Eye movements during reading proverbs and regular sentences: the incoming word predictability effect

Abstract

Reading is an everyday activity requiring the efficient integration of several central cognitive subsystems ranging from attention and oculomotor control to word identification and language comprehension. Effects of frequency, length and cloze predictability of words on reading times reliably indicate local processing difficulty of fixated words; also, a reader's expectation about an upcoming word apparently influences fixation duration even before the eyes reach this word. Moreover, this effect has been reported as non-canonical (i.e., longer fixation durations on word N when word N+1 is of high cloze predictability). However, this effect is difficult to observe because in natural sentences the fluctuations in predictability in content words are very small. To overcome this difficulty we investigated eye movements while reading proverbs as well as sentences constructed for high- and low-average cloze predictability. We also determined for each sentence a word at which predictability of words jumps from a low to high value. Fixation durations while reading proverbs and high-predictable sentences exhibited significant effects of the change in predictability along the sentence (when the successive word is more predictable than the fixated word). Results are in agreement with the proposal that cloze predictability of upcoming words exerts an influence on fixation durations via memory retrieval.

Keywords:

• eye movements
• reading
• proverbs
• incoming word predictability effect

Please note that this paper was previously published online in a different page layout.

Please note that this paper was previously published online in a different page layout.

Acknowledgements

Diego Shalom and Mariano Sigman were supported by Argentine National Research Council (CONICET). Gerardo Fernández was supported by the ANPCYT, PICT 2010 1421. Gerardo Fernández and Diego E. Shalom contributed equally to this work.

Notes

1. While there are studies reporting absence of evidence for parafoveal-on-foveal effects, we do not agree that these studies challenge the results of studies providing positive evidence for parafoveal-on-foveal effects. Obviously, the necessary conditions for parafoveal-on-foveal effects are not completely identified yet.

2. By now, positive predictability effects of word N+1 have also been presented, but not yet published, for reading English as well as simplified and traditional Chinese sentences (Kliegl, 2012).