Gender matters: From L1 grammar to L2 semantics

Abstract

The study investigates the effects of grammatical gender on bilingual processing. Native speakers of Russian (a gendered language) learning English and monolingual English controls performed a self-paced reading task in English (a non-gendered language). As predicted, bilingual speakers showed delayed latencies to gendered pronouns (he or she) that were incongruent with the noun's grammatical gender in Russian, indicating that first language (L1) grammatical gender assignment can be interpreted as biological gender in nonnative (L2) processing. The L1 gender bias was only found in sentences containing animate, but not inanimate, nouns. These results speak against the syntactic mechanism being solely responsible for gender biases, but rather support a semantic transfer account due to coactivation of linguistic and conceptual features as proposed in the sex and gender hypothesis (SAGH, Vigliocco, Vinson, Paganelli & Dworzynski, 2005). Overall, the study provides clear evidence for the L1 grammatical gender bias in bilingual processing, albeit constrained by animacy.

Gender congruency from a neutral point of view: The roles of gender classes and conceptual connotations.

“…However, our findings also deviate from this previous research in three important ways. First, we did not find the difference in effects for (GENERIC) ANIMATES versus NON-ANIMATES reported elsewhere (Cook, 2016;Saalbach et al, 2012;Vigliocco et al, 2005). Second, the effect in our study emerged both in an implicit task and for speakers of a language with three genders, for which such findings tend to be unstable or even absent (Cubelli et al, 2011;Sera et al, 2002).…”

“…From a theoretical point of view, two categories of nouns need to be distinguished: nouns referring to non-animates (mostly objects) that have grammatical gender but no sex (henceforth abbreviated as NON-ANIMATES), and nouns referring to animates (mostly animals) that have both grammatical gender and sex, with the two being potentially in conflict, such as when one particular gender is generically used for individuals of both sexes (GENERIC ANIMATES). Although our own work did not yield diverging effects for the two categories (Bender et al, 2016a), there is a body of research pointing to such differences-most notably that an effect may emerge for animals but not objects (Cook, 2016;Saalbach et al, 2012;Vigliocco et al, 2005), and that natural kinds are more strongly associated with female properties and artifacts more strongly with male properties (Mullen, 1990;Sera et al, 1994Sera et al, , 2000. Although our category of GENERIC ANIMATES is, by definition, confined to natural kinds and our category of NON-ANIMATES consists largely (albeit not exclusively) of artifacts, we ensured that items in each category evoke associations in line with the particular grammatical gender.…”

“…Even if this pattern is replicated with similar explicit tasks by studies that additionally employ implicit measures, findings with their implicit measures also suggest that the impact of grammatical gender is confined to the lexical level, as its effect is more stable for words than for pictures (Bowers et al, 1999;Ramos & Roberson, 2010;Vigliocco et al, 2005), is more likely to emerge for some categories (notably animals) than others (Cook, 2016;Saalbach et al, 2012;Vigliocco et al, 2005), depends on language context (Kousta et al, 2008;Vigliocco et al, 2004), disappears under articulatory suppression (Cubelli et al, 2011), or is not manifest on the behavioral level at all (Bender et al, 2011;Boutonnet et al, 2012). Finally, when investigating inferences on the biological properties of animals, effects of grammatical gender are found beyond those of the 1 This is true, of course, only in languages whose two genders are masculine and feminine.…”

Grammatical gender and linguistic relativity: A systematic review

“…Such results might be explained in terms of an effect of membership in the same grammatical category, independently of biological sex information (cf. Cook, 2016). Because this review is entirely concerned with this specific relationship such studies, though clearly interesting in their own right, were omitted.…”

Outliers in L2 Research in Applied Linguistics: A Synthesis and Data Re-Analysis

“…Lower word boundaries for excluding unusually fast RTs in SPR research are generally around 200 ms for word-by-word presentations (Jegerski, 2016; Jiang, 2007; Sagarra & Herschensohn, 2010). However, they can be 100 ms (Kim, 2018; Leal, Slabakova, & Farmer, 2017; Litcofsky & Van Hell, 2017), or even as low as 40 or 50 ms (Cook, 2018; Frank, Trompenaars, & Vasishth, 2016), despite research suggesting 100 ms as a minimum to respond to the identity of a signal (Luce, 1991). Rather than rely on conventions, some researchers provide research-driven justifications to determine a cutoff.…”