This is an electronic version of an essay first published in Representations 32 (Fall 1990), 157-174, and reprinted in Old English Literature: Critical Essays, edited by Roy Liuzza (Yale University Press, 2002), pp. 103-124.

Reading Cædmon’s “Hymn” with Someone Else’s Glosses

Kevin Kiernan

ACCORDING to the Venerable Bede, our first reliable English historian, English literature had a miraculous origin in the late seventh century in a religious somniloquy by an illiterate cowherd named Cædmon. Writing at least a half century after the miracle, Bede represents Cædmon’s Old English “Hymn” in only a Latin paraphrase in his Historia Ecclesiastica Gentis Anglorum. Our earliest vernacular versions of the “Hymn” appear, not as part of Bede’s text, but rather as notes later appended by scribes to two eighth-century manuscripts of the Historia Ecclesiastica.1 From its humble start as a marginal, secondary text, the vernacular “Hymn” first worked its way into the central, primary text by means of a tenth-century Old English translation of Bede’s entire History.2 It continued to appear, nonetheless, as a marginal text from the eleventh to the fifteenth century in Latin manuscripts of Bede. Nowadays scholars are generally convinced that we have inherited by this process authentic witnesses of Cædmon’s debut as a poet; in fact, they print the “Hymn,” in both scholarly editions and general anthologies, as the central text, with Bede’s Historia Ecclesiastica relegated to the margins. The textual history of Cædmon’s “Hymn” provides an unmiraculous case history of how re-productions of literary texts both purposely and unintentionally re-present our past.

Bede tells the story of Cædmon in book 4, chapter 24 of his Historia Ecclesiastica.3 Anglo-Saxonists feel sure that everyone knows and remembers this story, in modern translation needless to say, but it will still be useful to represent it here in summary form as our point of departure. 'We all realize of course that Bede’s Latin will lose some of its dignity in the process.

Cædmon was a layman who worked on an estate near the monastery of Whitby in Northumbria. Sometimes at parties (in convivio) his fellow workers would agree among themselves to liven things up by singing songs in turn. Yet whenever he saw the cithara coming his way, Cædmon would always duck out, not knowing any songs whatsoever. One night, when it is his turn anyway to take care of the cattle, he escapes to the stables, where he falls asleep and is urged by a visitor in a dream to sing about Creation. He complies, and Bede gives us a Latin paraphrase of his “Hymn”: “Nunc laudare debemus auctorem regni caelestis, potentiam Creatoris,” and so on. Before resuming the story, Bede makes it clear that he is giving “the sense but not the order of the words which [Cædmon] sang as he slept. “For,” Bede says, “it is not possible to translate verse, however well composed, literally from one language to another without some loss of beauty and dignity.” Following this somewhat academic intermezzo, Cædmon wakes up with the “Hymn” on his mind and, poetically inspired, adds some more verses to it. Cædmon then tells his boss the reeve about his gift, and the reeve takes him to Abbess Hild at the monastery for advice. After determining that God had granted him a special grace, Hild persuades Cædmon to join the monastery, where he launches a productive career, in the cloistered context of Whitby, as an oral composer of devout verse.

As far as we are concerned, this concludes the story of our first English poet. Writing ecclesiastical, not literary, history, however, Bede finishes his story differently. Many of our literary historians fail to tell the rest of the story, that our first poet opposed “with a flaming and fervent zeal” anyone who failed to submit to the monastic rule and that he died like a saint, predicting the exact hour of his death (EH, 418-21). These are not the high-points of his career for us.

At this proto-stage in English literary history, at any rate, all we had inherited of our first poet’s oeuvre was Bede’s Latin paraphrase of Cædmon’s oral performance while he dreamed.4 Scholars have not inquired too deeply into the exact nature of Bede’s source, no doubt because Bede specifies that he translated the short version Cædmon sang in his dream, not the longer one produced the next morning when Cædmon “added more verses in the same manner, praising God in fitting style.”5 As his Latin translation attests, moreover, Bede never had any intention of preserving the original Old English version. To put it bluntly, English literature did not for Bede fit into the grand scheme of things. Where, then, did we get our vernacular text? In light of its first preserved manifestation in a Latin paraphrase made more than a half century after the original illiterate dream, it is no small part of the overall miracle that Cædmon’s “Hymn” has come down to us in Old English in seventeen Bede manuscripts dating from the eighth to the fifteenth century. As wondrous as it may seem, moreover, new printed renditions continue to appear in the twentieth century. By taking a look at the manuscript tradition, we can perhaps learn something about our own Cædmonian dreams in the age of print.

