Last resort and no resort:
Whereas both Hebrew and Palestinian Arabic use resumption in object relatives, children with hearing impairment show completely different patterns of comprehension and production of object relatives in the two languages. This research compared the performance in the two languages, connecting the differences in the linguistic properties of resumption in the two languages to the observed differences in performance. In Hebrew, the resumptive strategy is optional in object relatives, and not always preferred. Young children produce many object relatives with a resumptive pronoun, but this tendency changes with age, and adults produce mainly object relatives without a resumptive pronoun. The assessment of relative clause indicates that Hebrew-speaking children with hearing impairment produce considerably more resumptive pronouns in object relatives than hearing children their age. In comprehension, resumptive pronouns provided an important clue to enhance the comprehension of relative clauses – Hebrew speaking children with hearing impairment understood object relatives with a resumptive pronoun significantly better than object relatives without a resumptive. In Palestinian Arabic, where resumptives are obligatory in object relatives, Palestinian Arabic-speaking hearing impaired children showed a completely different pattern from the Hebrew-speaking children. Although all object relatives were presented to them in comprehension with a resumptive pronoun, the presence of a resumptive pronoun did not help in the comprehension of relative clauses, and their comprehension of object relatives was very impaired. In production they were able to produce very few object relatives compared to their age-matched controls and compared to the Hebrew speakers. We provide an analysis for the different behavior between the two languages on the basis of the different syntactic properties of the resumptive pronouns. Resumptive pronouns are strong pronouns in Hebrew, and clitics in Palestinian Arabic. We assume that hearing impaired children have a general deficit in syntactic movement – which is independently motivated – and find additional support in the experiments presented in the paper. Following Hornstein (2001) and subsequent work, we contend that resumptive pronouns are a last resort strategy to rescue configurations in which movement is illegitimate. Since the hearing impaired children have a problem with movement, their pervasive use of resumptive pronouns in Hebrew follows straightforwardly. In Arabic, however, resumption is not helpful in salvaging derivations with movement, because the pronoun must enter the derivation already in the numeration, given its clitic properties. As such, it acts as a theta-role bearer and not as a last resort strategy to salvage an otherwise crashing derivation. As a result, whereas the resumptive pronoun in Hebrew functions as last resort for comprehension and production, the presence of a clitic in Palestinian Arabic does not provide any resort in either comprehension or production of object relatives, because it is only superficially similar to the resumptive strategy in Hebrew.