¹ A copy of the prose model is preserved in Oxford, Bodleian Library, Bodley 180 (B). The manuscript which contains the alternating Old English verse and prose is London, British Library, Cotton Otho A. vi (C). In the latter half of the seventeenth century Franciscus Junius made a transcription (J) of both manuscripts, which is now housed in the Bodleian Library as Junius 12. Since the fire of 1731 left C severely damaged, J is of great use in reconstructing the original text.

² Only one verse passage adds anything of substance to the prose, and it seems to have been taken from a Latin commentary. The verse passage is Metre xx. 166a–175a. A marginal note in one Latin commentary reads ‘caelum et terram mareque in modum ovi figurari’. Presumably on the basis of this simile the versifier added a short description of the earth, comparing its shape with the shape of an egg. On the use of commentaries, see Schepss, Georg, ‘Zu König Alfreds Boethius’, ASNSL 94 (1895), 149–60Google Scholar, and Donaghey, Brian S., ‘The Sources of King Alfred's Translation of Boethius's De Consolatione Philosophiae’, Anglia 82 (1964), 23–57, at 30–1Google Scholar. A recent study by Wittig, Joseph S. (‘King Alfred's Boethius and its Latin Sources: a Reconsideration’, ASE 11 (1983), 157–98)Google Scholar has thrown doubt on the traditional beliefs about the extent to which the author of the prose translation relied on Latin commentaries, but it does not address the question of the possible influence on this particular verse passage.

³ I wish to thank Alan Bliss, who first aroused my interest in Old English syntax and assisted me at every stage of this project. I also wish to thank Fred C. Robinson for his valuable advice and encouragement, John C. Pope, who kindly suggested numerous improvements, and R. Allen Shoaf, who helped put this paper in final form. The shortcomings that remain are my own.

All quotations from and references to the Metres are taken from Tie Paris Psalter and the Meters of Boethius, ed. Krapp, George Philip, The Anglo-Saxon Poetic Records 5 (New York and London, 1932)Google Scholar. Each reference gives the number of the Metre in roman capitals followed by arabic numerals indicating the line. (Krapp used arabic numerals for both.) Quotations from and references to the prose are taken from King Alfred's Old English Version of Boethius De Consolatione Philosophiae, ed. Sedgefield, Walter John (Oxford, 1899; repr. Darmstadt, 1968)Google Scholar. Each page and line reference to the prose is in arabic numerals, which indicate the starting point of the clause in question. The prose line reference does not necessarily indicate the extent of the clause. The abbreviations in Sedgefield's edition are silently expanded, and I have omitted from both editions the editorial signals for insertions, emendations and transitions from C to J. Other editions of the verse which I refer to include Sedgefield's (ibid.); Die Altenglischen Metra des Boetius, ed. Krämer, Ernst (Bonn, 1902)Google Scholar; and Die Handschrift von Exeter, Metra des Boetius, Salomo und Saturn, Die Psalmen, ed. Assmann, Bruno (Leipzig, 1898), pp. 247–303.Google Scholar

⁴ ASE 9 (1981), 157–82.

⁵ The question of particles is treated more thoroughly below, pp. 172–3. On the question of stress on auxiliaries, see Kuhn, Hans, ‘Zur Wortstellung und -betonung im Altgermanischen’, BGDSL 57 (1933), 1–109, at 52–7Google Scholar; he stated that although finite verbs in dependent clauses are more frequently stressed than those in principal clauses, auxiliaries in each kind of clause are often unstressed. See also Bliss, A. J., The Metre of Beowulf (Oxford, 1958; rev. ed. 1967), §12Google Scholar. Cosmos, Spencer (‘Kuhn's Law and the Unstressed Verbs in Beowulf’, Texas Stud. in Lit. and Lang. 18 (1976), 306–28)Google Scholar distinguished between auxiliaries and ‘fully lexical finite verbs’ precisely on the grounds that the latter normally occupy ‘positions of metrical prominence’. More recently Kendall, Calvin B. (‘The Metrical Grammar of Beowulf: Displacement’, Speculum 58 (1983), 1–30)CrossRefGoogle Scholar has argued that not only auxiliaries but also many alliterating finite verbs lack a full metrical stress. See also Lehmann, W. P. and Tabusa, Takemitsu, The Alliterations of the Beowulf (Austin, Texas, 1958), p. 6.Google Scholar

