Institutional Entities and Moral Obligation

On the society-centered theory, the question whether institutional entities have moral obligations is the question whether the ideal moral code would include principles “directly regulating” the actions of such entities.

I assume everyone agrees that the United States dropped an atomic bomb on Hiroshima. Agency individualists and the rest of us agree that this event was constituted in some appropriate sense by actions of individuals.⁸ What extremists deny is that this event was an action of the U.S. and that the U.S. is an agent. In general, they deny that true statements that appear to ascribe actions to institutional entities actually do so. They presumably would agree, however, that such statements refer to complex events that in some way involve the institutional entity and that also involve relevant actions of individual persons. Call complex events of this kind “institu­tional actional events.” And call an institutional entity that is involved in an actional event, in the way that the U.S. was involved in the bombing of Hiroshima, a “doer.” One question we have is whether actional events involving institutional doers are, properly speaking, actions of those entities — whether institutional entities are agents capable of performing actions properly speaking. I will return to this question.

To investigate whether institutional entities have moral obligations, according to the society­centered theory, I ask whether the ideal moral code plausibly would include principles or standards directly regulating institutional actional events.

To make the issue more vivid, it might help to think as social engineers who are designing the ideal moral code. The issue is whether we would be well-advised to design the code to include standards that directly regulate institutional actional events or whether we would do better to include only standards that directly regulate actions of individual persons. Suppose we aim to design a code such that people who subscribe to it will be inhibited from supporting or engaging in actional events that involve their state in attacking civilian populations. On the “individualist approach,” we might include in the code a standard such as, “No person who is in a position to support her state’s attacking a civilian population is ever to do so, and no person who is in an official position in a state is ever to perform an action that would be partly con­stitutive of her state’s attacking a civilian population.” On the “collectivist approach,” we might include the simpler standard, “No state is ever to attack a civilian population,” or even the standard, “Do not ever attack a civilian population.” On this approach, for reasons I will explain, we would also want to include in the ideal code a “transitional standard” to the effect that “If a collective entity is required to do something, then its members, and any relevant officers in the collective, are to do their part in bringing it about that the collective acts as required.” In this way, the ideal code would underwrite the existence of a “transitional duty.” The issue is whether to take the collectivist or the individualist approach.

This is a complex empirical problem. As social engineers, we would aim to encourage some kinds of institutional actional events and to discourage other kinds. On the individualist approach, the ideal code might accordingly include a large number of standards such as, “No person who is in a position to support a corporation’s polluting a waterway is ever to do so and no person who is in an official position in a corporation is ever to perform an action that would be partly constitutive of the corporation’s polluting a waterway.” The collectivist approach would seem to provide a simpler strategy. For, given that the ideal code would include the transitional standard, it could include a range of simpler standards directly governing institutional actional events of certain kinds, such as “No corporation is ever to pollute a waterway,” or, more simply, “Do not ever pollute a waterway.” The collectivist approach seems to allow for a simpler moral code, a code that would be less difficult to internalize and to have as the moral culture.

Consider, for example, the principles ofjust war theory. The familiar principles in the mor­ality of war speak both to the actions of individuals and to institutional actional events involving the state or its armed forces. For example, a state’s armed forces are obligated not to directly target noncombatants. On a collectivist approach, we can read such principles as directed both to the relevant institutions and to the relevant persons. On an individualist approach, we pre­sumably would need principles speaking to individuals when acting in their own right as well as when acting in ways that will contribute to an institutional actional event. So we would need a principle saying that no person is to directly target noncombatants as well as a principle saying that no person who is in an official position in a state or a state’s armed forces is ever to perform an action that would be partly constitutive of that entity’s directly targeting noncombatants.

The greater simplicity of the collectivist approach seems to argue in its favor. As I explained, the ideal code must be designed in a way that takes into account human psychology, including our cognitive limitations. It would seem less difficult and costly to internalize and teach a moral code designed in accord with the collectivist strategy than to internalize and teach one designed in accord with the individualist strategy. This suggests that the ideal code would be designed in accord with the collectivist strategy. If this is right, I take it, then the ideal code would include standards directly governing institutional actional events. Its rules would regulate institutional actional events in the way that they regulate actions of individual persons.

It is important, as I said before, that if the ideal code includes standards that directly regulate institutional actional events, it plausibly would also include the transitional rule. The transitional rule would say, roughly, “If a collective is required to do something, then its members, and any relevant officers in the collective, are to do their part in bringing it about that the collective acts as required.” The reason for including this rule is not hard to see. Collective entities are not independent agents. They do things only if relevant individual persons perform actions that constitute their doings. This means that standards that directly regulate institutional actional events will not motivate action unless they motivate the relevant individual persons, ideally by providing them with a reason to act.⁹ Inclusion of a transitional standard in the ideal code would address this worry. It would ensure that the relevant people have at least a pro tanto duty to do the things that would constitute the collective’s doing what the ideal code requires.

It is not clear, however, whether the fact that the ideal code would be designed in accord with the collectivist strategy means that, according to the society-centered theory, institutional entities have moral obligations. One might object that moral obligations are requirements governing actions. So even if some of the ideal rules directly concern institutional actional events, it does not follow that these events are actions, properly speaking, so it does not follow that any institutional entities have moral obligations.

We will return to this objection in a later section. As a first response, however, let me note that the objection is based on the twin assumptions that only actions in the proper or privileged sense can be governed by obligations, and that institutional actional events do not qualify as actions in that sense. These assumptions need defense.

7.4