Unfortunately, looking at this particular manuscript tradition is not a spectacle for casual observers. It too must begin, of course, more than a half century after Cædmon’s oral performances, sleeping or waking, and it ends a couple of centuries closer to us than to Cædmon. By sorting the manuscripts in a variety of inconsistent ways, we find that there is a Northumbrian group of manuscripts and a West Saxon group, provided we remember that West Saxon is a generic literary dialect in late Old English times; there is an early Northumbrian group of two and a different, very late, twelfth- and fifteenth-century, Northumbrian group of two; there is as well a fairly early “West Saxon” or “Alfredian” group, and a distinct, somewhat later and longer-lived group, also called West Saxon.6 Perhaps the most basic division revolves around the variant readings, ælda (ylda in the West Saxon manuscripts) and eorðan, both deeply fixed in the manuscript tradition. Two Northumbrian and six late West Saxon manuscripts agree that Bede’s Latin cliché, filiis hominum, exactly glosses “the sons of men,” ælda barnum or ylda bearnum, coming authentically from Cædmon,7 while two Northumbrian and seven West Saxon manuscripts agree that Cædmon sang an extraordinary phrase, “the sons of the earth,” eorðan bearnum, which Bede blandly glossed with a cliché.8 Of the ten manuscripts dating from Old English times, five belong to the ælda barnum or “sons of men” group, including the two eighth-century Northumbrian and the three late-eleventh-century West Saxon manuscripts, always with the West Saxon spelling, ylda, for Northumbrian ælda.9 The remaining five, belonging to the eorðan bearnum or “sons of the earth” group, are the tenth- and eleventh-century West Saxon versions in the Old English translation of the Ecclesiastical History.

Fairly soon, it seems, after Bede’s death in 735 some of his near contemporaries provide us with the first extant vernacular versions of Cædmon’s “Hymn” by adding them to the two eighth-century manuscripts of the Latin Historia Ecclesiastica. The Leningrad Bede almost certainly derives from Bede’s own monastery of Jarrow, while the Moore Bede may well have some connection with its twin monastery at Wearmouth.10 Like the related eleventh-century ylda manuscripts, neither of these eighth-century versions of Cædmon’s “Hymn” was ever part of Bede’s Latin text nor of the original design of the manuscripts. The central, overriding interests of early Northumbrian Christianity had the effect of literally marginalizing the English text of Cædmon’s “Hymn,” for it was the idea of the “Hymn,” not its language, dialectal idiosyncrasies, or even textual authority that was important in the context of the conversion of pagans. It is perhaps no coincidence that there was a resurgence of Latin manuscripts of the Historia Ecclesiastica and a return to the “marginalized” text of Cædmon’s “Hymn” during the eleventh century, when a well-established Anglo-Saxon Church again faced the task of converting pagans and apostates, this time among the Anglo-Danes and Scandinavian immigrants in the newly subsumed Danelaw. In both contexts, at any rate, an English version of Cædmon’s “Hymn” would not have been a burning priority.

There were conceivably many reasons, including the love of one’s own language, for individual scribes to transmit an English text of the “Hymn” in the margins of a Latin manuscript of the Historia Ecclesiastica. But in view of the unusually large number and overall uniformity of the surviving marginal texts, there must have been some institutional support for the practice--the propagation of copy texts, for example, for monastic libraries with Latin Bede manuscripts already in their collections.11 One motive particularly appropriate for manuscripts coming from Jarrow and Wearmouth must have been to strengthen the authority of Bede. What better way of showing Bede’s special veracity, or the prestige of a particular Bede manuscript, than by providing the source of his paraphrase in an endnote, as the Moore manuscript does? Mastering its source in high scholarly fashion, the scribe even adds the comment “Primo cantauit Cædmon istud carmen,” “Cædmon first sang that song.”12 Yet it would be wrong to conclude that the marginal notes were themselves meant to carry the authority they paradoxically add to the Latin Bede manuscripts. A closer look at the way the Moore manuscript produces the text of the “Hymn” as an endnote can help put all the marginal notes in perspective.

Moore Bede
Figure 1. The Moore Bede, fol. 128v. From The Moore Bede, ed. Peter Hunter Blair, Early English Manuscripts in Facsimile, vol. 9 (Copenhagen, 1959). Reproduced by permission of Rosenkilde and Bagger, Ltd.

A single scribe produced the Moore Historia Ecclesiastica using what Peter Hunter Blair has justly described as “an Anglo-Saxon minuscule of austere beauty” (MB, 15). The text ends on fol. 128v25 with a formal Explicit in red, but four lines, obviously written by the same scribe, are added at the bottom of the page with annals for the years 731-34, citing events that occurred after Bede completed his work but before he died. Turning the page, one is immediately struck by the reduced size of the script in lines 1-3, the Old English text of the “Hymn,” and in line 4, which begins with the attribution to Cædmon (fig. I). The small scale of these four lines sharply contrasts with the usual size of the script, which resumes in lines 5-12 for some chronological notes known as the “Moore Memoranda.”13 The contrasting scale alone should tell us that the chronological notes, like the main text of the Historia Ecclesiastica, had more prestige than Cædmon’s “Hymn.” The reduced size of the script for the “Hymn” reminds one of a gloss rather than a main text, and indeed in line 4 the scribe adds three apparently random glosses of hard Latin words--arula hearth, destina feur-stud, Iugulum sticung--only the first two of which appear anywhere in Bede. The glosses apparently mean “hearth,” “fire-proof stud,” and “pig killing,” judging by the meager information about these Old English words in our dictionaries, and at first glance would seem to have more to do with a pig roast than with epoch-making poetry.14 Strangely careful not to waste any space on a page that presumably was blank when lines 1-4 were written, the scribe also squeezes in “nota rubrica,” “mark in red,” at the end of line 4, apparently a reminder to someone to rubricate the manuscript.15 Today we wouldn't dream of printing such trivial items next to Cædmon’s “Hymn.”

