9£5 L4-2G Ml UC-NRLF B 3 S37 710 STUDIES IN LAYAMON'S VERSE A THESIS PRESENTED TO THE FACULTY OF THE GRADUATE DEPARTMENT OF NEW YORK UNIVERSITY UPON APPLICATION FOR THE DEGREE OF DOCTOR OF PHILOSOPHY BY SARAH J. McNARY. 1902 BALTIMORE J. H. FURST COMPANY 1904 STUDIES IN LAYAMON'S VERSE A THESIS PRESENTED TO THE FACULTY OF THE GRADUATE DEPARTMENT OF NEW YORK UNIVERSITY UPON APPLICATION FOR THE DEGREE OF DOCTOR OF PHILOSOPHY BY SARAH J. McNARY. 1902 BALTIMORE J. H. FURST COMPANY 1904 125 PREFATORY NOTE. The readings of ms. Cott. Otho. C. XII. have been for the most part disregarded, because Zessack has proved that it is not a recension of MS. Cott. Calig. A. IX., as Madden supposed, but is derived independently from a common original, and is not closely related either to the source or to the earlier manuscript. 296267 CONTENTS. PAGE. Chapter I. Previous Theories of L^amon's Verse, - - - 1 Chapter II. L^amon's Use of Alliteration and Rime, - 6 Chapter III. The Metrical Structure of the Brut, 17 Chapter IV. The Influence of Waco, 25 Conclusion, - ...- -36 BIBLIOGRAPHY. Authorities and Texts. Arthur, edited by F. J. Furnival, Early English Text Society Publications, No. 2, 1864. Ten Brink, B. : Geschichte der Englischen Literatur. Berlin, 1877. Altenglische Literatur (in Paul's Grundriss der German- ischen Philologie, Bd. II, p. 516). Strassburg, 1893. Einenkel, E. : J. Schipper, Englische Metrik in historischer und systematischer Entwicklung dargestellt (in Anglia, Bd. V, Anzeiger, pp. 30, 111). Halle, 1882. The Life of Saint Katherine. EETS. Pub. London, 1884. Ellis, A. J. : On Early English Pronunciation. London, 1869. Article on Metrics in the Transactions of the Philological Society, p. 442. London, 1875-6. Foster, T. G. : Judith. Studies in Metre, Language and Style. Quellen and Forschungen zur Sprach- und Kulturgeschichte der Germanischen Volker, LXXI. Strassburg, 1892. Fragment of the Song of Roland, ed. by S. Herrtage. EETS. Pub., Extra Series, No. XXXV. London, 1875-6. Guest, E. : A History of English Rhythms. A new edition, edited by the Rev. Walter W. Skeat. London, 1880. Hali Meidenhad, ed. by Cockayne. EETS. Pub., No. 18. Heath, H. F. : On the Old English Alliterative Line (in Transac- tions of the Philological Society, 1891-4). Keller, W. L. : Maistre Wace, eine stylistische Uutersuchung seiner beiden Roumane Rou und Brut. St. Gallen, 1886. Kluge, F. : Zur geschichte des reinies im altgermanischen (in Beitrage zur Geschichte der deutschen Sprache und Littera- tur, von Paul und Braune, Bd. IX, p. 422). Halle, 1898. The Lay of Havelok the Dane, ed. by W. W. Skeat. EETS. Pub., Extra Series, No. IV. London, 1868. de Lincy, Le Roux : Le Roman de Brut, par Wace, avec un Commentaire et des Notes. 2 torn. Rouen, 1836-8. viii Bibliography. Luick, K. : Heimische Metra (in Paul's Grundriss der Gerrnan- isehen Philologie, Bd. II, p. 994). Strassburg, 1893. Madden, Sir Frederic : La^anion's Brut, or Chronicle of Britain. Three Volumes. London, 1847. Marsh, G. P. : The Origin and History of the English Language. New York and London, 1862. Menthel, E. : Zur Geschichte des Otfridschen verses im Engli- schen (in Anglia, Bd. VIII, Anzeiger, p. 49). Halle, 1885. Mitford, W. : An Inquiry into the Principles of Harmony in Lan- guage and of the Mechanism of Verse, Modern and Ancient. 2d edition. London, 1874. Morley, H. : A First Sketch of English Literature. London, 1874. Le Morte Arthur, ed. by F. J. Furnivall. London, 1864. Paris, Gaston: Preface de la Traduction de Le Vers Francais ancienne et moderne, par M. Adolph Tobler. Paris, 1885. An Old English Miscellany, ed. by Rev. R. Morris. EETS. Pub. No. 49. Regel, K. : Die alliteration im La^amon (in Germanistische Stu- dien, herausg. von Bartsch, Bd. I, p. 172). 1872-5. Spruch und bild im La^amon (in Anglia, Bd. I, p. 197)- Halle, 1878. The Romance of Duke Rowlande and of Sir Ottuell of Spaine, ed. by S. Herrtage. EETS. Pub. Extra Series, No. xxxv. London, 1880. The Romance of Otuel, ed. by S. Herrtage. EETS. Pub. Extra Series, No. xxxix. London, 1882. Seinte Juliane, ed. by Cockayne. EETS. Pub. No. 51. Seinte Marharete, ed. by Cockayne. EETS. Pub. No. 13. The Sege of Melayne, ed. by S. Herrtage. EETS. Pub. Extra Series, No. xxxv. London, 1880. Sievers, E. : Altgermanische Metrik (in Paul's Grundriss der Ger- manischen Philologie, Bd. II). Strassburg, 1893. Schipper, J. : Englische Metrik, in historischer und systematischer Entwickelung dargestellt. Erster Theil. Altenglische Me- trik. Bonn, 1881. Grundriss der Englischen Metrik (Wiener Beitriige zur Eng- lischen Philologie, von Dr. J. Schipper, Bd. II). Wien und Leipzig, 1895. Bibliography. ix Metrische Randglossen (in Englische Studien, Bd. IX., Mis- cellen, p. 184; Bd. X, Miscellen, p. 187). Heilbron, 1886, 1887. Stephens, T. : The Literature of the Kymry. London, 1876. The Story of Genesis and Exodus, ed. by Rev. R. Morris. EETS. Pub. No. 7. London, 1876. Sweet, H. : History of English Sounds. Oxford, 1888. Tobler, A. : Le Vers Frangais ancienne et moderne, .... traduit par Karl Breul et Leopold Sudre. Paris, 1885. Trautman, H. : Uber La^amon's vers (in Anglia, Bd. II, p. 153). Halle, 1879. Zur alt- und mittelenglischen verslehre (in Anglia, Bd. V. Anzeiger, p. 111). Halle, 1882. Otfrid in England (in Anglia, VII. Anzeiger, p. 211). Halle, 1884. Zur kenntniss des altgermanischen verses, vornehmlich des altenglischen (in Anglia, Beiblatt V. p. 87). Halle, 1894-5. Zur kenntniss und geschichte der mittelenglischen stabzeile (in Anglia, Bd. XVIII, p. 83). Halle, 1896. Tyrwhitt, T. : Au Essay on the Language and Versification of Chaucer. Wissmann, T. : King Horn (in Quellen und Forschungen zur Sprachgeschichte, Bd. XVI.). Strassburg, 1876. Wiilcker, R. : Uber die quellen des La^amon (in Beitrage zur Geschichte der deutschen Sprache und Literatur, von Paul Braune, Bd. Ill, p. 524). Halle, 1878. Zessack, A. : Die beiden Handschriften von La^amon's Brut und ihr Verhaltniss zu einander. Inaugural Dissertation. Bres- lau, 1888. STUDIES IN LAYAMON'S VERSE. Chapter I. Previous Theories of La^amon's Verse. The nature and sources of the metrical form of La^amon's Brut have been the subject of much theorizing. The date of its pro- duction, the early part of the thirteenth century, at once suggests several possibilities as to the derivation of the verse ; and the length of the poem, 32,241 lines, and the apparent irregularity, give room for various theories and generalizations. No exhaustive investigation, however, was made until the laws of Old English verse were formulated. Mitford, 1 Tyrwhitt, 2 Regel, 3 H. Morley, 4 ten Brink 5 and Wissmann 6 regard the metre as of O. E. origin. Marsh 7 finds an intermixture of the A. S. and Norman systems of versification. A. J. Ellis 8 calls the verse little better than prose. Mad den 9 agrees with Guest 10 that the rimin g couplets of the poem were formed on a Latin rhythm with four accents. The more recent views 11 concerning La^amon form a chapter in the history of the war of Old English verse theories and of Middle English word-accent. Since the latter play so prominent a part in the literature of our subject, a brief statement of them may not be amiss. !p. 170. 6 King Horn, p. 57. 9 Vol. I, p. xxv. 2 p. 34. 7 p. 441. 8 p. 496. 10 pp. 406-16, 446-447. 3 Germ. Stud. I, p. 172. Anglia I, p. 197. *p. 73-74. 5 Gesch. der Engl. Lit., Vol. I, p. 237. See also Paul's Grundriss, Vol. II, p. 622. u Since 1876. 1 2 Studies in La Ramon's Verse. The backbone of the four-stress theory as applied to Old English verse and to its descendants in Middle English is the well known law of Lachmann : Every syllable immediately fol- lowing a long syllable or a short unstressed syllable, receives a subordinate accent ; and every syllable following a stressed short syllable remains unstressed. Inflectional and other weak endings must frequently be accented under application of this formula. The chief argument for its enforcement in Middle English descendants of the Old English metre — and this is the only kind of verse in which anybody tries to enforce it. — is the occurrence in such poems of rimes of (a) an inflectional with a root syllable, and (6) two inflectional endings. Much of the verse formed on Romance models l takes no account of the law, as even its supporters admit. And there is no evi- dence that it was observed in ordinary speech. The two-stress theory claims that ordinary word-accent and verse-accent should coincide ; that this is what happened in Eng- lish imitations of Latin and French rhythms, or the new metres would not have been intelligible to the hearers ; that these metres^ instead of accenting weak inflectional endings in the prescribed cases, tend to obscure and elide them ; and that no difference can be observed between such endings when they follow a long sylla- ble, or a stressed short one. Since Lachmann's law does not apply to these rhythms, to attempt to make the alliterative verse conform to it is illogical, especially since the two kinds of metres may be found in a single poem, as in the Bestiarius. 