Recovering Ergativity in Heritage Samoan.

Abstract

The unique language profile of a heritage speaker offers an ideal opportunity to investigate the persistence of specific linguistic systems in the face of language shift (i.e. reduced input). Ergativity, the morphosyntactic system of alignment that sets A (the subject of a transitive verb) apart from O (the object of a transitive verb) and S (the sole argument of an intransitive verb), has been shown to be a fragile linguistic feature particularly sensitive to changes in language input, as it is often late acquired by children (Ochs 1982, Bavin & Stoll 2013), and lacking in heritage grammar (Schmidt 1985; Montrul et al. 2012). This dissertation investigates whether ergativity persists in the grammar of Samoan heritage speakers in spite of the abrupt shift in language input during early childhood (i.e. from Samoan to English, an accusative language). Samoan, a relatively understudied language, exhibits a robust system of ergativity at both the morphological (i.e. case, agreement) and syntactic level (i.e. relative clauses, wh-questions, quantifier float) (Mosel & Hovdhaugen 1992). Four experiments were carried out investigating the production and grammaticality judgement of key ergative features in declaratives, wh-questions, and relative clauses in three distinct speaker groups: native, heritage, and L2. The findings from this dissertation suggest that ergativity in Samoan is indeed a fragile system particularly susceptible to decreased language input. However, in spite of an initial lack of ergativity in heritage grammar, key ergative features were recovered through targeted linguistic intervention (i.e. explicit modeling, recasting). The results demonstrate that heritage speakers were able to recover an underlying pattern of ergativity (i.e. extending ergative features to structures not included in the intervention), while L2 speakers were only able to acquire construction-specific features. These findings lend support for the Permanence Hypothesis (Brenner 2010, cited in Benmamoun, Montrul, Polinsky 2013), that is, linguistic knowledge acquired during critical periods of language acquisition persists throughout life.