Between-Language Interaction in Early Acquisition of Speech. An Analysis of Gliding by Bilingual Preschoolers in Southern Arizona

Children who grow up in Spanish-English bilingual families in Arizona, US, often start as Spanish-dominant speakers but become English-dominant as they advance in school (Genesee 2004). Two key factors in this language shift are campaigns promoting English monolingualism in schools (Fishman 2013) and lack of understanding of bilingual speech acquisition, resulting sometimes in therapy over-diagnosis (Yavas 1998). The present dissertation contributes to the understanding of bilingual speech sound acquisition, as it explores the distribution of inaccuracies of late-developing sounds by preschoolers undergoing this language shift. Specifically, this research study investigates the distribution of glides as substituted sounds. Glides present a conflict in the phonologies of bilingual Spanish-English speaking children because of their dual behavior in each language. Spanish glides [j] and [w] share acoustic and phonological features with the high vowels [i] and [u] (Hualde 1997), whereas English includes both sounds [j] and [w] in the consonant inventory (Ladefoged & Johnson 2014). When children have not mastered late-developing sounds, such as rhotics or laterals, they substitute these sounds with easier ones in their inventory. In English, the substitution pattern of gliding (“rabbit” pronounced as [‘wæbɪt] is frequently found in preschoolers, but it does not occur in monolingual Spanish, where the rothics [r ɾ] are replaced by [l] (“rojo” pronounced as [‘lo.xo]) or [ð] ([‘ðo.xo]), (Bosch 1983). Taking into account the phonological distributions of glides in English and Spanish, we posed the question, “Do bilingual children show a different substitution pattern of late-developing rhotics and laterals in English?” This study investigates gliding in Spanish-English bilingual children born and raised in a southwest border region of the U.S. We do so by comparing the single word outputs of 61 typically-developing children from the same preschool with (or without) language exposure. Only 22 of these children exhibited gliding (11 Spanish-English speaking children and 11 English-speaking children). A nonparametric Mann-Whitney U test indicated gliding occurred significantly more often in monolinguals than bilinguals, suggesting between-language interaction (p = .031). Results are discussed within the Processing Rich Information from Multidimensional Interactive Representations model (PRIMIR; Curtin et al. 2011a). Through mixed methods analysis with a sociolinguistic approach, this investigation shows evidence of how children create two different phonological systems while acquiring the language, as well as how these languages interact with each other