Group Agents and Their Responsibility

Abstract

Group agents are able to act but are not literally agents. Some group agents, e.g., we-mode groups and corporations, can, however, be regarded as functional group agents that do not have “intrinsic” mental states and phenomenal features comparable to what their individual members on biological and psychological grounds have. But they can have “extrinsic” mental states, states collectively attributed to them—primarily by their members. In this paper, we discuss the responsibility of such group agents. We defend the view that if the group members have accepted the group agent’s (attributed) attitudes and are committed to them, we can favorably compare the situation with the case of individual human agents and a group agent can be regarded as morally responsible for its intentional activities.

Responsibility of Groups: The Main Ideas

In this paper, we operate on the assumption that there are groups (e.g., corporations and especially we-mode groups) that can act intentionally and on that basis can be regarded as being responsible for their actions qua groups or collectives. (Tuomela 2007, 2013)

Some groups, e.g., we-mode groups, can be regarded as functional group agents that do not have “intrinsic” mental states and phenomenal features (comparable to what their individual members on biological and psychological grounds have), but they can have “extrinsic” mental states, states collectively attributed to them—primarily by their members.

A group agent such as a business company can normally be regarded as a single entity involving individual members, some constitutive features (e.g., goals, beliefs, norms, etc.) and sometimes more than that—e.g., in the case of corporations as including material possessions and legal properties.

We take the social groups under discussion here to be robust and resilient objectively existing social systems consisting of, or at least having, interrelated and dependent members (and the mentioned kinds of constitutive features). Their members (typically individual agents) can be regarded as intrinsically intentional agents if they are individual human beings. (Individual agents generally have also extrinsically intentional attitudes—e.g., we-mode attitudes in we-mode groups or other social role-based attitudes comparable with attitudes an actor in a theater play may have.¹)

Our claim, to be qualified below, is that a well-functioning we-mode group generally can qualify as a bearer of and be fit for responsibility. In it the members (when they function as group members and engage jointly in we-mode we-reasoning) are supposed to have collectively accepted the group’s ethos—its basic goals, views, norms, and practices (see Appendix). Thus the responsibility attributed to them qua group members will be responsibility viewed in analogy with the responsibility of ordinary intentional human agents as such (independently of whether or not they are functioning in a group context). A group’s responsibility, of course, is connected to its members’ collective responsibility at least when they act as group members, but—as will be seen—it still often differs from the members’ full-blown joint responsibility.

Even if groups (such as we-mode groups and corporations) cannot be responsible in the sense individual agents can be, the present view holding groups fit for responsibility in suitable cases still is intuitive. In contrast, the strict classical idea that assumes the actus reus and mens rea principles in unison (see below) fits individual responsibility in the case of standard individual action, but it does not apply to the group agent case, as will be seen. It was not originally created to fit either and it goes counter to common intuitions concerning group responsibility.

Note that corporate intentions and actions are not reducible to their members’ intentions and (participatory) actions in part because corporate actions require a suitable institutional context and the existence of certain constitutive rules (e.g., an ethos). The central loci of corporate moral responsibility are the group and its members, their attitudes and actions as group members (and, on the other hand, as private persons by the fact that they have volunteered to act as group members).²

Collectivist attributions of responsibility, attributions of the form “Group G is responsible for X,” “Danske Bank is responsible for misleading its customers” are typically made, at least in part, for epistemic reasons, that is, because large groups or corporations are opaque as to their internal structure. People outside the group are not aware of the chain of responsibility internal to the group or corporation.

However, we think there is more to be said in favor of such collectivist attributions in terms of the attributability view of responsibility. The central concern in views that hold responsibility to be a matter of proper attributability is whether the agent’s action or attitude discloses her evaluative judgments or commitments.

According to the attributability view, an agent is responsible if a fault or credit is properly attributable to her. If the action discloses evaluative judgments or commitments of the agent, it is thus attributable to her. That is, if an action of an agent is faulty and discloses the evaluative judgments or commitments of the agent, the agent is responsible for the fault. The responsible agent is the one whose evaluative judgments and commitments are disclosed or expressed by the action. This view obviously presupposes that entities fit to be responsible in this sense must be capable of intentional action and have the capacity to make evaluative judgments and commitments. That is, it is an agentive view.

If an entity is capable of intentional action and making evaluative judgments and commitments, then it is an agent and fit to be held responsible, at least in the sense of attributability.