2 The argument from rime is answered by the admission that unstressed syllables, alike in verse derived from Old English and in that from Romance sources, were rimed from Old English times to Shakespere, though never with great frequency. 3 The advocates of the four-stress theory have uniformly claimed that in La^amon the O. E. verse laws are in full force, and that consequently Lachmann's law of verse-accent is still observed. 1 Wismann, Anglia V, p. 476. For a critique of this theory, see Schipper, Eng. Stud. IX, p. 186. 2 Schipper, Wien. Beit. II, pp. 162-169. 8 Schipper, Eng. Stud. X, pp. 196-201. Studies in Lazamon's Verse. 3 The adherents of the two-stress theory have either traced La3a- mon's verse to a foreign source ; or seen in it the O. E. types modified slightly or to a considerable extent by French influence ; or derived it from the older Teutonic verse from which the Old English itself was developed. The four-stress theorists have not done much detailed investiga- tion of this work. No one has applied ten Brink's O. E. types 1 to the Brut. In trying to establish his theory of word-accent, which of course must apply to La^amon, Wissmann gives atten- tion chiefly to the Ormidum and the Poema Morale on the one hand, and King Horn on the other. The first of the avowed two-stressers to classify L^amon's metre is Trautmann. 2 But his position is anomalous — or was so before 1895, when he changed his entire theory of O. E. verse and became a four-stresser. His first idea was that the Brut gave evidence, not of development of a national form, but of derivation from a foreign source — the iambic dimeter acatalectic of the Am- brosian hymns. To secure the four accents which this verse demands, he was forced to adopt Lachmann's law, though denying its applicability to Old English or to prose discourse. 3 The pecu- liarities in the treatment of unstressed syllables 4 he explained by admitting the continuance of the O. E. practice in that particular. He reached his conclusion as to the source of the verse by claiming, in every important case, the agreement of La^amon's metre with that of Otfrid, which, with Wackernagel, he derived from the Latin. In a later article 5 he suggested that this form of verse in England originated in direct imitation of Otfrid rather than of the church hymns. His recent conversion to the four-stress theory, 6 however, has entirely changed this view of origins. He now derives directly from the Old English the verse which formerly haul's Grundriss, Vol. II, p. 516, also Heath, Phil. Soc. Trans., 1891-4, p. 382. 2 AngliaII, p. 153. s He distinguishes verse-ictus and word-accent, whence it is to be presumed that some of his stresses are subordinate. He uses the same stroke (/) for all alike. See Anglia V. Anz., p. 111. * (a) The omission of unstressed syllables in one or more feet of a verse, (b) The use of two unstressed syllables instead of one. 5 Anglia VII. Anz., 215, « Anglia Beiblatt V (1894-5), p, 87. 4 Studies in Lajamon's Verse. he sought to explain in other ways. 1 It is his opinion of the source, though, and not the characteristics of the metre in question, which is affected by his change of belief. Einenkel 2 supports Trautniann's first view as to the nature and origin of La^amon's verse. The number of cases of inflectional syllables riming either with like syllables or with monosyllabic words, and the fact that the syllables in question are only those regarded by the Lachmann law as capable of bearing accent, are proofs to him of the observance of that law. The rime was devel- oped, he claims, from the riming of end-syllables first, and of root- syllables later. The position of the alliterative letters in the line is an additional proof of the existence of four stresses. Menthel 3 examined 1500 lines from various parts of the poem. The results support Trautniann's statement as to accented syllables, and give the proportional occurrence of each peculiarity. Sweet 4 also follows Trautmann's view in the main. " In the M. E. four-stress metre, syllables that are quite stressless in ordi- nary speech can in verse take the full stress required by the metre." The chief upholder of the two-stress theory of the origin of the metre of the Brut is Schipper. In his earlier work 5 he asserts strenuously that the prevailing rhythm is the two-stressed half-line. Though occasional lines, when read singly, seem to conform to a different rhythm, when read in connection with a long passage they fall into the general scheme. These longer lines, some with three, some apparently with four accents, suggest the possibility that the short rimed couplet may have developed from the two- stressed verse by the emphasizing of a subordinate accent in the several-syllabled Senkung. In La^amon, however, the secondary accent retains its subordinate place. Though the French model may have come before the poet's eyes, it is his own national line that prevails. Long lines with a single alliterating word in each half-line are to be met most frequently. La^amon's work is a 1 Anglia XVIII, p. 96, foot-note. 2 Anglia V, Anz., p. 30. E. E. T. S., 80, pp. xxi-xxxix. 8 Anglia VIII, Anz., p. 49. 4 p. 163. 6 Altenglische Metrik, p. 146. Studies, in Lajamon's Verse. 5 battle-ground between alliteration and rime, but it is the allitera- tion which conquers. Luick, 1 though an adherent of the two-stress view of O. E. poetry, instead of deriving L^amon's verse directly from that source, connects it more closely with the eight-stressed Germanic Gesangvers, which he thinks must have survived in folk-literature. This long line in La^amon and similar verse became broken by the rime into two lines, and thus developed the national rime-verse. It is more measured (taktierend} than the O. E. metre. Though Luick does not expressly accept the law of Lachmann, his system of accenting virtually follows it. 2 Each line contains two princi- pal and two subordinate accents. Their relation to O. E. verse is shown by the fact that they follow five types resembling those of Sievers. 3 There are also lines which may be considered as modeled after the old form. These may be accounted for by the immediate influence of the Old English verse or survivals of it. 4 The rime in La^amon has quite superseded the alliteratiou, which is used merely as ornament. In his second work, Schipper 5 accepts Luick's claim for the grouping of the verse into types based on those of Sievers, but he criticises Luick's formulation of those types, rejecting those which demand a secondary accent, either at the end or within the line, on such inflectional endings as -e, -es, -en, -er. He bases this rejec- tion on evidence drawn from La^amon's rimes, which, except in a few unimportant cases, take account of the stem syllable and dis- regard the endings. Luick's theory of the origin of the verse he also rejects, because no traces of the old Germanic Gesangvers remain, because La^amon's metre can be immediately related to O. E. forms, and because the growing influence of AVace can be traced in the Brut in both verse-structure and rime. The change in Schipper's ideas consists in the admission of a more perceptible French influence, and in the recognition of addi- tional accents in the Old English forms as developed by La3 anion. 1 Paul's Grundriss, Vol. II, p. 994. 2 P. u. B. Beit. X, p. 209. 3 See especially Paul's Grundriss, II, p. 1002. * Luick divides M. E. verse into two classes, one derived immediately from O. E. , the other from the Gesangvers. 5 Wien. Beit. 2, p. 57. Studies in La^amon's Verse. Chaptee II. La^amon's Use of Alliteration and Rime. It is in the rime of the Brut especially that four-stresser and two-stresser alike find their strongest arguments ; and it is this element of the poem that most vitally helps to determine La^a- mon's method of composition and the source of his verse. The length of the work, and the evidence of change in the poet's tech- nique revealed by a mere cursory reading of different parts of the three volumes, necessitate long and careful statistical examination before conclusions can be trustworthy. The weakness of previous theories lies in the fact that they are not based on extensive inves- tigation. The present study aims to show the significance of La^a- mon's use of alliteration, to determine the function of rime in his verse ; to discover to what influences his technique was indebted ; and by deciding, as far as this poem is concerned, the vexed ques- tion of Middle English word-accent, to make as sure a basis as possible for a theory of his metre. The quality of La^amon's alliteration has been noted to some extent by Schipper and Kegel. 