Lexical and semantic representations in the acquisition of L2 cognate and non‐cognate words: Evidence from two learning methods in children

Abstract

How bilinguals represent words in two languages and which mechanisms are responsible for second language acquisition are important questions in the bilingual and vocabulary acquisition literature. This study aims to analyse the effect of two learning methods (picture- vs. word-based method) and two types of words (cognates and non-cognates) in early stages of children's L2 acquisition. Forty-eight native speakers of European Portuguese, all sixth graders (mean age = 10.87 years; SD= 0.85), participated in the study. None of them had prior knowledge of Basque (the L2 in this study). After a learning phase in which L2 words were learned either by a picture- or a word-based method, children were tested in a backward-word translation recognition task at two times (immediately vs. one week later). Results showed that the participants made more errors when rejecting semantically related than semantically unrelated words as correct translations (semantic interference effect). The magnitude of this effect was higher in the delayed test condition regardless of the learning method. Moreover, the overall performance of participants from the word-based method was better than the performance of participants from the picture-word method. Results were discussed concerning the most significant bilingual lexical processing models.

List composition effect on cognate and non-cognate word acquisition in children

“…A significant area of interest in the field of bilingualism and second language acquisition (SLA) concerns the study of how words of first (L1) and second (L2) languages are represented in bilingual memory, as well as how they are accessed and selected (e.g., Comesaña, Perea, Pineiro, & Fraga, 2009;Comesaña, Soares, Sanchez-Casas, & Lima, 2012a;Dijkstra, Miwa, Brummelhuis, Sappelli, & Baayen, 2010;Jiang, 2000;Jiang & Forster, 2001;Van Hell & Kroll, 2012). Based on the Revised Hierarchical Model (RHM) by Stewart of 1994 (see Kroll, van Hell, Tokowicz, &Green, 2010;Van Hell & Kroll, 2012, for more recent reviews), several studies have explored the lexical-semantic connections between words in both languages at different stages of L2 acquisition (e.g., Altarriba & Mathis, 1997;Guo, Misra, Tam, & Kroll, 2012;Kroll & Stewart, 1994;Sunderman & Kroll, 2006;Tokowicz, Kroll, de Groot, & van Hell, 2002).…”

“…Despite evidence supporting the tenets of the RHM, recent studies have challenged its predictions, since, on one hand, proficient bilinguals also seem to show lexical mediation when processing L2 words (Guo et al, 2012), and, on the other hand, at early stages of L2 acquisition, learners exhibit direct access to the semantic system from L2 words (see Altarriba & Mathis, 1997;de Groot & Poot, 1997;Sunderman & Kroll, 2006 with adults, also Comesaña et al, 2009Comesaña et al, , 2012a. The inconsistency of these results may be due to the influence of variables in L2 processing other than level of proficiency, such as the method used to learn the new words (e.g., Comesaña et al, 2009;Finkbeiner & Nicol, 2003), the type of word being learned (e.g., Comesaña et al, 2012a;Davis et al, 2010;Tokowicz et al, 2002;van Heuven, Dijkstra, & Grainger, 1998), the translation direction (e.g., Davis et al, 2010;Heij, Hooglander, Kerling, & van der Velden, 1996), or even age (Chen & Leung, 1989;Lotto & de Groot, 1998). The present study falls within this area of research, and aims to explore further the extent to which the list composition of words to be learned (using a list in which L2 cognate and noncognate words were learned separately [blocked condition] and also a list in which the same cognate and non-cognate words were learned jointly [mixed condition]) affects the establishment of L2 word-to-concept connections using a picture-based method.…”

“…Indeed, there is a huge body of evidence showing differential processing of cognate and non-cognate words in bilinguals. More specifically, cognate words are learned faster than noncognate words both in adults (e.g., de Groot & Keijzer, 2000;Lotto & de Groot, 1998) and in children (Comesaña et al, 2012a), and evoke faster response times and fewer errors in comparison to non-cognates, the so-called cognate facilitation effect (but see Brenders, van Hell, &Dijkstra, 2011, andComesaña et al, 2015, for inhibitory effects of cognates as a function of L2 proficiency and stimuli list [5] composition). These findings support a non-selective account of lexical access in bilinguals, as assumed by the Bilingual Interactive Activation model (BIA, , and its extension (the BIA+ model, see Dijkstra et al, 2010;Dijkstra & van Heuven, 2002).…”

“…Following this line of reasoning, in a subsequent study Comesaña et al (2012a) explored the efficacy of the L2 word-picture method over the L2-L1 word association method in the establishment of L2 word-to-concept connections when noncognate and cognate words were learned simultaneously. In this study, which followed the same procedure used in Comesaña et al (2009), results showed a faster and more accurate processing for cognate words in both methods, this being consistent with the literature.…”

The Use of Cognate Words and Interlingual Homographs to Investigate the Cross-Linguistics in Second Language Processing in Iran

“…Bilingual speakers rate cognate translation pairs as more similar phonologically and orthographically than translation pairs that are not cognates (de Groot & Nas, 1991;Elgort & Piasecki, 2014). Monolingual speakers are more likely to correctly guess the meaning of the cognate pairs when asked to translate words in an unfamiliar language (Comesaña, Soares, Sánchez-Casas, & Lima, 2012;Friel & Kennison, 2001;Kroll & Stewart, 1994).…”