¹ Exeter Cathedral, Dean and Chapter Library, 3501, fols. 8–130. (Fols. 0 and 1–7 were originally attached to another of the manuscripts given, like this one, to the cathedral by Bishop Leofric in the second half of the eleventh century.) See Ker, N. R., Catalogue of Manuscripts containing Anglo-Saxon (Oxford, 1957), p. 153Google Scholar, and the revised collation in his Medieval Manuscripts in British Libraries 11 (Oxford, 1977), 807–8Google Scholar. For the nature and extent of probable losses in the volume, see Pope, J. C., ‘Palaeography and Poetry: some Solved and Unsolved Problems of the Exeter Book’, Medieval Scribes, Manuscripts and Libraries: Essays presented to N. R. Ker, ed. Parkes, M. B. and Watson, Andrew G. (London, 1978), pp. 25–65Google Scholar.

² See below, p. 140.

³ Anglo-Saxon Poetic Records 3, ed. G. P. Krapp and E. V. K. Dobbie (New York, 1936), 3–4. Referred to hereafter as the Krapp–Dobbie text.

⁴ The Advent Lyrics of the Exeter Book, ed. Jackson J. Campbell (Princeton, 1959), p. 49Google Scholar, and Burlin, Robert B., The Old English Advent: a Typological Commentary (New Haven and London, 1968), p. 68Google Scholar.

⁵ See below, p. 145, for Das's article and the independent discovery of the error by Dr N. R. Ker, who first brought it to my attention and is thus chiefly responsible for the present article.

⁶ It was wisely chosen as the basic authority for Jr's, J. B. BessingerConcordance to ‘The Anglo-Saxon Poetic Records’ (Ithaca, New York, 1978Google Scholar). The publication of the concordance makes it especially desirable to correct the readings of the passage under discussion, since it is one of the very few passages where the readings of the ASPR are inaccurate.

⁷ Dobbie, who was responsible for the final form of this passage, called attention at the foot of the page to two emendations, þu for þa in 18a, we for þe in 31a. I have italicized the altered letters. Dobbie also noted that the final letter of forwyrneð in 20b was either d or ð explaining (p. 247) that the form is that of ð but the cross-stroke is uncertain. The bracketed dots indicate a missing letter in 20a and illegible letters elsewhere. Otherwise (aside from the modern punctuation, the metrical lineation, the letter forms and expanded abbreviations – here limited to ond for MS 7) the readings purport to be those of the manuscript. On the 0 of mundboran, in 28b, missing but not doubtful, see below, p. 154.

⁸ The Christ of Cynewulf, ed. Albert S. Cook (Boston, 1900Google Scholar; repr. from the corrected impression of 1909, with a prefatory note by J. C. Pope, Archon Books, Hamden, Conn., 1964), pp. xxxv ff. and 71–2. The antiphonal source of 416–25 (O admirabile commercium) was identified by Moore, Samuel, ‘The Source of Christ 416 ff.’, MLN 29 (1914), 226–7Google Scholar, and recently a partial source for 164–213 has been discovered in an antiphon on Joseph and Mary recorded by Alcuin. See Hill, Thomas D., ‘A Liturgical Source for Christ I 164–213 (Advent Lyric vii)’, MÆ 46 (1977), 12–15Google Scholar.

⁹ ‘O Key of David and Sceptre of the house of Israel, who openest and none shutteth, who shuttest and none openeth: come Thou, and bring forth the captive from the house of bondage, who sitteth in darkness and in the shadow of death.’ (Cardinal Newman's translation, quoted by Cook, , Christ, p. 71Google Scholar.) On the biblical sources of the antiphon see Cook, Ibid. p. 76.

¹⁰ On locan healdeð ‘guards the locks’, see Greenfield, S. B., ‘Of Locks and Keys – Line 19a of the Old English Christ’ MLN 67 (1952), 238–40Google Scholar. Toller, in his Supplement to the Bosworth–Toller Anglo-Saxon Dictionary, distinguishes the weak masculine loca, ‘a locked place’ from the neuter loc, ‘a lock’. If he is right we should take locan as accusative singular, ‘locked enclosure’, referring either to hell or to this earthly prison; but the distinction may not be absolute and a general plural ‘locks’ seems more natural in the context.

¹¹ See below, pp. 143–55, and the translation, p. 156. The implications of the imagery, briefly treated here, are more fully explored by Burlin, , The Old English Advent, pp. 70–7Google Scholar.

¹² See below, pp. 141–2.

