Ori Friedman and Alan Leslie, in their Cognition article “The conceptual underpinnings of pretense: Pretending is not ‘behaving-as-if’” (2007), argue against what they call behaviorist theories of pretense. As they characterize the debate, theories of pretense fall into the following two families:
Metarepresentational Theories: What is central to pretense is mentalistic: treating pretense as such through the possession and deployment of the concept PRETEND. (Leslie is the primary proponent of this view.)
Behaviorist Theories: What is central to pretense is behavioral: behaving “as-if” a scenario obtains–or rather, behaving in a way that would be appropriate if that scenario obtains. (On Friedman and Leslie’s characterization, pretty much everyone else–such as Perner, Lillard, Nichols and Stich, Harris, and Rakoczy–defends some kind of behaviorist theory.)
One central point in they make the paper is this: without making room for the concept PRETEND in their theories, behaviorists cannot adequately explain how children are able to recognize pretense as such. Here is how they put this central point on page 115:
But more importantly, the above makes clear what game the Behavioral theory perforce finds itself playing: namely, trying to get the child to think that someone is pretending without actually thinking pretend as such. If the Behavioral theory is to measure up to the phenomena of early human pretending, its success will depend on finding an exact conceptual paraphrase of PRETEND without using that concept. Moreover, the paraphrase must be strictly behavioral. … Propositional attitude concepts are the heart and soul of ‘theory of mind’ and utterly foreign to and rejected by behaviorism (Fodor, 1981; Ryle, 1949). ‘Pretend’ is just the name of a specific attitude.
In response, I have two small worries and a major one: