I read some literature on group agency and decided to develop an account of quasi-group agency, somewhere between private agency and group agency. Readers’ comments are welcome as always.
This paper is related to my earlier post on Rape Culture, which you can find here. I tried to give substance to those arguments by doing more research on group agency. The results are below.
Also, can I just say that Jose Medina is brilliant? I did not read him soon enough.
Quasi-Group Agency: Chains of Actors
Abstract: Group agency scholars typically posit two distinct types of agents: private agents who act on uncoordinated intentions on the one hand, and group agents whose members share common goals on the other hand. In justice scholarship, we find examples of a third category of agent, whose members are neither completely uncoordinated nor unified by consciously shared goals and common awareness—quasi-group agents. A salient example is a nascent social network, which has not yet developed coordinating structures or transparently shared goals. In this paper, I provide an ontological account of this intermediary category, which I call a quasi-group agent, and I show that quasi-group agents are not reducible to either private agents or group agents, as they have distinctive motivational and normative properties and are an essential part of our constantly-evolving social reality. I show that this ontic category can be used to describe both beneficial quasi-group agents, such as modern social movements, and malignant quasi-group agents, such as Rape Culture.
Key Words: Group agency; quasi-group agency; chains of actors; resistance; oppression; responsibility
The literature on group agency typically focuses on providing conditions of group agency, which succeed in distinguishing group agents proper from uncoordinated collections of individual or ‘private agents.’ Few theorists discuss in-between cases in which a set of private agents share coordinating states without consciously subscribing to a shared goal. A notable exception is Tuomela, who writes about “intermediary cases” in which a “collection of persons” share a “group ethos-to-be” and “can be expected to become a [group agent]” (2013: 129). Tuomela intends for his conditions of group agency to apply to these intermediary cases unproblematically, with the caveat that these collectives share only a weak form of shared intentionality (as we shall discuss); but they must still knowingly contribute to a shared ethos.
I do not think that this goes far enough in distinguishing these loose collectives from group agents. I will argue that quasi-group agents are a third ontic category with distinct individuating conditions, and that ignoring this ontic category can lead to a naïve representation of social reality and correlative naïve beliefs about the normative status of members of these collectives. By focusing on quasi-group agents as a distinct ontic kind, and outlining what distinguishes them from either private agents, who are completely uncoordinated, or group agents, whose members are consolidated by the shared goals and common awareness of inclusive members, we can learn something about how collectives operate in a fluid social space.
In particular, we can debunk common myths about the production of coordinated social movements from the mobilizing efforts of remarkable private individuals (epistemic and moral saints) who have incredible epistemic powers, organizing capacities, and charismatic influence. In reality, social movements are not (except perhaps in the rarest of cases) fomented by private agents, but rather, they evolve through an intermediary stage of agency in which chains of actors, dispersed across geographic and epistemic space, share a subset of motivational states—hopes, aspirations, values, implicit attitudes, transformative ideals—but are not fully aware of: their place in the chain, the transformative potential of the chain, the symbolic significance of their chained actions, whether others will reliably contribute to the chain, which people are members of the chain, and so on. Thus, they do not harbour ‘shared intentions’ or ‘common awareness’ in the sense typically assigned to group agents.
Although concrete examples are largely absent in the group agency literature, we find them in justice theory (e.g., Anne E. Kane 1997, Sally Haslanger 2017, Jose Medina 2012). Medina provides a particularly compelling account of the evolution of social groups, and I will rely on this historical account to develop conditions of quasi-group agency, which should be of interest to group agency theorists. I take List and Pettit’s four conditions of group agency as a point of departure for developing a set of conditions for quasi-group agency, which can be applied to what Medina calls “chained actors”—the partly disjointed, partly unified collection of agents who foment social movements (Medina 2012: 225). This metaphor of ‘chains of actors’ vividly captures the notion of beneficial and transformative social networks. I will add to Medina’s social schemata ‘malignant quasi-group agents,’ which collectively promote injustice and oppression, but are not tightly-knit or coordinated. Malignant quasi-group agents are often the impetus for the emergence of beneficial quasi-group agents, which arise to combat their harmful counterparts. They both, on my view, share motivational features that loosely unify their members.
In this paper, I will propose that chains of actors (i.e., quasi-group agents) satisfy the conditions of group agency offered by most theorists, minus the conditions of knowledge (of a shared goal, participatory status, or interdependence) and common awareness shared amongst inclusive members. Specifically, members of quasi-group agents: (1) together promote a common goal, (2) each do their part in promoting that goal, and (3) participate in the group at least in part because others do, and depend on the participation of others to sustain the group; but they may not know that they are satisfying (1)-(3), contra the standard view in group agency theory.
Specifically, in sections 2-3 I will discuss chains of actors as quasi-group agents and develop conditions of quasi-group agency by modifying List and Pettit’s influential conditions of group agency (eliminating the knowledge and common awareness conditions); in sections 4-6 I will show that my proposal applies to both beneficial quasi-group agents, like the incipient members of Black Lives Matter (prior to coordination), as well as malignant quasi-group agents, like members of Rape Culture. And in section 7, I will justify my proposal on grounds that chains of actors, whether beneficial or malignant, are a distinct ontic kind (intermediary between private agents and group agents), whose members have distinct motivational and normative features. Recognizing this is crucial to understanding the nature of our social reality, which involves constant conflict and negotiation between competing elements.
