Onomatopoeia_as_a_Figure_and_a_Linguistic_Principle

http://muse.jhu.edu/journals/new_literary_history/v027/27.3bredin.html H Bredin - New Literary History, 1996 - muse.jhu.edu Onomatopoeia as a Figure and a Linguistic Principle Hugh Bredin Figures [pic] It is easy to think of onomatopoeic words. Whizz, bang, splash, thump, will strike most English-speakers as typical examples; and once we are familiar with these, it is easy for us then to recognize others almost at will, and even to invent new ones if need be. An audience at the film How to Murder Your Wife needs no explanation why a cement-mixer is referred to in a cartoon strip as a gloppita-gloppita machine. The knowledge of how to speak a language seems to naturally involve a knowledge of whatever principle it is that underlies onomatopoeic idioms, coinings, and usages. 1 There is less unanimity, however, and more difficulty, when attempts are made to define onomatopoeia. A quick trawl through a number of standard reference books shows that, while everyone agrees that onomatopoeia is the name of a relationship between the sound of a word and something else, there are divergent views both on the second term of the relationship and on the nature of the relation itself. The second term of the relation is variously referred to as sounds, sense, referent, and what is denoted. The relation that obtains between the two terms generates an extensive and heterogeneous collection of names: imitates, echoes, reflects, resembles, corresponds to, sounds like, expresses, reinforces, and has a natural or direct relation with. 2 It looks suspiciously as if there is some confusion, or vagueness at least, about the concept of onomatopoeia. Even the nature of the confusion or vagueness is not clear. Some of the authors suggest that there is more than one type of onomatopoeia, since they distinguish between a strict or narrow sense, and a more general or broad sense, of the term. Others list more than one definition or sense of onomatopoeia without further explanation, as if the senses were roughly equivalent or not sufficiently different to warrant discussion. Others again provide a single definition and are content to leave it at that. The strict or narrow kind of onomatopoeia is alleged to occur whenever the sound of a word resembles (or "imitates") a sound that the word refers to. The words "strict" and "narrow" suggest that the sense in question is a kind of original usage or practice, in respect of which other [End Page 555] senses of onomatopoeia are metaphorical or perhaps extensional enlargements. However, if we go back in time to Quintilian's Institutio Oratoria, the work which laid the foundations for all subsequent descriptions and theories of figurative language, we find that onomatopoeia refers here to what its etymology implies: namely, the creation of a word ex novo. (Quintilian remarked in passing that the Greeks regarded word creation as a virtue, whereas among the Romans it was rarely acceptable: a fascinating glimpse into the contrast between the two great cultures of classical Europe.) Quintilian, it is true, gave as examples of created words what we would now call "onomatopoeic" words: mugitus for the lowing of cattle, sibilus for a hiss, and murmur. But his explanation for these coinages was rooted in a theory of language first expounded in Plato's Cratylus --that, at some earlier and foundational stage of history, language was invented by people (qui sermonem primi fecerunt) and that their inventive activities were motivated by a kind of fitness between the invented words and whatever they were names for (aptantes adfectibus vocem). The word used here by Quintilian for the objects named is adfectibus, which does not refer specifically to sounds but rather to any state or disposition of mind or body. Thus, while his examples may sustain something like the current notion of onomatopoeia, neither his definition nor his explanation of it do so. 3 It would be a substantial scholarly task to trace the origin and vicissitudes of the theory of onomatopoeia. Quintilian's work appeared in the late first century. Bede's early eighth-century De Schematibus et Tropis defines it in the so-called strict sense mentioned a moment ago. 4 Geoffrey of Vinsauf's thirteenth-century Documentum de Modo et Arte Dictandi et Versificandi reserves the name for a figure in which an onomatopoeic word is used metaphorically. 5 The earliest great English treatise on rhetoric, Puttenham's The Arte of English Poesie (1585), combines the idea that onomatopoeia means coining a new word with the further specification that the coinage must be a word whose sound is similar to the object that it names. 