Old English Manuscripts: The Scribal Deconstruction of “Early” Northumbrian
Kevin Kiernan
One of the few generalizations one can safely make about Old English manuscripts is that they are the principal source of all our erudition about Old English. Yet Anglo-Saxonists have seldom looked back at the manuscript foundations of our discipline with the same intensity and persistence we have applied to the modern texts and editions derived from them. The time seems right for a rejuvenation if not a renaissance. By reexamining these manuscripts in our postmodern lights, we are bound to discover things our predecessors overlooked or missaw, at least from the perspective of our current angles, which have the novel virtue of slanting in many different directions.
Consider, for example, the case of Early Northumbrian. We have known for a long time that the primary manuscript evidence for the early eighth-century Northumbrian dialect amounts to the two earliest copies of Cædmon's Hymn, nine lines of verse, which everyone uses in one way or another to date the remaining bits of evidence.1 These two Cædmon manuscripts, known as the Moore (M) and the Leningrad (L) manuscripts of Bede's Ecclesiastical History,2 both contain chronological notes scholars have used to date the manuscripts, and the dialects they preserve, very precisely in the first half of the eighth century—in 737 and 746. But in localizing and dating the dialect in this way, they have always had to discount some unhelpful manuscript evidence and rely, of course, on what was helpful. If we choose instead to accept all of the manuscript evidence at face value, however, we will have to face the prospect that Early Northumbrian is later than we think.
The old circle defining Early Northumbrian starts and ends with the internal dating of these two Cædmon manuscripts. A.H. Smith expresses the terse consensus in a one-sentence section on “Dialect” in Three Northumbrian Poems: “On grounds already stated,” he says, “Cædmon's Hymn (MSS M and L) is a priori Northumbrian and a comparative examination of the material in the foregoing paragraphs shows that Bede’s Death Song and the Leiden Riddle are also written in the same dialect.”3 Although he recognizes that the earliest copies of the five lines of Bede’s Death Song and the fourteen lines of the Leiden Riddle come from ninth-century continental manuscripts, Smith places all three poems in the eighth century and assumes without explanation that all three accurately represent the eighth-century Northumbrian dialect.4
In 1705 Humfrey Wanley became the first scholar to explain how internal dating of the Moore manuscript showed that it was written in 737. Wanley observed that, after Bede's Ecclesiastical History was copied, a contemporary scribe, if not in fact the same one (“si non eadem, saltem manu aeque antigua”), wrote Caedmon's Hymn on the top of what is now the last page of the manuscript, and that the same scribe who copied the History wrote the so-called “Moore memoranda” after it.5 These chronological notes record that Paulinus began baptizing in Northumbria 111 years ago; that there was an eclipse of the sun 73 years ago; that AElfwine, the brother of King Ecgfrith, was killed in battle 58 years ago; and that comets appeared and St. Egbert passed on to be with Christ 8 years ago. Wanley realized that only someone writing in 737 could accurately say such things. Unfortunately, the scribe gives us other information of the same ilk, which seemingly dates his work, as well, in 734, in 738, in 741, and in 748. Wanley was also the first scholar to brush these dates aside.
Wanley typically cites the Chronicle, rather than Bede, for most of the dates that confirm 737, but he is conspicuously less searching about the figures that do not confirm 737. Thus, for the first two retrospective dates of the Moore memoranda, the year when Paulinus began baptizing (626), and the year of the eclipse (664), he says “Vice Chronicon Saxonicum,” because these dates both lead us to the year 737. Yet he cites none of the obvious sources, including the Chronicle and Bede's History itself, for the next two entries, which say that Penda died 73 years ago and that Ecgfrith's battle occurred 63 years ago. For the first, Wanley passes over without comment the fact that 73 years after Penda's death (655) adds up to the year 734, not 737. For the second, Wanley merely subtracts 63 from 737 (“h.e. A.D. 674,” “this is A.D. 674,” he says), obscuring the fact that 63 years after the famous battle in which King Ecgfrith died (685) brings us to the year 748, not 737.