As we can plainly see from this manuscript, as well as from all the ones that add Cædmon’s “Hymn” to the margins of Bede’s Historia Ecclesiastica, Anglo-Saxon scribes did not view these versions, as we do today, as central texts. Figure 2, for example, shows the other early Northumbrian copy of the “Hymn” from the Leningrad manuscript. As in the case of the Moore Bede, this vernacular text of the “Hymn” was added to its manuscript, too, at a later time.16

Leningrad Bede
Figure 2. The Leningrad Bede, fol. 107r. From The Leningrad Bede, ed. O. Arngart, Early English Manuscripts in Facsimile, vol. 2 (Copenhagen, 1952). Reproduced by permission of Rosenkilde and Bagger, Ltd.

The possibility that the Old English versions we have inherited in this way began as glosses, or reverse translations of Bede’s Latin paraphrase, warrants more attention than it has yet received.17

The suspicion that all of our versions ultimately derive from the tenth-century Alfredian translation of Bede was in fact brought up repeatedly in the nineteenth century.18 Richard Wülker argued the case most effectively in 1876, showing why he thought the Northumbrian version at the end of the Moore manuscript was only a dialectal transliteration of the West Saxon version from the Alfredian Bede.19 His arguments, however, were insecurely based on inadequate representations of the “Hymn” and its manuscript setting before a facsimile was available. As a result, Julius Zupitza effectively demolished his theory by citing the internal dating of the Moore version of Cædmon’s “Hymn” in 737, about a century and a half before King Alfred came along.20 Wülker later conceded that the Northumbrian version in the Moore manuscript (the Leningrad version had not yet been discovered) was written in the eighth century, long before the Alfredian translation.21 His concession, however, does not eliminate the possibility that the earliest Northumbrian version might be a reverse translation of Bede’s Latin paraphrase. As David Dumville has recently observed, “There must be room for the conjecture that we have only nine lines of the Old English because that is all Bede gave in his Latin prose rendering: in short, the Old English is at least as likely to be a poetic rendering of Bede’s Latin as the source of his words; otherwise we could have expected more of the poem to have been given by the person who added these nine lines at the end of the Moore manuscript.”22

One way of approaching the question is to scrutinize Bede’s claim that he was unable to translate the “Hymn” word for word into Latin. If he had made a literal translation of the vernacular text that has come down to us, Bede would have ended up with the following text:

Nunc debemus laudare regni caelestic auctorem,
Nu scylun hergan   hefaenricaes uard,
Creatoris potentiam et illius consilium,
metudæs maecti   end his modgidanc,
facta Patris gloriae: quomodo ille miraculorum omnium,
uerc uuldurfadur   sue he uundra gihuaes,
[cum sit] aeternus Deus, auctor exstitit,
eci dryctin   or astelidæ;
qui primo tecti hominum filiis
he aerist scop   aelda barnum
caelum pro culmine,
heben til hrofe,   [haleg scepen,]
dehinc terram humani generis Custos
tha middungeard   monocynnaes uard;
[eci dryctin    aefter] tiadæ;
[firum foldu,   frea] allmectig.

The glosses show that Bede would have been able to “paraphrase” the “Hymn” word for word: “Nunc debemus laudare regni caelestis auctorem, Creatoris potentiam et illius consilium, facta Patris gloriae: quomodo ille miraculorum omnium, cum sit aeternus Deus, auctor exstitit, qui primo tecti hominum filiis caelum pro culmine, dehinc terram humani generis Custos creavit omnipotens.” The result may not be classical Bedan prose, but it is perfectly intelligible Medieval Latin, following the pattern of the vernacular, and after all, Bede was supposedly translating an Old English poem.23 But if Bede actually paraphrased this version of Cædmon’s “Hymn,” why did he precisely translate phrase by phrase for two-thirds of the poem and then leave out three half-lines of verse, haleg scepen, line 6b; eci dryctin, line 8a; and firum foldu, line 9a; as well as the two alliterative staves, aefter in line 8b and frea in line 9b? It seems especially strange for him to omit the new epithets for God, haleg scepen and frea, and to eliminate all the alliteration, the most salient feature of the verse. It is difficult to appreciate how Bede might think the inclusion of these things would spoil his translation.24 If we take the position, instead, that an enterprising Anglo-Saxon myth maker translated Bede’s Latin into Old English, we can see that he (or she) translated all of Bede’s Latin version and, compelled by the meter, boldly added a few half-lines and provided the necessary alliteration.