6 A recent compendium of sixteenth-century rhetoricians shows that the figure was defined then in several different ways. 7 The fact of the matter is that the notion of onomatopoeia has, throughout its history, generated just as much diversity as it does today. In view of these confusions, vaguenesses, and disagreements, both historical and actual, I now propose to undertake two tasks: firstly, to define and distinguish among different types of onomatopoeia; and secondly, to consider some general issues about language which are connected with the phenomenon (which may well be a universal one) of onomatopoeia. [End Page 556] Conceptual tidying always has to begin with the evidence of experience, which in this case consists of ordinary usage and items of general agreement. There is general agreement that it is correct to describe as onomatopoeic a substantial core list of English words. The reference books mentioned above provide us with the following list: whack, fizz, crackle, hiss, clip, clop, buzz, spatter, dingdong, moo, pop, whoosh, zoom, rattle, bang, moan, murmur, scratch, crash, cuckoo, sizzle, clatter, thud, clash, shriek, roar, swish, susurrus, whirr. I will therefore accept, without further argument at this point, that most or all of these words are, in some sense or other, onomatopoeic. There is also general agreement that, in the standard use of the word "onomatopoeia," it refers to a relation between the sound of a word and something else. Linguistic sound is, of course, a very complex sort of thing. The smallest linguistic sound segment is conventionally called a phoneme, but even a single phoneme is itself a complex construction of such features as voice, nasality, plosiveness, height, length, and many others. When phonemes are combined to form larger units, such as words and phrases, these units then acquire what are called prosodic features--pitch, loudness, and duration, for instance--all of which in turn have something to do with how language sounds. Thus, to fully describe the sound of even a single word, even a monosyllabic one, would be a lengthy task; and if we wanted to describe all the possible variations on the sound--the possible sound tokens as well as the sound type--it would become a huge task of indefinite length. It is best therefore to ignore these complexities and to consider the sound of words and phrases simply as the nexus of acoustic properties which constitutes them as objects of consciousness for a normal speaker of the language. To be precise, we should think of the relevant sound as an object of consciousness, not in or by itself, but rather as sound onomatopoeically related to something else. In onomatopoeia it is the sound as related to something that constitutes its essential nature. It is its being related to something that causes it to become an object of consciousness in the first place. To adapt the language of Genette, verbal sound in ordinary usage is transparent: that is, we can grasp the meaning of a word or phrase without consciously adverting to, or subsequently remembering, its sound. In onomatopoeia, the sound of words and phrases becomes opaque: our consciousness of the sound of a word, and of its meaning, are inextricably intertwined. Verbal sound, then, is the first of the two terms in an onomatopoeic relation. What is the second term' We can begin by describing the second term as a meaning, and then follow Frege in differentiating between meaning as reference and meaning as sense. The thing that a word refers to on a particular occasion of use is called its referent; the [End Page 557] thought that a word expresses on every occasion of use is called its sense. We can then follow Mill in further differentiating sense into denotation and connotation. The denotation of a word is the class of all its conventional referents; the connotation of a word is the concept instantiated by all the members of the class. Thus, if I point at a chair and say "This is a chair," the referent of the word "chair" is the chair at which I am pointing, its denotation is the class of chairs, and its connotation is (to quote Webster) the concept also expressed by the words "a seat typically having four legs and a back for one person." We now turn to the onomatopoeic relation between the terms as they have now been defined: that is, between verbal sound on the one hand, and, on the other, meaning differentiated as required into referent, denotation, and connotation. The evidence of usage, and the need for conceptual clarity, enables and requires us to distinguish three types of such relation, and thus three types of onomatopoeia. The first and most obvious type, which I shall call direct onomatopoeia, occurs whenever two criteria are satisfied: (1) the denotation of a word is a class of sounds; and (2) the