[In the essay below, Kiernan reviews historical and linguistic evidence which he contends indicates that Beowulf is contemporary with the extant manuscript.]

It may well be surprising that a study of Beowulf in conjunction with its unique ms represents a radical departure from all previous approaches to the poem. In fact, the Beowulf ms has scarcely been studied at all. It still holds a wealth of undiscovered paleographical and codicological evidence, which, under ordinary circumstances, textual scholars would have uncovered and weighed long ago, as a matter of course, for the purpose of founding a reliable text. This evidence has remained safely hidden away because most editors of the poem have relied on photographic fss of the ms, and, often enough, modern transcriptions of the fss, rather than on the ms itself. Their tacit justification for this decidedly unorthodox procedure is that the ms cannot possibly hold any relevant textual evidence that fss would not show as well. For, however variegated and contentious Beowulf studies are in all other respects, there was until very recently complete unanimity in the view that Beowulf is an early Anglo-Saxon poem preserved in a late Anglo-Saxon ms. The chronological gulf between the poem and the ms is usually reckoned to be two to three centuries. Under these circumstances, we are lucky to have an extant ms at all, but still rather unlucky to have such a late one. Surely, the broad paleographical features of a ms that ends an untraceably ancient transmission of the archetype are not textually vital. Beowulf and the Beowulf Manuscript challenges the unproven premise that Beowulf is an early poem. It argues instead, ultimately on the basis of extraordinary paleographical and codicological evidence, that the poem is contemporary with the ms.

The argument of the book is presented in three main stages. The first stage reconsiders the historical and linguistic evidence that has seemed to justify the neglect of the ms, and concludes that, historically and linguistically, both the poem and the ms could have been created in the early 11th century. The second stage is an extensive physical description of Cotton Vitellius A. xv., the composite codex in which the Beowulf ms now resides. The immediate relevance of this description is that it affords a clear view of the construction of the Beowulf ms, in relation to the construction of contiguous texts, and leads to the conclusion that the poem could have been copied in the early 11th century as a separate codex. The third stage studies the Beowulf ms as a contemporary ms of the poem. A variety of paleographical and codicological evidence shows that the poem may have been still undergoing revision while the ms was being copied, and that it was again undergoing revision later in the 11th century. In short, the evidence suggests that the actual creation of Beowulf, as we now know it, is partially preserved in the ms that has come down to us.

This relatively straightforward argument is necessarily embedded in a long and complicated train of subsidiary arguments that can be fruitfully summarized here. Part 1, “The Poem's Eleventh-Century Provenance,” argues that the paleographical dating of the ms, roughly between 975 and 1025, can be safely placed on historical grounds after 1016. Anglo-Saxon scriptoria during the reign of æthelred Unræd (978-1016) would not have copied a poem that praised the Danish Scylding dynasty, while the latest Scyldings, led by Swein Forkbeard, plundered and murdered throughout the country. By 1016 the Danes had conquered England. Swein's son, Cnut the Great, soon made England the center of his dynasty, and by the way provided a suitable historical context for copying a poem like Beowulf. But Cnut's reign (1016-1035) also provided an excellent environment for the creation of the poem, and we must not neglect the exciting possibility that the poem is contemporary with the extant ms. Until now, the origin of the poem has nearly always been restricted to the 8th century or earlier on the rough historical grounds that a poem eliciting sympathy for the Danes could not have been composed by Anglo-Saxons during the Viking Age of the 9th and 10th centuries. Historically, at least, there is a better argument for an 11th-century, post-Viking origin of the poem, since an 8th-century poem would still have to be transmitted by Anglo-Saxons through the Viking Age.

