Where syntactic interference persists: The case of Hebrew written by native Arabic speakers

In this paper we analyze the acquisition of Hebrew syntax in a corpus of essays, written by 22 native speakers of Arabic after studying Hebrew for a decade. Each of the participants contributed two essays to the corpus: one when they were in the 11th grade of high school, and a second essay a year later. We categorized the syntactic errors, and explored the relationship between persistence, interference and developmental errors. Statistical analysis showed interference to be involved in the vast majority of the errors that persisted most between the two time-points, whereas almost all the improvement over the year was in developmental errors with no interference. These results contradict a common claim that interference, initially predominant, decreases over time with relation to developmental errors. We propose that a key difference is that much of the theorization in Second Language Acquisition (SLA) is based on findings in unrelated language dyads, with English as L1 or L2, whereas the languages in our study are closely related, yet differ considerably in their syntax. We conclude that more research on syntactic interference in the acquisition of related languages is necessary in order to reveal findings diverging from many typically attested patterns.