Parafoveal preview benefit in unspaced and spaced Chinese reading

Abstract

In an eye movement experiment during reading, we compared parafoveal preview benefit during the reading of Chinese sentences either in the familiar, unspaced format or with spaces inserted between the words. Single-character words or the first of a two-character word were either presented normally or were replaced by a pseudocharacter in the preview. Results indicate that word spacing increased the parafoveal preview benefit but only for the one-character target words. We hypothesized that the incorrect preview of the first character of the two-character word prevented parafoveal processing of the ensuing character(s), effectively nullifying any benefits from the spacing. Our results suggest that word boundary demarcation allows for more precise focusing of attention.

Keywords:

• Eye movements
• Reading
• Parafoveal processing
• Chinese reading

The authors acknowledge support from the China Scholarship Council.

The research described in this article along with the write-up were supported by: the National Science Foundation of China [grant number 31200765]; the Natural Science Foundation of Shandong [grant number ZR2012CQ034] and Construction of Shandong Province applied basic type School Project for the first author; Key Research Topics of the Ministry of Education Award for the 11th Five-Year Plan of National Science of Education [grant number DBA090290] for the third author; the Social Science Foundation of China [grant number 10BYY029] for the fourth author; Economic and Social Research Council (ESRC) awards [grant number RES-000-22-4128], [grant number ES/I032398/1] and a Leverhulme Trust Grant [grant number F/00 180/AN] to the fifth author.

Notes

¹Note that there is also a productive research line examining whether (successful) word segmentation is an important determiner of the initial landing site in a word (e.g., Li, Liu, & Rayner, 2011; Yan, Kliegl, Richter, Nuthmann, & Shu, 2010).

²The somewhat surprising longer first-fixation duration on the one-character word (mean = 233 ms) than on the two-character word (mean = 224 ms) in the unspaced condition with the identical preview [t ₁(95) = 1.94, p = .06; t ₁(95) = 2.15, p < .05] can be explained by the number of fixations (respectively, 1.44 and 2.09 fixations). The first of two fixations on a word is usually shorter than a single fixation on a word, leading to shorter first-fixation durations on the two-character word (e.g., Rayner, Sereno, & Raney, 1996).