Richard W. Clement

    A recent book by Professor Kevin Kiernan, Beowulf and the Beowulf Manuscript (New Brunswick, 1981), has caused a great deal of discussion in the field of Beowulf studies. Professor Kiernan has moved the date of composition for the poem to the reign of King Canute in the early eleventh century. He believes that our unique manuscript, the Nowell Codex, or really just the Beowulf portion, is the poet's own working copy. His startling conclusions are based on a complex codicological and paleographical analysis. We must remember that as a result of the fire at Ashburnham House in 1731, the codex no longer has any trace of its binding or even of conjugation the Nowell Codex is a stack of disjunct leaves. Hence the collation of Cotton, Vitellius, A.XV, ff. 94-209 is not the simple and straightforward process normally associated with such procedures, but rather it demands the consideration of a whole range of codicological factors. The collation is exceptionally important to Professor Kiernan's thesis, as his conclusions are ultimately based on the requirement of a separate or discrete Beowulf: that is, that Beowulf begins a new gathering or quire, and that the Judith fragment which follows Beowulf was not originally part of the codex.

  The collation is not self-evident by any means. Förster, Dobbie, Ker, Malone, Boyle, and Kiernan have all come to differing conclusions.1 (We may, however, take Malone's collation, from his facsimile, to represent the traditionally accepted collation.)

    Kiernan's collation is the only one which has considered the hair-flesh sequences in the entire codex. In essence, his collation best explains many of the remaining physical aspects of the codex, yet Kiernan's codicological conclusions based upon his collation are highly conjectural. The diagrams at the end of this essay set out the hair-flesh sequences of the Nowell Codex following the official foliation of the British Library. The hair-flesh sequences clearly establish the larger structure of all the gatherings except quires 5 and 6. Quire 1 is a gathering of ten folios, all with the hair sides out. This first quire of the codex, the first quire of Scribe A's portion, has a make up which is identical to quire 12, the first quire of Scribe B's Beowulf portion. Quire 2 is made up of six folios: the hair-flesh sequence illustrates that no other configuration is possible. In this quire, however, the normal alternation of hair opposite hair, the flesh opposite flesh has not been maintained in the outer sheet. Here, where we might expect the flesh side to face outward if the alternation were maintained, the hair side faces outward, but of course this is not at all unusual, as it was customary in an outer sheet of a gathering that the hair side face outward to protect the quire. Indeed, this is true for all fourteen quires of the Nowell Codex. Quire 2, however, is actually an original gathering of four sheets: the outer sheet has been removed and the next sheet (now the outer sheet) reversed so that the hair side faces out. Conversely only the second sheet {nay have been removed; the result would be

    It has long been recognized that ff. 110-117 and 118-125 must make up two discrete gatherings. The manuscript foliation, and of course the text, reflects the transposition of the two gatherings by an early modern binder, and the hair-flesh sequences illustrate that ff. 110-117 and 118-125 do indeed constitute quire 3 (which is actually 4) and 4 (which is actually 3). As Kiernan has pointed out, both of these gatherings consist of three sheets augmented by a pair of singletons (ff. 3 and 6). If we remove the pair of singletons the two gatherings make perfectly regular gatherings of three sheets, or six folios, and as further corroboration, the inner sheets of both quires are positioned flesh side inwards, as is standard in three-sheet quires. Unfortunately, the hair-flesh sequences fail us in quires 5 and 6. There are two, or even three possible configurations, but I will return to these in more detail below.

    Quire 7 is seemingly a normal gathering of eight folios, except that folio 6 (f. 147) cannot possibly be conjugate with folio 3 (f. 144); folio 3 is hair side out, while folio 6 is flesh side out. Kiernan suggested that folio 6 must be a replacement leaf and this does indeed seem possible.

