In their keynote paper, Dijkstra, Wahl, Buytenhuijs, van Halem, Al-jibouri, de Korte, and Rekké (Reference Dijkstra, Wahl, Buytenhuijs, van Halem, Al-jibouri, de Korte and Rekké2018) present a computational model of bilingual word recognition and translation, Multilink, that integrates and further refines the architecture and processing principles of two influential models of bilingual word processing: the Bilingual Activation Model (BIA/BIA+) and the Revised Hierarchical model (RHM). Unlike the earlier models, Multilink has been implemented as a computational model so its design principles and assumptions can be compared with human processing data in simulation studies – which is an important step forward in model development and refinement. But Multilink also leaves behind an important theoretical advancement that was touched upon in extending BIA to BIA+ (Dijkstra & Van Heuven, Reference Dijkstra and Van Heuven2002): how linguistic context influences word processing. In their presentation of BIA+, Dijkstra and Van Heuven (Reference Dijkstra and Van Heuven2002) hypothesized that syntactic and semantic aspects of sentence context may affect the word identification system. Theoretically, this was an important step forward, as none of the bilingual word processing models (and few monolingual word processing models, for that matter) had incorporated linguistic context, and at that time only a handful of empirical studies had examined how linguistic context affects bilingual word processing. However, in the past 15 years a significant body of empirical work has been published that examines how semantic and syntactic information in sentences impacts word processing in bilinguals. These important insights are not incorporated in the Multilink model.

Studies that examined how linguistic context impacts lexical access in bilinguals have found that the language of the sentence context in and of itself does not affect cross-language activation of target words embedded in the sentence, and is not sufficient to restrict activation of lexical alternatives to the language of the sentence context, not even in cross-script bilinguals reading sentences that unambiguously point at one language (Jouravlev & Jared, Reference Jouravlev and Jared2014). However, semantic information has been found to modulate cross-language activation, suggesting that readers and listeners use semantic cues to guide lexical access (for a review, see Lauro & Schwartz, Reference Lauro and Schwartz2017). Specifically, semantic cues in a sentence can bias towards a particular meaning, and thereby constrain the activation of upcoming words, and the degree of language-nonselective access in bilingual processing. For example, Van Hell and De Groot (Reference Van Hell and De Groot2008), testing Dutch–English bilinguals, found that semantically high constraint sentence contexts (e.g., “The best cabin of the ship belongs to the . . . [captain]”), but not low constraint sentence contexts (e.g., “The handsome man in the white suit is the . . . [captain]”), eliminated the cognate facilitation effect in L2 lexical decision and reduced the cognate facilitation effect in forward (from L1 to L2) and in backward (from L2 to L1) translation production. These findings suggest that semantic constraints in a sentence context affect language selection via top-down activation of appropriate lexical entries and thus influence bottom-up processes of lexical access in bilingual word recognition and translation, as implemented in Multilink.

A related line of research studying prediction during L1 and L2 language comprehension suggests that bilingual readers and listeners may also use sentence structure and syntactic information to guide the selection of target words embedded in the sentence and constrain the degree of language-nonselective access (cf. Zirnstein, Van Hell & Kroll, Reference Zirnstein, Van Hell and Kroll2018). For example, Foucart, Martin, Moreno, and Costa (Reference Foucart, Martin, Moreno and Costa2014) had Spanish–Catalan bilinguals and French–Spanish bilinguals read semantically constraining sentences in Spanish in which the gender marking of the article preceding the sentence-final noun either matched or did not match the expected word (e.g., “The pirate had a secret map, but he never found the [masculine] treasure [expected target] / the [feminine] cave [unexpected target] he was looking for.”). Foucart et al. observed an effect of congruency of the article and the predicted noun (i.e., larger N400 effects were elicited on the preceding article when the gender did not match the expected word, and N400 and frontal positive effects were elicited for unexpected targets), suggesting that bilinguals can use semantic and gender agreement cues to predict upcoming target words. Interestingly, the means and visual inspection of the grant averages for the French–Spanish bilinguals suggested more pronounced effects for cognate than for noncognate target words, but the interaction between expectation and cognate status did not reach significance, which the authors attribute to lack of statistical power due to the low number of cognates and noncognates.

Bilinguals rarely read or listen to isolated words, and a substantial body of evidence indicates that semantic and grammatical cues potentially guide lexical access in bilingual word recognition and translation. A next step forward is to incorporate this knowledge in Multilink, and examine how contextual information affects its basic principles and assumptions regarding nonselective lexical access, resting level activation, and interactions among orthographic, semantic, and phonological codes.