The Indispensability Argument for the Doing/Allowing Asymmetry

There is worked into our moral system a distinction between what we owe people in the form of aid and what we owe them in the way of non-interference.

– Philippa Foot (1967), 4].

1 Thomson’s Challenge

In 1967, Philippa Foot published an influential article in which she brought up the question of “whether there is any difference, from the moral point of view, between what one does or causes and what one merely allows” (1967), 3. She then suggested that there is, and that this difference permeates our moral thinking. We may call it ‘Asymmetry’.

Asymmetry Other things equal, doing harm is harder to justify than allowing this harm to occur.¹

Appealing to Asymmetry, Foot maintained that our “negative” moral duties, our “duties of non-interference”, are weightier than our “positive duties”, our duties of aid.

Nine years later, and provoked by Foot’s article, Judith Jarvis Thomson kicked off the philosophical sub-discipline of trolleyology with an article on what came to be known as the “Trolley Problem” (Thomson 1976). Thomson confronted her readers with two moral scenarios that, despite sharing a very similar structure, aroused opposing moral verdicts. In the first scenario, a run-away tram is hurtling down the track and threatens to overrun and kill five people. You are a bystander who could throw a switch to turn the tram on a side-track, where it would overrun and kill only one person. So, you have two options:

Bystander’s Two Options You can

• (i) do nothing, letting five die, or

• (ii) throw the switch to the right, killing one.

Most people, Thomson noted, judge that you may choose option (ii).² But then she confronted her readers with a second scenario that seems structurally similar. We are, again, facing a run-away tram that is about to overrun and kill five people. This time, however, you are standing on a footbridge over the track, anticipating the disaster. Fortunately (well...), a fat man is on the footbridge with you and he invitingly leans over the railing. Again, there are two options.

Fat Man You can

• (i) do nothing, letting five die, or

• (ii) shove the fat man off the bridge down onto the track, thereby killing him, but also, since he’s very big, stopping the tram and saving the five.

Most people, Thomson remarked, judge that you must not choose option (ii).

This is startling. For are we not, in both scenarios, sacrificing one to save five? What explains our different moral verdicts about these scenarios? In honor of Foot—who lived in England, the country where trams are trolleys—Thomson called this problem the

Trolley Problem Why is it impermissible for the agent in Fat Man to choose option (ii), but permissible for the agent in Bystander’s Two Options to choose option (ii)?³

The Trolley Problem poses a threat to Foot’s original claim that Asymmetry permeates our moral Thinking. If her claim is true, why do we judge that the agent in Bystander’s Two Options may throw the switch? After all, doing so is a killing, whereas not throwing the switch would ‘only’ mean to allow the five to die. If Asymmetry really permeated our thinking, we would make a different judgment in this case.⁴

Several decades later, in a surprising turn of events, Thomson published another article in which she declared the Trolley Problem a “non-problem” (Thomson 2008, 367). She now argued that, contrary to first appearances, it is not permissible to throw the switch in Bystander’s Two Options. Suppose, Thomson now argued, you could only save the five by steering the tram onto a side-track on which you are currently trapped, thereby killing yourself. Would morality demand of you to pay this cost, to sacrifice yourself? No, said Thomson. Though sacrificing yourself would undoubtedly be a good deed, it surely is not morally required. If the only alternative is to sacrifice yourself, then it is morally permissible to let the five die (ibid., 365). We can all ask ourselves: Would we consent to self-sacrifice? Importantly, whatever our answer is, it will not conflict with the demands of morality. Either we act very altruistically and sacrifice ourselves, which is, of course, permissible, or we do not and let the five die, which is permissible as well. But there is one thing that would clearly be impermissible: somebody forcing us to sacrifice ourselves, forcing us to choose the ‘altruistic’ option. And this is the core insight of Thomson’s 2008 article: Nobody can legitimately force you, in the name of morality, to act in a way that goes beyond your moral obligations.

And, therefore, it would be wrong to throw the switch in Bystander’s Two Options. By steering the tram towards the one, you would be making him pay a cost morality would not require him to pay if the decision was his. Morality requires that it is up to him to make that decision. And thus the supposed conflict between our verdicts in Bystander’s Two Options and Fat Man is resolved. We were wrong all along: Choosing option (ii) is impermissible in both scenarios.

Thomson’s turn allowed her to hold on to Foot’s original idea that Asymmetry is a deep feature of our moral thinking. In Fat Man, for example, we judge that we may not shove the fat man off the bridge because doing so would violate a negative duty, whereas letting the five die ‘only’ violates a positive duty. So, our judgment is explained by the different weights we assign to negative and positive duties. The same explanations can be given for our moral verdicts in various other much-discussed cases: The judge—we think—may not put an innocent man to death in order to prevent some violent rioters from killing many; and the surgeon—we think—may not kill a patient in order to donate her organs to five other patients to save their lives.⁵ Because it explains our verdicts in cases of this kind so well, it really does seem, Thomson suggests, that Asymmetry is a deep feature of our moral thinking.

There is much to like about Thomson’s argument, but I will not discuss it any further. I presented the dialectic up to this point to introduce an important challenge that comes up at the end of her 2008 article:

It is one thing to say that there is a difference in weight between positive and negative duties, and quite another to say what this difference is. I know of no thoroughly convincing account of its source, and regard the need for one as among the most pressing in all moral theory (Thomson 2008, 372).

I will call this “Thomson’s Challenge”:

Thomson’s challenge We have to explain why Asymmetry is a deep feature of our moral thinking.

In this paper, I intend to provide an account of the source of Asymmetry, and I will explain why it permeates our moral thinking in the ways Foot, Thomson, and many others have described. I am going to argue, in short, that Asymmetry is a deep feature of our moral thinking because accepting it is indispensable for building and maintaining moral communities.

