Silvia Gennari & David Poeppel
University of Maryland
A long-standing debate in linguistics and psycholinguistics concerns how lexical meanings are internally represented. Lexical semanticists (e.g., [1][2][3]) claim that lexical meanings have internal structure that may differ in complexity across word types. In contrast, the atomistic view claims both on theoretical and experimental grounds that word meanings are atomic and not internally structured [4][5][6][7][8]. In this paper, I explore this issue by investigating how verbs of different complexity are processed and represented.
State- and event-denoting verbs are the most general verb types. Event-denoting verbs such as "kill" have internal structure because they denote (and entail) a change from an initial state to a resulting one (kill = being alive → being dead). In contrast, state-denoting verbs such as "contain" lack such structure. They simply denote facts, stable relations between their participants. Semantic theories explain this distinction by proposing lexical representations that differ in internal complexity: events are composed of multiple states, whereas states are not.
This difference in the representation of multiple vs. single states suggests that each verb type may involve differential processing cost depending on internal complexity; state verbs may be processed faster than event verbs.
To test this, we conducted a visual lexical decision task (n=50; 32 verbs per category). Event and state verbs were matched for frequency, word length and argument structure. Imageability ratings were collected independently and used as a covariate. Analysis of RTs (with verb-type/imageability as factors) revealed a main effect of verb types (p=.001 across subjects and items) but no interaction with imageability. As predicted, state verbs were processed faster than event ones (30 ms. difference).
We also conducted a self-paced reading study (n=30, 45 verbs per category in sentential contexts). Verb types were pair-wise matched by several criteria known to affect processing time: frequency and word length; number of syntactic frames; number of syntactic arguments; and preceding sentential context. An analysis of RTs at the verb position revealed a main effect of verb type (p= .01) (30 ms. difference). As before, event verbs engendered more processing cost than state verbs. Because argument structure was held constant, the effects cannot be due to structure building information.
This converging evidence suggests that event and state verbs are differentially processed and represented. Semantic properties of each type modulate lexical access and integration. This is consistent with theories that postulate differences in verb representational complexity and challenges the Fodorian unstructured approach to the representation of word meaning.
References
[1] Dowty, D. (1979). Word Meaning and Montague Grammar. Reidel, Dordrecht.
[2] Krifka, M. (1989). Nominal reference, temporal constitution and quantification in event semantics. In Bartsch et al., Semantics and Contextual Expressions. Foris.
[3] Verkuyl, H. (1989). Aspectual classes and aspectual composition. Linguistics and Philosophy, 12, 39-94.
[4] Fodor, J. A. (1970). Three reasons for not deriving "kill" from "cause to die". Linguistic Inquiry, 1.4, 429-438.
[5] Fodor, J. D., Fodor, J. A., & Garrett, M. (1975). The psychological unreality of semantic representations. Linguistic Inquiry, 6.4, 515-535.
[6] Fodor, J. A., & Lepore, E. (1998). The emptiness of the lexicon: Reflections on J. Pustejovsky's "The Generative Lexicon". Linguistic Inquiry, 29(2).
[7] Fodor, J. A., Garrett, M. F., Walker, E. C. T., & Parkes C. H. (1980). Against definitions. Cognition, 8, 263-367.
[8] Rainer, K., & Duffy, S. (1986). Lexical complexity and fixation times in reading: Effects of word frequency, verb complexity, and lexical ambiguity. Memory & Cognition, 14(3).