The chapter which follows this makes some notes towards a working taxonomy of poetic metaphor, analysing it at the levels of conceptual form and textual expression. The scheme I’ll initially present here is a rather anecdotal description of the ‘character’ of metaphoric ground (or ‘isomorphism’), the underlying conceptual mapping. This is one I used for years when teaching; it is itself largely a metaphorical description, which all non-scientific accounts necessarily are, but the ways in which it was patently inadequate struck me as interesting. An examination of its flaws led me to propose a more fundamental revision: a new scheme of four tropes.

Metaphor depends on there existing a ground between (or blend constructed of) two domains. This takes the form of an isomorphism, a contentual overlap, which defines a set of shared properties. The character of the isomorphism will identify the kind of properties shared by the T and V. This is not really a poetic or even a linguistic issue, but a conceptual-semantic one. For years I taught – mainly because I had to teach something – the little mnemonic quintet of the ‘five Fs’, which essentially sketch in ‘the nature of content’: form, function, flow, feel and free association. As useful as I found these categories over the years, I always knew them to be completely flawed. I mention them here only to introduce a second and more conceptually precise scheme, which I believe may point the way not only to a clearer understanding of metaphor, but also – because they essentially describe not just the nature of passive isomorphic content but the active semic means by which one thing can join another – a new scheme which replaces our old ‘four tropes’.

But let’s begin with my five fs. As will be obvious, the categories are porous, and can’t – and shouldn’t – be kept distinct: the kind of instantiations we deal with in the real world, and which the best poetry reflects, are usually complex peristases with functional, dynamic and tonal qualities, not single aspects of single things.

Form: A list of formal properties might answer the question What kind of thing is it? With ‘moon’, we might note the properties of crescentness, brightness, whiteness, distance, etc. Using these as a ground, we might form a strong comparison with a toenail clipping (crescent, white), or a boomerang (crescent, kind-of-distant), and so on.

Function: A list of functional properties might answer the question What does it do? or What is it for? or (if we want to indulge the anthropomorphic conceit of ‘the intentional stance’) What does it want to do? To stick with ‘moon’ – it sheds light, it draws tides, it drives humans mad, it sometimes occludes the sun in an eclipse, and so on. So we might compare it to ‘a strong lamp used in an interrogation’, ‘a hallucinogenic drug’, ‘a magnet’ or ‘a dark coin covering a bright coin’.

Flow: Though I really mean dynamic. A list of dynamic properties might answer the question In what way does it move or change over time? The moon floats, it follows an orbit, it librates, it waxes and wanes, etc. The slow rounds of a milk-float might coincide with the idea of lunar orbit; a dark mood might wax and wane like the moon. (Though note that things get messy immediately: in the first example, formal properties are borrowed – ‘whiteness’ – and in the second, tonal and lyric properties.)

Feel: Though ‘tone’ is the better word. A list of tonal properties might answer the question What does it feel like? This alludes to the more subtle and subjective tonal, emotional, sensual and aesthetic qualities of the T. The moon ‘feels’ cold, lifeless, distant (of course it literally is all of those things), but it also might remind us of death, of emotional sterility, of ‘distant relationships’. Incidentally, the role of the object as cultural symbol – e.g. the red rose as a symbol of love – is also subsumed into this category.1 These are sometimes claimed as pure Peircian symbols, or what I call ‘asymbols’ – meaning they have zero coincidence of attribute, zero ground, since the connection between signifier and signified is entirely arbitrary and learned. However, these symbols invariably turn out to have been formed, in part, by shared tonal concerns. To take a relatively recent example: in the allegedly pure ‘symbol’ the late musician Prince briefly adopted instead of his own name –  – one is obliged to observe that Prince and ‘the Prince symbol’ do have certain shared tonal attributes. Both demonstrate a certain pomposity, pretentiousness and baroqueness, and the symbol itself has isomorphic similarities with everything from astrological symbols and male-female biological symbols, to Jimmy Page’s old tag in Led Zeppelin – all of which were likely to be contributing factors in the new symbol’s design, and all evidence of its patent non-arbitrariness; the symbol was symptomatic of many things in Prince’s character, from his rock heritage and sexual identity to his mental state. Humans cannot select anything arbitrarily. Even an asymbol selected by an aleatoric process will immediately have some degree of iconicity as soon as it takes a form; the total infection of the human realm with a kind of primed, Higgs-field tonal ether means that any attempt at pure ‘asymbolicity’ is always doomed. However close it might approach it, the sign can never attain true independence from its referent.

Free association: A list of associative properties might answer the question What does it put you in mind of? This is a disease of degree, of course, the disease in question being unverifiable subjectivity, and ‘free’ connotations are the least-tightly bound to the semic core of the domain – at least as far as everyone else is concerned. The moon might remind you of a cold and unfeeling parent, and the reader might construct enough of a ground to make the trope work; but it might also remind you of Newcastle Brown Ale or blue sweatpants, for reasons which are wholly personal or random, and probably better kept to yourself. Vigilance is required on the poet’s part; too often wholly subjective associations are presented as if they were widely shared, the poet having mistaken their apparent strength for currency.

