I. INTRODUCTION
What makes careless conduct careless is easily one of the deepest and most contested questions in negligence law, tort theory, and moral theory. Answering it involves determining the conditions that make the imposition of risk unjustifiable, wrong, or impermissible. Yet there is a still deeper question. That risk can be impermissibly imposed upon others—that is, the very possibility of negligence—presupposes that imposing risk is the kind of thing that can be impermissible. It presupposes that the concept of a risk imposition can bear this crucial moral freight. This is obviously a basic and important presupposition of the law as well as of what might be called the morality of risking. Unless imposing risk can be impermissible, unjustified risking—which is to say negligence—is literally impossible. Yet what I call the moral significance of risking is usually overlooked and certainly undertheorized. Why does subjecting others to risk of harm call for justification in the first place?
There are certain grounds on which risk impositions seem quite obviously to call for justification. Subjecting another person to risk that ripens into the risked harm, for example, calls for justification just as any harmful conduct does. A risk imposition that induces fear in whoever is (or believes themselves to be) in its orbit also calls for justification, for instilling fear in others must always be justified. And risky conduct that prompts others to take disruptive steps to avoid the perceived risk calls for justification, too, because one must always account for disrupting the life of another. Risk impositions thus seem to rate morally and call for justification at least in part because they harm, induce fear in, or disrupt the lives of others.
As this much should make plain, however, it may be misleading to ascribe this moral significance to risk impositions as such. For the above are probably better described as cases of harming, inducing fear, and disrupting life, not as imposing risk. The moral significance that the above conduct bears is only indirectly related to the fact that the conduct (also) imposes risk. In any case, the central question remains: What of risk impositions that do not result in the harm that they risk, induce fear in others, nor disrupt anyone's life? Do they bear moral significance?
These are important normative questions for at least two related reasons. First, the class of such risks is massive: people in modern and modernizing societies live with and indeed impose upon each other a countless diversity of risks that usually remain inchoate and underappreciated, thus causing no resultant harm, fear, or disruption. Do these pure risks possess moral significance?Footnote 1 If so, then even they call for justification. If that is the case, moreover, then the imposition of risk as such calls for justification, which entails that all risk impositions, pure or not, can be wrong just in virtue of being risky, quite apart from any deleterious material effects that they may have. It would follow that there is a moral dimension to the imposition of any risk that cannot be reduced to any of the more familiar considerations noted above.
II. THE PUZZLE
If imposing pure risk is morally significant it means that the imposition of risk as such is a proper object of moral assessment, calling for justification such that any risk imposition is a candidate wrong and potentially impermissible.Footnote 2 It does not follow from this that the justification called for cannot be provided, for surely some risk impositions are permissible. Nor is there even any prejudice implied by deeming an aspect of some conduct morally significant, in need of justification, or “wrong-apt.” And although knowing what the moral significance of risking consists in may shed light on the proper standards governing permissible risk imposition, much as knowing why free expression matters sheds light on how and when it may be curtailed, all that strictly follows from ascribing moral significance to the imposition of pure risk is that risky conduct in and of itself is open to moral assessment, positive or negative, and so can be wrong.Footnote 3
As a way of getting at what is puzzling about the moral significance of risking, grant Samuel Scheffler's claim that morality is “pervasive” such that “no voluntary human action is in principle resistant to moral assessment.”Footnote 4 So characterizing morality might be thought to resolve the puzzle of risking's moral significance. For if morality is pervasive, it stands to reason that every action—risky action included—is open to moral assessment. On closer inspection, however, it becomes clear that morality's pervasiveness does not entail that, much less explain why, the imposition of pure risk calls for justification. More specifically, morality's pervasiveness does not entail that risking as such possesses the moral significance that attracts or warrants moral assessment. For the fact that some action is risky may not be a morally relevant feature, characteristic, or property of that action. Daydreaming, for example, can be wrong, as when one is supposed to be attending to a child learning to swim, but having pleasant thoughts is not what is morally significant about daydreaming; failing to pay attention is. Any action can be variously described. The question here is whether the description of an action as risky is morally relevant and thus whether the imposition of risk as such is morally significant. Citing morality's pervasiveness offers no resolution to that question.