⁶ By ‘verbal’ I mean a dependent infinitive or past participle; present participles and inflected infinitives preceded by to are automatically excluded. When there are two verbals only the first is considered, even if one is grammatically dependent on the other, as in the line, ‘þær he wolde a winnan onginnan’ (xxv.69). Stanley, E. G. (‘Verbal Stress in Old English Verse’, Anglia 93 (1975), 307–34Google Scholar, esp. 322–4), noted that infinitives are occasionally unstressed in other Old English poems, but the proportion is quite small.

⁷ The terms ‘common’, ‘demonstrative’ and ‘conjunctive’ come from Andrew, S. O., Syntax and Style in Old English (Cambridge, 1940)Google Scholar. See also Bliss's preliminary remarks, ‘Auxiliary and Verbal’, pp. 157–63, where he discusses in more detail a number of points only briefly considered here. Mitchell's, BruceOld English Syntax, vols. (Oxford, 1985)CrossRefGoogle Scholar, which came out after the completion of this paper, takes up these matters at §§3887–902 and §§3944–7. On rare occasions O precedes S and v; e.g. ‘and hire saule mon sceolde tedan to helle’ (Sedgefield 102.1).

⁸ See his ‘Zur Wortstellung und -betonung’.

⁹ ibid. pp. 8–10.

¹⁰ ibid. pp. 43–5.

¹¹ Altgermanische Metrik (Halle, 1893), §§22–9. In §24 he admitted some exceptions.

¹² ‘Auxiliary and Verbal’, pp. 160–2.

¹³ My count of auxiliaries does not include those where an infinitive is understood and excludes one auxiliary which is the second member of a pair: ‘wilt oððe most’ (XXIV.56a). Two trisyllabic auxiliaries are included among the disyllabic: tiliað (X.22b and XI.79b) and wilnige (XXIX. 1a).

¹⁴ Seven of the ‘later’ auxiliaries combine with a verbal to form a half-line and will be discussed below, pp. 177–82; another four (mage (VII.32b), sceal (XX.197b) and mæg (XXI.38b and XXII.13 b)) are breaches of Sievers's Rule; another two (ne magon (XXV. 59a) and meaht (XXVI. 107b)) are unstressed in breach of Kuhn's First Law. The other twelve are variously distributed.

¹⁵ Two ‘later’ disyllabic auxiliaries are unstressed in breach of Kuhn's First Law: willað (v.22a) and wolde (IX. 12a). The latter is discussed below, p. 176.

¹⁶ I consider the starting point of a clause to be the first word which performs a grammatical function in the clause. Expletives such as hwæt and eala and words used in direct address are outside the clause. On many occasions Kuhn's First Law seems to justify this treatment. According to the law a clause can have only one metrical dip either before or immediately after the first stressed word. In the passage ‘Eala, min drihten, þæet þu eart ælmihtig’ (XX. 1) there is a dip before the first stressed word of the second half-line. The poet treated this line as if it contained two clauses, or rather, as if þæt began the clause anew.

¹⁷ With one exception, ‘hæt eft cuman’ (XXIX. 83b), where the alliteration falls on h. Clauses with initial auxiliaries are discussed below, p. 175.

¹⁸ Other initial words of uncertain grammatical function in the Metres include ær, ærðæm, forhwam, forðæm, forðy, hu, hwa, hwær, hwæper, nu, se, sio, siððan, swa, þara, þæm, þær, þæt, þonne and þy. Bliss observed, ‘The clauses of doubtful status constitute a central problem in the study of Old English verse syntax; until it is solved the preference of an Anglo-Saxon poet for parataxis or hypotaxis must remain a matter of conjecture’ (‘Auxiliary and Verbal’, p. 166). Andrew, S. O. in Syntax and Style in Old English and later in Postscript on ‘Beowulf’ (Cambridge, 1948)Google Scholar took up the question of clauses of doubtful status in great detail, although his conclusions have not met with wide approval; see, e.g., the reviews of the former by Larsen, H. (JEGP 41 (1942), 85–8)Google Scholar, Macdonald, A. (RES 17 (1941), 499–501)Google Scholar and Potter, Simeon (MLR 36 (1941), 252–5)CrossRefGoogle Scholar and that of the latter by Malone, Kemp (ES 32 (1951), 116–19).Google Scholar

¹⁹ See Bliss, ‘Auxiliary and Verbal’, pp. 157–9.

