§ A.1 The two earliest witnesses to Bede’s Historia ecclesiastica and Cædmon’s Hymn are traditionally dated to 734 × 737 (M) and 731 (or 732)[1] × 746 (P). In both cases, this unusual precision is due to the presence of a series of chronological notes in the manuscripts, derived from dates given in Bede’s recapitulo, V. 24. These notes, the so-called Memoranda, provide an alternative to Bede’s usual practice of dating historical events from the year of the incarnation by providing instead a retrospective count of the number of years that have elapsed since the event in question occurred. Thus Caesar’s attempted invasion of Britain (dated to 60 B.C. in the recapitulo) is also described in P as having occurred ante annos DCCCVI, “806 years ago”; the baptism of Eanflæd by Paulinus (A.D. 626 in the recapitulo) is described in M as having occurred ante annos cxi, “111 years ago.”[2]
§ A.2 As numerous scholars have pointed out, this retrospective dating can be used to work out the precise year in which the Memoranda were calculated: Caesar’s attempted invasion was “806 years ago” only to somebody writing in 746; the baptism of Eanflæd was “111 years ago” only to somebody working in 737. Although the dates implied by each manuscript’s Memoranda show some internal variation (those in M range anywhere from 734 to 748, those in P from 746 to 861), both sets show a clear preference for one particular year: 737 in the case of M (five entries out of nine)[3] and 746 in that of P (thirteen out of fifteen legible entries). Since the calculations in P and M appear to have been added as or soon after the manuscripts were copied, these dates traditionally have been understood as recording either the actual year in which the surviving manuscripts were copied, or, particularly in the case of P where the Memoranda were added in a distinct hand, a terminus ante quem for the main text of the manuscript.[4]
§ A.3 This traditional view of the significance of the Memoranda as a guide to the date at which P and M were copied depends crucially on the assumption that the retrospective dates were recalculated each time the text was copied. That such recalculation sometimes did occur is suggested by the fact that M and P use different years as their consensus points of reference. That it did not always happen, however, is suggested first of all by the agreement of P and a third early Bede manuscript, London, British Library, Cotton Tiberius A. xiv (Tib. A. xiv) in one serious chronological error (see Arngart 1973, 49), and, secondly, by the fact that the Memoranda in M are found on the last page of the manuscript rather than beside the events described in Bede’s recapitulo. In the case of P and Tib. A. xiv, the shared error suggests either that the two manuscripts got their calculations from a common source, or that one of the two is a linear descendent of the other.[5] In the case of M, the fact that the Memoranda are found on the last page of the manuscript alongside glosses and other material (including Cædmon’s Hymn) more usually associated with the margins of the main text, suggests that the scribe responsible either extracted them from their proper place in his main text’s exemplar or collected them from the margins of some other manuscript of the Historia ecclesiastica (Kiernan 1990b, 50 and 52-53 [but cf. Kiernan 1990a, 162 n. 16]; Fulk 1992, 426). That this second possibility is the more likely is suggested by the glosses’ location and the evidence of witnesses to other recensions of the vernacular poem: a scribe copying marginalia from the same exemplar he used for his main text would presumably have preferred to copy them in situ; a scribe working in haste from different manuscripts of the Historia ecclesiastica, however, might be tempted to collect marginalia in a single place rather than collate his main text with the annotated copy. That scribes often added the vernacular Hymn from sources other than those from which they acquired their main text is suggested by the presence of a copy of the West-Saxon eorðan recension (a recension found otherwise only in manuscripts of the Old English translation of the Historia ecclesiastica) in To, a manuscript of the Latin Historia ecclesiastica, and the frequent appearance of related texts of the Hymn in apparently unrelated copies of the Historia ecclesiastica (see above, § 5.6). No matter how the text of the Memoranda and Hymn arrived in M, the procedure used seems unlikely to encourage recomputation.