The great problem of accepting an early 11th-century provenance of Beowulf is linguistic, not historical. On closer scrutiny, the linguistic arguments for an early date (or against a late date) are by no means decisive. The specific linguistic tests of the poem's antiquity are especially weak: the syntactical tests magnify the occurrence in Beowulf of some acknowledged archaisms found in unquestionably late verse; the phonetic-materical tests are based on subjective interpretations of the meter that require sweeping, yet inconsistently selective emendations to unrecorded, early, linguistic forms; and the lonely phonetic-morphological-orthographical test is based on a ms “ghost.” The most compelling linguistic evidence that Beowulf is an early poem that has endured a long and complex transmission is its rich mixture of linguistic forms. The language of the extant ms is basically Late West Saxon, but this base is permeated with apparent non-West Saxonisms and, more significantly, with earlier linguistic forms, all of which would seem to prove that the poem had passed through many dialect areas on the way to its present form. There are, however, other explanations for this mixture of forms that do not rule out an 11th-century provenance, and do not even require a linguistically diverse transmission of the text.

Late West Saxon was a standard literary dialect used throughout England in the early 11th century. Complete orthographical uniformity, however, even in West Saxon territory, cannot be expected in an age before printing and formal dictionaries, and all Late West Saxon texts exhibit a natural mixture of forms, including some early forms. If the Beowulf ms was copied in non-West Saxon territory, there would be good reason to expect the occasional intrusion of late, non-West Saxon spellings in the text, as indeed is the case. The mixture of forms in the poem is further complicated by the fact that Anglo-Saxon poets of all periods shunned the language of prose, and consciously employed an artificial, archaic, poetic diction, whose roots were in Anglian territory. This fact explains the persistent occurrence of certain early, non-West Saxon forms beside the basically Late West Saxon language of the ms as a whole. Finally, if the poet and his two scribes each spoke a slightly different dialect, a confusing mixture of occasional spellings in the ms might well have been inevitable. An 11th-century convergence of all of these factors in Beowulf explains the mixture of forms, and accordingly eliminates the need to presume a long and complicated transmission of the text. The 11th-century provenance of Beowulf is historically and linguistically possible.

Part 2, “The History and Construction of the Composite Codex,” prepares the way for a close analysis of the Beowulf ms. The composite codex, known as British Library MS Cotton Vitellius A. xv., to which the Beowulf ms belongs, has never been fully described in English, and a full description is desperately needed. The only existing study (in German) is not only out of date and rather inaccessible; it is also demonstrably inaccurate in dating, foliation, and collation. Moreover, it is rendered practically worthless as a reference, because among its many other errors it mistakenly uses a foliation of the codex that was abandoned in 1884, thirty-five years before the description of the codex was published. But a complete description of Cotton Vitellius A. xv. is needed for another, even more urgent reason. What is now the official foliation of the Beowulf ms is inaccurate, and it cannot be corrected without discussing Cotton Vitellius A. xv. as a whole. A new description of the entire codex, then, has the practical advantage of facilitating references to the Beowulf ms.

At present, it is very difficult indeed to make clear references to folios in the Beowulf ms, because a single, acceptable foliation has not been established. Thus, one fs uses the official foliation, a second fs uses a foliation written on the ms leaves, and a third fs uses the same ms foliation, while acknowledging the technical rectitude of the official foliation. As a result, it is not possible simply to refer, for example, to fol. 133 of the Beowulf ms, for it will not be clear which foliation it belongs to: fol. 133 is the second leaf of the Beowulf ms in the official foliation, but the fourth leaf in the ms foliation. Even those scholars who accept the official foliation have not been able to ignore the ms foliation because of the fss, and because the different numbers on the ms leaves always need to be explained. Consequently, those who accept the official foliation still use the ms foliation in tandem with it, so that fol. 133 is normally referred to as fol. 133(130). Obviously, a complicated system of reference like this vitiates the purpose of a foliation. What makes matters even worse is that the official foliation, in addition to perpetuating some of the errors of the ms foliation, has been pervasively wrong since a flyleaf that it counts was removed from the codex. All of these difficulties are easily resolved by abandoning the confusing official foliation and by correcting the ms foliation in the few places where it is wrong. Except for two numbers, the foliation written on the Beowulf ms is accurate. For convenience of reference it needs to be reinstated as the official foliation.

