The potential scope of error-based theories of language acquisition: exploring prime surprisal in transitive and dative structures

Error-based theories of language acquisition predict enhanced structure repetition after surprising as opposed to predictable sentences. These prime surprisal effects have been found with different age groups and languages. However, previous studies have generally concentrated on datives, while error-based theories suggest that other structures should also be sensitive to prime surprisal. To explore this we carried out a prime surprisal study with transitive as well as dative structures including 114 monolingual English-speaking adult participants. We detected reliable structural priming with both structures and observed a numerical prime surprisal effect in the dative dataset. However, we found no evidence for prime surprisal in the transitive study, where participants showed weaker rather than stronger priming effects after more surprising sentences. We conclude that either prime surprisal effects are only sensitive to overall (but not graded) predictability or that online processing differences between the two structures are behind these results.