University of Virginia Library

Wynkyn de Worde's "Sir Thopas" and Other Tales
by
Thomas J. Garbáty

Wynkyn de Worde, Caxton's successor,[1] has not been treated well by commentators. He has been termed a "mechanic" and "in no sense a scholar," knowing little about the literary value of books. He has been faulted for not having high ideals and for printing solely for profit. Evidently he was extremely careless in his work, reprinting typographical errors without change. Henry Plomer, the main authority, repeatedly accuses the printer of lack of taste, "He was utterly devoid of all artistic feeling. He had no literary tastes. . . . All that can be said of Wynkyn de Worde as a printer is that, when he liked, he could turn out good work."[2]

No doubt much of this is true. But there has also been some confusion regarding the printed corpus of Wynkyn de Worde. Vagueness and incomplete data have led to imprecision and misleading statements in reference books[3] and, more important, in those scholarly works which

the academic community has long accepted as the foundations on which its variegated structures could safely be built. No one is at fault in this. Although perspicacious, we have all been like the proverbial blind men who described individual parts of the elephant but thought that they were defining the whole. Our information has been incomplete. My purpose here is not to attempt this final definition. All I wish to do at present is to describe my own part of the elephant by which he appears smooth, to the point, and well formed.

It has long been thought that Wynkyn de Worde, in printing the fourth edition of Chaucer's Canterbury Tales (1498), did almost no independent work, but followed Caxton's revised second edition (ca. 1484) exactly. In 1892 Thomas Lounsbury wrote: "The text that was found in the second [edition of Caxton] was the one followed in the editions brought out by the printers who came immediately after, Wynkyn de Worde and Pynson."[4] Lounsbury, in his appraisal, anticipated W. W. Greg's thorough study of 1924, in which this scholar compared the first 116 lines of the Knight's Tale with the six earliest editions. His conclusion is well known. "Indeed, I may say at once that it is clear that no print after the first was set up from manuscript; each successive printer, whatever alterations or corrections he may have introduced, set up his edition from one or other of its predecessors."[5] Regarding Wynkyn de Worde in particular, Greg felt that there was no reasonable doubt that he printed from Caxton's second edition, as did Richard Pynson in 1492.

It may be that Henry Plomer relied on these two scholars when he wrote, "Further reprints of Caxton's publications, notably the Morte d'Arthur, the Golden Legend, and Chaucer's Canterbury Tales . . . are the outstanding features of De Worde's press in 1498 . . ." (p. 58). Of course De Worde himself authorized this idea by printing the Prohemium to Caxton's second edition in his own work, ending, "By Wylliam Caxton / His soule in heuen won." Greg, however, does admit that Wynkyn de Worde possibly may have had recourse to one, unknown, manuscript because of De Worde's order of tales (AB1DE1E2F 1F2GCB2 HI) which is the order of Harley 7334, minus the tale of Gamelyn. Any

manuscript of the Ellesmere type might have suggested the necessary transposition of tales from Caxton's edition. Greg sums up: "Wynkyn de Worde printed his edition from Caxton's second, and it is probable that he too had recourse to some manuscript source, though this cannot be identified." Caxton's first edition alone, "ranks with the manuscripts as a textual authority" (pp. 760, 761). I might add that Eleanor Hammond also noted Wynkyn's affinity in order of tales to Ha4.[6]

Two of my fellow editors on the Variorum Chaucer, Professor Roy Pearcy (University of Oklahoma, Reeve's Tale) and Professor Thomas Ross (Colorado College, Miller's Tale) have arrived at similar conclusions in collating their tales. Both feel that De Worde was completely dependent on Caxton.[7]

In collating Chaucer's tale of Sir Thopas for the Variorum, I have come to a different conclusion. True, I have examined only one part of the elephant, but this appears like pure ivory and superior to similar sections of any other edition until the one by Thomas Tyrwhitt (1775-8). Indeed, in some aspects the Wynkyn de Worde edition of Sir Thopas surpasses even the work of this great eighteenth-century editor.

Admittedly Sir Thopas has many unique features. It is a short piece, extending to only 276 lines (including the Prioress-Thopas and Thopas-Melibeus link). It is written in 6 line tail rhyme stanza (aabaab), and in several manuscripts, El, Hg, Ad3, Dd, Gg among the superior manuscripts printed by the Chaucer Society and/or used for the Variorum, the scribe places the tail "b" lines to the right of lines a,a, slightly raised from the second "a" line and bracketed together, according to the following scheme:

Virgules are used as punctuation in occasional manuscripts. Such a page would have posed an intricate spacing problem and evidence of great care in composition for the scribe.

