Page E. Piccinini

Voice onset time in Spanish-English spontaneous code-switching

Research on the phonetics of code-switching has focused on voice onset time (VOT) and has yielded mixed results regarding cross-language interaction, possibly due to differences in data used (scripted vs. spontaneous speech) and populations examined (L1 vs. L2 dominant, early vs. late bilinguals). Here VOT was measured in a corpus of spontaneous code-switching speech elicited from a homogeneous group of early bilinguals in conversation with and without distraction (completion of jigsaw puzzles). The distraction meant to increase cognitive load, a manipulation that could affect phonetic realization. Both English and Spanish VOT were shorter at code-switching points than in comparable monolingual utterances. English VOT lengthened overall under increased cognitive load (but remained shorter in code-switching as compared to the monolingual context). These results support previous findings of VOT shortening in code-switching for both English and Spanish, and confirm that the effect applies in the natural speech of early bilinguals.

The vowel spaces of Southern Californian English and Mexican Spanish as produced by monolinguals and bilinguals

The vowel spaces of Southern Californian English and Mexican Spanish were investigated using three groups of speakers: 11 English monolinguals (9 female), 10 Spanish monolinguals (7 female), and 11 Spanish-English bilinguals (8 female). Speakers produced six repetitions of the ten American English vowels [i, ɪ, ɛ, æ, ɑ, ɔ, ʊ, u, Λ, and ɜr] and six repetitions of the five Spanish vowels [i, e, a, u, and o]. Monolinguals produced vowels in one language; bilinguals produced vowels in both languages. Preliminary analysis shows for females Southern Californian English back vowels were less fronted compared to the results of Hagiwara (1997) from Southern Californian English, but more fronted than those of Hillenbrand et al. (1995) on General American English. Mexican Spanish back vowels [u] and [o] were substantially backed compared to Castilian Spanish vowels (Bradlow, 1995), while [i] was lower and more fronted. In general, Mexican Spanish vowels were produced more backed than Southern Californian English vowels in monolingual productions. Bilinguals produced their two vowel spaces with Spanish more backed and lower as compared to English.

The production of Spanish–English code-switching

It is generally assumed that in code-switching (CS) switches between two languages are categorical, however, recent research suggests that the phonologies involved in CS are merged and bilinguals must actively suppress one language when encoding in the other. Thus, it was hypothesized that CS does not take place abruptly but that cues before the point of language change are also present. This hypothesis is tested with a corpus of Spanish-English CS examining word-initial voiceless stop VOT and the vowel in the discourse marker “like.” English VOTs at CS boundaries were shorter, or more “Spanish-like,” than in monolingual utterances. Preliminary results suggest Spanish VOTs at CS boundaries were shorter than in monolingual utterances, thus even more Spanish-like than monolingual Spanish utterances. The vowel of “like” in English utterances was more monophthongal and had a lower final F2 as compared to “like” in Spanish utterances. At CS boundaries, “like” began similarly to the language preceding the token...

Perceptual warping of phonetic space applies beyond known phonetic categories: Evidence from the perceptual magnet effect

What is the mental representation of phonetic space? Perceptual reorganization in infancy yields a reconfigured space “warped” around native-language (L1) categories. Is this reconfiguration entirely specific to L1 category inventory? Or does it apply to a broader range of category distinctions that are non-native, yet discriminable due to being defined by phonetic dimensions informative in the listener’s L1 (Bohn & Best, 2012; Pajak, 2012)? Here we address this question by studying perceptual magnets, which involve attrition of within-category distinctions and enhancement of distinctions across category boundaries (Kuhl, 1991). We focus on segmental length, known to yield L1-specific perceptual magnets: e.g., L1-Finnish listeners have one for [t]/[tt], but L1-Dutch listeners, who lack (exclusively) length-based contrasts, do not (Herren & Schouten, 2008). We tested 31 L1-Korean listeners in an AX discrimination task for [n]-[nn] and [f]-[ff] continua. Korean listeners have been shown to discriminate both...

Cross language speech-in-noise perception by early Spanish-English bilinguals and English monolinguals

Bilinguals have shown a hyper-awareness of fine phonetic detail in speech, while also sometimes losing out on higher-level syntactic and semantic information in speech-in-noise studies. This study seeks to determine how bilinguals process speech in noisy environments across different language contexts. Specifically, this study tests whether bilinguals utilize certain phonetic cues to access higher-level information. Two experiments will be conducted. First, to determine how bilinguals process speech in noise in different language contexts, early Spanish-English bilinguals and English monolinguals learning Spanish listened to sentences mixed with white noise in English, Spanish, and code-switching (English to Spanish and Spanish to English) contexts. Preliminary results suggest early Spanish-English bilinguals perform significantly above chance on word identification in all contexts, performing best in the Spanish context. The second experiment will determine specifically which noise types (lower versus hi...

Predicting code-switches from phonetic information: The discourse marker like in Spanish-English code-switching

The present study investigated whether Spanish-English bilinguals (L1 Spanish, English dominant) use phonetic cues to anticipate code-switches. Listeners were presented with four sets of 10 utterances. In a given set all utterances began in English or Spanish. All utterances included the discourse marker like. In each set, half of the utterances continued in the same language after like and half switched languages after like. Listeners only heard up to and including like. Listeners evenly sorted the utterances into two columns, “continues in English” or “continues in Spanish,” to indicate which five utterances involved code-switching. Half of listeners received instructions in English and half in Spanish. Both sets of listeners were significantly above chance for stimuli beginning in English [p < 0.05]. Listeners who received Spanish instructions were also trending above chance for stimuli beginning in Spanish [p = 0.08]. This suggests listeners can use phonetic cues to anticipate switches from their dominant to their non-dominant language. Additionally, when language mode is the non-dominant language, listeners can also anticipate switches from their non-dominant to their dominant language. These results support a theory where both languages are somewhat activated at all times, allowing bilinguals to use phonetic cues to anticipate language switches.

Semantic contextual cues and listener adaptation to foreign‐accented English.

The context surrounding words can facilitate word recognition for native and non‐native listeners. However, for degraded speech signals, (e.g., with noise), non‐natives may lose the ability to take advantage of contextual cues unless the speech is clearly produced [Bradlow and Alexander (2007)]. This study investigated whether: (1) foreign‐accented speech degrades the signal such that natives show reduced ability to take advantage of context, and (2) both native and non‐native listeners adapt to foreign‐accented speech. Native and non‐native listeners were exposed to accented speech in two blocks. In each block, half the sentence final words were in high and half were in low predictability contexts. Listeners were asked to identify sentence final words. In a single talker condition, natives benefited from context and adapted to the accent; non‐natives could not use context but showed adaption. In a multitalker condition, natives neither benefited from context nor adapted to the accent. Non‐natives benefit...