Eye Movements and Cognitive Processes in Reading, Visual Search, and Scene Perception

Keith Rayner    Department of Psychology, University of Massachusetts, Amherst, MA 01003 USA

Introduction

During the past twenty years, a dichotomy has developed with respect to research on eye movements and reading. Namely, two different types of research have been undertaken. One type of research has focused on the characteristics of eye movements per se during reading, with the primary topic of interest being eye movement control. The second type of research has utilized eye movement data to infer something about the process of reading. The characteristic of this latter type of work is that the researcher seems to be only incidentally interested in eye movements per se, and primarily interested in what eye movements can tell us about reading. One interesting aspect to this dichotomy is that researchers in each of these groups don’t often seem to have much tolerance for the other group. As I just noted, those interested primarily in using eye movements to study reading seem to have little interest in the details of eye movements (such as where the eye lands, where it came from, saccade latency, etc), whereas those interested in eye movements per se seem to be constantly worried that effects that those interested in language processing obtain are somehow artifacts associated with eye movements. My own belief, and part of what I want to argue in this chapter is that advances are being made by both types of research group and that workers in each group should pay more attention to the other group.

To a large extent, the preceding paragraph is a caricature, and like most caricatures, it undoubtedly oversimplifies the true state of affairs. There are clearly a number of people interested in both aspects of the dichotomy I proposed. Indeed, in my own laboratory, we have engaged in both types of research. One interesting sociological phenomenon that I have sensed is that those at the extreme ends of the dichotomy typically perceive our work as being part of the other group. Thus, even though we have done a considerable amount of research on where the eyes land in words, saccade latency, and eye movement control, some researchers primarily interested in the oculomotor aspects of reading typically seem to think of us as working primarily on language processing, while those interested primarily in language processing often seem to think that we work primarily on oculomotor aspects of reading. As I hope to document in this chapter, considerable advances have been made on both fronts. My goal is to review some recent findings regarding eye movements in reading. I will primarily describe findings where there seems to be some agreement, but will not invariably do this. I will also note some instances where controversy exists. Near the end of the chapter, I will come back to the issue I began with (the dichotomy of research interests) and present some relevant data to discriminate between different types of models of eye movement control in reading. Finally, although the title of the chapter lists scene perception and visual search, I will focus for the most part on research on reading, but discuss those topics at the end of the chapter.

I will begin by discussing 5 issues: (1) the span of effective vision, (2) integration of information across eye movements, (3) eye movement control: where to fixate, (4) eye movement control: when to move, and (5) models of eye movement control. For the first four issues, I will primarily review prior research and make what seem to me to be appropriate conclusions. Since, to my mind, the most controversy exists concerning the last two issues, I will discuss some recent preliminary results that we have obtained which bear on the controversy. Prior to discussing the five issues, I will first provide a brief overview of some basic facts about eye movements and reading (see Rayner & Pollatsek, 1987 for more details).

Basic Facts about Eye Movements and Reading

During reading, we make a series of eye movements (referred to as saccades) separated by periods of time when the eyes are relatively still (referred to as fixations). The typical saccade is about eight to nine letter spaces; this value is not affected by the size of the print as long as it is not too small or too large (Morrison & Rayner, 1981). The appropriate metric to use when discussing eye movements therefore is letter spaces, and not visual angle (generally, 3-4 letter spaces is equivalent to 1 degree of visual angle). Because of the high velocity of the saccade, no useful information is acquired while the eyes are moving; readers only acquire information from the text during fixations (Wolverton & Zola, 1983). The average fixation duration is 200-250 ms. The other primary characteristic of eye movements is that about 10-15% of the time readers move their eyes back in the text (referred to as regressions) to look at material that has already been read.

Eye movements during reading are necessary because of the acuity limitations in the visual system. A line of text extending around the fixation point can be divided into three regions: foveal, parafoveal, and peripheral. In the foveal region (extending 1 degree of visual angle to the left and right of fixation), acuity is sharpest and the letters can be easily resolved. In the parafoveal region (extending to 5 degrees of visual angle on either side of fixation) and the peripheral region (everything on the line beyond the parafoveal region), acuity drops off markedly so that our ability to identify letters is not very good even in the near parafovea. The purpose of eye movements in reading is therefore to place the foveal region on that part of the text to be processed next.

It is important to note that as text difficulty increases, fixation durations increase, saccade lengths decrease, and regression frequency increases. More importantly, the values presented above for fixation duration, saccade length, and regression frequency are averages and there is considerable variability in all of the measures. Thus, although the average fixation duration might be 250 ms and the average saccade length might be 8 letter spaces for a given reader, for others these values might be somewhat higher or lower. This between reader variability (which also exists for regression frequency) is perhaps not as important as the fact that there is considerable within reader variability. In other words, although a reader’s average fixation duration is 250 ms, the range can be from under 100 ms to over 500 ms within a passage of text. Likewise, the variability in saccade length can range from 1 letter space to over 15 letter spaces (though such long saccades typically follow regressions). The reasons for this variability will hopefully become clearer later in this chapter.

The Span of Effective Vision

How much information does a reader acquire on each fixation?

This basic question has inspired a great deal of research in my laboratory and other labs. In order to investigate this question, George McConkie and I developed what has become known as the eye-contingent display change paradigm. Around the same time that we developed the technique, Steve Reder and Kevin O’Regan were also working on developing the technique. In this paradigm, a reader’s eye movements are monitored (generally every millisecond) by a highly accurate eye-tracking system. The eyetracker is interfaced with a computer which controls the display monitor from which the subject reads. The monitor has a rapidly decaying phosphor and changes in the text are made contingent on the location of the reader’s eyes. Generally, the display changes are made during saccades and the reader is not consciously aware of the changes.

There are three primary types of eye-contingent paradigms: the moving window, foveal mask, and boundary techniques. With the moving window technique (McConkie & Rayner, 1975), on each fixation a portion of the text around the reader’s fixation is available to the reader. However, outside of this window area, the text is replaced by other letters, or by xs (see Figure 1). When the reader moves his or her eyes, the window moves with the eyes. Thus, wherever the reader looks, there is readable text within the window and altered text outside the window. The rationale of the...