Moral responsibility for actions and omissions: a new challenge to the asymmetry thesis

Abstract

This paper presents a new challenge to the thesis that moral responsibility for an omission requires the ability to do the omitted action, whereas moral responsibility for an action does not require the ability to do otherwise than that action. Call this the asymmetry thesis. The challenge arises from the possibility of cases in which an omission is identical to an action. In certain of such cases, the asymmetry thesis leads to a contradiction. The challenge is then extended to recent variations of the asymmetry thesis defended by John Martin Fischer and Carolina Sartorio. Finally, a possible objection to the challenge is addressed.

1 Introduction: the asymmetry thesis

Some philosophers have defended asymmetrical requirements for moral responsibility for actions, on the one hand, and moral responsibility for omissions, on the other hand. In particular, some have argued that moral responsibility for an omission requires the ability to do the omitted action, whereas moral responsibility for an action does not require the ability to do otherwise than that action. Consider the following principles:

Principle of Alternative Possibilities (PAP): A person is morally responsible for performing an action only if she could have done otherwise than perform that action.

Principle of Possible Action (PPA): A person is morally responsible for omitting to perform an action only if she could have performed that action.

The philosophers mentioned above reject PAP but accept PPA. In other words, they accept the following thesis:

Asymmetry Thesis (AT): PAP is false, but PPA is true.

In this paper, I present a new challenge to AT. I then argue that the challenge extends to the two recent alternatives to AT that are defended by John Martin Fischer (Forthcoming) and Carolina Sartorio (2005), respectively. Finally, I consider a possible objection to the challenge. Before I introduce the challenge, however, it will be helpful to see why some philosophers have found AT attractive.

The first principle, PAP, is called into question by so-called “Frankfurt-style cases,” named after Harry Frankfurt because of his very influential paper (Frankfurt 1969). In that paper, Frankfurt presented a case very similar to this one:

Frankfurt-Style Case (FSC): Black wishes Jones to cast his vote for presidential candidate A. In order to ensure that Jones does this, he implants a chip in Jones’s brain which allows him to control Jones’s behavior in the voting booth. (Jones has no idea about any of this.) Black prefers that Jones vote for candidate A on his own. But if Jones starts to become inclined to vote for anyone other than A, Black will immediately use his chip to cause Jones to vote for candidate A instead. As it turns out, though, Jones votes for candidate A on his own and Black never exerts any causal influence on Jones’s behavior. (Swenson 2015: 1279–1280)

In this case, Jones is apparently morally responsible for voting for candidate A. Yet Jones cannot do otherwise than vote for candidate A. Thus, PAP appears to be false.

But now consider a case in which an agent omits to perform an action that he is unable to perform:

Sharks: John is walking along the beach and sees a child drowning in the water. John believes that he could rescue the child without much effort. Due to his laziness, he decides not to attempt to rescue the child. The child drowns. Unbeknownst to John, there is a school of sharks hidden beneath the water. If John had attempted to rescue the child, the sharks would have eaten him and his rescue attempt would have been unsuccessful. (Swenson 2015: 1280; originally from Fischer and Ravizza 1998: 125)

In this case, John omits to rescue the child from drowning, yet John does not appear to be morally responsible for this omission because of the presence of the sharks.¹ This case appears to support PPA, since it seems to show that an agent cannot be morally responsible for omitting to perform an action that she is unable to perform.

So, while Sharks suggests that PPA is true, FSC suggests that PAP is false. Taken together, these cases seem to support AT. But, as I will argue in the next section, it turns out that AT is false.

2 The challenge: action-omission identity

Some (but not all) omissions are identical to actions. Consider the case of a child playing hide-and-seek.² Imagine that, once she has found the perfect place to hide, the child holds perfectly still. This action—holding perfectly still—is identical to omitting to move (or refraining from moving, which I take to be an omission). Now consider a similar case but in which it is clearer that the agent is morally responsible for the action/omission (assuming, of course, that anyone is ever morally responsible for anything):

Ben: Ben’s friend commits a murder, and Ben plans to cover up the murder (even though he knows that this would be wrong). Ben is taken in for questioning, and he is told to raise his hand if his friend has committed the murder. In accordance with his earlier plan, Ben holds perfectly still, omitting to raise his hand.

If anyone is ever morally responsible for anything, Ben is morally responsible for holding perfectly still (and thereby covering up the murder). In this case, Ben’s action of holding perfectly still is identical to an omission, namely his omitting to raise his hand. And, though not all omissions are identical to actions, there are countless cases of action-omission identity like the case of Ben.

