Some notes on verbal music

The April 1913 issue of Poetry magazine contained this now-famous short poem by Ezra Pound:

In a Station of the Metro

The apparition     of these faces     in the crowd:
Petals     on a wet, black     bough.

This printing of the poem in the New Freewoman on 15 August is notable for reproducing Pound’s original spacing, an innovation he explained to Poetry editor Harriet Munroe in a letter of March 30th: ‘In the ‘Metro’ hokku, I was careful, I think, to indicate spaces between the rhythmic units, and I want them observed’[i]. I imagine that the explanation was intended to overcome objections by either Munroe or her printer (or possibly both) to Pound’s irregular requirements, and that similar objections may be why the layout was dropped in all other printings of the poem.

Pound’s explanation is interesting in its echoing of the last of the three principles that Pound, HD and Richard Aldington had agreed the previous summer: ‘As regarding rhythm: to compose in the sequence of the musical phrase, not in sequence of a metronome’[ii]. As I understand it, the sequence of the musical phrase represents rhythm, while the metronome represents formal metre, and Pound’s visual disruption of the lines of the Metro poem are a scoring of the ‘rhythmic units’ that comprise its music. Without the spaces, and if one is willing to distort any kind of reasonable reading, it’s just about possible to read the first line as an iambic hexameter:

The apparition of these faces in the crowd

The sequence of the metronome gone mad.

If, on the other hand, one takes Pound at his word and treats the spacing as a kind of mutual notation, the line resolves into a three-stress line, with the number of unstressed syllables in each unit both variable and relatively unimportant. This notation disrupts the tendency towards imposing regular metrical/metronomic units so as the highlight the underlying rhythmical music of the verse.

The apparition     of these faces     in the crowd:

Petals     on a wet, black     bough.

The second line has four stresses, again not tied to any formal metrical scheme, but with the three final stresses balancing the opening three unstressed syllables of line one.

Of course, Pound, HD and Aldington were not claiming to have discovered something new; they were concerned with recovering neglected approaches to verbal music. In this context, it’s interesting that Pound appended his ‘Treatise on Metre’ to ABC of Reading, with the inference that the ideas expounded in the treatise should be read as exemplified in the example poems earlier in the book.

The final paragraph of the treatise bears quoting in full:

Beyond which we will never recover the art of writing to be sung until we begin to pay some attention to the sequence, or scale, of vowels in the line, and of the vowels terminating the group of lines in a series.


Let’s look at a famous, and metrically conventional, stanza from the canon:

The curfew tolls the knell of parting day,

         The lowing herd wind slowly o’er the lea,

The plowman homeward plods his weary way,

         And leaves the world to darkness and to me.

As I touched on when discussing ‘In a Station of the Metro’ above, one key aspect of vowel music that is not as much discussed as it might be is the contrast between long and short vowel sounds, where long vowels are vowels pronounced as they are named (A = ape. etc.). There we see a regular major scale of vowel sounds running through the first three lines, with the long ‘O’ and ‘A’ sounds serving as pivotal points around which the tune revolves. In the fourth line, the substitution of a long ‘E’ for the ‘O’ acts as a resolution, a kind of deceptive cadence.

It’s interesting to see how Gray manages to marry the iambic pentameter exactly to the rhythm of natural speech (something that’s all too unusual) while using the long vowels to highlight the key tones in the verse: tolls/day/lowing/lea/home/way/leaving/me serve as a kind of emotional summary of these lines. And similar patterns run through the entire poem.

I’m not contending that he sat down and decided on a pattern of long and short vowels as the framework for the ‘Elegy’ (although he might have), but that his control of that verbal music that he was using that his ear naturally discovered such patterns in the language as if instinctively; but it isn’t instinct, it’s the learned behaviour of the poet as artist, or, if you prefer, technique.


The focus on patterns of stressed and unstressed syllables that underpins our understanding of English metrics is a consequence of the fact that English is primarily a stress-timed language, where, in speech, the unit of time is the stress, not the syllable, with the intervals between stresses being approximately equal in duration, regardless of the number of unstressed (pr short) syllables between them.

