This is an electronic version of an essay first published in The Iconic Page in Manuscript, Print, and Digital Culture, edited by George Bornstein and Theresa Tinkle. Ann Arbor: University of Michigan Press, 1998. 7-32.

Alfred the Great's Burnt Boethius

Kevin S. Kiernan

While medieval manuscripts are filled with beautiful and eloquent iconic pages, some of the most interesting problems of transmission are hidden away in the ugly, silent types. It would be hard to find uglier, less communicative iconic pages than many of those precariously preserved in British Library Cotton MS Otho A. vi, the only surviving copy of the Old English prose and verse translation of Boethius's Consolation of Philosophy by Alfred the Great (871-99).1 This tenth-century manuscript was severely damaged in the same infamous fire in 1731 that burned away the edges of the Beowulf manuscript. Nineteenth-century efforts at restoration, moreover, have compounded the fire damage by cutting up, covering up, or actually washing away parts of the text that managed to escape the fire. Technological advances in the twentieth century -- notably ultraviolet fluorescence, fiber-optic backlighting, electronic photography, and digital image processing -- make it possible to recover many of these losses. But editorial practice from the seventeenth century to the present has conspired with the vicious eighteenth-century fire and the virtuous nineteenth-century curators and conservators to render the iconic pages of this manuscript all but invisible in the editions we now use. The two standard editions, both monuments of modern Anglo-Saxon textual scholarship, are based on this manuscript, and yet one edition is entirely in prose, while the other is entirely in verse.2 These editions thus present ironic iconic pages that radically misrepresent the prosimetrical manuscript that preserved the text.

One can trace the reason for these curious editorial developments to two factors: (1) the inaccessibility of the tenth-century manuscript, which everyone thought was destroyed in the 1731 fire, until its burnt remains were recovered at the British Museum in the 1830s; and (2) an overpowering edition-in-progress of the twelfth-century manuscript by the great seventeenth-century scholar Francis Junius, with extensive collations from the missing tenth-century manuscript. Although there are no detailed descriptions of the appearance of Cotton Otho A. vi dating from soon after the fire, we may infer from the current state of the manuscript, with beginning and ending gone and significant sections of the middle lost, too, that Cotton's binding with its identifying spine was totally destroyed, leaving behind a nameless pile of brittle, charred leaves. According to Andrew Prescott, curator of manuscripts in the British Library, the books in this condition would have been among the crusts and fragments preserved in drawers with the single-sheet charters after the fire, until several years after the founding of the British Museum in 1753.3 Unidentified remains of this nature were removed to the "charter garret," where Cotton Otho A. vi disappeared and was forgotten for a century. Long after the fire the charred remnants were so inaccessible that the official British Museum catalog, compiled by then keeper of manuscripts Joseph Planta in 1802, still described it as lost.4 Until the mid-nineteenth century our only knowledge of Cotton Otho A. vi was indirect, through Junius and the seventeenth-century edition that eventually emerged from his transcript and collations.

Perhaps if he had lived in London instead of Oxford, Junius might have copied all of poor Cotton Otho A. vi before the disastrous fire, instead of copying the exclusively prose version in the twelfth-century manuscript, Oxford Bodleian Library MS Bodley 180 (Ker 305). We can at least be happy that, after he finished copying the prose manuscript, Junius collated his transcript with the tenth-century prosimetrical manuscript in London and tried to record the most notable variant readings in the margins of his transcript. Fortunately, the verse passages, known collectively as the Meters of Boethius, were all but impossible to collate with their prose counterparts. As they did not fit in the physical or figurative margins, Junius copied them in full on separate pieces of paper varying in size according to the varying length of the Meters. To hold all of this prodigious work, he assembled a great scrapbook, now MS Junius 12 in Oxford's Bodleian Library.5

In this notebook Junius created a sort of hypertext,6 linking all of the textual information needed for a bonafide scholarly edition of the twelfth-century prose version of Alfred's Boethius. In lieu of a hypertexted digital facsimile, the following illustration of two busy facing pages serve in miniature as a kind of icon of Junius's oversized "beta-version." Although this presentation renders the texts virtually illegible, one can see at a glance the essential nature of the document and perhaps appreciate the difficulties of reproducing it in a printed book of this size.7 Because he founded his edition on the later version in MS Bodley 180, Junius quite closely followed the design of the twelfth-century manuscript, as the transcription in the right facing page shows (fig. 1). Soon realizing that the Old English prose renditions of the Latin Boethian Meters were problematic, however, he constructed an interleaved guardbook for the transcript and pasted on the new facing pages the corresponding Latin Meters (but not the prose) to provide a means of comparison and contrast. As he compared the prose sections of Cotton Otho A. vi with his transcript, he used alphabetical sigla in the transcript to mark significant variants and then copied the variant with its key letter in the left margin.

With the verse sections Junius was forced to adopt a different procedure. Although he still identified the verse sections by sigla in the necessary places, Junius could not simply record significant variants in the margins of his transcript, because the prose translations were fundamentally different from the verse translations of the same Boethian Meters. His solution to copy the verse passages in their entirety and to paste them on facing pages was providential, because he thereby saved the entire set of verse passages for posterity when the originals in Cotton Otho A. vi were destroyed or damaged in the fire of 1731. It is likely that there would have been no Anglo-Saxon Poetic Records if Junius had not preserved the so-called Meters of Boethius. Because of subsequent fire and water damage, the editors would not have been able to read most of the meters without ultraviolet fluorescent effects. Here Junius copied Meter 3, only eleven lines long, on a long strip of paper he had to fold over to fit on the page next to the Latin excerpt. Beneath it, he hung the transcript of the long, fifty-seven-line Meter 4 from the bottom of the page. One must fold up this transcript to close the book. This somewhat reckless arrangement is particularly poignant when we realize that both of these Meters were totally destroyed in the fire.

