Non-selective lexical access in bilinguals is spontaneous and independent of input monitoring: Evidence from eye tracking
Introduction
It is estimated that over half of the world’s population speaks multiple languages (Grosjean, 2010). Despite the common intuition that bilingual speakers can function independently in their two languages, research has shown that they cannot completely “turn off” one language while using the other. For example, when Dutch–English bilinguals read in English, words that share form and meaning in the two languages (i.e., cognates such as ‘ring’) are recognized faster and more accurately as compared to noncognate words (Duyck, Assche, Drieghe, & Hartsuiker, 2007). This cognate facilitation effect suggests that lexical-semantic information in the non-target language is activated and affects the processing of the target language (Lemhofer and Dijkstra, 2004, Midgley et al., 2011, Van Assche et al., 2009). However, the origin of the cognate facilitation effect is obscured by the fact that cognates inherently have higher lexical frequency and have overlapping orthographic and phonological features between languages, which all potentially contribute to task performance. Nevertheless, other studies have shown that parallel language activation also characterizes word processing in bilinguals whose native and second languages have radically different lexical forms (e.g., Chinese and English; Moon and Jiang, 2012, Thierry and Wu, 2007, Zhang et al., 2011). Furthermore, this effect is not restricted to bilingual reading, but has also been found in listening (Lagrou et al., 2011, Spivey and Marian, 1999, Thierry and Wu, 2007, Wu and Thierry, 2010), speaking (Colome, 2001, Guo and Peng, 2006, Hoshino and Thierry, 2011, Kroll et al., 2008), and even sign language (Morford et al., 2010, Shook and Marian, 2012).
The finding that language processing in bilinguals is non-selective (but see FitzPatrick and Indefrey, 2010, Rodriguez-Fornells et al., 2002 for contrasting views) suggests that mental representations of the two languages are integrated (Dijkstra and Van Heuven, 2002, Green, 1998), a view that has been complemented by neuroimaging evidence showing cortical overlap for processing native and second languages (see van Heuven & Dijkstra, 2010, for a review). However, the demonstration of language non-selective lexical access in most previous studies has relied on explicit language tasks such as lexical decision, semantic relatedness judgment, or categorization. Hardly any research has investigated this effect in a non-linguistic context, when words are processed in an involuntary manner. As a result, it is unknown to what extent non-selective lexical access is dependent upon deliberate language processing. To fill this knowledge gap, in a recent study, we used event-related potentials (ERPs) to examine whether bilingual readers activate translations in the native language while they incidentally process words presented in their second language (Wu & Thierry, 2012). ERPs are an averaged brain signal recorded from the surface of the scalp and time-locked to a stimulus (visual, auditory, etc.) or response (button press, speech onset, etc.) of interest. The high temporal resolution (i.e., in the order of one millisecond) of ERPs affords fundamental insights into the chronometry of cognitive functions such as language processing, which is inherently fast. In the study by Wu and Thierry (2012), Chinese–English bilinguals were shown either a visual shape (i.e., a string of circles or squares) or a word of English in separate trials. They were instructed to press a given button when they saw circles, another button when the saw squares, and to withhold their response when they saw a word. In the critical condition, the Chinese translation of the to-be-ignored English word shared a phonological segment with the Chinese word for ‘circle’ or ‘square’. Although this manipulation did not affect behavioral performance, ERPs showed that an index of cognitive inhibition (the N200) was significantly increased in amplitude for critical words as compared to control words, suggesting greater involvement of the executive system. Since the study did not require participants to read the words or to engage in metalinguistic processing, the study demonstrated activation of native language translations elicited by incidental processing of words in the second language.
