Keil 1 Benjamin Keil University of Kansas, Ph.D. Candidate (Philosophy) Please do not cite or publish without permission. Immorality Without Harm? Walter Sinnott-Armstrong's book “Morality Without God” argues against divine command theory and, as part of this project, proposes harm-based morality as its replacement. I will examine Sinnott-Armstrong's harm-based account of morality and argue for its insufficiency in two respects. First, Sinnott-Armstrong's resolution of the doing/allowing distinction rests on an ad hoc limitation of moral obligation. Second, because Sinnott-Armstrong's system cannot condemn harmless immoralities, it is incomplete. Based on these flaws, then, I argue that Sinnott-Armstrong's system ought to be rejected. My arguments shall take the following outline. First, I will overview SinnottArmstrong's general account of harm-based morality. Second, I shall examine the doing/allowing distinction and argue that Sinnott-Armstrong's resolution of it rests on an ad hoc restriction. Third, I will provide an example of a harmless immorality and thereby conclude that Sinnott-Armstrong's moral system is deficient. Let us begin, then, with a preliminary explanation of Sinnott-Armstrong's moral system. Sinnott-Armstrong's project is to explain what makes immoral acts immoral. He begins by offering an example of an immoral act, and then explaining what makes it so: “Consider rape. Rape is immoral...The question is not whether rape is immoral, but why it is immoral. On the secular view, the answer is simple: Rape is wrong because it harms the victim for no adequate reason” (Sinnott-Armstrong 57, emphasis in original). On Sinnott-

Keil 2 Armstrong's view, then, what makes rape wrong is the nature of the harm done to another in the act's commission. Sinnott-Armstrong argues, further, that his proposed account explains not only the immorality of rape but the immorality of other acts as well. He writes: “Why is it morally wrong to kill other people? Because it harms them by depriving them of life. Why is it morally wrong to hit, kick, stab, or shoot other people? Because it harms them by causing them pain” (Sinnott-Armstrong 58). Through these and other examples, Sinnott-Armstrong argues that his system is a complete account of morality and, so considered, cannot only claim that an act's being harmful is sufficient for its immorality. If he did, his moral account would be incomplete and further would not be a replacement for divine command theory. Instead, Sinnott-Armstrong's position entails that harm is either a necessary condition or a necessary and sufficient condition for an act to be immoral. Having explained the broad outlines of Sinnott-Armstrong's moral system, let us focus on two of its important specifics. First, when considering a harm-based account of morality, it is necessary to understand exactly what harm is. Sinnott-Armstrong addresses this topic when he writes “Most people agree that harms include death, pain, and disability. Disability includes loss of freedom and maybe also false beliefs insofar as false beliefs make people unable to achieve their goals” (Sinnott-Armstrong 59). Sinnott-Armstrong relies on this “death, pain, and disability” definition throughout the rest of his work, as we shall see. A second important specification of Sinnott-Armstrong's moral account is the “without adequate reason” clause. This denotes that Sinnott-Armstrong does not consider all harms to be immoral; rather, those harms inflicted “without adequate reason” are immoral

Keil 3 (Sinnott-Armstrong 71). He clarifies this stance by noting: “It is obviously not always wrong to harm other people. Judges and police cause harm when they jail criminals. Doctors cause harm when they amputate limbs. But those harms are caused in order to prevent greater harm in the future” (Sinnott-Armstrong 71). That a harm is done “with adequate reason”, then, entails that such a harm is not immoral in Sinnott-Armstrong's account. We now understand both the generalities and two important specifics of SinnottArmstrong's moral system. I shall now offer two refutations of Sinnott-Armstrong's position. First, I argue that Sinnott-Armstrong's resolution of the doing/allowing distinction rests on an ad hoc limitation of its own moral obligation. Before I can argue such, however, I must explain what the doing/allowing distinction is. In general, the doing/allowing distinction turns on the question of whether, and to what extent, there is a relevant moral difference between doing something immoral and allowing something immoral to happen. In relation to harm-based morality, then, almost all agree that doing harm to another is immoral. What must be determined is whether, and to what extent, allowing harm to come to another is also immoral. Sinnott-Armstrong argues that, at least sometimes, both doing and allowing harm to others are equally morally condemnable. He writes “This does not mean that we have no obligation to strangers...The lesson for morality is that we also should not fail to prevent important harms to others when we have no adequate reason not to prevent these harms” (Sinnott-Armstrong 67). In Sinnott-Armstrong's system, then, we morally should not allow important harms to come to others if we have no adequate reason to prevent them. A tension develops, however, for it seems as if Sinnott-Armstrong’s system thus

