The General Prologue on CD-ROM (2000): Editor's Introduction
Elizabeth Solopova (University of Kentucky)
- The CD-ROM of The General Prologue
- This CD-ROM and the textual tradition of The General Prologue
- The Work on this Project
The General Prologue to the Canterbury Tales on CD-ROM offers the texts and the images of the forty-nine surviving fifteenth-century manuscripts and of the four early printed editions containing The General Prologue. It also includes word-by-word regularized and original spelling collations, spelling databases for each manuscript, and a database of all the spellings in all the witnesses, descriptions of the manuscripts prepared by Dan Mosser, software for textual analysis, an analysis workshop and a stemmatic commentary by Peter Robinson.
The manuscript transcripts reproduce the original spelling of the witnesses with abbreviations left unexpanded in accordance with the Canterbury Tales Project transcription system developed by Peter Robinson and myself and described in the ‘Guidelines to the Transcription of the Manuscripts of the Wife of Bath’s Prologue’ reproduced on this CD-ROM. Each transcript has a short introduction outlining the difficulties of the script which may have affected the reliability of some aspects of the transcription. Typically these difficulties are the lack of clearly distinguished emphatic letter forms or irregularity of word division in some of the scribal hands. The transcripts include notes which are linked by hypertext to the relevant lines and words, and provide commentary on textual and extra-textual features in the manuscripts not recorded in the transcripts themselves. The spelling databases group every occurrence of every spelling of every word in all the witnesses by headword and grammatical category. The witness descriptions are a result of a re-examination of the manuscripts by Dan Mosser and provide a wealth of paleographical and historical information. The CD-ROM includes a list of woodcuts from the printed editions with links to the corresponding images.
The most notable difference between this CD-ROM and the first CD-ROM published by the project, that of The Wife of Bath’s Prologue (1996), is the availability of two sophisticated computer programs which enable readers to make their own explorations of the textual tradition. The first of these pieces of software is Daniel Huson’s SplitsTree program. This program was developed for evolutionary biologists for reconstruction of trees of descent for organisms from information concerning the characteristics they share and do not share. The second piece of software is Peter Robinson’s VBase program, itself a derivative of his Collate program. This allows the reader to explore more precisely the relations suggested by SplitsTree by isolating the variants introduced by each presumed subancestor, and also offers means for identifying and coping with manuscripts which combine readings from different exemplars, either by contamination or shift of exemplar. Both programs are included on this CD-ROM and are linked to Analysis Workshop and Stemmatic Commentary sections, prepared by Peter Robinson. The analysis workshop introduces both programs, with many examples of their use and with exercises for the reader. The Stemmatic Commentary applies the picture of manuscript relations tentatively put forward in the Analysis Workshop to discussion of some points in the text where there is significant uncertainty or disagreement in the tradition.
The CD-ROM offers numerous choices. The reader can explore any of the fifty-three witnesses via the transcripts or images, compare several images and transcripts at the same time, take advantage of the complex searches which can be conducted with the Dynatext query language, or test further hypotheses with the additional tools for linguistic and textual analysis. In summary, users can construct their own methods for reading the General Prologue. We hope that this will be seen as a liberating and empowering feature: we linked together various parts of the edition in numerous ways to ensure that all the information can be quickly and easily accessed at any point of the many possible routes the readers may wish to take.
Our main task was to provide full access to the textual tradition of The General Prologue and tools for its further exploration. We believe that the publication of complete unedited texts of all the witnesses, and the availability of new techniques which assist evaluation of the manuscript support for each particular reading will give rise to further research on various aspects of content and language of this most widely read and studied part of The Canterbury Tales.