Junius duly noticed in his prefatory material that the Cotton manuscript, unlike the twelfth-century prose manuscript, had no table of contents. It did include an additional verse preface, however, which he saved for us, even adding metrical pointing to distinguish it from the prose in the manuscript and to facilitate a modern arrangement into lines of verse, such as the following:8

Ðus Ælfred us ealdspell reahte,
cyning Westsexna, cræft meldode,
leoðwyrhta list. Him wæs lust micel
ðæt he ðiossum leodum leoð spellode,
monnum myrgen, mislice cwidas,
þy læs ælinge ut adrife
selflicne secg, þonne he swelces lyt
gymð for his gilpe. Ic sceal giet sprecan,
fon on fitte, folccuðne ræd
hæleðum secgean. Hliste se þe wille!
(ASPR 5, 15 3)
[Thus Alfred told us an old story, / King of West Saxons showed off his craft, / skill of verse-making. His wish was wide / that to these people he should pass on poetry, / delight of folk, fusing his phrases, / so that dreariness might not drive out / the self-absorbed sort, when he gives notice only / to his own yelping. I yet must speak, /to fashion in fitts folk-shared philosophy, / tell it to the high-minded. Heed it, who will! (My trans.)]

In the left margin of his transcript of the prose preface Junius also recorded collations from the prose preface of Cotton Otho A. vi, confirming that the earliest manuscript contained both prose and verse prefaces.9

Following this prosimetrical preface Junius found and transcribed a long verse passage in Cotton Otho A. vi, corresponding to a much shorter prose passage in Bodley 180, on the invasion of Rome by the Goths. Once again frustrating an ordinary collation, this introductory matter has no basis in the Latin source. The prose version, moreover, was very different from the verse. These passages are too long to quote in full or discuss in detail here, but it is useful to show the most plausible direction of influence by comparing the opening passages. The first sentence of the prose simply tells how the Goths under Rædgota and Eallerica conquered Rome and were then succeeded by Theodoric, the tyrant who persecuted Boethius:

ON ðære tide ðe Gotan of Sciððiu mægðe wið Romana rice gewin up ahofon, 7 mid heora cyningum, Rædgota 7 Eallerica wæron hatne, Romane burig abræcon, 7 eall Italia rice þæt betwux þam muntum 7 Sicilia þam ealonde in anwald gerehton, 7 þa æfter þam foresprecan cyningum Þeodric feng to þam ilcan rice. (Sedgefield, 7)

[In that time the Goths from the nation of Scythia raised up war against the kingdom of the Romans, and with her kings -- they were called Rædgota and Ealleric -- stormed the stronghold of the Romans and held in control all the kingdom of Italy that is between the mountains and the island of Sicily; and then after these aforementioned kings, Theodric succeeded to the same kingdom. (My trans.)]

The destroyed verse passage corresponding to this opening sentence was thirty and a half lines long and contained many details with no parallel in the prose. Without Junius's transcript we would have never known that Alfred's burnt Boethius began as a poem:

Hit wæs geara iu     ðætte Gotan eastan
of Sciððia     sceldas læddon,
þreate geþrungon     þeodlond monig,
setton suðweardes     sigeþeoda twa;
Gotena rice     gearmælum weox. 5
Hæfdon him gecynde     cyningas twegen,
Rædgod and Aleric;     rice geþungon.
Þa wæs ofer Muntgiop     monig atyhted
Gota gylpes full,     guðe gelysted,
folcgewinnes.     Fana hwearfode 10
scir on sceafte.     Sceotend þohton
Italia     ealle gegongan,
lindwigende.     Hi gelæstan swua
efne from Muntgiop     oð þone mæran wearoð
þær Sicilia     sæstreamum in, 15
eglond micel,     eðel mærsað.
Ða wæs Romana     rice gewunnen,
abrocen burga cyst,     beadurincum wæs
Rom gerymed.     Rædgot and Aleric
foron on ðæt fæsten;     fleah casere 20
mid þam æþelingum     ut on Grecas.
Ne meahte þa seo wealaf     wige forstandan
Gotan mid guðe;     giomonna gestrion
sealdon unwillum     eþelweardas,
halige aðas.     Wæs gehwæðeres waa. 25
Þeah wæs magorinca     mod mid Grecum,
gif hi leodfruman     læstan dorsten.
Stod þrage on ðam.     Þeod wæs gewunnen
wintra mænigo,     oðþæt wyrd gescraf
þæt þe Ðeodrice     þegnas and eorlas 30
heran sceoldan.
(ASPR 5, 153-54)
[It was in bygone years that Goths from the East / out of Scythia carried shields, / oppressed with force many a nation, / two victorious peoples setting out southwards. / The kingdom of the Goths grew year by year. / They had by law two kings, / Rædgod and Aleric. The kingdom flourished. / Then was over Mount-Alp many / a Goth made full of boast, eager for war, / for folk-struggle. The banner waved / sheer on its shaft. The shooters intended / to overrun all of Italy, / these shield-warriors. And so they carried it out, / even from Mount-Alp to the great shore / where Sicily in sea-streams, / the great island, marks out a native land. / Then was the kingdom of the Romans won, / the choicest of cities broken open by battle-men, /Rome revealed. Rædgod and Aleric / advanced into that fortress. The emperor fled / with his nobles out among the Greeks. / The remnant might not withstand further warfare, / with battle of the Goths. The treasure of former men / the landprotectors unwillingly gave, / as well as sacred oaths. That was woe for two reasons. /Though the heart of the (Roman) men was with the Greeks, / if they dared to help the people-leaders. / It stayed that way for a while. The people were conquered / for many a winter, until fate decreed / that thanes and eorls had to obey Theodoric. (My trans.)]

We know that these prose and verse passages are closely related, because both incorrectly claim that the two Gothic kings ruled simultaneously. But the prevailing view that the Meters are simply callow, unpoetic expansions of the prose10 does not hold up to scrutiny in this example with no source in the Latin Meters. A mere versifier of the prose passage would not invent the name Muntgiop11 for the Alps, nor guess that the Roman emperor fled to Constantinople seeking help from the Greeks. If anything, the prose is a brief paraphrase of the verse.