The scribes who added the “Hymn” as footnotes or marginal notes all use the smaller script characteristic of glosses, which first become prevalent in Anglo-Saxon England during the eighth century, around the time of Bede.25 Elsewhere, interlinear continuous glosses ultimately end up as Old English poems in, for example, the Kentish “Hymn,” the psalms of the Paris Psalter, and the Meters of Boethius.26 With all due respect to the memory of Cædmon, it would not have taken a major poet to turn Bede’s paraphrase into the “Hymn” that has come down to us. As we have seen, a glossator with no poetic skills at all would have ended up with about two-thirds of the poem’s locutions, all metrically viable after the Latinate inversions had been turned around, from Nu [we] scylun hergan (Nunc laudare debemus) in line 1 to eci dryctin (aeternus Deus) in line 8, simply by making a straight-forward interlinear or marginal gloss. With Bede’s caveat before him about the difficulty of translating poetry word for word, an intelligent glossator familiar with the conventions of Old English poetry could finish the job in his sleep. The same argument holds for Bede’s translation from Cædmon to Latin, perhaps, but the exceptional occurrence of scylun, “must,” without the personal pronoun we, corresponding to Bede’s debemus, strongly suggests that the gloss moved from Latin to English.27

The vicissitudes of a living oral tradition give us plausible ways of explaining some substantive variations in the text of Cædmon’s “Hymn.” However, the extraordinary retention of the reading sculon, “must,” still without the pronoun we, in tenth- and eleventh-century manuscripts in the otherwise distinct eorðan group, convinces me that the Old English poem must have descended from eighth-century Northumbrian to late West Saxon as a changing written text, not an oral one. The oral tradition would not have perpetuated a formula native speakers could not understand, whereas the manuscript tradition did in fact perpetuate it, with the help of an accurate scribe, in what is considered the best copy of the Alfredian Bede, Bodleian Library MS Tanner 1o. The scribe of Corpus Christi College Oxford MS 279 first copied “Now must,” too, but then corrected the traditional blunder in his copy text to “Now we must,” a graphic illustration of what a native speaker of Old English ought to make of a first-person plural verb without a first-person-plural pronoun.28 This brave scribal correction of what we call “a deep-seated corruption” may also show how late scribes of the Old English Bede felt free to make intelligent revisions of Cædmon’s “Hymn.” As Katherine O’Brien O’Keeffe has observed about the tenth- and eleventh-century eorðan manuscripts, “In the variations ... weorc/wera/weoroda; wuldorfaeder/wuldorgodes; wundra/wuldres; gehwaes/fela; or/ord; sceop/gescop we see a dynamic of transmission where the message is not embellished but where change within the formula is allowed” (i5). We can sensibly attribute these variations to the “dynamic of translation” brought about by King Alfred’s native cultural program at the end of the ninth century.

Public relations in the Middle Ages in fact associated the English translation of Bede’s Historia Ecclesiastica with King Alfred himself. Although scholars now agree that all five of our manuscripts of the Old English Bede descend from a Mercian, rather than an Early West Saxon, archetype, it remains more than likely that Alfred’s reforms directly inspired the original translation.29 We know by definition that the Latin source was in front of the translator, and we may safely surmise from the well-attested manuscript tradition that many of these copies already contained Old English translations of Cædmon’s “Hymn” in the margins. With a copy of the Latin Bede and an Old English translation of the “Hymn” from the ælda/ylda group in the margins, we can easily explain the later development of the eorðan group of manuscripts. If he recognized that the marginal version of Cædmon’s “Hymn” was no more than a metrical, paraphrasing gloss of Bede’s Latin paraphrase, the Old English translator would have had the incentive to rework these materials. Encouraged by Bede’s comment that he was providing only a paraphrase of Cædmon’s “Hymn,” and not deterred by the clearly unauthoritative glosses in the margins, the first Old English translator of Bede’s Historia Ecclesiastica was free to produce a new and more compelling version of Cædmon’s “Hymn” than the copy text provided.

This late Old English translator of Bede takes some wise liberties with the Latin source. In the story of Cædmon, for instance, the phrase in convivio comes out as in gebeorscipe, “at beer parties,” and the cithara becomes the familiar Anglo-Saxon hearp, or “harp.” For our purposes, of course, the most important changes in this part of the Old English Bede are the ones directly relating to the new production of Cædmon’s “Hymn.” As we have already seen, the translator made his own metrical paraphrase more credible by rendering the biblical phrase filius hominum with the entirely original phrase eorðan bearnum, “the sons of the earth.”30 But he also made three big changes that for the first time accorded extraordinary authority to the original text of the “Hymn” by implying that Bede himself transmitted it in its original form. First, the translator renders Bede’s disclaimer iste est senses, “this is the general sense” of the “Hymn,” with the far more purposeful words þære endebyrdnesse þis is, “this is its proper disposition.” Second, he replaces Bede’s Latin prose paraphrase of the “Hymn” with a vibrant Old English poem no one had ever seen in the margins of the Historia Ecclesiastica. And third, the translator completely omitted Bede’s subsequent commentary: “This is the sense but not the order of the words which [Cædmon] sang as he slept,” Bede had said. “For it is not possible to translate verse, however well composed, literally from one language to another without some loss of beauty and dignity.” In these three changes, the producer of the Old English History gave his readers in the tenth and eleventh centuries the first really believable, seemingly authentic, text of Cædmon’s “Hymn” by concealing its actual history.

