Philosophers Index Abstract
- We suggest that much of the debate over simulation can be clarified by distinguishing various versions of the "theory theory" and simulation theory.
- We argue that merely showing that one of these versions of the theory theory is wrong wouldn't show that simulation theory is right. Further, we maintain that the different version of simulation theory must be assessed independently.
- We argue that it's still implausible to think that we predict behavior by means of "off-line" simulation. However, when we predict the grammaticality judgments or inferences that others would make, it's possible that we do use a type of simulation.
- In their earlier paper, "Stich (Stephen) & Nichols (Shaun) - Folk Psychology: Simulation or Tacit Theory" (1992) Stephen Stich and Shaun Nichols take great care to set up the debate between the two theories so that the fundamental issues can be brought out in the clearest and most empirically tractable way. One of their key claims about the theory theory is that the body of knowledge that underpins our folk-psychological practice must be implemented separately from any theory that is part and parcel of our decision-making system; otherwise the simulation theory wins (1995, p. 154, n. 7). In their new paper, 'Second Thoughts on Simulation' (chapter 5), they take up an argument put forward by "Harris (Paul L.) - From Simulation to Folk Psychology: The Case for Development" (1992/1995) to the effect that this requirement of separate implementation has the implausible consequence that we must postulate duplicate2 first-person and third-person systems. Harris argues, for example, that the requirement would force us to posit two separate systems of linguistic knowledge - one used in making judgements about our native language, and a second used for the prediction of the grammaticality judgements that our fellow native speakers would make (see also "Heal (Jane) - How to Think About Thinking", chapter 2). Stich and Nichols concede Harris's point about this particular example. But, they point out that the example concerns predictions of judgement and belief, rather than predictions of behaviour. So far as the prediction of behaviour is concerned, they use examples of what they call cognitive penetrability to argue for the theory theory, and against the simulation theory (pp. 99-100):
According to the theory theory, predictions about people's behaviour are guided by a rich body of mentally represented information (or misinformation) about the ways in which psychological states are related to environmental stimuli, other psychological states, and behavioural events. If that information is wrong or incomplete in various areas, then we should expect the accuracy of predictions in those areas to decline. According to the off-line simulation theory, we generate predictions of people's behaviour by running our own decision-making system off-line. If we are ignorant about how people's minds work, or if we have mistaken views, this should not affect the accuracy of our predictions about how people will behave, since our views about how the mind works are not involved in generating the predictions. So if the off-line simulation theory is right, what we don't know won't hurt us - predictions about people's behaviour are 'cognitively impenetrable'.
- The increasing complexity of the debate, and the departure from a simple theory theory versus simulation theory opposition, is illustrated once again here. Stich and Nichols, who would be labelled as theory theorists, are prepared to concede to the simulation theory the case of predicting a person's beliefs from information about her situation. Heal, who would be labelled as a simulation theorist, argues that this same case should be substantially conceded to the theory theory.
Footnote 1: Taken from "Stone (Tony) & Davies (Martin) - Mental Simulation: Introduction".
Text Colour Conventions (see disclaimer)
- Blue: Text by me; © Theo Todman, 2020
- Mauve: Text by correspondent(s) or other author(s); © the author(s)