How Attention can Alter Appearances

It has long been known that attention can change how things appear (e.g., Treisman, 2006). Traditional models of attention that focus on modulation of neuronal gain control, feature binding, or object tracking cannot easily account for the fact that attention can be voluntarily allocated in a way that markedly alters perceived features, including color, size, transparency, 3D shape/layout, orientation, and motion. Before delving into theory, we think it best that readers experience for themselves the problem that we feel calls for an explanation. To that end, we begin by presenting a broad range of examples where attention can change how a single input is experienced. Our goal is to present a theory of attention that can account for all of the effects summarized in the figures, as well as some new ones involving perceived colors in cases of overlapping transparent surfaces. Here, we suggest that attention alters perceived appearances (i.e., features or qualia) by defining the domain of automatic operations in the preconscious buffer—a window of time just prior to conscious experience. Gestalt psychologists (e.g., Koffka, 1935) noted that percepts can be grouped in various ways, and that this can lead to changes in subjective appearance. In addition, which grouping is experienced is to some extent under endogenous or volitional control. For example, one can see an “X” in Figure 12.1A or a “square” or any number of other groupings, depending on how one attends to the component ellipses. Since attention can be shifted endogenously, the groupings one experiences as a result of a change in the specification of figure-versus-ground relationships can also change endogenously. In Rubin’s well-known (1915/1958) face/vase example, a version of which is shown in Figure 12.1B, one can see either a vase or two profiles facing one another. Other examples in the literature suggest that attention does more than simply flip figure and ground. Attentional “flipping” can induce changes in