Resumen de (In)vulnerable agreement in incomplete bilingual L1 learners
This study investigates morphological vulnerability in incomplete bilingual L1 acquisition. It examines the production of L1 inflections by L2-dominant bilingual children, with the aim of exploring causes of difficulty in agreement marking. Spontaneous data (18 hours) from six Hungarian-English bilingual children, aged seven to nine, is used to compare the production of possessive inflections and verbal inflections, which are expressed by identical surface morphology in Hungarian. Drawing on recent analysis of Hungarian (Alberti, 1995; Szabolcsi, 1994), the paper asks where and why agreement morphology is susceptible to variability in bilingual children's weaker L1. Results indicate significant differences in the accuracy of agreement morphology. Inflections are almost error-free in possessive nominals and subject-verb agreement. However, they are frequently omitted in possessive be-clauses, which can be attributed to a combination of factors: (i) influence of semantic-syntactic properties of English have on Hungarian be possessive construction (transfer from L2); (ii) difficulty with long-distance agreement in Hungarian be-possessives (structural complexity of L1); and (iii) influence of Hungarian existential and locative constructions on possessive-be (ambiguity in L1). It is argued that morphological variability is selective, and the main source of vulnerability is the syntax-semantics interface where the weaker L1 is most susceptible to L2 influence.
