Abstract
This paper presents a new research paradigm for analysing human learning in dynamic task environments based on the theory of finite-state automata. Some of the advantages of the approach are outlined. (1) It is possible to design classes of formally well-described dynamic task environments instead of idiosyncratic microworlds that are difficult if not impossible to compare. (2) The approach suggests assumptions about the mental representation of a discrete dynamic system. (3) The finite-state automata formalism suggests systematic and appropriate diagnostic procedures. (4) Using finite-state automata to design dynamic task environments, one does not have to give up the “ecological validity” appeal of computer-simulated scenarios.
An experiment on the utility of an external memory support system with system complexity and type of memory support as independent variables is reported to illustrate the application of the formal framework. Systematically derived dependent variables reflect both system knowledge and control performance. The results suggest that the benefits due to the availability of the external aid vary as a function of the complexity of the task. Also, using reaction time measurements, priming phenomena have been found that point to the importance of sequentiality in the representation of discrete systems. It is concluded that the approach, although not entirely new in experimental psychology, awaits further exploration in research on human learning in dynamic task environments and promises to be a stimulating paradigm for both basic and applied research.
