Aloud is the only way you should read a poem. If hearing your own voice create the poem isn’t pleasing to you, the poem may not be a bad poem, and you may not be a bad reader, but one thing is certainly true: you’re out of touch with your tongue. If you’ve ever watched a baby learn to feed itself, you know that for a while everything goes into its mouth, usually along with a couple of fingers. Watching that same baby learn to talk reverses the process: for a while, everything — including those fingers — comes out of the mouth, every possible sound a human larynx can make. Or tongue can shape. The ones you share with the baby, the ones you respond with, enter its repertoire, and if the child hears sounds from more than one language, it can learn all the tongues it hears, perfectly grammatically, before ever entering the classroom. When you read a poem only with your eyes, you are throwing away that capacity to learn.
Yeah, poetry is a language — and verse is its grammar. And you can have bad poems, just as you can have poor conversations. A poem is a conversation between you and the person who wrote the poem, but you can’t take part in it silently. In fact, you can’t tell whether a poem is bad or not until you hear it, preferably in your own voice. Like that baby, you are trying out a new combination of sounds, and of words — even with a poem you think you know well from having scanned in its text. And like that child, you are hoping to hear back some confirmation that you are saying the words right. And if the poem is a good one, you will.
But just as there can be ungrammatical utterances, there can be ungrammatical poems. An ungrammatical utterance can speak as effectively as a grammatical one, but only in the mouth of someone who understands the grammar and knows when she’s breaking the rules. In my last turn I mentioned Frederick Rebsamen’s translation of Beowulf, and I want to return to it in order to talk about the rules of Anglo-Saxon alliterative verse. As I said then, I am writing a long poem in that prosody, and find some comfort in hearing that I am not the only one finding it difficult. The first difficulty Rebsamen comments on comes in his introduction: “Old English was a partially inflected language and a poet could therefore do things with word-order intolerable in Modern English, things that I have ‘corrected’ for the sake of clarity.[i]” Any speaker of English who tries to learn German, Latin or Russian understands what he means. English, in shedding its Anglo-Saxon inflections, developed a rigid word-order to compensate for the loss. So that’s one problem with writing modern English poems in Old-English prosody.
But it’s not the most difficult problem, especially when you’re not translating from Old English. As Rebsamen says, “Old English poetry has no stanzaic forms and no rhyme … except by accident” (xix). A line consists of two half-lines, each with two stressed syllables. The half-lines are separated by a pause in performance. A line thus has four stressed syllables, with a pause between the second and the third stresses. But it is not that simple. “There is no set number of syllables per line,” Rebsamen continues; “in Beowulf a normal line contains between eight and twelve.” What defines the line is alliteration on the strong stresses of the line, thus:
Then an alien creature cold wanderer could no longer endure from his dark exile bright bench-laughter borne to the rafters each night in that hall. The harp sounded the poet’s clear song. He sang what he knew. . . .” (lines 86-90 of the translation).
Note how, in line 86, the alliteration falls on the second and third stresses, “creature” and “cold.” There is a slight echo of “cold” in “could” on line 87, but it is not emphasized; at least to my ear, the first stressed syllable is “long-”. But note that it’s the second syllable of “endure” that alliterates with “dark” in the second half-line. And that’s a second feature of contemporary English that makes it harder to write alliterative verse: the English vocabulary has so far outgrown its Anglo-Saxon cradle that stresses can fall anywhere in a multi-syllabic word.
Simon Armitage offers a good example from his translation of Sir Gawain and the Green Knight.[ii] Quoting from his introduction: “A line like ‘and retrieves the intestines in time-honoured style’ (1612) might appear not to alliterate at first glance. But read it out loud, and the repetition of that ‘t’ sound — the tut-tutting, the spit of revulsion, the squirming of the warm, wet tongue as it makes contact with the roof of the mouth — seems to suggest a physical relationship with the action being described.” (viii)
So when Armitage translates line 1612 — “And has out the hastelettes, as hightly beseemeth[iii]”, or in the orthography used in the manuscript, “And hatȝ out þe hastletteȝ, as hiȝtly besemeȝ[iv]” — he has to change the alliteration from ‘h’ to ‘t’. Note that two of the three stressed syllables in his line are interior syllables, rather than the leading syllables of the Middle English line. But the Gawain poet was not working in Old English, but in a language now little-inflected, with a word-order nearly that of our own tongue
Rebsamen explains that “[t]he half-lines are tied together by alliteration of consonants or vowels, any vowel alliterating with any other vowel” (xix), but of course with the vowel having the stress in the word. Thus “awe” alliterates with “owe” and “eagle,” but not with “awake.” Moreover, “[t]he first stress of the second half-line, called the ‘head-stave,’ cannot alliterate with the second stress of that half-line, but it must alliterate with one or both stressed syllables of the first half line” (xx). He also sets forth the importance of primary and secondary stresses in a half-line, noting that the stress patterns include “bunching the two strong stresses at the beginning and then stepping down through secondary to weak, or bunching them both in the middle between weak stresses, or separating the two strong stresses with descending steps through secondary to weak, or approximating the Modern English iambic or trochaic measures. There are five of these patterns with a variation on one, some of them difficult to achieve in Modern English since secondary stress is not as clear or frequent today. . . .” (xx). And perhaps that last issue, with secondary stress and weak stress, is the hardest to address.
I know from my own attempts to write about Joseph Smith in Old English verse patterns that it is not easy to deal with the secondary stresses like the initial “could” in line 87 of Rebsamen’s translation above. I often find, in revising rough stone, rolling waters, that a line has only visual alliteration. When I speak it, it doesn’t sound right, and I have to work out where the stresses fall before I know why. Usually it’s because I confused secondary and primary stress in a line. Conversely, when a line sounds right despite not looking right, I have to study it out before I know why it seems good. The subject matter is no problem in a poem like that, and I have been more than once saved in moving forward in the narration by having to focus on the demands of alliteration and stress.
In fact, it happens often, when I am out walking, or running, or cutting up a stud, or repairing a chair, or hauling out the trash, that three or four new lines will deliver themselves while I am otherwise engaged. I capture them, on paper or in a voice memo, and take them home and look into the poem to see where they fit. I often have to cut out what I thought was pretty good before, to accommodate the new lines. I build the poem a line at a time.
But hold on, I hear you saying, that’s true of all verse.
[i] Rebsamen, Frederick. Beowulf : an updated verse translation. New York : HarperPerennial, 2004.
[ii] Armitage, Simon. Sir Gawain and the Green Knight. London : Faber and Faber, 2007.
[iii] Pearl ; Sir Gawain and the Green Knight / edited with an introduction by A. C. Cawley. — London : Dent, 1962. — (Everyman’s library ; no. 346), p. 111.
[iv] Sir Gawain & the Green Knight / edited by J. R. R. Tolkien & E. V. Gordon. — Oxford : Clarendon, 1930, p. 50.