That pretty much covers isomorphic ‘character’. However, this account is really only useful as a handy little mnemonic. As I’ve mentioned, poetic metaphor in the field will often combine elements of any or all five categories, and even the illustrative examples I gave were helplessly and freely borrowing from those categories; additionally, the metaphor will likely have dynamic, causal and contextual dimensions too, as well as areas of lyric, rhythmic and phonesthemic overlap. We embroider peristases, not isolated substantives: the moon will rarely be alone, but part of a larger imagistic landscape – ‘disappearing behind the rank clouds as a clear idea suddenly lost to confusion’, say. Poetic metaphor is no different from the kind of complex metaphor we often innovate in real life. The other day I tried to pour from a bottle of wine upon which the top had been screwed back on. My partner observed, ‘You’ve left the lens cap on.’ The metaphor here is formal, dynamic and tonal in nature. Yes, the mapping is A LENS CAP IS A SCREW TOP based on the ground ‘ABSOLUTE IMPEDIMENT TO AN ACTION WHICH REQUIRES UNIMPEDED FLOW’, but that wouldn’t have been as sharp a comparison unless it was understood with the dynamic and tonal isomorphism of ‘an impeded intentional action involving a befuddled agent’ and within the larger contextual domain of ‘drunken occasion which legitimates the use of distant metaphors’. (These subtler connections and dynamic contexts are often missing from accounts of the operation of conceptual metaphor; our descriptions are often suspiciously oversimplified.)

For years, I found these rough-and-ready divisions a fairly useful way of interrogating the metaphoric potential of two interacting domains. However, I’d now like to propose a more serious scheme. It will be self-evident that the categorical division between ‘functional’ and ‘dynamic’ attributes is rather arbitrary; it will also be clear that the business of ‘feel’ is a laughably subjective one, and that ‘free association’ is off the scale. Therefore I’d suggest we might categorise the properties of a domain according to the following types:

Formeme: A seme which describes contentual, formal and spatial elements.

Aeteme: A seme which describes functional, causal, temporal and dynamic elements.2

Patheme: A seme which describes all elements which we identify as carrying an isolable and distinct feeling, emotion, tone or ‘meaning’.

Aseme: An anti-seme of connection enforced by context or frame alone, i.e. between two elements via personal, culturally learned or arbitrary association. (Some will baulk at my conflation of these categories, but these associations are all ‘forged’ by outside forces, as opposed to the others, which are ‘domain-emergent’, and based on the intrinsic motivation of the sign.) Asemes are by definition content-less, empty, and mere vectors of connection.

The term ‘aseme’ reifies the unreifiable. An aseme carries a real connection for the author or a cultural group, but remains a nonsense for the reader, or a culturally excluded group. Old Betamax videotapes plunge me into the feeling of a very specific and miserable loss, for reasons too dull to explain; either way, I doubt you share my feelings for them. On the other hand, I know what you were trying to convey when you brought roses to the hospital bed of your new Amazonian girlfriend (who ignored a red traffic light while out riding her new bicycle, reasonably assuming it was some kind of street decoration), but being a newcomer to Europe, she is broadly confused by your gesture. Either personal or cultural aseme effectively operates as a Peircian symbol, i.e. it creates an ‘asymbol’.

By contrast, as we’ll see, we have evidence that we share pathemes, even if we cannot readily supply synonyms for them; we agree what things feel like. However, asemes are another matter. Your aseme, which, let’s say, is a tone that will forever arrive only with Après l’Ondée, and is one of unbearably sweet melancholy – because she wore it that afternoon, the day she told you that what she felt for you was deep and true, but she was nonetheless running off with a Polish lawyer who owned a big house in Thetford Forest. But for me, your merely telling me how it makes you feel is an arbitrary connection, although it may well taint my experience of the scent in the future. On the other hand, those roses she left for you – on this we can both agree – seem a sure symbol of her strong feeling, although we might now question its sincerity. Although the first association has been forged in an arbitrarily and personal way, and the second by culture – both associations are learned, and both the perfume and the roses are asemic. (All metonyms are ‘asymbols’ until you own the rule-book.)

Now we can begin to see how these semes may also propose the new tropic scheme I mentioned earlier. Formeme and aeteme represent the universal principles of (contentual-paradigmatic) static correspondence and (functional-syntagmatic) dynamic link respectively. These are the means by which metaphoric inter-domain and metonymic intra-domain processes are driven. Patheme and aseme represent such human meaning as falls outside these two axes. The patheme is the seme of ‘feeling’, of which more shortly; and if you recall my definition of abstracta as ‘qualia with names’, its potential relation to ‘the symbolic’ is obvious: I will argue that the patheme is simultaneously the seme of meaning, and points to the deep consubstantiality of ‘meaning’ and ‘feeling’ in human experience. Aseme covers the business of arbitrarily forged connections I alluded to in ‘free association’, and creates the ‘asymbol’ (i.e. the Peircian arbitrary symbol).

We now have terms for what a thing is (if we can use ‘is’ in the limited sense of ‘the sum of a thing’s formal and contentual attributes’); what a thing does; what a thing makes you feel; and what you associate (or we associate) with that thing through arbitrary and learned connection. These terms seem to correspond to four distinct semiotic categories:

  1. the formeme to the correspondence-function of the metaphoric (or ‘semi-iconicity’);
  2. the aeteme to the directly relational function of index;
  3. the patheme to the representational function of icon (this will take some explaining) and the animus/quale of the symbol, if we consider tone and abstract quale consubstantial categories;
  4. the aseme to the function of the asymbol, the pure Peircian arbitrary symbol.