We often attribute actions to groups. We accordingly assume that at least certain kinds of groups (e.g., we-mode groups) are capable of intentional action. (Tuomela 2007: 128–134, 2013: 160–161) Arguably, groups can also make evaluative judgments and commitments. We suggest that the group’s (group agent’s) evaluative judgments and commitments get their expression or are disclosed in the actions of a we-mode group. These judgments and commitments attributed to the group agent are generally based on the group members’ proposals for what the judgments and commitments of the group agent should be. The attitudes and actions of the group agent are not properly attributable to individual members. As claimed, there are actions that are properly attributable to a group, that is, intentional actions that disclose evaluative judgments or commitments of a group. It seems that the group is the primary candidate for being a responsible agent in such a case.

According to Robert Kane’s notion of ultimate responsibility that is geared to the individual agent case, we can be responsible for current actions, ones that are essentially determined by our character and values, as long as we formed that character ourselves by earlier free actions that Kane calls “self-forming” actions. (Kane 2007) A collective rendering of this seems to apply to groups: A self-forming action in the group case is the collective decision making related to the ethos of the group, which provides the group with an analogue of character and values of an individual agent. After their having formed the group’s ethos, the members collectively commit themselves to acting in accordance with it. As a result, the individual members are not allowed to act against the ethos, and the ethos can only be changed or dropped by collective decision making. Consequently, the ethos becomes the collective character of the group that determine the actions of the individual members qua group members. In the case of an autonomous we-mode group the individual members participate freely in the formation of the ethos and thus bear personal responsibility for their actions.³ Most centrally, a we-mode group is responsible for its actions, and consequently its members are jointly responsible as group members for them. Also, each member is responsible for those actions qua group members. As to dissidents’ actions, see Appendix.

There are still more complicated cases such as those in which the individuals collectively accepting the ethos for the (we-mode) group are not the same individuals as those who act in the name of the group. The outcome of this reasoning is that there are some individuals jointly responsible as a group for the actions of the acting group because of the first group’s role in the self-forming actions of the group; importantly, these individuals are not here the same individuals as the ones whose actions, qua group members, constituted the action of the group. Still, those who acted are collectively committed to the same “group character-giving ethos” when acting because of, and in accordance with, it, and bear responsibility for their actions as group members.

In what follows our discussion focuses on we-mode groups and corporations, that is, on group agents that are subject to attributions of responsibility and typically have an internal structure in virtue of which they can be characterized in terms of chains of responsibility. (Baier 1991)

In a hierarchically structured group individual members seem to have a double role when it comes to the normative evaluation of their actions qua members of the group. First, the members are internally responsible for discharging their role or task responsibilities in accordance with the ethos of the group. Second, they have an individual (moral) responsibility for their actions should they act in accordance with the ethos or not. If the ethos of the group is deplorable, morally unacceptable, the members seem to be in a position in which they, on the one hand, ought to act in accordance with the ethos because of their commitment to it and their collective acceptance of it. On the other hand, qua individuals the members of the group ought not to do what is morally wrong, and they might even be morally obligated to become whistle blowers. In such cases the all-things-considered assessment of what the members ought to do has to be made on a case by case basis. However, it may be that refusing to act in accordance with the deplorable ethos is so heavily sanctioned that it would not be reasonable to require the individual to refuse (this may be the case in an military unit).

There is an internal perspective and an external perspective on what is correctly to be attributed to the group and what not. Strictly speaking, if the members intentionally act against the ethos of the group, then they are not actually acting qua group members and the action is not to be attributed to the group. This is so from the internal point of view, but from the external point of view the epistemic access to the ethos and the authorization of the members may be missing and the action may be attributed to the group even though it is not the group’s action in light of the ethos. In a case like this the correct attribution of responsibility distributes among the members and they individually or jointly bear responsibility for such action.⁴

Responsibility of We-mode Groups

In this section, we will consider the responsibility of group agents such as we-mode groups. A we-mode group is a group that can act and does act as a group on the basis of its members’ group-based we-thinking and we-reasoning, or, briefly put, we-mode thinking. (Tuomela 2013) A we-mode group makes the members strongly interconnected—the group is supposed to function as a whole that consists of the individual members’ performing their parts of the group’s action.