Here is my plan: In the upcoming section, I introduce my main argument, the indispensability argument. Sections three and four are then devoted to an elaboration and defense of its premises. In section five, I present two noteworthy implications of the argument. In section six, I turn to an important objection that asks why we should think that the indispensability of Asymmetry has anything to do with its truth. After discussing and rebutting the objection, I quickly sum up my results in section seven.

Before we start, a final bit of terminology: What does it mean to allow something? When I use the expression “allowing an event E” in the following, I mean “intentionally forbearing to prevent E where one is capable of preventing it” (Foot 1967), 3]. The idea is that some causal sequence is already in progress independent of me, that I am aware of this sequence, that I could intervene if I wanted, but that I do not.

2 The Indispensability Argument

My strategy for meeting Thomson’s challenge is fairly straightforward. In a nutshell, I argue that Asymmetry is a deep feature of our moral thinking because it could not be otherwise. Accepting Asymmetry, I claim, is indispensable for creatures like us—creatures with our physiological and psychological makeup in a world like ours—at least in so far these creatures endeavor to build and maintain stable moral communities.

The indispensability I aim to carve out is psychological. It depends on a few general empirical claims about human nature, which renders my argument, in a way, contingent. However, as I hope to convince you in what follows, the claims I rely on are not only hard-to-deny but also general enough to address Thomson’s challenge with respect to all human beings—which is why I will ultimately conclude that Asymmetry is a kind of ‘moral axiom’ or ‘moral pillar claim’, something human beings need to accept in order to build everything else. Details will follow. For now, here is the argument.

The indispensability argument⁶

P1:

Caring about morality is indispensable for creatures like us to build and maintain stable moral communities.

P2:

Creatures like us can only care about morality if morality conforms to Asymmetry.

$\therefore$:

Asymmetry is indispensable for creatures like us to build and maintain stable moral communities.

The argument is valid, and so I will spend the greater part of the paper defending its premises. Before I do so, however, a few more preliminary remarks. Noteworthily, my argument deviates from the ones put forth in normative ethics. (I will not construct new trolley scenarios, and my argument will not rely on any appeals to your moral intuitions about throwing switches or shoving the obese off bridges.) Instead, my argument is an argument in moral anthropology or moral psychology. My goal is to explain why Asymmetry is a deep feature of our moral thinking. So, my interest is not (primarily) to advance the discussion between consequentialism and nonconsequentialism. However, my result will have some first-order normative implications. Since, as I argue, Asymmetry is indispensable, all those (for the most part: consequentialists) who reject it on the grounds of first-order moral considerations make a mistake. I will come back to what this means for consequentialism further below.

Moreover, my indispensability argument bears some similarities with an argument by Samuel Scheffler (2004), who also thinks that Asymmetry is an indispensable element of human life.⁷ It will be helpful to explain how our considerations differ.

In order to build moral communities, Scheffler says, we have to view ourselves as subject to norms of normative responsibility (2004, 220). He then claims that this presupposes (i) a distinction between doing and allowing and (ii) the normative relevance of the distinction from a moral point of view, that is, it presupposes Asymmetry.

The first presupposition—that we are committed to a doing/allowing distinction—results from the very idea of a moral norm: A moral norm tells us to regulate our behavior in specific ways. And bringing about this moral compliance is, necessarily, something we do. We cannot bring about our moral compliance by allowing, that is, by forbearing to prevent causal sequences that are already under way. I find this first part of Scheffler’s argument quite convincing.

But I have doubts about the second presupposition. Does taking yourself to be a subject to moral norms really presuppose that there is a moral asymmetry between doing and allowing harm? Scheffler thinks it does. He believes that rejecting Asymmetry would imply a purely instrumental view of one’s agency. A view which, he continues, is inconsistent with the idea of being an agent who is subject to moral norms (227–36).

I am not fully convinced by this second part of Scheffler’s argument. At least at a first glance, it seems unclear to me why it should be incoherent to internalize a moral norm while rejecting Asymmetry; even if moral norms address my agency, why couldn’t they coherently state that my doings and my allowings are equally normatively significant?⁸ But I do not need to discuss this any further for our purposes. It does seem clear that Scheffler aims for a conceptual indispensability; the idea is that, rejecting Asymmetry, we cannot coherently consider ourselves as agents that are subject to moral norms.

In contrast, I argue that Asymmetry is psychologically indispensable. This is a somewhat weaker claim but it is strong enough to reach a significant conclusion (I think). Also, one might suspect that Scheffler’s argument is too strong. It implies that all opponents of Asymmetry (mostly: consequentialists) make incoherent claims, are conceptually confused, or, at least, do not have an adequate grasp of the concepts they use. This is a drastic implication. Since my considerations are not conceptual but psychological, I do not have to claim that my opponents fail to understand what they are talking about. This, I take it, is a dialectical advantage.⁹

3 Defending the First Premise

In this section and in the upcoming one, I defend both premises of the indispensability argument. Here, again, is the first premise:

• P1 Caring about morality is indispensable for creatures like us to build and maintain stable moral communities.

Why ‘creatures like us’? I use this formulation to stress that my considerations apply to the real world we live in, and not to some merely imaginary, logically possible, utterly different world whose inhabitants might not even be remotely human.¹⁰ These creatures are capable of practical and theoretical deliberation. They have our present human constitution, that is, our physique and our psychological makeup, including beliefs, desires, hopes, fears, etc. Additionally, these creatures are inherently social; they cannot but live in communities.¹¹ This is partly due to the environment they live in; an environment where resources are limited and often hard to get by, where nature is beautiful but deadly, and where a significant amount of teamwork is required for staying warm during a cold winter night with enough food in one’s belly. In short, I am interested in a morality for creatures like us in a world like ours.