Like Wanley, modern scholars have wanted to overlook the conflicting evidence. Henry Sweet. who first identified the Early Northumbrian dialect using the Moore version of Caedmon's Hymn, simply ignores any year after 737. He asserts that “These chronological notes must have been written either in 737 or else between 734 and 737—most probably in 737 ... which is, of course, also the date of the Moore MS. of the History itself.”6 Smith adopts a more democratic stance in his edition, pointing out that, “as five of the nine [chronological entries] (first, second, fifth, seventh, and eighth) point to the year 737 their consistency easily outweighs the irregularities of the remaining four (which are probably due to errors of calculation)....”7 Peter Hunter Blair, leaning heavily on defective lost sources to account for the discrepancies, emphatically argued in his facsimile edition of the Moore manuscript that “beyond reasonable doubt ... the compiler of the Memoranda made his calculations from the year 737 and it seems difficult to escape the conclusion that 737 was in fact the year in which this document was compiled.”8 One can detect from these approaches that, despite the best wishes of some formidable scholars, it was never a simple matter to root Early Northumbrian in the early eighth century.
When the dialect was first defined, moreover, Anglo-Saxonists did not know about the analogous dating of the Leningrad manuscript in the margins of Book V of Bede's History.9 Next to entries in Bede's chronological summary, a different scribe entered his own chronological notes, repeatedly dating the manuscript in the year 746. Damage to this part of the manuscript makes it impossible to treat the dating comprehensively, but as in the case of the Moore manuscript, the Leningrad manuscript also contains some later dates, which scholars likewise assume are mistakes. The number 132, for example, beside the entry for the baptism of Eanflaed in 626 leaves us with the year 758, not 746.10 As soon as these notations came to light in 1912, scholars correctly saw that they paralleled the evidence in the Moore manuscript in some way, but curiously failed to draw the conclusion that the compiler of the Moore memoranda must have compiled his notes from the margins of a similar exemplar.
The direct source of the Moore memoranda can still be found, as only Olaf Arngart seems to have realized, in Bede's chronological summary in Book V, Chapter 24. After so many years of confusion, the time seems propitious to put the corresponding passages of the memoranda (fol. 128v) and the chronological summary (fols. 126 and 127) side by side:11
(1) “Anno dxluii. Ida regnare coepit a quo regalis nordan hymbrorum prosapia originem tenet . et xii annis in regno permansil“12 (fols. 128v5-6 and 126v9-10: “In the year 547 Ida began to reign, from whom the Northumbrian royal family trace their origin, and he lasted in the reign for twelve years”). The accompanying regnal list, from Glappa to Ceoluulf (fol. 128v6-9), is clearly the stuff that marginalia are made of:
Post hunc
glappa .i. annum.
adda uiii.
aedilric. iiii.
theodric. uii.
friduuald. ui.
hussa.uii.
aedilfrid. xxiii.
aeduini. xuii:
osuald. uiiii.
osuiu. xxuiii.
ecgfrid xv.
aldfrid xx.
osred. xi.
coinred. ii
osric. xi.
ceoluulf. uiii.
The reigns of the historical kings, from Edwin (aeduini) to Osric, are all except the two-year reign of Cenred (coinred) alluded to by Bede in his chronological summary (fols. 126v24-27, 126v28, 127r5, 127r7-16, 127r23, 127r25 and 127r30). Like the chronological summary, moreover, the regnal list also deliberately omits the names of the apostate kings, Osric and Eanfrid, between Edwin and Oswald, and incorporates the lengths of their reigns into Oswald's.13 Hence the successive reigns from Ida to the abdication of Ceolwulf add up to the year 737.
(2) “baptizauit paulinus . ante annos cxi” (“Paulinus baptized 111 years ago,” fol. 128v9); “Anno dcxxui eanfled filia aeduini regis baptizata cum xii. In sabbato pentecostes” (“In the year 626 Eanflaed, daughter of King Edwin, was baptized with twelve others on the eve of Pentecost,” fol. 126v24-25). 111 + 626 = 737.
(3) “eclypsis ante annos lxxiii” (“Eclipse 73 years ago,” fol. 128v9-10); “Anno dclxiiii eclypsis facta” (“In the year 664 there was an eclipse,” fol. 127r1-2). 73 + 664 = 737.
(4) “penda moritur ante annos lxxix” (“Penda died 79 years ago,” fol. 128v10); “Anno dclu penda periit et mercii sunt facti christiani” (“In the year 655 Penda perished and the Mercians became Christians,” fol. 127r1). 79 + 655 = 734.
(5) “pugna ecgfridi ante annos lxiii” (“Ecgfrith’s battle 63 years ago.” fol. 128v10); “Anno dclxxxu ecgfrid rex nordanhymbrorum occisus est” (“In the year 685 Ecgfrith, king of the Northumbrians, was killed,” fol. 127r16). 63 + 685 = 748.
(6) “ælfuini ante annos luiii” ("Ælfwine 58 years ago," fol. 128v10-11); “Anno dclxxviiii ælfuini occisus“ (“In the year 679 Ælfwine [was] killed,” fol. 127r12-13). 58 + 679 = 737.
