Most models of word recognition assume that lexical activation spreads to semantic associates. Gated semantic priming studies, which suggest that hearing just the beginning of a prime word activates a related target word (e.g., Marslen-Wilson, 1987) support this assumption. For example, when subjects are primed with the "silv" of "silver", lexical decisions on "bronze" are speeded. Although presenting gated stimuli makes it possible to gauge the activation of a word semantically related to the prime before the prime's offset, gated presentation may encourage subjects to place artificially heavy weight on phonological information available at the beginnings of words (Allopenna et al., 1998). Furthermore, it is difficult to obtain time course information, as only one temporal point can be tested on any given trial.
It has been argued (Allopenna et al., 1998 and Tanenhaus et al., 2000) that there is a direct relationship between eye movements and lexical activation. Thus, in theory, eye movements should reflect the activation of words semantically related to a given spoken word. In this study, we have found preliminary evidence that they do.
In experiment 1, 24 subjects were shown an array of four pictures and instructed to point at one of them (the "target"). The target (e.g., "key") was either presented with a semantically related item (e.g., "lock") and two unrelated items, or it was presented with three unrelated items. Subjects were more likely to fixate on an item semantically related to the target than an unrelated item in the same display. Examining the time course of subjects' fixations revealed that this pattern emerged about 570 ms after the onset of the target and continued until about 800 ms after the onset of the target. Allowing time (200 ms) to program and launch an eye movement suggests that semantically related items become active enough to draw fixations about 370 ms after the onset of the target (average target word duration = 517 ms).
In order to confirm that the results of experiment 1 reflected lexical processing, rather than strategic effects based on the presence of associated items in the visual display, experiment 2 asked whether eye movements would reflect the activation of an item semantically related to an unpictured phonological competitor of the target (i.e., would subjects look at "key" when asked to point at "logs", even if there were no "lock" in the display?). Pilot data suggests that subjects are more likely to fixate on an item semantically related to an unpictured phonological competitor of the target than at an unrelated item in the same display.
These results, from an independent paradigm, are consistent with the gated semantic priming data cited above. Furthermore, taken together with previous studies that used eye movements to examine spoken word recognition (e.g., Allopenna et al., 1998), they suggest that eye movement measures are particularly promising for experimental work that depends on documenting full time-course information of subtle lexical effects.
References
Allopenna, P.D., Magnuson, J., & Tanenhaus, M.K. (1998). Tracking the time course of spoken word recognition using eye movements: Evidence of continuous mapping models. Journal of Memory and Language, 38, 419-439.
Marslen-Wilson, W.D. (1987) Fuunctional Parallelism in spoken word-recognition. Cognition, 25, 71-102.
Moss, H.E., McCormick, S.F., & Tyler, L.K. (1997). The time course of activation of semantic information during spoken word recognition. Language and Cognitive Processes, 12(5/6) 695-731.
Tanenhaus, M.K., Dahan, D., Magnuson, J.S., & Hogan, E.M. (2000). Tracking the time course of subcategorical mismatches in lexical access in contiguous speech. Paper presented at the 13th Annual CUNY Conference on Sentence Processing.
Zwitserlood, P. (1989). The locus of the effects of sentential-semantic context in spoken-word processing. Cognition, 32, 25-64.