¹³ See the Preface to his edition, Fenix-Fuglen, it Angelsachsisk Kvad (Copenhagen, 1840), p. 13Google Scholar, where he speaks of his three visits to England in successive years, 1829–31, during the second of which he transcribed the book.

¹⁴ This brief extract and Grundtvig's own text of The Phoenix in Fenix-Fuglen are the only printed representatives of the transcript. I do not know whether or not it has been preserved among Grundtvig's papers.

¹⁵ Fenix-Fuglen, p. 13, where Grundtvig also claims credit for having suggested, in 1830, that the Museum should have its own copy of so important a book. The name of the otherwise unknown copyist, Robert Chambers, is entered as part of the record of acquisition of the transcript under the year 1831 in the Register of Acquisitions, a series of manuscript volumes in the archives of the British Library's Department of Manuscripts. I owe this information to the kindness of Miss M. A. F. Borrie of the Department.

¹⁶ As attested by his signed statement, dated ‘Febr. 24th, 1832’, on the second flyleaf of the transcript.

¹⁷ Codex Exoniensis, ed. Benjamin Thorpe (London, 1842Google Scholar). Thorpe's account of his transcription is on p. xi. His edition omitted a few of the riddles.

¹⁸ Three important collations of the Exeter Book were made during the last three decades of the nineteenth century: the first, of the entire book, by Schipper, Julius, who published the results, ‘Zum Codex Exoniensis’, Germania 19 (1874), 327–38Google Scholar; the second, of the long poems, 8r–76r, by SirGollancz, Israel for his two editions, Cynewulf's Christ (London, 1892Google Scholar) and The Exeter Book, pt 1, EETS o.s. 104 (London, 1895Google Scholar); the third, of the entire book, by Bruno Assmann for his edition of the Grein-Wülker Bibliothek der angelsächsischen Poesie 111 (Leipzig, 1898Google Scholar). (C. W. M. Grein's text of Christ in the 1st ed. of the Bibliothek 1 (Göttingen, 1857), 149 ff.Google Scholar, had relied on Thorpe's readings.) A few small observations by these scholars concerning Christ 18–32 are mentioned below in the course of the discussion, pp. 143–8. Cook, the most thorough of the editors of Christ in other respects, did not consult the manuscript.

¹⁹ The Exeter Book of Old English Poetry, with introductory chapters by R. W. Chambers, Max Förster and Robin Flower (London, 1933Google Scholar). Referred to hereafter as Exeter Book, facs.

²⁰ Ibid. pp. 69–82; the transcription of 8r9–23 is on pp. 69–70.

²¹ Some ultra-violet photographs of the page recently made for me by the photographer of the University of Exeter by kind permission of Mrs Audrey Erskine, archivist of the Cathedral Library, are much inferior, largely because of the poor conditions under which they had to be taken but partly because the hole at the end of line 12 has once more been covered and transparent strips have been glued over some of the darkest parts of the surface.

²² I owe these plates to the generosity of Dr N. R. Ker, who possesses the only copy of the photograph, among those made for and distributed by Chambers, that I have been able to trace. Dr Ker had his copy photographed at the Bodleian Library in May 1979, allowing me to have the negative and two prints. He has retained one print of the new photograph along with his original copy. The plates are derived from the negative.

²³ In the same note Chambers identified some letters in a damaged line of ior on the basis of a second ultra-violet photograph which I have been unable to trace. The damage in this case (caused by seepage of the same liquid that was spilled on 8r) has rendered portions of lines 153–4 of the poem illegible up to the present time. Both the manuscript itself and the facsimile have so far resisted efforts to distinguish enough vestiges of letters to support reasonable conjectures.

²⁴ MÆ. 2 (1933). 224–31Google Scholar Ker's new readings are listed on pp. 225–6. It is regrettable that the Krapp–Dobbie edition did not accept his correction of firina to firena (Christ 56, MS 8vi3–14), for he was surely right in pointing out, from the facsimile, that what looks like a second i is so spaced as to indicate an imperfectly preserved e. A spelling firina would have been eccentric both in this manuscript and elsewhere. Burlin, , The Old English Advent, p. 80Google Scholar, rightly prints firena.

²⁵ I examined the passage in the manuscript on 30 April 1979 and first saw the Ker copy of the photograph a few days later. Before that time I had studied, besides the facsimile, the incomplete reproduction of the photograph published by Smith, A. H., ‘The Photography of Manuscripts’, London Medieval Studies 1 (1938), pl. ix opp. p. 201Google Scholar. This is helpful for some of the difficult readings but shows only part of the page: lines 9–18 of the manuscript minus about one sixth of the writing at the right-hand end of each line (thus showing portions of sixteen verse lines from under, 14, to þat t-, 29). Efforts on the part of friends to trace the negative or copies once in the hands of Chambers, Smith and Dobbie have proved unavailing.