The amount of work Junius put into his scrapbook shows that he planned an edition, but he may not have proceeded, because his collation convinced him that the Cotton manuscript had, as he says, the best readings.12 Nonetheless, after Junius's death a twenty-one-year-old Oxford Saxonist named Christopher Rawlinson was able in 1698 to produce an excellent first edition of MS Bodley 180, the twelfth-century version, by basically printing Junius's scrapbook.13 The format Junius had adopted made it easy, if not virtually inevitable, for Rawlinson to gather the Meters together and print them as a separate group, detached from their original context in the tenth-century prosimetrical manuscript. After all, what else could or should he have done with them in his edition of a different manuscript? He could not simply put the verse versions in footnotes with Junius's collations from the prose sections of Cotton Otho A. vi, because some of the Meters would take up more than a page. Rawlinson's sensible editorial decision was to put them in an appendix of "Poetical Versions Taken from the Cotton Codex.14 His edition of Bodley 180 also includes Junius's light collations from Cotton Otho A. vi at the foot of each page.

While Rawlinson's edition of Bodley 180 preserved the Cotton Meters through the eighteenth century, the rest of Cotton Otho A. vi endured the fire and then languished in bits for the same century in the British Museum. Cotton Otho A. vi was one of the many Anglo-Saxon manuscripts first retrieved from the garret sometime after 1825 by then keeper Josiah Forshall. According to Prescott, the bear or tiger skin rug effect in some of the leaves was caused at this time when Forshall had the edges slit to help them lie flat, a destructive procedure later found unnecessary; much of the text was presumably washed away when the leaves were soaked, following a suggestion by Sir Humphrey Davy, in a solution of water and spirits of zinc (Prescott 405). For a brief period in the late 1830s and early 1840s the manuscript was apparently made available as loose unprotected leaves in a solander case, a box in the shape of a book, until it was inlaid by Henry Gough in 1843-44.15 During this time there was, of course, great risk for the brittle edges to crumble away from handling. In any event, soon after Sir Frederic Madden became keeper of manuscripts in 1838, the British Museum reluctantly agreed to an expensive restoration of some important manuscripts, including Alfred's Boethius.16 Madden knew the dangers of progressive deterioration of the brittle edges from his own experience collating the Beowulf manuscript in 1824-26. Soon after his appointment as keeper, Madden hired Gough, the same skilled binder who did the Beowulf manuscript in 1845, to inlay Cotton Otho A. vi. The manuscript he was expected to repair had large holes burned into many folios; the edges of all pages were burned away, often taking large and small parts of the text with them; and even some of the text on relatively undamaged folios was rendered partly or entirely illegible from water damage, either when the fire was extinguished or (more likely) when the pages were washed down to remove soot or, later, in preparation for its current inlaid binding, soaked in a solution of water and spirits of zinc to counteract the extreme shrinkage and wrinkling. In Gough's restored binding each damaged vellum leaf was separately inlaid in heavy paper frames, which were then rebound in a facsimile of the Cotton binding, with Cotton's gilded arms stamped on the cover.

The iconic page that emerges from the ashes is more than a page from a tenth-century Anglo-Saxon manuscript. In some important respects Cotton Otho A. vi is and always has been more a modern facsimile edition than a medieval manuscript. The sixteenth-century Cotton binding had caused the deliberate destruction of the medieval binding and had artificially combined formerly unrelated texts with Alfred's Boethius. Once the Cotton binding and two of the three formerly unrelated texts were destroyed by the fire in 1731, there was no compelling reason, at any rate, to restore it as Cotton Otho A. vi, as if it still were the same book.17 The method used to restore and conserve the manuscript nonetheless placed primary emphasis on the fact that it was a book, another icon, rather than two remaining texts that were first artificially combined in the sixteenth century. Thus, it was seemingly as important to restore the lost binding with the Cotton arms as to protect the exposed leaves from further deterioration. It is strange, then, that anyone who studies the manuscript in photographic facsimile, as most readers do, never sees the paper frames of the restoration binding. In facsimile the frames are cropped, making the leaves appear to float in space.

The frames, however, have become integral parts of the manuscript leaves. Folio 38 verso may be used to illustrate how the visible and the invisible tend to trade places as the text moves from manuscript to print. This folio, like all the folios now, is no longer physically defined by the Anglo-Saxons, who originally prepared the vellum and trimmed it into sheets, ruled the sheets, and folded the combined sheets into gatherings. The unique shape of this murky leaf today is actually defined by the eighteenth-century fire and the nineteenth-century paper mount framing it to prevent further crumbling of its seared and brittle edges. The framed leaf, as we have already seen in part, attests not only to its tenth-century manuscript but also to the seventeenth-century Junian transmission of collation, transcription, and edition. It also displays, of course, its subsequent transmission through the eighteenth-century fire and the nineteenth-century restorations. The paper mount of the restored binding functions as a mat in a picture frame, which highlights the main text but incidentally conceals surviving parts along the edges. Gough made the individual frame for this folio by tracing the uniquely damaged shape of the recto side on the heavy construction paper. He then cut out the center of the paper, following the jagged pencil outline but leaving a retaining space of about 2 mm. inside the traced line. Next he applied paste to this retaining edge and pressed the damaged leaf into place. He also used tape around the edges to secure the inserted manuscript leaf as well as onion skin over holes to prevent further crumbling of the brittle vellum. On folio 38 verso (fig. 2) one can see this opaque skin covering the hole in the top part of the leaf. Also visible on the verso, some later tape helps hold the leaf to the paper mount at the top and along the lower left side, where the old paste has begun to give way.

Cotton Otho A. vi in its current state has become an iconic page of a restored medieval manuscript. The fire damage in the middle of the page is most striking, telling the story of the fire, a terrible transmission from medieval to modern times. Its early modern provenance is painfully exposed in its burns, the losses at the edges, the violent shrinkage of the vellum and distorted misplacement of the text, and its complete disintegration in the middle. The tiny jagged edges of the recto page show the precise shape of the damaged leaf, pasted over and taped to the retaining edge of the paper binding frame. The smoother edges of the verso are produced by the retaining edges of the paper frames, which often hide bits of the text they cover. Most of the top line of folio 38 verso is almost entirely covered by this retaining edge. The most expressive part of the page is thus an unintentional tale it tells of the fire it survived and the importance the manuscript still holds for its owners, who have obviously gone to great trouble and expense to preserve what remains of each leaf in their uniquely miserable states. The modern facsimiles in microfilm and print reinforce its importance by testifying, less accurately than the inlaid manuscript leaves, to its continued transmission in modern times. For all it expresses through its physical condition the page seems to be by comparison almost silenced in the text it preserves, as the following transcription may illustrate:

[::: ::::: ::::::: :::::: :::::::]
[::::]ol fæst gereaht þurh þa stro
ngan meaht 7ge ende byrd swa swa
oðra sint wo[:::: :::::]f[:: ::::]
hit la þonne [::::: ::: ::::::]
gif hit [:::::: :::]
Her end[:: :: ::::: ::::::]
boc boeties 7onginð þridde.
Se boetius wæs oðre naman ha
ten seuerin[:: :]e wæs heretoga
romana
          A se wisdom þa ðis lioð
          asungen hæfde þa hæfde
          he me gebundenne mid
          þ[:]re wynsum nesse his san
[::: ::: :: :::] wæs swiðe wafiende 7 swi
[:: :]ustbære to geheranne mid
[::::]wearde mode 7 þa fulhræðe
[::: :: :]leopode to him 7 þus cwæð
[:::: :::]dom þu ðe eart sio heh
[::: :::::]r ealra werigra moda
[:: ::] me hæfst afrefredne æg
þer ge mid þinre smealican spræ
ce . ge mid þinre wynsumnesse þines
sanges to þæm þu me hæfst n[:]
aretne 7ofercumene mid þin[::]
ge sceadwisnesse þæt me nu ðincð ðæ[:]
te no þæt an þæt ic [::]s unwyrd aræfnan

This grim impression is at least in part fostered by the fact that there is so little recognizable formatting on the pages. Moreover, there are two types of gaps (or silence), one deliberate, left by the scribe awaiting the later insertion of a large initial Þ or Ð between lines 12-15; and the other accidental, left by the fire, mostly along the top and left side, and in the area of the taped hole between lines 4 and 7.

With the aid of ultraviolet fluorescence, fiber-optic backlighting, and image processing, we can restore many of the seeming losses around the hole and the edges. For example, one can make out some areas where the covered edge of the leaf shines through the paper mount, especially along the top and sides. At first glance 38 verso appears to contain only twenty-seven lines, instead of twenty-eight, but a closer look reveals some descenders of letters along the top covered edge. When illuminated with strong backlighting, nearly the entire line, [gif hiora modsefa meahte weorðan], is quite legible. Using the same method, many seemingly lost letters along the left margin are also visible and thus test the accuracy of Junius's transcripts. Ultraviolet fluorescence, moreover, does a very good job of restoring illegible text not covered by the paper frames. But even after these invisible parts (in italics here) have been made visible again, the text remains difficult to understand and not only because it is written in Old English.

[gif hiora modsefa meahte weorðan]
[stað]ol fæst gereaht þurh þa stro
ngan meaht 7ge ende byrd swa swa
oðra sint wo[ruld gescea]f[ta wære]
hit la þonne [murge mid monnum]
gif hit [meahte swa]
Her end[að nu seo æftre frofer]
boc boeties 7onginð [sio] þridde.
Se boetius wæs oðre naman ha
ten seuerin [us s]e wæs heretoga
romana
          A se wisdom þa ðis lioð
          asungen hæfde þa hæfde
          he me gebundenne mid
          þ[æ]re wynsum nesse his san
[ges þæt ic his] wæs swiðe wafiende 7 swi
[ðe l]ustbære to geheranne mid
[inne]wearde mode 7 þa fulhræðe
[ðæs ic c]leopode to him 7 þus cwæð
[eala wis]dom þu ðe eart sio heh
[ste frofe]r ealra werigra moda
[hu þu] me hæfst afrefredne æg
þer ge mid þinre smealican spræ
ce . ge mid þinre wynsumnesse þines
sanges to þæm þu me hæfst n[u]
aretne 7ofercumene mid þin[re]
ge sceadwisnesse þæt me nu ðincð ðæ[t]
te no þæt an þæt ic [ða]s unwyrd aræfnan

Even with the page restored with the help of MS Junius 12, MS Bodley 180, ultraviolet fluorescence, and fiber-optic backlighting, and with the added help of an Old English dictionary, there is virtually no punctuation and so no obvious way to tell where sentences begin and end. After the large A at v12 (which is not the start of a word, despite the capitalization), there is no punctuation at all until v24, where a point, or "period," finally occurs but does not mark the end of a sentence. In fact, there is no easy way to tell even where words begin and end, much less sentences, because lexical boundaries are frequently ignored. See, for example, 7ge ende byrd v3 for ond geendebyrd and fulhræðe v 18 for ful hræðe. Sometimes the ambiguity creates spurious textual evidence that early English words might begin with the non-English initial sequence ng- (ngan [38v2-3]), until it becomes evident without the help of hyphens that it is part of a word begun in the preceding line (strongan). Hyphens would also be helpful to identify words split by line breaks at v9-10, v15-16, v16-17, v20-21, v22-23, v23-24, and v27-28. Proper names are not distinguished by capital letters (cf. boeties [v8], boetius [v9], seuerinus [v10], romana [v11]). Further complicating the message for modern readers are the unfamiliar insular script in general and the special characters, the abbreviations, the seven-like ampersand, and the crossed thorn for þæt (that). Finally, it is not at all obvious that the page begins with a verse passage, because there is no difference in the writing of the prose and the verse sections -- no lineation, no line numbers, no marking for verse in any way.

