ABSTRACT
This chapter provides evidence that subject behavior can be under the control of hierarchical goal structures, as predicted by the ACT-R theory. The timing data confirmed specific predictions of the way goals are managed in the ACT-R architecture. With appropriate instructions, however, one can more or fewer guarantees that subjects will use goal structures to solve the Tower of Hanoi problem. The suspicious reader might wonder to what degree the subgoaling behavior that was observed simply reflected the logical structure of the Tower of Hanoi task. One way to address this question is to look at tasks that have the same logical structure as the Tower of Hanoi task, but where that structure is not so obvious. If one is to argue that subjects have the natural wherewithal to simulate goal stacks given just dependency information, then it is unclear what prediction would separate the proposal of simulated goal stacks from real goal stacks.