(7) “monasterium aet uiurae moda ante annos lxiiii” (“Monastery at Wearmouth 64 years ago,” fol. 128v11); the only place the name “Wearmouth” occurs in the Ecclesiastical History is in the chronological summary on fol. 127v6, “monasterii . . . ad uiuraemuda et Ingyruum” (“of the monastery at Wearmouth and Jarrow”). 64 + 674 = 738.14
(8 + 9) “cometae uisae ante annos uiii . eodem anno pater ecgberct transiuit ad christum” (“Comets seen 8 years ago, in the same year father Egbert passed on to Christ,” fol. 128v11-12); “Anno dccxxuiiii cometae apparuerunt sanctus ecgberct transiit” (“In the year 729 comets appeared [and] St. Egbert passed on,” fol. 127r29). 8 + 729 = 737.
(10) “angli in britanniam ante annos ccxcii” (“English in Britain 292 years ago,” fol. 128v12); “Anno ccccxluiiii . . . angli a brettonibus arcersiti [sic] Brittaniam adierunt” (“In the year 449 . . . the English came to Britain on the invitation of the Britons,” fols. 126v-6). 292 + 449 = 741.
The compiler of the Moore memoranda made knowledgeable and intelligent use of this nearby source. He was self-confident enough to paraphrase the conversion of Northumbria, for example, by shorthand reference to Paulinus, and he knew when Wearmouth was founded, perhaps because he knew Bede’s Historia Abbatum, perhaps because he was in some way connected to the monastery. But why would the compiler date a manuscript in this way? After all, a single entry would do the trick while eliminating the risk of miscalculations. There seems to be room, in other words, for an alternate theory to explain the retrospective dating on the early Bede manuscripts that use it. One possibility is that the dates, made and maintained by consecutive scribes from copy to copy, were meant to keep track of a manuscript tradition, rather than repeatedly and needlessly to date and redate a single manuscript by a single scribe.
If the hypothesis is correct, the Moore memoranda tell us that the manuscript tradition of this copy began in 734, the year before Bede died. Six more copies based on the same exemplar were made in 737, one of which marked the abdication of Ceolwulf by means of a regnal list. The next tallied copy, the one that records the founding of Wearmouth, was made in 738, during the reign of Ceolwulf's successor, Eadberht. Two additional copies were made in 741 and 748, still in the reign of Eadberht.15 But the memoranda taken together seem to indicate, as well, that the last copy, the Moore manuscript itself, was copied some indefinite time after 748 by a scribe who no longer understood the system of tallying copies in the margins. The Moore scribe apparently collected the marginalia in his exemplar and deposited them all in a heap at the end of his manuscript.
Thus, instead of inserting Cædmon's Hymn as a marginal note on the same page as Bede’s Latin paraphrase (fol. 91r), as the other Anglo-Saxon manuscripts of the Latin text do, the Moore scribe makes an endnote of our first English poem.16 He also apparently gathered together and alphabetized three rather specialized glosses from the margins of his copy-text: arula hearth, “the small altar on which Ræwald sacrificed victims to devils” (Book II, Chapter 15); destina feur-stud, “a church buttress made fire-proof by contact with Bishop Aidan” (six references in Book III, Chapter 17); and iugulum sticung, "pig-sticking," used in Bede only as the verb iugulare to describe how the Britons were “shughtered” and “butchered” (Book I, Chapters 13 and 15).17 And, of course, he combined the marginal chronological notes into what scholars have been calling the Moore memoranda.
The unchronological order of these notes in the memoranda implies, however, that the scribe gave us only a random selection of the original entries from his exemplar. For example, item 10, the arrival of the English in Britain, which occurs first of all the memoranda on fol. 126v4-6 of Bede’s chronological notes, occurs dead last in the memoranda (for. 128v12). As a result, the scribe has left us no way of knowing if 748 was the last date recorded in the margins of his exemplar. How do we date Early Northumbrian, then, from this manuscript evidence? Bernhard Bischoff has shown that the Moore manuscript was part of the library of Charlemagne’s Palace School in Aachen around the year 800.18 It is possible, therefore, that it was copied at the Palace School as late as the early ninth century by an expatriate Northumbrian scribe, with old ties to Wearmouth.19 If so, it is also possible that he wrote Cædmon’s Hymn in the same early ninth-century dialect he spoke every day to his fellow expatriates in Aachen, so that they might easily read and understand it. To be sure, if we go only by the most probable time period of his manuscript, the date of his “early” Northumbrian is a good deal later than 737.20
As this case helps illustrate, scholars have necessarily established Anglo-Saxon studies from inconsistent, anachronistic, and otherwise inadequate manuscript evidence. Working from the statistical base of the most consistent bits and pieces that have come down to us, our predecessors have confidently forged powerful, compelling theories recovering our lost past, letting us hear, for instance, the uncontracted syllables in the early Anglian dialects before the age of manuscripts and of Bede or the unsmoothed diphthongs in the reconstructed early West Saxon language of King Alfred. We should keep in mind, however, that these theories all depend on “normalizing” manuscript evidence, concealing inconsistencies, reconstructing early phases, recreating perfect sources. Today, if we are willing to know less about what is not preserved in our manuscripts, we can begin systematically discovering all of the different things that are in fact still in them.
