Onomatopoeia and metonymy

5. Conclusions
In this paper we aimed to discuss the position of onomatopoeia in linguistic theory and argue that the underlying conceptual mechanism behind onomatopoeic formations is metonymy. Accordingly, we proposed the following definition for onomatopoeia: onomatopoeic forms are novel or conventionalized words in which a part of the phonological form is perceived to be similar to the referent or to a sound metonymically associated with it. This implies that metonymy can be detected in onomatopoeia on two levels: a) the referential level (in the form of e.g., the product for the producer metonymy – c.f. cuckoo); and b) the structural level (in the form of the form for part of the form metonymy,¹³ whereby the whole word form stands for a couple of the sounds that do effectively participate in the process of imitation – c.f. click).
This view of onomatopoeia highlights the importance of the perception of similarity between the onomatopoeic word and its referent. The reason why perception is a crucial component in any definition of onomatopoeia is that in most cases the word form has undergone substantial modifications to adhere to phonological and morphological rules. Thus, many of the word forms that are designated as “onomatopoeic” (including those on the OED’s list) have become conventionalized to such a degree that the motivated link between form and meaning is purely a matter of subjective opinion. Thus, onomatopoeia are largely based on convention (just as most of language is), which is an often understated aspect of onomatopoeia, in comparison to the perceived “natural link” that we tend to feel in the case of such words. Attridge (1984, p. 1124) elaborates on this point by stating that onomatopoeia – similarly to any sign system – are conventional: onomatopoeic words can be interpreted with reference to the linguistic system within which they are produced. Thus, onomatopoeia also need to be learned; they have conventions in spelling (such as the common practice of using successions of letters in onomatopoeic words to represent an ongoing rush of sound, e.g., psst), and all sorts of (partly motivated, partly conventionalized) associations between sounds and letters. Last but not least, not only do we need to know what sound the onomatopoeic form is supposed to represent, we also have to have a certain degree of familiarity with the sound itself.
But why are onomatopoeia nevertheless felt to be natural in most cases? It is most probably due to the fact that we become used to our own language and do not have an objective view of it any longer. Thus, we come to think of the word form that is supposed to imitate a particular sound as just right, even if on closer inspection the word form does not even come close to the sound that it is supposed to resemble. In this respect, speakers “hear” the sound that are represented by onomatopoeia, which presupposes a reciprocal relationship between language and how reality is conceptualized.
Sounds do not carry meaning in themselves; only when they fit a particular meaning can the onomatopoeic nature of the word come into the fore. Thus, any word that is associated with a particular sound can be imbued with onomatopoeic resonance, even if there was no such resemblance in the first place. The drive to “fit” meaning to sound and sound to meaning is all-pervasive in language: even the slightest hint of resemblance or a peripheral property is enough for us to establish a connection.