1. INTRODUCTION
For many years now, moral philosophers have been debating the conceptual and moral status of the distinction between intending harm and (merely) foreseeing harm. And though some of the examples often discussed in this context have been examples of state action (such as the comparison of strategic bomber and terrorist bomber, of which more shortly), still, to the best of my knowledge, no explicit attention has been given in this context to the important particular instance of state actions. In this paper I argue that the general reasons to be suspicious about the moral significance of the intending-foreseeing distinction apply in an especially forceful way to state action.
The discussion proceeds as follows. In Section II, I present the intending-foreseeing distinction in detail and emphasize its initial appeal. I then proceed to survey, in Section III, some objections to the distinction's moral significance. In Section IV, I proceed to argue that however weighty these objections are in general, some of them are much more powerful when understood as objecting specifically to the application of the intending-foreseeing distinction to state action. I present a short conclusion in Section V and in Section VI, I present a research program: I hint at some of the theoretical discussions—in political philosophy and legal theory—that could benefit from thinking about the applicability of the intending-foreseeing distinction to the state. The list, it seems to me, is impressive enough to call for further work on these matters. In an Appendix I argue for a lemma used in Section VI, the claim that if neither the intending-foreseeing distinction nor any related distinction can be defended, consequentialism is the way to go.
II. THE DISTINCTION BETWEEN INTENDED AND (MERELY) FORESEEN CONSEQUENCES
A typical terrorist intends serious harm to noncombatant innocents. And while someone engaging in counterterrorist warfare may cause just as much harm to innocents as the terrorist does, he typically (so we would like to hope) does not intend such harm. True, given the relevant statistics, the counterterrorist engaging in risky, collateral-damage-causing activities knows it is very likely that he will cause serious harm to innocents; he expects such harm. But he does not intend it. And this is no mere psychological difference for it has practical implications: The terrorist typically does what he can to maximize the extent to which his actions harm innocents (or at least to insure that they cause enough such damage), whereas the (interesting) counterterrorist may very well make considerable efforts in order to minimize both the likelihood that harm to innocents occur and the extent of such harm should it occur. The terrorist and the counterterrorist—stipulatively thus described—can serve as an example of the intending-foreseeing distinction, and indeed, the intuitive moral difference between them may be thought of as evidence for the moral significance of the intending-foreseeing distinction.Footnote 1
Note that the distinction between intended and foreseen consequences is not the distinction between consequences the agent wants intrinsically and those she wants only instrumentally or extrinsically, because of their (for instance, but not necessarily, causal) relation with other consequences. The distinction is not, in other words, between ends and means. The harm to innocents is, after all, not the (interesting) terrorist's end but rather a means to the end of national liberation (or whatever). The distinction should be understood, rather, as the distinction between ends and means on one side, and side effects on the other. For the terrorist, the harm to innocents is an intended means to his end, whereas for the counterterrorist, the harm to innocents is a merely foreseen side effect.
In the following sections, I am going to be critical of this distinction, offering objections to its moral significance. In order to appreciate and evaluate these objections, though, it is important not to underestimate the role of the distinction in our commonsensical moral thinking. The two well-known cases—Trolley and Transplant—can work as well as any in making this clear.Footnote 2 In Trolley you have to decide whether to divert a runaway trolley about to crash into and kill five people to a sidetrack on which it will crash into and kill only one. In Transplant you have to decide whether to kill a healthy patient in order to harvest his organs and save five other patients desperately awaiting organ donations. Pretheoretically, we seem to think that it is morally permissible (and perhaps even required) to divert the trolley but that it is morally impermissible to kill the patient in Transplant.
The philosophical challenge here, of course, is to point to a rationale that could account for and justify this difference in our pretheoretical moral judgments, and at least one initially plausible attempt at this challenge points to the distinction between intending and foreseeing: For in Trolley, the death of the one is a merely foreseen side effect of diverting the trolley (and so of saving the five),Footnote 3 whereas in Transplant the death of the one is an intended means to the end of saving the five. If—as a result of the objections that follow or for some other reason—the intending-foreseeing distinction must be discarded as lacking any (intrinsic) moral weight, and if no amended distinction can be made to work, it is no longer clear that we have the resources to defend our pretheoretical moral distinction between Trolley and Transplant. And the same goes, of course, for many other pretheoretical judgments and distinctions.
Now, at the end of the philosophical day, perhaps these are all going to have to be viewed as bullets that have to be bitten, because the intending-foreseeing distinction (and related distinctions) cannot withstand criticism. But let us not kid ourselves about the significance of such a move. Abandoning this distinction (and related distinctions) is going to have to involve a revolution in commonsense morality, nothing less. And though such a revolution may be called for, its depth and scope should not be underestimated.
Let me note here that this common way of putting the distinction—between intended means and foreseen side effects—actually conflates two ways of understanding the distinction, two ways that the literature is often not careful enough in distinguishing. One way of understanding the distinction is as a distinction between, perhaps somewhat roughly, two mental states: that of intending and that of foreseeing. The other way of understanding it is as a distinction between two types of causal structures. On the former understanding, the intrinsically relevant difference between Trolley and Transplant is that in Trolley one merely foresees the death of the one, while in Transplant one intends the death of the patient. On the latter understanding, the relevant difference is that in Trolley the death of the one is a mere side effect, whereas in Transplant the death of the one is a means, lying on the causal way to saving the five.Footnote 4 Now, these two ways of understanding the distinction may very well be related. But they are still distinct.Footnote 5 And as will become apparent in what follows, it is important to keep the two different ways of understanding the distinction in mind, because some of the critiques of the distinction apply to one understanding but not to another.Footnote 6
Though there may, of course, be other ways of understanding the intending-foreseeing distinction, these two are the most common in the literature, and furthermore I am not aware of any suggested understanding of the distinction that is not seen upon reflection to be either an instance of the mental-state distinction, or an instance of the causal-structure distinction, or of some complex function of the two.Footnote 7 For the purposes of this paper, then, we can safely focus on just these two understandings of the intending-foreseeing distinction.
III. CRITIQUES OF THE MORAL DISTINCTION BETWEEN INTENDED AND (MERELY) FORESEEN CONSEQUENCES
The distinction between intending and foreseeing has been the object of much criticism in the past few decades. Some discussions concentrate on the conceptual question: whether we can even draw a coherent distinction between intended and foreseen consequences, and if so, how this distinction is best understood. This will not be my concern below. Indeed, I will assume that even if the distinction is vague and not at all easily made into a completely explicit set of necessary and sufficient conditions for either intending or foreseeing, still we are perfectly capable of drawing the distinction in many paradigmatic cases (such as, for instance, Trolley and Transplant, or the case of typical terrorist and counterterrorist acts). And this suffices, it seems to me, to show that the distinction is not without content (and not more problematic than many other conceptual distinctions we get by with rather well): Conceptual distinctions in general come rather cheap.