I would like to leave the manuscript tradition of Cædmon’s “Hymn” at this juncture, as it is about to drift off into the later Middle Ages, and look briefly at the printed tradition of the text in the multi-tiered institution of academia today. There were, of course, many texts produced in manuscript and print in the long interim that have helped shape our modern editions. What is immediately remarkable today is that the text has descended to fully fledged Anglo-Saxonists in only two “standard” editions, the prestigious, early Northumbrian, or “Bedan” version and the prestigious, “early” West Saxon, or “Alfredian” version, with copious footnotes to both recording the multitudinous variants from all the other manuscripts (fig. 3).31

Figure 3. From The Anglo-Saxon Minor Poems, ed. Elliot van Kirk Dobbie, Anglo-Saxon Poetic Records, vol. 6 (New York, 1942), 106-7. © 1942 Columbia University Press. Reprinted by permission of the publisher.

There is, most remarkably, no sign of the Latin or Old English prose of Bede’s Historia Ecclesiastica in these standard editions. The poem alone now comes to us handsomely bound, in this example as part of the six-volume set of The Anglo-Saxon Poetic Records. For us, the endnote or marginal note, Cædmon’s “Hymn,” has become the central poetic text while the old central prose text, Bede’s Latin History, has become the endnote or the footnote.

The text of Cædmon’s “Hymn” also descends to us in what is called “normalized” spelling, idealized linguistic forms meant to help students learn the largely unattested Early (with a capital E) West Saxon dialect of Old English associated with King Alfred (fig. 4).32

Seven Old English Poems
Figure 4. From Seven Old English Poems, ed. John C. Pope, 2d ed. (New York, 198 I), 4. Reproduced by permission of the estate of John C. Pope.

This new version of the “Hymn,” reconstructed in “normalized” Early West Saxon from the two early Northumbrian versions, is prominently centered on the page, while a modern translation of most of Bede’s Latin prose account of Cædmon’s story is marginalized in the commentary at the back of the book. Notice in the textual footnotes that John Pope has introduced no fewer than ten variants in his reconstructed Early West Saxon text from Bodleian Library MS Tanner 10, “the best of the West-Saxon texts.” Of the two “substantive variants” the most radical is the change in line 5b of eorðan to *ielda, a spelling of West Saxon ylda displaying the Early West Saxon shibboleth, ie. As Pope says in his preface, this ie spelling is the “most conspicuous mark of Early West Saxon” (viii). Indeed, without it, we wouldn't be able to tell that the reconstructed text was Early West Saxon at all: seven of the remaining “Early” spellings occur in one or more of the surviving West Saxon manuscripts of the “Hymn”; the other two variant spellings are found in other West Saxon manuscripts. Despite its conspicuousness, *ielda occurs nowhere in the three-million-word corpus of Old English prose and verse, not to mention any of the manuscripts of Cædmon’s “Hymn.”33 In contrast, the well-attested late West Saxon spelling, ylda, occurs often in Old English, eight times in the phrase “the sons of men,” in such divergent texts as Genesis, Daniel, Beowulf, the Paris Psalter, the Menologium, and A Prayer. As we have seen, it is not only a cliché but a loan translation of a Latin biblical cliché, filiis hominum. The translation ylda bearnum occurs six times in the six different copies of Cædmon’s “Hymn” found in the context of the Latin Historia Ecclesiastica.

In many ways, the most interesting and influential reproduction of the text of the poem today can be found in anthologies for general readers. As if by some miracle of literary history, Cædmon’s “Hymn” now appears as the first English text produced in the 2,616-page first volume, fifth edition, of the ubiquitous Norton Anthology of English Literature.34 Quite properly, the Norton editors try to give their general readers a simple, straightforward account of the prevailing critical position about the “Hymn.” Thus they present it in the immediate context of “An [sic] Ecclesiastical History of the English People,” without raising any questions about the antiquity, priority, or authenticity of the Old English poem. “The story we reprint,” they say, “preserves what is probably the earliest extant Old English poem... ... Bede tells how Cædmon, an illiterate cowherd employed by the monastery of Whitby, miraculously received the gift of song, entered the monastery, and became the founder of a school of Christian poetry” (19). They especially stress the idea that Cædmon, contrary to Bede’s insistence that “he was not taught the art of song by men or by human agency but received this gift through heavenly grace,” was an oral-formulaic poet:

Cædmon was clearly an oral-formulaic poet, one who created his work by combining and varying formulas--units of verse developed in a tradition transmitted by one generation of singers to another. In this respect he resembles the singers of the Homeric poems and oral-formulaic poets recorded in the twentieth century, especially in the Balkan countries. Although Bede tells us that Cædmon had never learned the art of song, we may suspect that he concealed his skill from his fellow workmen and from the monks because he was ashamed of knowing “vain and idle” songs, the kind Bede says Cædmon never composed. Cædmon’s inspiration and true miracle, then, was to apply the meter and language of such songs, presumably including pagan heroic verse, to Christian themes.35

In their efforts to distill an extremely complicated textual history and to introduce students to oral-formulaic theory, the editors have created a diminutive Old English monster. As a result, by looking more closely at this extraordinary production, we can get some idea of how we sometimes unintentionally rewrite literary history (fig. 5). The modern English translation effectively hides the unprecedented combination of languages, dialects, manuscripts, and versions in the Norton reproduction. Under the big general heading of “Old English Literature” the Norton first prints a modern English translation of “Anglo-Latin Literature,” Bede’s story of Cædmon. We can recognize the source by Bede’s remarks introducing and then commenting on his Latin paraphrase of Cædmon’s “Hymn.” Yet these remarks by Bede can appropriately apply only to his Latin paraphrase, not to what we are supposed to think of as Cædmon’s original Old English poem. Having studied the manuscript tradition, moreover, we know that none of the Anglo-Saxon Latin manuscripts contains an integrated copy of the Old English “Hymn,” only marginal copies. The Norton silently integrates its copy from the Old English Alfredian translation in Bodleian Library MS Tanner 10. The manuscript is hard to recognize at first glance because eorðan is for some reason emended to *ielda,36 which even in its attested forms of ælda and ylda only occurs in marginal copies in Old English manuscripts.