²⁶ I accept with some hesitation the traditional view that the scribe carelessly wrote þa for the obviously intended þu. Assmann, in his notes to the Grein–Wülker text (see above, n. 18), suggested that he had actually written þu. This seems possible, because the curved stain of the mug, crossing the letter, may have caused an original u to look like a; but when I examined the manuscript I concluded that a is somewhat more probable. That the scribe could be careless is shown by two slips on the same page: heafoð for beafod, line 2 of the manuscript, and crœstga for crœftga, line 7.

²⁷ Chambers, R. W. and Flower, , Exeter Book, facs., p. 69Google Scholar, indicate doubt by underscoring and bracketing ri.

²⁸ Ibid.

²⁹ See above, n. 18.

³⁰ ASPR 3, 247. As already suggested, it was presumably Dobbie who, after Krapp's death and the receipt of a copy of the ultra-violet photograph from Chambers, took full responsibility for this passage.

³¹ MLR 32 (1937), 79–80Google Scholar.

³² It is clear that he had not seen Ker's review of 1933 (see above, n. 24), in which upwegas and forwyrneð had already been noted. Neither had he seen the Krapp–Dobbie edition. Probably he had already returned to India when his article appeared. In the Preface (p. ix) to his Cynewulf and the Cynewulf Canon (Calcutta, 1942Google Scholar) Das explains that the thesis underlying it was worked out in 1925–7 at Dacca University under the supervision of C. L. Wrenn and revised for publication during the years 1934–6 while he was preparing an edition of Christ, Parts I and II, at University College, London, later approved for the Ph.D. degree at the University of London. This account helps to explain Das's unawareness of the Krapp–Dobbie edition even in the 1942 book. The main text of the latter depends on editions of the poem available in 1925–7 and is corrected only in the notes.

³³ See above, n. 10.

³⁴ When I saw his pencilled corrections of the transcription on that page in May 1979, I found that he had entered every one of the corrections and conjectural restorations, old and new, set forth in the present article, with the exception of on for to in 32a, and he agreed that on was to be preferred. He had not known of Das's article till I called his attention to it. So for Das's reading eadgu[ m] and also for ryht, m[ynd]giað and lœte [to l]ose his notes give independent confirmation.

³⁵ The final letter of lœmena, line 9 of the manuscript, 15a of the poem, as it stands in the editions, is a special case. It was tentatively taken as u by Thorpe. Actually, the scribe seems to have written 0 and corrected without erasure to u, thus producing a mixture of two forms that most readers mistook for a.

³⁶ In the prefatory note, p. [iv], to the reprint of Cook's Christ (see above, n. 8). The rest of the passage there quoted (18–26) follows the Krapp–Dobbie text, which corrects Cook's at many points but includes the errors already noted in 18b and 20a, of which I was then unaware.

³⁷ The Old English Advent, p. 68.

³⁸ See above, p. 140.

³⁹ Exeter Book, facs., p. 69.

⁴⁰ For examples, see Pope, J. C., The Rhythm of Beowulf (New Haven, 1942, 2nd ed. 1966), p. 333Google Scholar. Bliss, A. J., The Mitre of Beowulf (Oxford, 1958), p. 84Google Scholar, classifies as 2a3b.

⁴¹ Holthausen, F., ‘Zu alt- und mittelenglischen Texten’, Anglia Beiblatt 54 (1943), 31Google Scholar. The author refers to the facsimile but displays no knowledge of the Krapp-Dobbie text.

⁴² ASPR 3, 247. Dobbie's restoration was accepted by Schaar, Claes, Critical Studies in the Cynewulf Group (Lund and Copenhagen, 1949), p. 71Google Scholar (though he refers to Holthausen's m[ ynd] giað), and by Kuhn, Sherman M., ‘A Damaged Passage in the Exeter Book’, JEGP 50 (1951), 491–3Google Scholar.

⁴³ Burtin, , The Old English Advent, p. 69Google Scholar, avoids what may have seemed to him an impropriety by translating, ‘[we] are mindful of Him who created man’, but this hardly fits the normal sense of the verb or the context.

⁴⁴ Schipper, ‘Zum Codex Exoniensis’, p. 329, came close to recognizing the scribe's usual rounded d when he said, ‘der letzte der fehlenden Buchstaben scheint ein o zu sein’.