In addition to the peculiarities of page design, Sir Thopas has unusual textual problems which can be important shibboleths in bibliographical study. There are four short "tag lines" (1993, 2003, 2013, 2077).[8] Also, Sir Thopas, when interrupted by the Host, ends on a phrase (210) which is different or out in many manuscripts. The interruption and addresses to Chaucer by the Host are differently signalled or left out in the texts. Finally, several manuscripts have a gloss at line 1884.

From the evidence of these signals in particular and the whole text in general, I suggest that Wynkyn de Worde in the links and tale of Sir Thopas, as well as in a few other pieces to be mentioned below, ranks as a textual authority independent of Caxton. In fact, it seems from a comparison with the Hengwrt MS[9] that Wynkyn de Worde attempted, with Hg or some other MS of almost equal excellence, to print what we today would call a diplomatic text. I have collated the Hg version of Sir Thopas with twenty printed editions,[10] and then checked the variants with Manly-Rickert's Corpus of Variants. My evidence will be brief and will be presented in the following order:

• 1. Headings and Endings (Incipits and Explicits)
• 2. Gloss
• 3. Punctuation
• 4. Paragraph symbols
• 5. Page composition
• 6. Text

1. Headings and Endings: Manly-Rickert (III, 528) state, "Headings and endings are of little value for classification: the members of b̰ and d̰ can be roughly grouped by this method." They have not included incipits and explicits in their Corpus of Variants, since this work was done very thoroughly by Sir William McCormick (The Manuscripts of Chaucer's Canterbury Tales, 1933). The following is a comparison of Hg with Caxton 2nd edition, Wynkyn de Worde, Skeat, and Manly-Rickert-Robinson (as a group):

A careful search in McCormick of the 85 MSS and fragments of the Canterbury Tales extant has revealed only one MS which has any similarities with the edition of De Worde. This is Phillipps 6570 (Ph1), a defective MS of 24 ff. containing a section of the Pardoner's Tale, part of Group B2, and part of the Parson's Tale. B.M. Additional MS 9832, which contains the Legend of Good Women, belongs to this same MS. Ph1 shows the following incipits and explicits:

It should be noted that Ph1 also has the gloss at line 1884, a citation omitted in M-R.[11]

In the light of the exact reprinting of an unknown manuscript by Wynkyn de Worde, his incipits and explicits would seem to be trustworthy enough to identify his manuscript if we had such information available. Ph1 may have been involved in the descent of such a MS. It is obvious that De Worde did not follow Caxton except possibly in five words of the first incipit.

2. Gloss: Sir Topas has the gloss, ".i. Chaucer" above line 1884. This gloss is found in El, Hg, Ha2, Ht, La, Mc, Ph1, Ph3, Pw, Ry1 and Se. Wynkyn de Worde is the only editor to print this gloss.

3. Punctuation: Wynkyn de Worde is the only editor who has attempted to reproduce accurately scribal punctuation. The virgules seen in MSS like Hg, El and others are reproduced by WdW with full points. Of the 276 lines of Sir Thopas, De Worde reproduces the Hg pointing exactly in 218 lines; that is, 79% of the lines correspond, 21% diverge. Wynkyn de Worde has punctuation where Hg does not in 20 lines; Hg has punctuation where De Worde does not in 19 lines; De Worde has extra punctuation in 3 lines, and the rest show a divergency in punctuation placement. I have compared the variations of Wynkyn de Worde with the El pointing but find little consistent correspondence.

4. Paragraph symbols: The Caxton editions print Sir Thopas with no paragraphs, stanza indentations, or symbols. El, Hg, Ad3 of the Variorum manuscripts use paragraph symbols ¶ to introduce each new 6 line stanza. Wynkyn de Worde corresponds to Hg here with only 7 exceptions: in lines 1888, 1895 De Worde uses a space where Hg has a paragraph symbol (here De Worde corresponds to El) and in lines 1897, 2050, 2068, 2090, and 2102 De Worde has superfluous paragraph symbols.

5. Page composition: Wynkyn de Worde is unique among all editors in printing Sir Thopas exactly as we find it in the better MSS, with the tail rhyme "b" lines to the right and slightly above the "a" lines. As

noted above, the printer took unusual care to reproduce his manuscript exactly. The Caxton editions print the poem in a block and later editors in stanzas. None imitated the MS composition, and it is therefore obvious that only a MS set up in this fashion could have served Wynkyn de Worde as a base.

6. Text: My method has been to assign alphabetical letters to readings in the following way:

• a) WdW varies from Caxton but agrees with Hg.
• b) WdW varies from Caxton and varies from Hg.
• c) WdW agrees with Caxton but varies from Hg.

It will be helpful to have these formulas in mind when surveying the following summary. I have noted spelling variations only if they are significant, including seemingly slight variations where WdW differs from Caxton but the Pynson editions agree. Absolute—one might be tempted to say "servile"—dependence on Caxton by both Pynson editions (even with De Worde intervening), has been proved definitely by my collation. Certain important lines will receive special treatment below.

