Representation, Attention, and Perceptual Learning

What sorts of properties we perceive matters for understanding the nature of perception and the scope of perceptual justification. One way of arguing for the view that we can represent ‘high-level’ properties such as natural and artificial kinds is to appeal to Susanna Siegel’s method of phenomenal contrast, which contrasts the phenomenology of experts and novices. This argument can be strengthened by appealing to empirical work on perceptual learning and expertise that suggests the world really does look different to experts after training. Moreover, this appeal can diffuse several lines of objection to Siegel’s argument. Here I discuss two recent cases: Fred Dretske’s Goldilocks test, and Kevin Connolly’s attention shift argument. In both cases, a proper understanding of perceptual learning, and the role attention plays in such learning, explains how high-level properties can come to be represented in perceptual experience. Learned attentional patterns are themselves partially constitutive of representing high-level properties in perceptual experience.