucerian holograph has come down to us, any statements about Chaucer’s language have to remain rather speculative. Comparison of all other fifteenth-century witnesses to The Canterbury Tales can shed more light on the
CHAPTER 7 220
nature of the text that was in the archetype of the tradition, thus helping scholars to identify spelling variants that were authorial. Yet, unlike Gower’s Confessio Amantis, there is no Chaucerian manuscript that can be used as a reference text for establishing with any degree of certainty how other versions of The Canterbury Tales relate to the authorial copy.
During the past decade, several Chaucerian scholars have put forward the hypothesis that Chaucer supervised the production of the earliest manuscripts, Hg and El in particular (cf. Stubbs 2000, Blake and Thaisen 2004). This may imply that Chaucer gave instructions about spelling issues in El, as indeed we find illustrated in his poem to Adam Scrivener (see Chapter 2), in which Chaucer expressed his concern for the preservation of authorial spelling in other copies of his works.
Stubbs (2000: Observations, The Treatment of the Cook’s Tale in Hg and El) argues that very likely Chaucer oversaw part of Scribe B’s work in both Hg and El and that the copying of both texts overlapped to some extent. Otherwise it would be difficult to explain why in both manuscripts the scribe left a space after the unfinished Cook’s Tale: half a folio before the end of quire 8 in Hg, and two and a half folios before the end of quire 6 in El. It would seem that when Scribe B copied the few lines of this tale in both manuscripts, he still expected to receive the missing part, and for that reason he left some space for adding it once the exemplar was made available to him. Later on, however, probably after Chaucer’s death, it became obvious that he would no longer receive any further text for this tale. He therefore wrote only in Hg, in the margin of folio 57v, the note ‘Of this Cokes tale maked Chaucer na moore’ with the same ink that he used for copying the tales belonging to Section II, i.e. the Wife of Bath’s Tale, the Friar’s Tale and the Summoner’s Tale.
This suggests that the note was written during or just after the copying of Section II.
This section was most likely produced after the tales that display the dark brown ink used for most of the manuscript, those found in Sections I, IV, V and the first part of Section III (see Appendix 3). It follows that Chaucer probably supervised about two thirds of the production of Hg, and that in the same period of time the scribe also copied at least the first tales of El, from the General Prologue to the Cook’s Tale.
The latter hypothesis is substantiated by evidence from the illumination in El. In her study of the Ellesmere illuminators, Scott (1995:94) identifies the work of three different limners in this manuscript (cf. Chapter 2), observing that ‘the shift from Hand A and Hand B occurs at a point where the text, the Cook’s Tale, ends midpage on folio 47v’, thus marking an interruption in both the scribal and the decorative processes.
Even if Chaucer supervised the production of a great portion of Hg, it is unclear how much influence he had on the spelling used by the scribe, since significant variation is also found within Hg in the tales that were probably copied when Chaucer was still alive. I found it particularly striking that the last two tales, TM and PA, which are in prose, display numerous old-fashioned spelling variants, while the first two tales, GP and KT, exhibit a greater number of modern forms alongside the old-fashioned ones. This is further evidence that the scribe received The Canterbury Tales as a set of disordered fragments, but it could also prove that GP and KT were
not the first tales to be copied in Hg. In her discussion of the Cook’s Tale (see above), Stubbs argues that PA was probably copied first and set aside, and this would indeed explain why several old-fashioned spellings are found in this tale: at that point the scribe was still getting used to the handwriting of the exemplar. In contrast, when he copied GP and KT he was already familiar with the handwriting, and he proceeded more quickly through the material. This is an attractive theory, although I do not discard the more traditional assumption that the first tales in Hg, GP, KT and MI, were also among the first to be copied. In this case, the presence of old-fashioned variants in GP, KT and MI would be explained by the fact that at this point of the manuscript the scribe followed his exemplar more faithfully, while getting used to the handwriting on it. The prose tales, TM and PA, were thus copied later, and it is possible that their exemplar, or exemplars, simply contained a large number of older spelling variants, which were often preserved by Scribe B, especially in Hg. He would do so partly because TM and PA were long texts, which he copied almost mechanically, as I suggested in Chapter 3, and partly because the old-fashioned variants may have derived from another exemplar, possibly a revised one, that was used for copying parts of TM, as I argued in Chapter 5.