- Chains of actors
Recently, we have seen salient examples of social groups emerging and coordinating in response to conditions of political, epistemic, and moral injustice. Examples include Black Lives Matter and the Women’s March. We have also seen salient examples of antagonistic collectives that perpetuate injustice and oppression, but lack coordinating structures and precepts. One example of the latter case is Rape Culture, which is becoming a widely recognized phenomenon (more on which later). It is a substantive question how we should theorize these loose collectives in ontological and normative terms, seeing that they differ in critical respects from paradigmatic group agents, such as corporations and clubs, as well as (seemingly) private individuals. This is an important and undertheorized question, which has not received much attention in the literature on group agency (see List & Pettit 2011, Pettit 2007, Bratman 2013, Isaacs 2011, Tuomela 2013).
One place where we find explicit reference to uncoordinated, dispersed, and nascent social groups is Jose Medina’s work on epistemic responsibility and resistance (2012). He introduces the term “chained action” to describe the isomorphic actions of members of uncoordinated collectives (224). We can extrapolate that these collectives are ‘chains of actors,’ whose members act together in some meaningful sense, in spite of being more or less unstructured. Medina introduces the notion of chained actors to address what he takes to be a lacuna in the literature on group agency and group responsibility: the omission of a “hybrid, middle ground” between fully-formed group agents on the one hand, and individual actors on the other (224). This interim category consists of social networks that are neither coordinated groups with shared intentions, nor isolated individuals with radically disjunctive intentions.
Medina does not explicitly discuss the ontology of this category, but it appears to be a type of quasi-group agent, with is a composite (in some sense) of group agency and individual agency. One of my aims is to provide an ontological grounding for this category.
But why posit quasi-group agency, rather than maintaining a more parsimonious ontology? We will explore this question in more depth later, but we can note here that Medina introduces the notion of chained actors to capture the emergence of resistant and revolutionary social movements through an intermediary developmental stage, between uncoordinated private agents and coordinated groups. These networks create the social, political, and epistemic conditions that make coordinated movements possible. Identifying this stage of development also helps to debunk the popular myth that social movements emerge from the discrete actions and private decisions of individual “epistemic heroes” and moral saints (25), who spark resistance movements using their own agential resources. This familiar ‘boot-strapping’ theory represents a false picture of the world—a false ontology—which is widely accepted within mainstream American culture (Medina 2012, Hood 2012).
On scrutiny, prominent social activists begin as part of a chain of actors who support one another and echo each other’s symbolic gestures. Rosa Parks, for example, was one of many activists to resist segregation laws, and her political activism was enabled and supported by others: these actions would not have been realistically imaginable, intelligible or effective without some degree of community support. Normatively speaking, when we buy into the boot-strapping myth of heroes and saints, we misconstrue the role of responsibility in social life, giving too much credit to perceived saints, and short shrift to supporters; similarly, we tend to vilify leaders of nefarious groups such as neoNazis, while ignoring the complicity of background members. This vilification is dangerous not because it is too harsh on leaders, but because it ignores the role of enablers and abettors who are equally counterfactually significant. The hero-villain picture (I shall argue) is an example of The Fundamental Attribution Error (FAE) (see Harman 1999, Doris 1998), which is a pervasive cognitive distortion that induces us to exaggerate the significance of perceived intentional traits (such as saintliness and wickedness) while neglecting ‘background factors’ such as supporting figures, communities, collectives, and infrastructure. FAE gives rise to misperceptions of reality (i.e., false ontological assumptions), which in turn give rise to naïve and distorted moral judgments. Thus, the reasons to posit quasi-group agents are both ontological and moral: doing so helps us understand social reality as a fluid space with competing elements, and this in turn helps us allocate responsibility fairly to private, quasi-group, and group agents.
This is still a rather cursory description of the problem. I will expand on these thoughts later.
In the next section, I will define a chain of actors as a quasi-group agent that satisfies some, but not all, of the standard conditions of group agency offered by group agency theorists. Then I will apply this account to salient examples of beneficial and malignant collectives, to justify this framework and perhaps also shed light on these collectives.
- Conditions of quasi-group agency
What is a group agent?
There is considerable overlap in contemporary theories of group agency, though some proposals are more stringent than others. I will consider three views that share commonalities, but differ in salient respects—those of List and Pettit, Tuomela, and Kutz. I will argue that even the most lenient of these views is too stringent to capture the unique motivational psychology of quasi-group agents, but an attenuated version of List and Pettit’s conditions (and similar proposals) will suffice.
All group agency theorists agree that a group agent is a collection of individuals with intentional states, but they differ in their view of how these agents are coordinated or incorporated into a group. List and Pettit offer an influential account of group agency, which has the following four conditions:
- Shared goal: The [members] each intend that they, [as] members of a more or less salient collection, together promote [a] given goal. (Each member has this intention).
- Individual contribution. They each intend to do their allotted part in a more or less salient plan for achieving that goal.
- Interdependence. They each form these intentions at least partly because of believing that the others form such intentions too.
