Verb-based versus class-based accounts of actionality effects in children's comprehension of passives.

Several studies have shown that children perform worse on tests of passive comprehension when the verb is non-actional than when it is actional. Most existing accounts focus on the semantic characteristics of the class of non-action verbs in explaining this difference. An alternative is a "verb-based" account in which passives are initially learned verb by verb, and children hear fewer non-actional passives in their language input. An analysis of the passives heard by Adam, Eve and Sarah (Brown, 1973) found more actional than non-actional passives, consistent with the verb-based account. In a second study, children tested for passive comprehension were re-tested a week later. The verb-based account predicts that children should show a consistent pattern of responses for individual verbs on test and re-test. Such consistency was found, with some inconsistency due to improvement over the re-test. Further analyses showed no effects of affectedness in explaining children's problems with passives. Finally, we discuss whether a mixed model containing both verb-based and class-based mechanisms is required to explain the actionality effects.