[In the following essay, Dobbie offers an overview of the scholarship on Judith, discussing its biblical source, fragmentary nature, date and circumstances of composition, and poetic elements.]

JUDITH

[T]he 349 lines of Judith which are still preserved are no more than a fragment of what was once a much longer work. The original length of the poem may be estimated from two bodies of evidence: (1) the relation of the extant text to the poet's source, and (2) the section numbers in the manuscript.

It has long been recognized that the source of our poem was the Latin Vulgate text of the apocryphal book of Judith.¹ The poet, however, did not follow the Latin text closely but omitted many nonessential features of the narrative and introduced a number of expansions and transpositions designed to increase the dramatic effect of the whole. Because of this freedom with which he handled his original, it is not easy to determine the exact passage in the Latin text which corresponds to the beginning of the Anglo-Saxon fragment. It is, however, probable that ll. 1-7a of the poem are the conclusion of some observations by the poet on the heroine's prayer in Judith xii.8 of the Latin text. The only earlier passage of the Latin which seems to have been used in the extant part of the poem is Judith x.19, which corresponds to ll. 46b-54a of the poem; the fleohnet described in these lines is the same thing as the conopeum of Judith x.19, which is not referred to in the twelfth chapter of the Latin text. To all intents and purposes, the poem closes with the end of chapter xv of the Latin, when the defeat of the Assyrians has been accomplished; no reference is made to Judith's later life, and her long hymn of praise (Judith xvi.2-21 in the Latin text) is omitted, only a brief reference being made to it (ll. 341b-345a) before the poet closes his work with some moralizing remarks of his own.

The most striking modification of the narrative in the hands of the Anglo-Saxon poet is the reduction of the number of principal characters to three—Holofernes, Judith, and Judith's unnamed attendant. The other persons in the Latin text, the Jews Ozias and Joacim, Achior the Moabite, and Vagao, Holofernes's lieutenant (who in the Latin text discovers the headless corpse of his leader), do not appear at all. Though King Nebuchadnezzar is likewise not mentioned, it is not improbable that some allusion was made to him in the part of the poem now lost. It is to be noted, however, that Holofernes is referred to as ðeoden (ll. 11, 66, 91, 268), hlaford (l. 251), eorla dryhten (l. 21), sinces brytta (l. 30), winedryhten (l. 274), and by other terms ordinarily reserved for kings and princes, and it may be that the poet intended to represent him as king of the Assyrians as well as their general.

Besides simplifying the cast of characters, the poet has achieved greater economy in the narrative by reducing the numerous speeches of the Latin original to four, three of which are spoken by Judith herself. These four speeches are: ll. 83-94a, Judith's prayer, uttered before she kills Holofernes; ll. 152b-158, her announcement to the people of her return to Bethulia; ll. 177-198, her speech while showing Holofernes's head to her people; and ll. 285-289a, the announcement to the Assyrians, by one of their leaders (Vagao in the Latin text), of the death of their general. All of these four speeches which remain are much expanded from their originals and achieve a dramatic effect which is wholly lacking in the Latin text.

Among the more striking additions made by the poet are the extended explanation of the fleohnet and its uses (ll. 46b-54a), the prediction of the fate which awaits Holofernes in hell (ll. 112b-121), and the long description of the battle (ll. 199-246a, continued in ll. 261b-267a, 301-323a). The second of these passages has ample precedent in the work of earlier poets (as in Elene 762b-771, Juliana 678b-688a, and elsewhere); the battle scenes are in the best Germanic tradition and lose nothing by comparison with similar scenes elsewhere in Anglo-Saxon poetry.