While performing their awesome, traditional duty of inaugurating all of English literature, the nine strange lines of Cædmon’s “Hymn” are made stranger by the modern English glosses and the huge caesuras, which unintentionally present the poem as if it should be read in columns, each with a left justified margin, like the footnotes at the bottom of the same page, rather than from left to right, like Old English poetry.

Norton I
Figure 5. From The Norton Anthology of English Literature, ed. M. H. Abrams et al., 5th ed., vol. 1 New York, 1986), 20-21. Reproduced by permission of W W. Norton and Co.

An unsuspecting reader, an undergraduate taking a survey course in English literature, for example, or perhaps even a specialist in romanticism teaching the survey, can by an amusing coincidence make almost as good sense out of the text by reading the modern glosses in “column one” (all the a-verses) first, and then proceeding to “column two” (the b-verses). Here is Norton’s “Hymn,” following this naive or otherwise carefree procedure:

Now we must praise
the Measurer’s might,
the work of the Glory-Father,
eternal Lord,
He first created
heaven as a roof,
then middle-earth
eternal Lord,
for men earth,
heaven-kingdom’s Guardian
and his mind-plans,
when he of wonders of every one .. ...

It gets a little clumsy at this point, but by reading it in the educated way, overcoming the Norton format, one comes upon this clumsiness sooner:

Now we must praise heaven-kingdom’s Guardian,
the Measurer’s might and his mind-plans,
the work of the Glory-Father, when he of wonders of every one...

But take a closer look at this unfortunate formatting of the Nortonian text. Despite the slightly larger print of the Old English text, it is the modern English version that unintentionally governs the layout, from the gigantic caesuras to the spacing between words in the Old English text. Like an Anglo-Saxon interlinear gloss, individual Old English words, not an edited text, are written above the modern English text, which alone is punctuated to give sense to its words. Having begun in our earliest manuscripts as a marginal gloss of a Latin paraphrase and having achieved an apotheosis as a central text at least twice in its polyglossarial history, Cædmon’s “Hymn” has ended up in the Norton Anthology at least looking like the most prestigious gloss in all of English Literature. When the editors fix the formatting of Cædmon’s “Hymn” in a future edition, they will have the salutary chance to see again, to re-vise, the way our ancient and modern reproductions re-present (or gloss) our past.


A version of this paper was read in 1989 at Texas A&M University as part of a conference on “(Re)producing Texts/(Re)presenting History,” sponsored by the Interdisciplinary Group for Historical Literary Study. I would like to thank Barrett Watten for exceptional help in preparing this article for publication.

1 From dubious internal evidence the Moore Bede (Cambridge University Library MS Kk.5.16) is usually dated in the year 737 and the Leningrad Bede (M. E. Saltykov-Shchedrin Public Library MS Q.v.I.18) in 746. For a discussion of these dates see Kevin S. Kiernan, “The Scribal Deconstruction of ‘Early’ Northumbrian,” ANQ 3 (1990): 48-55.

2 Thomas Miller, ed., The Old English Version of Bede’s Ecclesiastical History of the English People, 4 vols., Early English Text Society, o.s., nos. 95, 96, 110, 111 (1890-98; reprint ed., Millwood, N.Y., 1978-88).

3 Bertram Colgrave and R. A. B. Mynors, eds., Bede’s Ecclesiastical History of the English People (Oxford, 1969), hereafter cited in the text as EH.

4 Early scholars credited Cædmon with composing many other extant Old English poems. In fact, the eleventh-century Bodleian Library MS Junius 11, one of the four great Old English poetic codices, was formerly known as the Cædmon Manuscript, and it still carries this inscription in its current binding. The facsimile by Israel Gollancz is likewise entitled The Cædmon MS of Anglo-Saxon Biblical Poetry: Junius XI: In the Bodleian Library (Oxford, 1927).

5 Bede was neither old enough nor near enough to Whitby to record the miracle in the stables. John Pope seems to leave open the possibility that someone else was on the scene. He says, “Since Bede makes it plain that Cædmon was unable to read and write when he entered the monastery, and gives no indication that he learned to do so later, we may assume that the hymn was first recorded by someone else, quite possibly during Cædmon’s lifetime if not on the occasion of the alleged miracle”; Seven Old English Poems, 2nd ed. (New York, 1981), 51; all references to Pope are from this edition. Bede also makes it plain that Cædmon’s second oral performance of the “Hymn” was different from the first, all encouraging one to wonder about the exact age and authenticity of our extant version(s).