The patheme

What follows is a short discussion of the most controversial member of this group, the patheme, whose properties are difficult to think of as anything we might consider a seme at all. (The aseme, the anti-seme of ‘meaninglessness’, merely links the two categories of arbitrary personal association and semiotic ‘asymbol’, which are structurally identical.) I have long been plagued by the thought that much of what ties a poetic passage together and moves us is ‘tonal’ in nature, yet this goes largely unmentioned and unanalysed. By ‘tonal’, I mean communicated through its feel, tone, ambience, atmosphere, ‘vibe’. The seme of ‘tone’, the patheme, is largely non-articulable; it is, I am convinced, the dominant seme of music. (Its inarticulable nature, its coterminosity with ‘feeling’ itself, is the reason works on the aesthetics of music are in relatively short supply; indeed, the entire subject often strikes us as almost redundant or tautological.3) As the logical positivists and ‘language school’ of analytic philosophy were eventually forced to admit, there is a whole realm of human thinking and feeling that is resistant or alien to linguistic articulation; there are, it seems to me, an infinite number of discrete human moods whose transcendental qualities seem almost defined by their being inarticulable, by the fact no word can successfully adhere to them, despite their utter specificity. I would claim that these feelings, while they cannot be adequately ‘described’ in words, can nonetheless be provoked by words.

To provide a means of entry into the idea, I offer this brief diversion into the sign-system of music, which is, I believe, entirely ‘pathemic’. Music’s signs are connected to its ‘referents’ neither by indexical nor by symbolic means, and they invoke neither formal nor causal attributes, but ‘emotional states’ – descriptively vague but experientially specific loci in the realm of feeling and the affections. The situation is complicated, however, by the fact that music has no referents in the linguistic sense at all, since it lacks anything like a denotative lexis. It creates its signs instead via a recursive process, wherein its isolable components are given meaning beyond their most generic associations only by self-contextualisation – i.e. the melodic, harmonic, rhythmic, timbral and thematic matrix which they simultaneously create and occur within.

For example: the ‘emotional’ function of the A note above middle C on the piano will change radically as another note is added to adapt the harmonic context, and provide a different emotional cue. Play the A, and the effect on the Western listener is approximately: hmm … (single note, no context, entirely neutral);4 add the C above, and the result might be mmm … (the root note of a minor third, a little sad); add the F below and we might get an awww … (the major third of a major triad, the ‘happy’ effect considerably heightened by the fact that one had ‘misinterpreted’ the sadness of the minor third); add the B♭ below the F and we might get an oh … (the addition of the low B♭ turning it into a sweet, lounge-bar chord, but also one built on the subdominant relative to the F triad, the less stable and more bittersweet chord of the plagal cadence); and if we add a G♭ below that, we get huh? (the low note resulting in a tense, unsettling and dissonant chord only jazz musicians would bother to give a name: G♭Δ7#9#11, for the record). I could describe its broad emotional effect of this last chord as before, but the point to note is that as harmony increases in complexity the number of words required to describe its effect will increase too, and explicit consensus regarding its emotional ‘sense’ is concomitantly harder to achieve. But the musical sign ‘440Hz’ has no intrinsic musical ‘meaning’ whatsoever, despite what certain synaesthetes might try to tell you.5 [For a note on music synaesthesia and how it might relate to poetry, see endnote 15.]

I realised recently that, for years, I have regarded the following explanation as something of a truism. After committing it to paper, it seems considerably weirder and more speculative than I had hitherto admitted to myself. As I discussed in the first essay, music has a miraculous ability to map itself to the landscape of human emotion and also to generate ‘emotional sense’. Some semioticians claim music is polysemic, i.e. that its meanings are simultaneously multiple; but this is down, I think, to nothing but the text-centred life of the individuals attempting the description. My own experience has always been that music conjures something absolutely specific but self-evidently inarticulable, and that this belongs to a realm of feeling to which language has little direct access. For this reason we tend to characterise it as ‘spiritual’ as much as emotional, in recognition of its plainly transcendent – or at the very least language-transcending – qualities.6

Poetry enjoys something of the same ‘spiritual’ reputation, and for much the same reasons. It attempts to say things that language cannot usually contain, through increasing language’s combinatorial, idea-blending power – which has the simultaneous effects of destabilising its consensual meaning and increasing its capacity for personal interpretation. The first essay described how poetry moves its whole internal value-system closer to that of music’s, reinforcing the synaesthetic link between sound and sense through the strengthening of the iconic function already present in speech. I think the solution to the question What kind of sign-system is music? lies in that very approach. Semiotics struggles to shoehorn music into anything even vaguely resembling the standard terminology we use to describe the operation of the sign, and ‘why stuff means other stuff’. The trouble is that music nonetheless appears to mean something, inasmuch as it produces effects that are clearly and directly related to our experience of it; it also appears – to judge by its replicable success, and our clumsy verbal adumbrations of our encounters with it – that we are experiencing roughly the same effects. The trouble is we can’t decide how it achieves this connection.

If music works in a way that is neither indexical nor symbolic, then pretty much all that’s left to us – without inventing another mysterious, etheric category of semiotic connection to accommodate the experience – is to claim that music is somehow iconic. As we’ve seen in the phenomenon of ‘iconicity’, an icon is a sign that shares enough of the content of its referent to be ‘motivated’, i.e. it is a sign so close in its attributes to the thing to which it refers that it immediately invokes that thing – as the imitative word ‘thump’ mirrors the sound of the action ‘thump’, or a picture of a chair immediately conjures up the chair itself.7

Here, we have a kind of empirical proof: the idea that the relationship between a musical passage and its emotional effect is iconic might chime deeply with our own experience. Where language refers to things and to concepts,8 music appears to ‘refer’ to emotions and feelings – but with such immediacy, such speed, we can hardly get a cigarette-paper between the music and our emotional response to it. ‘Refer’ seems a bad verb: the presence of music and the arrival of our feelings ‘about’ it seem almost simultaneous, and to elide the ‘about’ itself. Our feelings are less evoked or invoked than provoked, and the usual sign-system of signifier – (contextually derived index) – signified has been short-circuited, with the ‘feeling’ arriving as a kind of inseparable daemon of the sound itself, in a way that could be described as forcefully motivated. Only an iconic procedure could account for the instantaneity of the experience (unless we are to claim that there is no semiosis involved, and that the response is as purely physical as the pain derived from hitting your thumb with a hammer). But given that music’s ‘referents’, those provoked feelings of ours, are apparently so abstract and visceral … What on earth might that shared iconic content be? Where, in other words, is the sonic representation of our emotions that acts as just like the picture of our chair?

