The examples under (a) show the tendency noticeable already in the first hemistich of the Old English alliterative line to admit anacrusis. The examples under (b) and (c) may be looked upon as extended forms of types E and D.
§ 56. Several poems of somewhat later date deviate more frequently from these types than the Alisaunder fragments, chiefly in the following points:
The end of the hemistich sometimes consists of an accented syllable instead of an unaccented one; the thesis is sometimes monosyllabic instead of polysyllabic, especially in A, or the anacrusis may be polysyllabic instead of monosyllabic. Secondary accents are introduced more frequently into the second hemistich also, but by poets whose technique is careful they are admitted only between the two accented syllables. Owing to these licences, and to the introduction of polysyllabic theses, the rhythm of the verse sometimes becomes very heavy.
Belonging to this group are William of Palerne, Joseph of Arimathie, both belonging to the middle of the fourteenth century, the three editions of William Langland’s Vision concerning Piers Plowman, of somewhat later date, and a few minor poems. The Romance of the Chevelere Assigne, written in the East Midland district, at the end of the fourteenth century, and the works of the Gawain-poet, viz. Sir Gawain and the Green Knight, Cleanness, Patience, and the Legend of St. Erkenwald (Horstmann, Altengl. Legenden, 1881, p. 265), form the transition to another group of poems belonging to the North of England, but differing somewhat from the preceding with regard to their metre.
The most important amongst these is Langland’s great work, but it is at the same time most unequal in respect to its versification. In many passages, especially in the beginning of the several Passus, as they are called, the flow of the verses is very regular; in other passages the theses are frequently of such great length, and the arsis stands out so indistinctly, that the rhythm of the verse can only be made out with difficulty. Some examples taken from the B-text (c. 1377) may serve to illustrate this:
Extended second hemistich (Type A):
Extended first hemistich (Type A):
Such verses obviously contain only two beats in each hemistich, although at the same time some of the syllables forming the thesis may have a somewhat stronger accent than others. For as a rule such extended verses are succeeded by a normal line, clearly bringing out again the general four-beat rhythm, as is the case with the line (A + A) following immediately upon the last-mentioned example:
Type A is in Piers Plowman the usual one, but the types C and BC frequently occur. In the following examples we have type C in the second hemistich:
in the first hemistich it occurs rarely:
Type B C is frequently to be met with in both hemistichs; e.g. in the first:
in the second:
Masculine endings, however (originating from the dropping of the final -e in the last words of the types A and C, as e.g. in and drédful of síght Prol. 16, cristened þe kýnge xv. 437, as þe kýng híght iii. 9), occur very rarely here. They are, on the other hand, characteristic forms in another group of alliterative poems.
§ 57. These belong to the North of England and the adjacent parts of the Midlands.
In these districts the final e had by this time become silent, or was in the course of becoming so. Thus many verses of West-Midland poems were shortened in the North by omitting the final -e, and then these forms were imitated there. Hence the middle of the line was much less modified than the end of it.
Types A, C, B C, therefore, occur not only in the ordinary forms with unaccented syllables at the end, but also, although more rarely, with accented ones, viz. corresponding to the schemes:
A1, (×)–́××–́, C1, (×)××–́–́, BC1, (×)××–́×–́.
These forms of the hemistich first occur in the Destruction of Troy, a poem written in a West-Midland dialect very like to the Northern dialect, and in the North-English poems, Morte Arthure and The Wars of Alexander (E. E. T. S., Extra-Ser. 47). Examples of these types (taken from the first-mentioned poem) are: of type A1 in the second hemistich, for lérning of ús 32, þat ónest were áy 48; with a polysyllabic thesis, and lympit of the sóthe 36; with a secondary accent,with cléne mèn of wít 790; without anacrusis,[112] lémond as góld 459, bléssid were Í 473; in the first hemistich, with disyllabic anacrusis, þat ben drépit with déth 9, þat with the Grékys was grét 40; without anacrusis, Býg y-noghe vnto béd 397, Trýed men þat were táken 258, &c.; examples for C1 (only occurring in the second hemistich), þat he fóre with 44, into your lond hóme 611, ye have sáid well 1122, þat ho bórne wás 1388, of my córs hás 1865; examples for B C1, in the second hemistich (of rare occurrence), when it destróyet wás 28, and to sórow bróght 1497, þere þe cítie wás 1534.
