Effects of Language History on Sentence Recognition in Noise or Two-Talker Speech: Monolingual, Early Bilingual, and Late Bilingual Speakers of English

Abstract

Purpose Language history is an important factor in masked speech recognition. Listeners who acquire the target language later in life perform more poorly than native speakers. However, there are inconsistencies in the literature regarding performance of bilingual speakers who begin learning the target language early in life. The purpose of this experiment was to evaluate speech-in-noise and speech-in-speech recognition for highly proficient early bilingual listeners compared to monolingual and late bilingual listeners. Method Three groups of young adults participated: native monolingual English speakers, bilingual Mandarin–English speakers who learned English from birth (early bilinguals), and native Mandarin speakers who learned English later in life (late bilinguals). All participants had normal hearing and were full-time college students. Recognition was assessed for English sentences in speech-shaped noise and two-talker English speech. Participants provided linguistic and demographic information, and late bilinguals completed the Versant test of spoken English abilities. Results All listeners performed better in speech-shaped noise than two-talker speech. Performance was similar for monolingual and early bilinguals. Late bilinguals performed more poorly overall. There was evidence for a stronger association between masked speech recognition and English dominance for late bilinguals compared to early bilinguals. Conclusion These results support the conclusion that bilingualism itself does not necessarily result in a disadvantage when recognizing masked speech in noise and speech in speech. For populations similar to those studied here (highly proficient early bilinguals), it would be appropriate to evaluate masked speech recognition using the same simple stimuli and normative data used for monolingual speakers of English.

Navigating the bilingual cocktail party: Interference from background speakers in listeners with varying L1/L2 proficiency

“…Several studies have investigated bilinguals' performance in competing speech tasks, usually in comparison to monolingual participants when the target is presented in a single language (Van Engen, 2010;Calandruccio & Zhou, 2014;Regalado, Kong, Buss & Calandruccio, 2019). Van Engen (2010) and Calandruccio and Zhou (2014) found that simultaneous bilingual participants performed better when the masker language differed from the target language, consistent with the pattern observed for monolingual participants.…”

“…Van Engen (2010) and Calandruccio and Zhou (2014) found that simultaneous bilingual participants performed better when the masker language differed from the target language, consistent with the pattern observed for monolingual participants. On the other hand, Regalado et al (2019) found that monolingual, simultaneous bilingual, and late bilingual (L1 Mandarin, L2 English) participants performed better with noise maskers as opposed to speech maskers (all targets and speech maskers were presented in English) but they did not investigate the role of masker language.…”

“…However, to compare monolingual and bilingual samples, the target voice must use the monolingual participants' L1, which is typically English. Unless the bilinguals recruited are simultaneous bilinguals, English is usually the L2 of the bilingual participants, as exemplified in Regalado et al (2019), in which English was the L2 of late bilingual participants. Differing levels of proficiency relative to the target language both reduces the validity of the comparison between bilinguals and monolinguals and makes it difficult to determine how bilinguals might perform on a competing-speech task using their L1 as a target.…”

“…Differing levels of proficiency relative to the target language both reduces the validity of the comparison between bilinguals and monolinguals and makes it difficult to determine how bilinguals might perform on a competing-speech task using their L1 as a target. Regalado et al (2019) also noted worse performance in the late bilingual group under both masking conditions, but they determined that this was due to differences in language dominance rather than bilingual disadvantage, further illustrating the inevitable differences in performance that arise when L1 differs between groups.…”

“…One last aspect of the current findings that we should discuss in depth relates to the nature of variables that underlie the bilingual experience. Many papers published on the topic of bilingualism treat it as a categorical variable, dividing monolinguals from bilinguals (Bialystok, Craik, Klein & Viswanathan, 2004;Rogers, Lister, Febo, Besing & Abrams, 2006;Filippi, Leech, Thomas, Green & Dick, 2012;Krizman et al, 2017;and Regalado et al, 2019, to name a few). However, several research groups have argued against dichotomizing continuous variables of bilingualism (Luk, 2015;de Bruin, 2019;DeLuca et al, 2019;Kremin & Byers-Heinlein, 2021).…”