²⁰ An auxiliary begins a clause in the following half-lines: (a) asyndetic coordinate clause, 1.42a, 1.59b, 1.66b, 1.70a, 1.82b, XIII.28b, XX.144a, XXIV.8a, XXIV.11b, XXIV.17a, XXV.19a, XXV.64a, XXVI.72a, XXVI.79b and XXIX.83b; (b) subject pronoun, v.5a, v.29b, X.38a, XIII.33a, XIV.9a, XV.9a, XX.36a, XX.94a, XX.164b, XXIV.52a, XXVII.9a, XXVII.14a, XXIX.38a, XXX.9a and XXXI.8b; and (c) noun or noun phrase subject, 1.22a, 1.31b, IV.46a, VII.6b, VII.11b, VII.18a, VIII.33a, XI.22a, XI.23b, XI.55a, XI.64a, XIII.6a, XIII.23b, XX.107a, XX.145a, XX.150a, XX.153a, XX.241a, XXVI.51a, XXVI.90a, XXVI.113b, XXIX.34a and XXXI.12b.

²¹ Auxiliaries in principal clauses occur in the following half-lines: (a) initial stressed word, 1.11b, 1.18b, 1.28b, v.18b, v.19b, VI.4b, VII.13b, X.7b, XI.65b, XIII.37b, XX.79a, XXV.49b, XXV.59a, XXVI.84b, XXVI.85b, XXVI.107b, XXVII.21a, XXIX.86b and XXIX.88a; and (b) initial unstressed particle, Proem 8b, 1.72b, 1.76b, 11.4b, IV17b, v.1a, v.18a, v.26b, v.28a, VIII.2b, IX.9a, IX.12a, IX.45b, X.52a, XI.29a, XII.22a, XIII.1a, XIII.79a, XIX.17b, XX.34a, XX.200a, XX.219a, XXII.58b, XXIII.7a, XXIV.15a, XXIV.26b, XXIV.50b, XXIV.57b, XXV.37b, XXV.45b, XXVI.1a, XXVI.110b, XVII.18b, XXVII.31a, XXIX.12b and XXXI.1a.

²² The nine auxiliaries in the third word order are sceal (IV. 17b), wyrð (v. 18a), scealt (v. 26b), sceolde (IX.45b), meaht (XIX. 17b), mæg (XXV.37b), wile (XXIX.12b), sceolden (XXIX.86b ) and sceoldon (XXIX. 88a). On the suggestion of Professor Bliss, I have emended the punctuation of IV. 17 so that the clause begins ‘He gongan sceal’ rather than ‘geara gehwelce’. The clause preceding ‘He gongan sceal’ makes perfectly good sense when ‘geara gehwelce’ ends it. Fora parallel, see XIX. 27.

