Disentangling cross-language orthographic neighborhood from markedness effects in L2 visual word recognition

Abstract

Previous research has reported that lexical access in bilinguals is language non-selective. In the present study, we explored the extent to which cross-language orthographic neighborhood size (N-size) effects, an index of language non-selectivity, should be dissociated from markedness effects, a sub-lexical orthographic variable referring to the degree of language- shared (unmarked) versus specific (marked) orthography. Two proficiency groups of French/English bilinguals performed an English (L2) lexical decision task with three word and non-word conditions: (1) English words with large French N-size/unmarked orthography (price), (2) small French N-size/unmarked orthography (drive), and (3) small French N-size/marked orthography (write). Evidence was found for orthographic markedness effects, albeit with a different pattern for word and non-word processing: while marked words were facilitated (responded to faster and more accurately) compared to unmarked words, the opposite pattern emerged for non-words. The pattern of results was comparable in both proficiency groups. No evidence emerged for the influence of first language (L1) neighborhood on L2 word or non-word processing. Thus, the results emphasize the need to integrate orthographic markedness as a relevant psycholinguistic variable in bilingual models of visual word recognition such as BIA/+ and to take it into account when investigating cross- language effects and the issue of language non-selectivity during visual word recognition.

Eye Movement Measures of Within-Language and Cross-Language Activation During Reading in Monolingual and Bilingual Children and Adults: A Focus on Neighborhood Density Effects

“…The bilingual literature parallels the monolingual literature; it has reported mixed patterns of facilitatory, inhibitory, and null within-language and cross-language orthographic neighborhood density effects among healthy bilingual young adults (aged 18-30) across their L1 and L2 (e.g., Beauvillain, 1992;de Groot et al, 2002;Dirix et al, 2017, Commissaire et al, 2019Grainger and Dijkstra, 1992;Van Heuven et al, 1998;Lemhöfer et al, 2008;Midgley et al, 2008;Grossi et al, 2012;Meade et al, 2018;Mulder et al, 2018). As discussed previously, these between-study differences are likely driven by methodologyrelated factors, including whether orthographic neighborhood frequency was accounted for, as well as by participant-related factors.…”

Investigating L2 reading aloud and silent reading in typically developing readers and dyslexic adolescents from grades 6 to 9

“…Marked words are words with rather language‐specific (L1 or L2) sub‐lexical orthographic patterns (e.g., ‐ght and ‐oeu are specific to English and French, respectively), whereas unmarked words contain shared orthographic patterns across languages (e.g., ‐ain is common in both English and French). Bilingual research among adults has reported faster processing of L2 marked words compared with unmarked words in tasks that require judging the language membership of the word (i.e., language decision task, Casaponsa, Carreiras, & Duñabeitia, 2014; Vaid & Frenck‐Mestre, 2002) or deciding whether the item is a real L2 word or not (i.e., lexical decision task, Commissaire, Audusseau, & Casalis, 2019). Despite a lack of empirical studies on how L2 marked versus unmarked words are read by young L2 learners, either typical or dyslexic readers, marked words seem to be less easily learned compared with unmarked words when using a word‐learning paradigm (with an artificial language, Bartolotti & Marian, 2017).…”

Do both WRAP and TRAP inhibit the recognition of the French word DRAP? Impact of orthographic markedness on cross-language orthographic priming

“…Thus, these two studies focusing on neighbourhood size effects across languages strongly challenge the existence of cross-language lexical competition in pure monolingual tasks such as lexical decision, with null effects being obtained, perhaps due to confounds with orthographic markedness in previous studies (Commissaire et al, 2019;Mulder et al, 2018). It should also be noted that the few eye-tracking studies on the topic also reported puzzling results, as cross-language facilitatory neighbourhood size effects were found in more natural sentence reading experiments (Dirix et al, Experiment 2;Whitford & Titone, 2019), which also does not support the lexical competition hypothesis.…”

“…This hypothesis is supported by evidence of lexical competition effects between cross-language orthographic neighbour words such as fire — rire (laugh) in the English/French language pair (Bijeljac-Babic et al, 1997; Dijkstra et al, 2010; Dirix et al, 2017; Experiment 1; Grossi et al, 2012; Midgley et al, 2008; Van Heuven et al, 1998), a mechanism that is therefore implemented in several models such as the Bilingual Interactive Activation model (BIA, Van Heuven et al, 1998), and its extension BIA+ (Dijkstra & Van Heuven, 2002). However, the growing interest on sub-lexical orthographic coding, and especially on the orthographic markedness effect, raises new issues about language detection mechanisms (Casaponsa et al, 2014, 2015, 2020; Casaponsa & Duñabeitia, 2016; Commissaire et al, 2014, 2019; Hoversten et al, 2017; Oganian et al, 2016; Vaid & Frenck-Mestre, 2002; Van Kesteren et al, 2012) and also suggests that this variable could modulate the degree of language non-selectivity in bilingual lexical access (Casaponsa & Duñabeitia, 2016). This study aimed to investigate masked L2-to-L1 orthographic priming, an index of cross-language lexical competition, in French/English bilinguals while considering words’ orthographic sub-lexical properties.…”

“…To do so, examining orthographic neighbourhood effects across languages turns out to be a better index of lexical competition in the bilingual lexicon, thereby also providing a more conservatory test of language non-selectivity. And yet, empirical studies on cross-language lexical competition per se are relatively scarce in the last 30 years (Bijeljac-Babic et al, 1997; Commissaire et al, 2019; Dijkstra et al, 2010; Dirix et al, 2017; Grossi et al, 2012; Midgley et al, 2008; Mulder et al, 2018; Van Heuven et al, 1998) and contradictory findings have recently emerged (Commissaire et al, 2019; Mulder et al, 2018; see also eye-tracking studies by Dirix et al, 2017, Experiment 2 and Whitford & Titone, 2019).…”