The Worse than Nothing Account of Harm: A Fallen Hero

Abstract

Daniel Immerman has recently put forward a novel account of harm, the Worse than Nothing Account. We argue that this account faces fatal problems in cases in which an agent performs several simultaneous actions. We also argue that our criticism is considerably more powerful than another one that has recently been advanced.

Many prominent accounts of harm are “comparative,” in that they claim that an event (such as an action) harms a person if and only if it leaves them worse off than they are in a certain “baseline” situation. While various well-known attempts at specifying the relevant baseline are seriously problematic, Daniel Immerman (2022a, 2022b) has recently defended a novel alternative, which he claims is “the hero of this story” (2022a: 12):

The Worse than Nothing Account of Harm (WTNA): A's doing x harms B just in case it leaves B worse off than if A had done nothing to B. (Immerman 2022a: 12)

Like its competitors, Immerman says, WTNA should be understood as an account of all things considered, as opposed to pro tanto, harm. Immerman also stresses that “A's doing x” should be understood broadly; for example, it covers not only actions but also events like the falling of a tree branch. Immerman does not commit himself to any specific understanding of “if A had done nothing to B,” except that it should be taken to involve an implicit “during the same time period at which [A does x]” qualification (2022a: 13).

Immerman argues that WTNA nicely avoids the “preemption” problem, which is often regarded as the most important challenge for the currently most popular account of harm:

The Counterfactual Comparative Account of Harm (CCA): A's doing x harms B just in case it leaves B worse off than if A hadn't done x.¹ (Immerman 2022a: 6)

Immerman uses the following common illustration of the preemption problem:

Dark Knight. Bobby Knight gets mad at a philosopher, Phil, in response to a perceived slight. So he chokes Phil. Luckily for Phil, Knight is in anger management. It's taught him to better control his behavior. He choked Phil because he applied anger management techniques. If he hadn't, he would have dismembered Phil. (Hanna 2016: 263; original case in Norcross 2005: 165–66)

Intuitively, although the choking preempts the dismemberment, it still harms Phil. However, CCA implies that it does not, as Phil would have been even worse off without it. WTNA, by contrast, implies that the choking does harm Phil, as he would have been better off if Knight had done nothing to him.

Despite WTNA's significant virtues in this and other respects (Immerman 2022a: 12–16; 2022b: 29–36), it seems to us to have a fatal flaw that has not yet been noted. WTNA's fatal flaw is that it has absurd consequences in various cases in which the agent performs several simultaneous actions. Here is one example:

Hard Kick, Nice Song. Alexander kicks Brian very hard, which causes Brian very intense pain. Alexander simultaneously also sings a nice song for Brian, which causes Brian fairly intense pleasure. If Alexander had not sung for Brian, he would not have caused him pleasure in any way, but would still have kicked him.²

Clearly, Alexander's singing does not harm Brian. This is also what CCA implies. But WTNA implies, absurdly, that Alexander's singing does harm Brian. This is because Brian would have been better off if Alexander had done nothing to him. If Alexander had done nothing to Brian, then Brian would not have been kicked and thus not felt any pain.

In Hard Kick, Nice Song, WTNA overgenerates harm. There are also cases with the same structure where WTNA undergenerates harm. Here is one example:

Softer Kick, Nicer Song. Alexandra kicks Brianna fairly hard, which causes Brianna fairly intense pain. Alexandra simultaneously also sings a very nice song for Brianna, which causes Brianna very intense pleasure. If Alexandra had not kicked Brianna, she would not have caused her pain in any way, but would still have sung for her.

Clearly, Alexandra's kicking harms Brianna. This is also what CCA implies. But WTNA implies, absurdly, that Alexandra's kicking does not harm Brianna. This is because Brianna would not have been better off if Alexandra had done nothing to her. If Alexandra had done nothing to her, she would have been deprived of the song and the pleasure. Indeed, since Brianna would have been worse off had Alexandra done nothing to her, an account of benefit that mirrors WTNA implies, even more absurdly, that Alexandra's kicking benefits Brianna.

Again, Immerman stresses that WTNA easily handles preemption cases like Dark Knight. In response to a slightly earlier defense of WTNA (Immerman 2022b), we have argued (Johansson and Risberg 2023) that the account nevertheless faces problems in cases that involve something close to preemption. It seems to us, however, that our novel criticism of WTNA is considerably more powerful. Here is one of our earlier cases:

Birthday Night. Batman wants to give Robin some golf clubs for his birthday, but does not know very much about the subject. Thus, the night before Robin's birthday, he asks Catwoman, a golf expert, for help. Catwoman recommends that he give Robin an excellent set of golf clubs, which Batman does, leaving Robin very well off. If Catwoman had not made that recommendation, she would have told Batman to give Robin some cheaper golf clubs, which would have left Robin much worse off. If Catwoman had not done anything at all in response to Batman's request for help, however, Batman would have bought an absolutely fantastic set of golf clubs to Robin as a result of his ignorance. That would have left Robin even better off than he in fact is. (2023: 238)

Intuitively, we claimed, Catwoman's recommendation does not harm Robin. This is also what CCA yields. WTNA, however, implies that Catwoman's recommendation does harm Robin.