The hypothesis that Scribe B was a speaker of a dialect similar to Chaucer’s London dialect has also been proposed to explain the similarities between the author’s and the scribe’s spelling practices. This would indeed be the case if Adam Pinkhurst, alias Scribe B, was indeed originally from Surrey, as suggested by Mooney (2006). Alternatively, it is conceivable that the scribe had become so familiar with Chaucer’s language that he adopted several Chaucerian spelling features in his active repertoire, and, as I argued above, that he deliberately retained authorial spellings. Some other dialectal variants are, however, regularly employed in Hg and also El, and are rather puzzling because they apparently do not have any special purpose, such as creating rhyming pairs or characterising some speakers in the tales. The most salient ones belong either to the Northern dialect, as shown by agayn, or to the Western dialect, as shown by muche(l)and murye. The Northern features have been accounted for by the hypothesis that they entered Chaucer’s language during the time he spent as a young page in the household of the Countess of Ulster in Yorkshire (Chapter 5). There is, however, no such simple explanation for the Western features, unless one postulates that they were in the exemplar of Hg, thus implying that O is not the direct ancestor of Hg, but of a manuscript containing Western variants. Yet, this is a theory that requires further investigation and that could probably be taken up further once all digital versions of the tales become available.
A final remark ought to be made with reference to Chaucer as a speaker of Middle English, a language that was not yet standardised, and that did not have a fixed spelling. A middle-class man by birth, he held several official positions in the king’s service throughout his life, and successfully rose to the top of fourteenth- century society. As a result of the numerous occupations he had in the course of his life, Chaucer was in touch with different social milieus in England and also abroad.
In addition, his career as a writer started with the translation of works in French (The
CHAPTER 7 222
Romaunt of the Rose) and probably the composition of poetry in French, and carried on with compositions in the vernacular, for many of which he was indebted to French, Italian and English writers such as Froissart, Machaut, Dante, Petrarca, Boccaccio, Gower and Langland. His vocabulary must therefore have been very extensive and have included words from other languages as well as from other English dialects. Likewise, his spelling practice must have included different variants, belonging to different stages of the language, since his orthography probably varied over time − he was after all not a professional scribe − and the language, the London dialect in particular, was undergoing rapid changes at the time. This would explain the presence of spelling variation in his works and especially in The Canterbury Tales, the object of this study. It would also help to clarify why the scribe preserved several variants in his manuscripts even when, as in El, he tried to normalise the spelling: these variants reflected Chaucer’s usage. I suggested above that Scribe B was a Type C scribe. However, when he worked on the production of El, he did more than merely mix faithful copying with translation.
I have argued in this study that he was basically torn between two ideals: on the one hand he tried to preserve Chaucer’s spelling, because this would lend authority to his copy of The Canterbury Tales, and on the other hand he tried to limit the spelling variation in El, in order to impose a pattern on it which would make this manuscript better that others, at any rate better than Hg.
The last word on Chaucer’s language has not yet been said, and this study has demonstrated the intricacies of attempting to analyse Middle English texts that derive from lost original versions. The application of computer technology to all fifteenth-century witnesses of The Canterbury Tales has opened new avenues of research in the study of these texts as well as in the study of Middle English in general and Chaucer’s language in particular. In recent times, scholars have emphasized that Hg and El can provide substantial evidence of the archetypal language of The Canterbury Tales only if they are collated with all other fifteenth- century witnesses, since later manuscripts and incunabula are likewise valuable sources of information (cf. Horobin 1997:20, Thaisen 2005–2006:392, Partridge 2007:350). My comparison of Hg and El with other witnesses in the tradition has enabled me to compare data from these two manuscripts with data from other texts in a number of tales. Yet, I believe that Hg and El represent an important stage in the textual tradition of The Canterbury Tales, because of the closeness of Scribe B to Chaucer. Not only did the copyist have authorial manuscripts at his disposal but he also benefited from his personal acquaintance with the author, who, in what most likely was a humorous warning, wished him the scalle if he did not copy his works faithfully. Scribe B’s privileged position allowed him to make well-informed choices about the variants to use in his texts, so that he passed on to us two manuscripts that are crucial for studies on Chaucer’s language.