Focusing on risk as such or pure risk imposition underscores the puzzle of the moral significance of risking because it is not at all clear what principled objection anyone could have to being subject to conduct such as pure risk imposition that lacks any discernable effect. James Griffin captures the import of this concern when he rhetorically asks, “How could murder acquire moral status independently of the havoc it wreaks in human lives?”Footnote 5 Broadening Griffin's plausible thought, it is hard to make sense of the moral significance of killing apart from its impact on the life of the person killed and her friends and family. Generalizing the claim, it is hard to understand how any action could be morally significant if it has no material impact on anyone's life. This casts doubt on the moral significance of pure risk impositions, which have no material impact on anyone's life.
Still it is difficult to resist the conclusion that even pure risk impositions are morally significant and so potential wrongs. Take drunk driving. When a drunk driver gets behind the wheel of his car and begins driving, at least with others around, that conduct is even then intuitively wrong—not merely culpable—even if no one who is put at risk is aware that the driver is drunk. One fairly straightforward conclusion that has been drawn from this intuition, which appears to be at odds with Griffin's general stance, is that what is wrong with (and so necessarily morally significant about) drunk driving cannot be that it actually harms anyone. For the wrong of driving drunk is perpetrated just as the drunk driver begins driving, when nothing has yet happened to anyone in any conventional sense. Rahul Kumar's observation about crashless drunk driving is representative: “As the risk did not in fact blossom into an actual harm, or end up setting back one's interests in any way, any talk of one having been left worse-off as a result of the drunk driver's conduct would be, in this case, misplaced.”Footnote 6 Yet, Kumar concludes, “there is nothing suspect about the claim that one has been wronged by the drunk driver . . . simply in virtue of his having, without justification, taken your life in his hands by exposing you, even briefly, to so serious a risk.”Footnote 7 Drunk driving, according to this view, need not have any adverse effect on anyone to be wrong and a fortiori morally significant, pace Griffin.
I believe that there is much that is right in both of these outlooks, diametrically opposed though they may seem, but also a fair amount that is misleading. Most of what follows is an extended attempt to show that, suitably finessed, the positions that Griffin and Kumar represent are compatible and between them contain the truth about the moral significance of risking. Specifically, I argue that while imposing risk does not involve material harm, like the injuries suffered in a car accident, it can nevertheless constitute a setback to a nonmaterial autonomy interest of a certain kind. And it is this nonmaterial negative effect or harm that imposing risk can involve that, in my view, grounds the primary moral significance of risking.
III. COMPETING ACCOUNTS OF THE RELATIONSHIP BETWEEN HARM AND RISK
I locate the moral significance of risking in an as-yet-unspecified nonmaterial harm to autonomy that risking can involve. Harming, of course, is clearly morally significant, plainly requiring justification. Linking the moral significance of risking to harming therefore holds out the hope of transferring the obvious moral significance of the latter to the former. But my position is a delicate one, and the surrounding territory is fraught. Despite possible appearances, I wish to forgo the “out” that is on offer here—I do not believe that risk impositions inherit moral significance from harm in an obvious way. Before explicating my own position, then, it will be useful to canvass and reject two less-complicated ways in which risk might be thought to inherit moral significance from harm and to assess one further argument casting doubt on the project of locating the moral significance of risking in any form of harm. The first competing view locates the moral significance of risking in its perceived potential for harm, the second maintains that risks are themselves harms understood in preferentialist terms, and the third contends that risk impositions, while setbacks to interests, are nevertheless not harms. Besides helping to clear the field for my account, exploring these other positions clarifies certain desiderata of the account that I am after.