Summary: 1) In the 276 lines of Sir Thopas, Wynkyn de Worde agrees with Hengwrt (minor spelling variations excepted, but including paragraph symbols) in 238 lines (244 if paragraph sings or spaces are excluded). There is therefore an 86-88% correlation between Hg and WdW.

2) The number of lines in which WdW agrees with Hg but varies from Caxton is 90 or 33% of the lines. Four of these lines have two "a" readings, three have a "c" reading, and two have a "b" reading. There is, therefore, a total of 94 "a" readings.

3) The number of readings in which WdW agrees with Caxton but varies from Hg ("c" readings) is 16. Of these I consider only six (1912, 1923, 2017, 2021, 2048, 2113) as significant. These lines will be examined below.

4) The number of readings in which WdW varies from both Hg and Caxton ("b" readings, not including paragraph signs) is 16 (1902, 1921, 1926, 1941, 1972, 1984, 1993, 2001, 2003, 2013, 2014, 2038, 2050, 2077, 2123, 2124). These include the four crucial tag lines out, and three other important signal lines, all italicized and examined below.

5) The total number of readings in which WdW varies from Caxton is 110, in 104 or 38% of the lines.

Textual conclusions: Wynkyn de Worde did not use Caxton's second edition as a base for his own text of Sir Thopas although he may occasionally have referred to it (l. 1923). My information also shows that he usually anticipates or surpasses Tyrwhitt in better readings. The textual evidence corroborates the information concerning headings/endings, gloss, punctuation, paragraph symbols, and page composition to confirm WdW's independence of Caxton.

Is it possible to identify the manuscript that Wynkyn de Worde used? Although there is an unusually large correspondence with Hengwrt, there are enough crucial discrepancies (the out tag lines among others) to eliminate this MS as De Worde's specific or only base. Using Manly-Rickert's Corpus of Variants,[12] I have examined the occurrence of the most important signal lines in WdW.

• 1884 ".i. Chaucer": El, Hg, Ha2, Ht, La, M̰c, Ph1, Ph3, Pw, Ry1, Se
• 1941 "bay": b̰, Gg, Ha3, li, Ph1 (Cx1 = bay)
• 2014 "Thyn hauberke shall I percen": El, Gg, Ha2, Ha3, He, Ii, Ph1 (Speght 2,3 and Urry agree with WdW (M-R, IV, 2014: "The addition of 'Thyn hauberk' was apparently made by the ancestor of El and copied by the other MSS which have it. It is to be rejected.")
• 2107 "worly": Ch, El, Hg, Ph1, (WdW is the only editor to use this word)

The four tag lines:

• 1933 "so wilde" out: b̰, Ha2 (added later by original scribe), Ii, La (added later by original scribe), Ma, Ry
• 2003 "with mace" out: Fi, Ha2, (added later by original scribe), Ra2
• 2013 "Thy mawe" out: b̰, B̰o1, Ha3, Ii, Ma, M̰c, Ra1
• 2077 "In londe" out: b̰, B̰o1, Bo2, Fi, Gl, Ha2, Ha3, Ii, Ma, Mc, Ra3, Sl2

It will be noted that 3 of the 4 lines are out in b, Ii, Ha2, Ma. Ii also corresponds to lines 1941, 2014, and Ha2 to lines 1884, 2014 above.

Correspondence of "c" lines:

• 1912 "A Lord": b̰, Bo2, Dl, Ha3, M̰c, Mg, Ph3
• 1923 "brydgys": No MS authority
• 2017 "drewe": No MS authority
• 2021 "it" omitted: b̰, Bo1, Ch, Ds, Gg, Ha3, Ht, Ii, Ph1, Se
• 2048 "a lake": C̰n, Ḛn1, Gl, Ox, Ra3, To

Of 7 readings, there is a correspondence of 2 for b̰, Ha3, Se.

From a total of 13 variants counted, b̰, Ha3 and Ii show 6 correspondences, Ha2, Ph1 show 5, M̰c shows 4, Bo1, El, Gg and Se show 3. However,

only one of these MSS passes muster since all except one fail in one or more of at least three crucial ("a" line) points:

• 1) They lack the gloss at line 1884.
• 2) They write rode for glood at line 2094 (including El).
• 3) They have the final line 2108 "Tyll on a day" out. The one exception is Ph1, which, however, includes the tag lines.

It must be concluded, therefore, that Wynkyn de Worde did not use any specific MS found in M-R Corpus of Variants as his base. It is, of course, possible that he omitted the four tag lines through the exigencies of his page composition[13] and/or that he conflated several MSS and attempted an eclectic edition. However, in light of his strict adherence to the peculiarities of text and diplomatic printing, this theory does not seem probable.