sound of the word resembles a member of the class. To put it less technically, the sound of the word resembles the sound that it names. Some typical examples are hiss, moan, cluck, whirr, and buzz. It is immediately clear, of course, that none of these words is very like the sound that it denotes, and this point is reinforced when we consider other examples such as splash, rustle, zoom, bang, shriek, and thud. These latter have some acoustical similarities to their denoted objects, but not, if we are honest, very many. There are higher and lower degrees of onomatopoeic resemblance, and the number of words, such as hiss, which have quite a high degree of resemblance, is relatively small. At the very lowest threshold of direct onomatopoeic resemblance we find names for animal sounds. It seems to be a common human instinct to devise onomatopoeic names for these, but more than one person has commented on how differently the instinct works in different languages. Ernst Gombrich writes, "To me, at least, the cock says not 'cock-a-doodle-doo' as he calls to the English in the morning, not 'cocorico' as he says in French, nor 'kiao kiao' as in Chinese, but still 'kikeriki,' as he says in German." 8 The same minimalism is to be found in other kinds of sound names as well. Whisper and bisbiglio (whisper) sound onomatopoeic to English and Italian speakers respectively, but the only resemblance between either word and the members of its denoted class rests on the s sound. The same point can be made about other synonymous pairs of words such as scream and strillo, rustle and frusciare, buzz and ronzare. Even hiss, which to English ears sounds onomatopoeic to a high degree, translates into Italian fischio, which not only seems just as good, just as "onomatopoeic," but also, disconcertingly, means "whistle" as well. [End Page 558] The moral to be drawn here is that onomatopoeic words of this first type are heavily determined by convention, and not just by a "natural" resemblance between sounds. We are inclined to overestimate the natural resemblance and to underestimate or forget the convention, because we are so used to our own language and do not observe it from the outside, as it were. The conventionality of onomatopoeia is due not only to the fact that all language is conventional, but also to the phonemic restrictions imposed by languages upon their speakers. Seventy percent of languages have between 20 and 37 phonemic segments, and the largest recorded number in a single language is 141. Thus there are serious limits, imposed by the availability of suitable phonemes, on the capacity of a language to reproduce nonverbal sounds like hisses and whizzes and bangs. Further restrictions are imposed by the anatomical structure of the vocal organs in human beings. Some people can become expert at imitating animal sounds--birdsong, wolf calls--but this is mimicry, not language. Even the most onomatopoeic of words are experienced as onomatopoeic as much because of convention as because of their experienced acoustical resemblance to the sounds of which they are names. These considerations generate questions about the nature of acoustic resemblance, and, in particular, how it is that we become aware of the resemblance between an onomatopoeic word and what it names, how it is that the word's sound loses its transparency and becomes opaque. Whisper is "like" the sound it denotes because the s is "like" the sibilant sounds I hear from time to time if I listen to people whispering at the other side of a room. The s in sister is also like the sounds that I hear from time to time if I listen to my sister speaking. Why is it, then, that whisper is onomatopoeic and sister is not' The answer resides in the fact that whisper denotes a kind of sound and sister does not. That is, it is because a word first of all denotes a sound that it then becomes a candidate for being experienced as onomatopoeic. This is what is meant by the claim that convention is just as important as word sound in determining onomatopoeia. Convention means conventional denotation. If a word conventionally denotes a sound, we are then predisposed to become aware of any acoustical properties of the word which resemble the sound, however minimal that resemblance may be. In fact any word the members of whose denotation have an acoustical property can acquire in the linguistic consciousness an onomatopoeic resonance: conversation, chatter, hurly-burly, tempestuous, slither, loud, soft, applause, grumble, noise. None of these is remotely onomatopoeic, yet it is impossible to use them or hear them without being aware of a kind of onomatopoeic aura. It is as if our consciousness of words includes an instinctive desire to fit sounds with things, to