Returning now to the problem of the original length of the poem, we may apply the evidence of the Latin original in the following way. The extant fragment of the poem, corresponding to the Latin text from the middle of the twelfth chapter to the end of the fifteenth, embraces approximately 30 percent of the entire story.² If the proportion between the length of the Latin text and the length of the Anglo-Saxon was constant throughout, we would have, as the original length of the poem, somewhat less than twelve hundred lines. But in view of the freedom with which the poet handled his source in the part of his work which is still preserved, such an estimate is inconclusive, and corroborative evidence must be obtained from the section numbers in the manuscript. … [The] three full sections of the poem which still remain average about a hundred and ten lines in length; if all the sections were as long as these, the entire poem of twelve sections would have run to about 1320 lines, more or less, or about the same length as Elene. This is the figure adopted by Förster,³ and it is on the basis of it that he assumes the loss of three full gatherings, or twenty-four folios, from the manuscript before fol. 202.⁴ The difference between 1200 lines, indicated by the evidence of the Latin source, and 1320 lines, determined from the evidence of the sectional divisions, is not great, and the two figures may be reconciled by one of two assumptions: (1) that the poet was relatively fuller in his treatment of the early part of the story, perhaps supplying an introductory exposition similar to the one in Daniel, or (2) that the sections were of relatively shorter length toward the beginning of the poem.

We may, then, conclude with some confidence that in its complete form Judith was a poem of twelve or thirteen hundred lines. In marked contrast to this estimate is the statement by Cook that “the poem seems virtually complete as it now is.”⁵ Pointing to the double occurrence of the verb tweode, in l. 1b and l. 345b, as well as to the repetition of ll. 6b-7a in ll. 344b-345a, Cook adds that “the lines which here [in the extant text] stand first are echoed so significantly at the end that it is difficult to believe that more than a very few lines are missing.” But this judgment, which seems unduly subjective, is hardly sufficient to outweigh the conclusions reached above.

On the question of the date of the poem there has been little agreement. The older scholars, even including ten Brink,⁶ Wülker,⁷ and Ebert,⁸ grouped it with the Cædmonian poetry, not only because of the Old Testament subject matter but also because of the undoubted similarities in spirit and treatment to Exodus. Resemblances to the Cynewulfian poems have also been noted, particularly in vocabulary and style, and Sarrazin⁹ went so far as to ascribe Judith to Cynewulf. Neumann,¹⁰ on the basis of a much more thorough study of the poem than Sarrazin had made, also considered Cynewulf the probable author. But tests based on meter and grammatical forms all point to a much later date than either Cædmon or Cynewulf.

Metrically the most striking feature of the poem is the high frequency of expanded (three-stress) lines. There are no less than sixty-six regular expanded lines (ll. 2-12, 16-21, 30-34, 54-61, 63-68, 88-95, 97-99, 132, 272-273, 287-290, 337-348), besides two lines (ll. 96, 349) which have a normal first half-line but a second half-line which conforms to the expanded rather than to the normal pattern, and one line (l. 62) which appears to have been an expanded line but the second half of which has been lost.¹¹ The expanded lines in Judith are thus nearly 20 percent of the entire number, a much higher proportion than we find in any other Anglo-Saxon poems except the Dream of the Rood and Maxims I.¹²

Frequent attention has also been called to the relatively large number of end-rimes in the poem. We find seven cases in which the first half of a line rimes with the second half (grunde:funde, l. 2; sīne:wīne, l. 29; nēosan:forlēosan, l. 63; næs:wæs, l. 113; bewunden:gebunden, l. 115; gūðe:ūðe, l. 123; scaron:waron, l. 304) and one case in which the second half of one line rimes with the first half of the next (strēamas:drēamas, ll. 348f.). In addition there are a number of imperfect rimes, bedreste:gehlæste, l. 36, and hyrde:gest¯yrde, l. 60, where the vowels are not identical; rondwiggende:wēnde, l. 20, gecoste:eornoste, l. 231, þoligende:ende, l. 272, and gedyrsod:god, l. 299, where a stressed syllable rimes with an unstressed syllable.¹³