The more interesting question is, I think, whether the conceptual distinction we do in fact make is of any (intrinsic) normative significance.Footnote 8 And in what follows I discuss only objections that are meant to push us in the direction of giving this question a negative answer.
A. Counterexamples
In the Loop version of TrolleyFootnote 9 the two tracks—the one with the five and the one with the one—are linked at their far end, so that the trolley can go from one to the other. Now assume that in this case diverting the trolley will only have the intended good effect—saving the five—if the one is hit (and killed), because otherwise the trolley will continue down the looping track, and hit the five from the other side. Only if it crashes into the one will it slow down and not reach the five. In Loop, then, killing the one is no longer a merely foreseen side effect but is an intended means to saving the five, and this is the case, it seems, however exactly you understand the distinction between intending and foreseeing or between means and side effects. So assuming that the distinction between intending and foreseeing is what makes the difference between Trolley and Transplant, we should conclude that in Loop it is no longer morally permissible to divert the trolley.
But this result, argues Thomson, is doubly counterintuitive: It is counterintuitive, first, in that it seems to us permissible to divert the trolley even in Loop, and second, because it is astonishingly implausible to suppose that the difference between Trolley and Loop—that extra bit of track that connects the two tracks—is a difference that makes a moral difference. So what we have here is a counterexample to the claim that the intending-foreseeing distinction accounts for the moral difference between Transplant and Trolley.Footnote 10
B. The Moral Insignificance of the Causal Order
One way of understanding the intending-foreseeing distinction is, remember, as the distinction between two kinds of causal structure: In Transplant, the death of the one is on the causal way to the saving of the five. Not so in Trolley, where the death of the one is on a different causal route, causally after the means (diverting the trolley) that saves the five. But when thinking not so much about particular cases and our pretheoretical reactions to them but directly about the distinction based on the causal order, it is hard not to doubt the moral significance of this distinction. Why think that the causal order has such far-reaching moral implications? Upon reflection, causal-order facts seem of the wrong kind to have intrinsic moral weight. Facts about the well-being of people, for instance, seem to be of the right kind to make a moral difference. Facts about people's choices also seem reasonably good candidates to make a moral difference. And facts about some causal relations may also be of the right kind to make a moral difference—for instance, facts such as which action causes which effects on well-being. But facts about causal order—what is causally on the way to what—look simply (intrinsically) morally weightless.Footnote 11
To see this more clearly, think about a moral agent deliberating about a possible future action of hers—whether or not to act in a certain way. We give this agent some relevant information to help her make the morally right decision, and she can ask for more information if she thinks she needs it. If she then asks for further information regarding the effects of the relevant action (or inaction) on the well-being of some person, her question seems appropriate. If she asks for further information regarding, say, the distribution of hair on the head of one of the affected persons, her question is not appropriate (at least absent some story explaining why it is that the distribution of hair on that person's head is morally relevant).
Thinking about this agent gives us a kind of a test—I will call it “the appropriate-question test”—for the (intrinsic) moral relevance of a given consideration. And it is a test that focuses on the perspective that is morally most important—the first-person perspective of the deliberating agent.Footnote 12
Let us apply this test, then, in order to check whether causal order is morally significant. Think, then, of our agent, deliberating whether to press the button in front of her. We give her information about the states of affairs that will obtain if she does—and if she does not—press it. We tell her, for instance, that if she presses the button, certain good effects and also certain bad effects will follow (and that they will not follow if she does not press the button), and we describe these effects in detail. She then proceeds to ask whether the bad effect is on the causal way to the good effect. Is her question appropriate? Is it more like the question about further effects on people's well-being or more like the question about the distribution of hairs on someone's head? To my ears, her question sounds weird, surprising, indicative of rather disturbing facts about her moral character.Footnote 13 Given a full description of the relevant consequences, and without some further (for instance, instrumental) story explaining how the exact causal structure is morally significant, the causal order seems (to me) simply morally irrelevant. If you agree with me that the question about the causal order is inappropriate, you have strong reason to suspect that causal order in general, and in particular the distinction between means and side effects, is simply (intrinsically) morally irrelevant.Footnote 14
Let me emphasize here—in case you are not yet convinced—that what is at issue is not any old way in which the causal structure may be normatively relevant. What is at issue—and what the appropriate-question test is supposed to help us with finding out—is whether the causal structure is intrinsically morally relevant, whether, in other words, it is morally relevant regardless of its relations to other factors. So it will be no reply to the line of thought in the previous paragraphs to show that, say, causal-structure facts are correlated with other facts, themselves normatively significant, and can thus serve as reasonably good proxies for them. The question is, rather, whether—holding all other things equal—the causal structure itself makes a moral difference. And here the answer that seems to me overwhelmingly plausible is that it does not.
It may be objected that the appropriate-question test, as many other intuition pumps, is too sensitive to the way the case is described, or in this case to the way the relevant question is formulated. Thus, if the question is put in the terms I have been using—“Will the bad effects be on the causal way to the good effects?”—perhaps the question sounds inappropriate. But if the question is put differently—say, “Will I be using the harmed person merely as a means?”—the question sounds perfectly appropriate. Does this show that the appropriate-question test is misguided (because too easily manipulable), or indeed that I have been applying it tendentiously?Footnote 15
It will come as no surprise that I think the answer is “no.” True, proponents of the moral significance of the intending-foreseeing distinction (in either of its meanings) often use such Kantian locutions (“using merely as a means”) as ways of making this claim plausible. Thus it is often said that by diverting the trolley, we will not be using someone merely as a means (in Trolley, at least, for it is not at all clear this is the case in Loop), whereas in Transplant we will be using the person whose organs we take merely as a means. Or perhaps it can be said that in Transplant we would be appropriating someone's body against their will, which we would not be doing in Trolley.Footnote 16
But such ways of describing the case—though plausible-sounding—do not help, for two reasons. First, proponents of such ways of talking have to fill in the details in a plausible way, and it is notoriously hard to say what treating someone merely as a means exactly comes to, why exactly it matters morally, and why it applies to Transplant but not to Trolley (and the analogous point applies, it seems to me, to appropriation as well).Footnote 17 Such ways of talking, then, are best seen not as a solution to our problem, but rather as more elegant names for the problem itself. Second, and perhaps more relevant in the context of the appropriate-question test, such ways of talking are normatively loaded, and so formulating the relevant question in such terms illegitimately biases the appropriate-question test. In order to test for intrinsic moral significance, the question has to be put in the normatively thinnest possible terms. Only if the question—thus put—seems appropriate is the relevant factor intrinsically morally relevant.
