Semantic Modulation of Temporal Attention: Distributed Control and Levels of Abstraction in Computational Modelling

This paper presents a computational exploration of how meaning modulates temporal attention. This has been explored within the context of Barnard et al's key-distractor Attentional Blink (AB) and proposes a mechanism by which attention is captured by salient meaning. We also explore the applicability of formal methods to the abstract modelling of cognition. Such formal methods enable modelling of an important class of cognitive architectures in which multiple subsystems interact and control is distributed. Most significantly, we have used the formal methods framework to explore the problem of how to relate different levels of abstraction / description in cognitive modelling. Our modelling of Barnard's ICS (Interacting Cognitive Subsystems) architecture illustrates this approach. A major result of our work is the simulation of the key-distractor AB. This simulation is formulated in terms of the interaction between ICS' two central subsystems, which extract implicational and propositional meaning, respectively. Items related to a target category may be interpreted as implicationally salient, even though a later propositional check reveals they are not targets. Such items capture attention, creating a window of time in which the system is vulnerable to missing actual targets. We also explore how changes in semantic salience modulate the key-distractor AB, thereby clarifying the temporal properties of attentional capture by semantic salience.