And yet we should be able to determine the classification of the MS he used. The overwhelming correspondence with Hengwrt and the identification of the variants above lead one to believe that for the Sir Thopas section of his edition specifically (and possibly also for the other parts mentioned below) De Worde used a manuscript which figured in two of M-R's seven lines of descent for this tale:[14] 1) Hg-Bo2, Gg-Ph1-b̰* (b̰* = Ii-Ha3-b̰), and 2) El. The presence of b̰ in the first line of descent would correlate with some peculiar readings of De Worde from Caxton (l. 1923). In sum, Wynkyn de Worde may have used a "Hengwrt manqué," a slightly corrupt descendant of this fine manuscript.

It must again be stressed here, that these conclusions go against all the other evidence assembled by most reputable scholars, past and present. De Worde's use of an excellent MS for his base, almost totally independent of Caxton's second edition, his unique and precise reproduction in print of a MS page and MS punctuation, his anticipation even of Tyrwhitt in better and accepted MSS readings has been noted in the Sir Thopas section of his edition of the Canterbury Tales. Here he must take his place with Tyrwhitt as one of the fathers of the editorial method. In this section he shows vast superiority to Pynson and certainly proves an exception to Greg's statement that, among the early editions, no print after Caxton's first was set up from manuscript. Pynson set up his editions from Caxton, and Thynne from Wynkyn de

Worde, but Caxton's pupil, in this instance at least, was his own man.

But why here? And was it here alone? It is enticing to attempt to answer these questions by stressing the nature, difficulties, and challenges of the parody of Sir Thopas. This tale was, after all, important as Chaucer the pilgrim's own contribution. Also, no other work of the Canterbury Tales demands so much from an intelligent printer-editor, and Wynkyn de Worde may have risen to the challenge and searched for the best base text.

However, certain other facts of great interest, supplied to me by Mr. Howard Nixon, Librarian of Westminster Abbey, direct us to a less fanciful conclusion. I am very grateful for Mr. Nixon's permission to use his comments and shall quote from personal correspondence below. Mr. Nixon pointed out something obvious to anyone who does not neglect the forest for the tree, that Sir Thopas in this instance cannot be treated in isolation. The full points used there run through all of the Melibeus, Chaucer's prose tale, and they start shortly before the end of the Tale of the Prioress, which precedes the Prologue to Sir Thopas. The paragraphing in De Worde's Melibeus is different from Caxton's, and De Worde includes a six-line sentence at the head of the second column on r vib which is lacking in Caxton.

After the Melibeus, the Monk's Tale also seems to come from a source independent of Caxton as can be seen by the different order of the lives, for Caxton split up and rearranged the usual order of the Modern Instances. According to Nixon, "Caxton has the order which Chaucer probably intended, although it is not found in most of the best MSS. De Worde is evidently still following a manuscript of the Hengwrt group." In sum, it is obvious that part of the Tale of the Prioress, Sir Thopas, the Melibeus, and the Monk's Tale are all set from a manuscript and not from Caxton's edition. The reason for this involves another puzzle, and this in turn has led Howard Nixon to propose a solution which seems, under the circumstances, eminently sane and logical.

Incidentally, there is a bibliographical problem to be solved here. De Worde has a five-leaf section, u (not v, as Duff [E. Gordon Duff, The Printers, Stationers, and Bookbinders of Westminster and London from 1476 to 1535. Cambridge, 1906.] says), which starts in the middle of the Monk's Tale, 4 stanzas before the change in the order of the lives, by suddenly going into single column on u 1a and u 1b, then returning to double column on u 2. The Monk's Tale concludes on u 4a and the Nun's Priest's Tale begins on u 4b, continuing on u 5a and b and then straight on to x 1. Some very odd miscalculations must have been made, possibly because for this portion De Worde had both Caxton and his MS.

For I am fairly certain that the reason for this sudden 'improvement'

in De Worde's editorial method was that his copy of Caxton's second edition—which had been probably lying around the shop at the Red Pale for the last fourteen years—was defective at this point. De Worde had therefore to find a MS and print from that. And judging by the sudden appearance of the punctuation points in the Tale of the Prioress in the last stanza on r 1a of the De Worde edition, the defect began with a torn leaf ii 1 in Caxton, for this stanza is the fourth on the recto of that leaf. After that the whole of Sir Thopas, Melibeus, and the Monk's Tale may have been missing or De Worde's compositor may have gone on setting from the manuscript until he was told to stop.

We are left finally with the undeniable fact that De Worde was careful to choose a very good manuscript from a fine tradition for his text to replace Caxton, and that he has established himself thereby as the only printer who ever set up a page according to a rather complicated manuscript composition. In addition, no editor but he has ever accurately reproduced scribal pointing.

Wynkyn de Worde has been a much maligned old elephant, and he is long overdue some small place in the sun.

Notes