Phonemic category development in a teenage Mandarin-Japanese 2L1 bilingual: A longitudinal study with monolingual and 2L1 adult bilingual comparisons

Voice Onset Time (VOT) and onset f0 are known correlates of voicing distinctions in stops and both contribute to the production and perception of voicing (House & Fairbanks, 1953; Abramson & Lisker, 1965; Ohde, 1984). As the values of VOT and onset f0, which correspond to voicing categories, vary cross-linguistically, a second language (L2) learner has to acquire a novel use of these acoustic cues to produce and perceive voicing in their L2. Although the acquisition of the primary voicing cue, VOT, has been studied extensively in L2 research (Flege & Eefting, 1988; Flege 1991; Birdsong et al. 2007), little is known about the acquisition of onset f0. The present study compares the use of VOT and onset f0 in French and English speech produced by American learners of French (23). The study also examines evidence for phonetic drift in L2 learners by comparing their English productions to a monolingual control group (33). Results indicate that although learners’ VOT values in French were heavily influenced by English, their onset f0 production in both English and French were on target, showing that learners are able to manipulate the two cues independently of one another. Little evidence of the effect of learners’ second language on the first language was found.This study also examines the role of individual learning history on the realization of VOT and onset f0, determining that average number of hours speaking French and age of L2 acquisition (AOA) reported by learners shows the strongest correlation with the learner’s acoustic productions.

In English, voiced and voiceless syllable-initial stop consonants differ in both fundamental frequency at the onset of voicing (onset F0) and voice onset time (VOT). Although both correlates, alone, can cue the voicing contrast, listeners weight VOT more heavily when both are available. Such differential weighting may arise from differences in the perceptual distance between voicing categories along the VOT versus onset F0 dimensions, or it may arise from a bias to pay more attention to VOT than to onset F0. The present experiment examines listeners' use of these two cues when classifying stimuli in which perceptual distance was artificially equated along the two dimensions. Listeners were also trained to categorize stimuli based on one cue at the expense of another. Equating perceptual distance eliminated the expected bias toward VOT before training, but successfully learning to base decisions more on VOT and less on onset F0 was easier than vice versa. Perceptual distance along both dimensions increased for both groups after training, but only VOT-trained listeners showed a decrease in Garner interference. Results lend qualified support to an attentional model of phonetic learning in which learning involves strategic redeployment of selective attention across integral acoustic cues.

Voice Onset Time (VOT) and onset f0 are known correlates of voicing distinctions in stops and both contribute to the perception of voicing (House & Fairbanks, 1953; Abramson & Lisker, 1965). The values of VOT and onset f0 which correspond to voicing categories vary cross-linguistically. Second language (L2) learners often have to acquire a novel use of these acoustic cues to produce and perceive L2 voicing. The acquisition of primary voicing cue, VOT, has been studied extensively in L2 research (Flege & Eefting, 1988; Flege 1991; Birdsong et al. 2007) but less is known about the acquisition of secondary cues. The present study compares the use of VOT and onset f0 in French and English speech produced by American learners of French (22). The study also examines the role of back transfer in L2 learners by comparing their English productions to a monolingual control group (33). The results demonstrate that although learners’ VOT values in French were heavily influenced by English, their onset f0 production in both English and French was on target. Little evidence of learners’ second language affecting their first language was found. Individual trends, including the effect of L2 proficiency level will also be explored.

As many of its geographical and typological neighbors, Marathi demonstrates a four-way laryngeal contrast in stop consonants, which includes plain voiced and voiceless stops as well as aspirated voiceless stops and typologically rare aspirated (breathy) voiced stops. The present study investigates the acoustic correlates of the laryngeal contrast in word-initial velar stops in Marathi. Thirty three native speakers of Marathi were recorded in Mumbai, Maharashtra, India. Positive and negative Voice Onset Time (VOT), fundamental frequency at the onset of voicing (onset f0), and spectral intensity as a measure of breathiness during the following vowel were investigated as correlates of laryngeal distinctions. Preliminary results show that all four stop categories are well-distinguished via VOT and onset f0 measures. Initial spectral intensity measurements suggest that breathiness in Marathi stops is largely contained in the release and does not extend far into the following vowel, unlike what has been found in Hindi (Dutta, 2007). The results will be discussed with respect to the typology of voicing contrast across languages, specifically the relevance of VOT distinctions and the use of onset f0, and compared to relevant findings for typologically similar languages.

