02.11.2021 Marilyn Vihman

Effects of phonological structure on early word production


Studies of early word production in English and other European languages have long provided evidence that children begin with whole-word representations, as first suggested by Ferguson and Farwell (1975). This is primarily the effect of child responses to variegation, or the presence of more than a single supraglottal consonant within the lexical unit (Vihman et al., under review, a), which constitutes a challenge for both (i) the retention and representation and (ii) the articulation of novel word forms. 

The proposal that children begin with whole-word representations is based on several types of evidence, including the widespread occurrence of consonant harmony, in which the place and/or manner of articulation of one consonant spreads to another; variability in the production of the same consonant in different lexico-phonological contexts; omission of the initial consonant in words with final-syllable accent (as in French or Hebrew) or medial geminates (Finnish, Italian, Hindi); or the shift or spread of palatalization or duration from one consonant to another (Japanese). 

However, Vihman et al. (under review, a, b) found that children acquiring Mandarin show none of those indices of whole-word representation. Instead, those children tend to substitute whole syllables – the choice of syllable being idiosyncratic by child – for target syllables that share neither onset nor nucleus, a process not seen in English, Finnish, French or Japanese, for example. This strongly suggests that Mandarin learners are operating with ‘templatic syllables’, just as European or Japanese children appear to be operating with whole-word phonological templates (also idiosyncratic by child: Vihman, 2019). We conclude that Mandarin, with its highly frequent use of reduplication in speech addressed to infants, its ‘omnitonality’ (Matisoff, 2001), or tone on each syllable, highly constrained syllable structure, and relative co-articulatory independence of the syllable (Ma et al., 2015) shapes early word representations in a way that is unique in the child phonological data described so far. This makes the need for much wider cross-linguistic investigation particularly clear, if we are to draw better founded generalizations as to the developmental origins of lexical representation.



Ferguson, C. A. & Farwell, C. B. (1975). Words and sounds in early language acquisition. Language, 51, 419-439. 

Ma, L., Perrier, P. & Deng, J. (2015). Strength of syllabic influences on articulation in Mandarin Chinese and French. Journal of Phonetics, 53, 101-124.

Matisoff, J. A (2001) Genetic versus contact relationship: Prosodic diffusibility in South-East Asian Languages. In A. Y. Aikhenvald & R. M. W. Dixon (eds.) Areal diffusion and genetic inheritance: problems in comparative linguistics, pp. 291-321. Oxford: OUP

Vihman, M. M. (2019). Phonological Templates in Development. Oxford: OUP.

Vihman, M. M., Ota, M., Keren-Portnoy, T. & Lou. S. (under review, b). A challenge to whole-word phonology? A study of Japanese and Mandarin. 

Vihman, M. M., Ota, M., Keren-Portnoy, T., Lou. S. & Choo, R. Q. (under review, a). Child phonological responses to variegation in adult words:  A cross-linguistic study.