The most detailed existing description of the textual history of The General Prologue is found in Manly and Rickert’s The Text of the Canterbury Tales (1940). Manly and Rickert based their analysis on the study of forty-nine manuscripts: they did not include in their collation London, British Library MS Additional 10340 — a fragment of thirty two lines almost certainly written from memory — or any of the early printed editions apart from Caxton’s first edition. They believed that Caxton’s second edition could not provide textual information of any value because corrections from a superior source mentioned by Caxton were entered very inconsistently. As has already been pointed out this CD-ROM provides transcripts, images, and associated materials for fifty-three witnesses including Additional 10340, Caxton’s second edition — from our point of view an important source of textual information — as well as the printed editions by Pynson and Wynkyn de Worde.
Manly and Rickert’s analysis of the textual tradition of The General Prologue though correct in many details and full of useful insights is at times problematic. Predictably many of the problems lie in their research on the higher levels of the stemma where a textual critic has to work with scrapings of evidence, where minor findings loom large and seem to overturn the result of previous findings. It is almost certain that some areas of the early history of Chaucer’s text will remain dark in spite of our efforts and the advance of our knowledge. We hope however that the materials and the techniques we make available will encourage its further exploration, as our work on The Wife of Bath’s Prologue CD-ROM and its publication certainly did.
According to Manly and Rickert of the forty-nine manuscripts all but six — Hg, Ch, El, Gg, Do and To1 — are derived from the same common ancestor (1940: 3, 78-96): this is their ‘large composite group’. The exact relations among the surviving manuscripts are obscured by the loss of intervening exemplars, independent editing and contamination, and particularly by the later supply of lost leaves — something that happened probably more often in The General Prologue than in any other part of the Tales, due to its position at the front of each manuscript.
The loss of leaves is certainly the most serious obstacle in the way of textual research on the General Prologue: of the fifty three witnesses on this CD-ROM only thirty (Ad1 Bo1 Bo2 Cx1 Cx2 Ch Ds1 En1 En3, El Fi Ha2 Ha3 Ha4 Hg Ht La Ld2 Ma Mg Ph2 Pn Ps Pw Py Ry2 Se Tc1 To1 Wy) have a complete or near-complete text. In four manuscripts as we now have them, the beginnings are supplied in post-medieval hands: the first 66 lines of Ii, the first 79 lines of Lc, the first 133 lines of Ld 1, and the first 55 lines of Tc2. The substitution of the lost text with text from other copies evidently happened in the ancestors of the surviving manuscripts as well, causing the shift of exemplars.
The clearest problematic areas of Manly and Rickert’s analysis are the lack of evidence for their large composite group (which may lead us to doubt its existence; see the discussion of this group in section 3.2.1 of the Analysis Workshop and of their methods and results in section 3.4.3 of the Analysis Workshop); the position of Ha4 on the stemma; the insufficiently explained relationship between Hg and El, and Hg and Ch; uncertainty about the nature of the agreements between El and Hg and the pair Bo1 and Ph2 and the relation of this pair to Py; failure to take advantage of the evidence from Caxton’s second edition; and numerous questions raised by their independently derived manuscripts. Our edition does not necessarily promote a particular view of the textual tradition of The General Prologue. The primary aim of Peter Robinson’s Analysis Workshop is to demonstrate the computer assisted techniques of textual research and to encourage scholars to explore these techniques. The answers to the questions raised in the Analysis Workshop are provisional; they are given in order to be tested and in doing so to stimulate further discussion.
Full access to the textual tradition of the General Prologue is also much needed because the ‘standard’ editions of the
Canterbury Tales Project have their own problems, often admitted to by their editors. I will concentrate here on the limitations of Larry D. Benson’s Riverside Chaucer, the edition most commonly used at the university level (Benson 1987.) The editors of the Riverside Chaucer observe in their introduction that ‘[F. N.] Robinson, like all the editors who had preceded him, prepared a text after consulting only a few of the manuscripts’ (Benson 1987, 1119.) They further note the drawbacks of the Chaucer Society’s Eight-Text Edition — Robinson’s ‘primary source for readings’ — and the fact that Robinson inherited the late Victorian tradition which saw such codices as El, Gg and Ha4 as ‘the central form in which Chaucer’s text had been transmitted.’ The editors then conclude: ‘For our textual presentation, we adopt the same eclectic (and perhaps not completely consistent) procedures used in Robinson’s second edition. The text of the Tales remains based, as was Robinson’s on El. However in the light of Manly and Rickert’s elaborate demonstration, one can no longer, as Robinson recognized, follow El in every lection.’ The editors inform us that they ‘continue Robinson’s procedures by introducing a substantial number of additional Hg readings,’ but note that they believe the text they print still to be Robinson’s, and that they prefer to present a ‘hybrid’ rather than to ‘switch copy texts’, since they remain unconvinced by Manly and Rickert’s argument that El represents the text ‘editorially sophisticated.’ This description of editorial procedures and of the complicated ‘history’ of Robinson’s text before its appearance in the Riverside very well demonstrates the uncertainty which surrounds the text and the choice of the best editorial method.
