This is an electronic version of Kiernan, “Madden, Thorkelin, and MS Vitellius / Vespasian A XV”, The Library: The Transactions of the Bibliographical Society, VI, 8.2 (1986), pp. 127–32.

Madden, Thorkelin, and MS Vitellius/Vespasian A XV

Kevin Kiernan

The unpublished journals of Sir Frederic Madden in Bodleian MSS Eng. hist. C. 145 and C. 146 reveal that Madden made significant contributions to Beowulf studies before beginning his long and productive career (1828–73) at the British Museum. His most important contribution was achieved in 1824 in a meticulous collation of Thorkelin's edition with the Beowulf manuscript, a collation recently discovered by J. R. Hall in the Houghton Library at Harvard University.1 Madden's journal for 1824, in addition to documenting his collation between 21 June and 30 July, contains a brief description of Cotton Vitellius A XV that also deserves notice. In it Madden fills in some small gaps in our knowledge of the codex and shows that he was the first to recognize that only two scribes copied the entire ‘Nowell Codex’. The same journal may inadvertently help us find a more plausible date than 1787 for Thorkelin's transcript of Beowulf.2

Madden first saw Vitellius A XV (which he curiously calls Vespasian A XV) on 17 June 1824. On that day he records in his journal that he ‘sent for Vesp. A. 15, which I have not before had an opportunity of looking at’. His first impression is of ‘a thick 8vo. very dirty, & damaged, & many places almost illegible from smoke’, our only glimpse of what the book looked like after the 1731 fire and before the new binding tidied it up and gave it the proportions of a quarto.3 If he in fact sent for Vespasian A XV, he leaves no doubt that the Beowulf codex was delivered to him. ‘The whole of the contents are in Saxon’, he observes, ‘and as it contains not only the famous poem of Beowulf, but the fragm[en]t of Judith, & part of the Gospel of Nicodemus, I shall collate all this part with Thorkelin and Thwaites’.

The casual reference to ‘the famous poem of Beowulf’ seems premature at this early date, raising the possibility, at least, that the part of the journal in C. 146, like the part in C. 145, is a later revision. Madden admits that he recopied the first part of his 1824 journal in 1840 (p. viii), and that he added many remarks ‘to the rough draft . . . when I wrote the present fair copy’ (p. 54). In this part of the journal, then, we can understand an anachronistic reference to ‘the celebrated poem of “Beowulf”’ (p. 55), long before John Kemble named it that in his 1833 edition. Moreover, the fair copy preserves the time, in 1824, when the poem was still without an official name. Madden first calls it ‘a very curious Saxon poem relative to the Danes’, inserting the name ‘Beowulf’ only later (p. 42), in 1840, after mechanically copying the old descriptive title. Other emendations in the fair copy, belatedly changing ‘Thorkelin’ to ‘Thorkelin's Beowulf’ (pp. 85, 87) or to ‘Beowulf’ (pp. 145 twice, 146), tell a similar story.

Madden records that he did not have time ‘to take a particular description’ of the codex the first time he used it. However, the journal shows that he returned the next day, 18 June, spent two hours collating the Judith fragment with Thwaites's edition, and then proceeded to describe Cotton Vitellius A XV (once again referring to it as Vespasian A XV).

His account of the material preceding the Nowell Codex is of minor interest, making minor contributions. On the prefixed leaf entitled Summa expensarum, he lets us read the previously illegible second line by translating the word as ‘towns’. Wanley's translation, villarum, feodorum, was accurate after all, but its form did not help later scholars decipher what was in the manuscript. Förster's ‘Zollerhebungsstellen’ now seems to be an effort to decipher Wanley instead of the manuscript. My own guess, ‘Subdued[?]’, though based on the manuscript, now becomes a palaeographical irony.4 With Madden's help, at any rate, there can be no doubt that the manuscript reads ‘Townes’.

