Once relegated to the margins of linguistics, iconicity (Haiman 1985) and ideophones (Voeltz & Kilian- Hatz 2001; Dingemanse 2012; 2018) have slowly reclaimed their place as linguistic phenomena worthy of study. Most work has been synchronic and elicited in nature. However, in the past few years, diachronic approaches have started to take hold as well (e.g. Akita et al. 2014; Van Hoey & Lu 2019). This study is conducted along similar lines, but focuses on onomasiology through computational methodology.
Van Hoey & Lu (2018) found that Premodern Chinese literary ideophones in the semantic domain matrix of LIGHT (1) exhibit a prototypical polysemous structure as to their semantic preference, similar to Geeraerts’s (1997) study of vergrijpen. This study, then, first replicates that analysis, using skipgrams as the foundation of the computational methodology. After this it is possible to enlarge the scope of the lexical items under investigation, i.e. LIGHT ideophones. This quantification provides a firmer empirical basis for arguing that many different frames (see Kövecses 2017) are involved with LIGHT ideophones, such as (but not limited to) LIGHT, FIRE, SUN, MOON, STARS, RED, FLOWERS, and GOLD.
(1) a. huīhuī 暉暉 ‘bright’ with SUN radical 日
b. huīhuī 煇煇 ‘bright’ with FIRE radical 火
c. huīhuī 輝輝 ‘bright’ with LIGHT radical 光
If these frames are then taken as the source of exploration, it is possible to reverse the perspective from a semasiological (‘meaning’ of a lexical item) to an onomasiological one (‘names’ of a certain meaning). This allows a statistically normalized comparison between the different ideophones for different times per frame — effectively displaying the dynamicity of the system. They also form a basis for comparison with the ideophones that ‘survived’ into Modern Mandarin Chinese (and other Sinitic languages).
Lastly, this quantified and onomasiological road leads to a comparison of the above-mentioned frames to the radicals (‘semantic indices’) often present in Chinese characters, it can be seen that for SUN (日 in 1a), FIRE (火 in 1b) and LIGHT (光 in 1c) there are differing tendencies of ‘radical attraction’ throughout time, unlike what one intuitively would expect, namely LIGHT ideophones having a statistically significant preference for LIGHT radicals. On the contrary, they tend to have a greater affinity for the metonymy FIRE FOR LIGHT or SUN FOR LIGHT, rather than the non-metonymous LIGHT itself.
These important findings provide further evidence for the variationist nature of ideophones, both diachronic as well as synchronic (in a given time period), as well as their the prototypicality of their structure – an aspect of their meaning and usage that is often overlooked by formal approaches that only focus on their morphology and autonomous syntax.