    Quire 8 is also seemingly a normal gathering of eight folios. Although each of the four pairs of folios appears to be conjugate, the hair-flesh sequence is irregular. The quire was originally made up of three sheets (the inner sheet has the flesh side inwards),

    Quire 9 is a perfectly normal gathering of four sheets, or eight folios, which maintains a regular hair-flesh sequence. Quire 10, on the other hand, exhibits an irregularity in its hair-flesh sequence: one sheet (or possibly two singletons) made up either of folios 2 and 7 or of folios 3 and 6 has been inserted into what was originally a gathering of three sheets (again, the inner sheet has the flesh side inwards). If we remove this sheet, the quire is restored to its original regularity. Quire 11 is identical to quire 10.

    As mentioned before, quires 12 and 13 are identical in make up to quire 1. They are quires of five sheets or ten folios, all with the hair sides out. Quire 12 was the first full quire of Scribe B's Beowulf portion. Quire 14, which comprises the fragmentary Judith, is a gathering of four sheets or eight folios, all with the hair sides out. As the text of folio 202r begins in mid-passage, several quires, probably comprising twenty-four folios, are missing between quires 13 and 14. The text ended imperfectly on folio 209v, but someone in an early modern imitation Anglo-Saxon minuscule hand has supplied the end of the poem from a now-lost leaf. Dobbie believed that there was only a single leaf attached to the end of quire 14,3 but, although possible, this seems most unlikely: single leaves were rarely attached at the end of a gathering. Kiernan suggested that another work followed which was detached for

    The arrangement of the hair and flesh sides in ff. 126-141, quires 5 and 6, is ambiguous. There are three possible collations. (See the diagrams at the end of the essay.) The traditional collation divides these sixteen folios into two quires of eight; thus Beowulf begins on the seventh folio recto of quire 5. In this scheme, quire 5 is a perfectly normal four-sheet gathering. Quire 6 is an original three-sheet gathering which has been expanded by the insertion of two singletons (ff. 3 and 6).

    Kiernan, on the other hand, has suggested an alternative collation. His collation divides these sixteen folios into two quires, the first of six, the second of ten; thus Beowulf begins on the first folio of quire 5. In this scheme, quire 5 is a normal gathering of three sheets. Quire 6 is an original gathering of four sheets which has been expanded to ten folios by the addition of folios 4 and 7. Kiernan's quire 6, although possible, is highly unlikely. As we have seen, the codicological practice of the scriptorium which produced the Nowell Codex is varied, but there is no precedent for an expanded quire of four sheets. The physical structures of the quires are of three types: first, an original gathering of three sheets, which in every instance has been expanded to a gathering of eight folios by the

    On the other hand, quires 5 and 6 of the traditional collation fit this established custom very well. Quire 5 is a normal gathering of four sheets. Quire 6 is an original three-sheet gathering which has been augmented by the addition of two singletons. This alteration of three-sheet gatherings into eight-folio gatherings is certainly one of the most striking idiosyncratic codicological aspects of the codex.

    Kiernan has not been unaware of the problems with his anomalous ten-folio quire 6. At a recent seminar devoted to his work, he suggested another possible collation.5 He has coped with the problem of his anomalous quire 6 by converting it into an original three-sheet gathering expanded to an eight-folio gathering by the addition of two singletons, but he has created a far greater anomaly in the bifolium which he has placed between quires 5 and 6. This is an even greater departure from the scriptorial practice reflected in the other quires of the Nowell Codex. Needless to say, it is highly unlikely that this collation represents the actual construction.

    Kiernan's two collations, as unlikely as they are, cannot be totally excluded by incontrovertible physical evidence. Indeed he has brought to bear several points which he claims bolster his collation. These are inspired by his desire to see Beowulf from the three prose works (Christopher, Marvels of the East, and Alexander's Letter) on the one hand, and Judith on the other. I shall consider each one of these in some detail.

    Humphrey Wanley in his great Catalogus described Beowulf in the terms "+In hoc libro+ ...." Kiernan states that there "does not seem to have been any doubt in Wanley's mind that Beowulf began a new codex."6 However, surely this is not at all what Wanley meant. For instance, Wanley describes Junius 11, the Caedmon MS, as two libri and the Exeter Book as ten libri. As Wanley did not mean to suggest that either Junius 11 or the Exeter Book was made up of several discrete codices, neither did he mean to suggest that Beowulf formed a discrete codex.

