Experimental Philosophy: Thinking Things and Feeling Things

Casual reflection on the daily news reminds us that we often ascribe intentional mental states, such as beliefs and desires, to groups. For example, Apple thinks its smartphone patents are worth more than Microsoft's. And surely it's reconsidering its new iMaps.The commonality of such ascriptions is often taken to constitute important evidence, both for experimental philosophers exploring the folk metaphysics of mind, and for non-experimentalists, interested in collective intentionality. However, such ascriptions actually constitute important data for these projects only on the assumption that people are realists about group mental state ascriptions--that is, only on the assumption that people interpret such ascriptions as really attributing mental states to groups over and above their members. In our recently published paper, "Thinking things and feeling things: on an alleged discontinuity in folk metaphysics of mind," Adam Arico, Shaun Nichols, and I argue against realism and offer a distributive account of group mental state ascriptions. (Paper available at: http://www.springerlink.com/content/4r026767m4101381/ and in a penultimate draft here: Download TTFTPreprint.)

At the heart of our view is the claim that intentional state ascriptions to groups are unremarkable. Intentional and phenomenal mental states constitute one subset of those states that, according to folk metaphysics, cannot be appropriately attributed to groups over and above the members that constitute them. Just as a group cannot get drunk or be well-dressed over and above its members, so a group cannot want or believe over and above its members. On our view, when a state is not appropriately attributed to a group , verbal ascriptions of that state to a group are understood distributively in terms of attributions of states to (a set of) group members qua members of the relevant group. In effect, such ascriptions are understood as attributions to (a subset of) those who make up the group in their role as members of the group. And this verbal shorthand is tolerated insofar as it secures brevity at a tolerable cost to innaccuracy.

What then explains the purported acceptability of group intentional state talk? To see this we have to first recognize what makes us think it is appropriate to ascribe a state to a person as the occupant of any role whatsoever. Suppose John is both a husband and the mayor. It no doubt strikes us as more appropriate to say, "As a husband, John loves his wife," than to say, "As a mayor, John loves his wife," though, of course, John actually occupies the state whether or not we are thinking of him as a husband. Such asymmetries in perceived appropriateness seem to reveal little more than our presuppositions about what states are saliently associated with a given role. On our view, there is nothing special when it comes to occupying the role of a group member. Thus, verbal ascriptions of states to a group--which are understood as attributions of states to the members that constitute the group (under the role of group member)--will be deemed appropriate insofar as the relevant state is saliently associated with the role of being a member of the relevant group. For example, the belief that profit margins will increase is more saliently associated with the role of being an Acme Corp employ than is the experience of existential angst. Thus, it strikes us as more appropriate to say, "Acme Corp believes profit margins will increase," than to say, "Acme Corp is experiencing existential angst." Along these lines, we maintain that the purported asymmetry in favor of group intentionality state talk is due to nothing more than a selective consideration of cases.

All of this is discussed more carefully and in much more detail in the paper, where it is also supported by argument and experimental findings. Here we offer but one graph as a teaser. This study demonstrates the way in which salient association leads to higher appropriateness ratings for group state ascriptions across a variety of states (non-mental, intentional, and phenomenal). Light gray bars record mean appropriateness assessments of ascriptions to groups for which the relevant state is more saliently associated. Note that the trend for group mental state ascriptions is mirrored for non-mental states that obviously cannot be attributed to the group over and above the members (such as getting drunk). Note also that in this study ascriptions of phenomenal states to groups were deemed more appropriate than either intentional or non-mental states.