6 See E. V. K. Dobbie, The Manuscripts of Cædmon’s Hymn and Bede’s Death Song (New York, 1937), 10-48. “Alfredian” in this context refers to the group of five manuscripts preserving the Old English translation of Bede’s Ecclesiastical History, apparently made by a Mercian scholar working for Alfred. As Dobbie says, “I am using the term ‘West Saxon’ here, in accordance with the general custom, to indicate all the non-Northumbrian texts of the Hymn, without prejudice to the probable Mercian origin of the Alfredian Bede” (22, n. 31).

7 The two eighth-century Northumbrian manuscripts are the Moore (ælda) and Leningrad (aeldu) texts. Of the six late (tenth-to-fourteenth-century) West Saxon manuscripts (with the spelling ylda for ælda), all except the fourteenth-century manuscript contain the “Hymn” as a marginal addition.

8 The five tenth- and eleventh-century manuscripts of the Alfredian Bede all preserve the “Hymn” in the main translation. The two late (twelfth- and fifteenth-century) Northumbrian manuscripts both incorporate the Old English version of the “Hymn” in the Latin Bede.

9 The best recent discussion of the manuscript tradition is Katherine O’Brien O’Keeffe, “Orality and the Developing Text of Cædmon’s Hymn,” Speculum 62 (1987): 1-20.

10 See E. A. Lowe, “A Key to Bede’s Scriptorium: Some Observations on the Leningrad Manuscript of the Historia Ecclesiastica Gentis Anglorum,”' Scriptorium 12 (1958): 182-90. Colgrave and Mynors suggest that the scribe of the Moore Bede “was on a visit to Wearmouth or Jarrow, or had the loan of a copy from there in his own monastery for a limited time”; Ecclesiastical History, xliv. It is worth noting that the scribe goes out of his way to mention the founding of Wearmouth in the “memoranda” at the end of the manuscript (fol. 128v11).

11 M. B. Parkes has recently drawn attention to urgent requests from Boniface in 746 and 747, indicating “a heavy demand for Bede’s work on the continent”; “The Scriptorium of Wearmouth Jarrow,” Jarrow Lecture (1982), 15.

12 The Moore Bede, Cambridge University Library MS Kk.5.16, ed. Peter Hunter Blair, Early English Manuscripts in Facsimile, no. 9 (Copenhagen, 1959), fol. 128v4; hereafter cited in the text as MB.

13 See Peter Hunter Blair, “The Moore Memoranda on Northumbrian History,” item 6 in Anglo-Saxon Northumbria, ed. M. Lapidge and Peter Hunter Blair (London, 1984), 245-57.

14 In fact all three glosses derive from the text of Bede if Hunter Blair is right that Iugulum was listed for the finite forms of iugulare, “to slaughter”; Moore Bede, 13, n.14.

15 The same scribe apparently wrote “rubricaui rubrica numeros,” “I have rubricated the numbers in red,” in the left margin, beginning at 128v12. Explaining this note, Hunter Blair observes that the chapters were later numbered “in red ink in the margin throughout the manuscript”; Moore Bede, 28.

16 To my knowledge no one has noticed that the “Hymn” was added by a different scribe with similar but not identical handwriting. The distinguishing feature is the truncated descender on the letters f, p, r, and s, compared to the long descender on the letter g. In the main text, with exactly the same space between lines, the descender on these letters is invariably long and spiky. For the view that the handwriting is identical, see for example O. Arngart, The Leningrad Bede: An Eighth Century MS of the Venerable Bede’s Historia Ecclesiastica Gentis Anglorum in the Public Library, Leningrad, Early English Manuscripts in Facsimile, no. 2 (1952), 30; and A. H. Smith, ed., Three Northumbrian Poems (New York, 1968), 19.

17 Arngart reflects the current view. “Suffice it to say,” he says, “that it is now recognized to be the original of Bede’s Latin paraphrase, not as was once maintained an Anglo-Saxon translation of the passage in the Latin text”; Leningrad Bede, 30-31. Bruce Mitchell briefly raises strong evidence to the contrary when he says, “One can see that, if the Latin version with debemus had come first, scylun alone could be explained as a careless gloss for it; compare Coll 253 wyllap wesen wise, Latin uolumus esse sapientes.” But he immediately dismisses this possibility; “Cædmon’s Hymn, Line 1: What Is the Subject of Scylun or Its Variants?” in On Old English (London, 1988), 92.

18 John Lingard first proposed this theory in 1896 in The Antiquities of the Anglo-Saxon Church. “The Anglo-Saxon verses are found in King Alfred’s translation of Bede,” Lingard said, “and are generally supposed to have been transcribed by that prince from some ancient copy. I think it, however, equally probable, that they were the composition of the royal translator.” For convenience I cite from the later American edition, based on the 1810 British edition (Philadelphia, 1841), 316.

19 Richard Wülker, “Ueber den Hymnus Caedmons,” Beiträge zur Geschichte der deutschen Sprache und Literatur (Tübingen) 3 (1876): 348-57. Wülker notes that in 1851 Benjamin Thorpe also argued that the “Hymn” from the Alfredian Bede was a “retranslation” of Bede’s Latin paraphrase (352-53).