Another well known problem with the Riverside is inaccessibility of textual information. The fact that the textual notes quote only nine or ten manuscripts makes the evidence they provide unreliable. For example the first textual note for the General Prologue reads:
Aprill] Aprille Skt Pol Rob1; the -e appears unambiguously only in Ha Pw. One of numerous efforts by earlier editors to follow MSS that edited out Chaucer’s headless or nine-syllable lines, cf. 217 below.
A look at the all-witness spelling database on this CD-ROM shows that -e is unambiguously present in another eight manuscripts (Ad1 Ds1 En3 Fi Ph2 Pn Tc1 Tc2) not quoted by Riverside. More importantly three witnesses (one of which is Hg, the others Ld2 and Ch) have the reading Aueril: Aueryll Hg Ld2 and Auerell Ch. Though the variant Aueril appears to have poorer manuscript support in this line than the variant April, it can not be discounted since it occurs in such an important witness as Hg, and because so many manuscripts lack the first line of the Prologue, so producing a very distorted picture in terms of numbers anyway.
There are two further occurrences of the word ‘April’ in The Canterbury Tales: in line 6 of The Man of Law’s Prologue (MLP) where Hg has April and El has Aprill, and in line 546 of The Wife of Bath’s Prologue where Hg has Aueryll and El has Auerill. In the Wife of Bath’s Prologue the manuscripts are approximately equally divided between the two readings. The fact that the reading Aueril appears in Hg and El, both of which have a text of a very high quality in the second part of the Wife of Bath’s Prologue, makes its authorial origin here quite possible, and so also in other parts of the Tales. Chaucer himself could have used both variants in his verse. A better understanding of the status of Hg, of the relationship between Hg and Ld2, and particularly between Hg and Ch, as well as a better understanding of the textual situation in the MLP, will provide further insights into this problem. Whichever variant is authorial does not affect, however, the question of how the line was scanned. There is little doubt that both variants no matter how they are spelled were disyllabic, stressed on the first syllable, and little doubt that the line was trochaic. In all the contexts where these words occur in The Canterbury Tales, spelt with or without the final -e, or with a crossed single or double l, the metre requires their disyllabic pronunciation, as can be seen from the Hengwrt concordance (Blake 1994.) The -e which follows the ‘u’ in ‘Aueryl’ was silent: its function was to indicate that ‘u’ was a consonant rather than a vowel. This common Middle English convention can be seen in many examples in the spelling of the Hg, El and other scribes. The spelling database on this CD-ROM shows an almost unanimous agreement of the manuscripts in using ‘e’ after ‘u’ in unabbreviated spellings of such words as ‘euery’ or ‘euerich’. These words require a disyllabic pronunciation in all their contexts in The Canterbury Tales. Finally, from what we now know of Chaucer’s metrical system, trochaic lines in a iambic context are likely to have had their legitimate part to play. There is evidence that one of their stylistic functions was to mark the beginning of a new section and the first line of the poem certainly falls into this category. The information about substantive and spelling variation across the tradition necessary for examination of individual readings such as ‘April’ in this example is not available in the Riverside, or in any of the existing printed editions. See further the Stemmatic Commentary, on line 1.