1 Strictly correct alliterating long lines (with two alliterating letters in the first, and one in the second half-verse) are to be found, Schipper asserts, " oft genug." Quite as often the last of the three letters is in the last stress of the second half-verse. " Oefters alliterieren auch die zwei Hebungen des zweiten Halb verses mit einer des ersten." " Nicht minder oft " all four of the stressed syllables alliterate, either with the same letter, or with different letters variously placed. The most fre- quent of all is the single letter in each line. 1 Alteng. Met., pp. 150-153. Germ. Stud. I, pp. 172 ff. Studies in La^amon's Verse. 7 Regel's investigation has to do with the occurrence in the Brut of the old alliterative formulas, rather than with the position of the alliterative letters in the line. "With the foregoing statements may be compared the following tabulation and its conclusions : TABLE I. Placing of Alliterative Letters. Number of lines examined, ------- 5, 000 (1000-2000, 2000-3000, 13,000-14,000, 15,000-16,000, 30,000-31,000) Total cases of alliteration, .... 1570=63 per cent. Strictly correct alliterating long lines, ... 75=4.7 " First half-line regular, third alliterating letter on last stress of second half -line, .... 40=2.5 " One letter in first, 2 in second half- line, ... 90=5. 7 ' ' Two letters in 'each half- line, .... 173=11 " One letter in each half-line, 1134=72.2 " These results reveal a notable departure from the O. E. laws of alliteration. So far as the junction of short liues into couplets is concerned, La3amon seems to have felt any combination of alliter- ating letters permissible. The distinction between the first and second half-lines, which in the old verse led to so careful an avoid- ance of more than one letter in the second, he quite disregarded. But with all this freedom, alliteration in the Brut has not forgotten its old function as a means of binding together the two short lines. But to admit this fact is not to decide whether La^amon is to be classed as a distinctly alliterative poet or not. A comparison of the practice of other poets of his own time and of the following centuries will help to establish a conclusion. Studies in La^amon's Verse. TABLE II. Comparison of Alliteration and Kime in Other Poems. Date. Form. .2 » £3 couplets. 22 82 18 Genesis and Exodus, 1250 couplets. 47.2 28 75.2 24.8 King Horn, c. 1250 [ms.] (i 11 21 32 68 Havelok, c. 1280 a 38 41.8 79.8 20.2 Song of Poland, 1430-40 [ms.] a 66.4 66.4 33.6 Pouland and Vernagu c. 1330 [ms.] 121 stanzas 11.2 11.2 88.8 Otuel, c. 1330 [ms.] couplets 17 17 83 Duke Powlande and Sir Otuell, 1430-40 [ms.] 121 stanzas 32.4 32.4 67.6 Arthur, 1428 couplets 19 12.4 31.4 68.6 Sege off Melayne, 1430-40 [ms.] 12 1. stanzas 35.1 35.1 64.9 Le Morte Arthur (Furnivall, 1864), 15 c (?) 81. " 9.5 37.5 47 53 From these results we may see that, while individual poets vary- widely in their use of alliteration and rime, the Proverbs of Alfred and Genesis and Exodus, nearly contemporary with the Brut, have a larger per cent, of alliteration, while the 15th century Roland still has practically as large a per cent. Again, Rouland and Ver- nagu and Otuel, both early, have a small per cent., while the remaining poems, ranging from the 13th to the 15th century, have about half as much alliteration as the Brut {Le Morte Arthur, two- thirds). These poems, it should be noted, are all rimed, so that there can be no question, except in the case of the Proverbs of Alfred, resembling the Brut metrically, as to the subordinate func- tion of the alliteration. As compared with them, the Brut does not use alliteration to a sufficient extent to justify its classification as an alliterative rather than a riming poem. Whether La3amon as a whole preferred rime as a means of uniting the short lines, and how his practice varied in different parts of his work, are shown conclusively in the following table. The estimate, so far as rime is concerned, is somewhat conserva- tive. End-syllable rimes as a rule have not been admitted, except where alliteration seemed to bring the root-syllables also toward Studies in Lajamon's Verse. 9 intentional comparison. A less "conservative treatment would merely emphasize the conclusion. TABLE III. Proportion of Alliteration and Kime. Lines. Allit, without Rime. Allit. Rime. Rime ■without Allit, Neither Allit. nor Rime. Total. Allit. Total Rime. 1- 1,000 162 242 72 15 404 314 1,000- 2,000 122 248 109 15 370 357 2,000- 3,000 98 258 116 26 356 374 3,000- 4,000 73 276 136 12 349 412 4,000- 5,000 60 309 110 10 369 419 5,000- 6,000 65 306 113 14 371 419 6,000- 7,000 80 234 163 18 314 397 7,000- 8,000 54 280 154 9 334 434 12,000-13,000 54 258 174 9 312 432 13,000-14,000 42 248 200 11 290 448 14,000-15,000 57 252 180 6 309 432 15,000-16,000 72 260 158 8 332 418 16,000-17,000 89 288 111 10 377 399 17,000-18,000 46 268 17S 8 364 446 18,000-19,000 52 216 221 10 268 437 19,000-20,000 70 269 147 13 339 416 20,000-21,000 92 263 133 12 355 396 26,000-27,000 34 336 125 5 370 461 30,000-31,000 34 310 150 2 344 460 31,000-32,000 15 309 170 6 324 479 6,761 8,440 The most obvious conclusion from this table is expressed in the two totals of alliteration and rime. The latter is in excess by 1,679 cases. In other words, 67.6 fy of the 20,000 lines show alliteration, while rime appears in 84.4 c / . It is evident where the poet's preference lay. The reason for the difference between this result and the conclu- sions reached by Schipper l and Regel 2 is partly because of the greater number of lines here examined, for both critics show as much latitude in the admission of rimes as was permitted in this ^lteng. Met., pp. 150-153, 2 Germ. Stud. I., p. 160. 10 Studies in La$amon , s Verse. case, Regel even granting the riming of end-syllables. Their results also depend partly on the section of the work from which they drew their extracts. The table shows a curious variation in the proportion of alliteration and rime. While on the whole the alliteration tends to decrease and the rime to increase, the latter reaching its highest point at the end of the poem, the alliteration is at its minimum at lines 18,000-19,000 instead of at the end, where we should expect it. It is by no means easy to determine exactly what La^amon's ear • / accepted as rime. Schipper admits as rimes such cases as layen : / / / , . ! , londe 1 14,339-40 ; scipe : brohte 14,862-3; freondscipe : seoluen 1 5,226-7. But in the following words Einenkel 2 hears rime only in the last or inflectional syllables and not in the root-syllables : hauene : haeleSe 28,432-3 ; ihale^ed : ifuh^ed 29,433-4 ; clerekes : hokeres 29,789-90 ; sechien : susteren 28,782-3 ; iheled : neoSered 29,- 991-2 ; baluwen : ileoten 31,306-7. And these he classes as rimes of monosyllables with inflectional endings : bliSe-mod : iblissed 29,701-2 ; aerd : bidaeled 12,742-3 ; iset : isemed 27,430-1. In his clasification of the rimes of King Horn Wissman allows such words as softe : brihte, to rime on the root-syllables as well as on the endings, though his principles of accent accept the merely inflectional syllable rimes. According to Wissman' s classification, therefore, the words mentioned from Einenkel, as well as those from Schipper, can be regarded as full though of course imperfect rimes. 3 Examples are : pure rimes, riht : pliht 5043-4 ; broker : ofter 5,017-8 ; only root-syllable perfect, Lundene : punde 5,119 -20 ; assonant, orn : nom 5,009-10 ; clupte : custe 5,011-2 ; con- sonant, kene : idone 5,287-8 ; impure vowel and consonant, richer aefter 5,566-7 ; leode : londe 5,239-40 ; monosyllables with root- syllables, Rome-wal ronwalden, folc :comen 5,556-7, blod :bsedde 1 The stroke (/) refers of course to accent. 