Despite the fact it has no isolable signs equivalent to the lexis of language, music comes on for all the world like a coherent sign-system. It has complex rules of pitch relation, of consonance and dissonance; it has something very close to a grammar and a syntax; and for centuries now theoreticians have explored parallels between the rhetorical structure of music and literature. Beginning with Joachim Burmeister’s Musica Poetica in 1606, theorists of the baroque era were especially keen to pursue analogies between musical figure and rhetorical structure as a means of systematising the expression of an emotional state, or the way in which such an expression will imprint itself on the listener’s affections. (Rhetorical terms would seek out their musical analogies: ‘anabasis’ might refer to an ascending passage which expresses rising spirits or heightened images; ‘litotes’, to musical understatement; ‘aposiopesis’ to a tacet in all voices that echoes the inability of an emotionally overcome speaker to proceed, and so on.) I am convinced that melody and harmony are mapped to surface sense and complex meaning respectively in a non-trivial way.

This sign-system, then, shares many of its characteristics with the sign-system of language, from its division into phrases to its argumentative structure. And while music has no equivalent to lexis, its various gestures nonetheless seem ‘tied’ to specific emotional states as effectively as lexis seems to refer to objects and to concepts. This has arisen, I suspect, from three convergent frames. The first is physical law; the second is a prelinguistic, learned association between vowel and human emotion; the third is language acquisition, in both its realised and potentiated state. In the first instance, music draws on the natural phenomena of isochrony and the harmonic series to organise its sonic material. In the second, it draws on strong prelinguistic, presymbolic and likely intrauterine connections between intonational vowel and human emotion. Finally, it arranges or maps that material to the recursive (and probably neurally embedded) syntactic structure of language itself.

In language, consonant is the means whereby we win the differentiating power to create a lexis capable of indicating a vast range of distinguishable referents. Without consonant, we can have no denotative system. I mentioned before that speech is a complex form of song; I strongly suspect music is, in part, both perceived and processed by the brain as a form of de-consonated speech, and can be considered as one long, infinitely differentiated and complex vowel-sound, carrying just the same kind of timbral, intonational and amplitudinal information as our speech-vowel.9 While we know music engages many parts of the brain simultaneously, I feel that it is also understood as a linguistically processed intonational prosody, a speech-minus-lexis – and parsed as something incapable of denotation, but possessing the same potential for the carriage of emotional information as the vowel does in speech. Music purely connotes.

If vowel is the principal carrier of emotional content, and we do indeed process music as vowel-speech – this allows us our iconic operation. In the same way that we can closely follow the emotional contour of an argument on the other side of the hotel wall, so we heard vowels in the womb. Our intrauterine experience also contains the human voice, de-consonated by the wall of the womb – which acts as a low-pass filter, cutting out the higher frequencies of consonants. While the association between vowel and expression is, of course, complex and guided by the reinforcing contexts of facial expression, gesture, etc. (although some facial expression is innate), it seems entirely possible that iconic links between intonational contour and emotion could be forged there and then, possibly at the level of neurotransmitter. (Six-month-old babies respond to the beat of their mother’s speech; emotions are shared via the hormones associated with them; certain feelings could then be firmly and viscerally associated with the cadences of a happy conversation, etc.) This is a material description of an idea that partly coincides with Kristeva’s take on Plato’s chora, by which he originally intended the matrical space in which The Forms were held. Kristeva sees this as an intrauterine presymbolic function, one which opposes the phallocentric symbolicity of language itself.10 Thus the patheme might be thought of as a ‘choric’ seme. By the time we are born the link is effectively iconic, because feeling and intonation are regarded as aspects of the same phenomenon, and one induces the other.

The patheme relies upon an immediate identification between sign and feeling. It has no ‘verification process’, whereby we might check the ‘reasonableness’ of our interpretation. In the case of a metaphor, our vehicle and tenor might share a formeme or aeteme; we inspect our ground for shared content to reassure ourselves that, say, the fluffy correspondence we noted between this cloud and that sheep was a sane one. However, the ‘verification’ of the patheme is consubstantial with its experienced effect. Its exchange can only take place in a less symbolic, less mediated sign-system, in a far purer medium that will afford a very high signal-to-noise ratio.

I suspect by the time we emerge – albeit born with grammar-capability and largely ‘good to go’ – we have already forged a direct, unmediated, iconic and motivated mapping between intonationally discriminated vowel and emotion. Our time spent as prelinguistic infants would further reinforce this; the phenomenon of both phonological contour and grammatical structure preceding the acquisition of lexis is well known. (YouTube has several delightful examples of infants communicating with the phrase-structures and intonational contours of speech without any lexis whatsoever – sometimes with each other, in vigorous debate.)

Just as language acquisition not only equips us with a bunch of handy signs to indicate stuff, but also provides us with a network structure along which abstract thought itself can take place (here, I’m more or less following Daniel Dennett), so the neural patterns laid down by grammar and recursive syntax might supply the carriage and combination of something other than lexis. Fill those channels, waterways and chambers with motivated vowel sounds which are perceived as partly consubstantial with the emotions they also invoke, and you have a system with the same properties of grammatical structure and recursive combination as language; it might have no denotative power, but it possesses the ability both to provoke and reflect feeling – and to generate feelings previously unfelt. Music has a generative power through its ability to develop context recursively; like syntax, these structures are infinite, and emotional analogues are automatically conjured for any combination of sounds, even those produced by aleatoric means. (Here the iconic nature of the sign forces us to make the interesting move of projecting not only our own feeling, but also the introjected feeling of a maker we know cannot exist.) This means we have access, in music, to a realm of imaginative feeling additional to that provided by our lived experience.