Suppose, for instance, that we want to know whether the fact that an act will constitute a killing is intrinsically morally significant and that we plan to find out by employing the appropriate-question test. The way to formulate the question, then, is: “But will it be a killing?” Formulating the question as “But will it be a murder?” is cheating, of course, because “murder” is such a normatively loaded word that its appearance in the question corrupts the appropriate-question test. Similarly, then, the right question to put in our context is not “But will I be appropriating another person?” (for “appropriating a person” is a highly normatively loaded expression) or “But will I be using her merely as a means?” (for “using merely as a means” is—at least among moral philosophers—a highly normatively loaded expression), but rather something like “But will the harm I cause her be on the causal route to the good I cause him?” If you—like myself—are unwilling to answer this question in the positive, you must not hide behind tendentious, normatively loaded alternative descriptions of the case and should instead conclude that causal-structure facts are without intrinsic moral weight.
Of course, as with all controversies regarding whether something is intrinsically morally significant, it is hard to construct sophisticated arguments one way or another—pretty much the best that can be done is to get misunderstandings out of our way, focus as clearly as we can on the precise question in front of us, and give the most plausible answer. And people may differ about what it is, of course. But it is in this spirit that I offer the appropriate-question test—when the misunderstandings above are removed and we focus clearly on the question of intrinsic moral significance, the claim that causal structures are significant in this way seems (to me) just too much to believe.
Consider one last objection to the employment of the appropriate-question test as a way of showing that causal structures are intrinsically insignificant.Footnote 18 Suppose our agent, having been presented with all the information about her possible actions' effect on people's well-being and the like, proceeds to ask: “But wait a minute, am I in Transplant or in Trolley?” Is this question not appropriate?
It certainly sounds appropriate, but this is so, I want to suggest, simply because Trolley and Transplant are intuitively so different from each other. And of course, we knew that all along. The objections to the moral relevance of the intending-foreseeing distinction do not deny, of course, the commonsensical appeal of the distinction. Rather, they are arguments for why we may need to revise the judgments that seem to us so intuitive, because they do not seem to us defensible on reflection. Here as elsewhere, what we want from a moral theory is not just a reasonably good fit with our pretheoretical intuitive judgments. Rather, we want the theory to be good as a theory, to score reasonably high on the list of theoretical virtues. In order to do that it must, for instance, offer something by way of unity, coherence, explanatory power, and elegance, it must not be ad hoc, it must focus on factors that upon reflection seem to make a moral difference, and so on. What the discussion above shows is that a moral theory that makes use of the intending-foreseeing distinction (understood causally) fails this further test, and this even if it does fit many of our pretheoretical specific moral judgments. Just noting again that these judgments do seem intuitive carries at this stage of the debate little weight.
C. Remnants of Catholicism?
The history of the moral distinction between intending and foreseeing is bound with the traditional Catholic doctrine of double effect.Footnote 19 According to this doctrine—and this is very rough—one is allowed to bring about both a good effect and a bad one only if the good is greater than the bad and, furthermore, the bad is not intended as a means to the good but is merely foreseen as a side effect of bringing about the greater good.Footnote 20
This historical connection between the moral distinction between intending and foreseeing and the doctrine of double effect lends some support to the following suspicion: perhaps the moral distinction between intending and foreseeing is well motivated in the context of the beliefs and judgments that are a part of the Catholic (or perhaps more generally theistic) tradition but not in other contexts. And perhaps our intuitive judgments on particular cases that seem to assume something like that distinction are really remnants from a system of thought many of us no longer adhere to nor should we adhere to. It may be argued, for instance, that within a (certain kind of) religious context. the distinction can derive its appeal from a view of the good will as holy or pure (or perhaps the bad will as especially impure). But if we no longer think of the good will as pure or holy in any such way, perhaps the philosophical rationale for assigning moral weight to the intending-foreseeing distinction no longer exists. Similarly, perhaps within a religious context it may be argued that we mere mortals are responsible only (or especially) for our own intentions and everything else—side effects included—is up to God. But if we no longer believe in the kind of Providence this suggestion presupposes, perhaps we no longer have any reason to attribute moral weight to the intending-foreseeing distinction.
Of course, such historical speculations are never conclusive. Perhaps, for instance, from within a Catholic perspective the intending-foreseeing distinction is supported by one rationale, a rationale we should perhaps reject, but still the distinction can be supported by some other, nontheistic, philosophical rationale. Or perhaps the traditional Catholic rationale for the doctrine is one also available for nonbelievers.Footnote 21 Nevertheless, the kind of historical suspicion I gesture at here is not, I think, without force. For by suggesting a debunking explanation of our relevant intuitions—one that is perfectly compatible with them being utterly misleading—one makes them more suspicious and the option of discarding them more viable.
D. The Irrelevance of the Agent's Mental State
Think about a hard medical decision—say, whether to give a suffering patient a high dose of morphine in order to relieve his pain (at the price of his likely death). And let us assume that in the circumstances, the (medically and also morally) right thing to do is to give the morphine. Now add the following piece of information: the physician making the decision and administering the procedure enjoys perverted pleasures from killing patients. If he gives the patient the morphine, he will do it intending to enjoy these perverted pleasures. He foresees that the patient's pain will be relieved, but this is not why he acts as he does. Of course, now that we know these disturbing facts about the doctor and his relevant mental states, we will morally judge him accordingly and will no doubt try to let someone else decide about the appropriate procedures. But—and this is the crucial point in our context—should this information make us change our mind regarding the permissibility of the relevant action? Could facts about these mental states of the doctor giving the morphine make us take back our judgments that this is the appropriate action in the circumstances, even when all other factors are held equal? The answer, it seems, is “no.”
Thomson suggests that we learn from such examples that the agent's mental states are simply irrelevant for the moral permissibility of the relevant action. They are very relevant, of course, for the evaluation of the agent, but this is an entirely different story. And because mental states are irrelevant to the moral status of the action, the intending-foreseeing distinction, understood as a distinction between two mental states and applied to the moral evaluation of actions, is without moral weight.Footnote 22 Of course, as it stands this line of thought is too quick. Strictly speaking, what the example at most shows is that sometimes the agent's mental states are irrelevant to the permissibility of the relevant action, not that they never are. But the strength of the intuitive judgment Thomson uses, together with the distinction between the evaluation of the action and that of the agent, and given the absence of an obvious rationale for why it is that the mental states should be relevant to permissibility in some circumstances but not others—all these factors together strongly suggest, I think, the more general conclusion.
