Mental extrapolation in patterns constructed from memory

Abstract

Finke and Pinker (1982, 1983) showed subjects an array of dots followed by an arrow in a blank field, and asked them to determine whether the arrow pointed to any of the previously seen dots. Response times were linearly related to the distance between the arrow and the nearest dot, suggesting that subjects spontaneously used an internal scanning or extrapolation process to perform the task. We replicate and extend this finding by varying the retention interval, and by employing a paradigm in which subjects' eyes are closed and the arrows are described to them using a coordinate scheme. We also show that subjects are unable to predict the form of the data when the task simply is described to them. Results suggest that mental scanning along a straight path can be performed on images reconstructed from memory, and that it does not depend on the ongoing perception of a continuous surface, on physical eye movements, or on demand characteristics.In recent years, the phenomenon of mental image scanning has come to occupy a central position in the empirical and theoretical study of imagery. Kosslyn (1973) and Kosslyn, Ball, and Reiser (1978) showed that image scanning can be studied with considerable precision using chronometric methods. When subjects are asked to scan or shift attention between locations in an imagined scene, the time they require varies linearly with the Euclidean distances between the corresponding locations in the actual scene (r = .97 in Kosslyn, et al., 1978). Since then, data from imagery-scanning experiments have been used in three ways. First, the phenomenon itself has played an important role in debates over the format of image representation, with Kosslyn and his collaborators arguing that scanning implicates an

Is time to scan visual images due to demand characteristics?

“…The time to decide whether the arrow would point at a dot not only increased linearly with the distance to a dot, but the rate of increase was almost identical to that found in image-scanning studies. This result was replicated by Pinker, Choate, and Finke (1984), who had subjects form their images from long-term memory. Again, these results are very difficult to explain by appeal to the task-demands argument.…”

Different cognitive processes in two image-scanning paradigms

“…The results could not be explained by experimenter expectancy effects or task demands. These findings were confirmed in subsequent experiments (Finke & Pinker, 1983;Pinker, Choate, & Finke, 1984) and were taken as strong support for the claim that scanning reflects the spatial structure of image representations.…”

Analog properties of cognitive maps constructed from verbal descriptions

“…Imagery researchers generally interpret this finding as reflecting the structural isomorphism of visual images to the configurations they represent (cf. Finke, 1989;Finke & Pinker, 1982Kosslyn, Ball, & Reiser, 1978;Pinker, Choate, & Finke, 1984;Reed, Hock, & Lockhead, 1983). If a time-distance relationship is obtained in a mental-scanning task involving text-based representations, this would support the claim that cognitive maps constructed from descriptions possess properties that make them genuinely isomorphic to the described configurations.…”