Lexington, Kentucky
|
Notes
1 Although elsewhere in his Grammar acknowledging that “[t]he special problems of runic epigraphy are not in place in a grammar” (p. 27), Alistair Campbell uses the runic inscriptions on the Ruthwell Cross to exemplify the Northumbrian dialect (p. 4). However, he dates the linguistic forms of the inscription “slightly later” than the two early Cædmon manuscripts, which he says “belong to the first half of the eighth century.” Relying on this dating, he concludes that the runic inscriptions can be “assigned to the eighth century without hazard” (p. 4 and n 2). Old English Grammar (Oxford: Clarendon P, 1968).
2 The official names are Cambridge University Library MS Kk.5.16 for the Moore manuscript (M) and Saltykov-Schedrin Public Library MS Q.v. I.18 for the Leningrad manuscript (L). I would like to thank the officials of Cambridge University Library for letting me study the Moore manuscript during the summer of 1989.
3 (New York: Appleton-Century-Crofts, 1968), p. 36.
4 Barbara Strang would further enhance the image of these continental vestiges of Early Northumbrian. “Copied just after this period,” she says, “but perfectly [sic] representing the forms of the earliest Northumbrian, are two other poems [in addition to Cædmon’s Hymn], Bede’s Death Song and a riddle." A History of English (rpt. 1986; New York: Methuen, 1970), p. 362. Scholars are usually reluctant to accord general accuracy, much less perfection, to scribes of Old English texts.
5 “Deinde, eadem manu qua scribitur Codex, habetur series Regum Nordan-humbrorum, usque ad Resignationem Coelwlfi [sic], qui regnum suum Eadberhte cesserit. A.D. 737. Item & alice notuca Chronologica qua in eudem annum concurrunt.” Antique Literatura Septentrionalis Liber Alter. seu Humphredi Wanleii (1705), reprinted in English Linguistics: 1500-1800 (Menston, England: Scolar Press, 1970), pp. 287-288.
6 The Oldest English Texts, EETS, o.s. 83 (rpt. 1966; Oxford: Oxford UP, 1885), p. 148.
7 (3NT, pp. 20-210. For the origin of this reasoning, see J. Zupitza, “Über den Hymnus Cädmons,” Zeitschrift für deutsches Altertum und deutsche Literatur 22 (1878), 210-223. Although he was the first scholar to show that “[g]anz unbestreitbar ist allerdings das jahr 737 nicht,” Zupitza concluded that the other dates must be errors in calculation (p. 215).
8 The Moore Bede, Cambridge University Library MS Kk.5.16, EEMF 9 (Copenhagen: Rosenkilde and Bagger, 1959), p. 27. Hereafter MB.
9 Olga Dobiache-Rojdestvensky first brought this manuscript to the attention of Anglo-Saxonists in “Un Manuscrit de Bède à Leningrad,” Speculum 3 (1928), 314. Since then, Olaf Arngart has shown that BL Cotton MS Tiberius A. xiv not only follows the same system of dating as L, but also repeatedly derives the same date, 746, leading him to the conclusion that both manuscripts are transcripts of a copy originating in 746. “On the Dating of Early Bede Manuscripts,” Studia Neophilologica 45 (973), 51-52. Hereafter “Dating.”
10 There is also a marginal entry yielding a date of 861, a long way indeed from 746. Assuming a scribal blunder, Arngart argues that it “is clearly due to the inadvertent addition of the figures CXV at the end of the marginal numeral, for if these three figures are struck out, we get the sum DCC + XLVI, that is gain 746, as in nearly all the other extant entries of the summary” (“Dating” 49). If one is unpersuaded by an “inadvertent addition of the figures CXV,” 115 years to an otherwise simple addition of 700 + 46, another possibility is that all of the number were copied in 861, instead of 746. Arngart would agree, at any rate, that the numbers themselves are copies and as such do not help us date the manuscript (see “Dating” 52).
