Phonological differentiation before age two in a Tagalog–Spanish–English trilingual child

Abstract

This study focuses on a trilingual toddler's ability to differentiate her Tagalog, Spanish and English productions on phonological/phonetic grounds. Working within the articulatory phonology framework, the word-initial segments produced by the child in Tagalog, Spanish and English words at age 1;10 were narrowly transcribed by two researchers and her accuracy levels in employing/reproducing the specific gestural properties of word-initial consonants (e.g. constricting organ, location, degree and glottal gestures) were compared across languages. The child's phonetic inventories in each language were also compared to those produced by monolingual peers to assess whether trilingual exposure had any consequence on phonological production. The results indicate that the child's accuracy levels in word-initial segments differed significantly from language to language reflecting distinct levels of phonological development. The child's performance with organ and location was indeed higher in her stronger language(s), while success in reproducing constriction degree tended to be greater in those languages with a larger variety of fricatives. The child also produced a variety of language-specific phonemes that went beyond the average monolingual inventory, suggesting that multilingual exposure might lead to a heightened attention towards the phonological properties of these languages, thereby enhancing phonological production.

Keywords:

• trilingual development
• articulatory phonology
• language differentiation
• phonological production

Acknowledgements

I am very grateful to Kathryn and to her family who not only allowed me to undertake this investigation but also helped with the data collection and transcription. I also thank the members of my dissertation committee, Elaine Andersen, Jo Ann Farver, Toby Mintz, Carmen Silva-Corvalán and especially Dany Bird for their helpful comments and suggestions on earlier drafts of this paper.

Notes

1. Notice that because Tagalog, Spanish and English share many lexical items (Rau, 1992), each vocabulary value included not only words that were exclusively Tagalog, Spanish or English but also neutral lexical items (i.e. words common to or ambiguous among two/three languages). Neutrals were counted separately in each language as long as they were produced in the appropriate language context. This was done in order not to underestimate the child's lexical knowledge in each language, be it in the form of language-specific or neutral word.

2. In monosyllabic words (e.g. ) or in disyllabic items in which stress fell on the first syllable (e.g. ), the ‘target segment’ corresponded to the adult word's beginning consonant (e.g. /b/). In multi-syllabic words in which stress fell on the second or third syllable (e.g. ), the target segment was the stressed syllable's first segment (e.g. /t/). Omitting unstressed syllables in multi-syllabic words has been indeed documented as one of the most frequent phonological processes in early child language (Schwartz, Leonard, Folger, & Wilcox, 1980). In the case of consonant clusters, the child's word-initial phone was compared to the earliest acquired segment in the cluster, usually a stop in a stop + tap/glide/liquid sequence (for example [g] in grande, ‘big’, [t] in truck and [k] in Clifford) or a fricative in a fricative + tap/glide/liquid sequence (for example [f] in frío, ‘cold’, [θ] in three and [f] in fly; Stoel-Gammon, 1985).