Although the high frequency of expanded (three-stress) lines in Judith seems not to provide any certain evidence as to date, the end-rimes are of greater significance. Many years ago Friedrich Kluge, in a study of rime in Old Germanic verse,¹⁴ showed that while end-rime appears sporadically in the earliest Anglo-Saxon poetry,¹⁵ its frequency increases in the later periods, and that rime may, within limits, be used as a criterion of date. If we disregard the Riming Poem (which is a tour de force and not to be reckoned as in the same tradition as the other extant Anglo-Saxon alliterative poetry), we find that Judith far surpasses all the other extant poems, even the Battle of Maldon, in the frequency of its end-rimes. Kluge would therefore assign Judith to the tenth century. Luick,¹⁶ on linguistic evidence derived from his metrical study of the poem, also considers it a late work, though he does not suggest a definite period. Furthermore, in addition to the evidence which has already been presented, we find in Judith certain metrical irregularities which point to a late date. For example, in l. 279, his goldgifan gæstes gesne, we have four alliterating syllables; in l. 149, of ðære ginnan byrig hyre togeanes gan, we have two alliterating syllables in the second half-line and only one in the first; and in l. 55 we have the very unusual alliteration of sn with st. If these lines are not corrupt, they are evidence of a weakening of metrical standards which may be paralleled in the Battle of Maldon but is still foreign to the Battle of Brunanburh. So far as the metrical evidence goes, then, it favors dating Judith in the tenth century, and in the middle or late rather than the early part of that century. If we assume an earlier date for the poem, we are faced by the paradox of a poet who was so familiar with the older poetry that he could produce a work in the heroic style which falls little short of his models but was at the same time not always able to imitate, or not always desirous of imitating, the older and more rigid alliterative patterns. Such poets existed late in the tenth century (as witness the author of Maldon) but, so far as we know, not at any earlier period.¹⁷

The question of the original dialect of Judith also requires consideration. Foster¹⁸ called attention to the Anglian (or Kentish) form þēgon (West Saxon þagon), l. 19, as well as to several unsyncopated verb forms, sēceð, l. 96, hafað, l. 197, and the past participle ārēted, l. 167, which is required by the meter, inferring from these words that the original dialect of the poem was Anglian. His conclusions are accepted by Cook.¹⁹ If, then, the poet wrote in the Anglian dialect, he must have been a Mercian, since the existence of a poet of such powers in Northumbria in the late tenth century is very unlikely. But there is no need whatever to assume an Anglian origin for the poem. As Tupper has pointed out,²⁰ unsyncopated verb forms are by no means rare in West Saxon poetry; in the Meters of Boethius, for example, besides the more usual syncopated forms we find not only sēceð, Met. 19, 8; 19, 45; 22, 15, and hafað, Met. 8, 46; 9, 63, but also unsyncopated forms of many other verbs, such as scrīfeð, Met. 10, 29, wealdeð, Met. 21, 33, and weorðeð, Met. 13, 56; 18, 9; 28, 76. Nor does þēgon necessarily point to Anglian origin, since, like āgēted, Brunanburh 18, and bestēmed, Brussels Cross 2, it may be the result of close imitation of the older poetic formulas. Tupper's conclusion, that “there seems to be no good grounds for regarding the Judith as anything else than a West Saxon poem,” is amply justified by the available evidence.

Of the author of Judith and the circumstances of composition we know nothing at all, but some ingenious suggestions have been made. Cook in 1888 presented at some length the hypothesis that the poem was “composed, in or about the year 856, in gratitude for the deliverance of Wessex from the fury of the heathen Northmen” and that it was dedicated to Judith, the second wife of King Æthelwulf.²¹ Foster, accepting Cook's theory that the poem was written in honor of an Anglo-Saxon noblewoman but placing its origin in Anglian territory rather than in Wessex, argued that it was composed in commemoration of Æthelflæd, the “lady of the Mercians,” the heroic daughter of King Alfred who, after the death in 910 of her husband the alderman Æthelred, conducted with vigor the resistance of Mercia against the Danes.²² Both of these theories appeal to the historical imagination, but neither is supported by any concrete evidence. Cook's hypothesis puts too much weight on the identity of names and involves dating the poem much earlier than most present-day scholars are inclined to date it; Foster's is preferable on chronological grounds but still fails to inspire conviction. There is, after all, no need to assume that the poet had any contemporary figure in mind. He was not the first poet, or the last, to be attracted by the story of the Jewish heroine.²³

Notes

1. The Vulgate text of Judith is not only shorter than the Greek Septuagint text but varies from it in many details and apparently represents a new recension by St. Jerome himself, who says, “multorum codicum varietatem vitiosissimam amputavi.” The English Bible (either the King James version or the Revised Version of 1881-1895, in both of which the translation of Judith is based on the Septuagint) is therefore an even less satisfactory basis of comparison than in the case of other Anglo-Saxon biblical poetry.