Although they do provide punctuation, glossaries, explanatory notes, and the like, the existing modern editions are remarkably unhelpful guides to reading this particular manuscript. The prose edition by Sedgefield was published in 1899 to commemorate the millennium of King Alfred's death and has remained the "standard" edition for almost a hundred years now. Whereas the tenth-century manuscript began with a prosimetrical preface explaining the circumstances of the Alfredian translation, Sedgefield's edition provides only the prose part on the opening page while banishing the verse to a distant appendix at the end of the book. Whereas the tenth-century manuscript divides the translation into five books, roughly along the lines of its Latin source, Sedgefield's edition interpolates a forty-two-chapter structure, including a four-page table of contents (3-6), from the other, twelfth-century, manuscript. And, whereas the tenth-century, prosimetrical. manuscript integrates prose and verse throughout, following its Latin source, Sedgefield's edition removes all of the verse passages and replaces them with the corresponding prose from the same twelfth-century manuscript. As Sedgefield puts it:

This edition of King Alfred's most personal and therefore most important work, may be regarded as a compromise, aiming as it does at meeting the wants of the scientific investigator of Old English on the one hand, and of the student of our literature on the other. The former has long been calling for a text which should faithfully reproduce the Cotton MS., and in offering him such a text I ask him to bear in mind the difficulty of the task. Even when MSS. are in perfect condition it is almost impossible to avoid errors in copying, as will be readily granted by those who have had experience of this kind of work; but when, as in the case of the chief MS. of the Old English Boethius, fire and water have played havoc with almost every page, the labour of deciphering and the liability to error are immeasurably increased. Individual pages of this MS. have received as much as an hour's scrutiny, and this scrutiny was repeated three or even four times in a few instances. By taking advantage of the rare intervals of London sunshine during the winter and spring months, I found much decipherable which in ordinary light would have remained hidden. (Sedgefield, Preface, vii)

Consider Sedgefield's edition of folio 38 verso:

(...þis moncyn wære gesælig, gif heora mod wære swa riht 7 swa gestaðelod 7 swa geendebyrd swa swa þa oðre gesceafta sindon.) Her endað nu seo æftre froferboc1 Boeties, 7 onginð sio þridde. Se Boetius wæs oðre naman haten2 Seuerinus; se wæs heretoga Romana.
                                        XXIIn
§ i. Ða se Wisdom þa ðis lioð asungen hæfde, þa hæfde he me gebundenne mid þære wynsumnesse his san[ges, þæt ic bis] wæs swiðe wafiende 7 swi[ðe lu]stbære (hine) to geheranne mid [inne]wearde mode, 7 þa falhræðe 7 [ðæs ic] cleopode to him 7 þus cwæð: [Eala, Wis]dom, þu ðe eart sio heh[ste fro]fer ealra werigra moda3; hu þu me hæfst afrefredne ægþer ge mid þinre smealican spræce, ge mid þinre wynsumnesse þines sanges. To þæm þu me hæfst n[u] aretne4 7 ofercumene mid þin[re] gesceadwisnesse, þæt me nu ðincð ð[æt]te no þæt an þæt ic [ðas] unwyrd aræf[nan....]

Replacing the Cotton Meter, Sedgefield uses the italicized text in parentheses at the beginning to mark the text as coming from the later prose manuscript, Bodley 180. Sedgefield regularly indicates in the margins where the folios begin and end in the Cotton manuscript, but his practice of substituting prose passages from Bodley 180 for the verse of Cotton Otho A. vi means that there are no references to Cotton folio numbers in these places. As a result, there are no references at all to folios 37 and 38 between page 48, where the marginal note says "[folio] *36b C[otton Otho A. vi] about here," and page 50, where the note says "*39a C." In another departure from the Cotton format, the chapter heading XXII comes indirectly (and inaccurately) from the Bodley organization. In fact, on folio 30 the twelfth-century Bodley 180 has the XXII chapter mark before the notice of the ending of the book. Moreover, the scribe left a space for a specially drawn H in " [H]er endað seo æf[t]re frofer . . . ," showing that the change from book 2 to book 3 was meant to begin the new chapter. In his transcript Junius nonetheless placed the chapter heading after this passage. Note n, attached to the chapter number, identifies the following passage as coming from "Boeth. iii. pr. 1. 'Iam cantum illa finiuerat,' etc." -- that is, "when this song was finished." Footnotes 1-4 alert the reader to differences in Junius 12 and Bodley 180, giving the misleading impression that the edited text is primarily from Cotton Otho A. vi. The unobtrusive § i, marking the first prose section of book 3, is a quiet and confusing concession to the five-book structure emphasized in the layout of the Cotton manuscript.

Meanwhile, these verse passages expurgated from the Cotton manuscript end up incongruously numbered seriatim, following the misplaced verse preface in the appendix, named by Sedgefield "The Old English Version of the Lays of Boethius" (151-204). These disembodied poems have often become a text unto themselves, printed separately without their prose contexts, as in the standard edition of this modern reconception, The Paris Psalter and the Meters of Boethius, volume 5 of the Anglo-Saxon Poetic Records, in which their new context becomes the similarly extracted Old English verse glosses of a Latin psalter.18 It is remarkable that, proceeding from the practice begun by Christopher Rawlinson at the end of the seventeenth century, all modern editions have printed the Meters separately, as if they made sense as a group outside of their larger context. The separation makes sense for Bodley 180, because in that manuscript the Meters are indeed an early modern addition, or appendix, to the main text. But they should be recognized as a deletion, not an appendix, in any supposed edition of Cotton Otho A. vi. The practice of divorcing the prose and the poetry has even carried over into the facsimiles, as in the EEMF volume, in which only those manuscript pages containing verse are reproduced. The advantage of the facsimile edition over the printed edition in this respect is that the facsimiles perforce include the prose contexts at the beginnings and ends of the verse passages. In contrast to Sedgefield's, consider Krapp's edition of folio 38 verso, Meter 11 in his seriatim numbering but Meter 8 of book 2 in Boethius:

gif hiora modsefa meahte weorðan
staðolfæst gereaht þurh þa strongan meaht,
and geendebyrd, swa swa oðra sint 100
woruldgeceafta. Wære hit, la, þonne
murge mid monnum, gif hit meahte swa.
(ASPR 5, 170)

Krapp's entire edition of folio 38 verso is only five lines of verse, the same five lines that Sedgefield excises and replaces with prose from Bodley 180. In Krapp's edition there is no place to mention the passage from book 2 to book 3, only the passage from Meter 11 to Meter 12.