51 E. E. T. S., 80, pp. xxi fit. 8 King Horn, p. 53. Cf. examples of imperfect rimes in Judith (edited by A. S. Cook, p. lxx), hund:wand 110, herewaethan : onwrlthan 173, fymgeflitu : swyrdum 264, geheawum : beheafdod 289, fleam : eacen 292a 293a, sceacen : feaht 292, sigorlean : geleafan 345, laeg : gesaeged 294. Studies in Lcqamon's Verse 11 18,980-1, munt : strengSe 5,530-3; monosyllables with end-syl- lables, men : biwunnen 5,608-9, mon:speken 31,024-5. Mr. B. S. Monroe, in a recent study, 1 has decided that the spell- ing of the Brut is not phonetic. The diversity of characters in some cases found to represent a single sound, in Mr. Monroe's table, 2 indicates that a pair of words may be far from riming to the eye, or to the ear trained in O. E. and M. E. pronunciation, and yet they may have been true rimes to La^amon. In view of the foregoing opinions, the following tables will be seen to be conservative. TABLE IV. Various Kinds of Rime, 1 2 345 6789 10 11 m Monosyl- Monosyl- - h a 0> a «a »3 11 aj £t= a ot_ a lable and Root. lable and End syl- a. S « £ « a.d."S a o^ 3 --— u V a. a - — a 3 a Ph « 3 < O «u-3 W^ tf Ph oa Ph »-) Ph m 5,000- 6,000 106 44 99 34 25 57 31 3 23 2 2 18,000-19,000 191 29 88 48 20 36 19 9 19 6 31,000-32,000 150 4 113 58 28 37 64 9 9 10 This table is of value as an aid in criticising the computations of rime throughout this work. It indicates where errors may have been made by admitting too imperfect resemblances as rime ; and it also shows that, however rigid the exclusion may be, the number of possible rejections is so small (columns 5 and 6) that they would only slightly affect any given total. 3 That L^arnon's use of rime was intentional and studied, is 1 Studies in the Phonology and Vocabulary of La^amon's Brut Thesis for the Doctorate, Cornell University, 1901. Not published. 2 E. g. O. E. Sound. Lajamon. 3.j 26) CJtj O a(o) a, o, a?, e, ea, eo e, eo e (?o) e, a, le, ea, eo e + g ei ei, sei, ai a a > o a, a?, eo, o 3 Rimes noted in columns 7, 9, 11, have been disregarded in all other estimates in this work. 12 Studies in Layimon's Verse. proved by a study of his interpolations. If much of his rime is due to unconscious rather than conscious imitation, it would prob- ably decrease when he added to Wace material invented by himself or derived from another source, unless, indeed, that source was also in riming form. Extended interpolations, each of 25 lines or more in length, have been examined, selected from 9000 lines of the poem, and forming a total of 1396 lines. Of these 35 °fo are rimed without alliteration. Of the 9000 continuous lines from which these interpolations were taken, only 15 % show non-alliter- ated rime. Scarcely a single line of the interpolations is rimeless. Sometimes an almost continuously rimed passage, with comparatively little alliteration, may be noted in the expansions or additions. The number of j^actically continuous passages throughout the poem, which employ the same method of uniting the short lines, whether alliteration and rime, or rime without alliteration, is not insignificant. It is noticeable that more than five consecutive merely alliterative lines never occur, while there are passages of from fifteen to forty alliterative-rimed lines, and some merely rimed passages from ten to fifteeen lines in length. The longest con- tinuous passages are at the end of the poem. TABLE V. Passages Connected by either Eime or Eime and Alliteration. Aim, without Rime. Alliteration and Rime. Rime without Alliteration. Total No. connected passages. LINES. 3-5 lines. 3-8 lines. 8 + lines. Total. Number lines in longest passage. 3-8 lines. 8 + lines. Total. No. of lines in longest. 1- 1,000 14 32 8 40 (14) 2 2 (4) 56 1,000- 2,000 13 26 5 31 17) 30) 10 10 (4V 54 5,000- 6,000 2 23 12 35 12 12 (4) 49 12,000-13,000 14,000-15,000 1 2 32 32 7 8 39 40 16) 16 19 36) 40) 22 21 4 3 26 24 (10) (15) 11 12) 17 66 66 18,000-19,000 26,000-27,000 2 26 15 5 14 31 29 30 11 6 1 36 27 69 41 31,000-32,000 20 15 35 ( 25 4 29 64 A theory of the process by which Laaamon gained his technique has been growing with eveiy new detail in this investigation. A Studies in Layamon's Vertitr 13 further examination of one of the classes of rimes in Table IV will advance far toward the conclusion. The single class of rimes not noted by the critics quoted on p. 10 was that in column 6 — words riming only through alliteration. The ground for consider- ing such words rimes at all is their frequent occurrence, especially in the first part of the Brut, and evidently of set purpose. TABLI ! VI. Kelation BETWEEN ALLITERATION AND KlME. LINES. Number cases allit. of last words. Number good rimes among these. Per cent, good rin among last lette allit words. 1- 1,000 99 38 38 per cent. 5,000- 6,000 88 37 42 12,000-13,000 43 16 37 18,000-19,000 46 28 60 " 26,000-27,000 92 69 75 " 31,000-32,000 62 53 85 As we have already seen (Table III), rime was increasing from the beginning to the end of the poem, while alliteration was decreasing. The percentages in the last column of the above table indicate how one form was merged into the other. It was partly through the capping of the half-lines by words alike only in their initial letters that L^anion's conception of similarity of sound was developed, and his skill in the use of it was gradually attained. These statistics of rime and alliteration give many indications of the process by which La^amon learned his technique. Of the alliterative possibilities of his own language he availed himself with careless freedom, perhaps often almost with indifference, for the other similarities of sound were a greater pleasure to his ear, and it was these on the whole that he employed with the most con- scious effort. The laws of rime he knew at first very imperfectly. The acquaintance with the Latin hymns and with the rhetorical writings of Bede, which may have come to him as a part of his priestly training, had not taught him these laws, nor had the Welsh poetry that was probably often heard in the region around the Severn. The change in his methods during the course of the 14 Studies in La^amon's Verse. Brut controverts the theory of the survival of a national rime- verse. If there had been such a verse, as Luick believes, La^a- mon would have known its form from the beginning. But he had to learn to rime, as the early men in every nation learn, feeling through assonances and stumbling over consonances, not unsatisfied with a half-word rime, and sometimes bringing more than two words into a pleasant jingle of similarity. The rate of his progress is indicated in Table III, in the numbers registering his use of alliteration and rime throughout the poem, and in Table V, which tells how he gradually came to use longer passages joined through- out by the same means of union. It will be noticed that the difference between the same successive thousands is not great. It will also be observed that the most noticeable point is that represented by lines 18,000 to 19,000. Table IV shows that the highest number of pure rimes occurred there. Table V records that the number of continuous passages of the same form as to rime and alliteration also reached its maximum in that part of the poem. This peculiarity invites further investigation. Certainly it might be expected that rime should increase in amount and in excellence to the end of the poem, and that alliteration should decrease. Three explanations suggest themselves. 1. Possibly for the first 19,000 lines La3amon's effort to acquire skill in the use of rime was relatively great. Then, feeling fairly satisfied with the result of his labors, he may have relaxed his efforts somewhat, with the result shown in the tables — an increase in the alliteration, and a decrease in the excellence, though not in the amount, of his rime. 2. Or possibly he did not write the poem in consecutive order, and liucs 18,000-19,000 represent his latest and therefore most excellent workmanship. But this supposition is weakened by the testimony of Table VI, which by the regularity of succession indicates continuity of composition. 3. Or possibly the prepon- derance of rime in this particular place may give some clue to the character of the unknown sources of the material which La^amon added to Wace. The passage in question deals with the Arthur story, and it is to this story that La^amon made the most signifi- cant additions. It is not the province of this study to pursue Studies in La^amon's Verse. 15 such an investigation. It is bound, however, not to disregard the fact that the unknown sources may have been Celtic, 1 and that Welsh verse was intricately rimed. 