Hands Down: Cognate Effects Persist During Written Word Production.

Evy Woumans,Robin Clauws,Wouter Duyck

doi:10.3389/fpsyg.2021.647362

Abstract

Words that share form and meaning across two or more languages (i.e., cognates) are generally processed faster than control words (non-cognates) by bilinguals speaking these languages. This so-called cognate effect is considered to be a demonstration of language non-selectivity during bilingual lexical access. Still, research up till now has focused mainly on visual and auditory comprehension. For production, research is almost exclusively limited to speech, leaving written production out of the equation. Hence, the goal of the current study was to examine whether bilinguals activate representations from both languages during typewriting. Dutch-English bilinguals completed second-language written sentences with names of displayed pictures. Low-constraint sentences yielded a cognate facilitation effect, whereas high-constraint sentences did not. These findings suggest that co-activation of similar words across languages also occurs during written production, just as in reading and speaking. Also, the interaction effect with sentence constraint shows that grammatical and semantic sentence restrictions may overrule interlingual facilitation effects.

Highlights

If a bilingual reads, hears, or produces a word, do they activate a representation of that word in one or multiple languages? Much research has sought to answer this question and different theories have been put to the test
We have primarily focused on studies reporting cognate and interlingual homograph effects on words presented in isolation, as this was the prime interest of literature
A large body of research suggests that bilingual lexical access is language non-selective (Dijkstra et al, 1999; Costa et al, 2000; Lagrou et al, 2011a)

Summary

Introduction

Hears, or produces a word, do they activate a representation of that word in one or multiple languages? Much research has sought to answer this question and different theories have been put to the test. There is general consensus on an integrated lexicon (see Brysbaert and Duyck, 2010) or a segregated lexicon that is activated in parallel (Dijkstra et al, 2019). Evidence to support this idea comes from studies employing interlingual homonyms or cognates to demonstrate language non-selective activation (e.g., Dijkstra et al, 1999; Costa et al, 2000; Lagrou et al, 2011a,b). Cognates are words that share form (completely or predominantly) and meaning across languages, such as the Dutch and English word film. The cognate facilitation effect states that cognates are read (Cop et al, 2017), heard (Blumenfeld and Marian, 2007), and spoken (Costa et al, 2000) faster than noncognates by bilinguals speaking those specific languages, whereas cross-linguistic homophones seem to interfere with bilingual language processing, in listening (Lagrou et al, 2015) and reading

Objectives

Methods

Results

Conclusion