Phonological status, not voice onset time, determines the acoustic realization of onset f0 as a secondary voicing cue in Spanish and English

In English, fundamental frequency at the onset of voicing (onset f0) co-varies with the Voice Onset Time (VOT) of initial stops and provides an additional perceptual cue to the phonetic feature of voicing, especially when VOT is ambiguous. However, aerodynamic and physiological explanations of the onset f0/VOT relationship suggest that onset f0 should correlate with voicing only in languages such as English that contrast short lag (voiceless unaspirated) with long lag (voiceless aspirated) consonants, and not in languages such as Spanish that contrast prevoiced with short lag stops. Previous perceptual research supports this prediction: Spanish speakers with little English experience do not incorporate onset f0 in making voicing decisions, suggesting lack of a correlation in their ambient language. In contrast, Spanish speakers with extensive experience with English show an English-like pattern of onset f0 use, suggesting that exposure to the English pattern of covariation has influenced their perceptual weighting of these two cues. The present study compares the distribution and correspondence between VOT and onset f0 in syllable-initial bilabial stops ([b] - [p]) in Spanish and English. Implications for the typology of voicing contrasts and perceptual strategies for sound categorization in non-native language environments are discussed.

Onset pitch perturbations and the cross-linguistic implementation of voicing: Evidence from tonal and non-tonal languages

Word-initial stops in Mandarin and English show a distinctive phonological categorization but a similar phonetic realization along the VOT (Voice Onset Time) continuum. Previous research reported that native Mandarin adults produce measurably longer long-lag VOTs than native English adults. The present study examined whether and how the difference between Mandarin and English VOTs is manifested in monolingual children and Mandarin–English bilingual children. The participants included 15 five- to six-year-old sequential bilingual children, 24 corresponding monolingual children (15 Mandarin, 9 English), and 22 monolingual adults (12 Mandarin, 10 English). The bilingual children were divided into two groups (Bi-low and Bi-high) based on the amount of experience in English. Each participant was recorded producing 18 Mandarin words and/or 18 English words containing six stops in each language. The VOT values were measured from the beginning of stop burst to the onset of the voicing. The results showed that the language difference in VOT in the monolingual children was manifested in a pattern similar to the monolingual adults. However, Mandarin and English VOTs showed less separable distributions in the two groups of bilingual children. Further analysis suggested that both groups of bilingual children tended to separate Mandarin and English short-lag VOTs but only the Bi-low children showed different long-lag VOTs between the two languages. These results suggested that due to the bilingual effects and L1–L2 (first language – second language) interactions, even though the bilingual children tried to separate the two VOT systems, they implemented the separation in a different manner than the monolingual speakers.

Brainstem and early cortical auditory activities associated with language differences in acoustic cue weighting for voicing perception

The present study compares voice onset time (VOT) productions across age groups in Hindi and in English, which have four-way and two-way phonemic voicing contrasts, respectively. Both voicing lead times and post-release VOTs were taken from initial velar stops in isolated monosyllabic words. Data were elicited from monolingual adults and children (ages 2–6 years). Vowel quality effects were controlled. A two-way analysis of variance showed that for all age groups in both languages, phonemic voice category had a significant effect on VOT. Pairwise comparisons on the Hindi data, however, revealed that, although all pair had significant mean post-release VOT differences in the adult group, significant /k/ vs /g/ differences were not noted until 6 years of age. The /k/ vs /kh/ pair did exhibit significantly different mean VOTs from the age of 2, suggesting that when wider mean VOT differences are noted in the adult phonemes, the children are more likely to acquire the contrast at an early age. The more complex /gh/ was not adultlike, even at 6 years. The English subjects demonstrated significant differences at all ages, perhaps because the English mean VOT differences are at least as wide as those in the Hindi /k/ vs /kh/ contrast. In both languages, lead timings were not adultlike in the younger groups, which supports the claim that lead voicing is more difficult to acquire than either short- or long-lag timing. Repeated measures analyses showed significant interactions for VOT between age group and the target phonemic category, demonstrating a steady progression toward the adult model. The implications of these results for a universal phonetic category system are discussed.

Consonantal effects on onset f0 are implemented differently in different languages, namely, the duration of consonantal effect, the direction of f0 change and its perceptual importance vary cross-linguistically. This study aims to investigate consonantal effects of aspiration in a tone language, Cantonese. The results showed that aspiration had a raising effect on onset f0, that is, onset f0 after aspirated stops was higher than after unaspirated stops. Besides, the aspiration-related f0 perturbations can extend to around 100ms after voicing. However, unlike f0 as a secondary cue for stop contrasts in English, when voice onset time (VOT) becomes ambiguous, f0 was not strengthened for contrast enhancement in Cantonese as well as in L2 English. This indicates that although consonantal effects in Cantonese showed phonetically similar directions and comparable duration with native English, onset f0 was not used for phonological contrast enhancement.

