Lexical Access Dynamics in Bilingual Speakers: Evidence from the Multilingual Lexical Neighborhood Effect

Abstract

Bilingual word recognition engages simultaneous competition from phonologically similar words in both languages. We examined 58 bilingual English-French speakers using eye-tracking during visual lexical decision. Targets embedded in phonologically dense neighbourhoods (many similar-sounding words) showed increased fixation times and more saccade corrections. Cross-linguistic phonological interference emerged for French targets preceded by English primes, suggesting shared, language-nonselective phonological representations.

Introduction

A central question in psycholinguistics concerns how bilingual speakers maintain separate lexical systems while managing inevitable cross-linguistic interactions. Classic models propose either language-selective access, wherein only the relevant language's lexicon is activated, or language-nonselective access, wherein phonologically and semantically similar words from both languages compete during word recognition. Recent evidence from monolingual speakers demonstrates robust lexical neighbourhood effects, whereby recognition of a target word is disrupted by the presence of phonologically similar competitors (neighbourhood density effect). However, the behaviour of such effects in bilinguals, where the competitor set potentially includes words from both languages, remains relatively poorly understood.

Eye-tracking methodology provides unprecedented opportunity to examine the time-course of lexical access and competition dynamics with millisecond precision. The present study employed this technique to investigate whether bilingual speakers experience phonological interference from cross-linguistic neighbours during visual word recognition. We predicted that French-English bilinguals would show increased lexical competition effects when targets were embedded in phonologically dense neighbourhoods in the non-target language, providing evidence for language-nonselective phonological activation.

Method

Participants

Fifty-eight bilingual adults (mean age = 24.3 years, SD = 3.1; 34 female) who were native speakers of French who had acquired English as a second language beginning before age 12 were recruited from the local university community. All participants reported daily exposure to both languages and rated themselves as highly proficient in both (self-rating > 7 on 10-point scales). English proficiency was verified via the Nelson English Language Test (mean score = 92.3%, SD = 4.1). All participants were right-handed and had normal or corrected-to-normal vision.

Procedure

Participants viewed 240 target words (120 English targets, 120 French targets) presented one at a time on a computer monitor while eye position was tracked using an EyeLink 1000 tracker (sampling rate: 1000 Hz). Target words varied in lexical neighbourhood density in the presentation language (high density: 12+ similar words; low density: 0-4 similar words), and we manipulated cross-linguistic phonological overlap (high-overlap pairs such as English "pacer" and French "passer"; low-overlap controls). Participants made speeded lexical decision judgments (word vs. non-word) by button press. A block of 40 trials in the non-target language preceded each 60-trial block in the target language, priming the non-target language during lexical processing. We analyzed fixation duration, saccade amplitude, and fixation count within the critical 200-800 ms post-stimulus window.

Results

Reaction time analysis revealed a robust neighbourhood density effect for targets in the presentation language: high-density English words were recognized more slowly (M = 687 ms, SD = 134) than low-density words (M = 621 ms, SD = 118), t(57) = 4.31, p < .001. This effect was significantly larger for French-language targets in which cross-linguistic English phonological competitors were present (mean difference = 78 ms) compared to low-competitor trials (mean difference = 31 ms), F(1, 56) = 6.47, p = .014, partial η² = .10. Eye-tracking data revealed that high-density targets elicited longer initial fixations (M = 298 ms vs. 267 ms, t(57) = 3.84, p < .001) and more regressive saccades (M = 1.43 vs. 0.69 regressions per trial, χ² = 8.92, p = .003). Critically, the phonological neighbourhood effect in one language was enhanced when cross-linguistic neighbours were present in the non-target language, even when targets were unambiguously presented in the target language context.

Analysis of fixation patterns 200-400 ms post-stimulus (lexical selection window) revealed that cross-linguistic interference emerged even for participants with high English proficiency (r = -.18, p > .05), suggesting that language-nonselective phonological activation is not contingent on proficiency-related differences in language balance.

Discussion

These results provide convergent evidence from both reaction time and eye-tracking measures that lexical access in bilinguals engages language-nonselective phonological representations. The augmented neighbourhood density effects observed when cross-linguistic phonological competitors exist align with computational models proposing that phonological activation spreads automatically across both languages' lexicons during word recognition. The early timing of these effects (200-400 ms) suggests that cross-linguistic interference operates at the level of phonological access itself, prior to semantic processing or language-specific lemma selection.

These findings advance models of bilingual lexical organization and have implications for understanding how bilingual experience shapes the architecture of the mental lexicon. Future neuroimaging work employing fMRI or magnetoencephalography could identify whether language-nonselective phonological activation recruits overlapping neural populations for phonologically similar words across languages. Additionally, investigations of how bilingual experience modulates these effects in highly proficient versus early-stage learners would illuminate the neural plasticity underlying second language acquisition.

References

Dijkstra, T., & van Heuven, W. J. (1998). The BIA model and bilingual word recognition. In J. Grainger & A. M. Jacobs (Eds.), Localist connectionist approaches to human cognition (pp. 189-225). Mahwah, NJ: Lawrence Erlbaum.
Luce, P. A., & Pisoni, D. B. (1998). Recognizing spoken words: the neighborhood activation model. Ear and Hearing, 19(1), 1-36.
Vitevitch, M. S., & Luce, P. A. (1999). Probabilistic phonotactics and neighborhood activation in spoken word recognition. Journal of Memory and Language, 40(3), 374-408.
Costa, A., Caramazza, A., & Sebastián-Gallés, N. (2000). The cognate facilitation effect: implications for models of lexical access. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26(5), 1283-1296.
Kroll, J. F., & Tokowicz, N. (2001). The development of conceptual representation for words in a second language. Journal of Memory and Language, 45(3), 496-518.
Brysbaert, M., & New, B. (2009). Moving beyond Kučera 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, 41(4), 977-990.