Position and Identity in Visual Information Processing

Abstract

Information that is presented visually can be described in terms of its identity and in terms of its position, and a distinction can be drawn between what an item is and where that item is. For example, a letter displayed on a screen has both an identity (its name) and a spatial position; the spatial position can be specified either absolutely (the upper right quadrant) or relatively (beside the "x" and above the "y"). There is an obvious and intimate relationship between the identity component and the position component, and it is this relationship, between the the processing of position information and the processing of identity information, that forms the subject of the present thesis. First, the relevant I literature is reviewed. The relationship between position and identity is examined in the context of two major research areas: iconic memory and short term visual memory. Second, the concept of dimensional separability is considered with reference to the appropriate literature. The purpose is to indicate a theoretical framework within which the issue of concern may be profitably addressed. The key idea to be developed is that position and identity are asymmetrically separable dimensions. A small group of studies that offer tentative support to this conceptualization will be discussed. Third, the results of eight related experiments are reported. These experiments involve the recognition of position and/or identity information in a discrete trials procedure. The eight experiments fall into three separate groups. Experiments 1 to 3 examine the recognition of either position or identity information, with the two types of information presented in relative isolation. Several stimulus factors are manipulated in order to demonstrate differential effects upon the two dimensions. Experiments 4 to 6 examine the effect of the irrelevant dimension upon recognition of the relevant dimension. Experiment 4 uses a logically balanced set of stimuli, so that the irrelevant dimension is either consistent or inconsistent, whereas in Experiments 5 and 6 each dimension is examined in the context of consistent, inconsistent, or neutral information on the irrelevant dimension. Experiments 7 and 8 explore the integration of position and identity information by varying the task requirements. Shared attention conditions are contrasted with selective attention conditions to show the impact of attentional strategy. The thesis concludes with a general discussion of the results, and their accordance with the hypothesis of asymmetric separability.