Keil 4 contains an implausibly heavy moral burden. For indeed, as I write this philosophy paper, I am also neither helping the needy nor donating my life’s savings to Oxfam. In so doing, I am allowing important harms to come to others without adequate reason. Since allowing important harms to come to another without adequate reason is immoral for SinnottArmstrong, it thus seems morally obligatory to devote one's whole life to ameliorating others' important harms. Sinnott-Armstrong recognizes this requirement of his moral system, and thus he proposes to restrict one’s moral obligation as follows: This does not mean that we must always do everything that we can to minimize the amount of harm in the world. That would require too much. We do not have to give to CARE every time we receive a request. Nor are we morally required to volunteer at every soup kitchen. To fulfill that unlimited demand, we would all have to spend our entire lives helping the needy. Still we ought to do something...[A]lthough the limits are controversial, it should not be controversial that we do have at least some positive duties to help the needy and to educate the ignorant in order to prevent the same old harms of death, pain, and disability (Sinnott-Armstrong 68). Sinnott-Armstrong wishes to argue, thus, that his position does not entail doing everything possible to minimize harms experienced by others. However, by all initial appearances, Sinnott-Armstrong does not formalize any restriction on his system's moral obligations whatsoever. But even if this section of his argumentation is understood as offering a restriction on the moral obligations imposed by his system, I argue that such a restriction would be unprincipled and thus would be ad hoc. Allow me to briefly clarify my claim. A moral rule such as “Thou shalt not kill” can have principled or unprincipled restrictions. A principled restriction of the above rule might be “except in self-defense”; this recognizes that any moral rule must cohere with other important moral principles such as the moral right to not be touched without consent, and

Keil 5 such coherence is accomplished through a principled restriction on the moral rule's requirements. However, if I hate Blonde Bill and propose the restriction “unless the victim has blonde hair” only so I can kill him, what I offer is an unprincipled restriction. I argue that, if Sinnott-Armstrong offers a restriction on his system's moral obligation, it is unprincipled and thus is ad hoc. Let us examine Sinnott-Armstrong's argumentation in detail. First, he notes that the full obligation imposed by his system would “require too much” (Sinnott-Armstrong 68). “Too much” is, however, a comparative term – that is, Sinnott-Armstrong is asserting that the moral obligation imposed by his system is too much relative to some other standard. But Sinnott-Armstrong offers no clarification of what this standard is against which “too much” must be judged. He further fails to offer necessary or sufficient conditions by which we could evaluate whether a given moral obligation, in fact, requires “too much”. Most importantly, Sinnott-Armstrong offers no principled reason why a moral obligation should indeed be restricted whenever it does require “too much”. Since Sinnott-Armstrong offers no standard, no way for the standard to be applied, and no principled justification of his implicit standard, we must conclude that SinnottArmstrong offers no principled reason for any purported restriction put on his own system’s moral obligation. Sinnott-Armstrong's argumentation continues, then, by offering examples of what such an unlimited demand would look like. However, such examples by themselves are presumably not principled reasons to restrict his moral system's obligation either. Sinnott-Armstrong closes his argumentation on this issue by noting “To fulfill that unlimited demand, we would all have to spend our entire lives helping the needy” (Sinnott-