All this would leave us with a simple explanation: music is emotionally iconic. Music does not ‘represent human emotions’; music is exactly what human emotions sound like, and when combined with the ratiocinative power of linguistic structure, can drive our human emotions into more and more complex and nuanced patterns, quite literally allying our emotion with our intelligence. (I can find only a very limited amount of scholarly literature on the subject, but I’m sure others will also have observed that the ability to appreciate and ‘emotionally parse’ harmonically complex music often – though not always – goes hand in hand with a high degree of literacy.)11

We know that ‘poetry has a strong emotional effect’ on its readers. Much of this is simply down to its emotionally charged subject matter, which, if the process of mimesis is transparent enough, inevitably engages our own emotional memory in its powerful invocation and description of events, leaving our mirror neurons flashing like the lights on a Christmas tree; this relation between the depiction of ‘moving events’ and feeling seems straightforwardly iconic in its operation. However, it’s worthwhile being pedantically clear on this point. I use the word ‘depiction’ advisedly. While a depiction of a chair summons the chair iconically, a ‘depiction of an emotion’ has no iconic force, since an emotion can only be summoned in words via hypostatic symbol. Naming the concept ‘happy’ or ‘melancholic’ makes us feel neither of those things. In other words, telling me how you feel makes me feel nothing (which is why one encourages all apprentice poets to write with but not about their feelings); what provokes the emotion is the depiction of a situation or event, or the making of a song or argument, which incarnates, realises or activates it – and likely has also provoked it in the poet. Frost’s law of reciprocity, ‘No tears in the writer, no tears in the reader; no surprise for the writer, no surprise for the reader’, cannot be stated often enough, and is the fundamental rule of emotionally effective writing. The feelings which are provoked in the experience of the unsentimental reader are broadly those which mirror those the author felt during the process of composition itself.12 I suspect only mutant geniuses like Bach and Shakespeare were spared suffering the emotions their art provokes.

Another obvious way of conducting the patheme is the one music employs. As I’ve already discussed (and will cover from another angle in Part III), the increase in the salience, number and length of strong stressed vowels through lineation, metre, lyric selection and the preference for content over function words has an immediate effect: it simultaneously increases the line’s potential for the carriage of intonational information, and so emotional expression. Poetry cannot achieve the level of emotional iconicity music does; but it attempts to, and one aspect of the global shift from denotative to connotative function is to drag the stressed vowel centre-stage, so that the mind is primed for the reception and performance of just the kind of articulated feeling it associates with music. As we have seen, vowel is heavily exaggerated in poetry.

A crucially important aspect of this is the way rereading fixes intonational pattern, in a manner similar to the storage of prosodic pattern in the lexicalised phraseme. (This will be explored at length in the final essay. Lexis is stored along with its prosody; when a phrase become lexicalised, it undergoes the same intonational nailing-down as any other word. There are a thousand ways of stressing the phrase ‘close the kitchen window’, but barely more than one of ‘go figure’.) There is a rather astonishing psychological effect known as the ‘speech-to-song illusion’, which works as follows: a recording of a piece of conversational speech is played to a listener; then a random sentence or phrase is then isolated, and played several times over. The effect of this is to embed the intonational contour of the sentence as a memorised series of tones. When the full passage is played back to the listener, and the repeated phrase is heard again, the speaker appears to be clearly singing the phrase, in a kind of auditory version of the ‘persistence of vision’ effect.13 Repetition – of both reading and recitation – can also nail words to an intonational pattern closely resembling that of memorised song. One can see that poetry, with its built-in emphasis on song-like, strong-stressed vowels, is uniquely placed to take advantage of this. When memorisation takes place, that aspect of music which is non-sequential and atemporal (that part which, in the Rilkean conception, ‘sings across the gaps’) intrudes into the syntagm to make ‘one big word of everything’, rendering it simply song-like, with the nuanced and fixed prosody of a single word many syllables long. Song-poetry persists and exists in different parts of the brain from mere word-poetry. The auditory cortex is working a bit harder, the fun-centre of the nucleus accumbens is high on dopamine, and – if the poem is learned, say, while in love – the amygdala is meta-tagging the word-song as emotionally important, a feeling that will be recalled and revived when it is sung again. I have versions of Shakespeare’s Sonnet 18 and Yeats’s ‘Inisfree’ in my head that I can hum, and that exist separately as intonational contours, fixed patterns of semantic emphasis and emotional nuance; they feel to me as song-like as any songs I know.

This process will strongly reinforce the experience of the ‘pathemic’ through the codifying of emotional information – until, perhaps, it becomes a fixed performance; though in my own experience, intonational pattern just as often becomes a template to gently improvise within, and adapt to further performed interpretations. Add this to the mimetic power of vivid descriptive language, and we have, in effect, the definition of lyric potential – whose nature is fundamentally pathemic.

The details of a good poem are almost always bound by what I identify as the trope of ‘tonal isologue’, shared properties of ‘feel’ that overlap and create stylistic, emotional and tonal consistencies in any composition. Raising the stressed vowel-count is one direct means of permitting the carriage of emotional information in any spoken or internalised performance. The other has already been discussed at length: our examination of the phonestheme showed how shared connotations are often indicated through the presence of shared sounds. The patheme extends this idea, and states that the phonestheme – whether in the form of pre-existing sound-nodes in the language, or local innovations within the poem – conducts and activates shared qualia between words.

