The Battle of Maldon, page 10
We must assume an intentional divergence of prosody. This looser freer kind, as I have said, may have all along existed precisely as a less studied form, and have continued to exist as a popular or ruder form after the French invasion, when the more formal modes perished, reappearing in the various ME varieties. It may, however, of course have been derived from the deliberate acceptance by poets of the results of time and linguistic change upon the more compact phrasing of the older periods (which was the basis of the stricter verse). This action of change did not, as I have said, alter metre – it left the old kind intact for use if required. But it may have encouraged the development (by poets) of a freer looser kind, easier to fit to the normal wordier and more analytic syntax of the current speech, and so convenient for rapid and less studied composition. Thus mid prasse bestodon for prasse 68b; hafast ealle gemanode for ealle gemanodest 201b etc. But such elongated word-groups of course always existed: and whenever the freer type developed, it was rather by the purposefulness of craft than the compulsion of time. In Norse, a language that did not to the same extent or so quickly as English alter the compact brevity of older phrasing, we find new metres being devised by the deliberate selection and normalising of the maximum elements found in the older epic verse – the making of málaháttr. And in such intermediate and debatable cases as Atlakviða we have really a close parallel to the versification of Maldon. The action of time is seen in Maldon: but in the cases where it can clearly be detected it does not alter the rules, but allows the pattern to be filled with material which in the older stages of the language would not have fitted. That is quite a different thing. Thus g front and back had diverged and are no longer equated in alliteration. The gradation of stress in different parts of speech is also shifting. The strict subordination of one stress to the immediately preceding stress is probably making way for a more level stress, and so we get Ǽlfhere and Máccus with only alliteration on M. The verb finite is no longer so generally subordinate to nouns, and is especially when it precedes its subject evidently at least equal to it in stress. So we have STihte hi Byrhtnoth 127, 139 (helped by crossed alliteration geHLeop Eoh Ahte HLaford), 240. Some compounds, especially those ending in man which was very frequently the second element in such words, are weakened: brím-mèn becomes brímmen and so we have brimmen wodon 295b, a variety that in Beowulf in the first hemistich would need double alliteration, and in the second would require lightening in the second main lift. But these are no breaches of the rules according to the purpose and reason of them.
In European verse, then, and in English verse we can have at once a tradition of metre, and at the same time connected but independently a tradition of actual poems. In this way already before the wide diffusion or practice of writing archaic words and forms might be preserved in old poems, and perpetuated in newer ones – for metrical convenience, and as poetic diction without special metrical reason. In this way in England at any rate could not be preserved the whole grammar and language of the past – this or anything like it only occurred when poems got written, and so left the stream of a tradition more or less. For the language of verse was no separate species and its life was bound up with the vernacular. This tradition could not preserve obsolete sounds. Obsolete equations (alliteration or rhyme) when preserved in writing may suggest to enquiry that an old identity has been disturbed; but without writing nothing is preserved but the fact that there is a misfit.
We do find, of course, in later times – especially after the writing and reading of verse has become general, and even its composition with the aid of writing common – old collocations or equivalences, like rhymes maintained after they have in current pronunciation ceased to conform to the rules. Even so they are usually preserved for some metrical reason as the scarcity of rhymes to such useful words as heaven, love; though equivalence in traditional orthography may maintain some (such as war and star) which have no such excuse. But even so, even now with the spelling strongly impressed upon our minds, we do not usually in repeating older poetry ‘mispronounce’ war when we find it rhymed with star, nor broad to match road. And we could not do so if we had not a fixed spelling – at any rate we should not know which of an ill-fitting pair to alter.[14]
‘Tradition’ also is powerless by itself to enforce the use of archaic or poetical forms in all cases to the exclusion of current equivalents. It can only usually preserve them as optional. If certain words like heaþu are only found in OE in ‘poetic’ (originally dialectal) form, it is because they had no current equivalents in the language of the persons who made our copies. If we feel now any hesitation about mixing thou and you, or eth and s as in sitteth or sits, this is because of an actual book-knowledge of the past, and because of grammatical teaching. In oral tradition there is no defence against new forms that have come to conform to the rules. Only spelling protects us (if it is protection) against rhymes such as sword/fraud; and even our spelling boggles only about consonants,[15] it is too inured to discrepancy in vowel sounds to object to heart/part or warm/storm. In fact ‘archaisms’ preserved by tradition, affecting sounds purely, and not vocabulary or grammar, are rare and of small effect. To actual readers or hearers of verse, or to makers of it that still use them, they are not ‘archaisms’, but parts of poetic usage or diction – ‘licences’: conventionally permitted defects, not different from real ‘licences’ such as the inaccurate rhymes for home and love which repose in the main upon the sheer scarcity of good rhymes for these desirable words. They are either tolerated as misfits – or else are made the starting point of technical innovation: that is of alteration of the rules so as to admit them. They do not preserve the noises of by-gone days.