A. A Commonsense Account
The most straightforward and initially intuitive account of the moral significance of risk impositions traces that significance to the fact that such actions portend harm. This commonsense view holds that risking is morally significant because of its perceived likelihood of or potential for causing material harm to the person upon whom the risk is imposed. There is much to be said for so unadorned a view. For there can be no doubt that the seeming potential that risky actions have for causing material harm to people matters morally and is, in that broad sense, morally significant. The problem with the view, given the aims of the present inquiry, is that it isolates only a secondary kind of moral significance. Such an account can hold only that the reasoning of the person imposing risk, not the risk imposition itself, is potentially impermissible, which is another way of saying that the act itself of imposing risk is potentially culpable or blameworthy but not impermissible.
The account that I am after attempts to prescind as much as possible from the material harm that risks can ripen into. Instead it seeks to determine whether pure risk impositions, remaining inchoate, can be impermissible and, in that core sense, morally significant. In a moral or more broadly normative (including legal) context, any risk that is imposed must be understood subjectively or epistemically, making it an estimation from some vantage point of a probability of harm.Footnote 8 It is possible to restate the commonsense account, then, as the view that what makes imposing risk morally significant is that for all one knows, someone will be harmed by that risky conduct.
If for all one knows one could end up materially harming someone by one's conduct or if one goes ahead with some conduct without giving any consideration to the chance that doing so could materially harm someone, then going ahead with that conduct may well be culpable or blameworthy. Yet an account of moral significance that focuses on an agent's beliefs and reasoning, thereby undoubtedly yielding insight into the agent's potential culpability, cannot yield any insight into the potential impermissibility of the risky action itself. Whether risk impositions themselves are capable of being impermissible or wrong is, however, the central question of the moral significance of risking.
To appreciate the distinctions between culpability and permissibility, as well as between the evaluation of someone's reasoning and their actions, consider the following variations.Footnote 9 First imagine someone who engages in risky conduct such as driving a gasoline truck in order to earn a living while at the same time fostering a slim hope of having an accident and wreaking havoc on the public at large. This person clearly suffers from a character defect, to say the least. The defect, though, does not bear on the permissibility either of the reason for driving the gasoline truck (namely, to earn a living) or of the risky action itself (namely, driving the gasoline truck). There is nothing inherently wrong with trying to earn a living or with driving a gasoline truck, whatever else one might hope happens in the course of doing either. The driver's indisputable character defect is only that, though, which is not to denigrate the importance of character or to minimize the driver's depravity. This shows that it is possible to disconnect the evaluation of one's character from the evaluation of one's reasoning as well as one's actions. Is it also possible to disconnect the evaluation of one's reasoning from the evaluation of one's actions?
Next imagine someone who drives a gasoline truck in order to wreak havoc. Wreaking havoc is not (at least usually) a normative reason for anything. One who acts for that reason thus fails to act on a morally defensible reason ceteris paribus. Such reasoning is morally impermissible. If the driver's reasoning or operative reason for driving the truck is impermissible, as it is in this case, then the driver's action or conduct is quite clearly culpable.Footnote 10 Driving a gasoline truck for that reason is unabashedly blameworthy. This much shows that risky action is morally significant in a certain way; namely, in virtue of being conduct that is capable of being culpable. What it does not show, however, is that risky action can be morally significant in the primary or core sense of being potentially impermissible or wrong, which is the issue here. The driver's defective, impermissible reasoning does not bear on the permissibility of the risky action itself of driving the gasoline truck. Whether there is reason to drive a gasoline truck and what the moral significance of doing so consists in are questions that stand quite apart from why one might drive a gasoline truck and whether one's operative reasons for doing so are morally defensible. Conduct that is based on impermissible reasoning is culpable conduct, then, but culpable conduct is not necessarily impermissible conduct. Just as there can be blameless wrongdoing, there can be blameworthy “right doing,” as it were. It is, after all, a platitude that one can do the right (or permissible, or morally significant) thing for the wrong reason.
