Beowulf survives in a single manuscript dated on palaeographical grounds to the late 10th or early 11th century. The manuscript measures 245 × 185 mm.
The poem is known only from a single manuscript, which is estimated to date from close to AD 1000, in which it appears with other works. The Beowulf manuscript is known as the Nowell Codex, gaining its name from 16th-century scholar Laurence Nowell. The official designation is “British Library, Cotton Vitellius A.XV” because it was one of Sir Robert Bruce Cotton’s holdings in the Cotton library in the middle of the 17th century. Many private antiquarians and book collectors, such as Sir Robert Cotton, used their own library classification systems. “Cotton Vitellius A.XV” translates as: the 15th book from the left on shelf A (the top shelf) of the bookcase with the bust of Roman Emperor Vitellius standing on top of it, in Cotton’s collection. Kevin Kiernan argues that Nowell most likely acquired it through William Cecil, 1st Baron Burghley, in 1563, when Nowell entered Cecil’s household as a tutor to his ward, Edward de Vere, 17th Earl of Oxford.
The earliest extant reference to the first foliation of the Nowell Codex was made sometime between 1628 and 1650 by Franciscus Junius (the younger).:91 The ownership of the codex before Nowell remains a mystery.:120
The Reverend Thomas Smith (1638–1710) and Humfrey Wanley (1672–1726) both catalogued the Cotton library (in which the Nowell Codex was held). Smith’s catalogue appeared in 1696, and Wanley’s in 1705. The Beowulf manuscript itself is identified by name for the first time in an exchange of letters in 1700 between George Hickes, Wanley’s assistant, and Wanley. In the letter to Wanley, Hickes responds to an apparent charge against Smith, made by Wanley, that Smith had failed to mention the Beowulf script when cataloguing Cotton MS. Vitellius A. XV. Hickes replies to Wanley “I can find nothing yet of Beowulph.” Kiernan theorised that Smith failed to mention the Beowulf manuscript because of his reliance on previous catalogues or because either he had no idea how to describe it or because it was temporarily out of the codex.
It suffered damage in the Cotton Library fire at Ashburnham House in 1731. Since then, parts of the manuscript have crumbled along with many of the letters. Rebinding efforts, though saving the manuscript from much degeneration, have nonetheless covered up other letters of the poem, causing further loss. Kevin Kiernan, in preparing his electronic edition of the manuscript, used fibre-optic backlighting and ultraviolet lighting to reveal letters in the manuscript lost from binding, erasure, or ink blotting.
The Beowulf manuscript was transcribed from an original by two scribes, one of whom wrote the prose at the beginning of the manuscript and the first 1939 lines before breaking off in mid sentence. The first scribe made a point of carefully regularizing the spelling of the original document by using the common West Saxon language and by avoiding any archaic or dialectical features. The second scribe, who wrote the remainder, with a difference in handwriting noticeable after line 1939, seems to have written more vigorously and with less interest. As a result, the second scribe’s script retains more archaic dialectic features which allow modern scholars to ascribe the poem a cultural context. While both scribes appear to proofread their work, there are nevertheless many errors. The second scribe was ultimately the more conservative copyist of the two as he did not modify the spelling of the text as he wrote but rather copied what he saw in front of him. In the way that it is currently bound, the Beowulf manuscript is followed by the Old English poem Judith. Judith was written by the same scribe that completed Beowulf as evidenced through similar writing style. Worm-holes found in the last leaves of the Beowulf manuscript that aren’t present in the Judith manuscript suggest that at one point Beowulf ended the volume. The rubbed appearance of some leaves also suggest that the manuscript stood on a shelf unbound, as is known to have been the case with other Old English manuscripts. From knowledge of books held in the library at Malmesbury Abbey and available as source works, and from the identification of certain words particular to the local dialect found in the text, the transcription may have taken place there.