²³ Auxiliaries in dependent clauses occur in the following half-lines: (a) with a coordinating conjunction, IV.50b, v.24a, v.30b, VII.17a, VII.22b, VII.30b, X.15a, XI.46a, XI.52a, XI.68a, XI.70b, XI.75b, XIII.27b, XIV.4b, XIX.40b, XX.72a, XX.I39a, XX.257b, XXI.35b, XXIII.6b, XXIV.4a, XXIV.53b, XXVI.80b and XXX.I2a; and (b) with a subordinating conjunction, 1.27b, 11.14a, IV.15a, IV.49a, v.31b, v.40b, VII.10b, VII.29b, VIII.1b, VIII.16a, VIII.22b, IX.12b, IX.21a, IX.27b, IX. 53a, IX.58b, X.2b, X.12a, X.27a, X.39b, X.69b, XI.56b, XI.79b, XI.80a, XI.98b, XII.1a, XIII.32b, XIII.35b, XIII.41b, XIV.7a, XVI.1a, XVI.3b, XVI.8b, XV1.19b, XVII.23b, XVIII.9b, XX.21b, XX.65b, XX.70b, XX.115b, XX.I28a, XX.I29b, XX.244b, XXI.5a, XXI.38b, XXII.1b, XXII.19b, XXII.37a, XXII.42b, XXII.50b, XXII.52b, XXII.56b, XXIII.2b, XXIV.6a, XXIV.46b, XXV.22a, XXV.72b, XXVI.46a, XXVI.69a, XXVI.102b, XXIX.1a, XXIX.32b, XXX.11b and XXXI.1b.

²⁴ The six auxiliaries in the first word order are sceal (VII.30b), mæg (X.12a), sceal (XI.52a), wille (XII.1a, with accidental alliteration), wille (XVI.1a) and bid (XXII.37a). The clauses containing mæg and bid (X.12a and XXII.37a ) each begin with the second þeah of a þeah … þeah pair and must be construed as principal.

²⁵ Henceforth I call these half-lines verbal-auxiliary half-lines; they are 1.27b, 1.31a, 1.39a, 1.62b, 11.14a, IV.15a, IV.23b, IV.34b, IV.40b, IV.50b, v.18a, VII.5b, VII.10b, VII.29b, VIII.22b, IX.12a, IX.15a, IX.16b, IX.19b, IX.27b, IX.35b, IX.45b, IX.58b, X.2b, X.22b, X.64b, XI.28b, XI.70b, XI.75b, XI.79b, XI.98b, XIII.27b, XVI.19b, XVIII.9b, XIX.25a, XIX.32b, XIX.40b, XX.102b, XX.273b, XXIII.6b, XXIV.37b, XXIV.46b, XXV.72b, XXVI.18b, XXVI.72a, XXVI.80b, XXVI.82b, XXVI.104a, XXVII.7a, XXIX.12b, XXIX.39b, XXIX.86b and XXIX.88a.

²⁶ I follow the method of scansion used by Bliss in The Metre of Beowulf.

²⁷ D. Slay discussed similar instances of displaced but unalliterating verbs (‘Some Aspects of the Technique of Composition of Old English Verse’, TPS 1952, 1–14, at 13–14). He called them ‘mistakes’ with no ‘mitigating circumstances’ to explain the unusual position for the verb.

²⁸ ‘Auxiliary and Verbal’, p. 162, n. 18. The half-line in Beowulf is 1728b, where transpositon may be called for: ‘*hwilum he læteð on lufan hworfan’.

²⁹ Breaches of Sievers's Rule are not limited to verbal-auxiliary half-lines. Other auxiliaries include mæge (VII.32b), sceal (XX.197b) and mæg (XXI.38b and XXII.13b). This count does not include other finite verbs. In fact Sievers singles out the Metres as a poem that allows ‘zahlreiche Verstösse’ of his Rule (§29).

³⁰ Krämer emended to geþwened (XX.102b); if the reading in MS C is left unemended, geþawened undergoes resolution, so for metrical purposes the two forms are almost identical. Assmann and Krämer also emended weorðað (from C) to weorðeð for the sake of agreement in number.

³¹ Type IA appears nowhere else; type IA* occurs twice in the third word order: ‘secgan geherde’ (IX. 15 a) and ‘ðioton ongunnon’ (XXVI.80b).

³² So far in this study auxiliaries like ne sceal and onginð have been treated as monosyllabic and ne sceoldon and ongunnon as disyllabic; moreover, auxiliaries like hafað have been considered monosyllabic whether or not they undergo resolution. When these verbs are analysed metrically, however, the proclitics and prefixes must be taken into account, and verbs like hafað not subject to resolution must be treated as disyllabic. The differences arising from the two ways of treating the additional syllables are too slight to affect my conclusions, but the occasional differences should be noted.