Keil 6 Armstrong 68). Of course, this is indeed a consequence of his proposal. However, the mere fact that a moral system would require us to spend our entire lives helping the needy is not by itself a principled reason to restrict its moral obligations. Instead, it indicates a flaw in the moral system itself. So considered, the appropriate remedy is not to offer an ad hoc restriction on the unduly burdensome requirements of Sinnott-Armstrong's moral system, but to reject it altogether. In short, then, it is questionable whether Sinnott-Armstrong offers a restriction on the overly burdensome requirements of his moral system at all. But even assuming he does, the restriction offered is not principled and thus is ad hoc. Since having ad hoc limitations on moral rules is not a desirable feature of any moral system, Sinnott-Armstrong's account of morality ought to be rejected. I shall now offer my second argument: Sinnott-Armstrong's system of morality cannot condemn harmless immoralities, and thus is incomplete. Sinnott-Armstrong recognizes that such an argument, if successful, would undermine his system. He writes: “Real trouble seems to arise for a harm-based account [of morality] if any immorality is harmless. Then that account covers only part of morality, not all of it” (Sinnott-Armstrong 80). Perhaps unsurprisingly, I fully agree with Sinnott-Armstrong: if harmless immoralities did exist, this would undermine the sufficiency of his moral system. I shall now offer just such a case. The example I propose, unlike many in philosophy, is a real-world event and so, if successful, demonstrates the existence of an especial flaw in Sinnott-Armstrong’s moral system. According to a recent news article, a man was arrested for programming 100 Bank of America ATMs to, when a certain input was received, undergo a computer error. This

Keil 7 error triggered the ATM to dispense all its remaining cash. So once Bank of America installed the otherwise functional ATMs, the programmer/thief went to them, triggered computer errors, and stole over $300,000 (Zetter 1). I argue that, properly understood, this is a case of immorality without harm. My position does rest on a preconception, however, which I shall now make explicit: What the thief did was immoral. He stole $300,000. On the explicit assumption that the robbery was immoral, then, I now argue that it was harmless. There are two normal avenues of harm which are not open in the example I give. First, unlike a “holdup”, the robber did not physically enter the bank with a gun demanding cash. If our thief had, one could argue that the individuals in the bank were emotionally harmed by the holdup, and so Sinnott-Armstrong's harm-based account of morality would morally condemn what transpired. Alternately, in a normal holdup, the bank might have been damaged (perhaps the robber's gun would have been fired, and harm to property would have resulted). But due to the specifics of this case, these sorts of harms do not exist in our example. Second, unlike most electronic thefts, nobody's account or personal information was endangered by this robbery. More precisely, our thief did not take $300,000 out of various persons' bank accounts; the software error merely caused the ATM to dispense money. It did so, moreover, without harming anyone's account balance or compromising their personal information. If such had happened, a clear sort of harm would exist. But, once again, our example does not leave this avenue of harm open. So since these two ordinary sorts of harm do not occur in our example, where might

Keil 8 the harm lie? The obvious harm one would identify, in this case, is financial. The bank was, after all, robbed of $300,000 and this financial loss is a serious harm. I instead argue that, properly considered, Bank of America underwent no financial harm in this case. For indeed this bank, like all banks, carries robbery insurance. So when the crime was discovered, their loss was entirely compensated and so the bank was not financially harmed. Well and good, one might claim, but then surely the insurance company was financially harmed. After all, it did have to pay a large claim. This is incorrect, though, because from the insurance company's point of view the payout's cost will be recovered, by all reasonable expectations, through charging slightly higher insurance premiums to all the banks it insures. Thus the insurance company is not financially harmed by the robbery either. But this only means, it seems, that the financial harm has been shifted back to the banks – each one now has to pay slightly higher insurance premiums than it otherwise would have. But there is no reason to think this cost would be borne by the banks – instead, they will cover the added insurance cost by slightly raising the fees charged to all their customers. Considered thus, all the banks together were not financially harmed by our thief either. But then, if the customers ultimately bear the financial costs, surely they are the ones harmed by the robbery. At this point, though, the financial cost has been very widely dispersed. For it is not just distributed among all the banks covered by the insurance company. It is actually dispersed among all the customers at all those banks. At this point, then, each customer is paying just a fraction of a cent more than he otherwise would have. The extremely wide dispersal of the robbery's financial cost, then, means that nobody was, financially speaking, harmed by the robbery. And, of course, if nobody was harmed by the

