In speech perception tasks young infants show remarkable sensitivity to fine phonetic detail. Despite this impressive ability demonstrated at early ages, studies of word learning in young toddlers indicate that they have difficulty learning similar-sounding words. This evidence suggests that infants may not be using this speech-perception ability as they begin to learn words. The studies in this thesis were designed to test how infants' speech-perception skills are used in the early stages of word learning. Using a simple habituation procedure, we have shown in earlier work that 14-month-old infants, but not younger infants, are able to learn the association between novel nonsense words and objects (Werker, Cohen, Lloyd, Casasola, & Stager, 1998). The current series of experiments used this simple habituation procedure to test whether infants use minimally contrastive phonetic detail in the very early stages of word learning. In this thesis, I show that 14-month-old infants, who are on the cusp of word learning, while still able to discriminate phonetically-similar words in a speech perception task, do not incorporate minimally contrastive phonetic detail when first forming word-object associations. Infants of 8 months of age do, however, appear to use fine phonetic detail in a similar task. Taken together, these results suggest a decline in the phonetic detail used by infants as they move from processing speech to learning words. I hypothesize that this decline may occur as infants move from treating the task as one of speech perception to treating the task as one of word learning. / Arts, Faculty of / Psychology, Department of / Graduate

Thesis (Ph. D.)--University of Texas at Austin, 2003. / Vita. Includes bibliographical references. Available also from UMI Company.

Not available / text

Thesis (Ph. D.)--University of Washington, 2002. / Vita. Includes bibliographical references (leaves 187-207).

Lei, Ka Yan Margaret. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2007. / Includes bibliographical references (leaves 88-94). / Abstracts in English and Chinese. / Acknowledgments --- p.i / Abstract --- p.iv / 摘要 --- p.vi / Chapter CHAPTER 1 --- INTRODUCTION --- p.1 / Chapter CHAPTER 2 --- BACKGROUND --- p.8 / Chapter 2.1 --- Infants' perception of speech contrasts --- p.9 / Chapter 2.1.1 --- Discrimination of segmental contrasts --- p.10 / Chapter 2.1.2 --- Discrimination of suprasegmental contrasts --- p.16 / Chapter 2.2 --- The Cantonese tonal system --- p.22 / Chapter 2.2.1 --- Properties and classification of the Cantonese tones --- p.22 / Chapter 2.2.2 --- Distributional pattern of the Cantonese tones --- p.24 / Chapter 2.3 --- Tone perception in Cantonese-speaking adults and children --- p.27 / Chapter 2.3.1 --- Tone perception in Cantonese-speaking adults --- p.27 / Chapter 2.3.2 --- Tone perception in Cantonese-speaking children --- p.29 / Chapter 2.3.3 --- Discussion --- p.31 / Chapter 2.4 --- Tone acquisition --- p.32 / Chapter CHAPTER 3 --- METHODOLOGY --- p.34 / Chapter 3.1 --- Operant conditioning --- p.35 / Chapter 3.2 --- Background of the paradigm --- p.35 / Chapter 3.3 --- Subjects --- p.38 / Chapter 3.4 --- Stimuli used in the experiments --- p.39 / Chapter 3.3.1 --- Auditory stimuli --- p.39 / Chapter 3.3.2 --- Visual stimuli --- p.41 / Chapter 3.5 --- Experimental site --- p.43 / Chapter 3.6 --- Procedure --- p.46 / Chapter 3.7 --- Technical control of experiment --- p.49 / Chapter CHAPTER 4 --- RESULTS OF THE STUDY --- p.52 / Chapter 4.1 --- Details of experiment --- p.52 / Chapter 4.2 --- Assessment criteria --- p.55 / Chapter 4.3 --- Results --- p.58 / Chapter 4.3.1 --- Discrimination performance of Group I subjects --- p.61 / Chapter 4.3.2 --- Discrimination performance of Group II subjects --- p.68 / Chapter 4.3.3 --- Discrimination performance of Group III subjects --- p.72 / Chapter 4.4 --- Comparison with previous studies --- p.72 / Chapter 4.5 --- Summary --- p.73 / Chapter CHAPTER 5 --- DISCUSSION & CONCLUSIONS --- p.74 / Chapter 5.1 --- Strengths of the VRISD paradigm --- p.74 / Chapter 5.2 --- The use of operant conditioning --- p.76 / Chapter 5.3 --- Relative nature of tone --- p.77 / Chapter 5.4 --- Acoustic similarity between the tones --- p.78 / Chapter 5.5 --- Interplay between production and perception --- p.79 / Chapter 5.6 --- Association between tone perception and tone acquisition --- p.81 / Chapter 5.7 --- Tonal versus segmental discrimination --- p.82 / Chapter 5.8 --- Tone perception in connected speech --- p.84 / Chapter 5.9 --- Gender identification --- p.85 / Chapter 5.10 --- Conclusions --- p.86 / Bibliography --- p.88

Thesis (Ph. D.)--University of Texas at Austin, 2005. / Vita. Includes bibliographical references.