So if one believes that taking some action would subject another person to a risk or that for all one knows, doing something would impose some risk upon someone else, or if one acts without considering the risk to others that so acting poses, then taking that risky action is certainly morally significant in some sense. The sense in which it is morally significant is, however, different from the sense with which I am chiefly concerned. It bears on the considerations that one takes to be normative reasons for doing what one does or on those aspects of what one is doing that one did not take to be conclusive normative reasons against so acting. It focuses, in short, on the reasoning of the person undertaking the action—her operative reasons. My focus is not the considerations that anyone takes (or does not take) to be normative reasons, though, but the normative reasons themselves. To determine whether risky actions bear moral significance in the primary sense of being potentially impermissible, one needs to determine whether there are any normative reasons that can make risky action as such impermissible.
Summing up, the hypothetical gasoline truck driver in the two preceding cases is open to moral criticism regardless of whether the action of driving the truck can be wrong and in that primary sense morally significant. Alternately, the driver's character and his reasoning are morally defective. But the central question of the present inquiry is whether risky action is morally significant in the core sense of being potentially impermissible or wrong. The commonsense approach to the moral significance of risking cannot be successful, then, because it is the wrong kind of account.
B. Finkelstein's Preference-Based Account
If the perceived potential for future harm that risking involves is rejected as the basis for the moral significance of risking, what of accounts that locate that significance in harm more directly? Claire Finkelstein, for example, argues that risk impositions are themselves harms.Footnote 11 If her argument were to succeed, it would establish the moral significance of risking by easy inheritance, as mentioned above. Finkelstein's argument, in a nutshell, is this: just as we should prefer to have, say, a lottery ticket and the chance at a benefit that accompanies it, we should prefer not to be exposed to risk on account of its accompanying chance of harm. The fact, in short, that risk is dispreferable entails that risk is harm.
Before proceeding, it is worth noting that Finkelstein's account is not presented as an account of the moral significance of risking. She never explicitly addresses what I am calling the moral significance of risking. Rather, she argues that imposing risk constitutes harm. For this reason, my aims may appear to be at cross-purposes with hers. Contrasting my view with hers may therefore appear to obfuscate, not illuminate, my own position. I disagree and do not think that our aims are at cross-purposes, though they are indeed different. Implicit in Finkelstein's view that risks are harms is a claim about the moral significance of risking. If it is true that risk impositions are themselves harms, as Finkelstein argues, then, given the obvious moral significance of harming, it would follow that risk impositions are morally significant (at least in part) because they themselves constitute harms. A risk imposition would constitute a harm separate from that which is risked, of course, but it would still be a harm. This makes it possible to extract an account of the moral significance of risking from Finkelstein's argument about the harm of risk, though it is not her aim to provide one.
What does the work in Finkelstein's account is her preference-based conception of harm. On her view, that which one prefers (or should prefer) is beneficial, while that which one (rationally) disprefers is harmful. Harms are setbacks to interests or to well-being, and thus the proper conception of harm will be the mirror image of the best conception of well-being. Finkelstein clearly believes that well-being consists in preference or desire satisfaction and that accordingly, harm consists in having one's preferences or desires thwarted. It is precisely this commitment, though, that undermines her defense of risk as harm and thus her implicit account of the moral significance of risking, and for a number of reasons.
If Finkelstein endorses an actual preference conception of harm, then there is little that recommends it. Our preferences are often misguided. and we often prefer, through ignorance or whatnot, what is in fact bad for us. This is not a novel observation, but it is decisive.Footnote 12 If on the other hand the preferences that matter are rational preferences, which is more plausible, a new (and indeed novel) problem arises. Rational preferences are typically spelled out as fully informed preferences—the preferences that matter are the ones that one would have if one knew all the facts about what one might prefer. From that omniscient perspective, however, risk just disappears. It is only because we do not know what the future holds, after all, that we think and reason in terms of probabilities, and so of risk. In other words, we think and reason in terms of risk due to our bounded knowledge. A rational preference conception of harm assumes away those boundaries, but in doing so, it also assumes away the object of inquiry, namely, risk.Footnote 13 It may well be possible to craft an account of “sufficiently informed” preferences that have normative bite but which stop short of being fully informed, thus preserving some room for risk. In any case, this is but one difficulty with Finkelstein's approach and not, moreover, the fundamental one.