20 Julius Zupitza, “Über den Hymnus Cädmons,” Zeitschrift fur deutsches Altertum und deutsche Literatur 22 (1878): 2 10-23.

21 Richard Wülker, Grundriss zur Geschichte der angelsächsischen Litteratur (Leipzig, 1885), 117-20.

22 David Dumville, “‘Beowulf’ and the Celtic World: The Uses of Evidence,” Traditio 37 (1981): 148. Dumville is mistaken that Dobbie and Hunter Blair thought that the “Hymn” was added by another scribe (148, n. 182). E. V. K. Dobbie says it was added “apparently by the same scribe who wrote the text of Bede”; The Anglo-Saxon Minor Poems, Anglo-Saxon Poetic Records, no. 6 (New York, 1942), xcv. Hunter Blair concurs, referring to “the seeming identity of handwriting with the remainder of the manuscript”; Moore Bede, 27.

23 The only peculiar constructions involve the verbs, especially tecti, which Bede places after pro culmine.

24 Wülker agreed with Zupitza in 1885 that these parts of the text were left out because, as Bede said, “it is not possible to translate verse, however well composed, literally from one language to another without some loss of beauty and dignity”; see Grundriss, 119-20.

25 See Henry Sweet, ed., The Oldest English Texts, Early English Text Society, o.s., no. 83 (1885; reprint ed., Oxford, 1966), for editions of eighth-century glosses.

26 M. C. Morrell, A Manual of Old English Biblical Materials (Knoxville, Tenn., 1965), observes that “if the Old English metrical psalms really have an Anglian origin, they belong then to that first flowering of Christian piety that had its centers in the north, and the tradition of metrical translation is even older than the tenth-century Benedictine Reform” (149)...

27 It is easy to explain scylun as a careless gloss of debemus, but hard to account for a native speaker like Cædmon omitting the identifying pronoun. Mitchell reveals that there is no acceptable evidence supporting the editors’ view that “we could be unexpressed at the beginning of a poem in which it does not occur and in which there was therefore no first person grammatical referent”; “Line 1,” 90.

28 A second Alfredian Bede, BL Cotton MS B.xi., supposedly read Ne sculon, “not must,” but this part of the manuscript was destroyed in the Cotton fire of 1731. This misreading in fact comes from a sixteenth-century transcript, now BL Additional MS 43, 703, by Laurence Nowell, who presumably conflated Nu and we in the Cotton text.

29 See Simon Keynes and Michael Lapidge, Alfred the Great (New York, 1983), 29, 33-34; and Allen Frantzen, King Alfred (Boston, 1986), 8.

30 According to Dobbie,

The evidence of other poems would ... favor aelda barnum as the original reading of the Hymn at this point, for this formula is found not only frequently in Anglo-Saxon, but in the other Germanic dialects as well, whereas eorðu bearnum (or eorðan bearnum) is, so far as I know, unexampled elsewhere. And, though we cannot be quite rid of the possibility that eorðu was the original reading of the Hymn in 1.5, the weight of the evidence favors the conclusion that aelda barnum, as found in the original of M and L and in the hominum of the Latin text of Bede, were the words written [sic] by Cædmon himself.
Manuscripts, 48; Dobbie is not the only scholar to forget that Cædmon was illiterate.

31 These “standard” Bedan/Northumbrian and Alfredian/West Saxon editions are presented in the same way in Smith’s Three Northumbrian Poems (38-41) and Dobbie’s Anglo-Saxon Minor Poems (195-6).

32 The reconstructed dialect is more romantically than rigorously founded on characteristic spellings in unrelated manuscripts of texts closely linked to Alfred’s cultural reform. As Campbell says, “The ‘Early West Saxon’, which has come to be regarded as a grammatical norm, is based on the Parker MS. of the Old English Chronicle from the beginning to 924, the two oldest manuscripts of Ælfred’s translation of Gregory’s Cura Pastoralis, and the Lauderdale MS. of Ælfred’s translation [sic] of Orosius”; Old English Grammar, 8-9.

33 Antonette diPaolo Healey and Richard L. Venezky, A Microfiche Concordance to Old English (Newark, Del., 1981).

34 The Norton Anthology of English Literature, ed. M. H. Abrams et al. (New York, 1986), 19-22.

35 This modern analysis of the miracle is indebted to Fred Robinson and Bruce Mitchell, who likewise say the “Hymn” “attests to a minor miracle of literary history that cannot be denied: in these polished verses Cædmon demonstrated that the ancient heroic style was not incompatible with Christian doctrine and hence was worthy of preservation”; A Guide to Old English, 4th ed. (Oxford, 19886), 204. Bede would doubtless be chagrined to learn that modern readers rationalized his miracle as our first literary hoax.

36 The inspiration for the Norton conflation is presumably Pope’s “The Hymn Normalized in West-Saxon Spelling Based on the Northumbrian Version of MSS M and L,” Seven Old English Poems, 4, the only other place I know of where *ielda and Bodleian Library MS Tanner 10 can be found together. The grave accent on the e of heofon (line 6a) and the spelling ða for þa are minor errors in transcription of the Tanner MS, not odd features of some other West Saxon manuscript.