Next, let me turn to what I mean by ‘morality’ or a ‘morality system’. Any moral community needs ways to coordinate the behavior of its community members in some sort of social practice. This social practice needs at least two components. It needs, firstly, a set of widely accepted and internalized moral rules for cooperation and conflict management. Examples for such rules are “Do not steal!”, “Keep your promises!”, and “If others are in dire need of help and you could help them without taking significant costs upon you, help them!” I regard it as obvious that no group of humans could build a moral community without a set of widely accepted and internalized moral rules.¹²

Secondly, in order for these moral rules to be internalized and complied with, moral communities need to establish accountability practices. Rules erode if nothing happens in cases of non-compliance. To prevent that, community members have to demand compliance of each other. They have to have the (normative) expectation that everyone should comply with the rules (Darwall 2006; Wallace 2019). The flip-side of these demands or expectations is the social practice of holding accountable. That is, if someone (intentionally) violates a normative expectation, she is held accountable by the members of the community—typically in the form of reactive attitudes like resentment, indignation, and guilt (Strawson 2008). I regard it as obvious that human beings could not build moral communities if they did not direct expectations of this kind at each other. If there were no practice of holding each other accountable with respect to a set of moral rules, the rules would quickly erode and the moral community would falter.¹³

Let us call a system that comprises both of these elements a morality or a morality system.¹⁴ We can then say the following: In order for any group of people to count as a moral community, they need a morality system.

Next, what does it mean to claim that community members must ‘care’ about their morality system? On a first approach, if one cares about a morality system, the system is motivationally, affectively, and cognitively entrenched in one’s psychology. It mashes with what one wants, feels, and deliberates on. Two aspects of this psychological entrenchment are crucial for my argument.

First, most members of a moral community must conform to the moral rules most of the time. Again, I take this to be quite obvious. If they did not, they could not build a moral community. Their lives would, then, be shaped by a fundamental distrust towards others. A state of nature would always be within a stone’s throw; life would be nasty, brutish, and probably quite short. So, a basic trust in others’ compliance with a set of rules for regulating behavior is a sine qua non for a stable common life. And in order for this trust to exist, most members of the moral community must adhere to the rules most of the time.

Note the implications. If most members of the community adhere to the morality system most of the time, they have to commit a large part of their mental resources to morality. Each and every day of their lives, they have to constrain their own and others’ behavior in accordance with the moral rules. For example, they have to muster motivational resources to exert moral self-control. They have to emotionally respond to both their own and others’ behavior. And they have to dedicate cognitive resources to deliberations on what morality requires them and others (not) to do. Participating in the morality system consumes mental resources. And, again, if it were not the case that most members of any given group participate in the outlined ways, they could not build a moral community.

The second crucial aspect of the psychological entrenchment of morality is this: Community members must respond differently to the different severities of moral wrongs. Trivially, any morality system has to comprise more and less serious moral wrongs. For example, killing Chris is harder to justify than slapping him. Both action types are wrong but the former is a greater, more severe moral violation. These different severities find their expression in the associated accountability practices. Community members will resent killers more than face slappers, and (if almost all goes well) murderers will experience more guilt than face slappers. The rule against killing is more important, morally speaking, than the rule against face slapping.

This second aspect concerns, again, mental resource management. On the one hand, it is clear that moral community members must commit some mental resources to morality. On the other hand, they must also commit a significant chunk of these resources to the non-moral parts of their lives: sustaining themselves, working, making plans for the future, etc. Since mental resources are limited, the task is to strike the right balance. This is no easy task because, in certain scenarios, morality must draw on almost all of our mental resources, or else the moral community becomes unstable.

Consider an example. Martha the medical researcher has developed a new medication. A day before the start of its factory production, she discovers that the medication causes severe harm over the long term. In such a scenario, the moral community would expect Martha to muster a significant amount of mental resources to avoid this harm. Among other things, community members would expect her to empathize with her possible victims, to be compassionate, to deliberate her practical options, and to anticipate the guilt she would experience if she swept her discovery under the carpet. Martha’s fellow citizens would expect her not to be able to sleep particularly well that night (and at least a few nights to come). In other words, they would expect her to enter a kind of psychological moral crisis, a cognitively and emotionally straining exceptional condition, albeit one that can be overcome in an obvious way, namely, by doing what morality requires: Publish the discovery and stop the factory production.

The example illustrates a general point: Since it is within most community members’ power to cause significant harm to others—just think about the danger of having millions of people driving around in cars—they need to be ready to muster a significant amount of mental resources to avoid it. Importantly, this mental readiness requires the moral rules to be deeply internalized in (most) community members’ minds.

But not all moral rules are equally important. We should avoid throwing our waste in our neighbors’ trash bins, to be sure, but flouting this rule hardly warrants a psychological moral crisis. And so it is clear, I think, that the members of a moral community should not muster the same amount of resources to each and every one of their own or others’ moral transgressions, howsoever small. There has to be the right psychological balance. The morality system should occupy a central but no all-commanding place. There has to be, as I am going to call it, a balanced care about morality. And achieving this balance is precisely the point of the second aspect of ‘care’ mentioned above: Community members should respond differently to the different severities of moral wrongs. If they did not, the morality system would lead to a kind of mental overload; it could not play the role it needs to play in order for us to live together in stable moral communities.

Summing up, a balanced care about morality entails that people internalize the moral rules in such a way that they are ready to devote a possibly significant amount of mental (motivational, emotional, cognitive) resources to the moral cause. Without this readiness, they could not build and maintain stable moral communities. However, since mental resources are limited, the central challenge is to strike the right balance between spending these resources for morality and spending them for the non-moral parts our lives. There has to be a ‘balanced care’.

Before we move on to the second premise in the next section, let us quickly consider an objection. It could be claimed that there is another way for morality systems to avoid mental overload. Instead of the psychological entrenchment just described, we could minimize the amount of moral rules. If there were only very few rules to comply with, morality would not become all-commanding.¹⁵

This, however, is an implausible proposal. Creatures like us cannot build moral communities on the basis of a few simple moral rules that are easily applicable and whose compliance consumes only few mental resources. Take a seemingly straightforward moral rule like “Don’t kill others!” Even though it is part of any plausible morality system, there are exceptions to it. For example, it is not wrong to kill others in self-defense. And euthanasia is, arguably, morally permitted in some cases. The same holds for other moral rules. Lying is wrong, yes, but there are morally permitted white lies. And when a job interviewer asks you if you are planning to become pregnant in the near future, lying might be morally permissible as well. In short, even the application of seemingly straightforward moral rules requires a complex moral understanding (Scanlon 1998, 199–202).

Moreover, human interactions are just too complex to be governed by only a few moral rules. We need rules against killing, stealing, and mistreating others. And we need rules for helping others in need. All these rules come with (at least partially) oblique application conditions that require a complex moral sensitivity to grasp. Thus, a few moral rules would not get the job done. Life is too complex. And even if we somehow managed to reduce all the required moral rules to a single master rule—e.g., “Respect everyone’s autonomy!”—this rule would still require a significant amount of mental resources to be practically applied in all the various contexts of human communal life. So, reducing the amount of moral rules in order to avoid mental overload is no option for creatures like us in a world like ours.

4 Defending the Second Premise

Here, again, is the second premise of my argument:

• P2 Creatures like us can only care about morality if morality conforms to Asymmetry.

My strategy for defending (P2) takes the form of a reductio: If a morality system conformed to Asymmetry’s only plausible contender—Symmetry—morality would inadvertently generate mental overload.

Symmetry Other things equal, doing harm is as hard to justify as allowing this harm to occur.¹⁶

I am going to argue that creatures like us could not care about morality in a balanced way if morality conformed to this principle. Asymmetry will then remain the only relevant alternative. My argument has five steps.

First, proponents of Symmetry are committed to the following principle.

Rule-Symmetry For any moral rule R with the content “Don’t $\upphi$!”, there is an equally important moral rule R* with the content “Do not allow $\upphi$-ings!”

Proponents of Symmetry are committed to Rule-Symmetry because they deny that there is a moral difference between stealing something and allowing someone else to steal it. If they are right, the rule against stealing is, from a moral point of view, as important as the rule against allowing others to steal. Thus, Symmetry implies Rule-Symmetry.¹⁷

The second step of my argument takes the form of a sad but hard-to-deny claim:

Bad World Wrongful harms are ubiquitous.

Unfortunately, our world is full of (wrongful) harm. People frequently slap each others’ faces, put their waste in their neighbors’ trash bins, insult each other filthily, disregard others’ legitimate expectations, and snatch bubble gum at the corner kiosk.

Moreover, we know about the existence of much of this wrongful harm. Not all of it, to be sure. But quite a lot of it, at least in our immediate personal environments. For example, I know that a friend of mine cheats on his partner. I suspect that a neighbor down the streets sells Marihuana to twelve year olds. I know that a resident of my apartment complex, motivated by sheer antipathy, stealthily cuts off the blossoms of a plant that another resident decorates our entrance area with. I know that a former friend of mine shape-shifts into a despicable online personality and threatens proponents of the “radically left, woke PC-culture” on social media. And so on. All of these examples concern wrongful harms in my immediate environment. More or less depressingly, you will not have difficulties coming up with your own examples.

The third step of my argument makes use of an earlier point: Most of us care a lot about not causing wrongful harms. We are ready to commit a significant amount of emotional, motivational, and cognitive resources to morality. Consider how much you care about not causing harms of the kind mentioned in the previous paragraph. You would feel guilty about intentionally destroying your neighbor’s plants out of sheer antipathy. (And if you did not feel guilty, you would try to appease your conscience by justifying your clandestine shearing; you would try to convince yourself that your neighbor deserves this treatment.) You would also feel terrible about selling drugs to kids; you imagine and anticipate the suffering you might cause, you empathize with your (imagined) customers or victims. You are similarly put off by the idea of cheating on your partner and then having to meet their gaze without shrinking. You anticipate your partner’s resentment and you know that it would be fully justified given your breach of trust; which is part of your motivation not to breach it. Likewise, the idea of becoming a part-time internet troll whose sole purpose is to insult and threaten people with a different political orientation strikes you as utterly appaling. You know that if you did something like this yourself, the indignation you feel towards internet trolls would transform into guilt—the guilt of someone who is aware that the main purpose of their nightly hobby is to spread fear and hate. All of this—all these thoughts, these actual and anticipated emotions—contribute to your motivation to conform to (what you take to be) the moral rules.

So, we do care a lot about not committing wrongful harms. Our readiness to muster mental resources for the moral cause is quite remarkable; which is precisely as it has to be if we want to build and maintain a moral community characterized by mutual trust.¹⁸

Coming to step four: Further above, we said that Rule-Symmetry implies that these two rules are equally morally important:

${R_{M}}$ Do not mistreat others!

${R^*_{M}}$ Do not allow the mistreatment of others!