Even setting these problems and the potential rejoinders to them aside, it is simply not the case that what is dispreferable is necessarily harmful. If, for example, one has the choice of living either in a house that is wired with explosives or in one that is not, it is, of course in everyone's interest in some sense to live in the latter house.Footnote 14 This entails that one's interests in some sense would be set back if one ended up in the house wired with explosives. Though harm is a setback to a person's interest, one who ends up occupying the house wired with explosives would not suffer harm just because the interest picked out by the (rational) preference was dealt a blow. This is because the kind of interest picked out by a preference (or a dispreference) is the wrong kind to anchor harm. Living in a house wired with explosives would certainly be contrary to interest in the variously termed “subjective,” “epistemic,” “decision-theoretic,” or “decision-ought” sense: it would be contrary to interest qua irrational, given what one believes. What is decision-theoretic irrational is not, however, necessarily harmful.
Interestingly, Judith Jarvis Thomson anticipates an argument strikingly similar to Finkelstein's and draws a similar conclusion about it. Discussing the possibility of a right against risk impositions, Thomson considers the following two propositions: first, “if it would be bad for X to get a thing Z, and if Y makes it probable that X will get Z, then Y causes X to be at a disadvantage,” and second, “causing a person to be at a disadvantage is itself causing the person a harm.”Footnote 15 She rejects the case for a right against risk impositions so founded because, she maintains, “we cannot really say that causing a person to be at a disadvantage is itself causing the person a harm.”Footnote 16 On her view, the fact “[t]hat people prefer a minor harm to a risk of a major harm does not make the risk of the major harm itself be a harm.”Footnote 17 We are right to disprefer that which is disadvantageous, in other words, but it does not follow from that that disadvantages—that which we disprefer—are themselves harms. Finkelstein gives no reason to think otherwise. This alone casts doubt on her conclusion about the harm of risk impositions, and in conjunction with the other problems owing to her reliance on a preference-based account of harm, Finkelstein's argument that risks are themselves harms falls short. The moral significance of risking does not reside in our (rational) preference for avoiding risk.
C. Perry's Rejection of “Risk Damage”
The strongest case—and at this point, the canonical case—for denying that risk impositions are themselves harms is advanced by Stephen Perry.Footnote 18 Perry develops his argument to rebut the claim that risk constitutes tortious harm, but he generalizes the argument to cover all harm.Footnote 19 Thus he denies that risk impositions can involve even nonmaterial harm, as I maintain. I explain here why Perry's argument does not preempt my account of risking's moral significance.
Toward denying that so-called “risk damage”Footnote 20 is a compensable injury in tort law, Perry discusses Hotson v. East Berkshire Area Health Authority,Footnote 21 where the plaintiff suffered an injury that initially went undiagnosed by the defendant hospital and that later developed into a much worse condition. According to the facts of the case, there was a 0.75 risk that even with a proper initial diagnosis the injury would have deteriorated. Normally, that would have been sufficient to defeat the prima facie case of negligence. For as the 0.75 risk implies, the deterioration in the plaintiff's condition would likely have occurred with or without any carelessness on the part of the defendant, entailing that it could not be a but-for cause of the deterioration. But because when the injury was eventually properly diagnosed the deterioration was inevitable, making the risk of deterioration not 0.75 but 1, the plaintiff argued that that increased risk of deterioration was itself a compensable injury. The chance of avoiding the harmful deterioration “was an asset possessed by the plaintiff when he arrived at the authority's hospital. . . . It was this asset which [counsel] submits the plaintiff lost in consequence of the negligent failure of the authority to diagnose his injury properly.”Footnote 22 A trial judge and the English Court of Appeal accepted the plaintiff's argument, but Perry sides with the House of Lords, which rejected that argument.