The same modification of types took place later in other parts of the Midlands, as appears from two works of the early sixteenth century, Scottish Field and Death and Life (Bishop Percy’s Folio MS., edited by Furnivall and Hales, i. 199 and iii. 49). The last North-English or rather Scottish poem, on the other hand, written in alliterative lines without rhyme, Dunbar’s well-known Satire, The twa mariit wemen and the wedo, has, apart from the normal types occurring in the North-English poems, many variants, chiefly in the first hemistich, which are characterized by lengthy unaccented parts both at the beginning of the line, before the second arsis, and after it; frequently too syllables forming the thesis have a secondary accent and even take part in the alliteration, as e.g. in the following examples:
Sometimes the second hemistich participates in this cumulation of alliterating words, which not unfrequently extends over several, even as many as six or seven consecutive lines:
This explains how King James VI came to formulate the metrical rule mentioned above (p. 89) from the misuse of alliteration by the last poets who used the alliterative line, or the alliterative rhyming line to be discussed in the next paragraph, which shares the same peculiarity.
§ 58. In spite of the great popularity which the regular alliterative line enjoyed down to the beginning of the Modern English period, numerous and important rivals had arisen in the meantime, viz. the many even-beat rhymed kinds of verse formed on foreign models; and these soon began to influence the alliterative line. The first mark of this influence was that end-rhyme and strophic formation was forced upon many alliterative poems. In a further stage the alliterative line was compelled to accommodate its free rhythm of four accents bit by bit to that of the even-beat metres, especially to the closely-related four-foot iambic line, and thus to transform itself into a more or less regular iambic-anapaestic metre. The alliterative line, on the other hand, exercised a counter influence on the newer forms of verse, inasmuch as alliteration, which was formerly peculiar to native versification, took possession in course of time to a considerable extent of the even-beat metres, especially of the four-foot iambic verse. But by this reciprocal influence of the two forms of verse the blending of the four-beat alliterative line with that of four equal measures and the ultimate predominance of the even-beat metres was brought about more easily and naturally.
Alliterative-rhymed lines, the connexion of which into stanzas or staves will be treated of in the second part of this work under the heading of the ‘Bob-wheel-stanza’, were used during the Middle English period alike in lyric, epic, and dramatic poetry.
§ 59. Lyrical stanzas. The earliest stanzas written in alliterative rhyming lines were lyrical.
We must distinguish between isometric and anisometric stanza forms. In the former the whole stanza consists of four-beat alliterative lines, commonly rhyming according to a very simple scheme (either a a a a or a b a b). In the latter four-beat long lines as a rule are combined with isolated lines of one measure only and with several of two measures to form the stanza. The two-beat verses frequently have a somewhat lengthened structure (to be discussed further on sections on the epic stanzas), in consequence of which many of them having theses with secondary accents can be read either as even-beat verses of three measures or as three-beat verses on the model of those in King Horn. The four-beat alliterative lines, on the other hand, are mostly of more regular structure, the distances between the first and second arsis not being so unequal and the theses as a rule being disyllabic. The anacrusis too in these verses admits of a somewhat free treatment. The difference, however, between the first and second hemistich is less conspicuous than it was in those forms of the Middle English alliterative line before mentioned. Alliteration, on the other hand, is abundantly used.
The main rhythmic character of the verse is again indicated here by the frequent occurrence of the types A and A1. The types B C, B C1, C, C1, however, likewise occur pretty often, and the two last types present serious obstacles to the assumption that the lines of these poems were ever recited with an even beat. But how exactly these poems were recited or to what sort of musical accompaniment can hardly be definitely decided in the absence of external evidence.
The first verses of a West-Midland poem of the end of the thirteenth century (Wright’s Political Songs, p. 149) may serve as a specimen:
The second hemistichs in ll. 2 and 4 belong to type C. In other poems also, with lines of more regular rhythm (chiefly type A), this type may be met with now and then, e.g. in a poem published in Wright’s Specimens of Lyric Poetry, p. 25, especially in the second hemistich, e.g. haueþ þis mái mére, line 9, and þe gýlófre, line 40, þat þe bór béde, line 44.