experience [End Page 559] some sort of phonetic appropriateness in human speech. Even the most tenuous of connections between the acoustic structure of a sentence and the fact that it articulates is often transformed in our experience of it into a sense of fitness and rightness. It is arguable that onomatopoeia is not a trivial and incidental phenomenon of usage, but answers to a deep-seated need that lies at the heart of the linguistic consciousness. We want language to be onomatopoeic. A second type of onomatopoeia occurs whenever the sound of a word resembles a sound associated with whatever it is that the word denotes. Some examples of this are: cuckoo, bubble, smash, whip. None of these words has a sound that resembles the objects or actions that they denote. Cuckoo is the bird's name, but its acoustic resemblance is to the song that it produces, not the bird itself. Whip is like the sound made by a whip. Smash has a slight resemblance to a sound that may accompany the act of breaking or destroying. Bubble resembles neither the object nor the action, but has some similarity to the sound of a bubbling liquid. A famous historical example is barbarian, whose root, the Greek word barbaroi, was devised as a name for non-Greeks because their strange languages sounded to Greek ears like the stuttered syllables "ba-ba." If we call the first type direct onomatopoeia, this second type can be called associative onomatopoeia. Association is just as much a matter of degree as is acoustic resemblance. There is a close association of sound and object in the case of cuckoo, but a very slight association in the case of scratch or spatter. Association is also a matter of convention: the association of one thing with another thing may in some cases be predictable, given the relevant circumstances, but no association of objects is ever necessary or inevitable. The call of the cuckoo is so distinctive that it predictably motivates the name given to the bird in many languages, though even here the Swedish gök may sound less onomatopoeic to our ears than, say, the Latin cuculus or even the Irish cuach. Bubble is undoubtedly onomatopoeic, but the equivalent Italian noun, bolla, comes from the root bollire (to boil). That is, the English-speaking community's association of bubbles with the sound of boiling liquid was not the association made in the Italian-speaking community. Italians associated the word with the phenomenon, but not the sound, of boiling. Whip, interestingly enough, derives from Old English wippe, which means a quick movement or leap. It seems arguable in this case that the onomatopoeic character of whip is acquired and, to some extent, accidental: that the association of whip and sound came into existence after the formation of the noun, which originally had a nononomatopoeic motivation. Against this, however, it might be asked why this particular [End Page 560] Old English root was adopted by English-speakers. The contrast with Italian is again instructive. Frusta, a whip, comes from Latin frustare, which means to break. A similar root might have been chosen in English, such as brecan, to break; or even an alternative root such as streng, cord or string. It is at least possible, therefore, that wippe was adopted just because of its onomatopoeic possibilities. Associative onomatopoeia thus involves two levels of conventionality. There is, firstly, a conventional association between something and a sound (a bird and its song, a bubble and the sound of boiling liquid). Then, secondly, there is a conventional relationship of naming between a word and the thing named by it (cuckoo and the bird, bubble and a flimsy sphere of liquid). This is illustrated in figure 2. Figure 1 represents direct onomatopoeia. It will be useful to contrast associative onomatopoeia with another kind of acoustic association which is not onomatopoeic, though it might seem to be so at first sight. Consider the following sentence: 1. The coach's iron voice carried across the playing field. The point may be clearer still if we compare this with 2. The coach's hard voice carried across the playing field. In both of these sentences, an English-speaker will typically experience the words iron and hard as having what I called earlier an onomatopoeic aura. If we are put to the question, we might say that a hard voice sounds flat and steady, and that an iron voice has a kind of resonant clarity that the hard voice may lack. It is extremely difficult, in fact, not to interpret these adjectives as conveying something about how the coach's voice sounds, and this is why we might be tempted to ascribe to them a partly or slightly onomatopoeic character. If we consider these sentences more closely, however, it is immediately