Keil 9 robbery, Sinnott-Armstrong's moral account cannot morally condemn our thief's act. Two primary sorts of responses could be offered against the example of harmless immorality I offer. First, one could argue that our thief did cause some sort of financial harm. Second, one could argue that, although no financial harm took place, non-financial harm did. I will address these in turn. First, one could argue that for some reason, financial harm will be experienced in the example I have provided. Perhaps the insurance company will not ultimately pass its financial costs on at all, or perhaps the increased financial costs will be borne only by Bank of America's customers. Perhaps Bank of America will distribute the costs among its customers unevenly. Any number of similar stories may be told, but they all allege that some individuals would experience financial harm in the example I have provided, and thus that Sinnott-Armstrong's account could condemn the robbery as being immoral. I answer that such replies hide a problem. Indeed, these replies argue that financial harm did occur and hence the robbery was immoral. But, as I have argued at length, the costs could have been distributed otherwise such that there would be no harm. So even if one argues that there would be harm in my example, I show that harm did not have to occur. So considered, harm becomes an accidental rather than an essential feature of this robbery. Immorality thus becomes an accidental rather than essential feature of the robbery. Such replies, then, only succeed at saving Sinnott-Armstrong's position by conceding that, had different circumstances obtained, the same robbery would not have been immoral at all. Second, one could concede that there was no financial harm, but argue that some nonfinancial harm took place. I obviously cannot here respond to all possible stories one might

Keil 10 tell of non-financial harm, but I argue such replies also hold little promise. It might be argued that Bank of America's reputation was harmed through this robbery, and this reputational harm makes the robbery immoral. Once again, though, this makes the act's harm, and thus its immorality, an accidental feature of the problem. For if the robbery were never discovered or if Bank of America had covered it up, the same robbery would not have harmed the bank's reputation, and so would not have been immoral. Alternately, one might argue that even though the financial harm was recompensed, there still was harm. After all, if my car is stolen but my insurance company replaces it the same day, it seems I have still been harmed. So considered, this reply alleges that even though there was no unrecompensed financial harm, harm still occurred. Such an analogy is inappropriate, however, for $300,000 is a very small amount of money to Bank of America. So imagine instead that you own 5,000 cars and one is stolen but quickly replaced by your insurance company – it is less clear that you have been harmed at all. Moreover, to return to the example I provided, banks never have enough cash on hand to pay out all their deposits at once. So considered, even though the bank had $300,000 less cash after its ATMs were emptied, it underwent no greater risk of harm than that it normally bears. So even if recompensed harms still are harms, the specific nature of our example entails that such a counter-argument is not successful. In short, I have argued that two flaws exist in Sinnott-Armstrong's harm-based account of morality. I argued both that Sinnott-Armstrong's resolution of the doing/allowing distinction rests on an ad hoc restriction and second I provided an example of a harmless immorality, thus showing that his moral system is incomplete. On these grounds, then, I

Keil 11 argue that Sinnott-Armstrong's harm-based account of morality is sufficiently problematic so as to deserve rejection.

Keil 12 Bibliography:

Sinnott-Armstrong, Walter. Morality without God? Oxford: Oxford UP, 2009. Print. Zetter, Kim. "Bank Worker Pleads Guilty to Hacking 100 ATMs." Wired News. 13 Apr. 2010. Web. 21 Apr. 2010.