The Effects of Attrition on Grammatical Gender: A View from North American Icelandic

Abstract

Heritage grammars have been argued to differ with respect to whether they are an instantiation of divergent attainment or attrition. Attrition and divergent attainment are not mutually exclusive and can even co-exist with respect to the same or different grammatical phenomena, but teasing these apart requires longitudinal studies or carefully selected cross-sectional data (Montrul, 2008; 2016; Polinsky, 2011). In this article we present data from a longitudinal corpus of letters written by a speaker of North American Icelandic over a span of seventy-two years. The earliest letters suggest that the writer acquired Icelandic consistent with the baseline. However, in the last thirty years of writing, non-target forms emerge in the corpus. Morphosyntax, notably grammatical gender and inflectional morphology, is the most affected domain of grammar. In this article we focus on the nature of the changes attested for grammatical gender across time. Our results show that gender assignment does not undergo a systematic reanalysis. However, the non-target gender agreement indicates the overuse of an agreement default, which may reflect a trend towards a systematic reduction of the gender agreement system.