While some languages have prominent features of vowel length and of tone, the quality providing the most potential for linguistic contrast in English at the level of syllable is stress. While there are many other ways of looking at them, our sentences can be considered as arrangements of weak and strong stresses built up from a series of words, each of which has its own prosody: a ‘lexical stress’ that we learn simultaneously with its sense, and seem to store in the same location in the brain as a synaesthetic unit.
The complex phenomenon we breezily denote with the word ‘stress’ is an aspect of our speech composed of three constituent elements: loudness, pitch and length (which, within the flow of speech, can lead to a fourth: rhythmic displacement). When stress is salient, we call it ‘strong’; when not, ‘weak’. Sometimes only particular aspects of stress will be exaggerated, but we’ll nonetheless still characterise them as ‘strong’ or ‘weak’. Loudness is a common enough weak-strong-stress indicator in conversation, but both the yell for attention and the sotto voce euphemism are, through their shared quality of ‘salience’, both kinds of stress.1 (Significant increases or decreases in loudness are very rarely used as a salient foregrounding device in poetry, and are only ever indicated by typographical cue: this usually involves the novel use of font size, type or colour. The convention, such as one exists, is that ‘caps lock equals shouting’; far less frequently we might see small or faintly printed fonts indicating whispers. This takes advantage of our instinctive SMALL IS QUIET / BIG IS LOUD or VISUALLY FAINT IS QUIET / VISUALLY INTENSE IS LOUD synaesthetic metaphors.2)
However, in most English conversation, those small increases and decreases in loudness are accompanied by small rises and falls in pitch, and generally experienced as indistinguishable from them. In each strong stress, we see in the syllable a slight but perceptible rise in its loudness, a rise in its pitch and a slight lengthening of either its vowel or the temporal gap on either side of it. (Often when we detect such a space between words, we are actually hearing nothing of the sort; most speech blurs the sound of the last word into the next, and the definitional boundaries of each individual word are more psychological than physical, just as they are with the visual perception of discrete objects. Indeed, I suspect deliberate phonosemantic patterning has the effect of blurring discrete semantic boundary in a way directly analogous to the blurring of visual object-boundary in certain low-lit or painterly circumstances).
The classical iamb was measured as one short + one long vowel, equivalent in duration to a brach + two brachs: in musical terms, this is close to a single quaver triplet followed by two tied quaver triplets. The difference between the durations of the two syllables of our own stress-based iamb is nothing like so pronounced. The ratio is closer to 2/5ths to 3/5ths, and usually even subtler: it is more or less identical the so-called ‘swing quaver’ in jazz, where each alternate eighth-note is slightly delayed. And just like the swing quaver, it’s a flexible ratio which can be ‘heavy’ or ‘light’ depending on a number of variables – including dialect and emotional tone, both of which can also drastically affect the rate of syllabic delivery.3
This difference in vowel length would be noticed even less, were it not almost invariably emphasised by a slight rise in loudness (as in the jazz quaver) and in pitch. Pitch, in emotionally neutral and non-urgent conversation, is usually centred on a note in the lower third of the speaker’s natural register, and rises somewhere between the interval of a minor second and a minor third on the s-stressed syllable. (In a few UK accents – in parts of Fife and Aberdeenshire, for example – the s-stressed syllable drops below the pitch of the weakly stressed syllables, lending a ‘questioning’ character to the speech; certain Scandinavian accents do much the same.) This means strongly stressed vowel-sounds have a distinctive formant and an identifiable character, in contrast to weakly stressed vowels, which have a short neutral sound often close to a schwa; these weak sounds are relatively characterless, regardless of how the vowel is represented by its spelling.4
In the context of a normal sentence most monosyllabic content words – nouns, verbs, adjectives, adverbs and proper nouns – will take a strong stress, and most monosyllabic function words a weak one. Function words indicate how content words should relate and orientate themselves to one another. In adult speakers, function is a learned and automated procedure, and function words rarely need to be consciously processed, unless you’re reading Henry James. Their weak stress helps to ‘background’ them, and keep the brain’s CPU focused on the content. New or especially important content is flagged by an additional, contrastive stress which is intonational in character, and usually involves a sharp rise in pitch.