However, there’s more to the literary patheme than sound and vowel and consonant, and it may be transmitted by mimesis, semantic connotation, aesthetic feature, etc. – or any combination thereof. The point is that in every case the procedure will be, on closer inspection, iconic to some degree, i.e. carried via some form of direct representation. Whichever way we might analyse it, though, the fundamental characteristic of the patheme is that, like its musical equivalent: (a) it can only exist as a product of its context and not as an isolable sign; therefore (b) it can only connote; therefore (c) while it can be both indicated and conducted, it cannot be adequately paraphrased or have its complex, compound sense broken into constituents, and indeed often seems to resist linguistic description entirely. The patheme is the tonal aspect of sense. (At this point readers still sympathetic to my argument may see where this is all about to head: since the animus of the symbol is often indistinguishable from tone or feeling, the patheme is the semic means by which the symbolic is conducted.)

One might think of the denotation-free sign as a musical note shorn of its fundamental, leaving it with nothing but the higher partials of its distinct timbre – therefore unable to be woven into any harmonic or melodic structure, despite its possessing a clear timbral quality, one which other notes might share. It’s as if we could hear every one of a series of notes as being produced by a viola – but couldn’t clearly identify or name which notes they actually are. ‘Semantic timbre’ is excepted from both paradigmatic and syntagmatic (our harmonic and melodic, vertical and horizontal, if you will) analyses of texts. It can only be approached through retrospective verification, which is to say all we can do is make informal checks that two readers were indeed experiencing the same effects. The means of producing these effects is not mysterious. The means of defining them is mysterious, and not just hopelessly so: precisely so. The patheme is that very aspect of poetry that leads us still to proclaim the ‘heresy of paraphrase’; it is the quality without which the poem would be mere verse.

Meaning and feeling

I will resist exploring the matter in any great depth, as there is a book to be written on the subject – but I will merely observe that the patheme may also be a way of explaining the semiosis of the kind of symbol whose abstract animus is left unarticulated or felt to be inarticulable, i.e. those symbols which merely leave an aura of significance around a ‘totemised’ concrete detail. It seems to me that the brain processes such vague aurae as qualial effect, and may effectively ‘translate them into tone’. Suppose the patheme, our seme of feeling, is also the means by which we conduct uninterrogated meaning; inversely, this may propose that ‘interrogated feeling’ is therefore an alternative definition of ‘meaning’. Merely saying that ‘meaning is the epistemic corollary of the subservience of one domain to another’ describes the process, but does not address the mechanism by which meaning is conferred; the process will have its irreducible seme like any other. My strong hunch is that the seme of meaning is also – or at least can very usefully be designated as – the patheme, which means that the human (i.e. non-syntagmatic or non-paradigmatic) processes of meaning and feeling are, at their deepest level, consubstantial. We may have confused ‘meaning’ with the verbal articulation of ‘the feeling of meaning’.

The patheme is the seme of iconic function. But while music may be what our human feelings sound like, consider this for the last word in iconicity: ‘feelings’ are also simply what our feelings feel like. Antonio Damasio and others have provided plausible aetiologies of what we might call ‘meaningfulness’ through a description of the evolution of feeling: through the combination of learned cause-and-effect and memory (pain or pleasure centres are provoked by certain environmental conditions; these occasions are remembered; memory restimulates them), more complex emotions arise, with higher ‘feelings’ being the consequence of an awareness of those emotions. If one accepts that authentic ‘feeling’ might be a strong component of what we mean by ‘meaningfulness’, then you have something close to a proven link. This also points us back to the consciousness-creating feedback loop that I sense is also the key motif of meaning-creation. (Some current theorists in neuroscience feel that the only sensible measure of our ‘unique human consciousness’ is the extent of our awareness of our awareness; animals may be aware, but they seem not to be awake to the fact that they are so. For that same reason I very much doubt they ‘mean’ like us either.) So these feelings might be said to ‘possess’ meaning.14

However, I sense that what’s being missed is the extent to which that mere ascription is misread as inscription: the wilful designation of ‘feelings’ as meaning ful is intrinsicality by the back door. This strikes me as axiomatic: one decides on the meaning of the word ‘meaning’ by applying it to something that strikes one as self-evidently meaningful. My preference is first to say that feelings may rather be the means by which meaning is created, and then to look for an explanation with some predictive capability: how might meaning be created, from scratch, from one moment to the next? The origin of the process may well lie in the mapping of a self-reflexive, self-aware emotional procedure onto a semantic one via some exaptive expediency, resulting in the symbolic procedure we see today. But no description frees us from the obligation to understand the intermediary mechanism. For this we would have to understand the currency, the seme, the method of its conduction, the structural basis of its exchange.

A poet’s stake in all this is that it would make ‘meaning’ something capable of being consciously revised; poets sense intuitively that the dominant frames which govern received meaning can be switched, altered or subverted. If ‘meaning’ is an awareness of ‘sense’ that produces a conscious state (our use of the same word for ‘lower’ receptive faculties and basic understanding is no accident), then ‘an awareness of meaning’ may be the neutral definition of a ‘higher consciousness’; if, furthermore, the terms ‘feeling and meaning’ and ‘emotion and sense’ are linked by the fact that the former are merely a consciously aware version of the latter, we then have a definition of a ‘conscious emotional intelligence’, one that is likely the aim of all artists who seek to make the structure of both meaning and emotion manipulable.