In speech, vowels and consonants are two the basic sounds that combined result in lexically meaningful items in all languages. In tone languages, changes in pitch, tone differences also make meaningful lexical distinctions in spoken words. Young infants appear to have no trouble perceiving speech sounds and their production of sounds peculiar to their particular language environment proceeds relatively smoothly and rapidly compared with adults’ acquisition of foreign languages. One way of looking at how infants come to acquire speech sounds of their first language is by examining the speech input they receive. The term infant-directed speech (IDS) has been coined to describe the special way adults and even older children speak to infants. IDS is different to adult-directed speech in various acoustic/phonetic modifications, such as exaggerated prosody, increased pitch and vowel hyperarticulation (Burnham, Kitamura, and Vollmer-Conna, 2002; Kuhl et al., 1997). The exaggerated prosody and increased pitch appear to be related to the expression of affect and gaining infants’ attention (Burnham, Kitamura, and Vollmer-Conna, 2002), whereas vowel hyperarticulation appears to be related to infants’ speech development for a number of reasons. Firstly, investigating how adults speak to foreigners, Uther, Knoll, and Burnham (2007) found that vowels are also hyperarticulated in foreigner-directed speech as in IDS, while other acoustic modifications such as exaggerated prosody and increased pitch, related to affective and attentional factors, are not present in foreigner directed speech. Secondly, Liu, Kuhl, and Tsao (2003) found a positive correlation between vowel hyperarticulation and infants’ native speech perception; mothers who hyperarticulated their vowels more had infants who were better able to discriminate native consonant contrasts.\ While vowel hyperarticulation in IDS to 6-month-olds has been investigated in both tone languages such as Mandarin (Liu et al., 2003), and non-tone languages such as Russian, Swedish, American English (Kuhl et al., 1997) and Australian English (Burnham et al, 2002), no parallel studies have been conducted on the possibility of tone hyperarticulation in tone language IDS. If vowel hyperarticulation is related to infants’ language development then tones in tone hyperarticulated. The possibility of tone as well as vowel hyperarticulation in IDS of the tone language Cantonese, and the development of hyperarticulation across the first 12 months of infancy were investigated here using a longitudinal sequential cohort design. Two groups of native Cantonese mothers were recorded speaking to their infants, the first group at 3, 6, and 9 months, and the second at 6, 9, and 12 months. The study had four main aims (1) to investigate whether tone hyperarticulation occurs in IDS in a tone language Cantonese (2) to investigate whether vowel hyperarticulation occurs in IDS in Cantonese (IDS in this languages had not yet been investigated) and if 1 and 2 are the cases (3) to compare tone and vowel hyperarticulation, and (4) to chart the development of tone and vowel hyperarticulation across the infant’s first 12 months. Contrary to previous findings of vowel hyperarticulation in English Russian, Swedish, and Mandarin IDS to 6-month-olds (Burnham et al., 2002; Kuhl et al., 1997); vowel hyperarticulation was not found for Cantonese IDS. More detailed acoustic analysis examining different dimensions of the vowel space suggest that after the infant is 3 months old, mothers’ vowels begin to be hypoarticulated in IDS compared to ADS on dimensions of back versus front, and high versus low. This pattern of results is consistent with vowel perception studies which suggest that infants have already tuned into the native vowel categories by 4 to 6 months (Polka and Werker, 1994). Tone hyperarticulation, on the other hand, was indeed present at 3 months and increased to peak at 6 to 9 months before declining at 12 months. This pattern of tone hyperarticulation across the first year of infancy is consistent with infant language development – in which attenuation of perception of non-native tones had been found between 6 to 9 months (Mattock and Burnham, 2006). Moreover, detailed phonetic analysis revealed that while the level tones are more hyperarticulated than the contour tones, tones with similar onsets and offsets (i.e., the two rising tones) are actually hypoarticulated in IDS at 9 and 12 months, a time when infants have already tuned into native tones. Finally, results from a preliminary native speech discrimination study using the same infants provide some initial indication that mothers who hyperarticulate tones more also had infants who are better able to discriminate native Cantonese consonants. Together these results suggest that in Cantonese IDS vowels are underspecified whereas tones are consistently over-specified particularly at 6 months when infants are tuning into native tones. Moreover, during this initial period of tone acquisition, only level tones are over specified while tones with similar onsets and offsets are underspecified. It seems likely that for Cantonese language environment infants, during the early stages of language acquisition, pitch information specified by level tones is sufficient for initial acquisition of information about the Cantonese tone space and that information about vowels is not so essential at this time. These studies show that there is indeed tone hyperarticulation in IDS in tone languages, and that in order to make sense of the vowel hyperarticulation data in tone languages, it is important to investigate both vowels and tones in tone languages with complex tone systems such as Cantonese, instead of simply applying Anglocentric notions of vowel hypoarticulation. / Doctor of Philosophy (PhD)

