Advanced language attrition of Spanish in contact with Brazilian Portuguese

Language acquisition research frequently concerns itself with linguistic development and result of the acquisition process with respect to a first or subsequent language. For some, it seems tacitly assumed that a first language, once acquired, remains stable, regardless of exposure to and the acquisition of additional language(s) beyond the first one in childhood. Research on language attrition (language loss) questions the validity of this assumption and raises questions that will not only help in describing and explaining the nature of linguistic attrition, but also shed light on the mental (cognitive) representation of human language. The goal of this dissertation is to contribute to the general program of research that investigates possible domains of first language attrition and its cause(s). More specifically, I endeavor to test the predictions and theoretical tenability of the Interface Hypothesis (Sorace and Filiaci 2006) as applied to language attrition (e.g. Tsimpli et al 2004).

The Interface Hypothesis claims that certain linguistic properties, namely those at external interfaces such as the syntax/discourse interface, are especially vulnerable to optionality in language acquisition (see Sorace and Serratrice 2009). For attrition, it predicts that, upon sufficient exposure, linguistic properties that are dependent on interfaces between the linguistic computational system and external domains of cognition (such as pragmatics and discourse structure) are more vulnerable to erosion than those that lie internally to the linguistic system (e.g. syntax/semantic interface) or those that are purely syntactic in nature. Within this framework, attrition is hypothesized to either be due to direct interference from the L2 or due to linguistic processing deficits that are a byproduct of being bilingual. The comprehensive nature of this case study, which tests the L1 grammar of an adult native speaker of Spanish after 25 years of uninterrupted naturalistic exposure to Brazilian Portuguese across the different property types, not only allows for an examination of possible domains of attrition (e.g. external interfaces, internal interfaces, syntax) but also allows for teasing apart of the cause of attrition by combining both untimed and timed methodologies. Although the main focus of this dissertation is to test the limits and explanatory value of the Interface Hypothesis, the data will also be examined in light of other theories such as Paradis' (2004) Activation Threshold Hypothesis and Jakobson's (1940) Regression Hypothesis to determine the extent to which various theories might best explain the data to be obtained.