- Common awareness. This is all a matter of common awareness, with each believing that the first three conditions are met, each believing that others believe this, and so on.
By comparison, Tuomela’s view is somewhat more stringent than this, while Kutz’ is somewhat more lenient.
Tuomela’s view is more stringent in that his notion of a ‘shared goal’ entails a ‘we-intention,’ which represents the individual as inherently, not merely incidentally, part of the group. This gives rise to obligations to favour the group’s interests over private interests in all cases, which in turn engenders a shared sense of solidary.
Kutz’s view is more permissive in that he rejects the common awareness condition in favour of a more lenient ‘participatory intention condition,’ on which an inclusive member does not necessarily intend to promote or realize the group’s shared goal—she may, in fact, refuse to promote the group’s goals in a particular case (or set of cases), but is still a member, and responsible for the group’s actions, as long as she knowingly holds membership status. Kutz offers the example of the 800+ pilots involved in the Dresden firebombing squad, some of whom may have participated in spite of strong reservations against the airstrike, and any one of whom may have made no causal contribution to the outcome, viz., 22,000 to 25,000 civilians killed. On Kutz’ account, even non-contributing and privately dissenting pilots are members of the firebombing squad, and bear responsibility for all of the civilian deaths.
Even this attenuated criterion of participatory intentionality, however, does not adequately describe many chains of actors, whose members may not even self-identify as members of the chain. Yet, chained actors also do not exhibit the motivational psychology of disconnected private agents, since they share motivational states in common, which impel or inspire them to perform parallel actions.
I propose that we see a quasi-group agent (or chain of actors) as satisfying some of the standard conditions of group agency, though in a modified form. Specifically, I submit that a quasi-group agent must satisfy the first three of List and Pettit’s conditions, minus the constraint of knowledge or full awareness embedded in those conditions. This allows people to count as members of quasi-group agents even if they do not know that they are participating in a chain of actors, consciously intend to perform a role in the chain, or expect others to share their intentions and goals; but chained actors must share a set of implicit or less-than-fully-aware motivational states.
To be more precise, my proposal is that members of quasi-group agents, or chains of actors, must satisfy the following modified conditions of group agency: They must (1) together promote a common goal, (2) each do their part in promoting that goal, and (3) participate in the group at least in part because others do, and depend on the participation of other members to perform their role in the chain; but they need not know that they satisfy (1)-(3). This implies that we must drop condition (4) entirely, since quasi-group agents may not share awareness of each other’s shared intentions.
Chained actors do not promote a common goal by accident, however: they share (at a minimum) implicit states that give rise to symbolic gestures that express and promote a common ethos or aim. Thus, they are not simply distinct private agents.
To show that these conditions apply to chains of actors, we will examine two types of quasi-group agents: beneficial chains, which promote beneficial aims (such as justice), and malignant chains, which promote malignant aims (such as the oppression of vulnerable social groups). Medina focuses on the first type of collective, but nefarious collectives also act homologously on the basis of shared states. They are also a critical part of social reality, which involves ontological tensions.
- Beneficial groups: The nascent civil rights movement
To begin this analysis, let’s consider Medina’s example of the civil rights movement. Initially, various uncoordinated actors, including Rosa Parks, decided to resist segregation laws by enacting various forms of public protest, including refusing to sit in designated bus seats. These initial resistors were not part of a coordinated and self-aware group of social activists, but were chained actors who acted in a largely uncoordinated fashion, though with the same general goal in mind: resisting segregation. At this early stage, however, they may not have recognized their actions as part of a shared ethos, common to a chain of actors, or they may have had only partial and embryonic awareness of the collective saliency of their actions. Nor could they have understood, at this stage, the transformative potential of their chained actions, which were the initial conditions for consolidating into a coordinated movement. Thus, although these early activists were “unaware of [their] membership” in the chain, their chained actions comprise a retrospectively identifiable and “traceable performative chain, with each action in the chain having (subsequent) traceable effects in the subsequent actions of others” (Medina: 225-226).
How can we understand the common motivational features of these chained actors? In psychological terms, they can be seen as sharing a hope or aspiration for transformative change, but surely not knowledge or clear foresight of their action’s role in social change and the edification of mainstream society (more on which shortly). Thus, while chained actors share certain values, hopes, and (perhaps) incipient recognition of the transformative potential of their actions, they do not satisfy the conditions of shared goals and common awareness attributable to group agents.
Medina concurs that chained actors do not necessarily, or even characteristically, participate in a chain of actors in a fully conscious way. Parks, for example, is described by many historians as having been ‘selected’ by the community to represent the civil rights movement, rather than having consciously chosen to do so. The community itself, moreover, ‘selected’ her in a not-fully-conscious way, in part because she had characteristics that were viewed favourably by mainstream America (modesty, charm, intelligence, etc.). Thus, Parks’ role in the movement was not analogous to the appointment of a CEO to company, in which the candidate applies for the position and is considered by a hiring committee according to pre-determined selection criteria. But Parks’ role also was not the result of dumb luck or inexorable social forces: she and other chained actors intentionally chose to perform acts of resistance, though their intentions were not coordinated in the way that corporate choices are. The chained actions of nascent social groups are unified instead by shared hopes, ambitions, and a (potentially) nascent recognition of the transformative potential of their actions, though they are not known to be such by the actors, nor are they coordinated by transparent organizing structures and precepts.