The inadequacy of the current official foliation is best explained by discussing the flyleaves of Cotton Vitellius A. xv., which account for the marked discrepancies between the official and the ms foliations. A description of these prefixed leaves is useful in itself, for they have not been fully or accurately represented before, but a knowledge of their contents helps justify a return to the ms foliation. In addition, one of the leaves contains a scrap of new information about the state of the Beowulf ms before the fire of 1731. This subject is further investigated in a separate section on the history of the multiple foliations of the composite codex. Until now, scholars have mistakenly thought that the composite codex was not foliated in its entirety until after the fire. The discovery of two distinct foliations before the fire provides startling proof that the Beowulf ms was missing a folio at the time. A study of the various foliations (six different ones are documented and dated) shows precisely how three different folios from Beowulf were shifted from place to place from the early 17th century to the late 18th century.

To an extent, then, it is necessary to study Cotton Vitellius A. xv. as a whole in order to establish a simple and accurate foliation for the Beowulf ms. The utter confusion in the current system of foliation, and the history of the many abortive efforts to provide a correct foliation, are good indications of the astonishing lack of interest scholars have shown in Cotton Vitellius A. xv. and the mss that comprise it. The rest of Part 2 is devoted to describing the physical makeup of the various mss, and to defining their relation to the codex as a unit. To begin with, in Cotton Vitellius A. xv. Sir Robert Cotton artificially combined two totally unrelated codices, a 12th-century collection known as the Southwick Codex, and an 11th-century collection (which includes Beowulf) known as the Nowell Codex. This basic construction of Cotton's codex is self-evident. What is far more interesting, but not at all self-evident, is the basic construction of the Southwick and Nowell codices. The evidence shows that both are themselves composite codices. The original construction of the Southwick and Nowell codices has remained somewhat obscure to scholars because the fire of 1731 destroyed the threads and folds of the gatherings, reducing books to a stack of separate leaves. To recreate the Anglo-Saxon genesis of the book in which Beowulf was actually written, it is necessary to deduce the most probable construction of the original gatherings.

The task is comparatively simple in the case of the Southwick Codex, though no one before has ever deduced from the evidence at hand that the Southwick Codex is a Middle English composite of two late Old English mss, copied by the same scribe at different times in his life. The original gatherings of these two mss can be confidently reconstructed on the basis of sheet and quire signatures that were made in late Middle English times when the new book was rebound from old mss. But there are no sheet or quire signatures on the leaves of the Nowell Codex, and the job of reconstructing the original gatherings is considerably more complicated. Strangely, the most reliable method of reconstruction has been used to identify only one gathering. Each vellum sheet has, of course, a hair and a flesh side, and the difference in color and texture between them is noticeable. Gatherings can be identified, usually with virtual certitude, by collating the hair and flesh sides of separate folios, to see if two leaves that are presumed to be two halves of the same folded sheet are in fact conjugate. The method is not infallible, because the pattern of hair and flesh sides occasionally permits alternative gatherings, but usually there is other paleographical evidence that confirms one description over another. Moreover, the method can eliminate as impossible some established descriptions of the original gatherings; it can prove beyond doubt that some falsely paired folios are nonconjugate, and so could not once have been a single, folded sheet of vellum.

A close study of the original gatherings of the Nowell Codex leads to a revolutionary view of the Beowulf ms. The usual view is that Beowulf was copied as the fourth item in a basically prose codex. The traditional description of the gatherings seems to confirm this view by showing that the scribe began copying Beowulf within the last prose gathering, making the Beowulf ms an inextricable part of the prose codex. But a close analysis of the hair-and-flesh patterns throughout the Nowell Codex reveals that the scribe could have begun copying the Beowulf ms on a new gathering, while distinct differences in format and in execution indicate that the prose codex and the Beowulf ms were originally copied as separate books. It now seems that the Nowell Codex became a composite codex in two stages: first, the Beowulf ms was combined with the prose codex, probably soon after Beowulf was copied; then, undoubtedly, the Judith fragment was added on in early modern times at the end of this composite codex by ripping out a sheet from the Beowulf ms and using it as a cover for the late accretion.