2 All the statistics indicate that the evidence of rime as to word- accent may be accepted with a fair degree of confidence, since La^amon used rime consciously, not as a mere ornament, but as a means of joining his couplets. What information may be obtained from the rime as to the treatment of the end-syllable ? Previous investigators, as we have seen, are at variance. Einenkel 3 states that they occur frequently ; Schipper, 4 that they are very infrequent and open to question ; Menthel 5 asserts that they form 25 per cent, of the rimes at the beginning, and 50 per cent, at the end of the poem. Table YII confirms the fact of the increase, but differs from Menthel's percentages. These rimes form only 3^ per cent, of the rimes in the first thousand lines, and 13J per cent, of those in the last thousand. TABLE VII. Rimes of Inflectional Endings. LINES. Rimes of Inflec- tional syllables. Monosyl. with root of dissyl. Monosyl. with inflect. ending. 2 root syls but not endings. 1- 1,000 11 29 7 39 5,000- 6,000 16 13 4 44 12,000-13,000 12 11 4 47 18,000-19,000 19 23 6 29 26,000-27,000 18 7 11 23 31,000-32,000 64 9 19 4 The conclusion must be stated with a large if. The number of apparent inflectional-ending rimes is so small that it is by no means certain that La^amon intended them as rimes at all. If he did not, the conclusion as to word-accent established by this table is only strengthened. If he used them consciously, their decrease from the beginning to the end is very significant. When he was yet unskilled, and was willing to accept almost anything as rime, iWiilcker, P. u. B. Beit. Ill, p. 524. Stephens, pp. 484-9. 4 Wien. Beit. 2, p. 63. 5 E. E. T. S. 80, pp. xxi ff. 5 Anglia VIII, Anz. p. 49. 16 Studies in La^amon's Verse. it seldom occurred to him to rime inflectional endings. If these endings were ever accented in ordinary speech, or in O. E. verse and its descendants, with which he was certainly familiar, it would seem strange that he did not more readily avail himself of these very easy and numerous similarities. He at least did not develop his knowledge of rime by using them, as Einenkel believes. 1 He seems to have had a marked preference, at first at any rate, for root-syllable rimes over other irregular forms. It is not difficult to account for such an increase in end-syllable rimes without resorting to Lachmann's law. The change in accent tendency, by which these rimes would come more and more into the poet's use, may well be due to the French example. That L^amon should have perceived the possibility of throwing an accent upon the last syllable of English words, as he saw it so constantly done in French, is not unnatural. The conclusion to which this tends is that the Schipper theory of word accent is the one most nearly in accordance with L^arnon's practice. Weak inflectional syllables were not stressed in daily speech, whatever may have been the quantity of the preceding syllables. The prominence which they sometimes acquire by their employment in rimes is only momentary, and cannot form the basis of a theory of metre. 1( 'The rhymes upon the accent or minor tone no doubt first came into exist- ence." E. E. T. S., 80, p. xxxix. Studies in Lajamon's Verse. 17 Chapter III. The Metrical Structure of the Brut. The testimony of the rimes of inflectional endings sets aside all theories of L^anion's verse except Schipper's. 1 If La^amon avoided those rimes in the early part of his work, it was because he was not accustomed to stress the weak inflectional endings, -e, -en, -es, -ed, in his ordinary speech or when he recited poetry com- posed in the national metre. Since there is no proof in the Brut that these syllables were usually accented, and since, as Schipper makes clear, 2 the evidence of other Middle English verse is equally against accenting them, we shall hold the point settled. The Brut, then, is not written in four-stressed verse, for such readings as these of Trautmann and of all four-stressers are impossible : and sende 1 his sond6 25,341 burhcostned mid wepnen 25,440. 3 It is not written in verses containing uniformly two principal and two subordinate stresses. Whenever Luick's types necessitate the accenting of the inflectional endings, those types must be rejected ; thus : A (x)an kinge leof. 14,049. "With extensions to A'* (x)z(x)xxi(x)x(x) ]>e king sone up stod. 14,073. C* (x)xx^xx nes ]?er nan cristindom : Ca'* (x)xx^x^xx J?er ]?e king ]>at maide nom. 14,387-8. Schipper gives several additional examples of most of the types, which furnish some opportunity for criticism. It is not easy, for instance, to see why kinge in the above E* is subordinated : Haengest wes J>an kinge leof is quite as logical and probable a reading. Or why stod in A* is given a slighter stress than up. Alliteration suggests, J>e king sone tip st6d. Holden runinge (14,070) is read as a C*, and neowe trSenden as a D*. To make a distinction between these two lines is to split hairs. The follow- ing lines, classified by Schipper under A*, in Old English would be simply A ((x)^x(xxx)^x) : forS to )?an kinge (13,812), ne mihte we bilaeue : for liue ne for daeSe (13,875-6), J?at f6lc is isomned (13,856), and mid him brouhte here (15,088). It would not be difficult, but it might be unfruitful, to multiply illustrations like these. Enough has been said to indicate the danger of the over-enthusiastic application of new types. Schipper's forms do apply to La^amon's verse, as figures will presently declare (Table VIII), but they are not nearly so large an element in the Brut as their author would have us believe. There is a further reason for the reduction in the number of 1 Wien. Beit. 2, pp. 67-9. Studies in La^amon's Verse. 19 these types. Three of them must be accredited to their original formulator and reckoned as Old English. A*, C* and Ca* and D* are almost identical with Sievers's A*l, ^ix lx, C*l, ixllx and C*2 xux ^x.* The only diiference is in the. extra syllables marked (x) by Schipper. They are of course not rare in O. E. 1 The absence of a prepossession against O. E. forms, and a natural reading of certain classes of lines, together with a careful scrutiny of the lines which Schipper admits to be Old English, reveal these flaws in his summarizing classification : 1. Some two-stressed lines of the O. E. type are rimed : A Heo fusden from stronde C vt of Griclonde. 1109-10. A & alle bat bi-houeS ; A ba scipen to driuen. 945—6. 2. Many rimed lines with masculine ending have only one substress : riden benne ilke waei : be fdrx xOx Lx Lxl L D4 ^ | ^xi C*3 ixi | Cx D*3 :x ^x^ 20 Studies in La^amon's Verse. Schipper as containing two principal stresses (and presumably no substresses). The recapitulation of the Sievers types shows how incomplete such a classification is, in disregarding many of the substresses recognized in O. E. It also leaves entirely out of account the whole hypermetrical class (Schwettvers). These forms are so important that it will be well to state them in full. The following formulas are those recognized by Luick, Sievers and Schipper : ' / / AA sx . . . cx±x weaxen wltegbrogan. Gen. 45. A2A /x^x^x wserfaest willan mines. Gen. 2168. / v / i A* A /.xx^x.^x arleas of earde Jjinum. Gen. 1019. AB £.x . . . £.x.£ waesceth his warig hrsegl. Gn. Ex. 99. AC lx . . . llx wlftige to w6ruldnytte. Gen. 1016. AD lx . . . ^ix &&alde 6yrnwiggende. Jud. 17. AE / x . . . ^xx.i sweord and swatigne helm. Jud. 338. / / BA x.lx . . . lx.lx auyrged to widan aldre. Gen. 1015. BB xlx . . lx.l oferciimen bi]> he aer he acwele. Gn. Ex. 114. / / BC x . . lx . . . llx and nahte ealdfeondum. Dan. 454. / _ / BD x.lx . . llxx a^edon hie "Seer fimwerigne. Kr. 63. CA xclx.lx geseo'S sorga mieste. Crist, 1209. CC x . . . £.£.ox ne se bryne beotmaecgum. Dan. 265. CD xx^^ix )>& he J?yder folc samnode. Dan. 228. CE xi^ixi forj?6n waerlogona sint. Gen. 2409. i i BE xl . . . l\x.l ne }?earf he J?y edleane gefeon. Gen. 1523. (and a few others occurring very rarely.) Heretofore no one has applied these hypermetrical types to the Brut. But an examination of La^amon yields such examples as these : AA twd and feowerti wmtre. 10,232. 1 P. u. B. Beit. XII, p. 455 ; XIII, p. 389 ; XV, pp. 360-441 ; Sievers's Alt- germ. Met. 88-96 ; Wien. Beit., pp. 50-52. Studies in La^amon's Verse. 21 A2A (probably) G6dlac king on ueste. 4,526. AB alle )>at 3irnden his grrb\ 10,305. AC J>e gauel of Bruttlonde. 10,503. AD swiken )?ene king Basian. 10,615. AE & nom him oSerne cure. 11,255. BA ]>a ]?at child was of ]?rittene 3ere. 11,078. BB bitahte ]?an maidene an hond. 10,915. BC ]?at dursten him cZerf makien. 10,943. BD J?at ictimen was per Gillomar. 18,079. CA & $ef ^iues swi«e gode. 10,299. CC moni lond timbe-rowen. 114. CD Jns w6rd com to Oarrais. 10,634. BE & muchele monscipe biwon. 