Cases of action-omission identity present a challenge for AT, however, which can be seen by considering the addition of a counterfactual intervener to such a case:

Ben*: Black wishes Ben to cover up his friend’s murder. In order to ensure that Ben does this, Black implants a chip in Ben’s brain which allows him to control Ben’s behavior during questioning. (Ben has no idea about any of this.) Black prefers that Ben cover up the murder on his own. But if Ben starts to become inclined to raise his hand (to indicate that his friend committed the murder), Black will immediately use his chip to cause Ben to remain perfectly still instead. As it turns out, though, Ben holds perfectly still on his own, omitting to raise his hand, and Black never exerts any causal influence on Ben’s behavior.

To see why Ben* is problematic for AT, assume that AT is true. That is, assume that PAP is true and that PPA is false. According to PAP, Ben can be—and presumably is, in this case—morally responsible for holding perfectly still, but he cannot, according to PPA, be morally responsible for omitting to raise his hand. Ben can’t be morally responsible for omitting to raise his hand, according to PPA, because the presence of the counterfactual intervener, Black, precludes Ben’s ability to perform the action that he actually omits to perform (raising his hand). But Ben’s action (holding perfectly still) is identical to his omission (omitting to raise his hand). Thus, according to AT (which says that PAP is true and PPA false), Ben is both morally responsible for and not morally responsible for one and the same thing. But this is a contradiction, so AT must be false.

It is worth noting at this point that one way to avoid the challenge while maintaining an asymmetry view would be to restrict the scope of PPA to omissions that are not identical to actions. If PPA were so modified, my challenge would no longer apply to AT. The necessity of such a restriction of PPA would itself, I think, be a noteworthy result, but I doubt that many would be inclined to accept the new principle, if only for the reason that it would be problematically ad hoc. It is, then, worth considering alternative options. In the next two sections of this paper, I will consider two modified asymmetry theses that are not ad hoc, but I will argue that my challenge nevertheless undermines these asymmetry theses. In the last section of the paper, I will address a possible reply on behalf of asymmetry views.

3 Fischer’s new asymmetry thesis

Fischer used to accept AT, but he came to be convinced (for reasons independent of the argument of this paper) that AT is false.³ Nevertheless, Fischer maintains that there is some asymmetry in the requirements for moral responsibility for actions and omissions. In developing his more recent asymmetry thesis, Fischer appeals to the distinction between simple omissions and complex omissions. The former are “identical to, or fully constituted by, a bodily movement, where this can include the body’s keeping still (or a part of the body’s keeping still)” (Fischer Forthcoming: 12). Ben’s omitting to raise his hand in Ben* is an example of a simple omission. A complex omission, on the other hand, “is not identical to or fully constituted by (say) keeping my body still; it involves something more than the body—something about the relationship between the body and the external world” (Fischer Forthcoming: 12–13). Omitting to take a sip from my cup of coffee is an example of a complex omission, since it involves something more than my body, namely my cup of coffee. With the distinction between these two types of omissions on the table, we can consider Fischer’s new view:

Fischer’s New Asymmetry Thesis (FNAT): moral responsibility for actions (and, thus, simple omissions) does not require freedom to do otherwise, whereas moral responsibility for complex omissions does require such freedom. (Fischer Forthcoming: 18)

FNAT is better-equipped than AT to handle cases like Ben*. Since Ben’s omitting to raise his hand is a simple omission (since it is identical to Ben’s action of holding perfectly still), Ben can be morally responsible, according to FNAT, for both the action and the omission, despite lacking the freedom to do otherwise than what he actually does/omits to do.

But Fischer is wrong to think that only simple omissions can be identical to actions. Consider the following case (adapted from Clarke 2014: 27):

Brother: Joe is angry at his brother and decides not to call him on his birthday. Joe later changes his mind and plans to call his brother, thereby planning not to refrain from calling him (as Joe had previously planned). Joe calls his brother, omitting to refrain from calling him.

In this case, Joe’s omitting to refrain from calling his brother just is the action of calling his brother.⁴ But Joe’s omission is not a simple one, in Fischer’s terminology, since the omission involves something more than Joe’s body, namely the phone he uses to call his brother. So there can be complex omissions that are identical to actions.

