Abstract

Aims: We investigated: (1) whether differences in accuracy between heritage speakers (HS) and monolingual speakers (MS) signal differences in the path or merely in the rate of language development, and (2) whether, independently of these differences, HS become more accurate as they grow older. Methods: Using an elicitation task, we collected data from three groups of speakers of Greek: HS in the United States and Canada (78–226 months), MS of the same age (77–177 months), and younger MS (42–69 months). In terms of structures, we focused on two phenomena that are encoded differently in Greek and English: subject/object form in reference maintenance contexts and subject placement in embedded wh-dependencies. Data and Analysis: Data were analyzed with mixed-effects logistic regression models. Findings: We found that the heritage group had a lower accuracy and produced different error patterns than both monolingual groups. Specifically, only the heritage group produced non-felicitous lexical subjects/objects in reference maintenance contexts and ungrammatical preverbal subjects in embedded wh-structures. Accuracy, though, increased with age. Furthermore, current amount of heritage language (HL) input and generation, which were included as covariates, emerged as significant predictors in some or all of the conditions. Originality: The inclusion of a younger monolingual group helped us determine whether the different patterns observed in the language of HS are also attested in the language of MS at earlier developmental stages. The inclusion of a wide age range helped us determine whether, independently of differences in the path/rate of development, HS become more accurate as they grow older and accumulate the necessary amount of HL input. Implications: HS may go through developmental stages not attested in L1 acquisition. However, differences in developmental stages do not necessarily entail differences in the outcome of language acquisition. HS’ accuracy may continue to increase, provided that they continue using their HL.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call