Ben Acland, Nicole Baggette, Henry Bley-Vroman, Natalie Klein, Elspeth Llewellyn,
Gabrielle Osborne, Jacob Stiglitz, Andrew Waser, Robert Thornton, & Martin Hackl
Pomona College
robert.thornton@pomona.edu
Tenny (1995) distinguishes between eventive verbs, which specify motion along a path (e.g. dart, wander), and stative verbs, which do not (e.g. squawk, sleep). Prepositional counterparts to these two types of verbs are the PATH functions (e.g. across, to), which specify motion along a path, and PLACE functions (e.g. inside, at), which locate an object's static position in relation to another's (Jackendoff, 1983). Thus, eventive verbs typically take PATH functions whereas stative verbs typically take PLACE functions. We investigated how the spatial semantics of these types of verbs and prepositions are integrated during online language processing. Specifically, we examined cases in which an eventive verb is juxtaposed with a PLACE function. We reason that there are two mechanisms by which this conflict can be reconciled. (i) Verb priority: the verb determines whether the phrase is describing an action taking place along a path or at a location. Thus, the meaning of the preposition changes to accommodate the path bias of an eventive verb (e.g., ``to protect her nest, the bird darted at the hunter just now''). (ii) Preposition priority: The path of the verb is omitted and the preposition does not change in meaning (e.g., ``because he woke up early, the child wandered at the school last Tuesday''). The main difference between these two is that in (i), the meaning of the preposition is coerced into a PATH reading, whereas in (ii), the path is simply omitted. One prediction of this account is that reading times for (i) relative to a baseline should be increased as a result of the coerced meaning (e.g., Traxler et al., 2002), whereas reading times for (ii) should not be higher than baseline. This prediction was confirmed using a self-paced reading task, using stimuli like the verb priority items in (1) and preposition priority items in (2).
For the verb priority items, both of the inconsistent items (1b & 1c) have significantly longer RTs than the consistent ones (1a & 1d), whereas for the preposition priority ones, only (2c) is slower. Moreover, this effect is evident on the word after the preposition, supporting a highly incremental view of the integration of spatial semantics. The results will be discussed in relatation to recent psycholinguistic accounts of enriched composition (McElree et al., 2001; Pinango et al., 1999; Traxler et al. 2002).