Perhaps we should qualify this assertion in the following way: it may perhaps be thought that when all other things are equal, a state of affairs in which there are good-intending mental states is better than a state of affairs in which there are bad-intending mental states. Perhaps, in other words, something like the intending-foreseeing distinction can be incorporated into the conception of the good (though I do not know if this would be a plausible conception of the good). If so, then there is, after all, a difference between the state in which a good-intending doctor administers the morphine and one in which a bad-intending doctor administers it. And then it may no longer be true to say that the mental state of the action has absolutely no (intrinsic) bearing on the permissibility of the action.Footnote 23
Even if this qualification is not without weight, though, it is hard to believe that it has enough weight to outweigh some of the other relevant considerations—like that of the need to relieve the patient's pain or to avoid the harm to innocents in cases of counterterrorist acts. Furthermore, nothing like this line of thought can secure for the intending-foreseeing distinction anything like the deontological status it is often supposed to have. For if the moral significance of the distinction is incorporated into one's conception of the good, then in circumstances in which intending (and not merely foreseeing) harm will minimize future harm-intendings, we should go ahead and intend harm (so that, for instance, we should kill the patient in Transplant if doing so will somehow minimize the number of doctors killings patients in Transplant-like cases). And this, of course, is not a result many deontologists will be happy with.
The believers in the normative weight of something like the intending-foreseeing distinction are not completely without reply, then. For one thing, they can simply deny the substantive intuitions Thomson (and others) rely on. Or they can suggest better interpretations of the intending-foreseeing distinction, interpretations that perhaps deal better with this objection (more on this shortly). For now, though, let me just note again the strong suspicion that the intending-foreseeing distinction—at least if understood as a distinction between mental states—does not carry intrinsic normative weight, at least not with regard to the moral permissibility of actions.
E. Evading Responsibility
Think again of the appropriate-question test. What is the moral flaw exhibited by the person asking—while deliberating—about the causal order of the relevant consequences?
Intuitively, the flaw exhibited seems to me to be a kind of an attempt to evade responsibility. Attempting to find out about the causal order of the (foreseen) consequences looks here not so much as an honest attempt to unearth the morally relevant features of the circumstances but as an attempt to prepare a line of defense.Footnote 24 Once you know about the consequences of your actions (or inactions), an attempt to pick and choose among them those you are more and those you are less responsible for (those you intend and those you merely foresee, respectively) looks like an attempt to evade (full) responsibility for the merely foreseen consequences. A responsible agent, it can be argued, accepts responsibility for all (foreseen) consequences of her actions, both intended and unintended. This suspicion—that hiding behind the intending-foreseeing distinction is really just evading responsibility—is arguably at least a part of the rationale for the entrenched doctrine of the criminal law, according to which under certain circumstances foresight can substitute for intention.Footnote 25
The point here is not an empirical one, namely that people invoking the intending-foreseeing distinction are, as a matter of sociological fact, merely attempting to evade responsibility and are being disingenuous in invoking the distinction. My point here, rather, is that thinking about such cases in terms of responsibility (and not just in terms of permissibility) helps to emphasize a feature of the moral situation that is perhaps not as clear otherwise. Thinking in terms of responsibility, it just seems overwhelmingly plausible that agents are responsible for all the foreseen consequences of their actions. Furthermore, it seems just as plausible that they are equally responsible for all the foreseen consequences of their actions. But, of course, it is very hard (though perhaps not impossible) to reconcile such intuitions about responsibility with the differential treatment of intended and merely foreseen consequences when it comes to moral permissibility.
I think that this intuition—that a responsible person is (equally) responsible for all the foreseen consequences of his or her actions—accounts for much of the appeal of consequentialist positions. I return to this point below.
F. Interim Conclusion
These, then, are some difficulties facing the attempt to treat the intending-foreseeing distinction as intrinsically morally important. The critique discussed in Section III.B—about the moral irrelevance of the causal order—attacks the distinction understood as one between two kinds of causal structures; the critique discussed in Section III.D—about the irrelevance, for the moral status of an action, of the relevant agents' mental states—attacks the distinction understood as one between two kinds of mental states. And those in sections III.A, III.C and III.E—discussing Loop-like counterexamples, the suspicion that the intuitions underlying the distinction are remnants of theistic doctrines we no longer hold dear to our hearts, and the fact that according the intending-foreseeing distinction any moral weight is even more implausible once the situation is thought of in terms of responsibility—pose, I think, problems for the intending-foreseeing distinction however exactly it is understood. Given these objections and their (combined) force, how are we to proceed?
One possible way to proceed would be to present better—more precise and perhaps also more complicated—versions of the distinction or of close distinctions, versions that can cope with the difficulties mentioned. If, for instance, we understand the distinction as one between different kinds of mental states, then we should be sensitive to the fact that the motivational story is more complicated than the simplistic intending-foreseeing distinction seems to suggest.
What should we say, for instance, about a counterterrorist targeted killer whose end is, say, to ensure Israeli control of the occupied territories, who intends to do this by using a targeted killing in order to send a demoralizing message to the Palestinian population, but who never launches such an attack unless he is also reasonably confident that launching the attack will prevent casualties from terrorist acts? For such a counterterrorist targeted killer, the good effect of his action (saving lives of potential victims of terrorist attacks) is neither his end nor his means, but nor is it a mere motivationally inert side effect, for it does play some role in bringing him to action.Footnote 26
In order to deal with such—not uncommon—cases, we are going to have to develop conceptual resources and distinctions more subtle than the mere distinction between intending and foreseeing, and perhaps with those resources available to us we are going to be better placed with regard to some of the difficulties the traditional distinction is up against.
Similarly, if we understand the intending-foreseeing distinction as primarily involving the distinction between two kinds of causal structures, then it is again clear that there are many, many possible causal structures, and perhaps more intricate distinctions among them can help us better cope with some of the difficulties the original distinction faces.Footnote 27
But some of the discussed objections seem to be fairly general. Perhaps, for instance, greater ingenuity in the exact formulation of the sought-after distinction will nicely deal with Loop and other counterexamples. But it is hard to see how such ingenuity could help with the more general objections from the (intrinsic) irrelevance of mental states, from the (intrinsic) irrelevance of the causal structure (as revealed by employing the appropriate-question test), and from the suspicion that there is an evasion of responsibility going on. These problems, it seems to me, are pretty much here to stay.