³³ The half-lines with monosyllabic auxiliaries are ‘tosceaden wyrð’ (v.18a), ‘wile onlætan’ (XI.75b, with resolution of wile), ‘gesecgan ne mæg’ (XIX.40b), ‘mæg aweorpan’ (XXIII.6b), ‘ymbebæted hæfð’ (XXIV.37b, the medial e in ymbebated to be ignored in scansion), ‘gebidan ne magon’ (XXVII. 7a, with resolution magon) and ‘gestigan wile’ (XXIX. 12b, with resolution of wile).

³⁴ Some passages show changes from the prose that are so extensive that they can be most plausibly explained as the result of efforts to give them the preferred metre. See, e.g., VIII. 21b–23a (prose 33.29) and XI.69b–70 (49.24), where the versifier removed the negative proclitic and placed it in an introductory clause (not found in the prose) in order to preserve the negative sense. Another verse passage, VII.4a–6a (prose 26.24), shows how the poet added a prefix to the verbal, transferred the negative proclitic to an introductory clause and changed the tense of the auxiliary to give the half-line the metrical type IA*: ‘meahte asettan’ (VII.4b). Cf. the half-line ‘settan meahte’, five lines below. Each half-line's metrical type follows a pattern determined by the relative positions of the auxiliary and verbal. If a disyllabic auxiliary is first, the half-line is given metrical type 1 A*; if it is second, metrical type 2 A. In all of these examples the prose influence on the word order is negligible.

³⁵ Krapp and Sedgefield followed MS C in the spelling liban (IX. 58b), but for the sake of metre I follow Krämer and Assmann in correcting it to libban.

³⁶ The two trisyllabic words, tiliað and forgiten, undergo resolution.

³⁷ Krämer inexplicably placed se in the previous half-line, yet punctuated so that it remained in the clause with ‘gehaten wæs’ (X.46b). Half-lines with similar constructions but with a first word which is not a subject pronoun include ‘ac geþweorod sint’ (XX.72a), ‘ðæt forweorðan scylen’ (XXI.34b) and ‘swa bereafod sie’ (XXII.50b). In other similar half-lines, ‘Þa bioð gehyrste’ (XXV.8b) and ‘Þara is gehaten’ (XXVIII.25b), the alliteration falls on the initial particle, so that the following verb is unstressed. I consider these five to be exceptions to the variations of verbal-auxiliary half-lines discussed here.

³⁸ S. O. Andrew took exception to half-lines in Beowulf sach as ‘Ic gefremman sceal’ on the grounds that a conjunction must have been dropped from them. Otherwise, according to his strict application of the principles of syntax, there is no way to account for the conjunctive order (S [O…] Vv) (Syntax and Style, §70). Bliss noted a ‘slight preference’ for considering such half-lines subordinate but he allowed that they may also be principal (‘Auxiliary and Verbal’, p. 178).

³⁹ A Guide to Old English Revised with Texts and Glossary (Toronto, 1982) §162.4. See also the discussion by S. O. Andrew, Syntax and Style, §§121–7, and that in Mitchell's, Bruce ‘Adjective Clauses in Old English Poetry’, Anglia 81 (1963), 298–322Google Scholar. Mitchell's Old English Syntax has appeared since the first draft of this paper, and his sections on se þe clauses (§§2153–362) should be consulted, esp. §§2159 and 2204, for observations that differ from those presented here.

⁴⁰ A Guide, §163.1.

⁴¹ The reading of MS B is ‘ðe þ wille’.

⁴² Syntax and Style, §§121–7. Pace Andrew, false concord is not a compelling reason to suspect scribal corruption. Indeed, the consistency with which false concord occurs – particularly in what he calls exepegetical þara þe passages – suggests that such usage was acceptable to the poet.

⁴³ There are two exceptions, ‘weorðan sceolden’ (XXIX.86b) and ‘weorðan sceoldon’ (XXIX.88a), which reverse the word order of the prose to the third word order (Vvý) in order to form verbal-auxiliary half-lines. Since a second, dependent, verbal is added to each clause, however, they may be considered ‘newly introduced’ in a sense.

