Simulation and information

In the last few years, off-line simulation has become an increasingly important alternative to standard explanations in cognitive science. The contemporary debate began with Gordon (1986) and Goldman's (1989) off-line simulation account of our capacity to predict behaviour. On their view, in predicting people's behaviour we take our own decision-making system ‘off line’ and supply it with the ‘pretend’ beliefs and desires of the person whose behaviour we are trying to predict; we then let the decision maker reach a decision on the basis of these pretend inputs. Figure 4.1 offers a ‘boxological’ version of the off-line simulation theory of behaviour prediction.

The off-line simulation theory of behaviour prediction is a radical departure from the typical explanations of cognitive capacities. In explaining a capacity in some domain (e.g., our ability to solve mathematical problems), the usual strategy in cognitive science is to suppose that the subject has a body of information about that domain (see, e.g., Fodor, 1968a). For example, our ability to predict the motion of projectiles is thought to depend on a body of information about mechanics – a folk physics. Of course, much if not all of these information bases or theories may be tacit or ‘sub-doxastic’ (Stich, 1978). Further, different theorists have different ideas about how the information is encoded.