Some of Chaucer is missing

Authored by Christopher Cannon, Professor of English at New York University. Christopher is currently editing all of Chaucer for Oxford University Press with James Simpson, which will ultimately be published to Oxford Scholarly Editions Online.

My current project, producing a new edition of Chaucer for Oxford University Press, means that I'm thinking more than ever about the process of editing, the need for transparent decision-making, and how to present those editorial choices most helpfully to readers. The editing of Chaucer’s complete works from all extant manuscripts has really only been done twice before, by W. W. Skeat (1894) and F. N. Robinson (1933, with a revision in 1952) with a substantial revision of Robinson's second edition by Larry Benson and others in 1987 (as the Riverside Chaucer).  James Simpson and I are basing our edition on Skeat's. Much that Robinson and Skeat did does not need re-doing (it is hard to argue with the manuscripts they chose as the best for each text, not least because they generally agreed with each other). But there is much to decide about the details of each line of the text. Robinson tried to be faithful, above all, to readings in the surviving manuscripts, but this often means his 'Chaucer' doesn't scan smoothly. Skeat assumed that scribes made so many mistakes that it was just clear that they had ruined Chaucer's meter, and so Skeat sometimes printed words and word forms that no surviving manuscript contained in order to produce a regular line. So, to take one example, Robinson printed 'ferde' and 'there' in line 501 of The Book of the Duchess because these words have that form in Oxford, Bodleian Library, Fairfax 16, the manuscript he took to be the best surviving version of this poem:

But, if you pronounce the final –e's in 'ferde' and 'there', as Chaucer would have, the line has two extra weak syllables, which makes for a very bumpy tetrameter. Skeat fixed one bump by removing the '-e' from 'there':

But this still leaves an extra weak syllable, and a slightly irregular line, and so, if only to be true to Skeat's convictions, Simpson and I will go one step further and remove the offending '-e' from 'ferde' to yield a perfectly regular line:

The issues underlying these decisions are large: an editor of Chaucer must decide whether to print only what survives in authentic witnesses from Chaucer's day (or just after), and therefore assume that Chaucer sometimes wrote bad meter, or he or she must fix the obvious problems in the line and print a 'Chaucer' that editing has effectively invented. At the same time, each decision the editor makes to resolve these issues is nugatory. As I edit Chaucer, line by line, day by day, as much of my activity consists of putting in–e's, or taking them out, another large issue looms: who else will notice or care? In this day and age—since it has already been done twice—does editing of this kind really matter? 

And then I discover that some of Chaucer is actually missing. The gap opens at line 31 of the Book of the Duchess, where, if you are reading the poem in Fairfax 16, you see a very obvious change in script:

The nineteenth-century transcription of this manuscript, like all subsequent scholarship, describes this hand as 'quite late', opens a bracket here, and asks the obvious question 'Where' can these lines be 'from?':

This is not a question that Skeat allows any reader to ask, and his edition sails right over this bump, printing these lines as if they were seamlessly joined to what preceded them:

Robinson fails to make any mention of the problem in his notes, but it is not simply that there is a change in hand here. As Skeat does say (in the back of the volume), not only are these lines in a later hand in Fairfax—clearly added to the bottom of this page after it was thought to be complete—they can be found in neither of the other two surviving medieval manuscripts of the poem. They are there in William Thynne's edition of all of Chaucer of 1532, which seems to have been printed from good manuscripts since lost. But the gap that opens at line 31 continues to line 96, and so, no matter what you think of these 64 lines, they are missing from all the authoritative medieval witnesses to the Book of the Duchess.

But the Book of the Duchess also needs these lines: they contain the first half of the story of Ceyx and Alcione, and without their account of Ceyx's death at sea and Alcione's anguish over his continued absence, there is no obvious reason for the rest of the poem's subtle exploration of mourning and loss. The issues are once again large, but also the same issues produced by every wayward final -e: print the medieval witness exactly as it is found and produce a poem that makes no sense, or assume that Chaucer made sense and print lines that have no manuscript authority. The size of the problem has led more recent editors to the obvious solution: mend the gap, but make sure your reader knows what you've done. In their recent edition of Chaucer's Dream Poetry (1997), Helen Phillips and Nick Havely open a bracket at line 31,

which they then close at the end of line 96:

This is a deft blow to the Gordian knot, and it not only mends this tale but points a moral. Although Robinson never needed to think of it and Skeat was too firm of purpose to think it was necessary, every final –e that is put in or taken out—every small change made to ensure that Chaucer's meter remains regular—can be marked too. Simpson and I have worked out a system for such marking throughout. As we produce our text, line by line, and day by day, we will print the line we believe to be most authentically Chaucerian and we will tell our readers how we got there and why. It may not be that many more than us will care too much but out the changes. But since every subsequent reader will now know exactly what Chaucer is missing, and what is there, it has convinced us that editing of this kind really does matter.

Are you interested in using OSEO? Go straight in to some selected works, or watch our how to videos. If you don't have a subscription find out how to subscribe.