Temporal dynamics of the eye–voice span and eye movement control during oral reading

Abstract

The distance between eye movements and articulation during oral reading, commonly referred to as the eye–voice span, has been a classic issue of experimental reading research since Buswell (1921). To examine the influence of the span on eye movement control, synchronised recordings of eye position and speech production were obtained during fluent oral reading. The viewing of a word almost always preceded its articulation, and the interval between the onset of a word's fixation and the onset of its articulation was approximately 500 ms. The identification and articulation of a word were closely coupled, and the fixation–speech interval was regulated through immediate adjustments of word viewing duration, unless the interval was relatively long. In this case, the lag between identification and articulation was often reduced through a regression that moved the eyes back in the text. These results indicate that models of eye movement control during oral reading need to include a mechanism that maintains a close linkage between the identification and articulation of words through continuous oculomotor adjustments.

Keywords:

• Eye movement control
• Eye–voice Span
• Fixation–speech interval
• Oral reading
• Word identification

Acknowledgements

This research was support by NICHD Grant No. HD043405. We would like to thank Poujan Rahmani, Marika Inhoff, and Melissa Zeidler for their help with the analyses of speech data.

Notes

¹Outgoing saccades of less than 1 LS and more than 15 LS were also removed from the analysis of outgoing saccade sizes.

²According to Rayner and Pollatsek (1989), gaze durations during silent reading are typically between 250 ms and 275 ms, and saccade sizes range between 6 and 9 LS. The distribution of gaze durations in the current study is similar in shape to the distribution of gaze durations for silent reading (Rayner, 1998).