⁴⁴ ‘Verse Influences in Old English Prose’, Philological Essays … in Honour of Herbert Dean Meritt, ed. Rosier, James L. (The Hague, 1970), pp. 93–8, at 95Google Scholar. Campbell used the term ‘subordinate’ for the word order I call ‘conjunctive’.

⁴⁵ In one case the prose offers little help: X.35a–37b, prose 46.17–18. Judging from the sense of the passage as a whole, I interpret the first clause as principal and the second as dependent. While the position and stress of the auxiliary in many cases show the grammatical function of the clause, a clause of doubtful status with an auxiliary in the second word order (ýV) may be principal, dependent or an apo koinou clause. In considering clauses like this it may be wise to fall back on the suggestions made by Mitchell, Bruce in ‘The Dangers of Disguise: Old English Texts in Modern Punctuation’, RES n.s. 31 (1980), 385–413Google Scholar, where he argued against holding a strict distinction in some cases between principal and dependent clauses.

⁴⁶ See above, p. 175.

⁴⁷ See above, n. 22, and the discussion of verbal-auxiliary half-lines, pp. 172–82.

⁴⁸ See above, table 3.

⁴⁹ See above, p. 176.

⁵⁰ The five unstressed auxiliaries are sceal (11.2b), is (XVI.15b), was (XXVI.96a), hæfð (XXVIII.26b) and magon (XXX.18a). The corresponding prose clauses are 8.6, 67.32, 116.25, 126.9 and 141.19. Not every editor has agreed that each of the five verse clauses is dependent; three are punctuated differently in some editions: Assmann and Sedgefield punctuated 11.2b–4a as principal, Krämer and Sedgefield punctuated XXVI.96–7a as principal and Assmann and Krämer punctuated XXVIII.26b–27 as principal. Sedgefield punctuated three of the five corresponding prose clauses as principal, including the two corresponding to the verse clauses he considered principal (8.6 and 116.25 for the verse 11.2b–4a and XXVI.96–7a). The third principal prose passage is 141.19 (for XXX.18, which all editors have considered dependent), but it has a loose syntactic relation to the verse: one prose clause is expanded into two verse clauses. The results of this study favour the alternative punctuation for the three verse clauses mentioned above. Furthermore, the respective passages do not suffer in meaning or syntactic smoothness when these clauses are construed as principal. The other two clauses (XVI. 15b and XXX. 18) should remain dependent, because the passages where they appear would suffer from a change in punctuation. Consequently, simple repunctuation in accord with other editors' choices can reduce the total number of exceptions to two out of 104, a proportion that is even more one-sided than for clauses unambiguously dependent.

⁵¹ See above, table 4.

⁵² Note that the fifty-three initial auxiliaries are not included, but when they are, the fifteen asyndetic coordinate clauses significantly increase the number of instances of the first word order (vV) in dependent clauses. Revised to include initial auxiliaries, the distributions are for principal clauses vV 88, ýV 37 and Vvý 11 and for dependent clauses vV 26, ýV 108 and Vý 73.

⁵³ See above, tables 3 and 4 and pp. 186–8.

⁵⁴ At X.50a and X.51b.

⁵⁵ ‘Auxiliary and Verbal’, p. 178.

⁵⁶ Cf. the Appendix, ibid. pp. 180–2, where Bliss pointed out entirely different influences in Beowulf that may account for the presence of the prefix ge- on the past participle. Moreover, there are eleven half-lines, such as ‘habban ne mihte’ and ‘ðurhfon ne mihte’ (Beo. 462b and 1504b), which conform to metrical patterns avoided in the Metres.

⁵⁷ ‘The Old English Epic Style’, English and Medieval Studies presented to J. R. R. Tolkien on the occasion of his Seventieth Birthday, ed. Davis, Norman and Wrenn, C. L. (London, 1962), pp. 13–26.Google Scholar

⁵⁸ ibid. p. 16.