Medina specifies that chained actions manifest “a mixed and hybrid kind of agency in which intentional and non-intentional acts of individuals and groups become interwoven, enabling and constraining each other in complex ways” (244). A single act of protest is echoed by others and eventually solidifies into an ‘echoable’ symbol of resistance, which may be repeated by others, gradually increasing the public recognisability of that type of action. Adding to this, I have suggested that this ‘hybrid agency’ is psychologically realized as hopes, aspirations, and implicit or ‘patchy’ awareness of the radical social, political, and epistemic portent of these actions.
This supports the reading that quasi-group agents satisfy List and Pettit’s conditions of group agency, minus the requirements of knowledge and collective awareness; and it provides a descriptive rendering of the motivational psychology of chained actors.
In what follows, I will apply these criteria to the examples of Black Lives Matter in the nascent stage, and to Rape Culture in its current state, to show that these conditions do not just characterize the nascent civil rights movement—they have a broader scope. I introduce the example of Rape Culture because it is important to recognize that beneficial quasi-group agents and malignant quasi-group agents coexist as conflicting forces to be addressed and negotiated in our shared social space.
- Black Lives Matter: Early stages
Black Lives Matter can be seen as a modern example of quasi-group agency in my terms. Although it is now a coordinated group, it began as a response to the acquittal of George Zimmerman in the shooting death of Treyvon Martin in 2014.
This response was coordinated “by three black queer women who know what it is to have one’s humanity demeaned and despised: Patrisse Cullors, Alicia Garza, and Opal Tometi” (Gafney 2017: 205; see also Bradley 2016). They responded to what they insightfully perceived as a vilification of an innocent black man and an instance of systemic anti-black oppression, by creating the hashtag #BlackLivesMatter. These activists had epistemic insight into the insidiousness of white privilege and white ignorance, to be sure, but they could not have known at the outset that their symbolic act of resistance would have the transformative potential it did. They could not have known this because transformative change requires cooperation from a chain of actors who share situated insights, values, hopes, and aspirations, and no one can know in advance if others will rally to support a just cause in conditions of oppression.
I have proposed that chains of actors are united by hopes, which distinguishes them from both group agents and private agents. A hope is different from a conscious intention. Victoria McGeer describes hope as an “aspirational” attitude different from a belief, which involves a “decision to trust in the absence of prior belief” (2008: 243). A hope, that is, involves trust in others even when trust is not evidentially warranted. Jonathan Lear similarly describes a “radical hope” (a species of hope) as a motivating state that enables the agent “to go forward hopefully into a future that [one] would be able to grasp only retrospectively, when [one] could re-emerge with concepts with which to understand [oneself and one’s] experience” (2006: 115). A radical hope, then, ‘goes beyond’ the evidence and envisions a new social order. I am proposing that early-stage social activists share such hopes, which allows them to imagine the future as one in which there is justice and equality—that is, it enables them to envision a radically different social reality. A hope, as such, cannot be the content of a shared goal in List and Pettit’s sense, since a shared goal entails knowledge that others share this goal and will contribute to its realization, whereas chained actors can only hope that others will support their aims and interests.
Chained actors may also share insights and values, but they do not share knowledge or common awareness that these same states are shared by others and will be enacted by others in a coordinated movement. So, while chained actors share motivational states that structure their actions, they do not qualify as group agents.
- Malignant chains: Rape Culture
Malignant chains provide a striking example of non-intentional participation in a collective, because their members often act on implicit biases and other shared insensitivities, which the agents may not be aware of (I shall argue). I define Rape Culture as a chain of uncoordinated actors who promote rape myths in their words and deeds, though they may, in some cases, sincerely deny that they are perpetrating these myths and thus contributing to Rape Culture. Nonetheless, they are chains of actors in the sense that they share and express implicit rape-positive attitudes.
I am treating Rape Culture as a set of attitudes embodied in concrete intentional agents, who share distinguishing psychological and moral features. However, this term is typically used metaphorically to describe our entire culture—a culture in which rape is normalized and trivialized in a way that increases the incidence of rape. For example, Strain, Martens, and Saucier (2016) describe “rape culture” as a “society that excuses or encourages sexual violence,” in which “it is common for individuals to believe in rape myths—beliefs that are statistically false, but perceived as truths (e.g., ‘only certain types of men rape’ or ‘only certain types of women are raped’)” (87). These beliefs, they note, have significance behavioural effects:
Belief in rape myths has been established as a measurable construct (Burt, 1980 ; Payne, Lonsway, & Fitzgerald, 1999 ) that predicts various rape-related attitudes, including negative perceptions of women who have been raped (Anderson, Cooper, & Okamura, 1997 ; Hammond, Berry, & Rodriguez, 2011 ) and an increased self-reported likelihood (among men) of committing rape (i.e., rape proclivity; Bohner et al., 1998 ; Chiroro, Bohner, Viki, & Jarvis, 2004). (88).