The possibility of such cases raises a challenge for FNAT, however, which can be seen by considering the addition of a counterfactual intervener to such a case:

Brother*: Black wishes Joe to call his brother on his birthday. In order to ensure that Joe does this, Black implants a chip in Joe’s brain which allows him to control Joe’s behavior on his brother’s birthday. (Joe has no idea about any of this.) Black prefers that Joe call his brother on his own. But if Joe starts to become inclined to carry out his earlier plan (of refraining from calling his brother), Black will immediately use his chip to cause Joe to call his brother instead. As it turns out, though, Joe calls his brother on his own, and Black never exerts any causal influence on Joe’s behavior.

To see why Brother* is problematic for FNAT, assume that FNAT is true. That is, assume that moral responsibility for actions and simple omissions does not require freedom to do otherwise, whereas moral responsibility for complex omissions does require such freedom. According to FNAT, Joe can be (and presumably is, in this case) morally responsible for calling his brother, but he cannot, according to FNAT, be morally responsible for omitting to refrain from calling his brother. Ben can’t be morally responsible for omitting to refrain from calling his brother, according to FNAT, because the presence of the counterfactual intervener, Black, precludes Joe’s freedom to do otherwise than what is actually omitted (the complex omission of omitting to refrain from calling his brother). But Joe’s action (calling his brother) is identical to his complex omission (omitting to refrain from calling his brother). Thus, according to FNAT, Joe is both morally responsible for and not morally responsible for one and the same thing. But this is a contradiction, so FNAT must be false.

4 Sartorio’s new asymmetry thesis

Sartorio also rejects AT (for reasons independent of the argument of this paper), but has offered the following principle concerning moral responsibility for outcomes:

New Asymmetry (NA): An agent’s responsibility for an action can transmit to an outcome even if the outcome would have occurred anyway in the absence of the action. However, an agent’s responsibility for an omission cannot transmit to an outcome if the outcome would have occurred anyway in the absence of the omission. (Sartorio 2005: 470)

There is an asymmetry between AT and NA: the former is about the requirements for moral responsibility for actions/omissions, but the latter is about the conditions under which moral responsibility for actions/omissions transmit (or don’t) to outcomes. Despite this difference, cases like the ones I’ve discussed here reveal a problem for NA. Consider a modified case of Brother that makes explicit reference to a certain outcome:

Brother’s Phone: Joe is angry at his brother and decides not to call him on his birthday. Joe later changes his mind and plans to call his brother, thereby planning not to refrain from calling him (as Joe had previously planned). Joe calls his brother, omitting to refrain from calling him, with the result that Joe’s brother’s phone rings.

In this case, as before, Joe’s action (calling his brother) and omission (omitting to refrain from calling his brother) are identical. The outcome (Joe’s brother’s phone’s ringing) follows from both the action and omission (since they are identical). But this possibility raises a challenge for Sartorio’s NA, which can be seen by considering the following case:

Brother’s Phone*: Black wishes Joe to call his brother on his birthday. In order to ensure that Joe does this, Black implants a chip in Joe’s brain which allows him to control Joe’s behavior on his brother’s birthday. (Joe has no idea about any of this.) Black prefers that Joe call his brother on his own. But if Joe starts to become inclined to carry out his earlier plan (of refraining from calling his brother), Black will immediately use his chip to cause Joe to call his brother instead. As it turns out, though, Joe calls his brother on his own, with the result that Joe’s brother’s phone rings, and Black never exerts any causal influence on Joe’s behavior.

To see why Brother’s Phone* is problematic for NA, assume that NA is true. That is, assume both that an agent’s moral responsibility for an action can transmit to an outcome even if the outcome would have occurred anyway in the absence of the action and that an agent’s moral responsibility for an omission cannot transmit to an outcome if the outcome would have occurred anyway in the absence of the omission. In Brother’s Phone*, the outcome (Joe’s brother’s phone’s ringing) would have occurred anyway in the absence of Joe’s actual action, so his moral responsibility can (and presumably does) transmit from the action to the outcome. Yet because the outcome would have occurred anyway in the absence of Joe’s omission (the omitting to refrain from calling his brother), Joe’s moral responsibility for the omission cannot transmit to the outcome. But Joe’s action and omission are identical, so, according to NA, Joe’s moral responsibility both does and does not transmit to the outcome from one and the same thing. And this is a contradiction, so NA is false.

5 Objection: moral responsibility-attributions as intensional

One way to object to my argument against AT (and, mutatis mutandis, to object to my arguments against the other asymmetry theses) is to deny that it is a contradiction that the agents in the cases discussed in this paper both are and are not morally responsible for one and the same thing. Randolph Clarke defends a view of this type, first explaining how it applies in cases of action–action identity, then applying the view to a case of action-omission identity:

Can an agent both be and not be responsible for something that she does?