Here is another of our earlier cases:

Dark Knights. Phil meets Bobby Knight on campus and asks him for the way to the library. Annoyed that Phil talks to him, Knight chokes Phil. Had he not done that, he would have told Phil to go to the building to the left, where the library is in fact located. If Knight had not done anything at all, however, Phil would have gone into the building to the right. There he would have met Bobby Knight's equally aggressive brother, Billy Knight, who would have choked him. (2023: 236)

We claimed that in Dark Knights, just as in Dark Knight, the intuitively correct judgment is that Bobby's action harms Phil. Again, this is also CCA's verdict. WTNA, by contrast, implies that the action does not harm Phil.

We still find our earlier criticism fairly persuasive. It must be admitted, however, that WTNA's implications in Birthday Night and Dark Knights are not obviously unacceptable. Note, in particular, that in Birthday Night, there is already a process in motion – initiated by something like Batman's decision to give Robin a present – that is independent of Catwoman and that, if uninterrupted, would have resulted in Robin's being better off than he actually is. On the most natural understanding of the case, at least, Batman would have given Robin the fantastic set of clubs if Catwoman had not been present at all. Since Catwoman's action leaves Robin worse off than he would have been if this process had continued uninterrupted, it does not strike us as indefensible to claim that it harms him. Correspondingly, in Dark Knights, there is already a process in motion – initiated by something like Phil's going to campus – that is independent of Bobby and that, if uninterrupted, would have resulted in Phil's being no better off than he actually is. On the most natural understanding of the case, at least, Phil would have gone to the building to the right, and thus been choked by Billy, if Bobby had not been present at all. Since Bobby's action leaves Phil no worse off than he would have been if this process had continued uninterrupted, it does not strike us as indefensible to deny that it harms him.³

With regard to Hard Kick, Nice Song and Softer Kick, Nicer Song, by contrast, nothing like these responses is available. Clearly, there is no reason to think that Alexander's singing interrupts a process that is independent of him and that, if uninterrupted, would have left Brian better off than he actually is. Equally clearly, there is no reason to think that Alexandra's kicking interrupts a process that is independent of her and that, if uninterrupted, would have left Brianna no better off than she actually is.

More generally – and regardless of the comparison with our earlier cases – it simply seems to be a non-starter to try to defend the claims that Alexander's singing is harmful to Brian, and that Alexandra's kicking is harmless to Brianna. Someone who wants to defend WTNA thus has to argue instead that the view does not actually imply these claims. But this strategy is also unpromising. One possible idea, for instance, is that Alexander and Alexandra do not in fact perform multiple actions, but only one action each – roughly, that of singing-a-song-while-kicking-someone. Given this idea, WTNA implies that Alexander's action is harmful and that Alexandra's action is harmless, and avoids implying the claims that we have criticized – at least if it is also added that in each case, there are not even any multiple events such as the singing and the kicking. However, this idea evidently rests on a highly counterintuitive and unusual view of the ontology of actions (and events in general) – one to which WTNA adherents should surely avoid being committed. Moreover, the suggested idea leaves WTNA unable to accommodate the extremely plausible verdict that each one of Alexander and Alexandra does something that is harmful and something (else) that is harmless.⁴

Another type of response is to replace WTNA with some modified theory that better handles the cases that we have presented. For example, while WTNA as it stands addresses the question of what it takes for an event to harm someone, our arguments might be taken to suggest that a WTNA-style view about what it takes for a person to harm someone is more promising. Consider this view:

WTNA-2: A harms B just in case B is worse off than B would have been if A had done nothing to B.

WTNA-2 implies that Alexander harms Brian, which is clearly the right result. It also implies that Alexandra does not harm Brianna. While that result may initially seem plausible, we think that any such appearances are deceiving. For, as we have argued, it seems clear that Alexandra does something that harms Brianna – namely, kicking her. And it is highly plausible that facts about people harming people obtain in virtue of facts about events harming people: for A to harm B, we submit, just is for A to do something that harms B.⁵ (For instance, this view straightforwardly explains the incoherence of statements such as “Even though nothing you have done has harmed me, you have nonetheless harmed me.”) Thus, we should in the end say that Alexandra does harm Brianna, contrary to what WTNA-2 implies.⁶

If the verdict that Alexandra does not harm Brianna initially appeared plausible, that might be because there is a related verdict that is in fact plausible: that Alexandra's “total effect,” as it were, on Brianna is not harmful to Brianna. Consider, then, this variant of WTNA:

WTNA-3: A's total effect on B is harmful to B just in case B is worse off than B would have been if A had done nothing to B.