Now, consider what happens when we add Bad World to the picture. Then, the moral symmetry between doing and allowing generates a problem. ${R_{M}}$ is deeply entrenched in our mental lives; we care a lot about it. But, crucially, Symmetry implies that the ‘allowing rule’ against mistreatments, ${R^*_{M}}$, is equally morally important. It follows that both rules should play the same role in our mental lives. Why? Because, according to our understanding of ‘importance’, the importance of a rule is measured by the intensity of the responses with which its violation is met by the accountability practices of a given morality system.¹⁹ And since we need the members of a moral community to strongly respond to the violation of a set of basic moral rules, Symmetry forces us to respond with the same intensity to violations of the corresponding prevention rules. It follows that, from the moral point of view, we should feel as guilty about mistreating others as we should feel about allowing the mistreatment of second parties by third parties. It follows that we should blame the line-cutter as much as we should blame those who allow her to proceed. In general, it follows that from the moral point of view, committing any harmful wrong deserves as much blame as allowing it.

This implication is implausible. It is implausible because it is psychologically impossible. We could not muster the same amount of mental resources to avoid wrong $\upphi$-ings as we muster to prevent others from wrong $\upphi$-ings we are aware of. We could not do so because mental resources are limited and because we live in a bad world where harmful wrongs are ubiquitous. It is crucial to understand how breathtakingly far-reaching the implications of Rule-Symmetry are: It implies that I must care as much about preventing any mistreatment (I know about) as I care about not mistreating others myself. But I simply could not do that. If I attempted to meet this standard, I could not have a life. The morality system would become all-commanding; and creatures like us cannot live with an all-commanding morality system. So, it seems, a morality system that incorporates Symmetry renders a balanced care about morality impossible.²⁰

Which brings me to the fifth and final step of my argument. Proponents of Symmetry, I now want to suggest, entangle themselves in a practical self-contradiction. The goal that presumably motivates them is to develop the most plausible and internally coherent morality system. However, defending Symmetry, they lose sight of the psychological realities and practical necessities of human life. They fail to see that, for creatures like us, ‘Symmetry systems’ are incompatible with balanced care. Any such system vastly overextends our mental resources. Defending Symmetry in the name of a more plausible morality system is, thus, a practical self-contradiction.

Summing up, if a morality system conforms to Symmetry, it cannot function as the basis of a stable moral community. As long as we want to live in stable moral communities, we cannot but accept Asymmetry. And thus we have reached the conclusion of the indispensability argument.

We can now turn back to Thomson’s challenge.

Thomson’s challenge We have to explain why Asymmetry is a deep feature of our moral thinking.

The indispensability argument provides the explanation we are looking for. Asymmetry is a deep feature of our moral thinking because it has to be in order for creatures like us to live in moral communities. And since we do live in moral communities, it is no wonder that we accept Asymmetry.

Before I turn to some of the implications of the indispensability argument, let me address another objection.

4.1 Objection: But Why the Focus on Doing/Allowing?

Suppose one agrees that, to avoid mental overload, there has to be something that stops morality from overstraining our mental resources. Why does it have to be a distinction that appeals to doing and allowing, as Asymmetry has it? What justifies my focus on doing/allowing in the first place?²¹

Here is my reply. Moral norms address our agency, our becoming active. They aim for regulating our behavior such that we comply with morality. This is something we do. So, we could say, moral norms aim to regulate our doings.²²

Next, note that our entire moral practice, our practice of holding each other accountable, depends on a distinction: Some of our behaviors are genuine doings, others are more adequately described as happenings (e.g., sneezes). Now, clearly, we want to hold people accountable for what they did, and not for what merely happened to them. If I barge into you because I envy you for publishing a paper, I deserve blame. But if I barge into you because I faint due to a medical condition, I do not deserve blame. So, morality necessarily has a built-in focus on our doings and, thus, presupposes a distinction between doings and mere happenings.

But how does the notion of allowing enter the picture? Start by noting that the variety of human behavior is not exhausted by doings and happenings. There is another way in which we can become active. Suppose I steal a pack of bubble gum at the corner kiosk. This would be a morally wrong doing, theft. Next, suppose I accidentally bump into a shelf, whereupon, unbeknownst to me, a pack of bubble gum falls into my jacket pocket. This is no morally wrong doing, it is just an unfortunate happening. But now suppose I witness a small girl snatching a pack of bubble gum at the corner kiosk. Suppose further that only I saw it and that I decide to do nothing. Clearly, deciding to do nothing is a way of “putting my agency to use”. Also, it would be morally wrong. But it certainly wouldn’t be theft. The girl stole the bubble gum, not me. The theft was her doing. So, my saying nothing cannot be a morally wrong doing in the same sense in which my stealing the gum would be. Instead, deciding to say nothing is an allowing. It would be a way of being active, but, as we said above, in the sense of intentionally forbearing to prevent a causal sequence that is already under way. So, there are two ways in which we can put our agency to use, two ways in which we can become active. We could call these senses “$\hbox {doing}_{1}$” and “$\hbox {doing}_{2}$”, “strict doing” and “not-so-strict doing”, “primary” and “secondary manifestations of our agency” (Scheffler 2004), or, well, “doing” and “allowing”. Whatever we call them, the distinction between these two ways of putting our agency to use matters greatly to us.²³

So, why focus on doing and allowing to avoid mental overload? Because (1) a focus on agency is built into the very idea of a morality system, and because (2) once we realize that there are two ways of putting our agency to use (doing and allowing), we have to position ourselves with respect to their relative moral significance since this relative moral significance has massive implications for our moral practice—implications for the first-order level, as well as for the ‘feasibility’ of the whole moral enterprise.²⁴

But, an anonymous referee maintains, could we not avoid mental overload by drawing a different distinction such as, for instance, a distinction between strictly intended consequences and merely foreseen ones? This question appeals to the following principle that has been discussed as an alternative to Asymmetry in the literature:

Doctrine of Double Effect (DDE) Other things equal, harm that is strictly intended is harder to justify than harm that is merely foreseen.²⁵

I agree that DDE, by permitting us to care less about harm we merely foresee, could also get the job done and avoid mental overload. However, it seems to me that it does so in virtue of appealing to the same distinction, namely, the distinction between the two ways of putting our agency to use. After all, ‘strictly intending’ harm comes pretty close to ‘doing’ harm; and ‘merely foreseeing’ harm (without strictly intending and without preventing it) comes pretty close to ‘allowing’ said harm. So, I think, even though Asymmetry and DDE are not identical, they manage to avoid mental overload for the same reason. Moreover, and more importantly, it has convincingly been argued that DDE should be regarded as a “sister principle” of Asymmetry, and not an alternative to it (Woollard 2017).