evident that the sound of iron, like the sound of hard, has no intrinsic likeness to the imagined sound of the coach's voice. If this is doubted, try replacing iron by ironic and note how the supposed resemblance fades away. Iron is in fact quite a mild sounding word, no matter which of its dialectical variants (eye-irn, ion, eye-rn) we try out. An iron voice is authoritative and vaguely alarming, but the sound of iron is not. Why, then, do we feel that iron tells us something of what the coach's voice sounds like' The answer is that the coach's voice has an acoustical property, described above as a kind of resonant clarity, similar to that of iron being struck. But this similarity is an imagined similarity, not one that is available to the sense-perception of the reader or listener by listening to the sound of the word iron. The actual sound (iron) features not at all. The relevant relation obtains between the coach's voice and a [End Page 561] sound associated with the substance iron, not the word "iron." This is why it is not an onomatopoeic phrase. The appropriate diagram this time is figure 3. In a word, the phrase "iron voice" describes the coach's voice. It does not enact the voice, as it would if it were onomatopoeic. Onomatopoeia is a relation between signifier and signified in which the signifier is motivated, in part, by its sound. If there is no acoustic property in a signifier which motivates us to prefer it to some other signifier, it is not onomatopoeic, even though it may seem to us to have acquired some sort of acoustic rightness or inevitability in virtue of its reference, context, and descriptive vividness. [End Page 562] A third, and final, type of onomatopoeia I shall call exemplary onomatopoeia. Its foundation rests upon the amount and character of the physical work used by a speaker in uttering a word. Words such as nimble and dart require less muscular and pulmonary effort than do sluggish and slothful. Also, their stopped consonants encourage a speaker to say them sharply and quickly, whereas the latter two words can be drawn out slowly and lazily. The relevant properties of these words--the properties which constitute their acoustic character--have technical names such as voicing, stopping, plosiveness, stress, length, juncture; but even if we do not know their names the properties are responsible for familiar aspects of our everyday linguistic experiences. They are exploited with particular zest and expertise by good stylists, and it is worth repeating here a famous passage from Pope, something of a chestnut by now but still an object of admiration and a source of pleasure. 'Tis not enough no harshness gives offence, The sound must seem an Echo to the sense: Soft is the strain when Zephyr gently blows, And the smooth stream in smoother numbers flows; But when loud surges lash the sounding shore, The hoarse, rough verse should like the torrent roar: When Ajax strives some rock's vast weight to throw, The line too labours, and the words move slow; Not so, when swift Camilla scours the plain, Flies o'er th'unbending corn, and skims along the main. (Essay on Criticism, II, 364-73) This is a virtuoso demonstration of exemplary onomatopoeia, which differs from the first two types of onomatopoeia in two ways. Firstly, the [End Page 563] second term of the onomatopoeic relation is to be found in a word's connotation, not its denotation; secondly, the relation between the two terms is a relation of instantiation, not of resemblance. The word sound nimble does not sound like anything that can be denoted by the word, and it cannot resemble the idea connoted by it, since sounds and concepts cannot "sound alike"; concepts have no sound. Instead, the word sound instantiates or exemplifies nimbleness, since it is itself a nimble sound. Dart, similarly, has a quick, darting sound. Pope's lines "When Ajax strives some rock's vast weight to throw, / The line too labours, and the words move slow;" have a sound which encourages us to utter them slowly and with a certain self-conscious effort, and these are qualities (slowness, effort) which are connoted by the lines themselves. The properties that word sounds have in themselves, and the properties of word utterances that are involved in exemplary onomatopoeia, are not the same. The objectively describable properties of strives --monosyllabic, diphthongal, commencing with a voiceless and ending with a voiced alveolar sibilant, and so on--are one thing. The properties that strives has when it is uttered in an effortful manner--being slow, difficult, drawn out--are another. The former are properties of the type; the latter are properties of one of its tokens. The former do not necessarily cause the latter. They do, however, enable