It is not difficult to distinguish such rhymed four-beat alliterative lines from those of four measures which have fairly regular alliteration, for the long line of the native metre always has a somewhat looser fabric, not the even-beat rhythmic cadence peculiar to the iambic verse of four measures, and, secondly, it always has a caesura after the first hemistich, whereas the even-beat verse of four measures may either lack distinct caesura or the caesura may occur in other places in the verse as well as after the second arsis. This will be evident by comparing the following four-beat verses of the last stanza of a poem in Wright’s Spec. of Lyr. Poetry, p. 31:
with the following first four-beat alliterative lines of another poem (ibid. p. 25):
In similar lines are written several other poems, as Mon in þe mone (ibid. p. 110); Of ribaudz y ryme (Wright’s Pol. Songs, p. 237); and five songs by Laurence Minot (nos. ii, v, ix, x, xi), written in the middle of the fourteenth century.
§ 60. In other poems the four-beat long lines used in the main part of the stanza are followed by shorter lines forming the cauda, which in part are of a variable rhythmic cadence either of three beats (or three measures) or of two beats, as e.g. in the well-known poem in Percy’s Reliques, ii, p. 1.[113] The first stanza may be quoted here:
In the following stanzas of this poem the four-beat rhythm, although rarely marked by regular alliteration, is (in the main part or ‘frons’) still more distinctly recognizable, in spite of several rhythmically incorrect lines.
Second hemistichs of the type C1 are not infrequent, e.g. opon swývýng 9, sire Édwárd 46, o þy lýárd 47. Lines 5 and 7 are of a two-beat rhythm, l. 8 probably as well (cf. our scansion).
There is a decided similarity in regard to structure and versification between this stanza and that of a poem in Wright’s Pol. Songs, p. 153, although the long lines are divided in the middle by interlaced rhyme. This may be illustrated by its second stanza:
There is no line here corresponding to l. 5 of the preceding poem. Otherwise, however, the cauda of this poem is of a similar structure to that of the preceding one, at least in this and possibly in the following stanzas, whereas the last line of the first stanza has a two-beat rhythm, and in the others the last lines probably are to be scanned with three beats. The second line of the cauda of the first stanza of this poem belongs to type C. Another poem (Wright’s Polit. Songs, p. 155; Böddeker, P.L. no. iv) shows a very artificial form of stanza, either corresponding to the formula a a4 b2 c c4 b2 d d4 b2 e e4 b2 f f g g g f2 (if we look upon the verses as four-beat and two-beat lines, which the poet probably intended), or corresponding to the formula a a4 b3 c c4 b3 d d4 b3 e e4 b3 f f g g g f2 (if we look upon the frons as consisting of ordinary tail-rhyme-stanza lines of four and three even-beat measures).
The four- and two-beat cadence of the verses comes out still more clearly in the stanzas of another poem (Wright’s Pol. Songs, p. 187; Ritson, Anc. Songs, i. 51; Böddeker, P.L. no. v), the rhymes of which follow the scheme a a a4 b2 c c c4 b2 (extended tail-rhyme-stanzas). Some of its long lines, it is true, admit of being read as even-beat verses of three measures, e.g. and béo huere chéuentéyn 20, and móni anóþer swéyn 24, but the true scansion in all probability is and béo huere chéuenteȳ̀n (or chèuentéyn): ant móni anòþer swéyn, in conformity with the scansion of the following lines to cóme to parís: þourh þe flóur de lís 52–6, or wiþ éorl and wiþ knýht: with húem forte fýht 124–8.
As a first step to the epic forms of stanza to be considered in the next paragraph a poem of the early fourteenth century (Wright’s Pol. Songs, p. 212; Ritson, Anc. Songs, p. 28; Böddeker, P.L. no. vi) may be quoted:
The fifth line has one arsis only (as appears more clearly from that in the second stanza: wiþ Lóue), thus corresponding to the above-mentioned poems (pp. 99, 100); the other lines of the cauda have two stresses.
