Phonetic Development Of Voiceless Obstruents In Mandarin-speaking Children

This dissertation investigated the acquisition order of voiceless obstruents produced by Mandarin-speaking children,and examined whether children’s speech development unfolded in a continuous fashion or in a discrete manner.This study employed two kinds of methods:transcription and acoustic methods.The former relied on native adults’ auditory-based transcription,and the latter depended on both temporal and spectral parameters such as VOT,frication duration,F2 onset,spectral mean,standard deviation,spectral skewness and spectral kurtosis.Specifically,this study addressed the following three questions: 1)What are the acoustic properties of Mandarin voiceless stops,voiceless fricatives and voiceless affricates?2)What is the acquisition order of Mandarin voiceless stops,voiceless fricatives and voiceless affricates produced by children aged 3-5 years old based on the transcription and acoustic methods and how does the observed acquisition order relate to pertinent acquisition theories? 3)Does children’s speech develops in a continuous or discrete fashion by examining the acoustic similarities and differences between children’s and adults’ speech productions?Thirty-six children aged 3-5 years old and twelve adults participated in three speech production experiments.In the first experiment,both children and adults were required to repeat 18 disyllabic words with one voiceless stop in the word-initial position four times after hearing an audio prompt while watching the cartoon picture on a laptop screen.In Experiment Two and Three,they were required to repeat 15 disyllabic words with one target fricative and18 disyllabic words with one target affricate in the word-initial position following the same procedures,respectively.In each experiment,I first quantified the acoustic properties of Mandarin voiceless obstruents produced by the adults in order to form the baseline and criterion,examined the acquisition order based on the transcription and acoustic methods,and then analyzed the acoustic similarities and differences between the children and adults to examine the issue of continuity and discontinuity.The results of the first experiment showed that the accuracy rate of children aged three years old was significantly lower than that of the other groups for aspirated stops.However,no such significant differences were found for unaspirated stops.In addition,the results of acoustic analysis suggested that children aged three years old could distinguish all stops in terms of place of articulation,indicating that transcription methods might obscure the actual developmental route and rate.With respect to the acoustic similarities and differences for aspirated stops,there were significant age differences between the children and adults in terms of F2 onset,spectral mean,standard deviation and spectral skewness,with no significant age differences between child groups.However,no significant age effects werefound in terms of VOT and spectral kurtosis.Regarding unaspirated stops,it was found that only spectral kurtosis varied significantly as a function of age,without any significant age differences in terms of other acoustic parameters.In addition,the mismatches and variability diminished as age increased concerning the distribution of certain acoustic parameters.The aspiration pattern also showed gradual changes between child groups.These results indicate that children are gradually approaching adult-like patterns through constantly adjusting their articulatory precision with the development of speech motors.In the second experiment,both the transcription and acoustic methods demonstrated that labiodental fricatives,velar fricatives and palatal fricatives were acquired earlier than alveolar and retroflex fricatives.The two methods,however,did not agree on the acquisition sequence of alveolar and retroflex fricatives.The former revealed earlier acquisition of alveolar fricatives than retroflex,while the latter revealed an opposite pattern.Regarding the continuity and discreteness of speech development,findings suggested that there existed significant differences between the children and adults in terms of F2 onset,spectral mean and standard deviation.However,no significant differences were found between child groups in all selected acoustic parameters.Moreover,the results showed that gradual changes happened between child groups in terms of the distribution of several acoustic parameters,indicative of incremental approximation to adult-like patterns.In the third experiment,the results of the transcription analysis revealed that alveopalatal affricates were acquired at the age of three years old,and retroflex and alveolar affricates were acquired at the age of five years old.However,different acquisition orders were found by the acoustic analysis,which showed that for aspirated affricates,alveopalatal affricates were acquired at the age of three years old,retroflex at the age of four years old,and alveolar affricates at the age of five years old;regarding unaspirated affricates,alveopalatal affricates were acquired at the age of three years old as well,and both retroflex and alveolar affricates were acquired at the age of four years old.In addition,the results of acoustic analyses demonstrated that there were significant age differences between children and adults in terms of F2 onset and spectral mean,with no significant differences between child groups,whereas no significant age differences were found for duration spectral skewness and spectral kurtosis regardless of aspiration types.Significant age differences in spectral standard deviation were only found for unaspirated affricates.Children’s deviation from adults diminished over time in terms of certain acoustic parameters,and the aspiration pattern demonstrated gradual changes between child groups as well,indicating that children constantly fine-tune their articulatory precision to approach adult-like patterns.The observed acquisition order of Mandarin voiceless obstruents suggests that child speech development is a complex process,and is influenced by a combination of factorsincluding universality,oromotor maturation and ambient language.No single theoretical framework could provide a unified explanation for the phonological acquisition order.An integrated model,incorporating various factors with different weightings,is proposed to account for the acquisition order.In addition,the acoustic similarities and differences between children and adults provide further evidence for the continuous theory that children’s speech development unfolds in a continuous fashion,not in a discrete manner.