The discovery that Beowulf was probably copied as a separate codex strengthens the argument that Beowulf is an 11th-century poem preserved in an 11th-century ms. Certainly, it suggests that Beowulf was important to the scriptorium in which it was copied, and that it had an 11th-century audience who understood it and appreciated its merits. Textual scholars have always assumed that the poem was mechanically copied by scribes who, since they were largely ignorant of its meaning, were consequently lazy and inattentive. Yet one of the most striking indications that Beowulf at first existed as a separate codex emerges from the scribes' manifest efforts to provide an accurate copy of the poem. The first scribe carefully proofread his part of Beowulf, but did not proofread the prose texts; the second scribe carefully proofread his own part of Beowulf, and the first scribe's part, but again not the prose texts. Thus, the scribal proofreading of Beowulf alone strongly implies that the Beowulf ms once existed as a separate codex. More important, it is eloquent testimony that the scribes were neither lazy nor inattentive in copying the poem; on the contrary, they understood and appreciated it enough to want an accurate copy. In any case, Beowulf was obviously a special poem in the early 11th century, and all the indications are that it was intelligently copied in our extant ms as a separate codex.

Part 3, “The Beowulf Codex and the Making of the Poem,” studies the ms as a separate codex from the entirely new perspective that the ms and the remarkable poem it preserves share the same 11th-century origin. A belief in the absolute textual and paleographical authority of the ms has, to be sure, revolutionary implications for the study of the poem. Until now, it has been impossible to see the ms as anything other than a very late transcript of a very early poem, and this limited view has not only justified many scores of needless emendations, but has made the conscious neglect of paleographical and codicological evidence appear to be a sound editorial principle. By this view, the 11th-century scribes are so hopelessly distant from the archetype, and separated by so many intermediate copies, that they cannot have had any better knowledge of it than we do today. These assumptions are invalidated by postulating a contemporary ms. If Beowulf had no appreciable transmission at all, the degree of corruption reflected in all current editions must be challenged, for the causes of deep-seated corruption are gone. Moreover, the fresh conviction that the ms has a good chance of being right where it was always perceived to be wrong has a liberating effect on the most intractable cruces. At the same time, a rigorous textual conservatism becomes an exciting means of discovering new and intriguing variations in alliterative and metrical patterns, most of which have gone unnoticed in Beowulf because of the doubtful assumption that an early poet would adhere mechanically to standard patterns. Only by rejecting those emendations and interpolations based solely on an arid application of “rules” can the individual style of a late, traditional poet be studied.

The theoretical authority of the Beowulf ms is most effectively vindicated by the quality and extent of the scribal proofreading. There can be no doubt whatever that the ms was subjected to thorough and intelligent scrutiny, by both scribes, while the copying was in progress and after it was done. The nature of the proofreading firmly establishes the authority of the ms. The two scribes' written corrections and erasures unequivocally identify the kinds of errors each scribe was prone to make, and frequently show, as well, how alert the scribe was in the act of copying, for many corrections were made at the moment of the incipient error. The erasures (all of which were studied under ultraviolet light) are particularly informative in this respect, because the erased material can be used to recreate the causes of scribal error. All previous studies of scribal error in Beowulf are founded on editorial emendations, but this reasoning is circular, and hence untrustworthy. In most cases a conjectural emendation is only presumptive evidence of an error, and even when an error is certain, an emendation without the aid of another text is at best only a good guess at what the correct reading might have been. The scribes, on the other hand, have identified unquestionable errors, and have presumably corrected them on the authority of the exemplar.