The only other ‘particular description’ worth noting from the first part of the codex is Madden's transcript of the notice of ownership on the second folio of the ‘Southwick Codex’. The entire notice is still intact, but fire–damage and transparent tape have obscured the readings eadem and condigne on the edge of the manuscript. Förster could see only ‘cond::::’, but conjecturally restored condigne none the less.5 The late Neil Ker, in his Catalogue, mistook the first four letters for tand instead of cond, and so conjecturally restored tandem. After taking another look at the manuscript, Ker agreed that it ‘does not read “tandem” but quite clearly “cond” followed by three or four illegible letters. “condigne” is in three other Southwick manuscripts and was quite certainly the reading here’.6 Madden's transcript shows that condigne was legible before the leaf was mounted and taped in 1845 for the new binding.7 The case exemplifies the outstanding value of Madden's collation of Thorkelin's edition with the manuscript.

Madden's description of Vitellius A XV is most interesting for what it has to say, implicitly and explicitly, about the Nowell Codex, the part of the manuscript containing Beowulf. As is well known, the first ten folios of the Nowell Codex were out of order when the foliation (91–100) was written on the manuscript leaves sometime between 1793 and 1801.8 The disorder was still undetected in 1802, leading Joseph Planta to give faulty accounts of the condition of the St. Christopher fragment and of The Wonders of the East in his Cotton Catalogue.9 Madden not only gives accurate accounts of their condition but also adroitly avoids mentioning the folio number at the start of the Nowell Codex. He does this for one of two reasons: if the folios had been put in their proper order, he would have had to say that fol. 93 followed fol. 90b; if they were still out of order, he would not have wanted to perpetuate the confusion by citing the wrong number. While we cannot be sure which of these two problems Madden actually faced, we know he shrewdly solved both by mentioning only the preceding number, 90b, and writing ‘Then follows a Saxon legend of St Christopher . . .’. He always cites correct folio numbers for the remaining texts of the codex. His apparent concern about the accuracy of the foliation supports the idea that Madden revised this part of the journal in 1840, since by then he was in charge of repairing and binding the damaged Cottonian manuscripts. The repairs to this manuscript included putting the first ten folios of the Nowell Codex in their proper order.

Madden's description also incorporates the first expert palaeographical analysis of the Nowell Codex, an analysis sorely missed in the following century.10 Because it remained unknown, Stanley Rypins could still say as late as 1924:

It has for some time been recognized that Judith and the second portion of Beowulf are written in the same hand, but until quite recently no notice has been taken of the fact that the first 1,939 lines of Beowulf and the three articles immediately preceding the epic in the MS. volume are the work of a single scribe. Prof. Sedgefield is the first to note the identity of the hand of Alexander's Letter and the first hand of Beowulf. Mr. Kenneth Sisam first points out that this identity extends also to the hand of the St. Christopher fragment and of the Wonders of the East, and that certain conclusions depend upon this fact. No other writer, of the many who have examined the MS., has called attention to this important feature.11

In fact, Madden had confidently sorted out the matter of the scribal hands, text by text, a hundred years earlier.

After noting that the St. Christopher fragment is ‘in a smaller & more elegant char[acte]r’ than the handwriting of the Southwick Codex, Madden goes on to say that the Wonders of the East is ‘in the same writing’ and that Alexander's Letter is also in the ‘same hand’. Coming to Beowulf, he notices that Wanley's erroneous account of the contents was copied by Planta in the Cotton Catalogue, and comments caustically that ‘The truth is, no one had the curiosity & patience to read it thoroughly till Thorkelin undertook the publicat[io]n of it’. He then concludes his palaeographical analysis with Beowulf and Judith. Although he recognizes that the beginning of Beowulf ‘is written in the same small hand’ as the prose texts, it appears that he briefly overlooked the change of scribal hands on fol. 172 verso. He first wrote, at any rate, that the same hand ‘continues to t[he] end of fol. 198b’, the last page of Beowulf. However, he later corrected ‘198b’ to ‘172b’, so that his palaeographical account of Beowulf now reads:

It is written in the same small hand, which continues to t[he] end of fol. 172b where a thicker and worse letter is made use of till the end — viz. to fol. 198b.

His original mistake doubtless explains how Madden came to say that the first scribe wrote all of fol. 172b, rather than only the first three lines. It is unlikely that he actually failed to notice the obvious change in scribal hands on the page. To be sure, he immediately recognized the identity of the hands on the last page of Beowulf and the first page of Judith, despite their marked difference in appearance:

On the next fol: commences the poem of Judith — tweode gifena, &c. to the end of the codex, fol. 206b., in the same thick letter as the latter part of t[he] preceding poem.