As strange as it may seem, my presentation of the following passage is the first modern edition that simply edits the restored prosimetrical text of Cotton Otho A. vi in the most obvious way -- and as it occurs in the manuscript -- as a prosimetrum divided into five books instead of a prose text divided into forty-two chapters:

[gif hiora modsefa     meahte weorðan
stað]olfæst gereaht     þurh þa strongan meaht,
and geendebyrd,     swa swa oðra sint 100
wo[ruldgecea]f[ta.     Wære] hit, la, þonne
[murge mid monnum,]     gif hit [meahte swa.]
Her end[að nu seo æftre frofer]boc Boeties, 7 onginð [sio] þridde. Se Boetius wæs oðre naman haten Seuerin[us. S]e wæs heretoga Romana.
     A se Wisdom þa ðis lioð asungen hæfde, þa hæfde he me gebundenne mid þ[æ]re wynsumnesse his san[ges, þæt ic his] wæs swiðe wafiende 7 swi[ðe l]ustbære hine to geheranne mid [inne]wearde mode, 7 þa fulhræðe [þæs ic c]leopode to him 7 þus cwæð: [Eala, Wis]dom, þu ðe eart sio heh[ste frofe]r ealra werigra moda; [hu þu] me hæfst afrefredne ægðer ge mid þinre smealican spræce, ge mid þinre wynsumnesse þines sanges. To þæm þu me hæfst n[u] aretne 7 ofercumene mid þin[re] gesceadwisnesse, ðæt me nu ðincð ðæ[t]te no þæt an þæt ic [ða]s unwyrd aræfnan

Because the fire impinged, it is understandable that Rawlinson's all-prose edition held sway by itself for a hundred and thirty-one years. In fact, it remained the model for all our modern editions even after Cotton Otho A. vi was rediscovered. Junius and Rawlinson would no doubt be puzzled to find that their edition of Bodley 180 had become the model of editions of Cotton Otho A. vi. Yet, as J. S. Cardale admits of his 1829 edition, "The present edition is founded on Mr. Rawlinson's, but with such alterations as, it is hoped, will render it more acceptable to the Anglo-Saxon student." Knowing that the Cotton manuscript was always recognized as the best witness, Cardale creates the fiction of restoring the lost manuscript by integrating its readings, as transcribed by Junius, into the edition based on Bodley 180. "The readings taken from the Cotton MS. have been carefully examined, and in numerous instances introduced into the text," he says. But Cardale does not follow Rawlinson in providing all the Meters in his appendix, which instead provides a single "specimen of Alfred's poetry [that] corresponds with the third metrum of Boethius's fourth book" (397). He admits in a fine understatement that "some apology may be thought necessary for not adding the whole of Alfred's poetical versions of the metres. This had been originally intended, but it was found that the insertion of them, with such notes as appeared indispensable, would require a second volume."19 There is no apparent awareness on Cardale's part that a single edition of the Cotton manuscript would be a prosimetrum.

As late as 1835, Samuel Fox, the first editor of the unique verse passages from this manuscript and the first to number them seriatim, was of necessity still founding his text on Junius's seventeenth-century transcript while innocently misinforming his readers that the manuscript itself was destroyed in the fire.20 In his 1864 edition Fox almost as inaccurately claims that Cotton Otho A. vi had become easy to read after its restoration:

The MS. in the British Museum formerly belonged to Sir Robert Cotton, and was so much injured by the fire which destroyed a portion of his valuable collection before its removal to our great public library, that for many years it was utterly useless, consisting merely of detached fragments thrown together in a box, until, by the skill and industry of the Rev. Joseph Stephenson and the late John Holmes, Esq., it was rearranged in 1844, the detached parts being neatly put together within a border of new parchment, and is now rendered so perfect that most of it can be read with the greatest ease!21

Excising the Meters and printing them seriatim at the end, Fox also makes the preposterous assertion that "in this edition every word contained in both MSS. is given, and the variations, which are the result of a careful collation, are marked at the foot of each page.22 As Sedgefield observes, Fox apparently copied the Cotton Meters "from the Junius transcript, or from Rawlinson, without fresh collation with C" (xxii). It is a surprising fact that to this day the Meters of the tenth-century Cotton manuscript have never appeared where they belong in any edition of Alfred's Boethius, even in the most scholarly editions purportedly founded on the Cotton manuscript.

The justification for expurgating the Meters from our earliest extant version of Alfred's Boethius is based on an interpretation of the prose preface common to both tenth- and twelfth-century manuscripts. The preface says:

ÆLFRED KUNING wæs wealhstod ðisse bec, 7 hie of boclædene on englisc wende, swa hio nu is gedon. Hwilum he sette word be worde, hwilum andgit of andgite, swa swa he hit þa sweotolost 7 andgitfullicast gereccan mihte for þam mistlicum 7 manigfealdum weoruldbisgum þe hine oft ægðer ge on mode ge on lichoman bisgodan. Ða bisgu us sint swiþe earfoþrime þe on his dagum on þa ricu becoman þe he underfangen hæfde, 7 þeah ða þas boc hæfde geleornode 7 of lædene to engliscum spelle gewende, 7 geworhte hi eft to leoðe, swa swa heo nu gedon is. (Sedgefield, 1)
[King Alfred was translator of this book, and turned it from book-Latin into English, as it is now done. Sometimes he set word by word, sometimes sense out of sense, as he might most clearly and intelligibly present it, for the various and manifold worldly business that often consumed him in both mind and body. These busy cares are for us very difficult to number that in his time happened in those kingdoms which he had inherited. And yet when he had learned this book and translated it from Latin to English prose, he again reworked it for verse, just as it is now done. (My trans.)]23

Perhaps because Junius in his collation mistakenly recorded eft, "again," as efter, which looked like a late variant spelling of æfter, the word eft has been taken to imply not just "again" or even "afterwards" but also "after a significant period of time." Thus, Cardale renders this passage: "when he had learned this book, and turned it from Latin into the English language, he afterwards composed it in verse . . . " (iii). Like Cardale, all editors have explicitly or implicitly endorsed Wanley's unfounded supposition that Alfred first translated Boethius entirely into prose, while he fought the Danes, and then later, in tranquillity, made a complete verse translation (85). Because Bodley 180 is entirely in prose, the prevailing interpretation of the prose preface has been that Alfred first published an entirely prose version of Boethius and then made an entirely metrical version, which later scribes might or might not use in subsequent copies.