345. These types satisfy 10 % of L^arnon's lines, at a conservative estimate, and accepting the new Schipper types whenever they can be applied with a fair degree of ease. The following tabulation makes a conservative count of the Schwellvers. For the sake of comparison, the numbers of the Sievers O. E. and Schipper types are also given. The B and B* types are separated from the others in order that Schipper' s sum- maries, with their distinction between masculine and feminine lines, may be accurately tested. TABLE VIII. The Proportion of the Metrical Types. Sievers. Schipper. Schwell- ( > VERS. Simple. Expanded. Lines B Rest B* Best B* Best 1- 1,000 58 321 83 396 25 24 95 10,230-11,230 18,000-19,000 19,000-20,000 31,000-32,000 44 30 34 26 351 382 389 324 77 59 68 68 244 167 176 217 40 48 39 32 112 113 165 99 157 99 205 90 But there is reason for believing that this estimate of the pro- portion of the Schwellvers is too low. These types may be sub- 22 Studies in Laqamon's Verse. stituted for the Schipper types in many cases with great ease, and often in the same way for the expanded Sievers types. Such a substitution is eminently reasonable. It helps to do away with the subtleties of sub-stresses, and thus maintains the prevailing simplicity of the verse ; 1 and it emphasizes the strength of the Old English element. The reconsideration of the Schwellvers gives this result : TABLE IX. The Proportion of the Schwellvers. Links. Probably. Possibly. 1- 1,000 325 380 10,230-11,230 378 444 18,000-19,000 430 493 19,000-20,000 385 452 31,000-32,000 450 39.3 1 500 1,968 = 2,269 = There is further support for the Schwellvers scansion in the position of alliterating letters. It cannot be proved, of course, that an alliterating syllable in the Brut was a strongly stressed syllable. But in view of the large share that alliteration claims in this poem — a share so large, as we have seen, that students have been misled into believing it more potent than the rime — and in view of the prevailingly Old English character of the verse, it is extremely probable that La^amon used his alliteration to mark the most important words, instead of those with merely a substress. In the light of this opinion, the scansion of the following lines is clear : Nv ser8 mid Z6ft songe : be wes on feoden preost. al swa be 66c spekeS : be h6 to &isne in6m. 68-71. ful well pan like stade : 1 It i6 always possible to find diversities of metre in long poems. A. J. Ellis (Trans. Phil. Soc., 1875-6, p. 445) found 45 signs necessary for his description of modern blank verse. Schipper (Wien. Beit. 2, pp. 48-49) says that the Schwell- verse are not always to be distinguished with certainty from ordinary verses with long Auftakl. Decision depends on general verse-rhythm. Studies in La^amon's Verse. 23 bar R6me nou s£6nde3. 124-5. lie rnakede emie .sfronge castel : mid sMrke ston walle. 188-9. Cnihtes ^linnen riden : g&eres ^uunen gliden. 19,550-1. ba sae heoni saetten a that strond. 19,916. be /6lc fit of 16nde : /lah on delche deude. 31,845-6.* TABLE X. Schwellvers Established by Alliteration. Lines. Cases. 1- 1,000 73 10,230-11,230 116 18,000-19,000 66 19,000-20,000 83 31,000-32,000 141 479 It is obvious that, in order to furnish proof of a scansion of three strong stresses, there must be either three alliterating letters in a line, or two such letters and an unmistakably strong non-allit- erating syllable, or one alliterating letter and two other strong syllables. A glance back at Table I will recall how few, compara- tively, are the cases of even double alliteration, and will therefore emphasize the significance of the total stated above. All the probabilities, then, point toward the Schwellvers as an important element in the metre of the Brut. In the use of the hypermetrical lines the Brut is not unique, but has its place among other M. E. poems, in a definite succession from O. E. times. Luick classes it metrically with the Juliana, Marharete, Caterina, and Hali Meidenhad. He does not detect the presence of the Schivettverse in these works, but it is in this 1 The distinction in the placing of the letters in the 1st and 2nd half-lines, observed in O. E. verse, is not found in La;amon. His Schwellverse show the same licenses as the rest of his lines in this respect. 24 Studies in La^amon's Verse. element that they furnish the most significant groimd of compari- son with the Brut. The last poetic passage in the A. S. Chronicle has its place in the series, and the Proverbs of Alfred also. 1 TABLE XL S The Metricax Types en M. E. Poems. Brat. Alfred's Death. Marhar- ete. Juli- ana. Prov. Alfred. Cater- in a. Hali Meid. 0. E. types 16 242 238 176 182 186 Expanded fl 12 156 141 178 161 159 Schwellverse 6 = 102 = 121 = 146 = 152 = 155 39.3 1 to 17.6 i 20.4 * 24.2 i 29.2 i 30.4 i 31 * 45.4 The Schwellverse in these poems still tend to occur in groups, especially in the Brut, though they are often isolated. In none have they so artistic a function as Dr. Foster has shown that they exercise in the Judith. 4 But they spring more out of a syntactical necessity, in consequence of the more expanded forms of expres- sion that the language was developing. In La^amon, as we shall see, there was an additional reason for using extended forms. The types of Schipper, while they have not been entirely set aside, have been seen to hold a much less prominent place than he has claimed for them ; and the presence of the Schwellverse has been established. The new classification is as follows : 1. Two-stressed short lines of O. E. type, either merely alliter- ated, or rimed, or rimed with alliteration ; 2. The same extended, a. by one sub-stress as in O. E. expanded lines, and fre- quently by additional unstressed syllables also; 6. by two sub-stresses (Schipper's A'*, B*, C'*, E*) ; 1 The numbers of Schwellverse, in order of proportion, in O. E. pooms in which they are of most frequent occurrence, are as follows : Gn. Ex. 68 ; Eood 34, about 21.52 f ; Judith, 65.5, about 18.5 * ; Daniel 49 ; Christ 37 ; Genesis 31.5 ; Guth- lac G. 29 ( see Sievers, P. u. B. XII, p. 454 ; Foster, Judith, p. 35, foot-notes) . ' In all the poems except the first, which has only 34 lines, 500 lines were counted. 8 Expanded Sievers and Schipper types. As in the case of Lajamon, the major- ity of these might be read as Schwellverse. 4 Judith, ten Brink's Quellen und Forschungen LXXI. Studies in Layamon's Verse. 25 3. Lines of the O. E. Schwettvers type ; 4. Lines of four stresses not included in the foregoing. Chapter IV. The Influence of Wace. Up to this point, the influence of Wace has been mentioned in a general way as affecting La^amon's word-accent and his skill in rime. The first of these modes of operation must remain merely a matter of probability. The second may be demonstrated by a line for line comparison with the French original, as well as by the numerical increase in rimes already discussed. La^amon sometimes uses the very rimes of the French. This is notably the case when he simply transcribes a list of proper names : Brutus Uaert Escut. Margadud : Sisiluius. Regin. Bladud. Moruit. Lagon. Ebedloan : Ricar. Spaden. Gaul. Pardan. JEldad. Gangu. Xerin. Luor : Rue Assarac. Buel. Hector. 2,693-8. Cf. Wace, 1,581-6. See also La3amon, 2,703-12, and Wace, 1,589-8 ; La^amon, 5,259-64, and Wace, 2,913-18 ; La3amon, 24,331-2, 24,335-6, 24,343-6, 24,355-8, and Wace, 10,531-6, 10,549-52 ; La^amon, 24,399-24,404, and Wace, 10,587-90, etc. There are instances of other kinds of similar rimes : Lauine hehte his leuemon : J?ene castel he clepede Lauinion. 190-1. De La vine posa le nom Si l'apela Lavinion. Wace 71-2. heo beo^ to gadere icumene. 456. Se sont jost^ communement. Wace 232. In the nature of things these instances cannot occur often, 26 Studies in La^amon's Verse. La^amon too carefully refrains from the use of French words. In the 5,000 lines examined, only eight cases have been observed, in addition to proper names. But the fact that they are to be found at all shows clearly enough how La^amon learned riming. Similarities of sound in other positions than the ends of lines, produced by the repetitions of words or syllables, are found in La^amon and in Wace. Instances from Le Roman de Brut have been noted and classified by Keller. 