the latter. Given certain combinations of consonants and vowels in a word type, certain possibilities of utterance necessarily follow, though they may not necessarily be realized. It is perfectly possible that "When Ajax strives some rock's vast weight to throw" should be spoken rapidly and lightly. The meaning of the line, however, inevitably invites a reader to speak or imagine it slowly and effortfully. Here again we see how meaning impels us to seek out onomatopoeia where we can. Exemplary onomatopoeia can be extremely subtle and fluid, since it exploits the almost indefinite number of associations that can be made between sounds and things. It is instructive to consider how the s sound is used in three lines from Pope: (3) Soft is the strain when Zephyr gently blows (4) But when loud surges lash the sounding shore (5) The line too labours and the words move slow In each of these lines, s instantiates and reinforces three quite different qualities. In 3, it exemplifies softness and gentleness, qualities that quite clearly constitute a part of what the line connotes. 4 is much more problematic. The line connotes noise, and at first sight it might seem that s is a most unsuitable sound to convey an impression of noise. But there are different kinds of noises. We can contrast the sound of waves [End Page 564] crashing on a beach with the trickle of a mountain stream or a fountain, and contrast it again with the crashing and grinding of rocks, the piping of a bird, the slamming of a door, the roar of traffic, and so on and so forth. In fact the combination of s with open vowels very effectively captures the precise character of the continuous watery noise that the line compels us to imagine. 5 exploits yet another possibility inherent in the s sound: because it is a sibilant, it can be lengthened indefinitely and used, as it is here, to slow down the pace of the utterance; to exemplify, again, a property that is connoted by the sentence to which it belongs. This threefold contrast in the use of what is ostensibly the "same" sound illustrates sharply one of the main features of exemplary onomatopoeia. Verbal sounds in themselves have no particular semantic character, but they are able to instantiate an indefinite number of qualities. Once the utterance has been formulated, its connotations evoke and pin down the relevant onomatopoeic quality of the sounds used to articulate it. A diagrammatic representation of this kind of onomatopoeia is figure 4. Onomatopoeia has rarely, if ever, received the attention that it deserves. Rhetoricians are more impressed by figures that affect syntax and meaning than by figures affecting sound. Logicians and philosophers care mainly about propositions, since these are the bearers of truth value and the constituents of arguments. Saussure's principle, that the relation of sound to meaning is arbitrary, holds virtually universal sway. The main function of a verbal sound, it would seem, is that, because it is discriminable from other sounds, it can be used to give a sensuous and communicable existence to a nonsensuous and otherwise amorphous meaning. Distinguishable sounds as it were produce distinguishable meanings, out of a kind of conceptual continuum, by virtue of their very distinctness. This is why it is said that sounds are used to "articulate" meanings. Real experience of language, however, suggests that there is more involved in its sounds than the fact that they are instruments used for the purposes of articulation. No doubt word sounds are frequently "transparent," at least on the conscious level. But once we enlarge the focus of our attention to include the sounds of discourse--whether we do so as an act of will, or because we meet someone with a strange accent, or because the nature of the text forces its sound upon us (in literature, say, or in a good stylist, or a good speaker)--it is very difficult to experience the sound as arbitrary. Instead, we find ourselves passing judgments on the sound: that it sounds right or not quite right, or inappropriate or smooth or awkward. At the beginning of this paragraph, I initially wrote the phrase "articulatory instrumentality," and was instantly driven to find another way of saying the same thing that [End Page 565] sounded better. Who has not experienced a frisson of horror at learning that Arthur Conan Doyle considered naming his detective Sheridan Holmes' We know that in Keats's lines "Charm'd magic casements, opening on the foam / Of perilous seas, in faery lands forlorn," he initially wrote "ruthless seas" before changing to "perilous." "Ruthless" must have seemed a suitable word at first, since it echoed "the sad heart of Ruth" a few lines earlier. "Dangerous seas" would seem to be another, and