It is a pity that the terms of Madden's bequest of his journals to the Bodleian Library prevented Beowulf scholars from seeing these descriptions before 1920.

The most puzzling aspect of Madden's description of Vitellius A XV is that he twice calls it Vespasian A XV — once on 17 June 1824, when he first sent for it, and then again on 18 June, when he described it in detail. It is hard to believe that this meticulous collator and palaeographer repeatedly made such a careless mistake without cause. After all, the codex was clearly listed as Vitellius, A. XV. in 1802, in Planta's Catalogue, and Sharon Turner repeatedly called for the codex by this name in the Reading Room Register between 18 August 1800 and 6 September 1803.12 What confused Madden? Perhaps the shelf#8211;mark on the ruined binding could be misread as ‘Vespasian A. XV.’. We know from Madden's account that the codex was ‘very dirty, & damaged’.

In any case, if there was once some good, now vanished reason for a careful scholar like Madden to write down (twice) Vespasian A XV for Vitellius A XV, his otherwise inexplicable mistakes may explain why there is no record in the Museum's Reading Room Register for Thorkelin's transcript of Beowulf.13 According to Thorkelin, both transcripts of the poem were done in 1787, but there are compelling reasons to doubt that his own transcript, Thorkelin B, was done this early. One reason is that there are no Vitellius entries at all for 1787 in the Register, which shows instead that he was busy with other manuscripts in the first half of the year, and not even working in the Museum for the second half. Moreover, none of the Vitellius entries ever shows when, after his return to the Museum in 1788, Thorkelin might have copied the B transcript.14 On the contrary, if we go by the Register, Thorkelin used Vitellius A XV for part of the day on 28 May 1789, but after returning it the same day, he never again requested to use it. In view of Madden's two references, it now seems prudent to consider entries under Vespasian A XV, too, for possible signs of when Thorkelin might have copied the Beowulf manuscript.

The Register encouragingly enough shows Thorkelin requesting Vespasian manuscripts the next time his name appears in it (in June). More encouraging still, it shows that he sent for Vespasian A XV the next time, 17 June, and kept it out for a month and a half, until 3 August. This is remarkable because Vespasian A XV, Epistolæ S. Clementi . . . et aliorum episcoporum Romæ, was not a manuscript Thorkelin wanted to study or transcribe. He entered it by mistake in a notebook listing Cotton manuscripts he wanted to study, but deleted it when he learned its contents.15 Thorkelin first squeezed in and then crossed out the description ‘XV. Registr[um] Chart[arum] de Lewes’ between his listings for Vespasian A VIII and A XVIII. It is perhaps worth noting that he began to write ‘Vitellius’ for the ‘Vespasianus’ heading, but crossed out this mistake after writing only ‘Vite’ and the beginning of ‘l’. Even if we allow him a full month and a half to discover that Vespasian A XV was not the register of charters he thought it was, how do we explain that he again requested Vespasian A XV on 22 September and kept it out this time for over three months, until Christmas Eve 1789? In view of the Madden cases of mistaken identity, the explanation would seem to be that Thorkelin was actually using Vitellius A XV.

We have no way of knowing, now, how the Beowulf codex might have come to be mixed up with Vespasian manuscripts. Perhaps Thorkelin immediately exchanged Vespasian A XV for Vitellius A XV on 17 June 1789, the first time he saw the former codex, and the clerk in charge of the Register failed to record the transaction or simply misread a damaged shelf–mark. We do know that the course of Thorkelin's research from May to Christmas 1789 had Vitellius and Vespasian manuscripts coming and going at a good pace, increasing the likelihood of a clerical error. He returned five Vitellius manuscripts, for example, on 22 September, the same day he requested Vespasian A XV for the second time. In any event, we can easily understand why Thorkelin would be using Vitellius (if not Vespasian) A XV for long stretches of time in 1789. By linking Vespasian A XV to the Beowulf codex, Madden gives us a plausible new date for Thorkelin's transcript of the poem.