As already noted, if the intending-foreseeing distinction collapses under the weight of these (or other) objections, so do significant portions of commonsense morality. Perhaps at the end of the day there will be no alternative but to proceed in that direction. But the depth of the moral revolution this would call for inclines me not to be too willing to give up the intending-foreseeing distinction, even if at this point I do not know how to defend it against the mentioned critiques.Footnote 28
IV. THE SPECIAL CASE OF STATE ACTION
Assume for now that the objections in the previous section give us reason perhaps not to reject the moral significance of the intending-foreseeing distinction but at least to be suspicious about it. I now want to revisit two of these objections and to argue that we have even stronger reason to be suspicious of the moral significance of the intending-foreseeing distinction when applied to state action.Footnote 29
A. The Irrelevance of the Agent's Mental State
I have already given some reason to suspect that the agent's mental states are irrelevant as far as the moral permissibility of the action is concerned. But whatever you think about the strength of this reason, surely there can be no intrinsic moral significance in determining the permissibility of an action to the mental (or “mental”) states—such as intentions—of artificial agents such as corporations and states.Footnote 30
The problem is not that states and statelike agents do not have mental states. In many circumstances and for many purposes we are happy ascribing mental states to corporations, institutions, and indeed states. Perhaps this is all a great big mistake on our part. Or perhaps such mental-state ascriptions should be understood as mere metaphors. Or perhaps there is a perfectly acceptable, philosophically respectable sense in which, quite literally, states have mental states. I want to sidestep these issues here, though, both because discussing them would carry me too far into the philosophy of mindFootnote 31 and because deciding them is not necessary for the point I want to make here.
To show this, let me concede for the sake of argument that sentences ascribing mental states to states, corporations, and the like can be quite literally true. (At least, let me concede for the sake of argument that this is so for sentences ascribing intentions, which seem to be at least in part mental states. We need not ask whether states can literally suffer pains, see red, or whether, perhaps more generally, there is something it is like to be, say, France.) We are assuming, then, that states have intentions, a will, and the like. But now think about the nature of such mental states: they are determined by the intentions of individuals, by facts about decision-making mechanisms, by matters of institutional design, by internal power struggles. They are, in other words, highly complex and, in a sense, also artificial things. And though their intentions may nevertheless—as we are now assuming—merit being classified as bona fide intentions, still they seem very poor candidates for intrinsic normative significance. Perhaps, in other words, there is some intuitive plausibility to the thought that in the case of individuals there is something particularly good about the good will, that regardless of other factors, “like a jewel, it would still shine by itself, as something that has its full worth in itself.”Footnote 32 But this intuitive plausibility does not extend to the will of states and similar agents.
To repeat, my point here is not metaphysical—I do not want to present necessary conditions for having mental states in general, or intentions in particular, and then check to see whether states and perhaps other political institutions satisfy these conditions.Footnote 33 Nor is my point here epistemological—I am perfectly happy to assume that at times we know all there is to know about the mental states of political institutions. My point, rather, is normative: even if states and other political institutions have mental states, these are not plausible candidates for intrinsic moral significance. And if so, the difficulty for the intending-foreseeing distinction discussed in Section III.D above becomes much more pressing when dealing with state action. Even if the relevant agent's mental states are after all not irrelevant to the moral status of actions performed by individuals, then, it is exceedingly hard to believe that they are morally relevant when the action is performed by something like a state.
It may be tempting to try to solve the problem by taking the state out of the picture altogether and focusing instead on the mental states of the (real, natural, individual) decision-makers. Their mental states are, after all, the mental states of individuals and so are purportedly better candidates for normative significance. But this temptation should be resisted for the following two reasons.
First, even if we focus on the intentions and other mental states of the decision-maker, these are mental states she has in her capacity as a state official. And this already seems to weaken their status as plausible candidates for normative significance (the point in the next section—about evading responsibility—clearly applies).
Second, and more important, even if sometimes we can fairly clearly recognize the intentions and other mental states of real individuals underlying state action, often this is not the case. Think, for instance, about a general decision about counterterrorist policy made in the relevant political and military forums. These are often multimember forums (such as a cabinet), there is a strong feature of division of labor (one person or group of persons is responsible for getting the intelligence, another for analyzing it, yet another for making the decision, yet another for carrying out the mission), the institutional setup has significant weight in determining the output of the decision-making process, and so on. Whose intentions are to be attributed to the state once a decision has been reached?
As is often noted in the literature on legislative intent,Footnote 34 in multimember decision-making organs there is often no one whose intentions we can plausibly attribute to the organ as a whole. Members of such bodies vote on principle, on interests, according to strategic alliances, compromises, and so on. The institutional setup may have implications that are—in the specific case—quite arbitrary (say, what kind of majority is needed for what decision). And so on. Now, even in such circumstances—not uncommon circumstances, it seems to me, when we are talking about state action—we may be willing to attribute intentions as underlying the relevant state action. For now, then, I am assuming that even in such cases it will be true that the cabinet (and so, presumably, the state) intends one thing and not another. But it is very hard to believe that such mental states—the artificial functions of all these very different input intentions and other factors—are of intrinsic moral significance.
If you are not yet convinced of this point, think about a suggestion mentioned in Section III.D above. There, it will be remembered, I mentioned that one may want to incorporate a proxy for the intending-foreseeing distinction into one's conception of the good, so that a state of affairs that contains harm-intending may be at least somewhat worse than one that contains merely harm-foreseeing (with all other things held equal). I did not comment there on the plausibility of a conception of the good of this kind and I have no settled view on the matter. But let me note the following: even if such a conception of the good has some plausibility when the intention and foresight of individuals is concerned, it certainly does not when we are talking about a state.
Imagine two states of affairs, then, with all else equal, except that in one a state intends harm—in the compound sense in which states have (so we are assuming) intentions as a function of institutional structures, the mental states of individuals, and so on—and in the other, the state merely foresees it. Can it be plausibly maintained that the former state of affairs is worse than the latter (assuming, again, no other differences between them)? It seems to me that the answer is rather clearly negative. And this shows two things: First, that the (very partial) defense maneuver mentioned in Section III.D has no force when the relevant mental state is that of a state. And second, and much more important, that this is so serves to highlight the implausibility of any view that attaches intrinsic moral significance to the mental states of states (and the like).
This conclusion should be qualified: the extent to which what I have just said is true of state actions (and indeed of the actions of corporations and the like) differs from action to action and from decision to decision (and perhaps also from state to state). Some such decisions are rather clearly the decisions of one person (perhaps a “strong leader” of some sort). In such cases it seems that for many purposes we can attribute this person's relevant intentions to the state. (Though even then it is not clear that this is so—imagine a president who is a strong leader of this sort and who makes her decisions regarding counterterrorist acts on her own. And suppose that her intention in doing so is to gain popular support. Are we willing to attribute this intention to the state?) In other cases, where the decision is made in a manner more resembling the one sketched above, the point from the previous paragraphs applies.