To impose a risk is to impose a probability of harm, and on the most plausible reading of Hotson, Perry maintains, the claim of risk damage is based upon a relative frequency conception of probability. This entails that, on the facts of Hotson, if one hundred people had suffered the same initial injury, seventy-five of them would have gone on to suffer the deterioration anyway, while twenty-five of them would not have. Perry points out that this assessment is relative to the reference class that is chosen as the baseline, but that certain assessments will be more or less accurate. Specifically, if you assume knowledge both of the characteristics that determine whether any given individual of the one hundred would have suffered the deterioration anyway and of which particular persons have the causally determinative characteristics, it is possible to partition the hundred-person reference class into one group of seventy-five who would have suffered the deterioration anyway and one group of twenty-five who would not have, corresponding to the two probabilities. The Hotson plaintiff must be in one of the two groups, and with enough information, it would be possible to know which one. The risk damage claim, Perry argues, “must surely be independent of possession of knowledge of this kind,”Footnote 23 but clearly it is not. With enough information, it becomes apparent that on this hypothetical rendering of the facts, the defendant hospital simply caused some people but not others material harm on account of the misdiagnosis. When enough is known, Perry concludes, risk disappears entirely, and with it, risk damage, for we are left to speak only of material harm and its absence.
Though Perry's focus in the above discussion is on tortious harm, he believes the lesson that he draws generalizes to cover all harm. On his view, risk cannot be harm—period. Yet it seems to me that the provenance of Perry's analysis has implications for and, more specifically, limits the scope of the conclusion that he can rightfully draw on the basis of his analysis of risk damage. Tort law recognizes as harmful a certain kind of harm, and while that kind of harm certainly outstrips tort law, it does not exhaust the kinds of harm that there are. In short, Perry's rejection of tortious risk damage does not rule out the possibility that risk impositions are nevertheless authentically harmful.
Tortious harm is any harm that is compensable under a regime of tort law or is at any rate whatever should be compensable given the principles underlying tort law. What counts as tortious harm is therefore delimited by legal doctrine or, at a minimum, legal principle. In ascertaining tortious harm, one looks backwards, surveying the state that the supposedly injured person was in before the allegedly tortious behavior took place, the actual causal chain that resulted in the injury, and the apparent injury that was ultimately caused. Importantly, one also looks at the material condition or state that the person would have been in had the allegedly tortious behavior not occurred, for it is on that basis that the victim's harm and level of (potential) compensation is determined. In this way, tort law recognizes as compensable injury only material harm satisfying a counterfactual test.Footnote 24 In Perry's words, “[t]he idea is to put the plaintiff in the position that he or she would have been in had the tort not occurred, to the extent that this can be done with money.”Footnote 25 Perry is absolutely right about this, but the requirement that tortious harm pass a counterfactual test that tests for changes in a plaintiff's material position is also the reason Perry's argument against risk damage cannot be generalized to cover all harm.
The harm that risk damage, by hypothesis, embodies does not manifest itself in any material change to its supposed sufferer's life. Given that fact, and given the fact just adduced that tort law compensates one only if one's life ends up materially worse than it would have been in the absence of tortious behavior, tort law cannot compensate for (pure) risk damage.Footnote 26 This is to say that tort law will not treat it as harm, for harm is only cognizable by tort law if it is compensable. What is striking is that this is true even if there is such a thing as risk damage simpliciter—even if, that is, risk impositions really are in some (other) sense harmful. Nonmaterial harms do not rate on the particular counterfactual test employed by the law of torts, not because it employs a counterfactual test but because the counterfactual test employed by the law of torts tests for material setbacks. From a torts perspective, then, nonmaterial harms are for practical purposes illusory. Perry rightly rejects tortious risk damage in my view. By treating the material harm recognized by tort law as paradigmatic, however, he overlooks the possibility that nonmaterial harm, not conforming to tort law's template, may yet be authentic harm. His argument therefore does not show that the imposition of risk does not constitute harm of any kind.
