In the course of acquiring their first language, children develop a set of phonemes specific to that language. Learning two languages simultaneously requires the acquisition of two sets of phonemes. Do bilingual children keep these two sets of phonemes separate from each other, or do the phonemes of the two languages interfere with one another during any stages of language development? If there is interaction between the two systems, what is the nature of that interaction and how does it differ from that found in adult bilinguals? To address these questions, the current project examines how the two phonetic systems of bilingual children interact by comparing French-English bilingual children to age-matched monolingual children of both languages and adult French-English bilinguals. The study focuses on the acquisition of diphthongs, which are complex vowel segments that have not received much attention in the literature. Previous studies on the production of monophthong vowels have shown effects of language interaction in bilingual children, as evidenced by overlapping phonetic properties, which were significantly different from what was found for monolingual speakers. The results of this study will indicate whether the previously established findings on the acquisition of monophthong vowels can be generalized to that of the diphthongs. Additionally, by examining the acquisition of diphthongs, this project also addresses significant methodological issues on how to best capture the dynamic properties of complex segments by using the Smoothing Spline ANOVA, which is a relatively novel method for comparing differences in vowel production. This project will determine whether or not this technique is adequate for capturing diphthongs' dynamicity and whether or not it will produce more precise results than previous techniques.
The results of the current project will be relevant to parents of bilingual children, and to various associations working with bilingual children that may be curious about the progression of children's dual language acquisition. The results may also be relevant to clinical applications such as speech-language pathology. Given the phonetic complexity of diphthongs, the current study may reveal that they are generally more difficult to acquire than monophthong vowels. Thus, children may show speech delays involving the production of diphthongs. The results of the current study may become useful to clinicians for the construction of baseline data regarding the production of diphthongs. Speech therapists may then refer to this baseline, particularly when trying to differentiate between real clinical speech problems and potential byproducts of bilingualism, which may otherwise be interpreted as speech impairment.
Bilingualism has been the subject of a large number of studies in linguistic, the main objectives being not only to understand how the different languages of bilingual speakers interfere with one another, but also to examine how similar or different bilingual speech production is from that of monolingual speakers. The purpose of this dissertation research project is similar to that found in those previous studies, but with a primary focus on diphthongs, which have not received much attention in the acquisition literature. Diphthongs are complex vowels composed of two vocalic sounds, as in the word 'bye.' The same combination also exists in French, as in the word 'baille,' which is the third person singular form of the verb 'to yawn.' However, in English, the combination is considered part of the same segment or vowel, whereas in French, the combination is said to contain two distinct segments, a vowel and a consonant. As part of my dissertation, I carried out various acoustic analyses on data collected from adult speakers in each language and the results showed that these two categories have distinct phonetic implementations in terms of formant frequency patterns, but not of duration or intensity measures. Given these differences at the phonetic level, speech samples were also collected from bilingual children to examine whether they were able to maintain separation between the two categories. Overall, the French-English bilingual children recruited for the study varied greatly in their ability to maintain separation between these two categories; the acoustic details of their implementation depended on their language input and their age. In the great majority, all bilingual children were able to produce separate categories. However, those who attended bilingual schools in which they were taught in both languages generally produced the vocalic combinations with phonetic properties that were similar to monolingual children in both languages. The children who attended English-only schools, on the other hand, differed from the monolinguals in either language or both languages. In terms of age difference, the majority of the children in the younger age group showed productions that were different from monolinguals in both languages, suggesting they may use intermediate frequency values between the two categories. These younger children showed similar production patterns as the older children who attend English-only schools. This result was not surprising, given that both groups mainly receive French input at home only. The overall results seem to suggest that children who receive French input elsewhere than in the home setting were better at producing monolingual-like vowel-glide combinations. The finding is in line with previous studies, which showed that learners greatly benefit from variation in their source of input. By focusing on complex vowels, this dissertation also made methodology-related contributions to the field of phonetics. There is a long-standing debate in the phonetics literature regarding how to best represent the dynamic properties of diphthongs. This study used a relatively novel methodology called Smoothing Spline analysis of Variance (SSANOVA) to capture the dynamicity of these complex vowels, allowing the evaluation of their overall formant patterns rather than just comparing the onset and offset points. This method was very useful as it allowed an analysis of onset/offset regions of formant trajectories and a fine-grained analysis of glide transition between these two regions simultaneously. Although the primary investigation was bilingual acquisition, this dissertation also provided contributions to the study of monolingual language development. A developmental analysis, in which diphthong production was compared across three age groups, namely preschool children, school age children, and adults, was carried out and further confirmed the special status of diphthongs. The results indicated not only that there is no phonetic correspondence between the diphthongs and the monophthongs with which they are transcribed at all three stages of development, but also that diphthongs do not necessarily follow the same developmental trends as the monophthongs with respect to the decrease in formant frequency values and variability over time. With respect to broader impact, the results of the current study could be relevant to parents of bilingual children who may wonder whether or not bilingual education is useful for their children's acquisition of their native language. These results could also be relevant for speech-language pathologists, who treat bilingual children's speech delays in English diphthong production. The study suggests that a thorough investigation of a bilingual child's language input amount in both languages should be carried out in order to adequately discriminate between real clinical speech problems in diphthong production and byproducts of bilingualism due to dual language usage patterns.