Yet the prose preface explicitly and unequivocally states in both the Cotton and the Bodleian manuscripts that the following work contains both prose and verse. King Alfred, the prose preface in both manuscripts clearly states, geworhte hi eft to leoðe, swa swa heo nu gedon is, "reworked it for verse, just as it is now done." The finished work, the prose preface tells us, is prosimetrical. Because it refers to itself as a prosimetrum, the version in Bodley 180 most likely contains only prose, not because it goes back to the draft stages of Alfred's translation, as scholars have usually supposed, but because it retranslates Alfred's Meters into prose for a twelfth-century audience no longer tolerant of old Old English poetry. It seems that the scribe of Bodley 180 simply omitted the verse preface without sufficiently revising the prose preface to hide the excision. By the twelfth century, long after the Norman Conquest, Old English poetry had become foreign sounding in its rhythms and its excessive alliteration. Shorn of its eccentric Meters, Alfred's Boethius, on the other hand, was still well worth copying in the twelfth-century milieu that produced Bodley 180.

In its time and place the tenth-century prosimetrical version in Cotton Otho A. vi was a very special text, preserving the poems of a beloved king. There are paleographical and codicological codes that help us understand the significance of the confusing signs outlined earlier. The handwriting is a "well-formed Anglo-Saxon minuscule," attesting to a well-trained scribe in an established scriptorium.24 By collating the hair and flesh sides of the separate vellum leaves, one can still reconstruct the arrangement of vellum sheets in the ruined original gatherings, revealing that whoever put the quires together was careful to arrange the sheets so that hair faced hair and flesh faced flesh, to give the book a uniform look wherever it was opened. Some of the ruling on the page can be seen clearly on folio 35 recto toward the bottom, where the scribe has left four indented lines blank for a large initial. The arrangement of the sheets and the ruling suggest a division of labor in an organized scriptorium, while the blank space the scribe always leaves for the insertion of a special capital to mark off prose and verse sections is strong evidence in support of this inference. The same kind of space is left blank on folio 38 verso and, in fact, everywhere in the manuscript where either a prose section or a verse section begins. The extremely sparse coding is, paradoxically, evidence of a very high level of literacy among the relatively small groups who used vernacular manuscripts. These special Anglo-Saxons could read this manuscript without the many aids we need today, such as consistent spelling and word boundaries, capitalization of proper names, hyphenation and at least minimal punctuation, and marking for half-lines as well as full lines of verse.

The most important function of the formatting in this manuscript, for its scribes and contemporary readers, at least, was the signal of a change from prose to verse and from verse to prose. It is telling that it is the one medieval bibliographic code that is totally ignored in all modern editions. Another code, marking where one of the five books ends and the next book begins (see the large capital H, helping to mark the transition from bk. 2 to bk. 3 on fol. 38v, ll. 5 and 6), is overrun by the formatting into chapters of the twelfth-century prose manuscript, spelled out in its table of contents, in the standard prose edition and completely ignored in the standard verse edition.

The paper frames of the restored binding increase the original dimensions of the medieval book from octavo to quarto, just as Cotton's sixteenth-century binding increased the thickness of the medieval book by combining previously unrelated materials. Moreover, the paper is somewhat thinner and less flexible than the vellum it frames, and as a result the book no longer closes properly. The act of restoration produces many additional footnotes, by washing away the ink and requiring ultraviolet restoration or by chipping off edges of text and hiding remainders. The frames also make the text into a picture, or a picture of a text, rather than a text. As we have seen, on the versos the paper frame, like a mat in a picture frame, has cropped out some of the text along the edges, permitting new discoveries of hidden material with the help of backlighting. Gough has even drawn a framing line around each manuscript leaf, as if to emphasize their status as individual pictures (see fig. 3). The frame becomes a text, too -- one that in some respects has become more important than the manuscript leaf it is framing. The frame not only holds the foliation number, which places the leaf in a sequence; it also holds annotations identifying the fragmentary text in reference to some edition of Alfred's Boethius, in which the full text may be consulted. An annotation in the top margin, "Cap. XIX. XX," identifies the text as belonging to chapters 19 and 20 of a different exemplar. In the top left another note, "poetr. p. 162 col. 2 l. 30," points out that the passage is poetry that can be found on page 162, column 2, line 30, identifying this exemplar as none other than Rawlinson's late-seventeenth-century edition. Unless we know the edition, we are bound to be startled by the last notes in the left margin, "p. 163. col. 1 l. 28" and "prose p. 43 l. 30 cap. XX," pointing out that the poetry ends on page 163 but that the prose begins on page 43. If we were to update this notation for Sedgefield's edition, we could say just as incongruously that the Meter ends on page 166 and the following prose comes well before it, on page 46. It is hard to come to terms with the fact that this manuscript is, strictly speaking, no longer a medieval manuscript but, rather, a nineteenth-century facsimile edition of a medieval manuscript. As iconic pages, the framed manuscript leaves have actually hidden parts of the extant manuscript and, even more disconcerting, have finally led us away from the tenth-century prosimetrical manuscript itself to a seventeenth-century edition of a very different, twelfth-century manuscript, all in prose.25

Notes

1. N. R. Ker assigns the uniform handwriting of this manuscript to the fifty-year span around the middle of the tenth century (s. x med.) in his Catalogue of Manuscripts containing Anglo-Saxon (Oxford: Clarendon Press, 1957; reissued with supplement, 1990), Ker 167, p. 217. Humphrey Wanley, in his famous catalog, held that the manuscript was written during Alfred's lifetime or very shortly after his death ("quo vivente, aut saltem paullo post obitum ejus"). Catalogus historico-criticus (1705), English Linguistics: 1500-1800, no. 248 (Menston, Eng.: Scolar Press, 1970), 217.

2. Walter J. Sedgefield, ed., King Alfred's Anglo-Saxon Version of Boethius, de Consolatione Philosophiae (Oxford: Clarendon Press, 1899) (hereafter cited as Sedgefield); and George Philip Krapp, ed., The Paris Psalter and the Meters of Boethius, The Anglo-Saxon Poetic Records 5 (New York: Columbia University Press, 1932) (hereafter cited as ASPR 5).

3. See Andrew Prescott, "'Their Present Miserable State of Cremation': The Restoration of the Cotton Library," in Sir Robert Cotton as Collector, ed. C. J.Wright (British Library Publications, forthcoming), 391-454, esp. 399, 405 ff; hereafter cited as Prescott.

4. A Catalogue of the Manuscripts in the Cottonian Library, Deposited in the British Museum (London: n.p., 1802); Planta routinely refers to such manuscripts as "deest" or "desideratur."

