Cross-linguistic lexical and syntactic co-activation in L2 sentence processing

Abstract

This study investigates under which conditions the L1 syntax is activated in L2 on-line sentence comprehension. We study whether cross-linguistic syntactic activation of the L1 word order is affected by lexical activation of the first language (L1) by virtue of cognate words. In two eye-tracking experiments, German-English bilinguals and English natives read English sentences containing reduced relative clauses whose surface word order partially overlaps with German embedded clauses. The verbs used were either German-English cognates or matched control verbs. The results show lexical cognate facilitation and syntactic co-activation of L1 word order, with the latter being moderated by proficiency and cognate status. Critically, syntactic co-activation is found only with English control words. We argue that fleeting co-activation of the L1 syntax becomes measurable under higher demands of lexical processing, while cognate facilitation frees resources for inhibition of the L1 syntax and target-like syntactic processing.

The impact of a momentary language switch on bilingual reading: Intense at the switch but merciful downstream for L2 but not L1 readers

“…The purpose of these two experiments was to address whether cross-language activation during bilingual reading is sensitive to differences in language experience among bilinguals (such as the current exposure to the L2) and, crucially, to shifting language demands (such as a momentary shift in the language of the sentence). In line with the BIA+ model of word recognition (Dijkstra & Van Heuven, 2002) and many previous empirical studies (e.g., Gullifer et al, 2013;Hopp, 2016;Lauro & Schwartz, 2017; Pivneva et al, 2014;Titone et al, 2011;Van Hell & Dijkstra, 2002), we observed cross-language activation when French-English bilinguals read words within the context of sentences in their L2. Also in line with BIA+, evidence for cross-language activation was qualitatively much weaker when English-French bilinguals read in their L1.…”

“…Also in line with BIA+, evidence for cross-language activation was qualitatively much weaker when English-French bilinguals read in their L1. Crucially, in contrast with BIA+ but in line with general models of bilingual language control (Abutalebi & Green, 2007;2016;Green & Abutalebi, 2013), cross-language activation of the L2 was significant during L1 reading, but only following a prior switch into the L2. This suggests that the control demands placed on the system by a switch into the second language may have required bilinguals to suppress the semantics of the L1, allowing for greater cross-language activation of the L2 downstream from a switch.…”

“…Information about the time-course of cross-language activation is provided by measures of eye-movements during reading. Reading studies support the notion that cross-language activation occurs during the earliest stages of word recognition (indexed by eye-movement measures that include first-pass fixations, such as gaze duration) and persists into late stages into word recognition (indexed by eye-movement measures that include second-pass fixations, such as total reading time; e.g., Libben & Titone, 2009;Pivneva et al, 2014;Titone et al, 2011;Van Assche et al, 2009; for reviews on crosslanguage activation, see Kroll et al, 2013;2016;Lauro & Schwartz, 2017;Whitford et al, 2016; but see Hoversten & Traxler, 2016). The magnitude of cross-language activation is indeed related to individual differences in relative language proficiency, with larger magnitudes arising when proficiency in the target language is lower or when exposure to the non-target language is higher (e.g., Pivneva et al, 2014;Titone et al, 2011).…”

“…In addition to BIA+, more general models of bilingual language control, such as various instantiations of the inhibitory control model (Abutalebi & Green, 2007;2016;Green & Abutalebi, 2013), also assume that both languages are activated in parallel. Moreover, control accounts assert that cross-language activation of the L2 and L1 must be actively regulated in a manner that is proportional to the dominance of each language, with the more dominant language (typically the L1) requiring more control for successful suppression than the less dominant language.…”

“…Crucially, if switching also has consequences for the magnitude of cross-language activation, then homograph and cognate effects should be larger following a prior language switch. Both BIA+ (Dijkstra & Van Heuven, 2002) and control accounts (e.g., Green, 1998;Abutalebi & Green, 2007;2016;Green & Abutalebi, 2013) predict that cross-language activation depends on language exposure, sentence context effects (i.e., language switching), and task demands (i.e., trial order). However, the two theoretical accounts outlined above make different predictions regarding the time-course and directionality of the influences of language switching on cross-language activation.…”