1 He finds many examples of the recurrence of the same subject, verb, object, and phrases in the same sentence ; of repetitions of almost entire sentences ; of the same word in different constructions, and of the same root-sylla- ble in different words. These repetitions may occur in the same place in successive lines, or they may be arranged in chiastic form. La^amon's translation frequently reflects these peculiarities. His imitations as a rule pay little regard to the office in the sentence held by the repeated word in the original, therefore minute classi- fications of these phenomena are not desirable. He is content with the presence of a repeated sound in a translated passage, and often, when the recurring sound is inconvenient to manage in his corres- ponding sentence, he uses the device some lines farther on. In the instances most nearly approaching the French, the repe- tition of the verb is of frequent occurrence. Bi-leaf ]?e treuwe }>e bi-lef ]>ene a3~S. 4,340. Romp l'aliance, romp le foi. W. 2407. In his translation of the following, La^amon outdoes Wace (see 1. 2,001-17) for he repeats bi heold he eight times, thrice in con- secutive lines : Brutus esgarda les montaignes, Vit les values, vit les plaignes, Les marines et les boscages Et les 6ves et les rivages ; Vit les cans et les praaries, Vit les teres bien gaagnies, Et la terre bien avoier Et le pueple monteplier. W. 1,245-52. 1 Maistre Wace, eine stylistische Untersuchung. 1886. Studies in Lajcimon's Verse. 27 In the following, La^amon's restraint in the use of verbs is balanced by the parallel structure of the last two lines : Nu fusen we horn to : & stsercliche heom leggen on. & wrseken wunderliche : » ure cun & ure riche. & wreken }>ene muchele scome : J;at heo us iscend habbeoS. J>at heo ouer vSen : comen to Derte-muSen. & alle heo beo$ for-sworene : & alle heo beo$ for-lorene. 21,177-86. Vengies vos amis, vos parens ; Vengies les grans destruimens, Vengies les pertes, les travax Qu'il nos ont fait par tans assax Jo vengerai les felonies, Jo vengerai les foi menties ; Jo vengerai nies ancissors Et les pertes et les dolors, Et vengerai la revenue Que il ont fait a Destremue. W. 9,558-67. The common Old French maint La^amon renders in : moni heaued moni hond : fallen to foten. monie ]^ar fuhten : monie flaeni makeden. monie ]?ar feollen : 574—8. Maint home ester, et maint abattre, Et maint fair et maint combatre : Maint colp i recoivent et rendent, Li Troyen parmi les fendent, Maint en ont mort et abatu, Maint en ont pris et retenu ; Retenus fut Antigonus. W. 297-303. 28 Studies in La^amon's Verse. In the following instance the syntax of the repeated word is pre- served : grift he wolde lumen : nnriht he wolde scunien. grift he wolde habben : grift he wolde holden. 15,128-31. Pais desiroient, pais voloient, Pais amoient et pais querroient. W. 7,393-4. And in this the change is simply from adjective to adverb : J?at fseht wes swufte strong : & swufte stser & swufte longe. 4,170-1. Maint colp i ot pris et don6 Et maint home mort et navr6, Maint escru frait et depecie\ W. 2,283-5. The closest parallelisms are found in the case of entire lines in which, though no particular construction is necessarily copied, the effect of the whole is that of similarity : heo hefden wind heo hefden water : J>e heom wel ferede. Jmtti dawes & ]?ritti niht : heo ferden efer forft riht, 1,273-6. Taut ont siglS tant ont nagie - : Tant ont eu ore et vent, Que en trois jors rooidement De si qu'as pors d'Efrise vinrent. W. 704-7. Weore hit rih weore hit woh : 6,373. U fust a tort, u fust a droit. W. 3,430. A noticeable repetition of some sort is likely to be found in La^amon anywhere in sight of such a device in Wace ; as : muchcl dom, much el dune : rnuchel folkes dream. 1,009-10. Studies in La^amon's Verse. 29 immediately preceding his translation of Bien dit, bien dit, ce dient tuit. W. 560. These anaphoristic repetitions may not perhaps properly be called rime, but they are evidence that L^anion's ear delighted in similarities of sound, and that he was quick to use opportunities to secure them. True internal rimes are made by Wace, and imitated by La^a- mon : Armes quisrent et robes prisrent, Maisons arsent, homes ocistrent ; W. 9,474-5. Tant fist, taut dist et porcaca, Et tant pramist et tant dona, Et tant requist, et tant proia, Al roi Artur se concorda : W. 10,134-7. & Cestesburi castel : an Waladunes dune. 2,822-3. to munien his ikunde. 2,033. Jwlede ich on folde. 2,287. & ]>er he wonede in ]?on londe. 2,526. to somne heo comen sone. 19,138. Very common in L^amon are what Regel l calls Formeln des Binnenreims, and Guest 2 Sectional Rime ; such as widen and siden, grith and frith, &c. Various forms of chiastic rime are frequent in Wace, and these, too, are noted by the English poet : Mult me desdaigne, en mervillant, Et me mervel, en desdegnant. W. 10,923-4. Ne nul fors un, n'en sai nomer ; L'un sai nomer, ce vous puis dire. "W. 1,068-9. ^.174. 2 p. 122. 30 Studies in La^amon's Verse. Goruois un quens Cornwalois Mult prous et saiges et cortois. W. 8,689-90. Cors contre cours bataille enprist ; Chevalier ert vaillans et fors. W. 60-1. & swerie rne seftes : ]>e se^ heom seel iwurSen. 5,448-9. J>a Breunes hauede ihirde : his hirde-manne lare. 4,408-9. & ssei'S J>at he awundred is : wunder ane swrSe. 24,775-6. J>a clupede J>e king : kenliche lude. 21,295-6. he ferde ut of Cantuarie burie : mid balden his ferden. 7,438-9. Continuous rime, extending sometimes to ten lines or more, is too common in Wace and Old French poetry in general to need exemplification. Shorter series of assonantal end-words are to be met in the Brut also, and too often to seem quite accidental : & }if Jm Jus nult don : Jm scalt wurse underfon. for Oswy is a swulc mon : J>ine scome he wulle don. 31,583-6. Oswi hafde ernes sunen : J>e weoren swrSe prute gumen. and ma of his cunne : J>a weoren mod-fulle. 31,461—4. agaen he gon wende : in to ]?isse londe. and in ]>an norS ende : sette J>ene king Penda. to fleomeu Oswalde : ut of ]>issen londe. 31,347-52. Studies in Lajamon's Verse. 31 It is true that many of the peculiarities of La^amon's rimes are not unknown in Old English verse. The Formeln des Bin- nenreims are common there. Initial rimes, inverted rimes, and the riming of a word at the beginning of one half-line and the end of the next, are all noted by Kluge. 1 It is very possible that La^amon derived these elements of his technique from the litera- ture of his own country. But since these peculiarities are of such frequent occurrence in Wace, and are in many cases so directly imitated by La^amon, the conclusion seems inevitable that here at least he was indebted to the French. " Er lernte gar von Wace manche technische Neuerung, uberbot ihn sogar in anaphorischen Wiederholungen." 2 What effect had these imitations upon La^amon's metre ? Can an ear so sensitive as his have failed to be impressed by the metri- cal quality of the verse he was so constantly reading? All the forms of his lines have not been accounted for by referring them to O. E. progenitors. The line of four beats sometimes cannot be made to conform to an O. E. type. Schipper's modifications will not always satisfy it either : Ten ^er heo wes mid Locrine : ofte heo haefde seorwe & pine, fiftene }er and ni^en d&waes : 2514-16. ibrout ich habbe ]?es kinges brother. 725. $eouen niht & senne daei : )>e king swv&e seoc lai. 11,012-3. d\ Jjat J?et child mid seyen isseh : 11,062. and J?er wes Elene ]>e halie quene : 11,148. fe king lette wtirchen tw6ien imaken : 18,206. Nti wes ArSur god king : 19,960. d\ to-gadere M. wines lond. 31,295. 2 p. 422. a Ten Brink. Paul's Grundriss, Vol. II, p. 622. 32 Studies in La^amon's Verse. Penda his *w6ord Ht a-droh : 31,425. Single stresses in some of these lines might be subordinated, though usually the alliteration warns against such a change ; and even if it were made in spite of the alliteration, the result would not exemplify anybody's types. Whether masculine or feminine, the above verses are most easily scanned as simple four-stressed lines, irregular, to be sure, both in the number of their syllables and in the position of their beats, but in both these respects quite in accordance with the standards of Old English verse. They present, however, Old English standards applied to a new metrical ideal — French octosyllabics. If proof of this be needed, it may be obtained by noting the metre of La^amon's line for line translations from Wace. Some of these have already been quoted (pp. 25, 28). A reperusal of them discloses the fact that the majority of the English lines are four-stressed, whether or not they conform to recognized types. The list of names transcribed directly from Le Roman de Brut (see p. 25) indicates that La^amon did not shun the metre of his original. The exact number of line for line translations in 5,000 lines, and the number of expanded lines of all types among these, are noted in the following table : TABLE XI. Line for Line Translations. Lines. Total Transl. Expanded. 