Readers generalize adaptation to newly-encountered dialectal structures to other unfamiliar structures.

Growing evidence suggests that syntactic processing may be guided in part by expectations about the statistics of the input that comprehenders have encountered; however, these statistics and even the syntactic structures themselves vary from situation to situation. Some recent work suggests that readers can adapt to variability in the frequencies of known, but infrequent syntactic structures. But, the relation between adaptation to altered frequencies of familiar structures and learning to process unfamiliar, never-before-seen structures is under-explored. In two self-paced reading experiments, we investigated readers' adaptation to an unfamiliar structure used in some regional dialects of American English: the needs+past participle structure, such as using The car needs washed to mean The car needs to be washed. Study 1 used a novel Web-based recruitment method to target regions where participants were likely to be familiar (Ohio, western Pennsylvania) or unfamiliar (Colorado) with the needs+past participle structure. Participants unfamiliar with the structure initially read the structure more slowly, but over the course of the experiment came to read it more like the familiar participants. Study 2 further demonstrated that participants who have adapted to needs+past participle generalize this adaptation to a different, but related structure. These results suggest (a) that readers adapt to unfamiliar syntactic structures, (b) that, in doing so, they become more like existing users of those structures, and (c) that they can generalize this other structures that they may also be more likely to encounter. We discuss these results in the context of implicit learning accounts of exposure effects on syntactic processing.