Slide Slam A15
Influences on cognate facilitation in healthy bilinguals and bilinguals with aphasia
Manuel J. Marte1, Swathi Kiran1; 1Aphasia Research Laboratory Department of Speech, Language, & Hearing Sciences Sargent College of Health and Rehabilitation Sciences, Boston University
Introduction: Cognates share a continuum of overlapping phonological and orthographic features between languages, resulting in a facilitatory effect during lexical access (Costa et al., 2000). In bilingual individuals, a spectrum of language experiences (e.g., use, age of acquisition) may influence the extent of facilitation between native (L1) and second language (L2; Rosselli et al., 2014). Aphasia, an access deficit, tests the resilience of the facilitation effect (Kohnert, 2004). We sought to examine and compare the factors influencing naming accuracy in L1 and L2 on English-Spanish cognates and noncognates in healthy bilinguals (HB) and bilinguals with aphasia (BWA) while examining the influence of language experience. Methods: Twenty-seven bilinguals with aphasia (mean age = 53, SD = 16.17, range = 21 – 82) and 31 healthy bilinguals (mean age = 43, SD = 15.34, range = 18 – 82) completed the Language Use Questionnaire (LUQ; Kiran et al., 2010) and Boston Naming Test (BNT) in English and Spanish. BWA also completed the Western Aphasia Battery-Revised (WAB-R) in both languages. In matching for item difficulty, 22 cognates and 22 noncognate pairs were selected from the BNT (Gollan et al., 2007). Cognateness was measured continuously by first computing and then taking the average of normalized Levenshtein distance ratings between written and phonetic transcriptions of the aforementioned word pairs. Principal component analyses (PCA) were performed on L1 and L2 LUQ scores from all 31 HB and from a larger set of 59 BWA; 27 BWA were included in this analysis. This resulted in Background, Use, and Environment components in both L1 and L2 across both groups. Individual-specific factor loadings were extracted from PCA results and used in regression analyses. Logistic mixed-effects models were fitted to examine the effect of cognateness and language experience in both L1 and L2. In BWA models, we included Aphasia Quotient (AQ) per WAB-R. For each group, four models were constructed, two in each language that examined either BNT item difficulty or lexical frequency. Results: In HB, cognateness is a significant predictor of naming accuracy in both languages across all models. In contrast, results in BWA suggest that cognateness is not a significant predictor of naming accuracy in either language when controlling for lexical frequency. Furthermore, naming accuracy in BWA is influenced by item difficulty and lexical frequency, Use in L1 and L2, and AQ in L1 and L2. Likewise, accuracy in HB is influenced by item difficulty and lexical frequency, language experience components in both languages (Use and Environment), and interactions between language and these components. Cognate advantage was stable between languages within groups but comparatively diminished in BWA: L1 in HB (0.25, d = 0.57); L2 in HB (0.26, d = 0.57), versus L1 in BWA (0.09, d = 0.19); L2 in BWA (0.08, d = 0.17). Summary: These results suggest that (i) quantified language experience is an informative predictor of naming in both healthy and disordered bilingual populations, and (ii) cognateness yields only minor benefits for bilingual individuals with aphasia, exhibiting minimal predictive power on naming accuracy.