    Kiernan has noted that "Vi[tellius] A 15" appears at the bottom of folio 132r, the first page of Beowulf. He suggests that this signature "was written on the first page of Beowulf as instructions to Cotton's binder to include the poem in the composite codex."7 One must, however, query the date of this inscription. If this shelf mark was written as a guide to Cotton's binder, why are there no other similar inscriptions on the first folios of other discrete codices? We have several examples of Cotton's notes to his binder, but these are of some length and never (to the best of my knowledge) mention a shelf mark.8 The letters are certainly not in Cotton's hand. We have no idea when the codices which presently make up Vitellius A.XV were brought together, but if, for instance, they were bound by Burghley's binder

    Kiernan claims that there are two telltale signs on folio 132r that indicate it had once been the outer folio of a discrete codex. The first is a small soiled spot in the lower right hand corner of the folio which Kiernan identifies with the damage a thumb would inflict as it gripped the outer cover of a book.9 Even if this spot is the result of the friction produced by a thumb (and I am unable to see how a thumb could rest easily in that position), the folio simply does not have the appearance of an outer cover. Kiernan also believes that the heavy rulings on folio 132r indicate that it was an outer cover.10 As the first leaf and part of the outer sheet of a quire, Kiernan reasons it would have the heaviest rulings, as indeed folio 132r has. If we accept Kiernan's first collation, then the conjugate folio (f. 141v) ought to exhibit equally heavy rulings, but it does not. Likewise, if we accept Kiernan's second collation, then the conjugate folio (f. 133v) ought to exhibit the same heavy rulings, but it does not. If we accept the traditional collation, then the conjugate folio (f. 127v) ought also to exhibit the heavy rulings, but neither does it. Obviously then, folio 132r was individually re-ruled, as well we might expect on the first folio of a new item in a composite codex.

    Kiernan believes that the distinctive styles of capital letters in the three prose works on the one hand and Beowulf on the other reflect two discrete codices.11 While this is possible, it hardly constitutes compelling evidence; the differing styles of capitals may just as well reflect differences in the exemplar(s). Likewise Kiernan places great emphasis on the greater care with which the two scribes treated Beowulf.12 Certainly there are more corrections, and more evidence of intelligent proofreading, in Beowulf than in the three prose works. Again, Kiernan believes this difference in attitudes towards the texts not only separates the prose works from Beowulf as two discrete codices, but also separates them as two unconnected scriptorial projects. This differing scribal attitude might possibly reflect two separate projects of copying, but it may also simply reflect the differing attitudes of the scribes towards prose and verse: the technical requirements of verse require a far greater degree of accuracy in copying. As Kiernan has illustrated, the scribes were not unintelligent copyists; it is perhaps only natural that they should have taken certain pains over a poem such as Beowulf, and yet not bothered particularly with the prose works.

    Kiernan's most compelling argument for separating the prose works from Beowulf is, as he views it, the distinctively differing codicological formats of the two sections.13 With the exception of the so called "anomalous" first quire, all the quires before Beowulf are essentially gatherings of three sheets and therefore, argues Kiernan, quire 5 must also be a gathering of three sheets. As we have seen, however, quire 2 was originally a four-sheet gathering. Kiernan contrasts the supposed three-sheet format of quires 2-5 with quires 7-11, in which he claims to discern a four-sheet format. As I have illustrated, however, only quires 7 and 9 have an

    The Judith fragment, quire 14, is more problematic. Because there is a gap of unknown length between the end of Beowulf and quire 14, it is more difficult to make any conclusion based on continuity in the codicological format of Scribe B's portion. We have three quires which were made by or for Scribe B: quires 12 and 13 are five-sheet gatherings, all the hair sides out; quire 14 is quite likely to have originally been a similar gathering. Scribe B's consistent format is of five-sheet gatherings constructed with the hair sides facing out. It thus seems likely that Scribe B (or the person who made his quires) also made quire 1, as it also exhibits this highly distinctive and idiosyncratic format.