⁵⁹ See MacDonnell, A. A., A Sanskrit Grammar (Oxford, 1975), pp. 243–4Google Scholar. Lehmann, and wrote, Tabusa, ‘By the Indo-European and early Germanic linguistic practices, verbs of dependent clauses were stressed and might alliterate’ (The Alliterations of the Beowulf, p. 6)Google Scholar. I use Campbell's terms ‘epic’ and ‘lay’ for want of a better distinction. If the Metres belong to a different poetic tradition it may not be that of the ‘lays’ at all but some other. It is difficult to come to a definite conclusion on this matter because the evidence for lays is scanty. I have recently begun to examine auxiliaries and verbals in Cynewulf's four signed poems, where the auxiliaries do not behave quite as they do in the Metres or in Beowulf, but they more closely resemble the auxiliaries of the latter. Cynewulf may have been writing in the same tradition as that of Beowulf, or alternatively all three poets could represent three different individual variations of one ‘classical’ tradition. Indeed, all three poets shared a tendency to stress the auxiliary in a dependent clause. However, as I have shown, the Metres poet was more consistent in this pattern, and in Beowulf the stress appears to be the result of another tendency not related to grammatical function, as Bliss observed.

⁶⁰ Kuhn's Laws, which locate the metrical dip at the beginning of the verse clause, are perhaps the most succinct expression of this relation.

⁶¹ See above, p. 173.

⁶² ‘The Metrical Grammar of Beowulf: Displacement’.

⁶³ E.g., Hopper, Paul J. (The Syntax of the Simple Sentence in Proto-Germanic (The Hague and Paris, 1975), p. 56)CrossRefGoogle Scholar, applied Kuhn's conclusions for word order and stress in verse to his own examples from prose. In another study (‘Old English as an SVO Language: Evidence from the Auxiliary’, Papers in Ling. 5 (1972), 183–201) Lawrence Mitchell concentrated the bulk of his discussion on Old English prose, but concluded by applying his derivational process to verse.

⁶⁴ Toward a Grammar of Proto-Germanic, ed. van Coetsem, Frans and Kufner, Herbert L. (Tübingen, 1972), pp. 39–68.CrossRefGoogle Scholar

⁶⁵ ibid. pp. 243 and 244. He was certainly aware of the metrical form of the inscription; see his comments, The Development of Germanic Verse Form (Austin, Texas, 1956), pp. 28–9.

⁶⁶ ‘Auxiliary and Verbal’, pp. 173–4.

⁶⁷ See Campbell, ‘Verse Influences in Old English Prose’. The rhythmical influence of verse on prose has received a greater amount of scholarly attention. Whitelock, Dorothy (‘The Prose of Alfred's Reign’, Continuations and Beginnings: Studies in Old English Literature, ed. Stanley, Eric G. (London, 1966), pp. 67–103, esp. 98–101)Google Scholar discussed the varying degrees of stylization found in prose at the time of Alfred's translation of Boethius, though she did not directly deal with the possibility of verse influence on the prose. Funke, Otto (‘Studien zur alliterierenden und rhythmisierenden Prosa in der älteren altenglischen Homiletik’, Anglia 80 (1962), 9–36)Google Scholar found a tradition of alliteration and rhythmical phrasing already present in prose well before Ælfric and Wulfstan. An important study by Mclntosh, Angus (‘Wulfstan's Prose’, PBA 25 (1949), 109–42)Google Scholar outlined five rhythmical ways of writing that extend from ‘classical’ Old English verse to ‘ordinary prose’ and noted similarities between ‘debased verse’ and rhythmical prose. His conclusions on Wulfstan's rhythmical prose have recently been challenged by Hollowell, Ida Masters (‘On the Two-Stress Theory of Wulfstan's Rhythm’, PQ 61 (1982), 1–11)Google Scholar, who argued that larger syntactical units rather than two-stress phrasing are the organizing principle of Wulfstan's prose. By comparison, Ælfric's rhythmical prose, as John C. Pope explained in the introduction to his edition of Ælfric's homilies, more closely approximates to the rhythm of ‘classical’ Old English verse (Homilies of Ælfric: a Supplementary Collection, 2 vols., EETS 259–60 (London, 1967–1968), 105–36). Where such extensive metrical influence is present, it seems plausible that there is also some syntactic influence of verse on the prose.