2. In the Vulgate text of Judith, as far as the end of chapter xv, there are 315 verses, of which the part from xii.10 to the end of chapter xv contains ninety-five.

3. Die Beowulf-Handschrift [by Max Förster. Leipzig, Germany, 1919.], pp. 88f.

4. See above, p. xiii. Since the 8 extant folios of the poem contain 343 lines and the first word of another, 32 folios would contain between 1350 and 1400 lines. This number is close enough to the 1320 lines assumed on the basis of the sectional divisions.

5. Judith (1904), p. 21.

6. Geschichte der englischen Litteratur I (Berlin, 1877), 59f. But in the second edition, revised by Alois Brandl (Strassburg, 1899), Judith is grouped with Genesis B and other later poems.

7. Grundriss zur Geschichte der angelsächsischen Litteratur (Leipzig, 1885), pp. 140-142.

8. Allgemeine Geschichte der Literatur des Mittelalters im Abendlande III (Leipzig, 1887), 24-26.

9. Beowulf-Studien (Berlin, 1888), pp. 160-165.

10. Über das altenglische Gedicht von Judith (Kiel, 1892).

11. Although these figures are different from those given by Pope, The Rhythm of Beowulf, p. 100, his identification of the expanded lines does not differ materially from mine. His count of 136 verses (i.e., “half-lines” in the terminology of this edition) includes both halves of each of the 66 expanded lines mentioned above, or a total of 132, plus ll. 62a, 96b, 349a (for which he suggests the emenda tion [sæs] ond swegles dreamas), and 349b.

12. The Dream of the Rood has 31 expanded lines (plus two half-lines of the expanded type) in a total of 156, or approximately the same percentage as Judith; Maxims I has 69 expanded lines (plus a few half-lines) in a total of 204, or nearly 35 percent. Following Pope, I exclude Genesis B from consideration here, since it is a translation from Old Saxon.

13. In this account of the end-rimes I ignore suffix rimes, such as dēman: brogan, l. 4, þīnre: þearfendre, l. 85, etc. Foster, in his account of the rimes (Judith [T. G. Foster, Judith, Strassburg, Germany, 1892] pp. 28-33), treats these on an equal basis with the others.

14. Beiträge IX (1884), 422-450.

15. In Beowulf, for example, we find yrremōd: stōd, l. 726; wēn: gēn, l. 734; gefægon: geþægon, l. 1014; herepād: gebād, l. 2258, and wrecan: sprecan, l. 3172, besides a number of other rimes which go beyond the limits of a single line.

16. Beiträge XI (1886), 490f.

17. Brandl, in Paul's Grundriss der germanischen Philologie (2d ed.), II, 1, 1091, dates the poem in the tenth century, and C. W. Kennedy, The Earliest English Poetry [New York and London, 1940], p. 282, more precisely in the first half of that century. On the other hand, Kemp Malone, in A Literary History of England, edited by A. C. Baugh (New York, 1948), p. 68, assigns it to the ninth century, though he admits the possibility of a later date. G. K. Anderson, The Literature of the Anglo-Saxons (Princeton, 1949), p. 133, suggests the latter half of the ninth century. R. Imelmann, Beiblatt zur Anglia XIX (1908), 6, would date the poem as early as 850.

18. Judith, pp. 49-53.

19. Judith (1904), p. viii. Cook also regarded hēhsta, ll. 4, 94, and nēhsta, l. 73, as indications of a northern origin; but as Foster (Judith, p. 50) had already noted, such forms are not uncommon in late West Saxon.

20. Journal of English and Germanic Philology XI (1912), 82-89. See also Imelmann's arguments to the same effect, Beiblatt zur Anglia XIX, 2-3.

21. Judith (1888), pp. xxiv-xxxiv. Cook's further conjecture that the author was Swithhun, bishop of Winchester, is of no merit and need be mentioned only in passing.

22. Judith, pp. 89-90.

23. Ælfric's semi-rhythmical life of Judith, edited by B. Assmann, Angelsächsische Homilien und Heiligenleben (Kassel, 1889), pp. 102-116, seems to have no relation to the poem and need not be discussed here.