I hope these comments suffice to show that my focus on the moral asymmetry between doing and allowing is no arbitrary way of avoiding mental overload. Since morality essentially concerns our agency, and since there are two ways for putting our agency to use, the question of the relative moral significance of these two ways naturally raises itself. And so we have to address this question. Due to the nature of human agency, and due to the nature of morality (its ‘point’ is to regulate agency), we cannot avoid addressing the question of whether there is a moral distinction to be drawn here.

5 Implications

The indispensability argument has two interesting implications. The first one concerns the nature of Asymmetry’s justification. The second one concerns the plausibility of consequentialism, a family of views that typically reject Asymmetry.

5.1 Asymmetry Does Not Need Moral Justification

As I said before, the indispensability argument is no argument in normative ethics. While it justifies Asymmetry, it does not deliver a moral justification. Also, it does not appeal to our moral intuitions about individual cases, as many of the arguments in the literature do (Kamm 2006; Woollard and Howard-Snyder 2021). Instead, it establishes that creatures like us cannot but accept Asymmetry in so far they want to build and maintain stable moral communities.

This has a noteworthy implication: There is simply no need for a moral justification of Asymmetry. This is a desirable result because it saves us a significant amount of work in first-order, normative ethics. Since creatures like us cannot but live in moral communities, this is all the justification we need for designing morality systems that conform to Asymmetry. So, we can stop thinking of Asymmetry as a claim that can be morally objected to. We can stop construing exemplary scenarios that are supposed to awaken intuitions in one direction or another. Instead, we should think of Asymmetry as a ‘pillar claim’ or ‘axiom’ of morality. We need it to build everything else.

Also, note that accepting Asymmetry does not have far-reaching first-order implications—which I take to be desirable as well. For instance, it does not imply that it is morally alright to allow the starvation of a far-away stranger. Asymmetry only implies that it would be harder to justify to kill the stranger than to allow her to die (by, say, not sending her food intentionally). This is, of course, consistent with the claim that letting the stranger die is morally wrong. So, for all that Asymmetry implies, it might still be true that there are far stronger moral obligations to help those in need than common sense currently has it. How strong exactly? This will remain an important issue in normative ethics. So, even though the indispensability argument removes the justification of Asymmetry from normative ethics, there remain important first-order issues surrounding it—which is, I think, precisely as it should be.²⁶

5.2 Do We Have to Reject Consequentialism?

Consequentialists typically reject Asymmetry on the grounds that only consequences matter morally. From this perspective, it does not make any moral difference whether an action is a doing or an allowing; Asymmetry is false. But now I have argued that we cannot but accept Asymmetry. Does this mean that we cannot but reject consequentialism?

No, it does not. Consequentialists are cunny vixens. They tend to find a way out. They can simply embrace the indispensability argument and claim that, well, accepting Asymmetry has the best consequences. (Because rejecting it renders moral communities impossible.) Thus, the indispensability argument is no knock-down argument against consequentialism. It is no knock-down argument until somebody shows that the basic commitments of consequentialism necessarily commit its proponents to a rejection of Asymmetry. However, I do not think that such an argument can be made. Consequentialists care about the best overall consequences. If somebody convinces them that, for broadly empirical reasons, something they previously rejected leads to the best consequences, they will always be able to jump ship while remaining consequentialists by simply accepting the previously rejected thing for these broadly empirical reasons. They will go with whatever generates the best consequences; and if Asymmetry does, then they will happily accept it.

However, I think the indispensability argument reduces consequentialism’s intuitive appeal. Part of this intuitive appeal stems from cases in which it strikes (some of) us that it does not matter at all whether an agent does harm or merely allows it. Some consequentialists take the fact that we (or they) have this intuition as evidence for the truth of consequentialism. They claim that, intuitively, there is no moral difference between doing and allowing in certain cases, which is then taken as evidence that only consequences matter morally. However, the indispensability argument illustrates why we should not take the intuition that there is no moral difference between doing and allowing in these cases too seriously. Since we want to—or have to—live in moral communities, we cannot but accept Asymmetry. Thus, the existence of (contested) cases in which people’s intuitions point in a different direction bears little argumentative weight. If intuitions point towards the impossible, we may put them aside. So, while consequentialism is not refuted by the indispensability argument, it loses some of its intuitive appeal by taking Asymmetry on board.

6 An Objection: You Are Conflating Moral Psychology with Normative Ethics

Finally, let us turn to an important objection. It might be claimed that the indispensability argument rests on a fundamental confusion between moral psychology on the one hand, and normative ethics on the other.²⁷ Here is one way to reveal the (supposed) confusion: If sound, the indispensability argument only establishes that creatures like us are psychologically committed to Asymmetry. We cannot but live as if it were true. This is a psychological claim. But, crucially, the argument does not tell us that Asymmetry is, in fact, true. But this is, of course, what the whole debate is about. Thus, the indispensability argument misses its target; it does not even manage to address the relevant issue. In this spirit, a proponent of Symmetry could go along with my argument and say: “Since you are only talking about psychology, I do not have to budge on the ethical question of whether or not Symmetry is true. If your argument is sound, I only have to acknowledge that human psychology distributes our attitudes and attention in a way that does not always perfectly track the moral truths. But, since we have reason to believe that Symmetry is true, we should change our morality system accordingly—and then start worrying about our mental resources.”

