Secondary cues to the phonological voicing distinction: perception by bilingual speakers

The phonological distinction between /p, t, k/ and /b, d, g/ is implemented in different ways across languages. For example, whereas both Spanish and English have a contrast between two series of plosives, in Spanish, in utterance-initial position, this contrast is implemented as short VOT versus prevoicing, while in English the contrast is between aspirated (long VOT) versus unaspirated stops (Lisker & Abramson, 1964). Consequently, Spanish short lag /p/ (peso ‘weight’) may sound like /b/ (beso ‘kiss’) to English listeners, leading to potential difficulties in L2 acquisition. Onset f0 is an important secondary cue that may enhance the voiced-voiceless plosive contrast, and though it tends to physiologically co-vary with VOT, it has also been argued to be phonologized in at least some languages (e.g., Kingston, 2007; Dmitrieva, Llanos, Shultz, & Francis, 2015). Listeners are responsive to both phonologically relevant phonetic details and to non-contrastive variation, and bilingual speakers must learn to attend to the relevant acoustic properties in both languages. For both highly proficient speakers with distinct mental sound representations, and for less proficient speakers still learning to navigate their phonological systems, it is an empirical question to what extent learners encode acoustic details, in particular, the relationship between VOT and onset f0 for the word-initial voicing distinction. If onset f0 is phonologized, it is expected to influence categorization according to the realization of the voicing distinction in each respective language. Taking English and Spanish as an example, consonants within the short lag range represent two distinct phonological categories: voiced in English and voiceless in Spanish. A naïve English listener may therefore expect to hear low onset f0, whereas a naïve Spanish listener may expect to hear high onset f0. When presented with conflicting cues, a listener may be expected to resolve that conflict depending on the language they are presented with. This thesis examines the cue-weighting strategies utilized in three different languages: English, Spanish, and Mandarin Chinese, by L1-Spanish/L2-English, L1-English/L2-Spanish, and L1-Mandarin/L2-English speakers, through two-alternative forced choice tasks with a series of stops co-varying in VOT (4 steps) and onset f0 (4 steps). Results showed that listeners used different cue-weighting strategies depending on the target language presented, with all groups of speakers showing significant effects of onset f0 in categorization of tokens in English, but not in Spanish or Mandarin; in line with a proposal by Llanos et al. (2013), onset f0 may play a larger role in voicing decisions in English (relative to Spanish). As listeners were not similarly affected by onset f0 across languages, these results tentatively support the general hypothesis that onset f0 is phonologized and that perceptual cue-weighting of onset f0 as a cue to the voicing contrast may be attenuated in languages where f0 cues other important phonological contrasts, such as lexical stress and tone.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Linguistics

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2023

Author dc:creator

• Zhang, Jennifer Qian

Contributors dc:contributor

• Hualde, José I
• Shih, Chilin
• Shosted, Ryan
• Holliday, Jeffrey

• Cue-weighting
• Bilingualism
• Second Language
• Onset F0
• Vot
• Perception

Statement dc:rights

• Copyright 2023 Jennifer Qian Zhang

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/120282

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F120282