Cognates are studied in many psychological studies of bilingual language processing. Despite their frequent use, there is no clear operationalized definition of what constitutes a cognate. We conducted a literature search in three major journals to better understand how cognate status is typically defined and operationalized. In these journals, we analyzed similarity of cognate and non-cognate stimuli. We found that approximately 60% of the reviewed studies operationalized cognate status empirically. Stimulus analyses revealed a similarity continuum between cognates and non-cognates without a consistent cut-off. Based on these results, we make recommendations for future research.

To investigate how orthography and semantics interact during bilingual visual word recognition, Dutch–English bilinguals made lexical decisions in two masked priming experiments. Dutch primes and English targets were presented that were either neighbour cognates (boek – BOOK), noncognate translations (kooi – CAGE), orthographically related neighbours (neus – NEWS), or unrelated words (huid - COAT). Prime durations of 50 ms (Experiment 1) and 83 ms (Experiment 2) led to similar result patterns. Both experiments reported a large cognate facilitation effect, a smaller facilitatory noncognate translation effect, and the absence of inhibitory orthographic neighbour effects. These results indicate that cognate facilitation is in large part due to orthographic-semantic resonance. Priming results for each condition were simulated well (all r's >.50) by Multilink+, a recent computational model for word retrieval. Limitations to the role of lateral inhibition in bilingual word recognition are discussed.

To test the BIA+ and Multilink models’ accounts of how bilinguals process words with different degrees of cross-linguistic orthographic and semantic overlap, we conducted two experiments manipulating stimulus list composition. Dutch–English late bilinguals performed two English lexical decision tasks including the same set of cognates, interlingual homographs, English control words, and pseudowords. In one task, half of the pseudowords were replaced with Dutch words, requiring a ‘no’ response. This change from pure to mixed language list context was found to turn cognate facilitation effects into inhibition. Relative to control words, larger effects were found for cognate pairs with an increasing cross-linguistic form overlap. Identical cognates produced considerably larger effects than non-identical cognates, supporting their special status in the bilingual lexicon. Response patterns for different item types are accounted for in terms of the items’ lexical representation and their binding to ‘yes’ and ‘no’ responses in pure vs mixed lexical decision.

The computational BIA+ model (Dijkstra & Van Heuven, 2002) has provided a useful account for bilingual word recognition, while the verbal (pre-quantitative) RHM (Kroll & Stewart, 1994) has often served as a reference framework for bilingual word production and translation. According to Brysbaert and Duyck (2010), a strong need is felt for a unified implemented account of bilingual word comprehension, lexical-semantic processing, and word production. With this goal in mind, we built a localist-connectionist model, called Multilink, which integrates basic assumptions of both BIA+ and RHM. It simulates the recognition and production of cognates (form-similar translation equivalents) and non-cognates of different lengths and frequencies in tasks like monolingual and bilingual lexical decision, word naming, and word translation production. It also considers effects of lexical similarity, cognate status, relative L2-proficiency, and translation direction. Model-to-model comparisons show that Multilink provides higher correlations with empirical data than both IA and BIA+ models.

In two lexical decision experiments, we investigated how sentence language affects the bilingual's recognition of target words from the same or a different language. Dutch–English bilinguals read Dutch (L1) or English (L2) sentences, presented word by word, followed by English (Experiment 1) or Dutch (Experiment 2) target words. Targets were Dutch–English cognates or non-cognates in isolation or preceded by sentences providing a high or a low semantic constraint. English cognates were facilitated irrespective of whether they were preceded by high or low constraining English sentences (no language switch) or Dutch sentences (switch). For Dutch cognates, inhibition effects arose in low constraining sentences (irrespective of Dutch or English) and in English (switch) sentences (irrespective of semantic constraint). Thus, under mixed language conditions, sentence constraint modulates target word processing but does not always completely eliminate cross-linguistic effects. The results are interpreted in a BIA+ model that extends monolingual views on sentence comprehension.

This study investigated two prominent issues in the comprehension of language switches. First, how does language switching direction affect switch costs in sentence context? Second, are switch costs modulated by L2 proficiency and cross-linguistic activation? We conducted a self-paced reading task involving sentences that switched between participants’ L1 Dutch and L2 English. The cognate status of the main verb was manipulated to examine the influence of co-activation on intra-sentential switch costs. The reading times indicated the influence of switch direction: a cost was observed for switches into L2 but not for switches into L1, and the magnitude of the costs was correlated with L2 proficiency, indicating that switch costs in language comprehension depend on language dominance. Verb cognates did not yield a cognate facilitation effect nor did they influence the magnitude of switch costs in either direction. The results are interpreted in terms of an activation account explaining lexical comprehension based on L2 proficiency.

Researchers on bilingual processing can benefit from computational tools developed in artificial intelligence. We show that a normalized Levenshtein distance function can efficiently and reliably simulate bilingual orthographic similarity ratings. Orthographic similarity distributions of cognates and non-cognates were identified across pairs of six European languages: English, German, French, Spanish, Italian, and Dutch. Semantic equivalence was determined using the conceptual structure of a translation database. By using a similarity threshold, large numbers of cognates could be selected that nearly completely included the stimulus materials of experimental studies. The identified numbers of form-similar and identical cognates correlated highly with branch lengths of phylogenetic language family trees, supporting the usefulness of the new measure for cross-language comparison. The normalized Levenshtein distance function can be considered as a new formal model of cross-language orthographic similarity.

The present study aimed to determine how older bilingual subjects' naming performance is affected by their knowledge of two languages. Twenty-nine aging (mean age = 74.0; SD = 7.1) Spanish–English bilinguals were asked to name all pictures in the Boston Naming Test (BNT) first in their dominant language and then in their less-dominant language. Bilinguals with similar naming scores in each language, or relatively balanced bilinguals, named more pictures correctly when credited for producing a correct name in either language. Balanced bilinguals also named fewer pictures in their dominant language than unbalanced bilinguals, and named more pictures correctly in both languages if the pictures had cognate names (e.g., dart is dardo in Spanish). Unbalanced bilinguals did not benefit from the alternative (either-language) scoring procedure and showed cognate effects only in their nondominant language. These findings may help to guide the interpretation of neuropsychological data for the purpose of determining cognitive status in older bilinguals and can be used to develop models of bilingual language processing. Bilinguals' ability to name pictures reflects their experience with word forms in both languages. (JINS, 2007, 13, 197–208.)