A we-mode group is based on the group members’ we-mode we-thinking and we-reasoning with the result that all their we-mode attitudes and actions must (at least ideally) satisfy the three criterial elements of the we-mode the group reason, collectivity, and collective commitment conditions. (Tuomela 2007: Chap. 2, 2013: Chap. 2) The group reason element concerns the reasons “given” by the group to the members for their participation in its activities. The collectivity condition here refers to a kind of “being in the same boat” condition concerning the members, and collective commitment ties them to the group, especially its ethos, and the mutual commitment to the fellow members concerning the promotion of the group’s ethos involving its constitutive goals and other goals.⁵

A we-mode group, as treated in Tuomela (2007, 2013), is generally viewed as an autonomous egalitarian group where the only normative structural group connections, if any, between the members are based on group-internally constructed operative-nonoperative member-level normative connections.⁶

When a paradigmatic we-mode group acts, there are operative members acting for the group as ethos-respecting group members (in some cases all the members might be operatives). The operatives are in standard cases identified and collectively accepted by the group members for specific tasks. To put the matter in terms of joint action, the basic idea is that the members share a (we–mode) we-intention expressible by “We will do Y together” and, when the intention is satisfied, the joint action expression “We did Y together” applies. The intentional agent of the intention is “we” (that is, the group members forming a non-distributive “we”), and its content is the members’ jointly intentionally performed joint action.

Group responsibility is often taken to rely on the members’ attitudes and actions as group members.⁷ In the case of a we-mode group the group members are extrinsically involved in the group’s responsibility through the actions they perform qua group members. Yet if the members of a group act together or jointly in a strong sense so that we can speak of them acting as a group and as their being responsible as a group, it thus seems plausible to regard their group as being fit for responsibility.

A we-mode group that is not dominated by another group and that can itself determine its ethos is not literally an agent, but it can yet be regarded as a group agent on the basis of its capacity to act as a group (as a unit). As it is not a full-blown agent with a biological constitution it is not an intrinsically intentional agent (in the relevant sense). As a group, it can only have extrinsically intentional attitudes and mental states, viz. states that have been constructed for and attributed to it, typically by its members, while its members qua private persons normally are capable of intrinsic intentionality. Analogously, we argue that a group qua group cannot, so to speak, be “intrinsically” responsible (in the sense individuals are when acting as private persons) for its activities. The group members are capable of having intrinsically intentional mental states, but, strictly speaking, when functioning as group members they only operate on the basis of their extrinsic mental states deriving from the group’s “mental” states that are comparable with role states in a theater play. Note that the latter are attributed to the group by the members—via their proposals that are collectively accepted by the members as the group’s states. The members can have intentional joint mental states (e.g., intentions). But those states often involve compromises and the like. The compromises concern a group attitude based on the (partly) inconsistent proposals by the members. Those proposals concern the group attitudes, typically extrinsically intentional ones, on which the members’ functioning in their roles in the group are to be based. These “role attitudes” are typically extrinsically intentional—and not intrinsically created by their bearers. This situation arises because we are here dealing with the members’ proposals for group attitudes. Note that putting together the members’ attitude proposals may create consistency problems in addition to those that compromises involve. The collective acceptance of such proposals for the purpose of creating unique group attitudes does not always go smoothly. (List and Pettit 2011: Chap. 2).

Assuming that the above is right, suppose now that the group members have collectively accepted and thus created the group’s extrinsic “mental” states and have also themselves “internalized” them. Given this, we can see an analogy between the present kind of group and the case involving intrinsic attitudes and regard an autonomous we-mode group (and other similar groups appropriately organized for action) as being morally responsible for its intentional actions in an approximately intrinsic sense—see Appendix for precise conditions for responsibility (and recall note 3). As is the case with a group’s mental states and actions, also the group’s actually taking responsibility for its actions is analyzed through its members’ mental states and activities, their acting jointly as a group.

If a we-mode group with external and internal autonomy is normatively responsible for an action or outcome X, then in general no one of its members is solely normatively responsible for X as a group member. This claim is about responsibility as a group member but bypasses the question of his purely personal (or “private”) moral responsibility.⁸ To take an example of an internally non-autonomous hierarchical group with a dictator, for example, an army unit closely simulates this case. The members cannot voluntarily leave the group (or can only do so on pain of heavy sanctions). The dictator’s power can be enforced by means of strong punishment, in special cases even death, if the order is disobeyed. In such a group, to mention an idealized case, the dictator will normally alone be fully responsible for X as a group member, as the other members do not act freely and as they obey the dictator’s orders being coerced to do what they do as group members. Even so, they would bear at least some purely personal responsibility for their actions.

In a we-mode group the holistic idea of the members’ being to an extent responsible for the others’ undertakings qua group members holds true. All the group members acting appropriately as group members are, or at least ought to be, collectively committed to the group’s action, and they accordingly collectively bear responsibility for what the group did, and this includes mistaken actions and dissidents’ actions (see Appendix). In all, in a we-mode group (with or without an internally agreed operative-nonoperative division) the members are responsible for the group’s action both as group members and to an extent as private persons—first because of the holistic, interconnected nature of the we-mode group and second centrally because intentional group action requires the participation of group members as group members—when they function as sentient and morally sensitive human beings.⁹

Note, however, a member of an autonomous we-mode group (one capable of forming its ethos and, in principle, of acting freely) can at least partly escape attribution of responsibility to her qua group member in a case where the group is responsible for a blameworthy action X if she was not involved directly in the actual causal production of X and if in addition she publicly disassociates herself from the production of X (e.g., by explicitly publicly speaking against the production of X before its occurrence and perhaps even by disclaiming her membership in the group).

