Developmental differences in the ability to apply cognate knowledge as a cue to meaning of unfamiliar English words

Developmental differences in the ability to apply cognate knowledge as a cue to meaning of unfamiliar English words

First Author: YI-JUI Iva CHEN -- Rightpath Research & Innovation Center. University of South Florida
Additional authors/chairs: 
Maria Carlo; Ke Cheng; Christopher Barr; Diane August
Keywords: Bilingualism, Cognates, Spanish, Vocabulary, multilevel logistic regression
Abstract / Summary: 

This study examined the relationship between awareness of the cognate status of English words and Spanish-speaking students’ depth of knowledge of Spanish words. We hypothesized that the probability of a correct response on an experimenter-developed measure of English word knowledge would show an advantage for cognates after controlling for variability in English word difficulty. We further hypothesized that Spanish word knowledge would moderate performance on cognates but not on non-cognates. Additionally, we examined developmental differences in cognate awareness between 3rd and 5th graders, a period in which morphological awareness is in flux. Two hundred and fifty-eight students (third-grade n =130, fifth-grade n =128) were administered a series of measures of Spanish and English proficiency, and then they performed on an English vocabulary test with 25 cognate words and 24 non-cognate words. Four theoretical driven two-level logistic regression models were conducted. Bayesian information criterion and Chi-square difference tests were used to identify the best-fitting model. Statistically significant cognate by grade-level interaction, β = 0.36, odds ratio = 1.44, p < .001, cognate by Spanish proficiency interaction β = 0.14, odds ratio = 1.14, p = .027, and cognates by grade-level by Spanish proficiency interaction, β = 0.45, odds ratio = 1.56, p < .001 were found in the best-fitted model. Results suggest that Spanish word knowledge moderates performance on English words that are cognates, and that 5th graders exploit cognate relationships more readily than 3rd graders. Given the current findings, we proposed a development-by-L1 threshold hypothesis to examine the contribution of cognates to L2 lexical acquisition.