The Phonation Patterns Of Chinese And English Prosody By Native Speakers And L2Learners

Through a series of instrumental experiements, this thesis examines a duplex Chinese andEnglish prosody by both native speakers and L2learners. Based on phonetic andphonological principles, as well as from a perspective of signal processing, we investigatedthe suprasegmental features of narrow focus and rhythmic patterns of both Chinese andEnglish read speech of literary style. The first research concerns the “/a//C_C”construction at a narrow focus position and we found that different strategies are adoptedby English and Chinese native speakers and L2learners. F0, OQ and SQ are the three keyphonatory acoustic parameters under scrutiny within the scope of a three-dimensionalphonetic space. L2strategies are examined by measuring deviations from the controlledgroup to the experimental group. Since variable pairs cover L1-L2contrast, sentencelength, plus or minus narrow focus and sex difference, a24multi-variable matrix is foundover there. Therefore, internal variable comparison as well as a more complexmulti-variable interaction is investigated and analyzed. The matrix distance reflects eitherL1and L2’s traces and interaction in the interlanguage system. Materials are carefullyselected and experiments are designed specifically for obtaining both Chinese and Englishnative speakers’ and L2learners’ tokens of the two languages separately. In examiningtheir bidirectional processing strategy of prosodic features by the four groups of speakerstoward either languages, namely, Chinese and English, different phonation patterns areanalyzed and summarized. In speech production, relevant correlates or parameters of voicequality and respiratory features are collected and analyzed.For the English material, results on F0partly coincide with traditional research in that F0atnarrow focus position is higher than at non-narrow focus position. Besides, we discoverthat L1speakers are more sensitive in producing pitch contrast on narrow foci. No matterL1is Chinese or English, L1speakers produce more salient F0distinction for the contrastbetween narrow focus and non-narrow focus in their L1than in their L2. In other words,the F0contrast asymmetry on Chinese and English syllables of the same “/a//C_C”construction does not surpass the F0boundary in L1. The following formula can well account this symmetry on asymmetry–“L1’s control force> L2attraction force”.Speakers are more sensitive to the F0contrast for narrow focus/non-narrow focus in L1than be deviated by L2’s distraction force. Male and female speakers’ F0contrast onnarrow focus/non-narrow focus is different in absolute number scale but proportionate.Sentence length contributes to narrow focus realization in its duration. The duration ofChinese narrow focus syllable ranks as “+short&+narrow focus>+long&+narrowfocus>+short&-focus≈+long&-focus”. Chinese L2learners do better for longsentences than short sentences. The F0clustering sequence as from high to low–fCC>fAC> mCC> mAC.OQ results show strong sex difference. Male speakers’ OQ do not vary much at narrowfocus/non-narrow focus position. Both reveal a gradual increasing tendency which is atypical OQ feature at sentence final position. Female speakers’ OQ at non-narrow focusposition is comparatively lower than at narrow-focus position. Therefore females’ strategyat non-narrow focus position is represented by more breathy voice and less power.SQ also reveals sex differentiation. Male speakers’ SQ rise while females’ fall. Malespeakers’ SQ value is much higher than females’. Male speakers adopt intense lowfrequency power which is different from female speakers’ high pitch strategy to realizenarrow focus. L1/L2factors contribute to the realization of narrow focus. OQ and SQ as awhole is compensatory to F0.Analysis on Chinese material showed a stable F0at Chinese55tone on narrow focussyllables. English syllables contain a decreasing F0. We regard it shows a “tight” ChineseCV structure in contrast with a “loose” English CV（C） structure. American L2learners ofChinese have a less obvious narrow focus/non-narrow focus F0distinction than that intheir L1English. We think it shows the fact that L2effect cannot outwin L1scope. LikeChinese L2learners of English, American L2learners of Chinese also have acompensatory effect for OQ, SQ as a whole to F0. But the fitted tilt rate of their OQ andSQ are higher than Chinese L2learners of English. An exciting byproduct is found. Some tokens witness an ending “double-peak” EGG signal.We believe it shows two possibilities. One is that the ventricular folds come together in thevibration period of vocal folds. The other is that the vocal folds open again before acomplete closure.“Double-peak” or sometimes even “multiple-peak” become moreobvious towards sentence final position. It fully supports a governing of respiratoryprosodic features over the phonation of segments. Fujimura et al （2009） has also observedthis phenomenon in traditional Japanese Opera Noh.Laryngograph and respiratory belt are used to collect EGG signal and breath signal as twoimportant parameters concerning phonation in speech production process. EGG signaldetected by the laryngograph basically coincide with a traditional practice-H1–H2value,i.e., the first and the second harmonics difference, to examine a tense or lax vocal musclestate. And the further method is more accurate than a rough speculation of H1–H2harmonics difference for a more exact calculation than the latter’s LPC calculation method.Results show that speakers adopt specific voice quality and breathing features as a bettersolution of achieving L2nativeness. In addition, obvious different phonation types arefound between male and female learners. Average female index proved a pitch-increasestrategy while average male index proved an SQ-increase strategy. Namely male speakerscontrol their glottis to a much lower position to lengthen the vocal tract and produce moreenergy, which is represented by higher SQ, for a narrow focus token. Female speakersexpress narrow focus by raising F0.A respiratory belt study afterwards indicates that Chinese takes a “Strides&Swift-Paced”respiratory pattern which contains obvious major and minor breaths while English takes a“Moderate-Equal Step” respiratory pattern. Chinese language reflects high breast breathinhale resets at the left boundary of prosodic segments while English language tends tohave more moderate resets and lower exhale resets （syllable numbers and tempo to beconsidered in future research）. Speakers tend to borrow their L1respiratory pattern into their L2processing strategy. Results also show that this borrowing seems to be negativelycorrelated with their L2oral proficiency. The higher the speakers’ proficiency is, the lessborrowing there is.This study provides phonation evidence for narrow focus phonetic representation as well asquantified description of respiratory patterns of native speakers and L2learners’ prosodicmechanism of Chinese and English.