Poetry, I’d like to think, proceeds from a generous instinct, not a selfish one. Whatever private torments might have been assuaged in our writing, we want to give these damn things away in the end. To have someone else want your poem for themselves, it must be desirable; to be desirable, it must be beautiful, or interesting, or both; and for a reader to find it so, it must exhibit some of the symmetry of form and organisation we find in the natural world. This last statement might sound a bit of a reactionary leap, and of course it’s as old-school as it comes: it’s been a cliché since Plato and Aristotle to say that the reason we find a piece of art satisfying is because it is ‘imitative of nature’. However, I persist in thinking of the poem as kind of a human-made natural object, our ‘best effort’ that we quietly slip back into the world, and against which the world can raise no serious objection.1 I appreciate, however, that I’m circularly justifying the kind of lyric poetry I myself want to write and read, not the non-lyric kind I do not.
Poetry is often compared to music, but most of the comparisons tend to be facile, and many are plain false. In one important way, however, I think they’re closely analogous. We might define music as ‘those noises that we agree constitute satisfying or emotionally meaningful arrangements of sound’. When we examine such noises, and look at the way that one note-event follows another, we find that their sequenced patterns converge on the same fractal statistic (the 1/f ratio of spectral density) that we find in natural dynamic systems, in everything from quasar emissions to river discharge, traffic flow, sunspot activity and DNA sequences. This pattern is often referred to as ‘pink noise’. It lies somewhere between ‘white noise’, where the relation between one note and the next is uncorrelated and completely random – in acoustics, it’s that shhhh sound where all frequencies are heard simultaneously, at equal power – and correlated ‘brown’ or ‘Brownian noise’ (after Robert Brown, who discovered Brownian motion), where the pitch of the next note is wholly dependent on the position of the previous one, through the application of an inflexible rule. Music generated on a white noise algorithm is ugly in its unpredictability, and Brownian music is just as ugly in its predictability. But if we hit upon something in between, something ‘pink’, we tend to find it beautiful: we sense that it corresponds to an ideal balance of predictable regularity and surprise. [See endnote 3 for connections between white and brown noise and the paradigm/syntagm distinction.]
Analyses of static forms in nature – the outline of a landscape, for example – reveal correlated ‘Brownian’ patterns. Perhaps this explains why we find Brownian noise acceptable in static visual art, but not in a dynamic, time-based art like music. (The non-dynamic, static and visual aspect of poetry – its typographic arrangement on the page – is often tellingly ‘correlated’ in its stanzaic and lineated symmetries.) However, when it comes to natural dynamic processes, we find pink noise dominating. It appears to be the characteristic signature of complex systems, i.e. those which display non-random variation. The changing content of our sensory experience seems to hover around the pink noise mark. This sensory music is as much a product of the nervous system as of nature: the input received at our physical extremities can be near-chaotic white noise, but our brains, through the application of arbitrary but evolved ‘rules’, filter it down to pink – screening out the irrelevant noisy data and leaving only those patterns of change which have become useful to our specific evolved intelligence. The wholly dynamic, time-based medium of music is dominated by such a rule; the pattern of regularity and variation in its pitch and volume (just like human speech, incidentally) matches, perhaps more perfectly than any other kind of art, the spectral density of our flickering perception of the world.
It seems reasonable to assume that our brains also perceive the dynamic system of the successful poem as similarly balanced. (The poem itself can be thought of as operating, in its way, like a miniature nervous system, filtering the pink noise of our own perception even further, leaving only a pattern of locally significant data.) The best poetry has nothing so easily quantifiable as note pitch and length, but I’d suggest that were we able to measure accurately the ratios of its concrete and abstract speech, its light and its dense lines, the pattern of its metrical agreement and disagreement, we would see something identical emerge: a fair echo of nature, of its balance of correlated and uncorrelated, randomness and self-similarity. And, perhaps, a more crucial equilibrium: that of predictability and surprise, pattern and variation, familiar and unfamiliar, known and unknown.
Wholly familiar ‘Brownian’ poetry consists in the mere rehearsal of what the reader already knows to be the case, and it unfolds in a wholly predictable manner. It fails because it doesn’t surprise. The brain hears nothing but its dull coincidence with rules it already knows too well. ‘White’ poetry is all unfamiliarity and novelty and discontinuity, and fails because it does nothing but surprise. (This sounds just dandy; but there is nothing so predictable as an infinite series of exceptions – nor, as most readers correctly sense, is there anything so easy to create.) All its elements, moreover, are sounded at equal power, and with no sense of a background against which a salience might appear. If our aim really is epiphany, the poem must demonstrate a move from the known to the unknown, which we might define here as making an uncorrelated leap from a correlated position. But it can only do so by actually making it – and therein lies the risk and seriousness of our word-game.
The suggestion remains controversial, but were we able to measure and quantify these things in an accurate way, there would be nothing to stop us automating the process. While I believe that one day it will be entirely possible to write a great poem with a computer, we will probably have ‘gone biotech’ by then anyway – and in making the smart move from carbon to silicon, also quit all the pointless fear, sweating and coughing, and limited our orifices to an optimal minimum. Generative poetry is unlikely to seem such a big deal to our bionic scions. (Indeed such an elusive poetic algorithm already seems to exist in the parallel-networked meat-machine of our brains, and it doesn’t seem miraculous or impossible to us. It probably should.) We are far closer to devising such a successful generative process for music, though, and the fact that poetry and music are in many ways comparable systems might suggest that poets lack only a full description of their own compositional process. The perennial worry is that poetry written in this way would shed its ‘humanity’ – but you only have to look at the way traditional music skills have been ported over to programming to realise that those fears are quite unfounded. On the contrary: programmers invest their music with as much humanity and human expression as any other language, and the laptop turns out to be as humanly responsive as any other instrument. Our fear is just the standard human wariness over new means of production. Similar misgivings were initially voiced over cameras, typewriters, the pianoforte – and the printing of books themselves. Even if we are many decades from our robot Emily Dickinson, there is no good reason why computers should not be at least a little useful to the poetic art fairly soon. [For a note on poetry, computers and ‘humanisation’, see endnote 4.]
While we may be almost as many years from a comprehensive model of the poem as we are for one of the brain, certain aspects of the poetic art are more easily described than others. Just as music is amenable to fairly systematic description, so is the music of poetry. It’s in lyric that we can most clearly see our ‘pink’ balance: poetry naturally refines the music of language to something close to the ‘pink’ ideal – something correlated, modulated by something variable – and it’s that rule that I’ll spend the rest of this essay discussing. In its lyric aspect, poetry displays the most instantly recognisable emblem of its art. This takes the form of a strange default: a balance of variation and repetition, composed of shifting vowel and patterned consonant, of both an airy and a stop-heavy music.
1 The occasional use of the phrase ‘organic poetry’ for ‘free verse’ is just substituting an error – and a pretty stupid one – for a misnomer. In the organic, symmetry is everywhere. Once wholly ‘freed’ from every aspect of formal patterning, a poem may indeed be ‘organic’, but only like some kind of diseased amoeba. A better defence of the kind of free verse practice that abjures all formal patterning is the Lawrentian argument that it more closely represents the dynamic shape of spontaneous thought; but even this tends to ignore the fact that thought itself is highly rhythmic, and that spontaneous thought is often the least original we have. The ‘flash of inspiration’, welcome as it is, has given spontaneity an undeservedly good name. First thought = worst thought, speaking personally.