However, WTNA-3 is a much less interesting claim than WTNA. The central question in the harm debate has been the one that WTNA (as formulated at the outset) attempts to answer: that of what it takes for an event to harm a person. That focus, moreover, is well-motivated, since facts about events harming people seem more fundamental than many other harm-related facts, including not just facts about people harming people (as noted above) but also facts about the harmfulness of people's total effects on people. Surely, whether A's total effect on B is harmful depends on facts about whether and to what extent certain events (such as A's actions) harm B. Thus, having to settle for WTNA-3 would be a severe disappointment for those who took WTNA to be a promising candidate in the debate on harm.⁷

The last two revisions of WTNA that we shall consider do not disappoint in this way, as they do address the question of what it takes for an event to harm a person. The first of these tries to deal with cases involving multiple simultaneous actions by comparing the actual world with the nearest world in which A does not do x and also does not replace x with anything – or in other words, does not do anything instead of x:

WTNA-4: A's doing x harms B just in case it leaves B worse off than if A had not done x and also would not have done anything instead of x.⁸

We take it that while the notion of doing something instead of x is unclear in some ways, it is sufficiently clear for present purposes. (Plausibly, a paradigmatic case is one where A deliberates between doing x and doing some other thing, and ends up doing not x but that other thing.) Crucially, a world where A does neither x nor anything instead of x, need not be a world where A does nothing at all to B. In particular, in Hard Kick, Nice Song, there are possible worlds where Alexander neither sings for Brian nor does anything instead of singing, but still kicks Brian. It may be suggested, then, that the following counterfactual is true: if Alexander had neither sung for Brian nor done anything instead of singing, then he would still have kicked him. If so, WTNA-4 has the desired result that Alexander's singing does not harm Brian. Analogous remarks apply to Softer Kick, Nicer Song.

In fact, however, there is nothing in Hard Kick, Nice Song that requires this counterfactual to be true. Imagine, for example, that Alexander is determined to both sing or talk to Brian (the latter of which would not cause Brian any pleasure) and kick him, and is completely uninterested in doing only one thing to Brian. For instance, Alexander might be solely motivated by a desire to demonstrate his capacity for certain forms of multitasking. On this way of filling out the details of the case, it is unproblematic to suppose that if Alexander had neither sung for Brian nor done anything instead of singing (such as talking to Brian), then he would not have done anything at all to Brian (and would thus not have kicked him).⁹ On this supposition, WTNA-4 – unlike CCA – implies that Alexander's singing harms Brian. But that implication remains clearly false. Again, analogous remarks apply to Softer Kick, Nicer Song.

The final revision that we shall consider compares possible worlds in which A does only one thing – namely, the action whose harmfulness is being assessed – with worlds in which A does nothing to B:

WTNA-5: A's doing x harms B just in case B would have been worse off if A had only done x than if A had done nothing to B.¹⁰

WTNA-5 has intuitively correct results in both cases we have presented. In particular, since Brian would not have been worse off if Alexander had only sung for him than if Alexander had done nothing to him, WTNA-5 implies that Alexander's singing does not harm him. And since Brianna would have been worse off if Alexandra had only kicked her than if Alexandra had done nothing to her, WTNA-5 implies that Alexandra's kicking harms her.

Nevertheless, WTNA-5 is not plausible. To begin with, it would be extremely surprising if the actual harmfulness of an actual event could depend entirely on a comparison between two non-actual worlds. Moreover, there are other cases involving multiple simultaneous actions that WTNA-5 handles badly, and indeed worse than WTNA. Here is one:

Two Chemicals. Alexandria simultaneously injects two chemicals, X and Y, into Brianne's body. Receiving both these chemicals in combination causes a person enormous pleasure, but receiving only one of them causes enormous pain instead.

WTNA-5 entails that Alexandria's injecting X actually harms Brianne, since Brianne would have been worse off if Alexandria had only injected X than if Alexandria had done nothing to her. For similar reasons, it also entails that Alexandria's injecting Y actually harms Brianne. These results are highly counterintuitive, and the suggested explanation for them that WTNA-5 provides does nothing to soften the blow. Indeed, the appeal to facts about what would have been the case if Alexandria had only done one thing to Brianne seems entirely beside the point, since Alexandria did not in fact do only one thing to Brianne. WTNA, by contrast, implies that neither of Alexandria's actions harms Brianne, as Brianne would not have been better off than she actually is if Alexandria had done nothing to her.

It is important to note, however, that while Two Chemicals thus provides a reason to prefer WTNA over WTNA-5, it does not provide any reason to prefer WTNA over its more familiar competitors. In particular, since neither of Alexandria's actions leaves Brianne worse off than she would have been if it had not been performed, CCA also has the result that neither of them harms Brianne. And as already noted, CCA also handles with ease the cases that present problems for WTNA. Hence, if any account is “the hero of this story,” it is not WTNA.