Prof. Luick[114] looks upon the long lines of this poem and of several others (e.g. Wright’s Pol. Songs, pp. 69 and 187) as doubled native verses of the progressive or Layamon form, but rhyming only as long lines. This can hardly be, as the rhythmic structure of these verses does not differ from that of the other poems quoted above, which belong, according to Prof. Luick himself, to the class of the normal, lyric rhyming-alliterative lines.
§ 61. Narrative verse. Alliterative-rhyming verses occur in their purest form in narrative poetry, especially in a number of poems composed during the fourteenth and fifteenth centuries in stanzas of thirteen lines, and republished recently in a collective edition by the Scottish Text Society in vol. 27 under the title Scottish Alliterative Poems (ed. by F.J. Amours, Edinburgh, 1892). The poems contained in this collection are Golagras and Gawane (also in Anglia, ii. 395), The Book of the Howlat by Holland, Rauf Coilȝear (also in E. E. T. S., Extr.-Ser. vol. xxxix), The Awntyrs off Arthure at the Terne Wathelyne, The Pistill of Susan (also in Anglia, i. 93). Douglas’s Prologue to the Eighth Book of his translation of the Aeneid (although written in the beginning of the sixteenth century) likewise belongs to this group, as do also the poems of John Audelay, composed in Shropshire in the fifteenth century (Percy Soc. xiv, p. 10 ff.), and a poem Of Sayne John the Euaungelist (E. E. T. S. 26, p. 87) written in stanzas of fourteen lines in the North of England. The stanzas of all these poems—generally speaking—consist of two unequal parts, the frons written in alliterative lines, rhyming according to the formula a b a b a b a b, and the cauda which contains five or six lines, the first of which may either be a long line as in the frons, or, as in The Pistill of Susan, a short one-beat one, with four two-beat sectional verses following. Only in the last-mentioned poem does the cauda consist of six two-beat sectional verses.
The rhythm of this alliterative-rhyming metre may first be illustrated by the opening lines of Golagras and Gawane:
I.
II.
Lines like the four last quoted illustrate the normal structure of the rhyming-alliterative verse, especially the relationship of rhyme and alliteration to each other in monosyllabic and disyllabic words. It will be seen that the rhyming syllable, as a rule the root-syllable, or at least the accented syllable of the word, at the same time carries the fourth accent of the line, and in consequence the fourth alliterative sound. In all other respects the rhymed-alliterative verse is structurally similar to that without rhyme, and it is therefore evident that rhyme exercises no decisive influence on the rhythm of the verse. In this comparatively pure form—if we do not take into account the secondary accents occurring in the first hemistichs of the stanza in the later poem—are written the great majority of the lines in the earliest of poems mentioned above, viz. The Awntyrs off Arthure.
§ 62. The relation, however, between rhyme and alliteration and consequently the relation of the rhythmic accentuation of the words to their natural accentuation is less clear in the first stanza quoted above. The following verses rhyming together may serve to elucidate this:
or the verses Gol. 648, 650, 654:
In the first couplet the last syllable of the word gráciùs, although bearing only a secondary accent and forming the last thesis of the verse, rhymes with the last syllable of the word chéuailrús, which likewise in ordinary speech has a secondary accent, but here is the bearer of the fourth metrical accent of the verse. In the second couplet the syllable lyng of the word wéderlỳng, which has a secondary accent and forms part of the thesis, rhymes with the word hyng which has the rhythmical accent. In the last group of verses the last syllable of the words paramour, honour having secondary accents rhymes with the word hour, the bearer of the last rhythmical accent. Similar rhymes occur even in Modern English poetry, e.g. in the works of Thomas Moore: Váin were its mélodỳ, Róse, without thée or Whát would the Róse bè Únsung by thée?[115]
It also frequently happens that all the rhyming syllables, which have a secondary accent and occur in the thesis of a verse, belong to trisyllabic words, while the accented syllables in the arsis, whether alliterating or not, do not take part in the rhyme, e.g.:
Similar unaccented rhymes are also met with in disyllabic words:
Other rhymes of the same kind are sémbland: léuand, conséntand : endúrand, Gol. 428 ff., &c.