This study examined Korean three-way stop contrasts produced by North Korean (NK) immigrant speakers now living in South Korea, from the perspective of second dialect acquisition (SDA). Their production was compared with that of South Korean (SK) speakers. SK speakers and NK refugee speakers ( N = 22 each) completed three tasks designed to elicit careful and conversational speech: reading aloud lists of one-syllable words and short phrases and participating in a sociolinguistic interview. The potential acoustic cues voice onset time (VOT), F0, and H1-H2 were measured and analyzed for 14,478 stops. Results indicated that in conversational speech (interview), SK speakers neutralized VOT between lenis and aspirated and used F0 robustly to differentiate the two categories, while distinguishing the long VOT for lenis and aspirated from the short VOT for fortis stops. In careful speech (reading one-syllable words), SK speakers differentiated all three categories by VOT. In contrast, NK speakers distinguished all three categories by VOT in all tasks, except for neutralizing the contrast for fortis and lenis in phrase reading. Furthermore, F0 was not used as robustly by NK as by SK speakers. We also examined the effects of age of arrival (AoA) and length of residence (LoR) on NK speakers’ SDA. Our results indicated that the longer the NK refugees lived in SK, the more they could produce more SK-like stops. The results suggest that NK stop contrasts are likely distinguished by VOT, and these refugee speakers are in the process of acquiring SK stop patterns.

Cross-language differences in cue weighting are often attributed to differences in phonological inventories, in that listeners are assumed to give greater weight to those acoustic cues that best differentiate the specific phonological categories of their native language. However, subphonological properties of a given contrast may also affect weighting of acoustic cues for acoustic or perceptual reasons, in a manner independent of the structure of the listeners’ phonological inventory. Here, this possibility is investigated by comparing Spanish and English listeners’ relative weighting of two acoustic cues to the syllable-initial stop consonant voicing contrast /b/-/p/: Voice onset time (VOT) and fundamental frequency at the onset of voicing (onset f0). Spanish and English possess comparable word-initial stop consonant inventories, but differ in the phonetic realization of the voicing contrast in terms of VOT: Spanish contrasts a short lag (<10 ms) /p/ with a prevoiced (<0 ms) /b/, while English contrasts a long lag (> 50 ms) /p/ with a short lag (<20 ms) /b/. Preliminary results (N = 7) suggest that, although speakers from both languages depend mainly on VOT, Spanish speakers give more weight to onset f0 when VOT is ambiguous than do English listeners.

Many listeners report hearing difficulty, especially for speech in noisy environments, despite having normal audiometric thresholds. Recent work suggests that such cases may be caused by disruptions in coding of suprathreshold sounds at early stages of auditory processing. Specifically, differences in the function of auditory nerve fibers with lower spontaneous firing rates could disrupt coding of intensity changes for sounds in the typical range used for speech communication. In turn, this could affect perception of acoustic cues in speech that are dependent on intensity changes over time, such as voice onset time (VOT). The current study investigated a speech sound manipulation designed to counteract the effects of such disruptions by making intensity differences in the signal more pronounced. Listeners heard speech sounds varying along two acoustic dimensions that provide information about word-initial voicing, VOT and f0 onset, and categorized sounds as either voiced or voiceless. Results showed that the intensity manipulation reduced listeners' use of f0 for voicing categorization and also affected their use of VOT. This suggests that intensity manipulations can make acoustic cues like VOT more salient, providing a potential treatment approach for listeners who have difficulty coding intensity differences in speech.

The contributions of voice onset time (VOT) and fundamental frequency (F0) were evaluated for the perception of voicing in syllable-initial stop consonants in words that were low-pass filtered and/or masked by speech-shaped noise. It was expected that listeners would rely less on VOT and more on F0 in these degraded conditions. Twenty young listeners with normal hearing identified modified natural speech tokens that varied by VOT and F0 in several conditions of low-pass filtering and masking noise. Stimuli included /b/-/p/ and /d/-/t/ continua that were presented in separate blocks. Identification results were modeled using mixed-effects logistic regression. When speech was filtered and/or masked by noise, listeners' voicing perceptions were driven less by VOT and more by F0. Speech-shaped masking noise exerted greater effects on the /b/-/p/ contrast, while low-pass filtering exerted greater effects on the /d/-/t/ contrast, consistent with the acoustics of these contrasts. Listeners can adjust their use of acoustic-phonetic cues in a dynamic way that is appropriate for challenging listening conditions; cues that are less influential in ideal conditions can gain priority in challenging conditions.