Skip to main content
Open Access Publications from the University of California

UC Berkeley

UC Berkeley Electronic Theses and Dissertations bannerUC Berkeley

Costs and Cues in the Auditory Comprehension of Code-switching


Bilinguals alternate frequently between languages, but many psycholinguistic studies on code-switching have reported a “switch cost”, i.e. an increased processing difficulty, in production (Meuter & Allport, 1999; Thomas & Allport, 2000; Costa & Santesteban, 2004; Gollan & Ferreira, 2009, although see Kleinman & Gollan, 2016), recognition (Soares & Grosjean, 1984), and comprehension (Olson, 2017). This dissertation involves three experiments investigating the factors modulating switch cost in the auditory comprehension of Mandarin and English code-switched words. First, recent research suggests that subtle phonetic differences between the pronunciation of code-switched utterances and unilingual utterances might act as anticipatory cues to code-switches for listeners (Piccinini & Garellek, 2014; Fricke, Kroll & Dussias, 2016), which could mitigate switch cost. Second, an “asymmetric switch cost,” or higher switch cost for the dominant first language (L1) compared to the second language (L2), has been reported for auditory comprehension of Spanish-English code-switches (Olson, 2017). Additionally, Mandarin-English bilinguals judge switches from English-to-Mandarin as infrequent compared to Mandarin-to-English switches (Lu, 1991; Ong & Zhang, 2010). Thus, Mandarin-English switching could be subject to a cost asymmetry driven not just by dominance but by frequency.

Experiments 1 and 2 test the effects of withholding anticipatory phonetic cues on code-switched recognition by splicing English-to-Mandarin code-switches into unilingual English sentence contexts. Experiment 1 measured Mandarin-English bilinguals’ (N=42) reaction times in a concept monitoring task where they had to press a button when they heard a pictured object mentioned in an auditorily presented English sentence. The target word was either code-switched (i.e., in Mandarin) or unswitched. RTs were slower when the target was a code-switch, suggesting a switch cost. Experiment 2 tracked Mandarin-English bilinguals’ (N=41) eye movements during a task in which they were asked to fixate on the pictured object in a display that matched a code-switched (Mandarin) or unswitched target word in an auditorily presented English sentence. The average proportion of all participants' looks to target pictures corresponding to sentence-medial code-switches decreased when cues were withheld, suggesting that withholding anticipatory phonetic cues can negatively affect code-switched recognition. Therefore, in normal conditions, bilingual listeners use phonetic cues to anticipate an upcoming code-switch. Acoustic analysis of stimuli from Experiments 1 and 2 showed tone-specific anticipatory pitch coarticulation prior to code-switches, which might contribute to phonetic cuing.

Experiment 3 tests whether Mandarin-English code-switching might incur an asymmetric switch cost due to differences in dominant language and frequency, by comparing looks to English-to-Mandarin switches and Mandarin-to-English switches in an eye tracking study. Mandarin-English bilingual listeners (N=48) of varying language dominance scores (Birdsong, Gertken & Amengual, 2012) participated in an eye tracking task in which they were auditorily presented Mandarin and English sentences with and without code-switched target words. Compared to unswitched targets, the average proportion of all participants’ looks to target pictures corresponding to English code-switches was higher, but there were fewer looks to Mandarin code-switches. This suggests an asymmetric switch cost where English code-switches in Mandarin sentences are easy to process but Mandarin code-switches in English sentences are more difficult in terms of processing time. Mandarin-dominant bilinguals looked more toward targets corresponding to Mandarin words, suggesting a language dominance effect that is unexpected under influential models of code-switching.

These studies suggest that processing code-switches need not be costly: processing code-switched Mandarin words incurs a cost, which can be modulated by the presence of anticipatory phonetic cues depending on the speaker, but processing code-switched English words does not. Dominant language and the frequency of code-switch by language are influential factors in Mandarin-English bilingual auditory comprehension. Implications of these findings for models of code-switching and bilingual language control are discussed, along with possible mechanisms underlying the phonetics of code-switched speech.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View