That-trace effects in Najdi Arabic L2 learners of English A partial replication

The current study is a partial replication of Kim and Goodall (2024), who tested competing predictions of two prominent accounts of that-trace effects, which are argued to emerge due to either syntactic constraints or considerations of the production system. To tease apart these possibilities, Kim and Goodall examine L2 sensitivity to that-trace effects, as the two accounts implicitly have different expectations regarding L2 performance. Their results showed a non-native pattern of acceptability judgments for Korean and Spanish learners of English, whose L1s do not display that-trace effects, which are argued to support a production-based account. The current study extends their experiment to Nadji Arabic L2 learners of English, whose L1 critically exhibits that-trace effects, allowing us to probe whether previous findings can be accounted for by processing difficulties or L1 background. Our results indicated that despite L2 learners’ native-like sensitivity at the group level, lower proficiency was associated with nonnative-like subject extraction effects, in line with Kim and Goodall’s results. Overall, findings from an acceptability judgment task suggest that L2 sensitivity to that-trace effects does not involve transfer of a syntactic constraint, but is something that develops with proficiency, more in line with a production-based account.