[16] They cannot. Even if we pronounce war to rectify the rhyme with far, star we shall still not recapture the actual sounds of war, star in the days when they did rhyme, for those sounds no longer exist in English. It would be amusing if we could recall a man of the late tenth-century and get him to read out some Beowulf. In the first line, and still more when he came to line 151 gyddum geomore, þætte Grendel wan, his own pronunciation would destroy alliteration. He would no longer make lines like those himself. Yet absence of alliteration would be disturbing. And they would alliterate to the eye. Now we have good reason to suppose that there was some reading of verse aloud by book at any rate in the later period; the books would hardly else have been made. And the Saxonica poemata that young Alfred learned to read were certainly not all (if any) contemporary West Saxon productions.[17] Was there any lore that taught readers what to do? I doubt it. Of course when back g had ceased to be a spirant initially, the spirant had not vanished from the language. It was frequent medially – a parallel in reverse to our initial preservation of initial r and loss before medial consonants. And we can usually (though it is surprising what difficulty even this comparatively easy exercise in phonetics presents to those not practised in or conscious of sound-making) reintroduce the r where it has been lost. But generally our natural method in such cases of false equation is to make one of the terms the same as the other (in our current phonetics) – if we try to rectify matters at all. An alliteration was, as I have said, under much more urgent compulsion to rectify. I fancy then that in the tenth century Beowulf was read so that one or the other of the current pronunciations of g was carried through the line. The first that occurred, probably, unless one was still a familiar word, and the other no longer current; or if there was phonetic difficulty. I imagine therefore that Beowulf l.1 was given a stop g in geardagum; while l.151 was given a stop g in gyddum (the absence of stave in geomore would not matter), because a (front) spirant was difficult in Grendel. But this would be quite artificial – the accidental product of a special preservation from antiquity by means of letters – and having nothing to do with normal tradition. It would be felt as quaint, even awkward and unpleasing, and would not be imitated in actual writing, any more than some of the odd rhymes and ‘quaint’ final es of Chaucer are today – except for a joke or an exercise in ingenuity.
We will now consider in detail, because of their importance in an enquiry into OE verse tradition, the ‘irregularities’ of the prosody of Maldon. I shall pay chief attention to alliteration, for in this point Maldon is specially interesting, and to it we can apply with some confidence the principles derived from the above argument: that ‘metre’ is independent of phonetic change (though it may be affected by it with the consent of the poets); and that current phonetics, not an impossible tradition of lost sounds, nor even orthographic tradition, are reflected in poetic practice – above all in a poem in a free ‘popular’ and unstudied manner.
We are now dealing, nonetheless, with a document. However unstudied in composition the original poem may have been, the fragment of it that has survived has reached us only because written down, and copied after that. Maldon has therefore, since it left its author’s mind and mouth, been exposed to the same dangers as other written works and more scholarly verses. Though it has had certainly a much shorter line of descent than many of them. Let the proven corruptions in Brunanburh – of which we have several copies – which must also have arisen in a comparatively short time, warn us that this short descent cannot be expected to have protected Maldon entirely. It plainly has not. Besides in this case the actual MS is lost and we have only a print made from a copy in the eighteenth century – a series which greatly increases the chance of errors and verbal disarrangements affecting metre. For the moment there is no need to go further than to point to the omission of grimme at the beginning of l.109, or the absence of the second hemistich of l.172, to see that all is not intact. All the same we will handle this question of corruption conservatively. Nearly all the ‘irregularities’ are capable of easy emendation; but some are not, and that should give us preliminary pause.
This question of corruption introduces a point with regard to ‘tradition’ that has not yet been considered. So far we have assumed the power of oral tradition to hand on verse intact, except for unobserved phonetic change in the actual words – with perhaps occasional consequential but deliberate changes to rectify any damage to metre that phonetic alteration may have caused. But though this simple process may occur in favourable circumstances, the machinery does not always work so well.