My claim here, then, is not a dichotomous one. All I argue is that the more a decision-making forum resembles the picture sketched above, the less likely it is that the forum's “mental” states have intrinsic moral significance.Footnote 35 Derivatively, the further state decision-making is from the model of an individual decision-maker, the less plausible it is to apply to it the intending-foreseeing distinction as a distinction with any moral weight.
Let me address here three objections. First, it seems rather clear that the difference between foreseen and unforeseen consequences is morally significant—indeed, relevant to the question of moral permissibility—even in the case of state action. But such a difference is—at least partly—a difference in mental states. So it may serve as a counterexample to my claim that mental states of states lack intrinsic moral significance (unless, that is, I am willing to bite the bullet and deny the intrinsic moral significance of the foreseen-unforeseen distinction in the case of state action, which I am not willing to do).Footnote 36
In reply let me note the following. When we keep the nature of state intention and state foresight in mind, the distinction between them—so I have argued—seems a very unlikely candidate for intrinsic moral significance. But even when we keep in mind the nature of state foresight, still the foreseen-unforeseen distinction loses nothing from its intuitive weight. This suggests an important moral difference between the two distinctions. The moral significance of the intending-foreseeing distinction requires more than the mere functional role of intention and foresight. Not so for the foreseen-unforeseen distinction, where the mere functional role (whether a certain piece of information is somewhere included as input in the black-box-like procedure of state decision-making) suffices for its moral significance. Perhaps, then, the moral significance of the foreseen-unforeseen distinction is indeed a counterexample to the most general claim that states' mental states are intrinsically morally irrelevant. But it does not suffice to undermine the claim that the intention-foresight distinction lacks such moral weight. The anthropomorphist illusion—that the state is just a (giant) person like the rest of us—is needed in order to save some plausibility for the idea that the intention-foresight distinction matters with regard to state action, but it is not needed in order to sustain the plausibility of the idea that the foreseen-unforeseen distinction has such moral weight. When we are clear, then, about this anthropomorphist illusion being precisely an illusion, the former distinction collapses but the latter can survive.Footnote 37
The second objection I have in mind starts with the thought that there is nothing mysterious about the state and its decision-making mechanisms. The state, it may be thought, is pretty much just us, and its decisions represent us and our decisions. So it cannot be—this thought concludes—that the intentions of states are more suspicious candidates for moral significance than those of individuals like us.Footnote 38
Now, there are, of course, senses in which it is true to say that the state or the government is “just us,” and—at least in some states—there may be a sense in which the state or the government represents its individual citizens. But my point above was precisely that there is nothing self-evident about these relevant senses. If the state really is—in some sense—just us, then it is us in a very indirect, complicated, and possibly surprising way. And it goes without saying that the notion of representation is both normatively loaded and in need of serious unpacking. The sense, then, in which the state is just us, or at least represents us, does nothing to guarantee that the factors that are (presumably) morally relevant in our case are also relevant in the state's.
Finally, it may be thought that a certain way of thinking about state action that is currently fashionable—the way developed in expressive theories of law and of state actionFootnote 39—yields a nice rationale for the moral weight of the intending-foreseeing distinction in the case of state action, because, perhaps somewhat roughly, intending harm to someone may express a failure to appreciate the moral reasons that apply to one in a way that merely foreseeing similar harm does not.Footnote 40
Of course, I cannot discuss expressive theories of law and state action in detail here.Footnote 41 Let me settle, then, for the following points. First, expressivists have to develop an understanding of “expression” that will be broad enough to make the expressive view plausible and narrow enough to make it nonredundant. It is not clear that this can be done,Footnote 42 and if it cannot, the suggested rationale for the intending-foreseeing distinction collapses. Second, it is not at all clear why intending harm can express the wrong attitude when—in the very same circumstances—foreseeing it does not. In answering this question, it seems to me, expressivists are going to have to deal with all the issues discussed in this paper. Again, then, what we have here looks more like a name for the problem than its solution. Third, depending on the precise details of the relevant notion of expression, the suggested rationale for the intending-foreseeing distinction may show that it is extrinsically (though not necessarily instrumentally) morally significant, but what I am questioning in this paper is its intrinsic moral significance. Finally, and relatedly, some of what expressivists have to say here may be nicely dealt with not by according the intending-foreseeing distinction intrinsic moral significance but rather by the kind of consequentialist proxies for this distinction I proceed to discuss below.Footnote 43
B. Evading Responsibility
The general difficulty of the intending-foreseeing distinction here stemmed, you will recall, from the feeling that attempting to pick and choose among the foreseen consequences of one's actions those one is more and those one is less responsible for looks more like the preparation of a defense than like a genuine attempt to determine what is to be done. Hiding behind the intending-foreseeing distinction seems like an attempt to evade responsibility, and so thinking about the distinction in terms of responsibility serves to reduce even further the plausibility of attributing to it intrinsic moral significance.
This consideration—however weighty in general—seems to me very weighty when applied to state action and to the decisions of state officials. For perhaps it may be argued that individuals are not required to undertake a global perspective, one that equally takes into account all foreseen consequences of their actions. Perhaps, in other words, individuals are entitled to (roughly) settle for having a good will, and beyond that let chips fall where they may. But this is precisely what stateswomen and statesmen—and certainly states—are not entitled to settle for.Footnote 44 In making policy decisions, it is precisely the global (or at least statewide, or nationwide, or something of this sort) perspective that must be undertaken. Perhaps, for instance, an individual doctor is entitled to give her patient a scarce drug without thinking about tomorrow's patients (I say “perhaps” because I am genuinely not sure about this), but surely when a state committee tries to formulate rules for the allocation of scarce medical drugs and treatments, it cannot hide behind the intending-foreseeing distinction, arguing that if it allowsFootnote 45 the doctor to give the drug to today's patient, the death of tomorrow's patient is merely foreseen and not intended. When making a policy-decision, this is clearly unacceptable.
Or think about it this way (I follow Daryl Levinson here):Footnote 46 perhaps restrictions on the responsibility of individuals are justified because individuals are autonomous, because much of the value in their lives comes from personal pursuits and relationships that are possible only if their responsibility for what goes on in the (more impersonal) world is restricted. But none of this is true of states and governments. They have no special relationships and pursuits, no personal interests, no autonomous lives to lead in anything like the sense in which these ideas are plausible when applied to individuals persons. So there is no reason to restrict the responsibility of states in anything like the way the responsibility of individuals is arguably restricted.Footnote 47
States and state officials have much more comprehensive responsibilities than individuals do. Hiding behind the intending-foreseeing distinction thus more clearly constitutes an evasion of responsibility in the case of the former. So the evading-responsibility worry has much more force against the intending-foreseeing distinction when applied to state action than elsewhere.