IV. AUTONOMY, OPTIONS, AND NONMATERIAL HARM
Between the various flaws and limitations of the forgoing arguments there is both space and pressure to render a different kind of harm-based account of the moral significance of risking. As no sound argument can be offered to explain what negative material impact risking has on those subject to it, making sense of risking's moral significance requires widening the parameters on how risk could affect life in a morally relevant way. I believe that one can accomplish this by widening both what one means by life and what can count as affecting that wider conception of life, and arguing for an autonomy-based account of the moral significance of risking.
When Griffin rhetorically asks how murder could acquire its moral status apart from its terrible impact on life, he has in mind a straightforward understanding of what life consists in, what we might call “biological life,” as well as a straightforward conception of what the relevant impact on life consists in, which I am referring to throughout as a “material” impact or harm. So murder is wrong (and so necessarily morally significant) at least in part because of the way it materially affects our biological life. But one's interests can outstrip one's biological life, and more matters morally than matters materially.
Consider these facts in turn. The wider conception of life that accounts for the capaciousness of one's interests we can call “normative life.” It is these two different conceptions of life, biological and normative, that Thomas Nagel eloquently trades on in maintaining “[a] man's life includes much that does not take place within the boundaries of his body and his mind, and what happens to him can include much that does not take place within the boundaries of his life.”Footnote 27 It is worth noting that if one is unaware that one is being subject to risk, as is true of pure risk cases, risk will also not affect one's subjective experience, or what we might call one's “experiential life,” which is a subset of one's biological life just as biological life is a subset of normative life. Thus, in the course of defending the possibility of posthumous harms, Joel Feinberg contends, “[b]ecause the objects of a person's interests are usually wanted or aimed-at events that occur outside his immediate experience and at some future time, the area of a person's good or harm is necessarily wider than his subjective experience and longer than his biological life.”Footnote 28 I would only amend Feinberg's statement to broaden the scope of his claim, so that it recognizes that the area of one's harm is also wider, as it were, and not just longer than one's biological life.
Relating this to the present discussion, we can say that just because being subject to risk cannot in and of itself affect one's experiential or biological life, it does not follow that it cannot still affect one's normative life. The question now is whether risk actually does or can affect anyone's normative life. Any impact, whether from risk or anything else, must impact normative life. If it does not, there is no sense at all in which risk impacts a person's life and thus no way in which risk is morally significant in the central sense of being potentially impermissible. By widening what kinds of effects on life matter beyond material effects, however, we can recognize how risk impositions relevantly affect the normative lives of people even if not materially, and thus how they are in fact morally significant.
Imposing risk does just this in virtue of narrowing the risked person's otherwise “open future,”Footnote 29 to borrow Feinberg's turn of phrase. One can do harm in subjecting a person to risk, for it effectively attaches sanctions to or normatively forecloses certain options that would otherwise be available to the individual, thereby narrowing the risked person's set of worthwhile opportunities. Narrowing one's open future diminishes one's autonomy, suitably understood, and it is in this that the moral significance and thus the potential impermissibility of pure risking lies.