Consider now the responsibility of groups for their actions. “Actus reus” and “mens rea” are classical principles and requirements for intentional responsibility as well as blame and punishment. Actus reus is the requirement of presence of the responsible agent’s own action, and mens rea is the requirement that the action was performed intentionally by the agent in question. These two principles have been discussed in the case of the responsibility of individual agents, for which case they seem largely appropriate. Manuel Velasquez puts together the classical actus reus and mens rea principles as follows: “Moral responsibility is the kind of responsibility that is attributed to an agent only for those actions that originate in the agent, in so far as the action [is] derived from the agent’s intentions (the mens rea requirement) and from the same agent’s bodily movements (the actus reus requirement).” (Velasquez 1983: 114) The requirement of the presence of bodily movements is clearly too strict in general, but otherwise the account is acceptable as an ideal. Yet, we argue that it goes against the common view that groups are often responsible for what they do and cause.

In the strict classical account under discussion, moral responsibility of an act or outcome can be attributed only to the agent who originated the act in his own body in cases of bodily action in the relevant activities of his brain and body parts over which he has direct control. This requirement cannot be literally satisfied in the case of a group or a corporation (a corporate agent), for it does not have a body that it could move. Collectivities like corporations act only through their members’ actions. But, strictly speaking, those actions are not the corporation’s actions. A corporation’s actions are constituted or brought about by the members’ (bodily) actions that are not in its direct control, but at best under its indirect control. (However, the members have direct control of their participatory actions.) Thus a corporation is not fit for responsibility in the strict classical sense (that relies on actus reus and mens rea).

Corporations and “tightly connected” groups such as we-mode groups can yet be responsible for their actions in a slightly different sense through their members’ participatory actions: Assume that the members on the basis of their joint intention realize their intention and act jointly as a group (rather than in a weaker sense of sharedness or of interaction). This can normally be taken to entail that the group acts intentionally as “one agent,” and is thus responsible for its action.

Of course, the members’ joint action does not quite amount to the group’s action in full analogy with an individual agent’s intentional action, as the biological and psychological unity between the intention and ensuing action in the individual agent’s case is clearly different. But if we can realistically assume that a group’s action here is what the members do as a unit or as one agent, we seem to get close to the case of an individual agent’s case and other cases satisfying the actus reus–mens rea unity requirement. Yet full unity cannot be obtained on conceptual and metaphysical grounds: the group members’ mental states and actions in general cannot be aggregated or otherwise combined to become, respectively, the mental states and actions of a group agent. (Tuomela 2013: Chap. 5) The actus reus–mens-rea requirement does not in general apply to the group case.

Even if groups (such as we-mode groups and corporations) cannot be responsible in the sense individual agents can be, the present view holding groups fit for responsibility in suitable cases still is intuitive. In contrast, the strict classical idea (that assumes actus reus and mens rea in unison) fits individual responsibility in the case of standard individual action, but it does not apply to the group agent case. It was not originally created to apply to the group agent case either and, as a consequence, it runs counter to common intuitions concerning group responsibility.

In the present account, then, there are the two aforementioned conceptually involved elements—group members and the group—as central elements of group responsibility, and we can refer to the group level (the group viewed as a group agent), the collective or jointness level (viz. the members viewed collectively or jointly), and the purely individual level (individuals viewed as separate individual group members or as private persons).

Our analysis of group responsibility engages each of these three levels. In normal cases in which a we-mode group is responsible, the members are jointly responsible for what the group does intentionally (see Appendix for other possibilities), and every member qua group member is responsible for the group’s actions and indeed for every other group member’s participatory actions.

The above idea of the responsibility of a group and also of its members for one and the same outcome has been disputed in the literature. The core of the criticism is that this kind of dual responsibility idea is redundant as only individual members’ responsibility really counts in attributions of responsibility: If the individuals are collectively causally responsible for an outcome, the group cannot at the same time be causally responsible for it. The alleged causal group responsibility is taken by the critics to be redundant and hence it suffices to deal with the individual group members’ causal responsibility and control. But we dispute the redundancy claim: both the group and its members generally bear responsibility (see below).