In all such cases the natural accentuation of the words is not interfered with by the rhythm of the verse.
The kind of irregular rhyme most frequently occurring, however, is that which is formed by the unaccented syllable of a disyllabic word (the first syllable of which alliterates and bears the last arsis of the verse) rhyming with a monosyllabic word which likewise bears the fourth rhythmical accent of another alliterative line (or the second of a short line forming part of the cauda) and takes part in the alliteration as well, as e.g. in the rhymes Túskane: sane: Brítane: gane and súmmovne: rénovne: crovne of the above-mentioned stanza of the poem Golagras and Gawane.
It is not likely that a complete shifting of accent in favour of the rhyming syllable ever took place, as the first syllables of the words usually take part in the alliteration, and therefore have a strongly marked accent. Sometimes, it is true, in the poems of this epoch, unaccented syllables do participate in the alliteration, and in the case of the words Tuskane, Britane, summovne, renovne their Romance origin would explain the accent on the last syllable; but these words, both as to their position and as to their treatment in the line, are exactly on a par with the Germanic rhyme-words in ll. 870–2:
In both cases we thus have ‘accented-unaccented rhymes’ (cf. Chapter I in Book II), which probably were uttered in oral recitation with a certain level stress. This is probable for several reasons. First it is to be borne in mind that Germanic words in even-beat rhythms of earlier and contemporary poems were used in the same way, e.g.:
Only in these cases the rhythmical accent supersedes the word accent which has to accommodate itself to the former, while in the uneven-beat rhythm of the four-beat alliterative line the word-accent still predominates. In the even-beat lines, therefore, the rhythmical accent rests on the last syllable of a disyllabic rhyme-word, but in the alliterative lines it rests on the penultimate. In the case of words of Romance origin, however, which during this period of the language could be used either with Germanic or with Romanic accentuation, the displacement of the word-accent by the rhythmic accent in non-alliterative words may in these cases have been somewhat more extensive; cf. e.g. rhymes like rage: curáge: suáge Gol. 826–8; day : gay: journáy ib. 787–9; assáill: mettáill: battáil R. Coilȝear, 826–8, &c. (but ȝone bérne in the báttale Gol. 806).
As a rule, however, for these too the same level-stress accentuation must be assumed as for the rhyme-words of the first stanza of Golagras quoted above (p. 102)
§ 63. This is all the more probable because, in these alliterative-rhyming poems, there are many sectional verses corresponding to the old types C and C1, these answering best the combined requirements of alliteration and of end-rhyme, for which frequently one and the same Germanic or Romanic word had to suffice in the second hemistich, as e.g. in the following sectional verses rhyming together:—What is þi góod réde: for his kníȝthéde: (by crósse and by créde) Awnt. of Arth. 93–7; (and bláke to þe bóne): as a wómáne ib. 105–7; enclósed with a crowne: of the trésóne ib. 287–91; of ane fáir wéll: téirfull to téll: with ane cástéll : kéne and crúèll, or, as Prof. Luick scans, kéne and cruéll (but l. 92 crúel and kéne) Gol. 40–6; at the mýddáy: (wént thai thar wáy) Howl. 665–7. &c.
Also in the even-beat metres the influence of this type is still perceptible; cf. rhymes like
which are of frequent occurrence.
For the rest both in these alliterative-rhyming poems and in the poems with alliteration only the types A and A1, B C and B C1 are frequent. These alliterative-rhyming lines have this feature in common with the pure alliterative lines, that the first hemistich differs materially from the second in having often an anacrusis of several syllables (initial theses) and somewhat lengthened theses in the middle of the line, and in permitting such theses with only a secondary accent to take part in the alliteration. All this tends to give a somewhat heavy rhythmic cadence to the whole line.