Lexington, Kentucky


1 As Hall will show in a forthcoming study, this valuable collation allows us to restore with confidence many letters now gone from the manuscript. Madden's journals are on microfilm at Harvard University (Harvard Film R12, Red 2).

2 The description of the codex is in Bodleian MS Eng. hist. C. 146, pp. 30–32; cited by permission of the Keeper of Western Manuscripts, Bodleian Library, Oxford. Concerning the inaccurate date of Thorkelin B, see my article ‘Thorkelin's Trip to Great Britain and Ireland, 1786–91’, The Library: The Transactions of the Bibliographical Society, VI, 5 (1983), 1–21.

3 Madden later remarks (p. 30) that Thorkelin in his edition wrongly called the manuscript a quarto and mistakenly referred to it as ‘Vitell. A. IX’. Today we tend to agree with Thorkelin about the original dimensions of the codex, yet Madden helps us realize here how small the book must have appeared in its burnt binding. We can safely dismiss the reference to ‘A. IX.’ as a late, wild guess by Thorkelin, since we know that he failed to record the number ‘XV‘ on the title–pages of his transcripts. See The Thorkelin Transcripts of Beowulf in Facsimile, edited by Kemp Malone, EEMF (Copenhagen, 1951), pp. B and C.

4 The Summa expensarum is the second prefixed leaf, verso, of the codex. For a reproduction of the page and a discussion see Kiernan, Beowulf and the Beowulf Manuscript (New Brunswick, 1981; reprinted 1984), pp. 78–80; hereafter cited as BBMS.

5 See Die Beowulf–Handschrift, Berichte über die Verkandlungen der Sächsischen Akademie der Wissenschaften zu Leipzig, Philologisch–historische Klasse, 71 (1919), 74. Today ‘the passage can be read in its entirety by holding the leaf up against strong light’ (BBMS, p. 110; there is a reproduction of the notice on the facing page).

6 Cited from a two–page report for the Toronto symposium on the dating of Beowulf (20–23 April 1980), by permission of Neil Ker's literary executor (Professor A. G. Watson, University College, London). For the earlier reading, see Ker, A Catalogue of Manuscripts containing Anglo–Saxon (Oxford, 1957), p. 279.

7 Madden dates the binding in his ledger, Cottonian MSS., Repairing and Binding Account, now in the departmental archives of the Department of Manuscripts, British Library (see BBMS, p. 69, note 7).

8 For a history of the numerous foliations used for Cotton Vitellius A XV, see BBMS (pp. 71–110 to). Madden himself apparently corrected the old foliation when the manuscript was rebound in 1845.

9 A Catalogue of the Manuscripts in the Cottonian Library, Deposited in the British Museum (London, 1802), pp. 380–81; see BBMS (pp. 99–103).

10 Madden bequeathed his journals and papers to the Bodleian Library on the condition that they were to remain unopened until 1920 (DNB).

11 ‘Introduction’, Three Old English Prose Texts in MS. Cotton Vitellius A XV. EETS 161 (1924), pp. In Die Beowulf–Handschrift, Förster credits Eduard Sievers with first noticing (in 1872) that Judith was written by the same scribe who copied the second part of Beowulf (p. 31).

12 The Register recording Turner's use of the MS is in BL Add. MS 46516 (see fols 87v–88, 157–58, 180v–81, and 180v–81).

13 The Register for the years of Thorkelin's work at the British Museum is in BL Add. MSS 46513 and 46514.

14 The relative dating of Thorkelin A and B is fully discussed in my book, The Thorkelin Transcripts of Beowulf, Anglistica xxv (Copenhagen, 1986), pp. 14–54. There is evidence outside the Register that Thorkelin A was made in the second half of 1787, while Thorkelin was in Scotland.

15 See his Biblioth. Cotton. (fol. 51. Thorkelin's personal papers, formerly tied in a loose bundle numbered Rigsarkivet 6431.5, were recently reorganized and boxed as 6431.5 and 6431.6; most of the material relating to his trip to England, including this Cotton notebook, is now in 6431.6.E.3, a box labelled ‘Vedr. G. J. Thorkelins rejse til Storbritannien og Ireland 1785–91’ (i.e., ‘concerning Thorkelin's trip to Great Britain and Ireland’).