Group responsibility typically connects to the members’ responsibility by entailing their responsibility qua group members, or, as we may put it, it entails that the members qua members are jointly or collectively responsible for the action for which the group is responsible. The members’ responsibility need not always be responsibility qua group members, but might be their responsibility qua private persons in the case of a group with loosely connected members (e.g., consider an organized tourist trip to Paris by some people, or think of an I-mode group). In all cases the members are also privately morally responsible for their own actions.

Our general view of a group’s attitudes is that they are irreducible to the members’ attitudes (Tuomela 2013: Chaps. 3, 5, 2016), and it can accordingly be suggested that the same holds for the case of group responsibility.

Basically, the group’s (or group agent’s) control can be taken to amount to the group’s filtering for what is in accordance with the group’s ethos and excludes other possibilities.¹⁰

This kind of filtering approach involves a kind of concrete planning by the group for what its members should bring about by their actions as group members. Such bringing about of a planned and intended outcome by the group in many cases requires additional planning and decision making by the group concerning what is to be done, when, where, and how, etc. In all, the group members do what they do based on the background platform that the group, so to speak, represents when viewed as an (extrinsically) intentional agent. The group not only plans and initiates its action but also monitors and controls that the group members carry it out when the circumstances are feasible. Loosely, at least in simple cases the group is constituted by its members and its main principles (ethos), and the members not only plan and reason but also realize the plans through their actions with the idea that the group is constituted by “us-together” and its ethos. If the group disintegrates after its intentionally performing, say, a blameworthy action (that perhaps originally was meant to be a praiseworthy action in accordance with the group’s ethos but was not performed with the care it should have been performed), the members are generally collectively responsible for it. Here the blameworthy result occurred because they had not acted properly as group members in accordance with the group’s plan. The fact that the ethos of the group in question is fully shared (or so we presently assume) and mutually known by the group members to be shared can be taken to play the role of filtering the actions and leaving out only the unfeasible ones, and leaving the group perhaps only with a single action normatively required by the ethos or possibly by the group leader and assigned to the group on which to act.

The group is sometimes a mere dispositionally passive cause but typically an occurrent active cause of the outcome in question—through its members’ functioning in the right way as ethos-furthering group members—in some cases based on a leader’s instructions or orders.¹¹

Assuming that the members of a (we-mode) group identify with the group and thus are disposed to act in accordance with its ethos, the group necessarily figures in the group members’ minds and actions and motivates the members’ thinking and acting in the direction required or triggered by the ethos. The central idea here is that the group serves to initiate and maintain the members’ action through the mentioned psychological mechanism. The group, through its ethos suitably internalized by its members and through their accordingly motivated actions, can be said to intermediately cause the group’s relevant action (the “occurrent” or “active case”) or at least be a standing cause (the “dispositional” or “passive case”) for it. Such a standing cause may become manifest in appropriate circumstances. As a result the group will be responsible both for its praiseworthy and its blameworthy actions and their outcomes which take place through the members’ actions. As the group’s actions are constituted by its members’ actions, the group will be responsible also for its members’ participatory actions (and lack of them in other cases). This kind of group responsibility involves that the members of a we-mode group are responsible for their own participatory actions as well as typically to an extent responsible also for the other members’ participatory actions (e.g., in situations in which some of them require help). We suggest that the members may be regarded as responsible for what the group has intentionally done even in cases where the group disintegrates after the action has been performed and where the members had valid excuses for non-participation in a case of a blameworthy group action or group-induced outcome.

David Copp (2007) has argued that group responsibility does not always entail group members’ responsibility. There might even be cases where a group as a group is responsible for its action while its members are not responsible in any relevant sense.¹² The basic reason for this is that while the members were required to participate in the performance of the action, they may have had private reasons (viz. non-group reasons) against participating in the group’s action (e.g., one concerning the prevention of a morally disastrous outcome) although they were normatively required to participate. Those private reasons were of the kind that they justified non-participation by overriding the group-based reason for participation. An example would be one where all or some of the members would have been required to stop the group’s blameworthy action, but those members yet did not perform their ethos-based participatory actions and had their overriding private justification for their intentional non-participation. The overall result is that in cases where the group disintegrates (ceases to exist) after its blameworthy action there is no existing group agent to blame nor is any group member to blame as all were excused in Copp’s special case.¹³ However, such cases are rare. We suggest that the members may still be regarded as collectively responsible in cases where the group disintegrates after the action has been performed. This seems also to accord with common practices in ordinary life.