The four tropic categories

These reflect the four ways in which trope is generated; they consist of two universal categories and two human-specific categories. These categories have many other characteristic rules, operations, expressions and aspects, but these seem to me prime: (a) operative principle; (b) seme; (c) typical semiotic form; (d) operation; (e) domainial character; (f) principal trope-type; (g) generic role; (h) function; (i) metrical-prosodic role. Each function or aspect to a degree implies all the others in the same column; therefore when one role is altered, all associated roles are at least modified.15

1 2 3 4
(a) paradigm (a) syntagm (a) iconicity (a) contingency
(b) formeme (b) aeteme (b) patheme (b) aseme
(c) isomorphism (c) index (c) icon (c) emblem
(d) correspondence (d) relation (d) representation (d) arbitrary association
(e) inter-domain (e) intra-domain (e) supra-domain (e) extra-domain
(f) metaphor (f) metonymy (f) symbol (f) asymbol
(g) content (g) function (g) quale (g) arbitrary association
(h) form (h) cause (h) meaning (h) arbitrary link
(i) strong event/strong (i) weak space/weak (i) intonation/accent (i) arbitrary stress

1 We seem to process these cultural asymbols in a way very close to reading. Danish researchers have recently shown that when the brain encounters items used in a ‘pure-symbolic’ manner – flowers left on a doorstep, say, as opposed to the same flowers growing wild – the left fusiform gyrus and the inferior frontal cortex light up, areas associated with reading and semantic meaning, respectively.

2 I have made all the Mr T jokes in advance, so spare me. ‘Ergeme’ was one possibility, but the Greek root – aitia, ‘cause’ (as in ‘aetiology’), seems both accurate and appropriately primal.

3 Roger Scruton has made one of the few meaningful contributions to the field – see The Aesthetics of Music (Oxford: Oxford University Press, 1997) – though his analysis of songs by Nirvana, R.E.M. and the like shows up a problem of methodology. For Scruton, beauty or meaning in music resides primarily in those aspects that can be adequately described by pitch, note-length and the Western conventions of expression and performance represented by the symbols and marks of traditional notation. (Where that would leave, say, that period of Aphex Twin’s career when he decided not to use notes is anyone’s guess.) However, traditional notation does not account even tolerably well for the subtleties of rhythmic displacement and emphasis, and it is especially poor at describing dynamic changes in timbre; since these are the principal means by which any sophisticated information is carried in folk, electronic, dance or experimental music – none of which tend to demonstrate very much in the way of traditional harmonic or melodic complexity – this is close to asserting the superiority of one musical genre over all others, which sounds like a form of codal interference or incompetence; although there are many like Scruton who articulately contend that the intrinsic superiority of Western classical music lies in the sophistication of its compositional method, I would still counter that this position is only really tenable if one either ignores or is simply repelled by other musics. (It seems a cultural prejudice one can immediately dismantle by merely listing those artists who move between culturally prestigious and ‘lower’ musics – say, Keith Jarrett, Chris Thile or Squarepusher.) However, we are all unconsciously drawn towards explanations which least challenge our own tastes and prejudices, and methodology can also produce a form of confirmation bias; I very much doubt this book has escaped that charge either.

4 Unless the listener is a musical synaesthete and A has an ‘asemic’ link to the colour purple, the texture of cheesecloth, feelings of sexual inadequacy, or all three – ‘A’ can carry complex information only in its articulation and its timbral composition, not in its pitch. A single pitched note has few associations beyond the metaphorical mapping ‘high’ or low’, and its corresponding place within the human vocal range, with its vague connotations of gender, age, body-size, etc.

5 We instinctively sense that music is structured by a directional metaphor. Harmony is gravitational or ‘telluric’, and works from the bottom up, not the top down: the lowest note in a chord gives it its fundamental colour. Ernst Levy in A Theory of Harmony (New York: SUNY Press, 1985), and others have posited the existence of alternative, negative harmonic structures based not on the overtone but the (largely unperceived) undertone series. If ‘meaning is harmony’ there may also be a telluric structure at play in language, with ‘denotative centre’ standing in for something like ‘root note’; a ‘negative semantics’, where connotations would override and revise fundamental primary denotation, may also be possible through some mental training – but perhaps it already exists in the form of symbolic procedure, as outlined earlier.

6 Much of what follows was initially prompted by Susanne K. Langer, whose long-neglected Philosophy in a New Key and Form and Feeling were both way ahead of their time. ‘Because the forms of human feeling are much more congruent with musical forms than with the forms of language, music can reveal the nature of feelings with a detail and truth that language cannot approach.’ Susanne K. Langer, Philosophy in a New Key: A Study in the Symbolism of Reason, Rite, and Art, 2nd edn (Cambridge, MA: Harvard University Press, 1951).

7 Iconicity is a business of degree, and the isomorphism linking the V and T of all ‘X is Y’ type metaphors is, in a straightforward sense, just its ‘iconic’ content. If the iconicity is high, then V and T can be merely stated for the connection to be obvious; if it is low, it will need to be established through further context – in other words, ‘metaphors which depend on secondary connotations for their ground must articulate them’. However, the V-sign cannot be so motivated that its uncontextualised occurrence alone will be sufficient to invoke the T-signified, unlike the Peircian asymbol and its pre-established asemic link. (For example, in our hypothetical bad poem, let’s say we just have the image of a lawn being mowed. We suspect this is a V representing ‘genocide’ by the anchorage of the title, ‘The Aftermath’, and the poet’s having qualified the grass-blades as ‘trembling multitudes’. But without those contextualising elements, the lawn is a mere detail; by contrast, there is no way the guy with the hood and sickle is merely in the poem to mow the grass.) The only aspect of language which is motivated is its mimetic function: we read descriptions of things and events as possessing sufficient iconicity to invoke those things and events themselves. This is only possible though the fundamental ‘operative conceit’ of the deep confusion of word and concept, via their quasi-asemic, semi-symbolic link. (The part-arbitrariness of this link I explored in Part I, arguing that it is strongly reinforced by an underlying non-arbitrary phonosemantic iconicity.)