this time a three-syllable, alternative. But "perilous" was the right word, and Keats found it in the end. Joyce tells us that Leopold Bloom enjoyed "liver slices fried with crustcrumbs." An ordinary writer would have had them fried "in breadcrumbs," but Joyce, like Keats, found the right words. In fact sound does matter in language, and one constituent of our consciousness as language users is an awareness of the fit between sound and meaning. Onomatopoeia is one species of such a fit. It is the kind of fit that obtains whenever the sound of a word enacts some aspect of its meaning. No doubt onomatopoeic words are also descriptions (or names used as descriptions), but they describe things in a mimetic or enactive mode. For this reason, they are a particularly concentrated and effective product of a deep-seated need to coordinate words with things. I noted earlier how meaning seems to drive us to onomatopoeia: even the smallest pretext, such as a slight resemblance, an association, a peripheral property, is enough to spark off in our linguistic experience an awareness of phonetic mimesis, no matter how slight. To repeat an earlier remark: it is as if we want language to be onomatopoeic. It is inevitable that an analysis of onomatopoeia should make some reference to sound symbolism. This topic was given its classic formulation by Otto Jespersen, who quoted Humboldt's view that "language chooses to designate objects by sounds which partly in themselves, partly [End Page 566] in comparison with others, produce on the ear an impression resembling the effect of the object on the mind." 9 Jespersen did not advert to the oddity of this claim, which suggests that there can be a resemblance between an aural sensation and, say, a concept or an emotion which is present to consciousness. Jespersen's examples of sound symbolism, however, have become standard, and the one which has provoked most debate is the apparent correlation between, on the one hand, high tones (such as i) and brightness, cheerfulness, small size, weakness, daintiness; and on the other hand, low tones (such as u) and darkness, gloominess, large size, strength, coarseness. Examples of the first of these are light, thin, tiny, petite, and examples of the second are gloom, grumble, hunk, muck. It is easy, of course, to find counterexamples, and Jespersen cited several counterexamples himself. The most that he was prepared to claim was that "there is something like sound symbolism in some words" (397). More recently, a very exhaustive survey of sound symbolism has been included in Jakobson and Waugh's The Sound Shape of Language. It is clear that these authors believe that there is such a thing as sound symbolism, which they define as "an inmost, natural similarity association between sound and meaning," 10 and they cite an impressive quantity of experimental data which seem to support their belief. They make three main points about sound symbolism which are worth repeating here. The first point, which they take from the French phonetician Maurice Grammont, is that the semantic significance of sounds is a potentiality, which becomes manifest only when it is awakened by the meaning of the text. This is exactly parallel to a point made above about onomatopoeia: that it is because a word first of all has a certain meaning that it then becomes onomatopoeic. Verbal sounds are onomatopoeic, not in themselves, but only as related to a relevant aspect of meaning. The second point is that sound symbolism is a relation that holds, not between sounds and meanings, but between sound-relations and meaning-relations. It is not that i evokes bright and u evokes dark, but that the relation of i to u correlates with the relation of bright to dark (or small to large, dainty to coarse, light to heavy, up to down, quick to slow, sharp to blunt, and so on). Their third point is that the relation of sound to meaning has its roots in the phenomenon of synaesthesia. The theory of synaesthesia supposes that there is some kind of systematic interconnection between aural and visual perceptions, and perhaps among sensations of all kinds. Jakobson and Waugh cite the famous instance, in Locke's Essay, of the blind man who said that the color scarlet was like the sound of a trumpet. Synaesthesia, if there is such a thing, would provide not just an [End Page 567] explanation of sound symbolism but also a reason for its existence: sound symbolism would no longer be just a set of correlations based upon the empirical observation of a number of languages, but rather a natural and inevitable product of the human psyche. If synaesthesia is a fact, sound symbolism would seem to be its necessary consequence. I am unable to express any opinion either about