Responsibility of Corporations

Our discussion in this section concerns social organizations and it concentrates on social corporations. The purpose of the discussion is to illustrate the role of internal and external authorization in business corporations or in a public organization like the state, and thus also to bring philosophical theorizing closer to real life cases. Our discussion is based on typical corporations with a linear hierarchical power structure, but the basic features of our account apply also to other authority structures.¹⁴

A social organization consists of an ethos (basic goals, values, standards, and so on) and a set of positions that are normatively interlocked with each other in terms of normative “task-right” systems. (Tuomela 2002: Chap. 7) Each position involves a task-right system for the position holder concerning what tasks he ought to do and what he may do when acting in his position. The interrelations between the positions contain power relations and informational relations. For instance, a position-holder can have the power to order the holder of another position to do something, and different position-holders may have differing amounts of knowledge concerning the organization’s activities.¹⁵

Some organizations can be viewed as functional we-mode groups, viz. in cases where their members (or a substantial amount of them) collectively construct its ethos in we-mode terms and employ we-reasoning. For example, a union of nurses or of professors might be examples of organizations that can be viewed as a functional we-mode groups. The professional rights and duties of the position holders in question are vested by relevant authorities with the power to regulate and control their practices.

Groups with a traditional linear hierarchical power structure can be described as having an internal chain of responsibility which can be latched on the task-right system of the group in terms of which the positions of the hierarchical power system are defined. (Tuomela 2007: Chap. 2, 2013: 167–172; Baier 1991) Following Kurt Baier’s formulation, a simple chain of responsibility can be characterised as follows: Person A is responsible to person B. A is task-responsible for doing or seeing to it that X. A can discharge her task responsibility or fail to perform the task expected from her. B has a task to supervise A who is responsible to B, and it may be that B in her turn is supervised by C and thus B is task-responsible to C for supervising A. Part of the supervising person’s task-responsibility is to hold A responsible for discharging her task-responsibility. Holding A responsible consists of making him give an account of how he has handled his task responsibility. If A has discharged the task responsibility, no more needs to be done, but if A has failed in discharging the task-responsibility, he is answer-responsible for it, that is, A must explain why he failed in performing the task-responsibility. If A cannot give an exonerating, exculpating, or excusing explanation for the failure, then A is culpably responsible for not performing the task-responsibility. If A is culpably responsible for the failure of discharging the task-responsibility then B ‘s task responsibility is to move to stage two of holding A responsible, namely, to rectifying measures (dress him down, admonish, reprimand, warn, fire or punish).¹⁶

To comment on some features of authorization, let us first consider groups with internally authorized operative members, especially leaders. Such leaders are assumed to be collectively authorized (and given the power to act) by the group members. These leaders have the legitimate authority and right to decide (concerning a specified set of topics) about group activities—e.g., about what the group ought to do in a certain situation. Such decisions entail correlated obligations for the members to act accordingly. There is, then, appropriate justification for the non-operative “rank and file” members to take the operative members’ (leaders’) decisions (directives, orders, etc.) as “desire-independent” preemptive reasons for action when acting as full group members.

As to externally authorized groups (viz. groups in which an external agent, individual agent or group, has acquired the power to control the group in question), also in this case there is an obligation for group members to comply. In externally authorized groups normative justification may be lacking (think of dictatorship based on brute force), but the members may still, perhaps grudgingly, obey the externally based directives and thus give factual legitimacy to them.¹⁷ A business corporation is typically based on external authorization, an externally created, hierarchical authority-power structure.

Let us now consider social organizations consisting of hierarchically embedded groups (or subgroups or teams, if you like) where only the largest collection of such groups (a “supergroup” such as a society or state) possibly is internally authorized, and the (hierarchically ordered) subgroups derive their authoritative power from the supergroup concerning the group-relevant topics in question.

We argue that a democratic state is an example based on internal authorization at bottom, although it superficially shares several features with externally authorized cases. In this kind of case the parliament and the government in principle are under rather direct democratic control but, in contrast, the state institutions (e.g., the educational system—schools, universities, etc.) are under the authority of the relevant ministries that derive their authority from the highest, supergroup level. Baier characterizes authority and responsibility relations in a democratic society pithily as follows:

…in a democratic society the members play two quite different and opposed roles, and so does the government. While as subjects the members of the society occupy the lowest rank in the political order, as sovereign they “collectively” occupy the highest rank and are therefore responsible for its subordinates, including the government and its organs. While as ruler of its subjects the government occupies the highest rank of the political order, as servant, representative, or agent of the sovereign body of citizens it is subordinate and responsible to that sovereign. (Baier 1991: 215)

In such a system, for example, a university professor will have to take orders from the administrative bodies of the university, and their activities again are authorized and constrained by the orders of the ministry of education. Accordingly, there is society-level authorization even for a professor’s daily activities and he could in some contexts act as if he had taken part in the authorization of those (lower-than-highest-level) leaders whose orders he is supposed to obey, although he has his say only in state level elections and in the determination of the political values and views on the supergroup level.¹⁸ In real life there in many cases is a “democracy gap” between “local” authorization of the direct kind we may have in small groups and the more “global” case with indirect contact with authorities—think of the supergroup case.¹⁹

A corporation is a social organization that can be viewed as a group agent with a task-right system. In view of the positional power structure of the organization, those higher up in the hierarchy are responsible for the position holders’ actions who function lower in the hierarchy (who in turn are responsible for doing what they have been ordered to do). This kind of power and responsibility structure obtains because the consequences of power exercise flow downward in the hierarchy. In contrast, in the owners’ (shareholders’) group, everyone is an operative member.

