A subtle and often unrecognized assumption lies behind the use of the terms salience and perception when they appear in the context of human information processing. Although it has become commonplace to blend the language of low-level physiological process with the language of high-level cognition, there is a substantive theoretical step to be taken before we can claim that, for instance, the ability to notice a blinking light or to pick out a single voice in the crowded room is somehow continuous with the ability to sense the attitude of an airplane from only instrument readings or see a changing battlefield dynamic given the visual clutter of the plan view display of a common operating picture. A significant research opportunity exists to make the relationship between low-level perception and high-level information processing explicit and quantifiable. To do this we propose to reconcile a leading-edge theory of human perception in complex work environments with the process-level mechanisms of the state-of-the art cognitive modeling architecture. Specifically, we will establish a formal mapping between the constructs of Wickens SEEV statistical model of perception and select generative structures of ACT-R, an executable, computational theory of cognition.
Keywords: Information Value; Ontologies; Semantic Web; Decision Support
