«Genuinely collective emotions∗ Bryce Huebner Department of Philosophy Georgetown University Abstract: It is received wisdom in philosophy and the ...»
Genuinely collective emotions∗
Department of Philosophy
Abstract: It is received wisdom in philosophy and the cognitive sciences that
individuals can be in emotional states but groups cannot. But why should we
accept this view? In this paper, I argue that there is substantial philosophical
and empirical support for the existence of collective emotions. Thus, while there
is good reason to be skeptical about many ascriptions of collective emotion, I argue that some groups exhibit the computational complexity and informational integration required for being in genuinely emotional states.
Does the United States regret its decision not to intervene in the Rwandan genocide?
Were the Teamsters angry about the recent decision to open American borders to Mexican trucking companies? Did the Republican Party feel upset about its lack of success in the 2008 election? Most people would offer the same reply to all three questions: of course not! Commonsense, as well as the received view in philosophy and cognitive science, holds that individuals can have regrets, but the United States cannot;
individuals can be angry, but the Teamsters cannot; and, although the members of the Republican Party were probably upset by these results, the Republican Party itself was not. The dominant view in philosophy, as well as in commonsense psychology, is that the collective term in such claims should be read as a plural term, and the accompanying statements treated as generic statements that call for a collective or distributed reading that adverts to the mental states of the individuals in these groups. But why should we accept the view that individual people can be in emotional states while groups of people cannot? My goal in this paper is to show that we should not be so willing to adopt this view. Indeed, I argue that there is substantial philosophical and empirical support for the existence of collective emotions.
1. Commonsense resistance to collective emotions We often speak and write in ways that appear to ascribe emotions to a various human and non-human entities, as well as objects. “Susanne regrets her decision to live in this neighborhood”; “Germany regrets its genocidal past”; “My cat is unhappy when she finds her food bowl empty”; and, “My car was angry when I finally started it after over a year”.
This paper has had a long history, and has gone through many revisions on the basis of the compelling criticisms ∗ offered by numerous friends, acquaintances, and anonymous referees. Without the helpful criticisms that were voiced by these people, the central arguments in this paper would be far less compelling. Special thanks are due to Adam Arico, Bill Blattner, Dan Dennett, Aaron Garrett, Marcus Hedahl, Justin Junge, Joshua Knobe, Mark Lance, James Mattingly, Ram Neta, Mark Lange, Bill Lycan, Jesse Prinz, Andrea Scarantino, Mikko Salmela, Susanne Sreedhar, Justin Sytsma, and audiences at Georgetown University, Georgia State University, and The University of North Carolina – Chapel Hill.
But while such sentences occur in ordinary language, this does not establish that Susanne, Germany, my cat, or my car can be in emotional states. While some sentences that include mental states terms are intended as claims about psychological states, many others—despite their similarities in surface grammar—are not intended to convey anything that is literally true. As philosophers have long noticed, the ordinary usage of mental state terms ranges over a heterogeneous hodge-podge of genuine mental state ascriptions, dubious attempts at such ascriptions, instrumentally useful metaphors, and generic claims of various sorts (cf., Dennett, 1989). But, this leaves philosophers with a complex task of distinguishing those cases in which mental state ascriptions ought to be regarded as literally true from those cases in which these ascriptions are mere rhetorical flourish.
Of course, philosophical arguments have been marshaled to suggest that our practices of holding groups accountable for their actions, and expecting that groups should apologize for their morally problematic behavior, provide a plausible basis for positing collective guilt and remorse (cf., Gilbert 2001, 2002; Tollefsen 2006; and the essays in May and Hoffman 1991; see Kutz 2001 for a counterargument).1 However, any appeal to collective emotions is likely to seem at least prima facie implausible. Faced with the suggestion that “Microsoft believes the time is right for world domination”, it is reasonable to take the claim at face value, and doing so will likely yield plausible predictions about Microsoft’s future behavior. However, faced with the suggestion that “Microsoft feels melancholy when reflecting on the loss of innocence that has accompanied its rise to power”, it seems more reasonable to feel apprehensive and to demand an explanation. Of course, there are numerous differences between these claims, but the type of state that is being ascribed is clearly salient. Thus, although collective beliefs, desires, and intentions have seemed plausible enough to warrant philosophical attention, collective melancholia has seemed absurd enough to be rejected without reflection.
As described, we would expect Microsoft to feel a particular way, but it is not the sort of entity that can feel anything at all. With such worries in mind, Margaret Gilbert (2002,
119) argues that we must distinguish emotions from feelings, that feelings are inessential, though common concomitants of emotional states. This being the case, she argues that collective emotions require no specific phenomenology; they are unconscious emotions.
On first blush, Gilbert seems to have the force of commonsense on her side. In a recent study, Joshua Knobe & Jesse Prinz (2007) found that people tend to judge that sentences like ‘ACME Corporation regrets its recent decisions’ sound OK, but sentences like ‘ACME Corporation feels upset’ sound weird.2 But, while such data provide some support for A brief note regarding terminology: I use the term ‘collective’ in a more inclusive sense than it is typically used in the literature on collective intentions, plural subjects, and shared commitments. While discussions of collective mental states in philosophy tend to focus on states that are shared by groups of people, the position that I develop—building on theories of distributed cognition in the cognitive sciences—focuses on states and processes that include people as well as the technological structures in which they are embedded. As there is no obvious term that can be applied to groups plus technologies, I retain the term collective mentality to refer broadly to the states of groups as well as groups plus their technological environments.