The hired persons might not be allowed to directly change the ethos of the organization, but they may influence or induce ethos change, although the owners’ group is in the last instance responsible for such changes.

An organization’s decisions are generally constituted by its authorized operatives’ (leaders’) joint decisions; and when it acts, it performs its actions in virtue of its authorized operatives’ actions. The operative members may vary from one occasion and task to another, and they may also include non-member operatives. Not all operatives need to be group members (cf. hired lawyers, cleaners, truck drivers, etc. appropriately authorized for their respective tasks).

The shareholders generally have the ultimate power in a business company that has a linear top-down power structure. They thus collectively have the power to select a governing board for the company, and derivatively they have the power to select the other functional position-holders—from the CEO down to a salesperson and truck driver. The shareholders’ group may in some cases (e.g. family enterprises) operate as a we-mode group. Their group is a core–level subgroup within an organization, while the functioning operatives (the CEO, various kinds of managers, workers, and what have you) form its hired personnel and indeed another subgroup taking care of the daily affairs of the organization. It is in virtue of their action that specific actions are attributed to the organization. In general, they represent the organization in daily life.

In a typical business corporation, all the position-holders are hired ones and fill some codified positions. Of the mentioned subgroups, the board of directors and the upper-level managers (who relatively independently carry out their company tasks) may work together in a non-autonomous we-mode sense (although they need not on conceptual grounds). Also, the remainder of the hired personnel can in principle act together in a similar sense. Yet the business corporation under the present conception is a group of groups (or subgroups) of which only the subgroup of shareholders can function in an internally autonomous and authorized sense. The subgroup consisting of the board of directors and of (high-level) managers is externally authorized as is the subgroup of selling personnel (and what have you at the lower levels).

External authorization involves at least the possibility to exercise power in a sense that contradicts a subgroup’s autonomy (the capacity to function without interference by an external authority). This is relevant to the problem of whether corporations can be we-mode groups in a full sense. We submit that even when the members of the subgroups would adopt we-thinking and identify themselves with the organization, they cannot form a full-blown paradigmatic we-mode group. At best, they can only form a non-autonomous we-mode group. (Tuomela 2013: Chap. 2)

We argue that the we-mode conditions (the group reason, collectivity, and collective commitment conditions) may be “approximately” satisfied in the externally authorized corporation case, although the full kind of autonomy that we-mode collective acceptance ideally can be taken to involve still is missing. Thus some externally authorized groups (such as business corporations) can in some cases be similar to (paradigmatic) we-mode groups if the group members have the right kinds of attitudes based on the full we-perspective and have the possibility to exit the group. (See Tuomela 2013, for paradigmatic we-mode groups) However, they neither choose the leaders nor the ethos of the group, and in many cases they submit themselves to the leaders just for money (or something comparable).

In some cases even an entire corporation (viewed as a group of hierarchically connected groups) can be regarded as an “approximate” we-mode group—or a collection of paradigmatic we-mode (sub)groups with ethoses compatible with the corporation’s ethos. In those cases the approximate we-mode groups can function as proper paradigmatic we-mode groups as long as that is not blocked by the upper-level officers and ultimately by the owners’ group.

It may be useful here to look at some examples from real life. As to collective responsibility in the case of (autonomous) business corporations (such as Apple, BP, and Tesco), the only autonomous party is the owners’ (e.g. shareholders’) group, as we have seen. The rest of the personnel in the corporation are dependent on this group. The whole corporation can, primarily through its autonomous owners’ group, be regarded as autonomous and responsible to the external society for the actions of the corporation. Example: BP is responsible for bringing about the recent Mexican Gulf oil catastrophe even if the members causing the catastrophe by their actions might not have acted with proper information about the circumstances in question. Yet in legal courts such matters as the members’ informational status and their relevant intentions might be significant for responsibility and sanctions.²⁰ Note that the position holders at the lower-levels are accountable to the next level above for their actions. The normative hierarchy of the position holders here stands in the way of everyone being responsible for the whole group’s actions (this is a disanalogy in relation to the standard we-mode case). Although there is top down responsibility (e.g., leaders being responsible for what the lower level rank and file members do), in the opposite direction there is only answerability—all lower level position holders are answerable to the upper level for their performing their tasks appropriately.