However, this is not a universal language feature. The romance languages, Finnish and Japanese, for instance, are syllable-timed, with each syllable having approximately the same duration, and so you arrive, if I am allowed in a very over-simplified path, at syllabic metres, perhaps the most well-known being the 5 – 7 – 5 haiku, if we allow that the mora is more or less similar to an English syllable. Since the Renaissance, some English poets have experimented with syllabics, and this became more common in the 20th century, with poets like Dylan Thomas, Louis Zukofsky and Marianne Moore being among the most successful. Here’s the opening stanza of Moore’s ‘Bird-Witted’:

With innocent wide penguin eyes, three

large fledging mocking-birds below

the pussy-willow tree,

stand in a row,

wings touching, feebly solemn,

till they see

their no longer larger

mother bringing

something which will partially

feed one of them.

The stanza pattern of 10 lines with 9/8/6/4/7/3/6/4/7/4 syllables per line and a rhyme scheme of A-B-A-B-C-A-D-E-A-C is repeated five more times. But the stress patterns don’t repeat. Take the first line above:

With innocent wide penguin eyes, three

Four stresses, five unstressed syllables. Now the first line of the second stanza:

Toward the high-keyed intermittent squeak

Again the same number of stressed and unstressed syllables, but the pattern of alternation is completely different. It is the tension between the syllabic form and stress pulse that creates the unique music of this poem, variation and repetition, repetition and variation.


In a letter to Richard Eberhart dated 23 May 1954, William Carlos Williams talked about his ‘new measure’, the variable foot he believed bridged the gap between verse that had become too ‘free’ and the need for poetry to carry a tune: ‘by its music shall the best of modern verse be known’. The measure is linked to the triadic line he’d been working in since the 1940s, as illustrated by the examples he includes in the letter, and to his insistence on the American Idiom, ‘a language we hear spoken about us every day’.

The idea seems to be that each line is a metrical unit, a ‘foot’ that might be more like a yard, or a unit of breathing. Here’s a well-known example from ‘The Descent’, the first time he really used the triadic stanza form, which was included in the 1954 collection The Desert Music:

The descent beckons

                as the ascent beckoned

                            Memory is a kind

of accomplishment

                a sort of renewal


an initiation, since the spaces it opens are new places

                inhabited by hordes

                            heretofore unrealized

of new kinds—

                since their movements

                            are toward new objectives

But what’s going on? Williams is, it seems, counting neither stresses nor syllables. Neither are the individual lines, nor the triads, necessarily closed units of sense. Like Moore’s regular stanza above, the triadic form sets up an expectation of pattern for the reader, but that expectation is not fulfilled by a pattern of stress that mirrors the form. Instead, it seems to me, what Williams is doing is setting syntactic expectations that tend to be disrupted by the line and triad breaks o that, for instance, the word ‘even’ has two sets of syntagmatic relations with the word about it, one referring backwards, the other forwards, so that we have two senses to hold in our reading mind at the same time: ‘A sort of renewal, even’ and ‘even an initiation’. The music, the measure here is a kind of logopeia, not melopeia. However, there is a definite melodic structure here that depends not on the beat, but on the ‘pitch’ elements of language: assonance, dissonance and rhyme. This can be seen in the first triad not only in the obvious repetitions (descent/ascent/beckons/beckoned) but also in the sound variants on the ‘e’ vowel, with similar but different patterns in the second and third triads.


The essence of verbal music, of any music, is repetition and variation, the creation, disruption and resolution of the ear’s and eye’s expectations. This can be accomplished through rhythm, pitch (as I understand it in poetry, vowel and consonant patterns and the contrasting of long and short vowels), or any combination of the two. f course, I’m focusing here on poetry in verse; prose, concrete and visual poetries raise a whole range of other questions, but again the answers can, I believe, be resolved by appeal to these two primary elements.

[i] Letters

[ii] “A Retrospect” Pavannes and Divagations (1918)