Before discussing the relative merits of simulation and theory-theory we need to establish principled grounds for distinguishing the use of simulation from the use of a theory in order to avoid collapse between these two positions. My proposal is that simulation be characterised as the use of predication-implicit knowledge about the mind. On the basis of this characterisation I conclude that every use of a theory of mind involves an element of simulation, since our folk theory typically exploits predication-implicit knowledge about the content domain. However, I then go on to argue on empirical grounds that simulation cannot be the only ingredient in how we make mental attributions. The only viable position is a simulation-theory mix.

Threat of collapse: a brief review

Our folk psychology provides us with the notion of role- or perspective taking. So we thought we knew what simulation (Gordon, 1986, or ‘replication’. Heal, 1986) was supposed to be and how it differed from mental state attributions on the basis of a theory. The difference can be easily explained with simple examples. More concerned about urban safety than Grizzly bears my favourite example involves walking through a dark alley and noticing a seedy looking character closing in from behind. How does one feel? How will one react? Answering it by theorising one draws on knowledge that people are afraid of being mugged in dark alleys by seedy characters.