We can immediately see that any prosodic analysis which treats stress as a fixed phenomenon is going to create a terrible oversimplification. Depending on what kind of stress we encounter, its constituents – loudness, pitch and rhythmic displacement – will be mixed in a very different way, and have a very different quality. Consider, for example, the stress I just asked you to put on the word ‘kind’: the use of italics implies a foregrounded rise in vocal pitch on the italicised vowel – perhaps as much as the interval of a sixth from the median pitch – and a short (mora-length) pause after it, before the next weak stress. The stress-shape of speech should really be analysed in terms of its suprasegmentals – the sound-qualities which extend beyond isolable phonemes, e.g. pitch, timbre and so on. Poetic prosody is just a useful, simplified and codified way of speaking about a phenomenon that has received a far better scientific description in the field of phonology. In poetry we focus on how written poetic cues are translated into real-world, spoken stress – but not on the precise nature of that spoken stress itself, since there can be no ‘ideal performances’ of the line, even though one performance may be strongly proposed.
For this reason, the way I’ll represent sense-stress pattern in metrical analysis will be necessarily crude. Our metrical symbols are intended only to describe a strong-weak pattern; however, pitch – a crucial aspect of stress when we analyse sense – is not something that can be accurately indicated in this way; nor can vowel length. (Our Germanic speech-rhythm is primarily qualitative, but not wholly so; and there are metrical circumstances – such as very short lines – where quantity, in the form of vowel length, can become salient too.) However, the use of strong-weak symbol patterns is an extremely useful shorthand in determining underlying metrical templates, and noting which lexical stresses deviate from them; but since this doesn’t tell us why they deviate, the symbols themselves should not be reified and then talked about as if they had any real-world existence. If we are interested in gaining an understanding of real stresses, we have to consider not only their strength but also their quality – and since that performed stress is intimately bound up in ‘understood sense’, our final prosodical analysis must never become divorced from a discussion of sense and contextual meaning.
Despite T. F. Brogan’s earlier warning, I feel that adding an intonational aspect to notation is necessary if we are going to be true to that aim. Words spoken with no pitch accent or de-accent sound literally ‘robotic’. They communicate some sense, but leave us unable to judge whether the speaker either understands or identifies with it. Our understanding and interpretation of a line is revealed and broadcast in our vocal performance. If we are to regard the poem as a speech act, we must make some account of this performance, or at least its sensible variables. As ever, the purpose of this approach is to restore some agency to the humans on either side of the poem, and their role in the production of its meaning. In a poetic context, ‘stress’ is always a performative and interpretative act.
1 Another reason for my queasiness over the word ‘unstressed’: there are few words so apparently understressed as the soundless mouthing of the word ‘cancer’, but a form of stress it is.
2 Mark Z. Danielewski’s mind-crushingly disturbing House of Leaves provides one of the cleverest examples of this technique: the word ‘house’, wherever it appears, is printed slightly lighter and slightly askew (and blue, in the hardback version), a technique that seems innocuous enough – but as the house slowly gains in menace, the reader begins to actively dread its next appearance.
3 The most important factor determining the degree of iambic ‘swing’, as we’ll see, is whether the iamb occurs in a line derived from 4-strong, or in iambic pentameter; in the latter, the iamb can be flattened out to something close to the even jazz rhythm known as ‘straight 8s’. Whether the jazz quaver developed in imitation of African-American English speech is moot, but it was likely influenced by it (although certain folk musics which fed into the mix have a strong swung quaver too: that heard in traditional Scottish music is often so pronounced it sounds like John Knox angrily stapling pamphlets together). The iamb surely does not originate, as certain authors still sweetly claim, in the flub-dub of the heartbeat. If it did, we would expect francophone hearts, in contrast to anglophone hearts, to deliver a nice, even ‘boum-boum’.
4 However, I once met someone who pronounced all schwa with a touch of the vowel colour of their written form (i.e. they would say ‘vowel colour’ just a little like ‘vow-ell cul-oor’) – the result of a childhood spent largely in isolation with only books for company.