C. A Speculative Historical InterludeFootnote 48
Discussions of the intending-foreseeing distinction are often conducted in the context of a critical evaluation of utilitarian or more broadly consequentialist moral theories. Such theories—at least in their more simple, direct versions—cannot accord intrinsic moral significance to the distinction between intending and foreseeing. And this is taken either as a reason to reject the moral significance of the distinction or (perhaps more commonly) as a reason to reject consequentialism.
I have already emphasized how central something like the intending-foreseeing distinction seems to be to our commonsensical moral thinking. Has this fact escaped Bentham and Mill, whose role in the development of modern utilitarianism cannot be overestimated? Perhaps at the end of the day the intending-foreseeing distinction has to be rejected, perhaps because inconsistent with sound consequentialist thought. But surely this would be at the very least an intuitive price, and consequentialist thinkers should at the very least address this difficulty. But Bentham and Mill never discuss—as far as I know—the intending-foreseeing distinction and the challenge it poses for their utilitarianism, and this even though they must have been aware of the distinction and its intuitive force (for instance via the Catholic doctrine of double effect).Footnote 49 What explains this silence?
Bentham and, to an extent, also Mill were, as I understand them, primarily interested in developing a theory of political morality,Footnote 50 the kind of moral theory that could guide legislators and other public decision-makers. And I have just been arguing that in such a context the appeal of the intending-foreseeing distinction is greatly diminished compared to other, more private contexts. If this is right, perhaps this can partly explain why Bentham and Mill—primarily interested as they are in political morality—are so indifferent to the intending-foreseeing distinction. And if the hypothesis that the intending-foreseeing distinction is much less plausible when applied to state action can explain this fact, perhaps this in turn lends some support to that hypothesis itself.
V. CONCLUSION
I have already noted that given the centrality of the intending-foreseeing distinction (and related distinctions) to our commonsensical moral thinking, I would hesitate to discard it, even if at the moment we are not sure how to cope with some of the challenges it faces. A similar point can be made—though perhaps in a weaker, more qualified way—with regard to the relevance of this distinction to state action. True, some of the general challenges to the moral significance of the intending-foreseeing distinction are especially powerful in the context of evaluating state action. In that case too, though, it is unclear how—and even whether—it is possible to defend some of our firmly held pretheoretical judgments without relying on something like the intending-foreseeing distinction.Footnote 51 Here, too, then, we must not be too hasty in discarding this distinction. All I argue is that we should be very suspicious of it, indeed more so in the case of state action than in general.Footnote 52
Consider again the comparative discussion of typical terrorist and counterterrorist actions. If the intending-foreseeing distinction has no (or almost no) moral significance when applied to state action, and if related distinctions likewise fail, all that seem to remain as determining the moral status of state action are consequentialist considerations (for an argument for this claim, see the Appendix). If so, the state is morally justified in launching a counterterrorist attack likely to cause significant collateral damage whenever consequentialist considerations recommend doing so. The intending-foreseeing distinction is supposed to be a part of a deontological framework that includes a constraint against maximizing the good, so that some actions (like killing the one in Transplant) are out of moral bounds even if they do maximize the good. Take away the intending-foreseeing distinction, then, and the moral realm is left entirely open to good-maximizing considerations.
This point can be made in terms of responsibility as well: without the intending-foreseeing distinction, just as the state cannot evade responsibility for, say, the harm to the innocents surrounding the target of a targeted killing, so, too, it cannot evade responsibility for the harm to innocent potential victims of terrorist acts, should the targeted-killing mission be called off and that potential target be allowed to carry on with his terrorist plans.
And so a surprising asymmetry emerges:Footnote 53 if we should be more suspicious of the intending-foreseeing distinction when applied to state action than when applied to the acts of an individual terrorist (or even a terrorist organization, as long as it is not sufficiently statelike in the relevant respects), then perhaps in situations that are exactly alike in overall (morally relevant) consequences, still a terrorist act would be wrong and a counterterrorist act permissible not because the terrorist intends and the counterterrorist merely foresees harm to innocents, but rather because the terrorist intends harm to innocents (as one hardly ever should), whereas the intending-foreseeing distinction does not apply to the state, which may therefore be justified in its action on purely consequentialist grounds.
This asymmetry is surprising already as it stands. Some of its implications are even more surprising, though. For the asymmetry seems to imply that it is possible that a typical terrorist act could be more easily justified when performed by a state than when performed by an individual (in the former case all that has to be shown is that the good will be maximized if the terrorist act is performed, whereas in the latter it will presumably have to be shown that the good to be achieved is beyond a high threshold of a deontological constraint against intending harm). Indeed, this may be thought of not so much as a surprising result but rather as a reductio.
It is thus important to note that perhaps these disturbing implications can be resisted. For one thing, and as already noted, some state decisions resemble decisions of individuals sufficiently closely for much of the discussion not to apply. And even when it does apply, so do further considerations. For instance, a fuller story would have to take into account not only the decision-making forum but also the moral status of the individual persons asked to carry out the decision and of their actions. And if what the state calls upon them to do is morally wrong, this may have further implications for the moral status of the original decision. So it is not as if state action is completely divorced from the actions of individuals and the interconnections between them are morally significant.Footnote 54
Second, some other moral considerations that apply more to state action than to other actions may counterbalance the surprising asymmetry. Chief among those, perhaps, are second-order institutional considerations. Perhaps, for instance, many of the actions that may be ruled morally impermissible when performed by individuals because they involve intending harm will be ruled morally impermissible in the case of state action because allowing the state to act in those ways will require granting officials powers they are likely to abuse, so that on the whole it is better (in consequentialist terms) not to let the state act in those ways. Perhaps, in other words, the asymmetry between the moral standards applicable to states and those applicable to individuals is consistent with fairly wide overlap in the moral status of similar actions when performed by the two different kinds of agents (though for different reasons). I return to this point below.
And, of course, another possibility remains. Perhaps we should acknowledge the fact that the intending-foreseeing distinction is without moral weight in the case of state action. And then perhaps we should take the implausibility of the asymmetry between state and individual action as further reason to reject the distinction's moral significance in the case of individuals as well.
