888 - Utku Türk / When do we plan agreement?
888 - Utku Türk / When do we plan agreement?
Tuesday April 15, one of several certain things is the defense of Utku Turk's 888, "When do we plan agreement: Evidence from agreement attraction and unaccusativity," to the committee of Colin, Ellen and Jeff. When do they plan to agree? Come at 1:00 and find out.
Abstract
This paper investigates the timing of the planning of verbal agreement in sentence production by examining agreement errors in environments where the syntactic relation between the subject and the verb is manipulated. Recent work in sentence planning has shown that when the syntactic relation of the verb and the subject is tighter (unaccusative verbs subjects as their complement), participants diverged from their usual incremental planning and plan the before they start uttering sentences along with the subject in sentences like The octopus below the spoon is boiling (Momma & Ferreira, 2019). More importantly, the planning of the verb early did not mean that the modifier NP, spoon, is planned early too. We present two picture description experiments in which we manipulated the number of the modifier noun and the type of intransitive verbs in sentences like The doctor(s) by the wizard is running/boiling. We argue that if the agreement related features are planned early and eagerly, we should not see agreement errors with unaccusative sentences, as the verb is planned before the modifier noun. However, if the agreement is planned later, not when the verb is retrieved, we should see agreement errors with both unaccusative verbs and unergative verbs , which has their subjects as their modifier and not their complement. Both experiments (N=74, N=59) showed that the agreement errors as well as agreement related timing effects are comparable in both unaccusative and unergative verbs. However, our onset latencies, i.e. sentence initiation measure, consistently showed that the number of the second noun affected the planning of the verb early on, but only in unaccusatives. We take our results as evidence that the agreement is planned late given the comparable patterns of errors. However, this error patterns might be related to other factors than the assignment of the number feature on the verb, such as retrieval of the correct auxiliary form.