Folk psychological and phenomenological accounts of social perception

Abstract

Theory theory and simulation theory share the assumption that mental states are unobservable, such that mental state attribution requires an extra psychological step beyond perception. Phenomenologists deny this, contending that we can directly perceive people's mental states. Here I evaluate objections to theory theory and simulation theory as accounts of everyday social perception offered by Dan Zahavi and Shaun Gallagher. I agree that their phenomenological claims have bite at the personal level, distinguishing direct social perception from conscious theorizing and simulation. Their appeals to phenomenology and other arguments do not, however, rule out theory theory or simulation theory as accounts of the sub-personal processes underlying social perception. While I here remain uncommitted about the plausibility of sub-personal theorizing and simulation, I argue that phenomenologists must do more in order to reject these accounts.

Keywords:

• folk psychology
• phenomenology
• simulation theory
• social perception
• theory theory

Acknowledgements

Earlier versions of this paper were presented at UCSD's Experimental Philosophy Lab and the 2008 Pacific division meeting of the American Philosophical Association. I appreciate the comments provided by those present, particularly Adam Arico. I would also like to thank Bill Bechtel for his comments on this paper and related work.

Notes

Since even conscious theorizing must ultimately be characterized sub-personally, the idea must be that the sub-personal processes underlying conscious theorizing are very different from the sub-personal processes enabling direct social perception, such that ‘theorizing’ is never an appropriate characterization of the latter.

One of the more interesting advances by theory theorists is the characterization of theories in terms of Bayesian networks. This has enabled the creation of computational models of cognitive phenomena such as action understanding (Baker, Tenenbaum, and Saxe 2006). While these models remain rather simple at this point, they more precisely characterize the nature of ‘theorizing’ in particular domains, and permit more fine-grained comparisons with human data. Accordingly, they also provide critics such as Zahavi with clearer targets against which to launch objections.

For more on the role of models in cognition, and the sense in which models are isomorphic to what they represent, see Waskan (2006).

Grush (2004) explicitly argues for such an account of neural representation, applying it to motor control, imagery, and perception.

This phrasing is closer to Robert Gordon's version of simulation theory than Alvin Goldman's. Whatever Gallagher's stance on their competing characterizations of simulation, I read Gallagher's objection to subpersonal pretense as cutting against both Gordon and Goldman on sub-personal simulation – see his discussions of Goldman (Gallagher 2007, 361) and of Gordon (Gallagher 2007, 361n10). Below I attempt to undercut Gallagher's objection by appealing to Goldman's (2006) definition of simulation. I do not here delve into the differences between Gordon and Goldman on simulation, or whether Gordon would endorse such a response to Gallagher. I want to thank Marc Slors for his comments on this point.

Gallagher's point here is that the content of neural states fails to meet the pretense condition. He also contends that neural mechanisms considered as vehicles of representational content fail the pretense condition, writing that ‘neurons either fire or they do not fire. They do not pretend to fire’ (2007, 361). As Gallagher seems to recognize, simulation theory is concerned with matters of representational content more than representational vehicles. Thus I will focus on Gallagher's claim that the content of sub-personal, neural mechanisms cannot meet the pretense condition for simulation theory.

My argument here should not be read as endorsing the standard, simulationist interpretation of motor-based accounts of social perception. There are several different versions of these models, and how they should be interpreted relative to theory theory and simulation theory is not so straightforward. See Herschbach (2008a, 2008b).

Hybrid theory-simulation accounts, such as Goldman's (2006), do, however, acknowledge a role for representations of folk psychological generalizations. For example, ‘theory-driven’ simulation uses theoretical knowledge about the target system to generate appropriate input for the simulation process.

How exactly this resulting state should be characterized is a matter of dispute. Simulation theorists (as well as theory theorists) usually describe this as the production of a belief about the other's mental state (e.g., Goldman 2006). But phenomenologists (e.g., Gallagher 2008) contrast the ‘non-conceptual’ experience of direct social perception with the ‘conceptual’, belief-based understanding of reflective simulation and theorizing. Although this issue seems important to adequately describing the personal-level phenomena of social understanding, it requires delving into the thorny issue of the nature of concepts, and is thus beyond the scope of this paper.

See Note 9.

Additional information

Notes on contributors

Mitchell Herschbach

Email: [email protected]