Finally, some readings accepted by the Riverside require re-examination, which this CD-ROM can enable. The reading ‘atte’ which appears in line 29 in The General Prologue in the Riverside Chaucer occurs only in ten out of thirty four manuscripts (Cx1 El Ha4 Ht Lc Pw Ry2 Tc1 and in Cx2 and Wy dependent on Cx1. The readings in Ld2 atte the and Tc1 ate are uncertain.) On this and related readings with the alternation at the / atte see the discussion on line 29 in the Stemmatic Commentary. In line 131 the Riverside reading is:
The variant ‘ne’ occurs in nineteen out of thirty nine witnesses, namely in Bo2 Cp Ds1 El En1 Fi Gg Ha2 Ha3 La Lc Ld1 Mg Mm Pw Ry1 Ry2 Se Sl2. Eleven of these manuscripts belong to Manly and Rickert’s group cd, and both Fi and Ld1 may also have the reading from a cd exemplar. Support for this reading outside these is limited to El Gg Bo2 Ry1 and the closely related En1 and Ds1. In this case, as in line 29, the Riverside follows El without attempting to examine the complete textual evidence. See further the Stemmatic Commentary, on line 131.
The search for the ultimate source of the textual tradition or for ways of improving the standard edition of The Canterbury Tales should not overshadow the value and interest of every one of the fifty-three texts of The General Prologue. This value can not be measured only by how well they reflect Chaucer’s original. As a product of medieval culture Hg is no more important than (for example) Ht whose cheeky readings often resist classification under the convenient label of scribal corruption. One really wonders about the scribe’s agenda when instead of the expected
And bathed euery veyne in swich lycour
in line 3 one reads
And bathed euer in wyn in suche licoure
or (in line 68) instead ofAnd thogh that he weere worthy / he was wys
And thought that he was worthi for he was wise
or instead of (line 470):
Gattothed was she / soothly for to seye
Catte tothed she was sothli to sey
A principal point of contact between Chaucer and his critics is the interest in textual authority, though the critics can be said to share this interest out of necessity rather than by choice. The persistence with which the question of authority appears in Chaucer’s work shows its tremendous cultural significance. Chaucer seems to have been on the whole very sceptical about authoritative texts and about the possibility of successful transmission of information. Nor, it seems, was he accustomed to creating authority by finishing and publishing his work. His editors from medieval scribes to modern scholars devised many ‘authoritative’ texts, often eclectically created from parts of ‘less authoritative’ texts. At least some of the methods of endowing the text with authority are common to both the fifteenth and the twentieth centuries. They include, predictably enough, external beauty, an impressive critical apparatus, and surprisingly the regularisation of metre. The General Prologue on CD-ROM has not abandoned the ‘single text’ method entirely, nor denied authority completely: after all the base text with which it opens is a lightly edited text of Hengwrt. In the supplementary materials we often find ourselves advocating or undermining the authority of the existing editions, or that of the existing (and sometimes even only postulated) manuscripts. However the present edition gives the less ‘authoritative’ witnesses full scope to present their evidence. In the same way a fuller and better account of previously neglected textual evidence can make a single text of a standard edition seem less tyrannical.
Since the history of the Canterbury Tales Project as a whole is described in the editor’s introduction to The Wife of Bath’s Prologue on CD-ROM (Robinson 1996), I will briefly relate here the progress of our work on the manuscripts of the General Prologue. I started working on the General Prologue in Oxford in March 1996 after the completion of The Wife of Bath’s Prologue on CD-ROM. Lorna Stevenson joined me in Oxford in April that year. Our work was funded by a three-year grant from the Leverhulme Trust, for work on the manuscripts of ‘Fragment One’ of the
Canterbury Tales Project. This grant paid the salary and direct costs of the work. A grant from the Oxford University Research and Equipment Committee purchased necessary equipment, and a grant from the News International fund administered by the Oxford University English Faculty covered the rent for an office at the Centre for Socio-Legal Studies at Wolfson College. Both Lorna Stevenson and I were employed by Sheffield University which held the grant from the Leverhulme Trust (with further support from the British Academy) and provided continuous administrative and academic support.
The first stage of work, the transcription and collation of all the manuscripts, was finished in November 1997. We are indebted to the Sheffield branch of the Project for the first two or three transcripts of manuscripts Ad3 Cp El Gg Ha4 Hg and La. All the manuscripts were transcribed by Lorna Stevenson and myself and proof-read at least three times against the paper copies printed from the microfilms. The following witnesses were also checked against their originals: Ad1 Ad3 Ad4 Bo1 Bo2 Bw Ch Cp Dd Do En1 En3 Gg Gl Ha2 Ha3 Ha4 Ht Ii La Ld1 Ld2 Ma Mm Ne Pw Ra2 Ra3 Ry1 Ry2 Se Sl1 Sl2 Tc1 Tc2 and To1. The transcripts were collated by Peter Robinson and myself using Collate, the program developed by Peter Robinson for collation of large textual traditions. As a part of the collation process every word in every manuscript was assigned to a lemma, and its grammatical information, such as the part of speech and grammatical form, was recorded for later use in the spelling databases. A material help in this process was a lemmatization and regularization of some manuscripts of the entire General Prologue prepared by Estelle Stubbs and Michael Pidd, of the Project’s Sheffield team. The spelling was regularized to create the regularized collation for users interested only in substantive variation and for the purposes of the variant database and stemmatic analyses.
We started the proof-reading of all the materials in April 1998. This required a word-by-word check of all the collations and the spelling databases; validation of the various translation phases the text underwent in its journey to electronic publication; testing of the many kinds of link provided on the CD-ROM. These checks were first done against printouts, and then finally (and to a background of increasingly-tight deadlines) against a succession of proof CD-ROMs. I am grateful to Estelle Stubbs for the proofreading she did at these final stages of work and to Adrian Welsh of De Monfort University for undertaking much of the data conversion and validation for the CD-ROM.
I am grateful to the Leverhulme Trust, Sheffield University and the English Faculty of Oxford University for their sponsorship of this project. Without the support of the Leverhulme Trust, in the ‘Fragment One’ project and in the earlier Computers and Variant Texts and Collate projects, this project (and indeed, the whole Canterbury Tales project) simply could not have happened. I am deeply indebted to Sheffield University for the generous support given to me on numerous occasions. I would like to thank the owners of the manuscripts of the
Canterbury Tales Project (listed in the section ‘Manuscript Acknowledgements’) for their permission to reproduce the images of the manuscripts on this CD-ROM and for giving Lorna Stevenson and myself the privilege of checking our transcripts against the originals. We have benefited from the expertise of the staff of many libraries and archives in England: we are very grateful for their generous help. I would like to thank particularly the staff of the Bodleian Library where so much of the work on this project was done.
I am grateful to the contributors to this CD-ROM, to the members of the
Canterbury Tales Project Project at Oxford, Sheffield and Brigham Young Universities, and to the research associates of the Project: Norman Blake, Linda Cross, Simon Horobin, Beverly Kennedy, Darin Merill, Dan Mosser, Stephen Partridge, Michael Pidd, Peter Robinson, Lorna Stevenson, Estelle Stubbs, Claire Thomson and Paul Thomas. My work has profited enormously from them and would not have been possible without their contribution. I am indebted to Norman Blake and Peter Robinson for the support they gave me as general editors of the Canterbury Tales Project. I am particularly grateful to Lorna Stevenson with whom I worked closely for two years and who was both a supportive friend and a stimulating and inspiring colleague.
I am deeply indebted to my colleagues and friends: Anne Hudson, Tony Hunt, Linne Mooney, Eric Stanley, Peter and Sylvia Neumann. It is no exaggeration to say that without their support generously offered on countless occasions, this project would not have been finished. I am very happy to take this opportunity to thank Michael Fraser and his family for all their love and support. My greatest debt in this as in every other achievement is to my sister Catherine Fet — an admired example and always dependable support.