I believe so. Attributions of moral responsibility, it seems to me, are intensional. One can be responsible for that thing described one way but not responsible for that thing described another way.

When I raise my arm on some occasion, I raise it with some specific speed and trajectory. My action of raising my arm is, I think, (identical with) an action by me of raising it with that specific speed and trajectory. There aren’t two actions of arm raising that I perform on that occasion. And yet I might be responsible for raising my arm then but not responsible for raising it with that specific speed and trajectory. (Given my ignorance of and inability to control the exact speed and trajectory with which I raise my arm, there’s good reason to deny that I’m responsible for the latter.)

Now, consider the case of the child holding still for several minutes while playing hide-and-seek. Suppose that, had she not intended to do so, some external force would have caused her body to remain perfectly still then. She might be responsible for her action of keeping her body still. Her not moving on this occasion is simply her action of holding still, described in terms of something it isn’t—an instance of moving. And yet, it appears that she isn’t responsible for not moving. No matter how hard she had tried to move, she wouldn’t have been able to do so. (Clarke 2014: 158)

For the sake of argument, let us assume that Clarke is right and that there are cases (both action–action identical and action-omission identical) in which one is morally responsible for something under some description but not morally responsible for the same thing under another description. Now, when I consider why this might be so, I am inclined to think that Clarke’s parenthetical remark (about the case of action–action identity) provides the correct explanation: “Given my ignorance of and inability to control the exact speed and trajectory with which I raise my arm, there’s good reason to deny that I’m responsible for the latter” (Clarke 2014: 158). In other words, for actions/omissions that are described in a sufficiently fine-grained way, agents will inevitably fail to satisfy at least one of the two widely accepted conditions on moral responsibility, the epistemic condition and the control condition.⁵ If an agent is ignorant of the exact speed and trajectory with which she raises her arm, or if she lacks control over the exact speed and trajectory with which she raises her arm, then that fact explains the agent’s lack of moral responsibility for the action so described.

But there is, I think, a crucial difference between Clarke’s example of action–action identity and his case (as well as the ones discussed throughout this paper) of action-omission identity. In the former case, the agent’s level of knowledge/control varies from the action described in the fine-grained way (raising my arm at some exact speed and trajectory) to the action described in the course-grained way (raising my arm). In the case of child playing hide-and-seek, however, the child’s level of knowledge/control does not vary from action to omission.⁶ Given this, I find it implausible to maintain that, were she unable to refrain from holding still (and thus unable to move), the child would only be responsible for the action and not the identical omission.⁷ More generally, when an action and omission are identical and an agent meets the very same epistemic and control conditions for both the action and omission, it is implausible to grant responsibility for the action description but to deny it for the omission description. And if there is any case of action-omission identity in which the agent is morally responsible for both the action and the omission, this is problematic for the asymmetry theses considered here, for we can tinker with that case—adding a counterfactual intervener—and end up with a contradiction. It seems clear to me that there are such cases, so the asymmetry view faces a serious challenge.

6 Conclusion: Frankfurt’s observation

Given that cases of action-omission identity present a challenge to the various asymmetry theses considered here, I believe that we have reason to be skeptical of asymmetrical requirements for moral responsibility for omissions and actions. In his discussion of a certain asymmetry thesis, Harry Frankfurt made a related observation:

The idea that there is an important asymmetry of this sort between actions and omissions strikes me as rather implausible. There appears to be no fundamental reason why instances of performing actions should be, as such, morally different from instances of not performing them. After all, the distinction between actions and omissions is not a very deep one. Indeed, it is often a rather arbitrary matter whether what a person does is described as performing an action or as omitting to perform one. So it would be surprising if the moral evaluation of actions and of omissions differed in any particularly significant way. (Frankfurt 1994: 620)

Even if we grant that some omissions are not identical to actions (which seems plausible to me, though Frankfurt appears skeptical), clearly there are some cases in which omissions just are actions. In such cases, Frankfurt is right that the distinction (between action and omission) is not very deep. Moreover, in such cases, whether or not we evaluate the person as having acted or omitted is often arbitrary, as in the case of the child playing hide-and-seek. It should not be very surprising, then, that we can construct scenarios, like the ones developed in this paper, that put pressure on asymmetrical requirements for actions and omissions.