This is a good illustration of the attitudes that define Rape Culture. But I am using the term ontologically, not merely metaphorically, to denote a set of individuals who share and express (and thereby promote) rape-positive attitudes. To capture only this subset of agents—exclusive of resistors and neutral parties—we must modify the above description in two ways. First, we must redefine Rape Culture as, not an all-encompassing set of social practices and institutions, but rather a nested subculture embodied in uncoordinated agents who promote rape myths more than the average person, and thus contribute to male dominance. Second, we must amend the condition of belief in rape myths (as the central vehicle for Rape Culture posited by Strain and colleagues), to mere acceptance of rape myths, which can be either implicit or explicit. This allows us to include not only unapologetic rapists and misogynists in the category of Rape Culture, but also those who would deny being members.
- Rape culture as a subculture
Although Rape Culture surely affects and encompasses all of us, it is not perpetuated by everyone to the same extent. Rape myths face resistance from social activists, such as Black Lives Matter (which is committed to feminism, women’s rights, and LGBTQIA rights) and various feminist organizations. Thus, if we want to see Rape Culture as an ontological category with human members, we cannot see it as encompassing all of us to the same extent: we must see it as an uncoordinated set of dispersed members who disproportionately accept and express rape myths. This excludes those who reject and resist these myths, as well neutral parties who neither contribute to the existence of Rape Culture, nor actively resist its central ethos.
Rape Culture, then, can be seen as a chain of actors who share rape-positive attitudes.
- Implicit rape-myth acceptance
Members of Rape Culture may not know that they are members, not because rape culture is a nascent group—on the contrary, it is a deeply-entrenched and stable chain with sweeping influence; but members of this chain do not necessarily share explicit attitudes and beliefs about the permissibility of rape. It suffices that they share implicit, but motivating, rape-positive attitudes. Thus, we must change ‘belief in rape myths’ to mere ‘acceptance of rape myths,’ which can be either explicit or implicit.
An implicit attitude is a relatively fast, automatic, and unconscious motivational state, which can give rise to overt behaviour in response to salient eliciting conditions. Neil Levy calls implicit states “patchy endorsements,” which have propositional structure like beliefs, but are not responsive to other states, and thus can motivate behaviour independently. Implicit states also are not uniformly expressed in every context, but appear to have activating conditions. For example, male subjects were twice as likely to report that they would engage in non-consensual sex (rape) when asked to fill in a questionnaire in their dorm rooms while looking at pornography than when in a laboratory setting (Ariely 2006). However, implicit states, as we shall see, are not completely uncontrollable: their expression in overt behaviour is amenable to indirect control via remediating measures.
Since implicit states are not fully conscious, they may be genuinely denied as motivating by those who express them, if the person is ignorant or unaware of the content or normative valence of his choices and actions. Rape-myth acceptance as a mental state can be compared with Charles’ Mills’ notion of White Ignorance, which seemingly implicates implicit states. Mills describes White Ignorance as resting on “self-deception, bad faith, evasion, and misrepresentation” (17), which are, cognitively speaking, implicit biases. Medina similarly characterizes White Ignorance as involving a combination of first-order insensitivity to racist stereotypes, and meta-insensitivity to these first-order states. This suggests that White Ignorance implicates implicit states, and is to this extent first-personally opaque—though, as we shall discuss shortly, it can be ‘revealed’ through indirect means, such as exposing oneself to different perspectives. It seems accurate to see rape-myth acceptance as similar to White Ignorance in being realized largely in implicit states.
To illustrate this, consider Brock Turner, the Stanford University student who raped an unconscious woman after a party and was caught in the act by two passers-by. In Turner’s court statement, he argued that his unconscious victim was consenting; he denied that he had committed rape, but instead referred to to his act (of rape, objectively speaking) as a mere “lapse of judgment”; and he appealed to “the stress of school and swimming” as an excuse for his behaviour (Paiella 2016). These ideas—that an unconscious person can consent to rape, that rape is a mundane indiscretion as opposed to a crime, that rape is caused by stress rather than an intertwined set of rape-normalizing attitudes—all contribute to Rape Culture, to systemic male dominance and the victimization of women. Brock Turner was, by all appearances, not fully aware that he was invoking rape myths in his statement, but it is clear to any enlightened (non-ignorant) observer that this is precisely what he was doing. His father later reinforced these rape myths in a letter to the judge, where he similarly refused to name the act as rape, trivializing it as a mere “20 minutes of action,” for which his son had, allegedly, already suffered enough (Cleary 2016). Brock Turner and his father thus expressed rape myths in their words and actions, though they would both explicitly deny this interpretation of the situation.
This shows how people can be merely implicit members of Rape Culture: they may express rape-myth acceptance in their overt behaviour, even though they do not explicitly endorse these myths. Since they express similar symbolic actions of oppression, they can be characterized as chains of actors inclusive to Rape Culture. These actors: (1) each promote shared (perhaps implicit) states, which normalize and trivialize rape and preserve male privilege; (2) each do their part (perhaps implicitly) in promoting this shared aim, and (3) participate in the group at least in part because others do, and depend on the participation of others to play their role. The third condition is satisfied because people like Brock Turner could not intelligibly invoke rape myths as a justification or defense of their behaviour if others did not share these attitudes, giving them a semblance of credibility. If no one perpetuated these myths, it would be self-evident that they are false and pernicious.