§ A.4 Apart from some loss of precision, the possibility that the Memoranda were copied rather than recalculated has little significant effect on our understanding of these witnesses’ dating. Both manuscripts can be assigned with relative certainty to the eighth century on palaeographic, and, in the case of M, which appears to have been at the library of Charlemagne’s palace school in Aachen by ca. 800 (Bischoff 1966-1968, 56-57), external grounds (see, among others, Hunter Blair 1959 [for M] and Parkes 1982, Wright 1961, and Lowe 1958a [for P]). If the glosses have indeed been copied unchanged from earlier exemplars, then they simply serve to set a terminus post quem for the work of their respective scribes rather than an indication of the precise year in which the manuscripts were copied. In the case of M, where both the glosses and the main Latin text of the Historia are in the hand of what looks very much like a single scribe (see above, § 4.30 and n.), this terminus presumably (though not necessarily) applies to the date of the codex as a whole, suggesting that the entire manuscript was copied sometime between 737 and ca. 800 (see Hunter Blair 1959, 27). In the case of P, where the Memoranda are in a distinct hand, the terminus applies in the strictest sense only to the Memoranda and other marginal material copied by the same scribe. While the main text of the manuscript might well have been copied at approximately the same time as the chapter headings and Memoranda, it is not necessary to assume that it was. The scribe responsible for the Memoranda might equally well have added his material to a manuscript that had been copied some years before.
§ A.5 There is, moreover, some evidence to suggest that this was so. For while it is an impressive book, the St. Petersburg Bede is not a very consistent one (see Parkes 1982, 6-12). The main text of P has been copied in four quite distinct hands. The last two scribes of the manuscript (Scribes C and D) copy their text at a steady 27 lines per page on thick parchment. Scribe C copies the shortest portion of the surviving text, beginning on f. 64r [recte 65] and ending mid-sentence at the bottom of f. 68r [recte 69r]; Scribe D begins at the top of f. 68v [recte 69v] with the expected next word of the sentence, and continues until the end of the manuscript on f. 161r [recte 162]. In contrast, the work of Scribes A (ff. 1-32v) and B (ff. 33-63v [recte 64v]) in the manuscript’s first eight quires is far less consistent. The number of lines per page ranges from 27 to 33 and there is a high proportion of overruns at the ends of columns (see particularly Parkes 1982, 6 and n. 38). The parchment used is also notably thinner. As Parkes argues, this suggests that the work of Scribes C and D represents an earlier stage in the manuscript’s production and that Scribes A and B were attempting to supply “eight quires to the beginning of a pre-existing manuscript (up to and including Book III, cap. xix)” (Parkes 1982 6). Since the material added by the scribe of the Memoranda appears throughout the stints of all four of the main text scribes, it is clear that this fifth hand did not set to work until after Scribes A and B were finished.
§ A.6 How long the gap was between the work of Scribes C and D, A and B, and that responsible for the Memoranda is a matter for speculation. On the one hand, as numerous palaeographers have noted, the hand of Scribe D is quite different from that of Scribes A, B, or C. Where A, B, and C use a fairly uniform pointed minuscule, that of Scribe D is rounder and differs in the formation of several letters and abbreviations, most notably the lower left limb of the letter <x>, which curls outwards in the work of Scribe D, but inwards in that of A, B, C, and the scribe responsible for the chapter numbers and Memoranda (see particularly Lowe 1958a, 188). Lowe, who believed that the date implied by the Memoranda represented the date at which the manuscript was copied, nevertheless felt that hand D had “all the appearances of being saec. viii1” while the “general impression made by the script of the first three hands (ff. 1-68r [recte 69r]), one must admit, is saec. viii2, if not saec. viii ex.” On the other hand, however, it is clear that Scribes C and D were contemporaries: Scribe D takes over from C at the top of a verso and in the middle of a sentence. Given the strength of the similarities between A, B, and C, but the evidence that C and D were working at the same time, it is perhaps best to follow Lowe and Parkes in assuming that the difference in style between C and D represents a change either in generation or school, with Scribe C representing the newer style of minuscule in use at Jarrow, and Scribe D the training of an earlier generation or a different centre. If this is indeed the case, then the manuscript can perhaps be dated, independently of the Memoranda, to s. viiimed—not early enough to allow us to accept the otherwise intriguing suggestion that Bede himself was the manuscript’s fourth scribe (e.g. Lowe 1958b and Parkes 1982, 5-6 and 7-11, esp. n. 45), but still early enough to place it within a quarter century of his death.