11 A measure of the general confusion can be seen in Hunter Blair’s comment that the memoranda “amounts only to eight lines, but apart from its bearing on the date of the Moore MS., it has an importance out of all proportion to its length, both as a source of information for which there is now no earlier record and as an example of primitive Northumbrian historiography.” “The Moore Memoranda on Northumbrian History” (hereafter MM), item VI in Anglo-Saxon Northumbria, ed. M. Lapidge and P. Hunter Blair (London: Variorum Reprints, 1984). p. 245. In fact, apart from the dubious section of the regnal list, from Glappa to Hussa, the memoranda merely repeat information Bede provides on the immediately preceding folios.
12 The Latin transcriptions are taken directly from the manuscript, following scribal punctuation and capitalization, but expanding abbreviations (q[u]o, origine[m], and an[nis] occur only in the memoranda). As Hunter Blair points out, the d and x in dxluii “are so close together as to suggest that the scribe originally wrote c and later corrected this to x in fainter ink” (MM 246). Not recognizing the immediate source of the memoranda, however, he failed to notice that the same kind of x occurs in the roman numeral on fol. 126v9.
13 Hunter Blair notes that “[t]he omission of the names of the two apostates from the official lists is strongly reminiscent of the Roman custom of damnatio memoriae” (MM 248-249).
14 Bede’s date is “anno ab incarnatione Domini sexcentesimo septuagesimo quarto, indictione secunda, anno autem quarto imperii Ecgfridi regis” (“in the year of Our Lord 674, in the second indiction and fourth year of the reign of king Ecgfrith”) Historia Abbatum auctore Baeda, Chapter 4 (p. 368), Baedae Opera Historica, Vol. 1, ed. Charles Plummar (rpt. 1946; Oxford: Clarendon P, 1896). Molly Miller follows Hunter Blair's practice of dating a priori when she says, “since the base-date is 737, this means a foundation [of Wearmouth] in 673.” “The Dates of Deira” ASE 8 (1979), 57.
15 The same explanation accounts for the coincidence of dating in 746 for the Leningrad and Tiberius manuscripts of Bede’s Ecclesiastical History. Indeed, we might well expect to find multiple copies belonging to the same manuscript tradition, descending from a copy originating in 746. 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), p. 15; cited later as “Scriptorium.”
16 He provides an appropriate bibliographical reference when he says, “primo cantauit Caedmon istud Carmen” (“Cædmon first sang that Song”). I argue that all our Old English versions descend from Bede’s Latin paraphrase in “Reading Cædmon’s Hymn with Someone Else’s Glosses,” forthcoming in Representations.
17 Bede's Ecclesiastical History of the English People, ed. Bertram Colgrave and R.A.B. Mynors (Oxford: Clarendon Press, 1969). I follow Hunter Blair's suggestion that Iugulum was listed for the finite verbs (MB, p. 13, note 14). E.V.K. Dobbie first noted that destina occurs 6 times in III. 17. The Manuscripts of Cædmon's Hymn and Bede's Death Song (New York: Columbia UP. 1937), p. 11. All of the occurrences of destina refer to the same buttress Aidan leaned on, rendering it fire-proof.
18 “Die Hofbibliothek Karls des Grossen,” Das Geistige Leben, ed. Bernhard Bischoff (Düsseldorf: L. Schwann, 1965), p. 56.
19 Parkes has observed that “[t]he features of the handwriting of Moore appropriate to an early date may equally be explained as representing the work of a somewhat later scribe less advanced than those at Wearmouth-Jarrow.” He speculates that Alcuin may have had the manuscript sent over from York, presumably after he was settled in Aachen in 794 (“Scriptorium” 27, note 35). Another possibility, recently proposed by Peter Godman, is that Alcuin brought the manuscript to the Franklish court in 781 or 782. Alcuin: The Bishops, Kings, and Saints of York (Oxford: Clarendon P, 1982), p. 131 note.
20 Although holding to the early eighth-century dating for Early Northumbrian, Arngart appears appropriately nervous after showing that “the marginal numerals are not to be reckoned with for the dating” of the Leningrad Bede; he ends saying, “Still, of course, the dated MSS, of Bede have been the fixed point in deciding the relative chronology of early Old English sound change....” (“Dating,” p. 52). The elipsis is Arngart's.
|