VI. WHY THE APPLICABILITY OF THE INTENDING-FORESEEING DISTINCTION TO STATE ACTION IS OF MORE GENERAL INTEREST
The intending-foreseeing distinction is everywhere, so studying it and its implications is of quite general interest. Surprisingly, though, and despite a growing literature on the intending-foreseeing distinction in general, no sufficient attention has been given to the problems facing this distinction in the specific case of state action.Footnote 55
Think, for instance, about the distinction often made in First Amendment doctrine between content-based and non-content-based restrictions. If the state restricts my exercise of my right to free speech because of considerations having to do with the content of my speech, its action will be more strictly scrutinized by the courts compared to the case of a restriction for other considerations (such as the cleanliness of a terminal being the reason not to allow the handing out of flyers in the terminal).Footnote 56 But if there is no normatively significant distinction between the state intending and foreseeing a relevant consequence, then there seems to be no normatively relevant difference between the state intending to restrict a specific speech and the state intending to maintain cleanliness but foreseeing the restriction on speech.Footnote 57
Or think about the liberal doctrine of state neutrality, the thought that government should be neutral on the question of the good life.Footnote 58 The literature often distinguishes between a requirement of neutrality in reasons (so that the state is not allowed to act for a nonneutral reason) and a requirement of neutrality of effect (so that the state is not allowed to act in a way that will have nonneutral effects). Now, as is often noted, a neutrality-of-effect requirement is much too stringent, indeed almost impossible to comply with.Footnote 59 So the neutrality requirement still on the table is a requirement of neutrality in reasons. But the very distinction between a neutrality-in-reasons requirement and a neutrality-of-effects requirement seems to rely on a distinction between intended and unintended (yet often foreseen) consequences.Footnote 60
In a very different context—the theory of criminal law in general and criminal punishment in particular—there is sometimes a need, at least for retributivists, to distinguish between the intended punishment and the foreseen but unintended consequences of the punishment, so that retributivism requires equality (given the severity of the offense) in the former but not in the latter.Footnote 61 But, of course, this distinction seems to rely on an intending-foreseeing distinction and its application to the state.
On the other hand, most discussions within the economic analysis of law seem to assume that nothing like the intending-foreseeing distinction has any normative weight. For economic analysis of law presupposes—in line with the utilitarian tradition of which it is very much a part—that there is no morally relevant difference between what the law does and what the law allows to happen and between the intended and unintended (yet foreseen) consequences of a relevant legal arrangement.Footnote 62 All that matters for the desirability of a legal arrangement, economic analysts of law typically assume, is how the world will (or can be expected to) be given the relevant legal arrangement and be given alternative legal arrangements.
This is all very sketchy, of course, and much more needs to be said about each of these examples before anything like a comprehensive examination of the intending-foreseeing distinction and its implications in the political and legal contexts can be achieved. And many more examples can be brought here as well.Footnote 63 But enough has been said, I think, to show how ubiquitous the implications of such a discussion could be.
Earlier I said that even if the intending-foreseeing distinction has no moral weight when applied to state action, perhaps there are other considerations, considerations that apply uniquely or most forcefully to state action—such as second-order institutional considerations—that could serve as proxies to the intending-foreseeing distinction, allowing us to get results rather similar to the ones we would have gotten had we applied the intending-foreseeing distinction to the state. This point may give rise to the suspicion that none of this is of any practical bearing—practically speaking, it may be thought, what matters is what the state is allowed and not allowed to do, not what the reasons are supporting these judgments. But this is not so. True, even if there is no moral weight to the intending-foreseeing distinction when applied to state action, there may be other reasons to distinguish, for instance, between content-based and non-content-based restrictions of speech (perhaps, for instance, the authority to issue content-based restrictions is more easily abused).Footnote 64 But it may be of crucial importance—even practically—to be clear about the rationales for this doctrinal distinction. For the rationales may help us determine the scope of the doctrine and its best interpretation.
Even if, then, consequentialist-friendly considerations can approximate the intending-foreseeing distinction in the political sphere, it may be important for theoretical as well as practical purposes to appreciate their true nature. For this reason, too, then, it is important to pay more theoretical attention to the intending-foreseeing distinction (and related distinctions) in the special case of state action.
APPENDIX: WHY WITHOUT THE INTENDING-FORESEEING DISTINCTION, CONSEQUENTIALISM IS VERY HARD TO RESISTFootnote 65
According to consequentialists, the morally right action is, roughly, the one that is expected to promote the good. All it takes, then, in order to reject consequentialism is to believe in some deontological constraints, that is, in some rules to the effect that there are some actions it is morally impermissible to perform (at least sometimes) even when performing them will maximize the good.Footnote 66 And such constraints need not be put explicitly in terms of intention and foresight. So it is not immediately clear why a nonconsequentialist needs anything like the intending-foreseeing distinction, why without it consequentialism is the way to go.
In order to see that this is nevertheless so, think of an example of a deontological constraint, say a constraint against lying. Now think of the well-known type of case in which the way to minimize constraint violations is to violate that very constraint. So we are to imagine a situation in which, for some reason, if I lie now, fewer lies will be told (this one included) than if I do not lie now. If we want to justify the judgment that at least sometimes a constraint is not to be violated even in order to minimize constraint violations, then we had better distinguish morally between the relation between me and my action and the lying (right now), on the one hand, and the relation between me and my action and future lyings, on the other. Otherwise, if no such moral distinction is to be made, then by not lying now, thereby causing (or allowing) many more future lyings, I am responsible for all those future lyings and so I am acting wrongly. And, of course, in order to distinguish between the relation between me and my lying right now and me and future lyings, the intending-foreseeing distinction, or perhaps some closely related distinction, is needed.
Should we, then, say that at least sometimes we should not violate a constraint even in order to minimize constraint-violations? Well, I do not think that deontologists, as a matter of logical necessity, have to say this. Depending on the details of one's conception of the good, it is entirely possible that there are actions we should at least sometimes not perform even when performing them will maximize the good, and that nevertheless we should also violate a constraint whenever doing so will minimize constraint violations. But—as I think all deontologists agree—such a way out, though perhaps logically sound, is morally unsound. It seems inconsistent with if not deontology itself, at least the philosophical and moral motivations underlying deontology. And if so, absent the distinction between intending and foreseeing, or some other distinction that could distinguish between my relation to my lying now and my relation to (my or others') future lyings, consequentialism is hard to resist.Footnote 67
Note that this little argument is perfectly general and can thus be applied even to deontological theories that are not explicitly formulated in terms of anything like the intending-foreseeing distinction (as can be seen from the example of a constraint against lying). As long as it is at least sometimes morally impermissible to treat someone merely as a means even when doing so will minimize the number of (similarly severe) cases of people treating others merely as means, Kant needs something like the intending-foreseeing distinction. As long as it is at least sometimes morally impermissible to treat someone in a way that violates principles that nobody could reasonably reject even when doing so will minimize the number of times in which people are so treated, Scanlon needs something like the intending-foreseeing distinction.Footnote 68 And so on.
Whatever your favorite deontological theory, then, as long as you believe that there are deontological constraints that at least sometimes should not be violated even in order to minimize violations, you need something like the intending-foreseeing distinction. And so, if that distinction collapses, and if no similar distinction can be defended, it is very hard not to be a consequentialist.
