University of Virginia Library

The First Edition of The Abbey of the Holy Ghost
by
Curt F. Bühler

In the Spring of 1952, the Pierpont Morgan Library acquired the only known copy of a previously undescribed English incunable. Such an event, especially in the case of a work of literary interest written in the vernacular, is not one of routine occurrence; it has consequently been thought proper that this new "find" be given a somewhat more extensive discussion than the familiar "previously unknown" incunabulum normally deserves. The results, it is believed, will not only fix the approximate date and the proper order in which the several editions appeared, but will also indicate to bibliographers the value of textual study in the solution of bibliographical problems.

In the form advocated by the present writer (Standards of Bibliographical Description, 1949, pp. 3-60), the new incunable may be described in these terms:

Two other fifteenth-century editions of the Abbey of the Holy Ghost have long been known and are described by E. Gordon Duff[1] under his numbers 1 and 2. I have been able to consult the edition identified as Duff 1 by using the facsimile prepared from the copy in the University Library Cambridge by Francis Jenkinson (published by the Cambridge University Press in 1907), while the Folger Library has kindly supplied me with a microfilm of their copy of the remaining edition (Duff 2). With a view towards determining the order and approximate date of the editions, it will be necessary to subject these not only to a strict bibliographical scrutiny but to a thorough textual examination as well. We may best begin by considering the physical "make-up" of the editions which, for the sake of convenience, will hereafter be cited as C (Cambridge), F (Folger) and M (Morgan).

As we may see from the description given above, M was printed with only one column to the page while the other two editions were each set up with double columns. One fount of type only (Wynkyn de Worde no. 4) was used in the printing of M, whereas two different sizes of type are found in both other editions (types 2 and 4 in C; types 4 and 6 in F). Furthermore, C and F have a constant number of lines (29) to the page; M usually has 28 though thrice 29 lines were needed to accommodate the text. Similarly, on signature a4 verso of M, de Worde miscalculated the amount of text he could crowd onto a page and was obliged to set a portion of the word "ravisheth" in the position normally occupied by a catchword; thus, by printing "ra||uyssheth||", he was able to get all the text on that page. Again, while c4 and d4 are missigned in M as noted in the above description, there are no errors in the signing of the other editions. Finally, the text of M ends on the verso of the last leaf, whereas in both other editions the narrative concludes on the recto and is followed in each case by a printer's device. All these points suggest that M is the most primitive (and consequently the earliest) of the three editions.

The evidence of the woodcuts, while inconclusive, at least confirms the belief that C was printed before F. The woodcut of the Crucifixion (Hodnett 374)[2] has suffered greater material damage at the top of the

block in F than in C; in addition, the Folger copy shows the loss of a border-line at the top and another at the bottom of this cut.

If we now turn to the textual examination, it is clear that six different arrangements for the successive order of appearance of the editions are possible. Two series (C → M → F and F → M → C) can be eliminated immediately, both for typographical and for textual reasons. Since C and F were each printed with double columns and always agree page-for-page (and mostly line-for-line), it is self-evident that the one was set from the other with no single-column edition intervening. The text too corroborates this theory, as one example will suffice to indicate. In line 4 of sig. c4v, M has the text "þt were callid Merci & Pyte," while both C and F omit the word "callid". Now one will find it difficult, probably impossible, to believe that the second compositor could supply a needed word which in a third printing was then again omitted, thus accidentally duplicating the reading of the editio princeps. Yet one would be obliged to accept this coincidence if the order of appearance was either of those given above. No; one can only conclude, on the basis of reasonable probability, either that C was set from F or that the contrary took place.

The next two possibilities (C → F → M and M → F → C) are equally unacceptable, for C and M often have the identical text from which F alone departs. For example, in M (b1, line 4) one reads: "For no man I trowe myght al fully it fele"; with this C completely agrees. In F (a6v, β, 17-18), however, we find: "for I trowe þt no man myght all fully it fele". To cite another example, M (a6v, lines 14-19) prints the following:

While C offers this same text, F (a6v, α, 16), sensing an omission, alone supplies the word "understanding", giving the reading "lyght of vnderstandynge".