Let us pursue this idea a little further. If we incorporate Symmetry into our morality system, we get an increase in significant moral failures (because people are, now, as morally responsible for their allowings as they are for their doings). Call these additional significant moral failures the ‘added failures’. We now face the challenge of explaining how, despite the added failures, mental overload can be avoided and balanced care remains achievable. We somehow have to cut the ties between, on the one hand, the added failures and, on the other hand, the moral community’s legitimate responses to these failures. How could this be done?

I see one way to do so. We could introduce a distinction between, on the one hand, a criterion of wrongness and, on the one hand, a moral decision procedure.²⁸ We could then claim that an agent’s blameworthiness—that is, the currency determining the legitimate responses the agent may be met with—depends solely on her compliance with the moral decision procedure; and not on the moral status the criterion of wrongness assigns to her doings and allowings. Blameworthiness would, then, only be a matter of complying with the decision procedure; and this procedure could, in principle, allow agents to ignore the added failures in their moral decision making.

New possibilities arise. We now get cases in which, even though an agent’s action is wrong, blaming her would be illegitimate (because she did follow the moral decision procedure). In short, we get “blameless wrongdoing”.²⁹ Supposedly, this new category allows us to avoid an overload of mental resources while holding on to Symmetry. Because now we can coherently claim that, even though an agent ‘committed’ an added failure—which was wrong—she is not blameworthy for it. Mental overload is avoided because we have loosened the formerly closer tie between, on the one hand, the moral status of an agent’s behavior and, on the other hand, the community’s accountability practices.

I do not find this objection convincing. There are three independent problems with it. I present them in reverse order of importance.

The first problem concerns the notion of “blameless wrongdoing”. It has been argued that the notion is incoherent (Mason 2002). But even if it is coherent, the notion is at odds with our ordinary moral thought and talk. Common sense morality entails a conceptual link between, on the one hand, intentionally committing a significant wrong and, on the other hand, being a legitimate target of moral blame (Copp 2018; Darwall 2006; Wallace 2019). The category of blameless wrongdoing severs this conceptual link with seemingly startling consequences. I could constantly and intentionally commit significant wrongs that it would be illegitimate to blame me for. This strikes me as a weird conceptual implication. Thus, the introduction of blameless wrongdoing would have to be flanked by a non-trivial amount of conceptual engineering. I consider this a high theoretical cost.

A further remark: From a common sense perspective, the introduction of added failures and blameless wrongdoing into our morality system would be practically irrelevant for human communal life. If I ‘commit’ an added failure, there are no practical consequences whatsoever, neither for my relations to others nor for my relation to myself. So, it is questionable whether moral communities would gain anything of any practical importance by introducing added failures and blameless wrongdoing into their morality systems.

The second problem with the objection concerns an important question we have hardly discussed up to this point: Why think that Symmetry is true? Clearly, proponents of Symmetry must provide us with some reason to think so. Typically, their arguments rely on moral intuitions about hypothetical scenarios. Consider Thomson’s first tram/trolley scenario again.

Bystander’s Two Options You can

• (i) do nothing, letting five die, or

• (ii) throw the switch to the right, killing one.

Intuitively, many consequentialists would argue, it is obvious that we should throw the switch and thereby save the greater number. The negative duty against killing could, they continue, not possibly outweigh the prospect of saving five times as many people. As Alaistair Norcross puts it, when non-consequentialists insist on the impermissibility of throwing the switch, they are “fetishizing certain causal processes” (Norcross 2008, 80). Since Asymmetry generates these (supposedly) unintuitive results—Do not destroy one life to save five!—it is false. So, Symmetry must be true.

However, there are structurally analogous cases in which it seems much less fetishistic to insist that doing harm is harder to justify than allowing it. These cases invoke intuitions that point in precisely the opposite direction. In Thomson’s second scenario, Fat Man, Asymmetry generates the intuitively correct result: We may not push the fat man off the bridge. It therefore seems reasonable not to rely on our moral intuitions to adjudicate between Symmetry and Asymmetry. If there is a chance to decide the issue without appeal to moral intuitions, we should take it. The indispensability argument provides this chance.

Finally, here is the third and most important problem I see with the objection that, even though we may be psychologically committed to Asymmetry, Asymmetry could still be false. Let me start by asking: Why should it not be enough to show that creatures like us in a world like ours must build morality systems on the basis of Asymmetry? Would we add anything of any philosophical import by stating that—besides its indispensability—Asymmetry is also true? What exactly would we be adding? And if we cannot but accept Asymmetry as part of the structure of our morality systems, why even worry that it is false? What would that even mean?

It seems clear that proponents of the objection should be able to explain why the falsity of Asymmetry would be so worrisome; why and how it would matter that a claim we cannot but accept is false. In order to do so, they must first explain what the truth of Symmetry supposedly consists in. They need, in other words, an account of moral truth. Without such an account, the objection remains hollow.

Now, there are several possible directions for an account of moral truth: a correspondence with mind-independent moral facts. A correspondence with the advice of an ideal observer; its capturing self-evident moral intuitions; its being part of the result of a reflective equilibrium procedure; and so on. All of these proposals come with their own problems, of course. Discussing them would lead us too far astray. But I want to emphasize that it is not sufficient to point at the logical possibility that Asymmetry could be false, even though we are psychologically committed to it. As long as we are not told why and how the falsity of a claim we cannot but accept should even matter, this strikes me as an empty complaint. Without the support of a plausible account of moral truth, the objection is utterly weightless.

