Meanings below the morpheme: sound symbolism
The question of what level of grammatical structure a meaning should be attributed to may often be problematic, and boundary cases, where meanings seem to straddle several different grammatical units, occur quite frequently. One such boundary case is sound symbolism, (also known as ideophony or onomatopoeia). This is the existence of semi-systematic correspondences between certain sounds and certain meanings, usually within the domain of the individual morpheme, such as English clash, clang, clatter, etc. Such associations may sometimes have a clear imitative basis, as with English click, thwack, meow, etc. Sound symbolism is by no means limited to English, of course. In Ilocano (Cordilleran, Philippines), for instance, a high front vowel is often used in words denoting high pitched sounds, as in (18):
Here the choice of vowel imitates the characteristic timbre of the sound referred to. Similarly, the alveolar fricative is often found in words representing rustling sounds or the sound of water:
A possible connection might be discerned here between the acoustic quality of the fricative and the irregular, ‘perturbed’ sound of the referent. But the imitative basis of such associations is often less obvious, at least to English speakers. Egbokhare (2001: 90–91), for example, documents the fact that many words indicating ‘smallness’ contain kp in Emai (Niger Congo, Nigeria):
In all these cases we have a sound-meaning correspondence which exists below the level of the individual morpheme. Neither the high front vowel nor the alveolar fricative in Ilocano, nor kp in Emai can, formally, be considered as individual morphemes, since one cannot remove them from the ideophonic words in (19)–(20) and retain possible roots to which other morphemes could attach. Yet the correspondence is widespread: although not every s in Ilocano is used in words referring to rustling sounds (cf. sarotsot ‘quick succession’, Rubino 2001: 315), the correspondence is systematic enough to allow a hearer who is unfamiliar with karasakas, for instance, to infer that the word probably refers to some sort of sound. Reference to a rustling sound can therefore be considered as, in some way, a semi-predictable part of the meaning of a unit which is neither a word nor a morpheme. Yet it is only in the words in which they occur that this meaning exists: in describing sound symbolism in Emai, it is necessary to specify that there are many words containing kp which do not refer to small objects (e.g. úkpun ‘cloth’; ókpósó ‘woman’; Schaefer 2001: 344). Sound symbolism can therefore be considered simultaneously as a property of a word and of the relevant submorphemic unit, and the description of sound symbolism in these languages must invoke both lexical and submorphemic units: the reference to sound is conveyed by a particular segment or sequence of segments, but only in certain words.