5. There is a facsimile of these parts of Junius's transcript in Fred Robinson and E. G. Stanley, Old English Verse Texts from Many Sources. A Comprehensive Collection, Early English Manuscripts in Facsimile 23 (Copenhagen: Rosenkilde and Bagger, 1991) (hereafter cited as EEMF 23).

6. The concept of a medieval hypertext is perceptively discussed by C. M. Sperberg-McQueen, "Text in the Electronic Age: Textual Study and Text Encoding, with Examples from Medieval Texts," Literary and Linguistic Computing 6.1 (1991): 42.

7. The best way to study these pages is in a digitized image, which permits separate examination of individual parts, close collation of relevant material, and magnification of details. In my own experience studying the large manipulable electronic images was far easier than studying the manuscript in situ.

8. For a facsimile, see EEMF 23:5.1 Oxford, Bodl. Libr., Junius 12, fol. iiibr: The Meters of Boethius, Proem beg. l. 5. The cumbersome modern foliation number, iiibr, underscores the complexity of what I have characterized as a hypertext.

9. Simon Keynes and Michael Lapidge refer somewhat ambiguously to Bodley 180 as "the sole manuscript which preserves this prose preface," in Alfred the Great: Asser's 'Life of King Alfred' and Other Contemporary Sources (New York: Penguin Books, 1983), 298 n. 3. Both prose and verse prefaces in Cotton Otho A. vi were destroyed in the fire after Junius respectively collated and transcribed them. Allen Frantzen is mistaken when he says that Cotton Otho A. vi "does not include the [prose] preface, which must have been composed after the prose translation of the meters had been versified." King Alfred (Boston: G. K. Hall, 1986), 44.

10. Kenneth Sisam represents this view when he says, "It is generally agreed that [Alfred] made the prose translation, or rather paraphrase, of both the Prose and the Metres; that the Anglo-Saxon verse metres were produced by simply turning this prose into metre; and that the result has no merit as poetry." Studies in the History of Old English Literature (Oxford: Clarendon Press, 1953), 293.

11. See muntgeof in Joseph Bosworth and T. Northcote Toller, eds., An Anglo-Saxon Dictionary (1898; reprint, London: Oxford University Press, 1973); the variant spelling -giop in this text, with palatal g and final p, is perhaps a phonetic representation of -Alp.

12. Junius refers to it as "Cottonianus optimæ notæ codex MS" in Junius 12, fol. iiibr.

13. An. Manl. Sever. Boethi Consolationis Philosophiæ libri v. Anglo-Saxonice redditi ab Alfredo, inclyto Anglo-Saxonum rege. Ad apographum Junianum expressos...Oxoniæ...MDCXCVIII.

14. "VERSIONES POETICÆ é Codice Cottoniano desumptæ," 150-98.

15. See Prescott, 414-16. Solanders were invented by Daniel Solander, a Swedish botanist, who settled in England in 1759 and was appointed to catalog the Natural History collections at the British Museum. He went on Captain Cook's voyage on the Endeavour in 1763 and later became keeper of printed books. Solanders are still in general use in the British Museum, particularly in the Prints and Drawings collections.

16. See Prescott, 416; and R. W. and Gretchen P. Ackerman, Sir Frederic Madden: A Biographical Sketch and Bibliography (New York: Garland, 1979), 22-23.

17. The two short texts from the middle of the codex, Excerpta quædam Historica de S. Edwardo Rege & Confessore (fol. 141) and Historiola de quodam insano ad bonam mentem restituto per merita S. Edwardi (fol. 142), were lost either by fire or carelessness before the book could be rebound. The manuscript codex the Anglo-Saxons called froforboc, "the book of comfort or consolation," was in any case very different from the one Cotton rebound with three unrelated texts about Edward the Confessor. The book we now call Cotton Otho A. vi is likewise very different from the one that faced the fire in 1731.

18. While making available formerly inaccessible Old English texts, the Anglo-Saxon Poetic Records in many cases stripped Old English poetry of its manuscript context of prose and, in the tradition of Sir Robert Cotton, often recombined them with completely unrelated texts.

19. King Alfred's Anglo-Saxon Version of Boethius' De Consolatione Philosophiae (London: 1829), preface, unnumbered pp. ii and iii.

20. Samuel Fox, King Alfred's Anglo-Saxon Version of the Metres of Boethius (London: William Pickering, 1835), iii-iv.

21. Preface, iii. This is a garbled, secondhand account. Although the restoration was completed in 1844, Gough inlaid the vellum leaves in paper frames, as is evident in the illustrations for this essay. The reference to Holmes and Stephenson is puzzling. While there is some slight evidence that Holmes was briefly assigned to work on the restoration, Stephenson worked at the British Museum only between 1831 and 1834. There is ample evidence, moreover, that Sir Frederic Madden took over the restoration himself. See Prescott, 414-16, and 445 n. 119, 120.

22. Preface, iv. Sedgefield discloses the actual extent of Fox's collation. "Up to Chapter xxix," he says, "only the readings from C [i.e., the Cotton manuscript] excerpted by Junius and printed in Rawlinson's edition are given in footnotes; afterwards these latter are supplemented by some more variants taken directly from C" (xxii).

23. The preface concludes with a prayer:

7 nu bit 7 for Godes naman he halsað ælcne þara þe þas boc rædan lyste, þæt he for hine gebidde, 7 him ne wite gif he hit rihtlicor ongite þonne he mihte; forþamþe ælc mon sceal be his andgites mæðe 7 be his æmettan sprecan þæt he sprecð, 7 don þæt þæt he deþ. (Sedgefield, 1)
[And now he prays and before God's name begs everyone of those who wish to read this book, that he pray for him and not reproach him if he [the reader] understands it better than he could. Because each man must according to the measure of his intelligence and his leisure speak what he speaks and do what he does. (My trans.)]

24. Ker, Catalogue of Manuscripts, 218.

25. I wish to thank Robert Edwards for his incisive comments as a respondent to an earlier version of this essay at the Iconic Page symposium; and Andrew Prescott for providing me with a prepublication copy of his revealing study of the restoration of the Cotton Library.