Bridging phonology, meaning, and written form across time: Introducing a database of Chinese ideophones — CHIDEOD (2019)
Ideophones have been receiving renewed interest in the last three decades (see e.g. Voeltz & Kilian-Hatz 2001; the Iconicity in Language and Linguistics series). However, linguistic studies have been synchronic in nature for the most part (Dingemanse 2018). We address this diachronic research gap by constructing a database for literary Chinese ideophones (following Mok 2001; Van Hoey 2018) that explores the variables such as:
1. the phonology across different stages of Chinese
2. the morphology of ideophones (different types of partial and full reduplication),
3. sensory domain mappings based on a cross-linguistic implicational hierarchy (cf. the
author; Dingemanse 2012),
4. Chinese characters and their radical support (see below).
Through this database it is possible to investigate the two different modalities of phonological form and written form, an analysis licensed by the Chinese model of Meaning//phonology / writing (xíng-yīn-yì 形-音-義 ‘shape – sound – meaning’, cf. Baxter & Sagart 1998, as opposed to the mere meaning-form pair of most Construction Grammar approaches).
PHONOLOGICAL FORM
Currently, there are 1918 ideophones in the database, of which 52% are of the full reduplication type, and 47% belong to partial reduplication type. This latter group of partially reduplicated ideophones captured our attention: morphologically speaking, 36% of these are base-reduplicant, 10% are reduplicant-base, and for 51% the base is unclear. Looking at vowel alternations of reduplicants (like English zig-zag), 35% types exhibit this pattern. However, there appears no (productive) difference in what ideophones with or without nucleus alternation can depict, see Table 1.
WRITTEN FORM
While there is evidence (Horodeck 1987; DeFrancis 1984; Unger 2004) that the sound part of a Chinese character plays a more important role in their processing, one would expect that for ideophones, being “marked words that depict sensory imagery” (Dingemanse 2011; 2012) make statistically more use of semantic radicals functional components. This so-called radical support is statistically significant in the partial reduplicated types in the database (p = 0.0002468). Interestingly, they radicals follow a Zipf-distribution, with WATER being the most commonly used radical, e.g. wù-jué (< Middle Chinese mit-kwet) 沕潏 ‘bubbling forth from a fountainhead’ in which 氵 is WATER.
This database and the preliminary conclusions drawn from it provides a base line, as well as new opportunities, for future comparative and diachronic studies of phonological, semantic, and orthographic nature with respect to iconicity.
