Every spoken language has stops in its consonant inventory, and stop-vowel syllables such as [pa] and [ta] are among the earliest linguistic sounds to be identified in the babbling and first words of typically-developing children. A large majority of spoken languages also have at least two series of stops that contrast in their associated laryngeal gestures. This dissertation investigates how the acoustic details of the laryngeal contrast are related to the order in which children master the different stop phonation-type categories in their native languages. Analyses of a cross-sectional database of productions collected from Korean-, Japanese-, English- and Greek-acquiring children (2;0-5;11) supports some well-established claims about universal characteristics of children's stop phonation-type mastery patterns across languages, and also suggest the potential role of language-specific acoustic properties in explaining seemingly exceptional mastery patterns.
A starting point for this crosslinguistic comparison is Jakobson's (1941/1968) claim that there are implicational universals in the mastery of stop phonation types: the aspirated or voiced categories are mastered after the voiceless unaspirated category when a language has a contrast involving either aspiration or voicing. Using the acoustic measure of Voice Onset Time (VOT: the latency between oral constriction release and the onset of voicing), studies of many languages have shown that voiceless unaspirated stops (with short lag VOTs) are mastered before one year of age, whereas voiceless aspirated stops (with long lag VOTs) are not mastered until two years of age in English (Macken and Barton, 1980a), Cantonese (Clumeck, Barton, Macken, and Huntington, 1981), and Thai (Gandour, Petty, Dardarananda, Dechongkit, and Mukongoen, 1986), and truly voiced stops (with lead VOTs) are mastered even later, at around age five in French (Allen, 1985), Thai (Gandour et al., 1986), Spanish (Macken and Barton, 1980b), and Hindi (Davis, 1995). Kewley-Port and Preston (1974) ascribed this universal order of mastery to the challenging aerodynamic and laryngeal control required to produce long lag VOT or lead VOT relative to short lag VOT.
I test this universal tendency of children's stop mastery and its relationship with the associated VOT properties in three relatively understudied languages (Greek, Japanese and Korean) and English. While the voiceless vs. voiced stops in English are successfully distinguished along the VOT range, the stop phonation-type contrasts in Greek, Japanese and Korean do not neatly fit into the three-way differentiation among truly voiced stops, ordinary voiceless stops, and aspirated stops. We address three specific puzzles that arise from the VOT characteristics of stop categories in the target languages in relation to children's production accuracy, and examine the effects of various acoustic parameters in predicting the order of mastering stop categories.
The analysis tools include native speaker transcriptions of the stops produced by 100 children (2;0-5;11) acquiring Korean, Japanese, English or Greek to estimate the relative production proficiency, and the acoustic properties of children's stop productions measured with VOT, H1-H2 and f0. The stop productions collected from 20 adult speakers in each target language were examined to establish the objective norms of acoustic variables used in the native language.