8 From Schopenhauer onwards, arguments have occasionally been advanced for the idea that music might have a semantic and referential function like that of language. I cannot take any of this seriously, and find no evidence whatsoever for such a position within my own direct experience. Supporters of the notion seem to argue only for its theoretical possibility. Other than its programmatic and onomatopoeic effects, music simply lacks a system of isolable signs and concomitant valencies to map itself to the realm of discrete concept. Besides, the most basic model of music as a denotatory sign-system would also have to assume things like universal perfect pitch, as well as a period of innocent language acquisition, where the function of the sign is primarily indexical until the links are forged. (See my earlier, admittedly contentious statements on presence and absence.) This would allow initial arbitrary connections to be made – between, say, the second inversion of the B♭ triad and that armchair, or the tierce de picardie in the key of D and an overweening self-confidence. However, this seems effectively no different from forging a sign-system based on, say, various patterns of fluff, paperclips or small change.

9 While I appreciate all arguments from experience should probably be disregarded, I should mention that I feel I instinctively know this from my own work as a jazz musician (one reason I have hitherto – and erroneously – regarded this as a commonplace). In the improvised ballad solo, especially, the rate of note-delivery is often very close to that of conversational prosody, and my experience of improvising in these circumstances is entirely ‘rhetorical’. To play a solo on a long-note, human-voice-range, monophonic instrument over, say, the melancholic, plaintive harmonic argument of Bill Evans’s ‘Time Remembered’ feels more or less identical to the sensation of making a speech whose intention is to convince a listener of a truth; I feel my aim is to argue my point, articulate my emotional position, with as much nuance, conviction, reasoned debate and perhaps ‘manipulative’ tenderness as I can summon.

10 Julia Kristeva, ‘Motherhood According to Giovanni Bellini’, Desire in Language: A Semiotic Approach to Literature and Art, ed. Leon S. Roudiez, trans. Thomas Gora and Alice A. Jardine (New York: Columbia University Press, 1980).

11 A further and necessary complication is the mapping between complex and nuanced meaning and musical harmony, which I suspect we perceive as directly analogous ‘vertical’ sophistications – one more example, if you like, of the projection of the axis of selection into the axis of combination, and a historically parallel development to poetry’s use as a literary and not merely lyric form.

12 Sentimental readers move themselves. Nonetheless, there’s a balance to strike here. Young poets are often encouraged to ‘take risks’, but by ‘risk’ what’s often intended is ‘experiment’. However, ‘experiment’ is mostly a pretty safe activity: the worst that can happen is the reader will get bored, confused or mildly offended. The most fruitful risks will often involve writing very simply, just before it shades into simple-mindedness; with prophetic force, just before it shades into pretentiousness; and at the extremes of feeling, just before it shades into sentimentality. I tend to be of the opinion that when you write, you should always risk looking like an idiot, a pretentious fool or a sap – or, indeed, a failure. Readers read poetry to take them closer to something they’d not usually be prepared let themselves feel: they ‘gear up for it’, and say – right: now I will read a poem.

In Japanese restaurants, great respect is accorded to those chefs who can prepare the puffer fish called the fugu. To qualify, you have to train for years; most candidates never even get to sit the exam – and even there, 90 per cent fail. The reason that it takes so much training is not so the chef can avoid cutting into the fish’s poison sac; that’s easily learned. And certainly, if you cut it too close, you will kill the diner: but the reason the diner has ordered the fish is for the ‘high’ of the toxin. If the chef can cut it just close enough, a little of the toxin will seep into the flesh of the fugu, and give the diner the strange sensation they’ve paid so heavily to experience. So it is with the poetic art. One slip of the knife, and ‘the reader will die on you’, which is to say they will find you too stupid, or sentimental, or ridiculous to trouble themselves with, and close the book. But cut just close enough, and they’ll be rewarded for their investment. So we should, by all means, risk: but it’s our reputations that should be risked. And as for the reader – their mental health should be at risk, their sleep should be at risk, their unchallenged assumptions about the world should be at risk, just as the poet’s will have been. But if the poet isn’t walking the highest tightrope they dare … mediocrity is guaranteed.

13 The effect reinforces the argument that a deep structural connection exists between language and music:

To conclude, this illusion is in line with what philosophers and musicians have been arguing for centuries, that strong linkages must exist between speech and music. We still need to determine the neural processes that are responsible for this striking perceptual transformation. However, the present experiments show that for a phrase to be heard as spoken or as sung, it does not need to have a set of physical properties that are unique to speech, or a different set of physical properties that are unique to song. Rather, we must conclude that, assuming the neural circuitries underlying speech and song are at some point distinct and separate, they can accept the same input, but process the information in different ways so as to produce different outputs. As a further point, this illusion demonstrates a striking example of very rapid and highly specific perceptual reorganisation, so showing an extreme form of short-term neural plasticity in the auditory system.

– Diana Deutsch, Rachael Lapidis and Trevor
Henthorn, ‘The Speech to Song Illusion’ (invited
lay language paper presented at the 156th meeting
of the Acoustical Society of America, Miami, FL) Journal
of the Acoustical Society of America, 2008

14 Antonio Damasio, Looking for Spinoza: Joy, Sorrow, and the Feeling Brain (London: Vintage, 2004).

15 This table is ripe for crankish expansion: one could cheerfully start listing everything from the fundamental interactions (these fit rather well, alarmingly) to the Four Horsemen of the Apocalypse (ditto), but I have restrained myself. Nonetheless I am convinced that some kind of unified poetic field lurks behind it.