Chapter 4 How to read minds Tim Bayne Introduction Most animals have mental states of one sort or another, but few species share our capacity for self-awareness. We are aware of our own mental states via introspection, and we are aware of the mental states of our fellow human beings on the basis of what they do and say. This chapter is not concerned with these traditional forms of mindreading—forms whose origins predate the beginnings of recorded history—but with the prospects of a rather different and significantly more recent form of ‘mindreading’: the capacity to ascribe mental states to a creature on the basis of information derived from neuroimaging. The thought that we might be able to read minds by inspecting brains has stimulated philosophical interest for decades (Dennett 1978), but with recent advances in neuroim- aging this idea has now passed from science fiction and into science: mindreading—or ‘brain decoding’, as it is also known—is now a burgeoning industry. 1 Here are three examples of mindreading—or at least attempted mindreading. In one study, Haynes and colleagues asked subjects to decide either to add or subtract two numbers that had been presented to them (Haynes et al. 2007; see also Haynes, this volume). On the basis of fMRI data, the experimenters were able to determine with up to 70% accuracy whether the subjects would sum the presented numbers or whether they would subtract one number from the other. In another study, Spence and colleagues suggested, on the basis of neuroimaging evidence, that a woman who had been convicted of intentionally ind- ucing illness in a child may have been innocent (Spence et al. 2008). In a third study, Owen and colleagues concluded that a vegetative state patient was conscious on the grounds that she showed neural activity in brain areas implicated in motor imagery and spatial navigation when instructed to either imagine herself playing tennis or visiting the rooms of her home (see also Boly et al. 2007; Monti/Vanhaudenhuyse et al. 2010). These studies are of great interest in their own right, but they also raise more general questions about the nature and scope of brain-based mindreading. One set of questions concerns methodology. How might one justify the ascription of a mental state to a creat- ure on the basis of neuroimaging data? A second set of questions concerns the scope of mindreading. Under what conditions, and with respect to which kinds of mental states, might mindreading be possible? A third set of questions concerns the interaction between 1 For other examples of mindreading see Chadwick et al. (2010); Dehaene et al. (1998); Haynes and Rees (2005, 2006); Kamitani and Tong (2005); Polyn et al. (2005); Richiardi et al. (2011); and Shirer (2011). 04-Edwards et al-Ch-04.indd 41 04-Edwards et al-Ch-04.indd 41 3/15/2012 10:46:40 AM 3/15/2012 10:46:40 AM OUP UNCORRECTED PROOF – FPP, 15/03/2012, CENVEO