One practical limitation of ERPs is that eye movements generate artifacts that interfere with the measurement of relevant brain activity. To control such artifacts, in Wu's (2012) study, all stimuli were presented at the center of the screen. Thus, in the incidental word presentation trials, participants had to fixate the stimulus for a period of 500 ms, a process which may have triggered in-depth word processing despite the fact that it was to be ignored. In the present study, we overcome this limitation by using eye-tracking in an adaptation of the previous paradigm. Eye-tracking offers the full flexibility of measuring ocular responses in the context of visual displays in which task-irrelevant words can be presented amongst distracters. Here, each trial involved the presentation of a 4-item stimulus grid, scattered far enough from one another as to require eye movements during visual search. In filler trials, three of the stimuli were English words and the fourth was a target, made of circles or squares (i.e., a nonverbal visual target, Fig. 1). In the test trials, all four stimuli were words of English, one of which featured a phonological overlap with the Chinese word for ‘circle’ or ‘square’ when translated in Chinese (i.e., a target-related word). Participants were instructed to press one of two keys, depending on the shape, to terminate filler trials, and to press a third key to terminate test trials where no shape was present. Since participants had control over trial termination, they were expected to look at an English word only for the time necessary to distinguish it from target shapes when scanning through the grid. The task, therefore, encompassed nothing but a simple, unbiased visual search process. Assuming that language non-selective lexical access is independent of explicit language processing effort, access to the Chinese translation of English words would interfere with the nonverbal task when translations are connected to the visual shape targets (e.g., causing longer fixation duration and more fixations; Spivey & Marian, 1999). The absence of a difference between critical and control words, together with previous literature, would suggest that language non-selective processing in bilinguals may be contingent upon explicit linguistic processing and/or individual input monitoring.
Section snippets
Participants
Twenty Chinese–English bilinguals (13 females; mean age = 22, SD = 1.9) and twenty native speakers of English (11 females; mean age = 21, SD = 2.4) gave written consent to take part in the experiment approved by the ethics committee of Bangor University. Participants were right-handed undergraduate or Masters students, and had normal or corrected-to-normal vision. All bilingual participants spoke mandarin Chinese as their native language and English was their only foreign language. They began to learn
Results
Behavioral data were analyzed using repeated measures ANOVA with trial type (experimental versus filler) as the within-subject factor and group (English controls versus Chinese–English bilinguals) as the between-subject factor. We found a significant main effect of trial type on reaction time, F (1, 38) = 36.389, p < .001. No effect of group, F (1, 38) = .873, p = .356, or interaction, F (1, 38) = 2.952, p = .094, was found. Follow-up analysis showed that both English, t (19) = 8.701, p < .001 and bilingual
Discussion
The present study used eye-tracking to examine automatic cross-language activation in bilinguals engaged in a non-linguistic, visual perceptual task. The experimental manipulation failed to elicit significant behavioral differences between groups, as was the case in several previous studies involving the same subtle cross-language manipulation (e.g., Thierry and Wu, 2007, Wu and Thierry, 2010). There was also no between-subject difference in the filler trials, suggesting that the two groups did
Conclusion
We examined incidental language processing in bilinguals using a visual perception task. Eye-tracking data shows that bilinguals activated translation equivalents in the native language, which were not presented in the experiment. Such implicit effect, found when the participant is fully in control of filtering the input, shows that language non-selective access in not conditional upon lexical information being actively processed for meaning (e.g., during a semantic task) but even when only
Acknowledgements
Y. J. W and F. C. conceived the experiment, collected the data, and analyzed the data. Y. J. W, G. T., C. L., and F. Cr. wrote the manuscript. Y. J. W is funded by the British Academy/Leverhulme (SG120227). G.T. is funded by the Economic and Social Research Council (RES-000-23-0095). Y. J. W and G. T. are funded by the European Research Council (ERC-StG-209704). C. L. and F. C. are funded by the Economic and Social Research Council/ Engineering and Physical Sciences Research Council
References (33)
Lexical activation in bilinguals’ speech production: Language-specific or language-independent?
Journal of Memory and Language
(2001)- et al.
The time course of visual word recognition as revealed by linear regression analysis of ERP data
NeuroImage
(2006) - et al.
Effects of word length and frequency on the human event-related potential
Clinical Neurophysiology
(2004) - et al.
Language selection in bilingual word production: electrophysiological evidence for cross-language competition
Brain Research
(2011) - et al.
Language selection in bilingual speech: Evidence for inhibitory processes
Acta Psychologica (Amsterdam)
(2008) - et al.
Bilingual lexical activation in sentence context
Journal of Memory and Language
(2006) - et al.
Bimodal bilinguals co-activate both languages during spoken comprehension
Cognition
(2012) - et al.
Language comprehension in the bilingual brain: fMRI and ERP support for psycholinguistic models
Brain Research Review
(2010) - et al.
Language non-selective lexical access in bilingual toddlers
Journal of Experimental Child Psychology
(2012) - et al.
Unconscious translation during incidental foreign language processing
NeuroImage
(2012)
Cross-linguistic activation in bilingual sentence processing: The role of word class meaning
Bilingualism: Language and Cognition
Moving beyond Kucera and Francis: A critical evaluation of current word frequency norms and the introduction of a new and improved word frequency measure for American English
Behavior Research Methods
The MRC psycholinguistic database
Quarterly Journal of Experimental Psychology
The architecture of the bilingual word recognition system: From identification to decision
Bilingualism: Language and Cognition
The BIA model and bilingual word recognition
Visual word recognition by bilinguals in a sentence context: evidence for non-selective lexical access
Journal of Experimental Psychology: Learning, Memory, and Cognition
Cited by (36)
Language experiences and cognitive control: A dynamic perspective
2020, Psychology of Learning and Motivation - Advances in Research and TheoryCitation Excerpt :Among the different types of inhibitory control (response inhibition, and interference control which in turn consists of cognitive inhibition and focused attention, Diamond, 2013), focused attention has been frequently tested in bilingual research. It has been argued (e.g., Wu, Cristino, Leek, & Thierry, 2013) that a bilingual's two languages are activated even if they are using just one language, and so they have to focus their attention on the target language. Typical tasks to test focused attention include the tasks of Flanker, ANT, Simon, Stroop and antisaccade (see Dong & Zhong, 2019 for a comparison of these tasks).
Limitations of translation activation in masked priming: Behavioural evidence from Chinese-English bilinguals and computational modelling
2018, Journal of Memory and LanguageCitation Excerpt :Bilinguals have the unique ability to translate words between their languages. Although translation seems a deliberate and conscious process, recent research has shown that first language (L1) translation equivalents are automatically activated during second language (L2) word reading (e.g., Meade, Midgley, Sevcikova Sehyr, Holcomb, & Emmorey, 2017; Morford, Wilkinson, Villwock, Piñar, & Kroll, 2011; Thierry & Wu, 2007; Wu, Cristino, Leek, & Thierry, 2013; Wu & Thierry, 2010, 2012a, 2012b; Zhang, van Heuven, & Conklin, 2011). In an event-related potentials (ERP) study, Thierry and Wu (2007) presented proficient Chinese-English bilinguals with pairs of English words and asked them to judge whether these word pairs were related in meaning or not.
Investigating the time-course of phonological prediction in native and non-native speakers of English: A visual world eye-tracking study
2018, Journal of Memory and LanguageCitation Excerpt :It is possible that Japanese-English bilinguals did not activate Japanese at all during the experiment, because all the experimental setting was in English. However, this explanation is not likely given the evidence that L2 speakers activate their L1 even when L1 is irrelevant (Thierry & Wu, 2007; Wu, Cristino, Leek, & Thierry, 2013). Alternatively, a Japanese phonological competitor effect may have been too weak to affect eye movements – a possibility that seems reasonable given the relative weakness of the English phonological competitor effect (about 10% difference in fixation proportion, for about 150 ms).
Do you hear ‘feather’ when listening to ‘rain’? Lexical tone activation during unconscious translation: Evidence from Mandarin-English bilinguals
2017, CognitionCitation Excerpt :In their seminal work, Thierry and Wu (2007) showed, using ERPs, that Mandarin-English bilinguals implicitly accessed the Mandarin translation equivalents of targets when making semantic relatedness judgments on English pairs (e.g., 邮政-邮件 were activated when judging post-mail), as indexed by an amplitude reduction of the N400 component only for target pairs whose Mandarin translations shared first characters (e.g., 邮), compared to pairs whose translations did not share any characters in common. Even though this subtle effect appears to be more detectable in ERP measures compared to behavioral data, as indicated by Wu and Thierry (2010), Wu et al. (2013) demonstrated implicit access to translation equivalents even in a visual search non-linguistic task. More specifically, they showed that Mandarin-English bilinguals fixated more frequently on English words (e.g., reason) whose Mandarin translations (e.g., yuan yin) began with a morpheme which was a homophone of the Mandarin word for circle (i.e., yuan) or square (i.e., fang) when searching for a circle or square in arrays of English words that sometimes contained one of these shapes.
SECOND and THIRD LANGUAGE LEARNERS' SENSITIVITY to Japanese PITCH ACCENT IS ADDITIVE
2019, Studies in Second Language Acquisition