Scanning visual mental images: some structural implications, revisited

People talk about many things, and most of the time conversations are about other people, objects, or events that are not directly perceivable. This is not true for spatial language use. An important characteristic of spatial communication is that it often relies on a tight coupling between linguistic and visual information. An utterance such as "Your car keys are on the coffee table" could be seen as a linguistic starting point or an imperative from one person, which will result in a visual search by another person. This coupling has been captured in a detailed and rigorous fashion by Carlson and Logan in a theoretical framework, experimental work, and a computational model (e.g., Carlson, 1999; Carlson-Radvansky & Irwin, 1994; Logan & Compton, 1996; Logan & Sadler, 1996; Regier & Carlson, 2001). To describe this work in detail is beyond the scope of this chapter, but it is important to note that it predicts how visual attention is directed over a visual scene to establish a link between a conceptual representation (of the spatial sentence) and a perceptual representation (of the visual scene). However, these ideas become hard to apply in a situation where a spatial sentence is encountered without accompanying visual information. This could happen when a person hears a route description of a town that he or she is about to visit in the near future, or when someone is incapable of receiving visual information (i.e., when this person is blind). The question of how sighted and blind people represent and memorize these "isolated" spatial descriptions is the theme of this chapter.