Optimal Viewing Position in Words and the Strategy-Tactics Theory of Eye Movements in Reading

Work on eye movements in reading received a considerable new impetus starting in about 1975, when several research groups began using computers to control eye movement experiments “on-line” (McConkie & Rayner, 1975; O’Regan, 1979; Reder, 1973). The idea of a “moving window” allowing only a limited amount of text to be visible around the instantaneous eye position was an appealing technique that allowed researchers to measure the size of the “perceptual span,” that is, the zone around the fixation point from which different kinds of information is gathered. My own feeling at that time, and I believe also the feeling of other workers, was that the most reasonable thing that the eye might be doing in reading would be to adjust its saccades so that from moment to moment, the edges of successive perceptual spans would just touch, or overlap by a fixed amount. In that way, at moments when the text is easy to predict, or the words easy to recognize, the eye would be able to make larger saccades than when the eye was moving into less predictable or more difficult territory. This view of reading could be called the “perceptual span control” hypothesis (Jacobs & O’Regan, 1987). The perceptual span control hypothesis is at the root of a very large body of research in which perceptual span is measured by making perturbations in the information available in parafoveal vision (for a review see Rayner & Pollatsek, 1987): in this research, changes in eye movements are assumed to indicate changes in perceptual span. However, it has not usually been realized that this assumption can only be true to the extent that perceptual span actually directly determines eye movements.