The exact theory of the versification of King Horn remains yet to be established. Luick in his article in Paul's Grundriss offers the very ingenious hypothesis that in the beginnings of English as well as of German rimed verse, we have before us the coming to light again of the primitive Teutonic measured song verse.' This hypothesis, though ingenious and plausible, does not admit of verification, and it is perhaps safer to adhere to the view of Schipper (Grundriss der englischen Metrik), who sees in Layamon's verse the direct traditional descendants of the OE. types, and in King Horn a further development of the versification of Layamon. We see then, probably, in the versification of King Horn a transitional stage in the development of native English metre, connecting, as we have seen, more closely with the future than with the past. It was probably the occurrence in each verse of two syllables marked from the other syllables by a stronger stress, that gave rise to a feeling of uniformity in rhythm. This tendency toward uniformity in rhythm was fostered by the regular introduction of rime, for since the riming syllable naturally bore one of the two verse accents, and since the riming syllables in two riming verses would occupy the same relative position, hence in a riming verse the second of the two verse accents must balance with that in the other verse of the pair, and the balance established between the second pair of accents would naturally lead to a complete balance between the two verses. In other words the two verses would be levelled to the same rhythm. The regular introduction of rime was, no doubt, attended by the gradual loss of alliteration, which would cease to be significant as marking the verse accent, since it could hardly be made to fall regularly on the same syllable with the rime, and would hence be merely an unorganic adornment of the verse. As the position of the two verse accents came to be a fixed one, there seems to have been a tendency by raising some of the syllables bearing merely a logical stress, to rhythmic importance, thus to bring about a verse with regular measure. The most natural products of this development are the two types: (1) with three accents and feminine rime, the natural product of the OE. A, D, and C metrical types, (2) with four accents and masculine rime, the natural product of the OE. B and E types. These forms of verse were very similar, as Schipper has pointed out (as above, § 39), to two popular Romance forms of verse-namely: the first form, three accents with feminine ending, to the half verses of the Alexandrine; and the second form, four accents with masculine ending, to the verses of the short riming couplets and to the first member of the septenar. The King Horn. § 6. Versification. xxiii development toward regular measure, which had its origin as explained above, was furthered by the influence of the Romance and Mediæval Latin forms of verse. In certain ME. poems, notably the Bestiary, there are to be found verses constructed regularly after Romance or Medieval Latin models along with native forms in all the stages of development: 1. His muð is yet wel unkuð Wið pater noster and crede; Faren he norð, er fare he suð Leren he sal his nede. vv. 112-15. 2. De mire muned us bis little wile. vv. 273-6. De leun stant on hille And he man hunten here. vv. 1-2. The native forms must have been influenced by this close association with foreign forms. To these conditions and to this course of development we must probably attribute the origin of the versification in King Horn. The rime has become a regular and essential element, the alliteration, a rare and unessential element in the verse. The forms mark a transitional stage in development, but are more closely related to the new than to the old. There has been a half-hearted attempt to introduce regularity of measure, but the rhythm of the OE. types has still influenced the ear of the composer. The most frequent verse form is the one with three accents and feminine rime, about 1300 verses (Schipper). This is developed from the OE. through a stronger accent on one of the original theses; e. g., king he was biweste so longe so hit laste, vv. 5, 6 C, where the measure has been developed from the OE. A. type through stronger stress on was and so respectively. Sometimes the original OE. A. type