Seeing and Windows of Integration

I am grateful to Bradley Richards and J. H. Taylor for their thoughtful critiques and for the chance to clarify and re-think the main line of argument in ‘‘The Grain of Vision and the Grain of Attention’’ (Block 2013). My article concerned peripheral vision. Much of normal vision is peripheral. The fovea is the central area of the retina that is needed for fine discrimination, for example in reading. It subtends only 2◦ of visual angle, a bit more than the width of the thumbnail at arm’s length, so if you are looking at something the size of a hand at arm’s length, on any one fixation most of it is seen nonfoveally. All perception degrades in acuity in the periphery because cone cells in the retina decrease with eccentricity. But object-perception in the periphery suffers from more than a degradation in acuity; it suffers from crowding. Crowding is a phenomenon of peripheral vision in which ‘‘things . . . lose the quality of form . . . without losing crispness . . .’’ (Lettvin 1976). It is widely agreed that the explanation of crowding is that vision involves assigning features to objects; but there are minimal ‘‘windows of integration’’ within which the visual system cannot determine which features are to be bound to which objects. (Binding is the process by which the visual registration of a blue square and a red circle involves blueness being attributed to the square instead of the circle.) The windows of integration grow larger farther from the fixation point—that is, where the eyes are pointing. See Figure 1 for an indication of the size and shape of these windows. In the periphery, the windows can be large enough so that normally two or more objects are in the same window. I argued that there is conscious object-seeing without object-attention in a phenomenon that I called identity-crowding. Identity-crowding is the special case of crowding in which the crowded items are all the same so that there is no issue of determining which features are bound to which objects. See Figure 2 for an example. As Taylor and Richards note, my argument was that in identity-crowding, the subject consciously detects the crowded object (in the sense of distinguishing consciously between presence and absence), consciously differentiates the object from the background, consciously discriminates the crowded object from other objects and consciously identifies the crowded object, so it is difficult to see a rationale for denying that