In a corporation, and especially an approximate we-mode one, there is also codified “social commitment” between the position-holders: a position-holder may be responsible to another one (if they have interlocking positions or are somehow in a suitable power relation to each other). In any case, a position-holder is committed to the company to perform her tasks. All the position holders hired by a company can be operative agents for some tasks (e.g., decision making or for concretely carrying out the decisions). However, hired agents who are not position-holders but are only temporarily connected to the organization in order to perform some tasks are not operative members but rather are more like the company’s “means” or “tools.” The ultimate core of an organization is constituted on the one hand by its owners (if it has owners in the legal sense) and on the other hand by the positions that constitute it. The owners determine (often via suitable operative members) the ethos (and thus the general content) of the organization, and the position–holders act to achieve and/or maintain it. Some position-holders may in fact have as their task to reformulate the ethos when that is needed.

In hierarchical organizations (such as a typical business corporation or a university) power exercise is directed downward: upper-level members or groups control the level below. There may not be much difference on the level of practical activities in these cases, although the business corporation is often externally authorized while a university might be internally authorized. In both cases a group member may somewhat, metaphorically speaking, have to relinquish some of his “natural” authority involving often a simple tradeoff, e.g., instrumentally offer one’s services for money or for some comparable good. Often the same happens also because people may want to get a task performed which requires cooperation by several individuals, as emphasized earlier. This kind of (possibly only temporary) giving up part of one’s de facto authority is typical and indeed necessary for much of social life. A requirement of autonomy in an ontological sense (that is free from contextual empirical constraints) or in a moral sense can still be retained, but as a desideratum rather than a right that others are obligated to respect in all situations.

Appendix

A detailed account of the responsibility of a group agent in the blameworthiness case.

Let us assume that there is a normative standard, S, which is deontically relevant to Y and that the group understands this and is “deontically sensitive” to this fact. In general, S can be taken to specify whether Y is required (namely, in terms of stit, (O(stit(g, Y))), permitted (P(stit(g, Y))), or forbidden (F(stit(g, Y))) or whether some other kind of deontic relationship obtains. Whatever task g adopts for itself, it will be regarded as prospectively responsible for it relative to S.

Based on the collectivity condition and the collective commitment requirement, a we-mode group is at least to an extent responsible for its members' actions as group, and each member of a we-mode group is prima facie responsible to an extent for any other member’s actions as a group member.

Here is an account from Tuomela (2007: Chap. 10), of retrospective normative group responsibility in the blameworthiness sense in the case of a we-mode group or a related kind of organized group (cf. Mäkelä and Tuomela 2002, for an earlier version):

(NRGB) Group g is normatively retrospectively responsible in the blameworthiness sense for a blameworthy action X relative to a deontically relevant normative standard S, where X is a group action performed by g [see (IGA) of Tuomela 2007: Chap. 10] or an individual group member’s (or members’) action as a member of g or an outcome of either kind of action, and S is the normative moral part of g’s ethos if and only if (or to the extent that)

• 0. it was the case prior to action that, compatibly with the normative standard S, group g ought to have seen to it that Y or brought about Y, where Y is an action or outcome of action that is incompatible with X and is to be performed in the group context in which the members are acting in as members of g (in a sense allowing dissidents); and they knew that or, because of the rationality or reasonability standards related to S, they ought to have known that, and

• 1. g intentionally brought about X, or saw to it that X, or intentionally let X happen; or

• 2. g intentionally omitted preventing X (outcome) from coming about or from obtaining or continuing to obtain; or

• 3. g brought about X (or let it happen) as a foreseen but perhaps not an intended consequence of its intentional action (in contrast to (1)); or

• 4. g brought about X as a consequence of its intentional action, or it let X happen while it did not foresee X happening but reasonably ought to have foreseen it; or

• 5. g brought about X, or let it happen, by mistake when intentionally trying to bring about Y as a group, and there was either recklessness or negligence involved in its so acting; or

• 6. (6)(a) when intentionally trying to bring about Y, g ended up bringing about X, or it ended up letting X happen, because of its (dissident) members’ action that it ought to have prevented but did not, although it was capable of it or it reasonably ought to have been capable of it, or (b) the group members had by their attitudes violating S and favoring actions of kind X (perhaps indirectly) facilitated the dissidents’ action that served to bring about X.

In this account, the first clause (0) is thus assumed to hold true no matter which of the disjuncts (1)–(6) are satisfied.