Alliteration can fall on a final consonant which, in pronunciation, attaches itself to a following word beginning with a vowel; C 1779 thus alliterates on n:
Withinne an houre of the night an entre thay hadë
Alliterative needs may be satisfied by an unaccustomed stress on a prefix rather than on the root syllable, something which (especially in the case of Romance loan words) figures so frequently as to suggest that it was an accepted practice in alliterative poetry:43
Then cònfourme thee to Cryst, and thee clene makë (1067)
Prèsented him the presoneres in pray that thay token (1297)
The possibility of stress on prefixes gave poets useful options. For instance, some words occur with alliterative stress falling either on the prefix or on the root: compare ‘He dèvised his dremës’ (1604a) with ‘Devìsed he the vesselment’ (1288a); ‘Was dìsstryed with dìstresse’ (1160a) with ‘I schal strenkle my distrèsse’ (307a); ‘Oght that was ùngoderly’ (1092a) with ‘Thou art a gome ungòderly’ (145a);44 and ‘He brek the barers as bỳlive’ (1239a) with ‘to sech him bylìvë’ (1615b).
Alliterative stress falls generally on open-class words – classes of words to which new items are easily added (nouns, adjectives, adverbs and most verbs). Closed-class words (grammatical-function words such as prepositions, pronouns, auxiliary verbs) do not normally bear alliterative stress. Not all open-class words bear stress, however (even if they alliterate: as in G 2: ‘The burgh brittened and brent to brondes and askes’). This is frequently the case with such common words as ‘thing’, ‘man’, ‘lord’, ‘take’, ‘have’, with adverbs of degree, quantifying adjectives, and with adjective + noun and adverb + verb combinations:
And if He lovës clene layk that is our lord richë (1053)
So brod bulde in a bay that blonkes myght rennë (1392)
Schot with his schulderes his fayrë schelde under (G 2318)
The man marred on the molde, that myght him not hydë (Pa 479)
Though other patterns can occur in Sir Gawain, the aa/ax rule is followed in Cleanness and Patience with such regularity as to justify editors in regarding as corrupt any non-aa/ax lines (e.g. C 345, 422, 427) and in emending them wherever a plausible emendation suggests itself.
Recent research by Hoyt Duggan45 and by Thomas Cable46 has established certain rules governing the metrical shape of the b-verse (the second half of the line): one of the two stresses must be preceded by a ‘long’ dip – that is, by more than one unstressed syllable – and only one such ‘long’ dip can figure in the verse. The b-verses of the lines quoted above and below furnish examples, and also illustrate the operation of another rule: the line must end in one (and only one) unstressed syllable. That this is a rule (and not merely a strong tendency, as Duggan himself describes it) is readily demonstrable. Where the poet had a choice of forms, the disyllabic one is regularly selected at line-ending position: compare line-ending selven (1745), spylled (1248), schined (1532) with non-line-ending self (1418), spylt (1220), schon (G 772, 956). Some disyllabic forms are found only at line ending: lys [lies], for instance, appears as lygges only at line ending (1126 and 1792). Similarly, now only ever appears at line ending as nowthe (Pa 412; G 1251, 1784, 1934, 2466). No syllable bearing secondary stress can occur at line ending: thus suffixes such as -ing or -ly/lych do not figure in that position, nor does the second syllable of such words as warlow or coltour; in short, virtually the only vowel sounds that can occur at line ending are schwa (the indistinct vowel heard in the second syllable of, for instance, china) and unstressed i (as in the second syllable of merry).
With regard to the a-verse, a long dip (at least two unstressed syllables) between the two stresses is the norm, and this poet makes deliberate and demonstrable adjustments to achieve it. For instance, adverbs and adjectives in -lych(e) are used in these poems only for metrical reasons: unlike -ly (the poet’s usual, unmarked form, which was always monosyllabic in later Middle English), the suffix -lych permitted the addition of inflectional -e (though this has been mostly lost from the -lych spellings of the manuscript), so producing a disyllabic suffix.47 This is clear from both a- and b-verses:
In His comlychë court, that kyng is of blissë (546)
And other luflychë lyght, that lemëd ful fayrë (1486)
Of that wynnelychë Wye that wonës in heven (1807)
At this causë the knight comlychë hadë (G 648)
The poet’s other forms of these words (lovely/lufly, comly and wynne [joyous, glorious]) would not have permitted the -e inflection for adverb (at G 648), plural adjective (at 1486) or weak adjective (at 546 and 1807) that in these lines ensures the inter-stress long dip in the a-verse. The insertion of redundant for before to + infinitive is likewise, in this manuscript and elsewhere, metrically motivated, and that too occurs to ensure the inter-stress two-syllable dip in the a-verse:
And had a wyf for to welde, a worthilyche quenë (1351)
Now is set for to serve Satanas the blakë (1449)
Insertion of ful [very] – before, for instance, a post-positioned adjective – is likewise a common metrical manoeuvre used to the same end:
Was long and ful large and ever ilyche swarë (1386)
Then a dotage ful depe drof to his hertë (1425)
The disyllabic dip between the two alliterating staves is dispensed with only where the want of it is compensated by an opening dip of four or more syllables – e.g. ‘And there was the kyng caght’ (1215a) – or by a verse-ending dip that is long (two or more syllables) and/or heavy (i.e. a syllable with secondary stress such as is banned from line ending), e.g. ‘In the burgh of Babyloyn’ (1335a).
Since English (unlike French) is a stress-timed rather than a syllable-timed language, the alliterative metre is in some ways more natural to it than are octosyllabic or decasyllabic iambics. But there is another sense in which it is the metre ‘native’ to English. The four-stressed alliterating line had been the metrical norm in Anglo-Saxon verse before the Norman Conquest, and that Old English alliterative line is evidently the ancestor of the Middle English one. Some of the alliterative synonyms used in Old English verse also continue into Middle English alliterative poems, though a number of them were, by this time, otherwise rare or archaic usages. The synonyms for ‘man’, for instance – burn, freke, gome, hathel, lede, renk, schalk, segge, wye – are all found in both Old and Middle English alliterative verse.
Historically and prosodically ‘natural’ to English though it may be, that word should not be taken to imply that alliterative verse is therefore any less ‘artful’ or rule-bound than rhymed verse, though the absence of rhyme and regularly recurring identical feet may make it look at first blush ‘freer’ than, for instance, the rhymed iambic pentameters of Chaucer. As will by now be clear, there are complex rules regarding rhythm and alliteration that impose no less pattern on normal or prose usage – and which will, for instance, occasionally necessitate the same artful deviation from normal word order as occurs in rhymed verse. When, for example, Belshazzar’s scholars fail to interpret ‘What tythyng ne tale, tokened tho draghtes’ (1557), and he sends for men ‘That were wyse of wychecraft, and warlowës other’ (1560), the word order in the b-verses is dictated by the alliterative aa/ax pattern (which, in the strict form of the metre observed in Cleanness and Patience, cannot be varied to aa/xa), and by the requirements for a long dip in the b-verse and for a single unstressed syllable at line ending, which would preclude ‘tho draghtes tokened’ or ‘other warlowes’.