Eye Movements and the Modulation of Parafoveal Processing by Foveal Processing Difficulty: A Reexamination
White, Sarah J., Rayner, Keith, Liversedge, Simon P., Psychonomic Bulletin & Review
Henderson and Ferreira (1990) found that foveal load (manipulated via word frequency) modulates parafoveal processing, thereby affecting the amount of preview benefit obtained from the word to the right of fixation. The present experiment used the eye-contingent boundary paradigm and, consistent with Henderson and Ferreira, showed that foveal load modulated preview benefit for participants who were not aware of the display changes during reading. Also, for these participants, foveal load modulated preview benefit regardless of fixation durations on the foveal word. For participants who were aware of the display change, preview benefits occurred regardless of foveal processing difficulty. These results have important implications for understanding the way in which foveal load influences parafoveal processing during reading.
During reading, information is extracted from each fixated word and from the word to the right of fixation. A fundamental issue for understanding eye movement control in reading is whether the processing of nonfixated text remains constant, or whether it is influenced by foveal processing load (Liversedge & Findlay, 2000; Rayner, 1998; Starr & Rayner, 2001). The results of an important experiment by Henderson and Ferreira ( 1990) suggest that parafoveal processing is modulated by foveal processing load. This notion is a major constraint for the architecture of the E-Z reader model of eye movement control (Reichle, Pollatsek, Fisher, & Rayner, 1998; Reichle, Rayner, & Pollatsek, 1999, 2003). Other research suggests that a similar phenomenon may occur in tasks other than reading as well (e.g., Mackworth, 1965; Williams, 1988).
In the present study, we used a manipulation similar to that of Henderson and Ferreira ( 1990) in order to address three important issues. First, the modulation of parafoveal processing by foveal load is so central to our understanding of eye movement control in reading that it is crucial to establish that this phenomenon is reliable. second, Henderson and Ferreira excluded trials in which participants detected saccade-contingent display changes, whereas we tested whether foveal load also modulates parafoveal processing for participants who detect the changes. Third, Schroyens, Vitu, Brysbaert, and d'Ydewalle (1999) suggested that the effects of foveal load on parafoveal processing are due to spillover processing following short previous fixations. In the present experiment, we also tested whether foveal load modulates preview benefit regardless of preceding fixation durations.
Studies have shown that the processing of a parafoveal word facilitates the processing of that word when it is subsequently fixated. This preview benefit (Rayner & Pollatsek, 1989) is measured using the saccade-contingent boundary change technique (Rayner, 1975). This technique involves changing the preview (which may be correct or incorrect) when the eye crosses an invisible boundary prior to fixation such that the word is correct when it is fixated. The difference in reading times between when the preview is correct and when it is incorrect provides the preview benefit (which measures the extent to which the processing of the correct preview facilitates processing once the word is fixated).
Henderson and Ferreira (1990) used the boundary paradigm to show that foveal processing difficulty reduces parafoveal processing. They compared reading times on critical target words (e.g., despite] when the preview of that word was correct (despite) and when it was incorrect (zqdioyv) and when the word prior to the critical word was frequent (chest) and when it was infrequent (trunk). They showed that the preview benefit for the critical target word was larger when the previous word was frequent than when it was infrequent.1 That is, the preview benefit was smaller when foveal processing was difficult than when it was easy. Two other studies yielded similar results (Kennison & Clifton, 1995; Schroyens et al. …