That M and C are textually related—and that the one was set from the other—is convincingly demonstrated (if further proof be needed) by a curious slip common to both; this does not occur in F. On b6v, lines 2-5, M reads:

With this reading C agrees, while F, in accordance with tradition, correctly asserts that, before the coming of Christ, Adam had languished in Hell 4,600 and more years ("and there hath be now .iiij. M. yere and .vi. C. and more"). Thus, if we assume that F intervenes between C and M, we must believe that the error in time was corrected for the Folger text only to have the third edition, by sheer coincidence, make the identical error found in the first printing.[3] The points here made surely eliminate these two possibilities.

We thus arrive at the last two alternatives.[4] The series F → C → M seems improbable for typographical reasons. The state of the woodcuts assures us that F is the latest of the three, and the typographical "set-up" of M is certainly more primitive than the sophisticated press-work of the other two editions. Textually too, this is unlikely, for M and C exhibit older dialectal and orthographic variants ("her", "syndre", and the habitual use of "wol") where F offers the more modern forms ("theyr", "sondry", and usually wyl").[5]

By this process of elimination we arrive at the conclusion that the order M → C → F is the only possible one for the appearance of the three editions of the Abbey of the Holy Ghost. Furthermore, there are an adequate number of pieces of evidence to support this contention. In illustration of this, the text in M (d3, lines 11-14) may be cited:

The term "too short to the hole" had been used a few lines earlier in very similar circumstances; in that case, all three editions employed this same phrase. In the instance previously cited, however, both C and F substitute (d2v, β, 3) "to sort for þe hole", an apparent modernization introduced by the later printings.

The gradual deterioration of the text (M → C → F) is amply demonstrated by one enlightening example.[6] In M (a6, lines 15-19) one reads:

This appears in C (a6, α, 17-24), where the omission is clearly due to haplography, as:[7] The compositor of F, sensing a corruption of text, set about "correcting" it, and with this in mind he substituted "losyd" for "loued" (loved). In the special Biblical sense of "to loose" (NED vb. Ic) this emendation gave meaning to the passage though it did not result in a return to the original sense. The explanation for the texts of C and F can only be made evident by assuming the priority of M. Surely it is clear, from this lengthy analysis, that the newly discovered copy of the Abbey now in the Morgan Library belongs to the earliest known printed edition of this text.

A fortunate set of circumstances makes it possible to supply a likely date for the several editions of the Abbey of the Holy Ghost. As we have seen, the Morgan copy of the Abbey was printed in a single column with de Worde's type 4; Duff I was set up in two columns with types 2 and 4; and Duff 2 was issued with double columns using types 4 and 6. Now the first edition of Bishop John Alcock's Mons perfectionis (Duff 12) was also printed in one column with de Worde's type 4, while the second edition (Duff 13) has double columns with types 2 and 4; a third edition was issued by de Worde after he had left Westminster (STC 281). Are we not justified in assuming that the Abbey and the Mons perfectionis, being similar tracts of devotional literature, were issued as companion pieces and that they were designed to be sold either separately or together?[8] Since the first edition of the Mons perfectionis has the printed date 22 September 1496, and the second bears the date 23 May 1497, it is with

some confidence that one may assert that the first two editions of the Abbey were also produced, respectively, in September 1496 and May 1497.

The third edition of the Mons perfectionis presents an added complication in that it was printed after 1500, at a time when Wynkyn de Worde had left Westminster and had moved to "the City." According to Sayle (no. 144),[9] the volume states that it was printed in London "and fynysshed the xxvii. daye of þe moneth of Maye" 1501. We may well suppose that this book was intended to go on sale together with the third edition of the Abbey of the Holy Ghost; consequently it is likely that little time had elapsed between the printing of these two volumes. On the basis of this reasoning, there is some justification for believing that the third Abbey (Duff 2) may well have been the last book printed by Wynkyn de Worde at Westminster, just as the third Mons perfectionis is judged to be the first one issued by his new press in London; in any case, no earlier book with a London imprint has yet been discovered.[10]

This discussion, then, offers what seems to be fairly conclusive evidence in support of the contention that the Morgan copy of the Abbey of the Holy Ghost belongs to the first printed edition of this work (at least, of those that are now known to us) and that this work was probably printed by Wynkyn de Worde not much later than 22 September 1496. In concluding this somewhat lengthy account, one further observation may be made. If Duff's bibliography (or if the Gesamtkatalog der Wiegendrucke) should ever be reprinted, the present edition of the Abbey will head the list of entries as the very first number, with the Morgan copy cited as the only known example.

Notes