ÖйúAƬ

One of the first bits of wisdom imparted to students in introductory linguistics classes is that there is no intrinsic connection between a word's pronunciation and its meaning: Japanese "bara" is just as good a name as English "rose". In the terminology of Ferdinand de Saussure's enormously influential Cours de Linguistique Generale, the link between the signifier and the signified is arbitrary, and the eponymous protagonist of Plato's Cratylus serves as a favourite foil for suggesting that the association between form and meaning is "natural" rather than merely "conventional". But the dictum of arbitrariness comes with an equivocation: onomatopoeic words are obvious exceptions. Onomatopoeia is the most familiar subcategory of sound symbolism, which can be defined broadly as any nonarbitrary connection between a vocalisation and its meaning. The standard pedagogical ploy is to assure students that the sound-symbolic vocabulary in any language is a negligible fraction of the lexicon, but the suspicion lingers that sound symbolism should not be dismissed quite so perfunctorily.

Sound Symbolism is a collection of papers from a conference held at the University of California at Berkeley, and the papers are divided into seven groups: (1) native American languages north of Mexico, (2) native languages of Latin America, (3) Asia, (4) Australia and Africa, (5) Europe, (6) English, and (7) the biological bases of sound symbolism. The introduction by the editors provides a carefully thought-out typology of sound-symbolic phenomena as a framework into which the papers that follow can be integrated. The introduction could stand on its own as a scholarly essay were it not for the requisite references to the papers. There are also numerous references in the papers to the introduction and to each other, and these give the book a cohesion that is unusual in conference proceedings.

The editors' sound-symbolism typology has four basic categories along a continuum ranging from least arbitrary to most arbitrary. The first category is "corporeal sound symbolism": the use of sounds or intonation patterns to express internal emotional or physical states. At one extreme, this category includes involuntary, "symptomatic" sounds such as coughing, and it ranges through what is known as "paralanguage": the expressive intonation and expressive voice quality which, in everyday parlance, are referred to jointly as "tone of voice". Although the phenomena in this category fit within the broad definition of sound symbolism given above, the editors are careful to point out that even tone of voice is outside the domain of what most linguists would consider sound symbolism proper. Nonetheless, they include this first category in the typology because it provides a link to the biological roots of sound symbolism.

The second category is "imitative sound symbolism": onomatopoeia. In the delightfully apt terminology of Richard Rhodes, these imitations of environmental sounds range from "wild" to "tame". Wild onomatopoeia is acoustically more accurate but involves sounds that are outside the inventory used in conventional speech. Typical examples are good imitations of the sounds produced by animals or machinery. Such sounds are difficult to convey in writing; cartoonists' attempts (such as "pffft" for the sound of an airtight package being opened) look outlandish and are only approximations. Tame onomatopoeia, on the other hand, stays within the boundaries of conventional speech. Words such as "clang", "buzz", and "meow" are easy to write and are often in dictionaries. To appreciate the degree of conventionalisation (arbitrariness) in tame onomatop-oeia, one can imagine a Japanese speaker and an English speaker both imitating the sound of a pig; they would probably produce quite similar sounds. But the tame onomatopoeic words in the two languages are quite divergent: English "oink-oink" versus Japanese "buu-buu".

The third category is "synesthetic sound symbolism": the use of sound to symbolise nonacoustic phenomena such as movement, size or shape. Movement is often accompanied by sound, of course, and objects of a certain size or shape often make characteristic sounds, so onomatopoeic words are frequently extended to non-acoustic uses. As an example, Robert Oswalt cites the use of English "whack" as a noun or verb denoting a blow rather than the sound of a blow. True synesthetic sound symbolism, by contrast, does not involve this kind of mediation by onomatopoeia. It depends instead on a perceived iconic relationship between the pronunciation of a word and some nonacoustic characteristic of what that word denotes. And there is an additional complication: the iconic relationship can be based either on the acoustic signal that results from a pronunciation or on the articulatory movements involved in producing it. Furthermore, the two modalities can motivate contradictory associations. Many languages appear to exploit a perceived connection between small size and vowels like the ee in bee, and between large size and vowels like the ah in bah. The iconic basis for this connection seems to be acoustic: one of the prominent resonances (the second formant) that gives a vowel its characteristic auditory quality is relatively high pitched in vowels like ee and relatively low pitched in vowels like ah. The association of high pitch with small size and low pitch with large size seems to be rooted in biology (as explained below). As Gerard Diffloth shows, however, the situation is just the opposite in the Bahnar language of Vietnam: vowels like ee are associated with large size and vowels like ah with small size. The iconic basis for the Bahnar system, Diffloth argues, is articulatory: in ee the tongue fills a larger proportion of the volume inside the mouth, whereas in ah the mouth feels almost empty.

The fourth category in the editors' typology is "conventional sound symbolism": language-specific associations between certain sounds or sound combinations and certain meanings. A stock example of this category in English is the initial gl consonant cluster in words such as "glitter", "glisten", and "glow", which suggests a meaning that Leonard Bloomfield specified as "unmoving light". As the editors note, conventional sound symbolism is very close to the arbitrary end of their continuum.

As to the biological basis for the acoustically-based example of synesthetic sound symbolism given above, John Ohala proposes what he calls "the frequency code". Based on the ethological work of Eugene Morton, the frequency code offers a unified explanation for some widespread linguistic uses of pitch. The key is that, in all species, there is a correlation between the frequency (ie pitch) of a vocalisation (from low to high) and the size of the vocaliser (from large to small). Two additional assumptions are required: the pitch of resonances, like basic pitch, can convey the size of the source; and it is a plausible metaphorical jump from size of vocaliser to size of what is denoted. Given these two assumptions, the tendency to associate vowels having high-pitched resonances (like ee) with smallness and vowels having low-pitched resonances (like ah) with largeness is natural.

One caveat - this book is not aimed at a general audience; most of the contributions presuppose solid training in linguistics. But it will almost certainly be the standard reference on this topic for years to come.

Timothy J. Vance is professor of Japanese, Connecticut College, US.