synaesthesia or about the universality or otherwise of sound symbolism in human language. But I am convinced that onomatopoeia and sound symbolism must be kept distinct. If we take the paradigm case of sound symbolism to be a correlation between i and u on the one hand, and, on the other, relations such as bright-dark, light-heavy, quick-slow, and the like, then sound symbolism is in essence a relation between ranges of sound and fields of meaning. Onomatopoeia, in contrast, involves a relation between a specific verbal sound and a sound that is in some way an aspect of meaning: a denoted sound, a sound associated with the denotation, or a sound property constituting part of the connotation, of the onomatopoeic word. Onomatopoeia is a relation of sound to sound. Sound symbolism is a relation of sound to semantic fields such as brightness and darkness. The onomatopoeic relation is a relation of similarity or instantiation; the relation in sound symbolism is a relation of association (grounded, according to Jakobson and Waugh, in synaesthesia). It is possible that both onomatopoeia and sound symbolism could be included eventually in a single unified theory, a theory capable of specifying general rules governing the relation between meaning and what Jakobson and Waugh happily call the sound shape of language. We would expect the theory to account for those occasions on which language appears to be mimetic as well as designative. Even within such a theory, however, onomatopoeia would still have a separate place, since it is a distinctive relation and would have its own distinctive rules. Much more work has to be done before we can say with any certainty whether sound symbolism is a universal cross-linguistic phenomenon. Onomatopoeia is a different matter. Every human language is capable of denoting sounds and of denoting and connoting sound properties, and onomatopoeia is therefore a universal possibility in all languages. I will therefore finish dangerously by formulating the following hypothesis: whenever some aspect of a word's meaning can be identified as a sound or a sound property, there is a universal human disposition to seek out some aspect of the word's sound that resembles or instantiates it. If we should find such a resemblance or instantiation, we experience the word as onomatopoeic: that is, we experience the word as a sign which enacts its meaning as well as articulating it. The hypothesis comes complete with rules: onomatopoeia is a possibility whenever a word (1) denotes a sound, (2) denotes something associated with a sound, or [End Page 568] (3) connotes at least one property such that a verbal sound can also have the property. I want to claim, in a word, that onomatopoeia is, in some sense or other, a linguistic universal. Queen's University, Belfast Hugh Bredin is Senior Lecturer in Scholastic Philosophy at the Queen's University of Belfast, Northern Ireland. He has translated works by Umberto Eco, and has published several articles on literary theory and Þgurative language. Notes 1. While I was preparing this paper, I asked my nine-year-old son (whohas never seen the film) whether he could guess what kind of machine a gloppita-gloppita machine was. After thinking for a while, he said, "A cement-mixer'" 2. The standard reference books are M. H. Abrams,A Glossary of Literary Terms, 5th ed. (New York, 1988); Chris Baldick, The Concise Oxford Dictionary of Literary Terms (Oxford, 1990); Karl Beckson and Arthur Ganz, Literary Terms: A Dictionary, 3rd ed. (London, 1990); J. A. Cuddon, A Dictionary of Literary Terms (Harmondsworth, 1982); J. R. Harmsworth, Dictionary of Poetical Terms (Toronto, 1972); The New Princeton Encyclopedia of Poetry and Poetics, ed. Alex Preminger and T. V. F. Brogan (Princeton, 1993); Katie Wales, A Dictionary of Stylistics (London, 1989). 3. Quintilian discusses onomatopoeia in hisInstitutio Oratoria, tr. H. E. Butler (London, 1920-22), 8, 6, 31-32. 4. Bede, the Venerable,De Schematibus et Tropis, in English as Concerning Figures and Tropes, tr. Gussie Hecht Tannenhaus, in Readings in Medieval Rhetoric, ed. Joseph M. Miller, Michael H. Prosser, and Thomas W. Benson (Bloomington, 1973), pp. 111-12. 5. Geoffrey of Vinsauf,Documentum de Modo et Arte Dictandi et Versificandi, tr. Roger P. Parr (Milwaukee, 1968), p. 61. 6. George Puttenham,The Arte of English Poesie, ed. Gladys Doidge Willcock and Alice Walker (Cambridge, 1936), p. 182. 7. Lee A Sonnino,A Handbook to Sixteenth-Century Rhetoric (New York, 1968), pp. 132-33. 8. Ernst Gombrich,Art and Illusion, 3rd ed. (London, 1968), pp. 306-7. 9. Otto Jespersen,Language (London, 1922), p. 396; hereafter cited in text. 10. Roman Jakobson and Linda Waugh,The Sound Shape of Language (Brighton, 1979), p. 178.