The vast evidence of scribal proofreading clearly illustrates the importance of a close paleographical investigation of the ms. Surely, no valid assessment of the reliability of the scribes (or the authority of the ms) can afford to ignore the scribes' written corrections and erasures, and yet no editor has ever taken them into account. On the contrary, there are several cases in all editions of the poem where a scribal correction has itself been subjected to a conjectural emendation. In such cases we can be sure that the ms reading, however difficult it may be, is right, and the emendation is wrong. Paleographical study of the scribes' work also reveals that the second scribe's connection with the ms was not limited to copying and proofreading. Apparently the ms remained in the second scribe's possession, presumably as part of a monastic library, for this scribe continued to work with the ms long after he had originally copied it. He has restored readings that were later damaged through accident or by ordinary wear and tear, most notably on the last page of the ms, where he freshened up a badly faded text. But the most extraordinary instance of his later work is that he made a palimpsest of an entire folio, and copied on it a new text.

The palimpsest provides startling paleographical evidence that Beowulf was revised in the course of the 11th century, long after the original text was copied. It is certain that the entire folio, containing lines 2207-2252, was scraped and washed down after the binding of the ms, for the palimpsest is part of the outside sheet of a gathering, yet the vellum's surface contrasts sharply with its conjugate leaf. Scholars have long believed that all of the original text on the folio in question was freshened up by a later hand, but a recent study has shown that the handwriting is still in fact the second scribe's, in a later stage of development. In any event, there is no credible reason for the palimpsest other than revision. A full paleographical and codicological investigation supports this conclusion in various ways. An objective transcription of the new text on the folio discloses a number of anomalous linguistic forms, which can be interpreted as signs of later attrition in the standard literary dialect, a process that accelerated as the 11th century advanced. A closer look at the badly damaged condition of the text, particularly at the textual lacunas, shows that the revised text was shorter than the original text, that parts of the revised text were erased for some reason, and that a full restoration of the revised text was never carried out.

The incipient state of the text on the palimpsest, and the fact that it contains in any case a late revision, opens the possibility that the Beowulf ms is in effect an unfinished draft of the poem. As incredible as it may seem, there is considerable paleographical and codicological support for the view that the Beowulf ms actually preserves the last formative stages in the creation of the epic. Three lines of text thematically related to the new text on the palimpsest have been imperfectly, but deliberately, deleted on the next folio, verso. The erasing was never finished, though it seems likely that the vellum was being prepared for a new text, as well. Presumably, both folios are part of the same revision. An analysis of the construction of the Beowulf ms provides a possible explanation for the purpose of this revision. The palimpsest begins a self-contained unit of the ms. It is the first leaf of the last two gatherings. The number of sheets in these last two gatherings, the manner in which the sheets are arranged, and the number of rulings to the sheets, are unique in the Beowulf ms. It is possible, then, that this part of the ms formerly existed separately, and was artificially appended to the extant ms. If so, the revised text on the palimpsest may have been written to provide a smoother, more natural transition between the two, originally distinct, and perhaps even totally unrelated mss.

The theory is based on paleographical and codicological grounds, and it is defended with other paleographical and codicological facts. But it is surprisingly corroborated as well by the three-part structure of the poem:

1. Beowulf's fights in Denmark with Grendel and Grendel's
dam;
2. His homecoming, and report to Hygelac;
3. His fight with the dragon and his death in Geatland.

Indeed, many critics have argued that the first and last parts once existed as separate poems, or oral narratives, and that the homecoming was composed at a later stage to link the Danish and Geatish narratives.

The paleographical and codicological evidence leads to precisely the same conclusion. The palimpsest is the first folio of the dragon episode. And there is paleographical and codicological proof that the gathering immediately preceding the palimpsest, which holds the text of Beowulf's homecoming, was copied by the second scribe after he had copied the last two gatherings of the ms. The obvious conclusion is that the Danish and Geatish exploits of Beowulf were first brought together in the extant ms by the second scribe. The aesthetic fusion of these parts does not reflect a dim, romantic view of a non-Anglo-Saxon past, but rather a vivid imaginative response to chilling contemporary events. The fall of a great and noble hero, and the imminent extinction of the race he ruled, was well understood by this 11th-century Anglo-Saxon who had recently seen the fall of Alfred's house and the subsumption of his homeland in the Danish empire. The second scribe begins to look like “the last survivor of a noble race,” while the Beowulf ms, the treasure he continued to polish after the death of his old lord, no longer looks like a reproduction.