    Kiernan has suggested that there is a format difference in the size of the writing frame of the Judith fragment.

What makes this especially valuable as evidence is that the number of lines per page in Beowulf varies in different quires from twenty lines, to twenty-one lines, to twenty-two lines, and yet the written space between the first and last rulings, regardless of the number of lines per page, is uniformly between 17 and 18 cm., usually about 17.5 cm. This shows that even when the scribes departed from the normal number of lines (twenty), they took care, for aesthetic reasons, to make the written space of all pages throughout the Beowulf codex a uniform size. In the case of the Judith

In fact, the written space in the Judith fragment (measured between the first and last rulings) is between 16.6 and 17.1 cm; the average is 16.9 cm. Kiernan claims that the average for this same measurement in the quires of Beowulf is 17.5 cm. The actual size, however, is 17.1 cm., which is only 2 mm. difference from the measured average in the Judith fragment. It is obvious that Judith was ruled to fit the same format as the three prose works and Beowulf. Far from separating Judith from Beowulf, this evidence in fact supports the traditional belief that Judith was part of the same copying project which produced Beowulf.

    One possible impediment to this scheme is the presence of a worm-hole in quire 13, which has no equivalent in quire 14. In other words, when this worm ate his way through the final gathering of Beowulf, quire 14 was not next to it. But since the first part of Judith has been lost, quire 14 was not originally adjacent to quire 13. It is in the now-lost intervening folio(s) that we would have found traces of the worm. Father Leonard Boyle, on the other hand, suggested that the worm had feasted on quire 13 before any other gatherings had been joined to it, and he may well be right.15 We cannot know how long the finished unbound quires may have been stored. Certainly quires 12 and 13 were finished before Scribe A's gatherings and a hungry worm would have had ample opportunity to bore into quire 13. Thus, although we cannot be sure when the worm-hole was made, its presence fails to exclude Judith as the fifth member of this composite codex.

    Quire 14 (and however many lost folios) was probably copied before quires 12 and 13 as Leonard Boyle has suggested. Kiernan has questioned Förster's conclusion that Judith was once composed of twenty-four additional folios before quire 14.16 It may be that Judith is nearly complete and therefore that the outer sheet is all that is missing from quire 14, or any number of complete quires (four sheets, three sheets, or more likely Scribe B's normal format of five sheets) may have disappeared. For the purposes of this essay it matters not at all whether the missing folios contained only Judith or perhaps some other works as well. Whatever has been lost comprised Scribe B's first stint of copying. As both Boyle and Kiernan point out, the final folio of Beowulf (f. 210v) is crowded with text; the scribe obviously had no more folios, as he would have had with Judith still to copy. He must, therefore, have copied quires 12 and 13 after Judith. Again, both Boyle and Kiernan have demonstrated that the final portion copied by Scribe B was in quire 11, where he took over from Scribe A. Once again, it is obvious that Scribe B was faced with only so much parchment and a certain number of lines to be written. His solution here was not to crowd them as on the final folio of Beowulf, but to add one line of text to four folios (177v-179r). Scribe B must, therefore, have copied Judith first (probably in quires of five sheets with all the hair sides facing out), then quires 12 and 13 of Beowulf, and finally the remainder of quire 11. Scribe A, on the other hand, seems to have copied his material straight through from quire 1 to 11 before breaking off. If Förster was correct in estimating that one gathering of ten folios has been lost from the beginning of Christopher, this gathering may well have been one of Scribe B's distinctive five-sheet quires. Indeed, the first extant quire is just such a gathering. Surely it is no accident that both scribes began their stints with these distinctive gatherings of

    In summary, the collation of the Nowell Codex is as Kemp Malone described it in his facsimile: 110, 26, 3-118, 12-1310, 148. Christopher, Marvels of the East, Alexander's Letter, Beowulf, and Judith make up a single, unified, composite manuscript the Nowell Codex.

Illinois State University