Debate over oral tradition
The question of whether Beowulf was passed down through oral tradition prior to its present manuscript form has been the subject of much debate, and involves more than simply the issue of its composition. Rather, given the implications of the theory of oral-formulaic composition and oral tradition, the question concerns how the poem is to be understood, and what sorts of interpretations are legitimate.
Scholarly discussion about Beowulf in the context of the oral tradition was extremely active throughout the 1960s and 1970s. The debate might be framed starkly as follows: on the one hand, we can hypothesise a poem put together from various tales concerning the hero (the Grendel episode, the Grendel’s mother story, and the fire drake narrative). These fragments would have been told for many years in tradition, and learned by apprenticeship from one generation of illiterate poets to the next. The poem is composed orally and extemporaneously, and the archive of tradition on which it draws is oral, pagan, Germanic, heroic, and tribal. On the other hand, one might posit a poem which is composed by a literate scribe, who acquired literacy by way of learning Latin (and absorbing Latinate culture and ways of thinking), probably a monk and therefore profoundly Christian in outlook. On this view, the pagan references would be a sort of decorative archaising. There is a third view that sees merit in both arguments above and attempts to bridge them, and so cannot be articulated as starkly as they can; it sees more than one Christianity and more than one attitude towards paganism at work in the poem; it sees the poem as initially the product of a literate Christian author with one foot in the pagan world and one in the Christian, himself perhaps a convert (or one whose forebears had been pagan), a poet who was conversant in both oral and literary composition and was capable of a masterful “repurposing” of poetry from the oral tradition.
However, scholars such as D.K. Crowne have proposed the idea that the poem was passed down from reciter to reciter under the theory of oral-formulaic composition, which hypothesises that epic poems were (at least to some extent) improvised by whoever was reciting them, and only much later written down. In his landmark work, The Singer of Tales, Albert Lord refers to the work of Francis Peabody Magoun and others, saying “the documentation is complete, thorough, and accurate. This exhaustive analysis is in itself sufficient to prove that Beowulf was composed orally.”
Examination of Beowulf and other Old English literature for evidence of oral-formulaic composition has met with mixed response. While “themes” (inherited narrative subunits for representing familiar classes of event, such as the “arming the hero”, or the particularly well-studied “hero on the beach” theme) do exist across Anglo-Saxon and other Germanic works, some scholars conclude that Anglo-Saxon poetry is a mix of oral-formulaic and literate patterns, arguing that the poems both were composed on a word-by-word basis and followed larger formulae and patterns.
Larry Benson argued that the interpretation of Beowulf as an entirely formulaic work diminishes the ability of the reader to analyse the poem in a unified manner, and with due attention to the poet’s creativity. Instead, he proposed that other pieces of Germanic literature contain “kernels of tradition” from which Beowulf borrows and expands upon. A few years later, Ann Watts argued against the imperfect application of one theory to two different traditions: traditional, Homeric, oral-formulaic poetry and Anglo-Saxon poetry. Thomas Gardner agreed with Watts, arguing that the Beowulf text is of too varied a nature to be completely constructed from set formulae and themes.
John Miles Foley wrote, referring to the Beowulf debate, that while comparative work was both necessary and valid, it must be conducted with a view to the particularities of a given tradition; Foley argued with a view to developments of oral traditional theory that do not assume, or depend upon, ultimately unverifiable assumptions about composition, and instead delineate a more fluid continuum of traditionality and textuality.
Finally, in the view of Ursula Schaefer, the question of whether the poem was “oral” or “literate” becomes something of a red herring. In this model, the poem is created, and is interpretable, within both noetic horizons. Schaefer’s concept of “vocality” offers neither a compromise nor a synthesis of the views which see the poem as on the one hand Germanic, pagan, and oral and on the other Latin-derived, Christian, and literate, but, as stated by Monika Otter: “… a ‘tertium quid’, a modality that participates in both oral and literate culture yet also has a logic and aesthetic of its own.”