A caveat is in order. Since everyone is exposed to rape myths, some degree of rape-myth acceptance is pervasive. But membership in Rape Culture must be limited to those who perpetuate rape myths more than the average person; these vanguards of the movement are sustaining and reinforcing the status quo, while others are either neutral (not making a difference either way) or resistant. While there is no precise calculus for picking out members of Rape Culture, an enlightened person can identify members by their disproportionate degree of overt rape-myth acceptance.
To expand on this point, we can look at List and Pettit’s three corporate roles, which include (i) leaders, (ii) members, and (ii) enactors. Leaders “determine the group’s procedures for forming its beliefs and desires and taking its actions,” members play a non-governing role in the group, and enactors “carry out the group’s wishes,” even if they are not inclusive members (164). (Enactors may be facilitators, corroborators, or abettors, for example). We cannot apply these categories to chains of actors, since there is no formal structure within the chain, but we can identify more and less central members of the collective based on how much the individual perpetuates the chain’s values and attitudes—in this case, the extent to which the individual causally or symbolically contributes to the promotion of rape myths and the (related) preservation of male privilege. Turner was a very central member, as he actually committed rape and then denied it, while more mundane (less violent) contributors may be seen as peripheral or marginal by comparison. Thus, there are correlatives to List and Pettit’s three categories, in terms of the significance of an agent’s contribution to the chain: a chained actor may be central, middling, or marginal.
- Why posit quasi-group agency?
These considerations show that incipient social movements, as well as stable but uncoordinated collectives, can be described as quasi-group agents. But one might question why we should posit quasi-group agency at all, rather than retaining a more parsimonious ontology with only two categories: group agents and private agents. To show that quasi-group agents are conceptually and normatively significant and irreducible, I will return to, and expand on, the rationales outlined above.
I will offer both ontological and normative rationales for positing quasi-group agents.
- The evolution of group agents & shared motives
As Medina observes, group agents are not created by the individual choices of private agents; they typically evolve through interim chains of actors, who provide the underlying social conditions for mobilization. These actors share values, insights, hopes, and aspirations, which motive parallel acts of resistance. These actions lay the foundation for coordinated activism and social change. Malignant collectives can also consist of chains of actors who share an overlapping set of implicit and explicit attitudes, values, and interests, which are expressed in patterns of overt behaviour. While it is possible for a single individual to create a corporation or club through their own private initiative, this is not how transformative social movements come into existence, nor how malignant chains manage to thrive with relatively impunity, safely ensconced within a collective of allies. A realistic social ontology must make space for these chains of actors, and the tensions between them.
- Shared uncoordinated motives
We have already seen how chains of actors fail to satisfy standard conditions of group agency, which is what we would expect of nascent and unincorporated chains. But chains of actors also are not identical to private agents, inasmuch as they do not act on discrete, unrelated, and purely private motives. As we saw, members of nascent social movements share situated attitudes, values, hopes and interests, though they may not know (in full consciousness) that there are others who share, and are prepared to perform and echo, these attitudes, creating the basis for a more coordinated movement. Malignant chains of actors also share motivational states, which may not be explicitly realized. The actions of chained actors are not coordinated through formal structures and procedures like a corporation, but they also are not radically disjunctive like the intentions of private agents. For instance, if several private actors commit a murder, they may do so for financial gain, in retribution, for status, in a state of psychosis, or for any number of unrelated reasons. These individuals do not share motivational state as a result of their social position and shared experiences—they act on their own, uncoordinated, private intentions.
- The Fundamental Attribution Error (FAE)
Seeing chains of actors as disjointed private agents is misguided, since individuals do not produce social movements on their own. This bootstrapping myth is belied by social psychology and an accurate understanding of how social movements arise. The idea that epistemic heroes and moral saints alone foment social movements, and that epistemic villains and moral rogues alone produce malignant social groups, are popular but mythical beliefs. In reality, no one can foment a social movement alone, since the ideals of the movement lack ‘uptake’ or public recognizabilty and credibility in the absence of multiple repetitions of the same gesture in public space. The hashtag #BlackLivesMatter would not be widely recognized as a symbol of racial justice, and effectively serve this purpose, if it had not been shared by many people and incorporated into a broader narrative about anti-black racism in America. Similarly, Brock Turner’s court statement would not have made sense as an intended defence, and earned him a mere six months in prison, if the rape myths that he invoked were not familiar within our culture and regarded by many as truths.
Our susceptibility to hero and villain myths is, I have suggested, an example of FAE, a cognitive bias that induces us to ignore background factors and exaggerate the role of charismatic and/or villainous leaders. While it is true that leaders often play a disproportionate causal role in collectives, they cannot sustain the collective on their own. FAE provides an error theory for the myth of superhuman leaders. But FAE is not an inevitability: while it is part of normal human psychology, it is not beyond our control. Indeed, implicit states in general appear to be susceptible to indirect control, even if they are not directly introspectible and sheddable. Jules Holroyd notes that we have, at least, “indirect and non-immediate control over the influence of implicit biases” on our behaviour (2012: 286, see also Kelly et al. 2010). That is, while we cannot reflectively access and directly change the state of our motivational psychology, we can take steps to mediate the expression of implicit biases in our overt behaviour, over time. Studies show that exposure to countersterotypical exemplars or members of stigmatized groups has a remediating effect, as does the use of implementation intentions, which help the agent re-condition her habitual responses to eliciting conditions (Holroyd 2012, Kelly et al. 2010).