⁴⁵ Grundtvig, according to Müller, had hete..ofe, and Chambers apparently corrected ceofe to ceose.

⁴⁶ ASPR 3, 247. The suggested b misled Kuhn in his attempt at a reconstruction, ‘A Damaged Passage in the Exeter Book’.

⁴⁷ Exeter Book, facs., p. 70, where ce is underscored and bracketed.

⁴⁸ Critical Studies in the Cynewulf Group, p. 72.

⁴⁹ The o of to, though imperfect, can be made out fairly well. I can distinguish nothing where t ought to be and only a discontinuous vertical for l, without a finishing curve on the line. In fact a segment of the vertical appears to extend a little below the line. Some of what is visible may therefore be an irrelevant stain.

⁵⁰ The Bosworth–Toller Anglo-Saxon Dictionary, under los, gives examples from the Old English Bede, King Alfred's version of Gregory's Pastoral Care and the Rushworth Gospels.

⁵¹ Holthausen, F., ‘Zu Crist I, v. 32’ (‘Zur altenglischen Literatur, IV’, Anglia Beiblatt 18 (1907), 201–2Google Scholar). He properly dismisses two apparent exceptions in the poetry, Guthlac 549a No by him to deað and Juliana 86b þe to geweald, where editors have agreed that both grammar and metre call for emendation to dative deaðe and gewealde. A third example, however, Guthlac 1343b laðspel to soð, is metrically correct as it stands and is not emended by Krapp and Dobbie. Here we are not dealing with the usual idiom to soðe as Holthausen supposed but with to as the adverb ‘too’. The hateful news of Guthlac's death is too true. Dobbie mentions Holthausen's article in his note on 32 (ASPR 3, 247) without expressing an opinion.

⁵² See pl. V, line 20, and pl. VI, line 10, the penultimate word. A slightly darkened crack in the parchment extending leftwards from the top of what I take to be o may have been mistaken for the cross-stroke of a t, it but seems likely that the early readers jumped to conclusions because of the modern use of ‘to’. Chambers, and Flower, (Exeter Book, facs., p. 70Google Scholar) bracketed and underscored to as not distinguishable under ultra-violet light. As already mentioned, Dr Ker agreed with me (by letter and in conversation) that his copy of the photograph indicated on.

⁵³ Grein, C. W. M., Bibliothek der angelsächsischen Poesie 1 (Göttingen, 1857), 150Google Scholar.

⁵⁴ The Advent Lyrics, p. 84.

⁵⁵ Codex Exoniensis, p. 3.

⁵⁶ Dichtungen der Angelsachsen, stabreimend übersetzt 1 (Göttingen, 1857), 149–50Google Scholar.

⁵⁷ The Exeter Book, pt 1, p. 5.

⁵⁸ Cook, , reviewing several translations (The Christ of Cynewulf, pp. 78–80)Google Scholar, gives partial approval to both Grein and Gollancz, though he construes gedo in his glossary as imperative and quotes, as if it were an equally possible alternative, a translation proposed to him by James W. Bright that includes a radically different interpretation of 30b: ‘Make us worthy of this (what has preceded), us, whom he hath [denied] shut out from glory, when we were doomed in wretchedness, deprived of our home (heaven), to sojourn in this narrow world (earth).’ On p. 80 Cook explains, ‘With tð wuldre forlét Professor Bright compares wiðwurpon tō weorce, 3, and thence infers for forlǣtan the sense “reject”, “exclude”, “deny”.’ I doubt, however, whether the parallel is sufficiently exact, whether forlet can bear the sense attributed to it and whether, in the context, this interpretation of 30b fits as well as does that of Thorpe and Gollancz. So far as I know, Bright's interpretation has had no supporters since Cook.

⁵⁹ See above, n. 4.

⁶⁰ The Advent Lyrics, p. 84. 1 quote the more rigorous version in the notes. Facing the text, p. 48, he introduces an ‘us’ that alters the construction of þa þe, forcing it to claim usic as antecedent, as in Thorpe's version.

⁶¹ Klaeber, Fr., in a review of Cook's Christ (JEGP 4 (1902), 108Google Scholar), strongly supports this interpretation for 57b with comparable examples of to plus noun in Old English and Old Saxon. Cook's note on 57b had suggested the interpretation as a possibility, attributing it to ‘the translators’. The context of 30b is sufficiently different to make an adverbial use much less certain.

⁶² The Old English Advent, p. 69.

⁶³ Ibid.

⁶⁴ The Christ of Cynewulf, p. 77.