Not available / text

Infants with hearing loss (HL) are being diagnosed and fitted with amplification earlier than ever before. In order to acquire speech and language abilities that are on par with their normal-hearing (NH) peers, hearing-impaired (HI) infants require amplification that provides appropriate exposure to the sounds of their native language. To date, no research has addressed this issue and thus, there is a need to examine the type of amplification that is most suitable for infants during the early stages of language acquisition. In this thesis, three different amplification types were compared: one that preserves the natural spectral shape of speech (unmodified speech); a second that emphasises high-frequency speech information (positive spectral tilt); and a third that emphasises low-frequency information (negative spectral tilt). NH infants were tested to examine how each of these amplification types affects speech perception and to explore whether infants at different stages of language acquisition find modified spectral tilt a help or hindrance in perceiving native-language speech contrasts. A visual habituation (VH) procedure was used to test 288 6- and 9-month-old NH infants on their ability to discriminate the high-frequency fricative contrast /f/- /s/; the mid-frequency approximant contrast /l/-/r/; and the low-frequency vowel contrast /5/-/N/ under modified spectral tilt conditions. For each speech contrast, 96 infants were tested in one of three conditions: (a) unmodified spectral tilt (n = 32); (b) with a positive 6 dB/octave spectral tilt (n = 32); or (c) with a negative 6 dB/octave spectral tilt (n = 32). The results showed that both 6- and 9-month-olds discriminated the three speech contrasts in the unmodified condition. However, when the contrasts were spectrally modified, a consistent developmental trend emerged. Six-month-olds’ discrimination performance improved when the spectral tilt modification amplified the relevant frequency information. That is, for the consonant contrasts (fricatives /f/-/s/ and approximants /l/-/r/) 6-month-olds performed best when high-frequency information was emphasised, whereas for the vowel contrast (/5/-/N/) the best discrimination performance was found when low-frequency information was enhanced. Nine-month-olds, on the other hand, showed no evidence of discriminating any of the contrasts when spectral tilt was modified. For all three contrasts, the older infants’ best discrimination performance was in the unmodified spectral tilt condition. The findings reflect the early stages of linguistic development. Six-month olds, whose acoustically driven speech perception operates in a language-general mode, demonstrated a broad-based ability to discriminate speech sounds. They not only accommodated speech with positive or negative spectral tilt, but a facilitation effect was observed when relevant frequency information was amplified. In contrast, 9-month-olds, whose perception is linguistically driven, found that spectral tilt modifications hinder speech perception. It seems that because older infants are attuning to native phonemic categories, their attention is constrained to native language spectral profiles only. Thus, even those spectral tilt modifications that should have been beneficial for perception were eschewed by 9-month-olds because they were incompatible with the older infants���� narrow focus on spectrally intact native speech sounds. Only NH infants were tested in this study, so it is not possible draw firm conclusions about amplification schemes until HI infants are tested too. Nevertheless, the results have a number of implications for infants with HL. The demonstration that spectral tilt modifications interfere with older NH infants’ speech perception suggests that future research should investigate whether HI infants also show this developmental pattern. That is, do HI infants progressively attune to the native language, and if so, does this affect their perception of spectrally modified speech sounds? Moreover, if HI infants are to attune to the native language and acquire speech and language in the same way as their NH peers, then the research reported here suggests that, throughout infancy, HI infants will need access to amplified speech that maintains the spectral shape of natural speech. Thus, the current research demonstrates, for the first time, a developmental difference in the way that infants perceive spectrally modified speech, one that is closely linked to the infant’s stage of native-language attunement. This research provides a solid foundation on which to conduct further research with HI infants and offers preliminary suggestions with regard to amplification and intervention to ensure that HI infants have the opportunity to emulate the native-language attunement process and subsequently achieve language development outcomes comparable with their NH peers. / Doctor of Philosophy (PhD)

Thesis (Ph.D.) -- University of Western Sydney, 2008. / A thesis submitted to the University of Western Sydney, College of Arts, MARCS Auditory Laboratories, in fulfilment of the requirements for the degree of Doctor of Philosophy. Includes bibliographical references.