§ 64. The same difference is perceptible, as Prof. Luick was the first to show (Anglia, xii, pp. 438 ff.), in the single two-beat lines of the cauda, the three first (ll. 10–12 of the whole stanza) having the looser structure of the extended first hemistichs of the long lines, while the last two-beat line (line 13 of the whole stanza) has the normal structure (commonly type A, A1, as e.g. Birnand thrétty and thré Gol. 247; Of góld that wes cléir ib. 1) of second sections of the long line, as is evident from the first stanza of Golagras and Gawane quoted above (p. 102). In this concluding line, however, other types of verse peculiar to the second hemistich of long lines may also be met with, as e.g. C, C1, BC, BC1, e.g.: For thi mánhéde Awnt. of Arth. 350; Withoutin dístánce Gol. 1362; As I am tréw kníght Gol. 169; Couth na léid sáy ib. 920; In ony ríche réime ib. 1258, Quhen he wes líghtit dóun ib. 130.
In other poems the group of short lines rhyming according to the scheme a a a b and forming part of the cauda is preceded neither by a long alliterative line nor by a one-beat half section of it (as in Susan), but by a complete two-beat sectional verse, which then, in the same way as the last verse rhyming with it, corresponds in its structure to that of the second hemistich of the long line; as e.g. in The Tournament of Tottenham (Ritson’s Ancient Songs, i. 85–94), rhyming on the scheme A A A A b c c c b (the capitals signifying the long lines), and in The Ballad of Kynd Kittok, possibly by W. Dunbar (Laing, ii. 35, 36; Small, i. 52, 53; Schipper, 70).
In Sayne John the Euaungelist the ‘cauda’ has the structure of a complete tail-rhyme-stanza, the order of rhymes of the whole stanza being A B A B A B A B c c d c c d
§ 65. In connexion with this it is particularly interesting to note that such two-beat sections of the alliterative line are also used by themselves for whole poems written in tail-rhyme-stanzas (as was first shown by Prof. Luick, Anglia, xii, pp. 440 ff.); cf. e.g. the translation of the Disticha Catonis (E.E.T.S. 68), the two first stanzas of which may be quoted here:
In the same stanza The Feest (Hazlitt, Remains, iii. 93) is written.
Still more frequently such lines were used for extended tail-rhyme-stanzas rhyming on the scheme a a a b c c c b d d d b e e e b, as e.g. in a poem, The Enemies of Mankind, of the beginning of the fourteenth century, published by Kölbing (Engl. Studien, ix. 440 ff.).
The first stanza runs as follows:
Here, again, the difference between the lines on the pattern of the first hemistich of the long line, which form the body of the stanza (a a a, b b b, c c c, d d d), and those on the pattern of the second hemistich used as tail-rhyme lines (b, b, b, b) is plainly recognizable.
The same is the case in other poems written in this form of stanza, as e.g. in the Metrical Romances, Sir Perceval, Sir Degrevant (Halliwell, Thornton Romances, Camden Society, 1844, pp. 1, 177) and others; cf. Luick, Anglia, xii, pp. 440ff., and Paul’s Grundriss, ii a, p. 1016. But in these later works, one of the latest of which probably is the poem The Droichis Part of the Play, possibly by Dunbar (Laing, ii. 37; Small, ii. 314; Schipper, 190), the two-beat lines are frequently intermingled and blended with even-beat lines, which from the beginning of the fifth stanza onward completely take the place of the two-beat lines in the last-mentioned poem. Likewise in the ‘Bob-wheel-staves’, i.e. stanzas of the structure of those sixteenth-century stanzas quoted above (§§ 60, 61), the cauda, as is expressly stated by King James VI in his Revlis and Cavtelis, is written in even-beat lines of four and three measures, though the main part of the stanza (the frons) is composed in four-beat rhyming-alliterative lines (cf. Luick, Anglia, xii, P. 444)
§ 66. In the contemporary Dramatic Poetry this mixture of four-beat (or two-beat) alliterative lines with lines of even measures is still more frequent, and may be used either strophically or otherwise.
In the first place, we must note that in the earlier collections of Mystery Plays (Towneley Mysteries, York Plays, and Ludus Coventriae) the rhyming alliterative long line, popular, as we have seen, in lyric and in narrative poetry, is also used in the same or cognate forms of stanzas.
But the form of verse in these Mysteries, owing to the loss of regular alliteration, cannot with propriety be described as the four-beat alliterative long line, but only as the four-beat long line. In many instances, however, the remnants of alliteration decidedly point to the four-beat character of this rhythm, as e.g. in the following stanza of the Towneley Mysteries (p. 140):