How did Maldon, for instance, reach a written form? It may have come straight from the author. Some clerk reverencing the memory of Byrhtnoth may have heard (or heard of) the poem celebrating his last battle, and knowing its maker have taken pains to take it down. But it is more probable that it had already gained some currency, and passed through several mouths, before this happened. Now oral tradition of this sort even when running through only a few years (especially dealing with such ‘topical’ matter) is not only less reverent of verbal detail than is scribal, and far less than editorial, it is actually not capable of uniform fidelity – except in special cases, where special efforts are made (as in the case of liturgical or sacred poems). It is prone to error. Even in a period when through practice memories are acquisitive and tenacious, reciters, though they may be themselves makers and conversant with rules, suffer from minor lapses of memory, and moments of inattention to form – when the meaning survives rather than the exact expression. This is the common experience of all who attempt the learning and recitation of verse. It is precisely in such gaps that are likely to occur substitutions of synonyms, slight disarrangement of words, and hasty patching of the sense with a line or so that is barely metrical, or is in a different metrical manner. These stand out from the main texture, and so catch the eye or ear of the later enquirer whose attention is concentrated primarily on metre. But it is necessary that he should carefully consider what is the proportion of the really ‘bad’ lines – as judged by the general level – and how far they contain in themselves any real difficulty for a composer of the capacity revealed by the poem as whole. For example, outlandish names; translation (as for instance in hoc signo vinces which gives Cynewulf some trouble in Elene); the necessity of sticking fairly close to actual words in reporting speech (to which is allied the general difficulty of oratio recta and dialogue in verse, and especially alliterative verse), and so on.
If these ‘bad’ lines in a longish piece are very few[18] in proportion, and are all lines which present no inherent obstacles (such as could, for instance, be easily remedied by the substitution of a synonym, or a slight rearrangement of words) – it is a proper assumption for the metrical historian, if not for the textual editor, that in this small residuum he has the results of the imperfections of repetition. There are 325 lines in Maldon. Of these some as received definitely break essential rules – rules the poem normally obeys: namely 45, 75; 224, 271, 288. From this small number,[19] 75 is probably to be deducted (see below). This leaves four in which the cause or nature of the corruption (if any) is not obvious on ordinary editorial principles – which are thus for an editor’s purpose ‘genuine’. But I should feel inclined to ascribe precisely these four to imperfect repetition; and 271 I suspect to be a line not by the original author at all (but of this more below). None of them present any inherent difficulty at all. 45, 224, 288 are all easily emended or rewritten with words the author knew.
I have attributed imperfection to repetition not composition because, on principles already argued, such imperfections could only arise in composition by a less skilled man than this author, or by such a man in peculiar circumstances: namely when extemporizing. Now extemporizing produces exactly the same defects as appear in the ‘bad patches’ of repetition. Only, unless with a very skilled person, and one moreover practiced in the art of stringing conventional half-lines together, they will occur more often. The defects are the same because the process is: they occur when the meaning required is in mind but not its metrical expression, and the mind is not agile enough to find metrical words to fit in the brief time allowed. But Maldon is not an extemporized poem. It is or was too long. It is traditional in language, and it uses a good many ‘stock phrases’, but is plainly not just a string of these. Compare 975 Death of Eadgar: 7 þa wearð eac adræfed, deormod hæleð / Oslac of earde ofer yða gewealc / ofer ganotes bæð, gamolfeax hæleð / wis 7 wordsnotor, ofer wætera geðring / ofer hwæles eðel, hama bereafod. This is in strict metre – but it presents the kind of thing an extemporizer with any knowledge of traditional verse could go on doing as long as breath lasted, and without a single breach of rule. I could do it myself.
Extemporizing was, of course, practised. It was one of the ways the craft was learned, and skill in it exemplified. We hear of it wherever the making of vernacular verse was a widely esteemed and popular accomplishment – as in Iceland. But although some metrically intricate stanzas are (in literary texts) ascribed to impromptu utterance, even to repartee, we must make due deduction for touching up, literary invention, and exaggeration of the speed of composition. And in any case we must imagine that really extempore utterance in anything like a difficult metre was limited to very special persons, and in them to rarer occasions and short stanzas. It was not a normal method of composition, even in celebration of immediately topical events. Cædmon was not disgraced because, when the harp came to him, he could not then and there sing without any previous thought.[20] But he did not occupy his quiet in the cowshed with making anything up for later occasions. As now verse was naturally composed in privacy, in the watches of the night, and trotted out later at the symposium. To this natural method we have many references: not only in the case of Cynewulf, the polished (if unexciting) poet – who says ‘At times I pondered and arranged my thought anxiously by night’ (Elene 1239 ff.); but also in the case of the maker of a popular lay such as Havelok, whose author asks prayers for him that haveth þe ryme maked / and þerfore fele nihtes waked. To return again to the skáld Egill: he said in pretended anger of his friend Einarr Skalaglamm, another poet, whom he accused of wanting to get a poem out of him: ‘Does he think I shall sit up all night on that account and write verse about his shield?’ (lxxviii). Einarr himself in a strophe referring to his own famous poem Vellekla says: ‘I wrote verse concerning the prince meþan aþrer svofo’ (ibid).[21]
The maker of Maldon (originally 400 lines long [pencilled above:] – probably more like 600) at the very least, we guess, and possibly much longer) did not extemporize this poem, and doubtless fele nihtes waked: at least this is a natural conclusion from its manner, subject, and general excellence.