But now suppose someone presents such an account. Suppose someone proves that, say, the objective, mind-independent moral facts yield the falsity of Asymmetry. Even in this case, the indispensability argument would remain sound. It would still be true that the practical necessities of human life require our psychological commitment to Asymmetry. So, even in this case—even if it were an objective, mind-independent moral fact that Asymmetry is false—this would not, and could not, change a thing about the practical necessities of human life. We would still have to build and maintain moral communities. And to this purpose, we would still have to accept Asymmetry. I thus take it that, even if someone came up with an account of moral truth according to which Asymmetry is false, it would not change anything about what I have argued in this paper. We would be ‘pragmatically justified’ in ignoring any claim to the effect that Asymmetry is false. This falsity, whatever it might consist in, simply would not—could not—matter for creatures like us in a world like ours.

This suggests two things. Firstly, the objection we are currently considering rests on a problematic presupposition, namely, that it would (somehow) matter if a claim we cannot but accept is false, even if this falsity could not have any implications with regard to the practicalities of human life. And, secondly, a broadly ‘pragmatist’ account of moral truth seems correct, at least in the case of Asymmetry. According to such an account, some moral truths are grounded in the contingent practical necessities of human life.³⁰ Once we understand that we cannot but accept Asymmetry, we have everything we need. According to pragmatism, the fact that we are committed to Asymmetry rules out its falsity; once we acknowledge the indispensability argument, there is no further question about Asymmetry’s truth-value.

All in all, this strikes me as an adequate reply. But even if you do not share my pragmatist inclinations, my earlier point remains in force: Proponents of the objection have to explain why exactly we should even care about the potential falsity of a claim we cannot but accept; they have to explain why this falsity would, and could, matter. Without such an explanation, the objection remains hollow.

7 Conclusion

In this paper, I have proposed a solution to

Thomson’s challenge We have to explain why Asymmetry is a deep feature of our moral thinking.

I suggested that Asymmetry is a deep feature of our moral thinking because it could not be otherwise. Accepting it is indispensable for building and maintaining stable moral communities—at least for creatures like us in a world like ours. To support this conclusion, I presented and defended the indispensability argument.

The indispensability argument

P1:

Caring about morality is indispensable for creatures like us to build and maintain stable moral communities.

P2:

Creatures like us can only care about morality if morality conforms to Asymmetry.

$\therefore$:

Asymmetry is indispensable for creatures like us to build and maintain stable moral communities.

In defense of P1, I claimed that members of a moral community must care about morality in a balanced way. While they must at all times be ready to devote a possibly significant amount of mental resources to morality, morality must not become all-commanding—it must leave individuals enough mental resources to live the non-moral parts of their lives. Otherwise, people just stop caring about morality, which would render a stable moral community impossible.

In defense of P2, I claimed that any morality system that rejects Asymmetry generates mental overload and, thereby, renders a balanced care about morality impossible. A morality system that conforms to Symmetry—that is, to the idea that doing harm is as hard to justify as allowing it—would entail that we have to care as much about not doing harm as we have to care about not allowing this harm to occur. This, however, would vastly overstrain our mental resources because of (a) the ubiquity of the wrongful harms we know about and (b) the necessity of caring a lot about not causing wrongful harms ourselves. Since the argument is valid, the conclusion follows.

I then turned to two noteworthy implications. Firstly, Asymmetry is not in need of any moral justification because it is a ‘pillar claim’ or axiom for constructing morality, understood as a social practice we cannot live without. This means that the truth or correctness of Asymmetry is no adequate object of discussion in normative ethics. Secondly, Asymmetry can, in principle, be accepted by consequentialists. However, the view loses part of its intuitive appeal by incorporating it. Part of consequentialism’s rationale stems from examples that suggest the truth of Symmetry. If consequentialists incorporate Asymmetry, they can no longer motivate their views with examples of this kind.

I then turned to an important objection. It stated that the indispensability argument misses its target because it confuses moral psychology with normative ethics. While it might be true, the objection went, that we are psychologically committed to accepting Asymmetry, Asymmetry could nevertheless be false (and Symmetry true). I discussed the objection in detail. The most important point was this: Without a worked-out account of moral truth that explains why the falsity of a claim we cannot but accept should matter to us, the objection remains hollow.

I closed by pledging allegiance to a broadly pragmatist account of moral truth. According to this account, the practical necessities of human life can determine moral truths in at least some cases; once it is established that we cannot but accept Asymmetry because we need it to build and maintain something we cannot live without—a moral community—there is no further question about moral truth. Once we realize that a commitment to Asymmetry is a precondition of stable moral communities, we have all the justification we need.

A final remark: Some moral philosophers may wrinkle their brows because my argument is based on contingent anthopological facts; and not, say, on pure laws of reason, self-evident moral intuitions, or a mind-independent moral reality. This, they may think, has an inadequate implication, namely, that Asymmetry is contingent—whereas moral truths are often taken to be necessary. However, note that the anthropological facts I appealed to are both hard-to-deny and unlikely to change very soon. They are robust facts about human nature. Even more, they are mind-independent in the sense that they do not depend on the contents of human mental states. It is not open to us to change the contents of our attitudes or desires such that, after the change, we can do without Asymmetry. So, according to the broadly pragmatist picture I have outlined, Asymmetry can be considered a ‘robust truth’. It is, if you will, ‘anthropologically necessary’.³¹ I believe that this is all the robustness we can achieve in moral philosophy. But, in any case, it is all the robustness we need for stable moral communities.