[1]Bede records that he finished the Historia ecclesiastica in 731 (V. 23): Hic est inpraesentiarum uniuersae status Brittaniae, anno aduentus Anglorum in Brittaniam circiter ducentesimo octogesimo quinto, dominicae autem incarnationis anno DCCXXXI, “This is the state of the whole of Britain at the present time, about 285 years after the coming of the English to Britain, in the year of our Lord 731” (ed. and trans. Colgrave and Mynors 1969). Earlier in the same chapter, however, he refers to the defeat of the Moslems in Gaul, an apparent reference to the battle of Tours in October 732: Quo tempore grauissima Sarracenorum lues Gallias misera caede uastabat, et ipsi non multo post in eadem prouincia dignas suae perfidiae poenas luebant, “At this time a terrible plague of Saracens ravaged Gaul with cruel bloodshed and not long afterwards they received the due reward of their treachery in the same kingdom.”
[2]Editions of the Memoranda can be found in Kiernan 1990b, 50-52 (M), Hunter Blair 1984, 246 (M), and Arngart 1973, 48 (P and Tib. A. xiv). The Memoranda in Tib. A. xiv are discussed below.
[3]The edition of the Memoranda in Kiernan 1990b lists ten items; items “eight” and “nine”, however, refer to closely associated events (the appearance of comets in the year in which Saint Egbert died) and share the same retrospective date: 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” (f. 128v, manuscript lines 11-12; as ed. and trans. in Kiernan 1990b, 52).
[4]Kiernan proposes an alternative to the traditional view that the retrospective dates in the Memoranda were meant to indicate the year in which they were compiled. Rather than mistakes, Kiernan suggests that the “discrepancies” among the retrospective dates actually reflect different years of calculation and present a record of the years in which each of the ancestor copies in each tradition were made. Thus in the case of M, the “memoranda tell us that the manuscript tradition of this copy began in 734, the year before Bede died,” that “six more copies based on the same exemplar were made in 737,” that an eighth “was made in 738,” and that “two additional copies were made in 741 and 748” (Kiernan 1990b, 52).
There are at least three problems with this argument, apart from the fact that Kiernan cites no parallels from any other text:
The first involves the dating of the first manuscript in each tradition. If the dates “were meant to keep track of a manuscript tradition” (52), why do they not begin with the first copies produced after Bede finished the Historia ecclesiastica in 731 or 732? The tradition recorded in M begins ostensibly with a copy made in 734; the tradition recorded in P and Tib. A. xiv with a manuscript from 746. These manuscripts must have been descended from earlier copies. Why was this fact not recorded?
The second problem involves the assumption that the Memoranda were “made and maintained by consecutive scribes from copy to copy” (52). This seems to require us to assume that the manuscripts involved were copied sequentially rather than in parallel: i.e. that the manuscript copied in 748 in the M tradition has a record for the manuscript copied in 741 because the manuscript copied in 741 was its immediate ancestor. While this is not unreasonable perhaps when we are dealing with one or two manuscripts in a single year, it becomes increasingly difficult to maintain in the face of years such as 737 (in which Kiernan’s system requires us to assume that five successive generations of manuscripts were produced in a single year in the M tradition) or 746 (where the Memoranda in P suggest thirteen generations were produced at Wearmouth-Jarrow). If any of these successive generations produced multiple offspring (a fact that would not be recorded in the surviving Memoranda), the number of manuscripts we are required to assume these scriptoria produced increases rapidly. Assuming that the tally was kept in the exemplar rather than the copy and only accidentally found its way into M only partially solves this problem: we are still left with a very uneven production schedule, and need to explain why scribes would want to keep track of the daughters of any given copy of the Historia ecclesiastica.
Finally, there is the question of the production schedule these Memoranda imply. While we might well expect to see certain peaks in activity in response to requests for copies of Bede from other libraries or the continental missions, it is surely odd that these peaks are restricted to a different single year in each tradition. Why would at least five generations of manuscripts be copied at the scriptorium that produced M in 737, but only one in 738 and none in 739? And why did the scribes of the M tradition produce no copies at all in the year in which those responsible for P and Tib. A. xiv produced thirteen generations?
[5]For opposing views on the two possibilities, see Arngart 1973, and Parkes 1987, 26-27; the debate as a whole is discussed in Fulk 1992, 426.
As Arngart notes, the error in question, involving the dating of Claudius’s invasion of Britain, dated by Bede to A.D. 46 and retrospectively dated in P and Tib. A. xiv to [ante annos] DCCCXV, “815 years ago” (implying a gloss written in 861), cannot be explained by an independent graphic error: the correct figure for somebody writing in 746 should be DCC (Arngart 1973, 49).