The matter is somewhat simplified in the case of FAE, since knowledge of history and social movements should moderate our bias toward over-representing perceived villains and heroes, to the exclusion of supporting and enabling members, as well as background infrastructure. We can recognize the role of leaders and prominent figures in chains and groups, while still keeping an eye on supporting figures.
These considerations also have implications for whether agents bear responsibility for their role in chains and groups, which we will discuss in the next subsection.
- Normative (Responsibility)
List and Pettit offer a definition of responsibility that resembles the control view defended most famously by J.M. Fischer (2011, 2006). On this view, an agent is responsible for a morally-significant choice, only if the agent had control over that choice (List & Pettit: 158). Saying that an agent A is responsible for choice C implies that A is morally appraisable (blameworthy or praiseworthy) for C (List & Pettit: 154).
List and Pettit say little about differential responsibility for different roles in a group (‘role responsibility’), except to say that responsibility “may vary with [the members’] roles in the group” (164). Intuitively, it seems that a members’ responsibility should be proportionate to the member’s causal contribution to a group action A—although if Kutz is right, a member can be responsible for A even if he did not causally contribute to A, if the person is a voluntary group member (recall the Dresden firebombing example). Kutz says that acknowledging complicity forces us to reject the control condition of responsibility. While I agree with Kutz that participation confers responsibility in a certain sense, I disagree with his claim that this forces us to reject the control condition. The way that I interpret complicity cases is to see group membership as signalling support for the group’s ethos, and thereby causally contributing to the public credibility and acceptance of that ethos, enabling the group to realize its shared ethos in a particular action. Thus, a member may be indirectly responsible for a group action A, even if she did not enthusiastically support or causally contribute to A. This is because group membership enables the performance of collective actions that realize the group’s shared ethos.
This interpretation of indirect-member-responsibility is compatible with a control condition, if group members are held (indirectly) responsible for the collective actions of the group only if they can voluntarily renounce their membership, and thus have control over whether or not they are members. (Someone forced to join a club at gunpoint, or under other coercive circumstances, is not a voluntary member).
Now, List and Pettit say that group agency is morally potent (in part) because the group can be held responsible, even if no individual member is responsible. Individual members might be non-responsible if “they are blamelessly ignorant of any harm collectively done” (166), or (presumably) unaware of any good collectively done. List and Pettit apparently see ignorant members as, in some cases, incapable of controlling the group’s actions, and thus not responsible for them.
To my mind, it is a mystery how a group agent can satisfy the control condition when no individual member does. Attributing responsibility to corporations, independent of any individual scaffolding for this attribution, requires an ontological leap of faith that contradicts the authors’ commitment to a naturalistic ontology.
I think that we should relax the conditions of individual responsibility to prevent this gap from arising. Moreover, this move is consistent with the majority opinion in responsibility theory, which is that individuals can be responsible for ignorance and neglect (e.g., Holroyd 2012, A. Smith 2005, H. Smith 2015, Arpaly 2015, Sher 2010). One way that an individual can be responsible for ignorance and neglect is if she had indirect control over a certain action or outcome. An agent has indirect control if she is in a position to foresee and influence an action or outcome, or she was in such a position at some point in the past; it is not necessary that she be able to foresee/influence a particular action or outcome on the basis of her extant motivational psychology. This indirect reading of the control condition inculpates leaders (executive members, typically) in the decisions and actions of their corporations, under ordinary circumstances, seeing that they are in a position to govern the corporation responsibly, irrespective of whether they actually do. The exception is if someone interferes to hijack the corporation’s decision-making process, in which case the intervener is responsible for the corporation’s actions. In either case, some intentional agent is responsible for the corporation’s actions. This eliminates the gap between corporate responsibility and individual responsibility.
The indirect control view also explains the common-sense intuition that people like Brock Turner can be responsible for expressing rape-positive attitudes: because we ordinarily (except in cases of brain damage and so on) have indirect control over the cultivation and expression of our implicit states over time. Holroyd points to cognitive exercises such as exposure to countersterotypical exemplars and implementation intentions, which produce quick results that can be measured by experimenters; but long-term exposure to counterstereotypical cues and ideologies of resistance, etc., is likely to be even more effective at suppressing and de-sensitizing malignant implicit biases, making them less salient. It is hard to imagine that Turner would have raped an unconscious woman and then invoked a litany of rape myths in his court statement if he had made radically different life choices, such as: taking women’s studies classes, reading feminist philosophy, exposing himself to feminist discourses and ideals, and so on. Was it insurmountably difficult for Turner to take these steps? Surely not. Turner, after all, enjoyed many intersections of privilege as a white cisgender male. If he didn’t understand consent or respect, this was due to his own choices. Given that he had, at least, indirect control over his choices and motivational states, he is responsible for expressing them.