1- 1,000 11 6 1,000- 2,000 99 76 10,230-11,230 16 13 18,000-19,000 17 14 19,000-20,000 11 9 31,000-32,000 9 9 The smallness of these numbers is not surprising, when it is remembered that L^amon's constant practice was to draw out and augment Wacc's matter in every way, by interpolations from other sources, and by splitting up and expanding the octosyllabic lines. The number of considerable interpolations, according to Madden, in the 60,000 lines, is recorded in the following table : Studies in La^amon's Verse. 33 TABLE XIL Intekpolations. LINES. 1- 1,000 45 ( excl. introduction ). 1,000- 2,000 39 10,230-11,230 281 18,000-19,000 356 19,000-20,000 691 31,000-32,000 514 Total, 1,926 This constantly increasing amount of new matter as constantly contracts the field for parallelism in translation. The approach to the forms of Wace, however, is discernible in some extended pas- sages that only approach Wace's syntax. These are naturally in the first part of the poem. Pp. 54-5, 84, 114, 115, 116-7 and 123, all in Vol. I., contain examples of these long lines. La^amon's more usual practice is to follow the French closely for a few lines, and then to drop into his shorter, more familiar verse, probably because he finds the long line difficult to maintain. The following passage, with its original, illustrates this : Li rois a le brief escot6 ; Grant merveille li a samble Que li Troyen se revelent Et que de francise l'apelent. Fol hardiment, ce dit, ont pris ; Et en fole oevre se sont mis. Ses dus, ses princes, ses barons Et tos ses homes a somons. Gent a ceval, et gent a pie Vers eels de Troye a cevalcie. W. 253-62. J>e king nom ]?at writ on bond : & he hit wro"Sliche bi-heold. seolcuS him Jnihte : swulcere speche. Jm he alles spac ; 34 Studies in La^amon's Verse. mid Jn'jete he spilede. To wro]>er heore hele : habbeft heo such were idon. Mine Jn'alles i mire }>eode : me suluen ]n'etia$. wide he sende ^eoud J?at loud : for he was leoden king. ]>at come to hirede : riche men & we^len. al ]>at wapmon-cun : )>a mihte beren wapen. vppen lif & uppen leomen : al ]>es londes folc. An horsen & an foten : fortS heo ifusten. 484-503. To appreciate the extent of La^amon's imitation of the French verse, a clear idea of the character of that verse is essential. Schipper l holds that, like the mediaeval Latin verse from which it was developed, it is accentual. "In der mittelalterlich-lateinischen Poesie, sowie audi in der romanischen, ist . . . . eine regelmassige Aufeinanderfolge von starker und schwacher betonten Silben oder von Hebungen und Senkungen Gesetz, die beide vou gleichen Wert fur den rhythmus sind." Quoting two couplets of Wace as illustration, he concludes : " Wir haben hier ein Versmass von im Ganzen jambischen rhythmus vor uns." Wace's editor calls the verse syllabic rather than accentual ; and most of the proso- dists agree that, though closely related to the accented iambic dimeter, by Wace's time the syllabic character of French verse was established, and that the only necessary stresses were those on the last syllable and at the caesura after the fourth syllable. But there are some differences of opinion as to the last point. Tobler's views on the subject are as follows : "Mais la versification francais n'a jamais connu un principe semblable au principe fundamental de cette espece de vers latins qui exigeat que des syllabes tonique alternassent invariablement avec des syllabes atones." 1 Wiea. Beit. 2, p. 79. 2 Studies in La$amon's Verse. 35 " Naturellenient, vu son peu delongeur, ce vers [le vers octo- syllabique] n'a pas de cesure ; on bien il faudrait, pour lni en trouver, reunir des vers de quatre syllables qui se suivent, rimant deux a deux, de maniere a faire des vers de huit syllabes avec rime interieure : .... il est fort douteux que meme pour la periode la plus ancienne, la cesure dans le vers de huit syllabes ait 6t6 autre chose que l'effet du hasard, ou plutot autre chose qu'un produit direct de la nature du vers et du langage." l But Gaston Paris affirms 2 " 1' existence d'une ensure, a l'epoque primitive, dans les vers de huit syllabes .... elle ne peut, a mon avis, s'expliquer par le simple rhythme de la langue elle-meme, comme M. Tobler est porte a le croire, car ce rhythme n'a pas changS depuis le xn e siecle, et cepeudant deja dans AVace on ne trouve plus trace de l'accentuation reguliere de la quatri^me syllabe du vers octosyllabique. Cette accentuation reguliere se retrouve dans les vers latins rythmique qui correspondent a nos octosylla- biques et n'est qu'un reste d'une periode plus ancienne, ou l'alter- nance des syllabes atones et toniques a du 6tre beaucoup plus constante." "Wace's line, then, had not more than two stresses, and was mainly syllabic. There is no evidence that La^amon perceived the syllabic char- acter of Wace's verse. Certainly he did not imitate it, for his own verse does not improve in syllabic regularity. The largest number of couplets in which both members contain the same, or nearly the same, number of syllables, is in the first 1,000 lines, where there are 412 such couplets; while in the 11th thousand there are 312, in the 19th 312, in the 20th 341, and in the 32nd 340. However the line may have sounded to French ears, it was its iambic quality, or rather simply its capacity for receiving more stresses than the O. E. line he knew, which struck La^amon's attention and aroused his instinct or his desire to imitate it spas- modically. Of metric regularity in any other than the O. E. sense he had no conception ; therefore, as we have seen, many of the lines traceable to T^ace may be designated by a modified form of an O. E. type. It is significant that the scribe or scribes who copied 1 pp. 123, 125. * Preface, p. xi. 36 Studies in Layamon's Verse. his work avoided the long lines, perhaps because they were unfamiliar with the French. The younger MS., Cott. Otho, C. XII., shortens 17.5 fo of the lines, while it lengthens only 6 f , and one half of these are produced by combining two or more lines, in accordance with a general process of shortening the work. It may be that scribes of the older MS. did the same thing, and that the original contained more long lines than survive for our counting. It is clear, then, that the influence of Wace on La^amon is noticeable, and that La^amon was often conscious of it. Its effect was the development of rime and consequently the decrease of alliteration, and the occasional lengthening of the Old English line to one of four stresses. It is also clear that to imitate the French verse constantly and consistently was not in the power, or accord- ing to the desire, of the English poet. His sense of the Old English verse was on the whole the stronger. Conclusion. The verse of La3amon presents an interesting spectacle of the unfused mixture of the old and the new, of the activity of instinct and purpose, habit and conscious effort. The old verse forms and the old alliterative formulas are stamped upon his expression, but they take shape in his writing with no strict sense of the laws of their ancient usage. As he ponders over Wace he learns what rime is ; and his pleasure and skill in it grow as his work grows. He learns, too, a new form of verse, but he uses it sparingly, and with increasing distrust. Though his poem reflects his acquisition as well as his inheritance, it is only one new element, the rime, which he elects to make a vital part of his technique. The rhythms of his national poetry are too strong in him to be up- rooted. Its oldest forms come most naturally from his hand ; its newer, more expanded lines he receives from his contemporaries, develops and makes his own under a new necessity for fuller speech, and occasionally oversteps with a liberty born of a new example. The verse that results is not unique, for the influences that formed it played upon his fellows also. They stand together as representatives of an old order still unconquered, yet bearing the sure signs of a future inevitable yielding to the new. 14 DAY USE RETURN TO DESK FROM WHICH BORROWED LOAN DEPT. This book is due on the last date stamped below, or on the date to which renewed. Renewed books are subject to immediate recall. ID** &m mcWL E UP DEC 8 '6 4 '9 AM - - — fed 2 liases nEC 1 3 1968 1 2 B0TD ID Dur prH r. | ttq V1270-3PM2 \iLL 1 0CT2b '12 g < subject to recall afte BET 15 7 2 -3 PM 54 1KB lMAR 2 1> tooj (E4 LD 21A-60m-4,'64 (E4555sl0)476B General Library University of California Berkeley