Montag,12.01. 09, 16 Uhr, Raum F 125

Georg Zeißner:  Modelling consciousness as the source of human flexibility in complex problem-solving

Modelling consciousness as the source of human flexibility in complex problem-solving Present cognitive architectures like ACT-R (Anderson, 1996) postulate that humans solve problems by retrieving adequate behavioral programs from their (procedural) knowledge and applying them to the task at hand. Apart from some quantitative and experience-dependent fine-tuning (Lovett & Anderson, 1996) the program is not changed in the process. In the announced talk an ongoing PhD project will be presented, which testifies one major shortcoming of the so-called „program-hypothesis“ and aims at amending this theoretical deficit. In our design the subjects were tasked with a complex turn-based problem in the form of running a farm in a simulated simplified eco-system. In order to obtain online data of the accompanying cognitive processes the subjects were asked to think aloud during the task. The resulting data of the problem solving process, although being interspersed with shorter episodes of unreflected „program-application“, is rich in changes of pace and strategy which we think that the aforementioned “program-hypothesis” fails to account for. This talk will present the basics of a theory that places consciousness (in the sense of „conscious meta-cognition“) at the source of the observed flexibility. To illustrate the principles of the theory an autonomous agent will be introduced, which models the problem-solving behavior of individual subjects on the same task.