Turner is, furthermore, responsible for more than the effects of his actions as a private agent, since he acted as part of a chain, and so his actions were more consequential than those of a lone private agent: they were more effective at normalizing rape myths and victimizing women, since they were part of a broader system of interconnected practices. Turner thus bears responsibility for a consequence enabled by many actors, which could not have been effectuated by any sole person.
Members of beneficial chains of actors also have a degree of indirect control over the acquisition and expression of their shared insights, hopes, and aspirations, and are thus responsible for the expression in these states in their chained actions, and for the aggregate effects of these chained actions—often, radical social change. Again, no private agent can effectuate the transformative effects enabled by chained actors.
The denial of group responsibility absent individual responsibility does not mean that group agency is redundant. Rather, members of group agents can bear responsibility for the effects of the group, which no private agent has the capacity to bring about; and likewise, chained actors have the capacity to bring about effects that no private agent, or group agent, could bring about—often, radical social change.
- Concluding remarks
I have argued that chains of actors are quasi-group agents, which satisfy List and Pettit’s conditions of group agency minus the constraints of knowledge and common awareness. Members of quasi-group agents have sui generis motivational and normative properties, which distinguish them from both private and group agents. This category is critical to understanding social reality as fluid and tension-laden. I made some programmatic remarks about responsibility in cases of group and quasi-group agency, but much more could be said on this point. I leave this for another day.
Ariely, D., & Loewenstein, G. (2006). The heat of the moment: The effect of sexual
arousal on sexual decision making. Journal of Behavioral Decision Making, 19(2), 87-98.
Bratman, M. E. (2013). Shared agency: A planning theory of acting together. Oxford
Bradley, R. N. (2016). Afterword: How I Use# BlackLivesMatter as an Entry Point
and a Disruption Tool. Meridians: feminism, race, transnationalism, 15(1),
Cleary, T. READ: Full Letter to the Judge by Dan Turner, Brock’s Father. Heavy. 29
Aug 2016. Accessed 08 May 2017. http://heavy.com/news/2016/06/brock-turner-father-dad-dan-turner-full-letter-statement-stanford-rapist/
Doris, J. M. (2002). Lack of character: Personality and moral behavior. Cambridge
Fischer, J. M. (2011). Deep control: Essays on free will and value. Oxford University
______. (2006). My way: Essays on moral responsibility. Oxford University Press.
Gafney, W. C. (2006). A Black Feminist Approach to Biblical Studies. Encounter,
Harman, G. (1999, January). Moral philosophy meets social psychology: Virtue ethics
and the fundamental attribution error. In Proceedings of the Aristotelian
society (pp. 315-331). Aristotelian Society.
Haslanger, S. (2017). Racism, Ideology, and Social Movements. Res Philosophica,
Hood, B. (2012). The self illusion: How the social brain creates identity. Oxford
Isaacs, T. (2011). Moral responsibility in collective contexts. OUP USA.
Kane, A. E. (1997). Theorizing meaning construction in social movements: Symbolic
structures and interpretation during the Irish Land War, 1879–1882. Sociological Theory, 15(3), 249-276.
Kutz, C. (2007). Complicity: Ethics and law for a collective age. Cambridge University
Lear, J. (2006). Radical hope: Ethics in the face of cultural devastation. Harvard
Levy, N. (2015). Neither fish nor fowl: Implicit attitudes as patchy endorsements.
Noûs, 49(4), 800-823.
List, C., & Pettit, P. (2011). Group agency: The possibility, design, and status of
corporate agents. Oxford University Press.
McGeer, V. (2008). Trust, hope and empowerment 1. Australasian Journal of
Philosophy, 86(2), 237-254.
Medina, J. (2012). The epistemology of resistance: Gender and racial oppression,
epistemic injustice, and the social imagination. Oxford University Press.
Mills, C. (2007). White ignorance. Race and epistemologies of ignorance, 247.
Tuomela, R. (2013). Social ontology: Collective intentionality and group agents.
Oxford University Press.
Strain, M. L., Martens, A. L., & Saucier, D. A. (2016). “Rape is the new black”: Humor’s
potential for reinforcing and subverting rape culture. Translational Issues in
Psychological Science, 2(1), 86.
Paiella, G. 2016. Here Is Brock Turner’s Statement to the Judge. In New York
Magazine, 08 June 2016. Accessed 08 May 2017.
 There may also be morally neutral chains but I will ignore them, as they are not amenable to moral appraisal and are thus less interesting and less politically salient.
 In fact, ordinary hiring processes are biased, but hiring committees at least apply some transparent criteria, in addition to acting on implicit bias.
 See Neil Levy 2017 for an account of implicit biases as ‘patchy endorsements,’ with propositional structure like beliefs, but which are not responsive to other mental states.
 No one can say for sure whether Turner knew he was committing rape, but we can assume for the sake of argument that he was acting under a degree of (what I will later argue is culpable) ignorance. What I am trying to show is that if someone promotes rape myths unwittingly in his overt deeds and actions, this suffices to classify him as a member of Rape Culture. It goes without saying that a self-aware rapist is a member of Rape Culture.
 I am going to adopt List and Pettit’s control view. There is not space to survey the practically innumerable variations on moral responsibility, so I will set them aside. Assessing non-control theories of responsibility vis-à-vis group agency is a topic for another paper.