is preserved; e. g., Hi slózen and fúzten þe nízt and þe úzten, 1473-4 C. But that this was not considered normal is shown by the fact that the other two texts, L and H, have made these two verses quoted, fit into the new normal form, by adding a new syllable in each verse, so that we have in MS. L, He smýten ánd he foúten þe nýzt and éke pe oúzten, vv. 1473–4 L. Cf. also H. The next most frequent type is the one with four accents and masculine rime; e. g., Here sone hauede to name horn; Feyrer child ne micte ben born, 9, 10 L. Less frequent types are; that with three accents and masculine ending, e. g., þu art gret and strong, Fair and euene long, 99-100 C; and that with four accents and feminine rime, e. g., To depe he hem alle brozte, His fader dep wel dere hi bozte, 951–2 C (but cf. L and H, which have more normal forms). While nearly all the verses may be made to fit into one of the types mentioned above, there are some which do not fit naturally into any one 2 of the new types, but which seems rather to be a stereotyped form handed down from OE. tradition; e. g., Bi pe se side (OE. C type) 35, of alle wymmanne (OE. C type) 71, Wringinde here honde (OE. E type) 118, Bi pe se brinke 151, In to a galeie 199, He was pe faireste 187 C. (OE. C types). (Cf. L which tries to make this verse fit better into the new versification, For þat he was fayrest), We ben of sodenne 189 L, Of Crist ene blode (OE. type E) 191 C. And þi fairnesse 227 C. poru out westnesse 228 L (MS. C adapts the verse by changing the westnesse of L. H. to Westernesse). Compound proper names seem to have been a source of confusion. Should both1 elements of the name receive stress, primary and secondary, as in OE., or should only one? Notice the struggles of the scribes with verse 169: Hy metten wiþ almair king C, Metten he with aylmer king L, metten hue Eylmer, þe kyng H. Also 257. Ailbrus gan lere C, And aylbrous gan leren L, Apelbrus gon leren H. On the whole the scribes have been fairly successful in making the native material fit into the new forms, but not unfrequently may be detected traces of the rhythm of the native OE. types, especially of the C type. § 7. DIALECT. In what dialect King Horn was originally composed, it is not easy to determine. This is a particularly difficult matter because the real pronunciation is disguised behind a great diversity of written forms. Under the circumstances the only safe guide is to be found in the rimes. Even these are very unsatisfactory since they are too few to permit any safe generalizations. For instance, it is impossible to apply satisfactorily Prof. Hempl's -wo-, -wō- test (cf. Journ. of Germ. Phil. I, pp. 14-30). In a similar way it is impossible to apply Pogatscher's ingenious test by means of the shortened product of WG. 4, WS. & (cf. Anglia, xxiii, pp. 301 ff.) because of want of rime material. Another difficulty in using the rime-test is the double pronunciation indicated, notably in the case of WS. -eald-, éa- as the result of contraction (e. g. WS. sléan), and of words with initial palatal 3- (e. g., WS. geong). Cf. examples below. From a consideration of the phonology of the poem Wissmann concludes (King Horn, Untersuchungen, Strassburg, 1876, p. 33) that, "Im Allgemeinen ist der Charakter des Vocalismus ein südöstlicher, der jedoch von dem kentischen in vielen Punkten sich unterscheidet. Die grösste 1 The rimes throughout indicate that the second syllables in compound words and the more important suffixes still bore an accent. Cf. 169-70, 199–200, 209-10, 219-20, 1353-4, etc. King Horn. § 7. Dialect. XXV Wahrscheinlichkeit hat Essex als Gegend der Entstehung für sich." A further investigation reveals to me no reason for dissenting from this view. Some of the more prominent features of the phonology are as follows: In all of the three MSS. the sign a has been disused. In its place occurs, now a, now e, so that the indication of pronunciation is often ambiguous. That the letter a sometimes denotes the a sound seems certain (cf. Wissmann, Untersuchungen, as above, p. 10). The original pure ǎ, as in some districts of America, had nearly disappeared, or been lengthened, or become o or part of a diphthong. The letter a was thus left free to denote the sound, though sometimes assisted in this function by the letter e. OE. ǎ and OE. œ (ê) shortened. In the North and