Cognates interfere with language selection but enhance monitoring in connected speech

Abstract

The current study investigated the contribution of phonology to bilingual language control in connected speech. Speech production was elicited by asking Mandarin-English bilinguals to read aloud paragraphs either in Chinese or English, while six words were switched to the other language in each paragraph. The switch words were either cognates or noncognates, and switching difficulty was measured by production of cross-language intrusion errors on the switch words (e.g., mistakenly saying (qiao3-ke4-li4) instead of chocolate). All the bilinguals were Mandarin-dominant, but produced more intrusion errors when target words were written in Chinese than when written in English (i.e., they exhibited robust reversed dominance effects). Most critically, bilinguals produced significantly more intrusions on Chinese cognates, but also detected and self-corrected these same errors more quickly than with noncognates. Phonological overlap boosts dual-language activation thus leading to greater competition between languages, and increased response conflict, thereby increasing production of intrusions but also facilitating error detection during speech monitoring.

What reading aloud reveals about speaking: Regressive saccades implicate a failure to monitor, not inattention, in the prevalence of intrusion errors on function words

“…Both these findings reduce the likelihood that inattention during reading leads to intrusion errors in the read-aloud task. Previously, we suggested that reversed dominance and default language effects reflect global inhibition of the dominant language to allow language mixing (Gollan & Goldrick, 2018;Gollan, Schotter, et al, 2014;Ratiu & Azuma, 2017); the replication of this result in a different population of bilinguals implies broad applicability of this bilingual control mechanism (see also Li & Gollan, 2018). The persistent susceptibility of function words to intrusion errors is informative because in English, Spanish, and French, function words tend to be shorter than content words.…”

“…Reversed dominance effects. In previous studies (Gollan & Goldrick, 2016Gollan, Schotter, et al, 2014;Gollan et al, 2017;Li & Gollan, 2018;Ratiu & Azuma, 2017), bilinguals more often mistakenly replaced a word written in the dominant language with its nondominant languagetranslation equivalent than the opposite, an effect we refer to as reversed dominance. We replicated this pattern; Chinesedominant bilinguals were more likely to produce English intrusions, saying the English translation instead of the written Chinese word, than vice versa.…”

Language selection contributes to intrusion errors in speaking: Evidence from picture naming

“…By manipulating the language context, the paradigms successfully simulated daily-life scenarios where language intrusion is more likely to occur. Compared to other tasks such as the reading aloud of texts (Gollan & Goldrick, 2018;Gollan et al, 2014;Li & Gollan, 2018;Schotter et al, 2019), the current paradigm is better suited to investigate failures to 'stay' rather than failures to switch. However, in the current study, bilingual participants were still asked to use both of their languages in quick succession, which makes the repeat trials still intrinsically different from the 'staying in the same language' situation in daily life.…”

“…It is worth noting that language intrusions have also been investigated extensively using a reading aloud task, where participants are asked to read aloud mixed-language paragraphs (Gollan & Goldrick, 2018;Gollan, Schotter, Gomez, Murillo & Rayner, 2014;Li & Gollan, 2018;Schotter, Li & Gollan, 2019). However, the fact that people can read aloud non-existing words suggests that reading aloud does not necessarily involve concept and lemma selection.…”

“…As discussed so far, language intrusion takes place in daily life – though not very frequently (Muysken, 2000; Poulisse, 1999) – as well as in laboratory experiments, such as in the picture-word interference task (Boukadi et al, 2015; Hermans et al, 1998; Klaus et al, 2018), the cued language-switching task (Meuter & Allport, 1999; Zheng et al, 2018a; Zheng et al, 2018b), and the reading aloud task (Gollan & Goldrick, 2018; Gollan et al, 2014; Li & Gollan, 2018; Schotter et al, 2019). Studying why intrusion errors happen can help us better understand how bilinguals exert control over the bilingual word production system.…”