As has often been noted, the comparison between ‘regrets’ and ‘feels upset’ is not clearly the right comparison— and comparisons that differ only in the inclusion of the ‘feels’ locution must be examined to determine how robust this difference really is. Taking into account the various subtleties of mental state ascriptions, subsequent experiments have revealed that the phenomena are more complex than these initial data suggests. Justin Sytsma & Edouard Machery (2009) found that commonsense psychology draws a distinction between individual-appropriate and group-appropriate behaviors; Adam Arico found that contextual information has a significant impact on commonsense judgments about the sentences ascribing phenomenal states to groups; and my colleagues and I (Huebner et al 2010) found a significant difference between the ascriptions of mental states to groups in East Asian and Western settings using stimuli that used ‘minimal pairs’ of sentences that differed Gilbert’s claim, no one should be persuaded of the existence of collective emotions by statistical regularities in commonsense judgments. After all, even if some ascriptions of collective emotions ‘sound plausible’, and even if they are ‘read literally’,3 there are likely to be plausible non-mentalistic and literal interpretations of such sentences. Such alternative interpretations must be ruled out before we attempt to draw ontological conclusions from these ordinary language data.
The most striking problem with ascriptions of collective emotion is that they fail to distinguish holistic and collectivist views of these mental states (cf., Pettit 1996).
According to the holist, an individual’s mental states depend on her social associations;
according to the collectivist the groups themselves can be in genuinely mental states. As Robert Wilson (2001) argues, ascriptions of mental states often function as claims about certain psychological states of individuals that tend to be manifested only within the context of particular group relations. If individual mental states depend on social relations in this way, then they must be understood relationally, as opposed to as intrinsic properties of individuals. But this fact is far from sufficient for licensing the ontologically robust claim that some groups as such can be in mental states. Unfortunately, many arguments for collective emotion can only establish the weaker holistic claim.
Karl Jaspers (1947) argued that there was a sense in which every German should feel co-responsible for the atrocities perpetrated by the Third Reich. Building on this claim, Larry May argues that a ‘metaphysical guilt’ for collective actions arises where group membership yields a ‘shared identity’ and where a person “did not but could have (and should have) responded differently when faced with the harms committed by his or her fellow group members” (May 1991, 240). Similarly, Tollefsen (2006) defines collective guilt as “the guilt one feels in response to the harms committed by one’s group”. But none of these claims demonstrate the truth of the stronger collectivist claim. While collective guilt is clearly a social phenomenon, these descriptions of collective guilt suggest nothing more than individual emotions that are manifested “as social abilities, as ways of negotiating aspects of the social world” (Wilson 2004, 418). While there is little doubt that members of various groups commonly experience such feelings, there is no reason to suppose that these are emotional states of a collectivity.
To demonstrate the existence of such genuinely collective emotions, it is necessary to show that there are emotional states that are not merely states of individuals in aggregation. In attempting to develop a non-aggregative account of collective emotion, Margaret Gilbert (2001, 2002) distinguishes three types of ‘guilt feelings’ that are present in the context of a various groups. ‘Personal feelings’ are the result of actions that an individual herself carries out. There are many cases in which we must advert to socialrelational properties in individuating these states; but individuals experience personal feelings because of their own contributions to some collective end. ‘Membership feelings’ only in the agent to which a mental state was applied, though we also replicated Knobe & Prinz’s original effect.
Yet, as I have argued elsewhere (Huebner, 2010), commonsense judgments about mental states and processes are likely to be highly malleable and, as such, are more likely to tell us far more about ideology than ontology.
That is, while such experiments might tell us something important about the default strategies that people employ in evaluating the plausibility of various sorts of mental state ascriptions, they are far less likely to be informative when we turn to questions about the furniture of the world (Thanks are due to an anonymous referee who pushed me to clarify my views on this point).
Faced with such data, philosophers are likely to worry that the people in these experiments are interpreting these sentences in some metaphorical or figurative sense. After all, there are many ways of using emotion terms in meaningful ways without ascribing emotional states. However, when Arico and his colleagues (unpublished data) asked participants to categorize sentences as figurative (e.g., “Einstein was an egghead”) or literal (e.g., “Carpenters build houses”), sentences that attributed mental states to individuals (e.g., “Some millionaires want tax cuts”) and groups (e.g., “Some corporations want tax cuts”), tended to be categorized as ‘literally true’.
are the result of collective actions even where the individual that feels them did not contribute to the specified action. Jasper’s claim—and May’s claim following him—is that a sort of moral taint affects the member of a group who did not respond as she should have to the morally problematic actions of her group. The fact that a person can say of her group that it carried out some action, while she was a member of that group, provides grounds for a feeling of guilt even in those cases where she did not herself behave in a morally problematic way. Finally, the non-aggregative alternative suggested by Gilbert (2001, 139) is termed ‘collective feelings’: For us collectively to feel guilt over our action A is for us to be jointly committed to feeling guilt as a body over our action A. This brings us closer to a genuinely collective emotion. However, even here it is unclear how such a state is supposed to be understood as a collective emotion as opposed to a distribution of individual states that are directly sustained by the causal and conceptual connections between the individuals in a collectivity.
On Gilbert’s view, the emotional states of a ‘plural subject’ are understood as resulting from the commitments that each of the individuals has made to feel guilty or to feel remorse. On the basis of this model, non-aggregative collective emotions result from the ways in which the members of these groups coordinate their actions to produce some intentionally specified behavior. However, as Robert Rupert (2005) has argued, from the standpoint of psychological explanation, we gain no additional explanatory power by appealing to collective mental states in these cases.
After all, every step in the construction of such representations, as well as every step in the causal sequence alleged to involve the effects of those representations, proceeds either by brute physical causation (e.g., photons emitted from the surface of the page stimulate the reader’s retinal cells) or by causal processes involving the mental states of individuals (Rupert 2005, 5).