the Midland, OE. œ and ê (umlaut of WG, ai) shortened, appear as a, OE. (WS.) & (= WG. â) shortened usually as e. In the West-Southern and Middle-Southern, (1) early writings have e (œ, ea), (2) later writings have a. In Kentish and East-Southern the prevailing vowel is e. (Cf. Morsbach, §§ 96–105.) In K. H. OE. ĕ appears (1) in C usually as a (one exception bed 536), (2) in H as e, e. g., sumwet: net 725–6, (3) in L as a or e. OE. (i-umlaut) shortened seems to have been written the same. Cf. 5-6, 653-4, 1249-50, with some variations from the rule in 21-2, 553–4, 1305-6, 701-2 C H. The pronunciation of this shortened OE. ❀ (i-umlaut) seems to have been e. Cf. geste feste 553-4, 1305-6, biweste laste 5-6. Apparent evidence to the contrary are haste : laste 653-4 C L (but beste: leste H), and icaste : ilaste 701-2 C H (but keste L), hadde ladde 21-2, hadde : dradde 1249-50 C L, but hedde ; dredde 1249-50 H. : : Note 1. OE. (WS.) & must have had a close pronunciation () if we may judge from the rimes; here : lere 241-2, lede: 3ede 309-10 C, ete: suete 1349–50, lere : yfere 257-8, swete: forlete 231-2, seche: speche 183–4, 483-4, etc. Or perhaps we must conclude that ệ close and ê open were not carefully distinguished in rime, for cf. stede drede 273-4 C, and Note 2. : Note 2. OE. a when lengthened in open syllables seems to have had an open ê sound. Cf. makede: verade 179-80, pere: fare 497–8 L H, speke: take 567-8, pere: aylmere 537-8 L,C H, 1613-14, 3ate: late 1123-4 C, 1593-4 C, brake: gate 1157-8 C, lede : made 1501-2 LH, slape: rape 1531-2 C. Cf. also the ai : ei rimes. L and H write ai, ay, ei, and ey without distinction. Cf. 1087-8 L, 1361-2 C, 1399-1400, etc. Note 3. Pogatscher's ingenious test (Anglia, xxiii, 301 ff.) can not be applied here, because, so far as I can see, there are no instances of rimes with shortened OE. & (WG. a). This & with original length occurs in rime, now with a lengthened in open syllable (cf. Note 2, above), now with é. Cf. seche: speche 183-4, swete: forlete 231-2, etc. On the whole, then, we may conclude that it is possible to assume for K. H. the East-Southern product e, but that if we do so we must also assume either inaccuracy in the rimes or a mixed dialect. WS. ea before + consonant is written, sometimes eld, sometinies old. It seems also to have had a double pronunciation. Both pronunciations are supported by rimes. Such rimes as welde: zelde 513–14 C H, felde : welde 451-2 H, bihelde : felde 901-2, support one pronunciation based on the OE. (WS.) breaking ea before + cons., while Admirad: bald C, amyraud baud L, Admyrold: bold H 95-6, seem to testify to the unbroken sound in OE. lengthened before -ld to â and then opened to ộ. For other instances with varying spelling cf. 17-18, 323-4, 397-8, 639– 40, 1499–1500. In v. 497 the L reading talede seems to represent the OE. broken form as opposed to the unbroken form tolde in C and H. OE. ĕ. There are many instances of e: i rimes. But it is seemingly impossible to determine thereby much concerning the dialect. (Cf. Morsb. § 109, 114, N. 1.) For examples of this rime, cf. wille : telle 383-4, 1015-16 C; stille : duelle 393-4 C; pikke: nekke 1327-8; snelle: wille 1581-2 C, etc. The form sigge seems to belong especially to the South-East. (Morsb. 114, N. 1, 109, N. 4, also Wissman, King Horn, p. xiv.) Cf. K. H. vv. 1367-8, ligge wipsegge C, ligge sigge L; lygge: wipsugge H. OE. §, umlaut of u offers many difficulties. It is represented in writing by y, i, u, e. The rimes show the prevailing sound to have been e; e. g., Suddenne : kenne 155–6, 923-4, pelle : fulle 421-2, leste beste 505-6, also 617-18, 671-2, 647-8, 703-4, 917-18, 919-20 L, 805–6, 795-6, 1479-80, 1637-8, 1341-2, 1367-8, etc. But cf. y i in kesse: ywisse 461–2 C H, lizte : drizte 1405-6 C. That y: i rimes should occur, might be expected in view of the vague distinction between e and i as shown by the e: i rimes, but the number of y: e rimes attests to a pronunciation e. This is the strongest available evidence that K. H. was composed in the south-eastern district. That the dialect of King Horn is a mixed dialect is supported by the treatment of a above by the double pronunciation of WS. -eald, and by further double pronunciations. OE. (WS.) slêan, flêan seem to have had double pronunciations. The pronunciation is attested to by the rime, |