I take Joseph Raz's influential conception of autonomy as a starting point. On his view, autonomy requires plotting one's own life and having a range of acceptable options from which to do so.Footnote 30 In the present context, a salient determinant of an option's acceptability is its safety, that is, whether it can be exercised without thereby suffering material harm. How does imposing risk narrow one's options and so diminish autonomy? It is easiest to answer this question by analogizing imposing risk to laying a trap.Footnote 31 Laying a trap in itself materially affects no one, but it can nevertheless impinge upon a person's nonmaterial autonomy interest. This is because the trap takes away the option, or more accurately renders unacceptable the exercise of the option, of stepping where the trap has been set. On this picture, and employing Raz's conception, one's autonomy would be utterly annihilated if enough traps were laid, for that would afford one too few safe passages—one's range of options would fall below the threshold of overall acceptability. Autonomy, so conceived, is more fragile than this suggests and Raz himself may believe, though, for autonomy comes in degrees and thus can be curtailed or diminished even when not outright annihilated. What this means is that even a single well-placed trap, or even foreclosing a single option, can conceivably diminish one's autonomy, which is far from saying that it necessarily does.Footnote 32
Imposing risk is like laying a trap in the following way. Like laying a trap, imposing risk does not itself materially harm anyone. Instead, just as someone's safe courses are winnowed down when one lays traps, so, too, another person's safe options are narrowed when one subjects them to risk. While risk is usually (and rightly) regarded as a probability of harm, what is relevant about risk here is an entailment of this, namely, that a risk imposition also connotes the possibility of harm. Laying a trap creates the possibility that someone will become caught in it, just as imposing risk creates the possibility that someone will suffer a material harm, namely, whatever material harm is the object of the risk. The possibilities of material harm in these cases just are the courses or options that would be materially harmful were one to take or exercise them. When one is subject to risk, in other words, certain of the possible things that one might choose to do (or ways that one might choose to be)Footnote 33 are no longer viable because materially harmful. There are any number of prosaic examples that illustrate this: sharing the road with others limits how and where one may drive, taking certain antibiotics limits what one may ingest or how much sunlight one may get, working with electricity demands single-minded attention that prevents multitasking, and so forth. When one is subject to risk, then, it is akin to having certain items removed from the menu of choice-worthy options with which one was originally presented. Risk impositions, in short, normatively foreclose formerly safe possibilities.
If when one is subject to risk, one could be materially harmed, and if that means that a formerly safe option available to one is no longer available in the relevant sense, then being subject to even a single risk imposition can impair one's autonomy and can constitute a nonmaterial harm that must be justified and whose imposition may be impermissible. As the harm of a risk imposition has no material profile, it need not pass any counterfactual test for a material setback to be validated as authentic harm. If one is put at risk but exercises the very options that one would have exercised absent the risk, which, as it happens, are safe, it should be clear that one suffers no material harm—a counterfactual test calibrated to detect material changes in position does not register any misfortune here. One nevertheless suffers the harm of having one's autonomy curtailed, for the risk imposition narrows one's safe options. Or again, if one's open future is assailed through risk impositions, but one still manages to live exactly the life one would have led if one's future had remained wholly open, then in all material respects one's position remains unchanged, yet one's autonomy has still been curtailed. This should make plain that there is more to harm than material harm. One might say, in a rough and ready way, that escaping any injury threatened by a risk imposition does not entail that one has escaped harm. In these cases one is lucky not to have been materially harmed, to be sure, but equally clearly one is still nonmaterially harmed in virtue of the negative impact the risks have on one's autonomy. Material impacts are not the only morally relevant kind of impacts on normative life. Autonomy matters, too.
It is the bare curtailment of autonomy that risking can involve that calls for justification, grounding the moral significance of risking as potentially impermissible. Focusing only on what materially affects or happens to a person overlooks a morally relevant feature of anyone's situation, namely, what options the person has or had. It is the possession of these options that define one's sphere of autonomy, whose integrity is of momentous moral concern. Pure risk impositions are morally significant precisely because of their negative but nonmaterial effect on autonomy. One way of putting the point is that understanding the harm of risk imposition involves a shift from asking what material position one would have been in absent the risk imposition to asking what material position one could have been in absent the risk imposition.
V. TAKING ANOTHER'S LIFE IN YOUR HANDS: FINESSING THE ACTION AND EFFECT OF RISKING
The account of the moral significance of risking that I defend accommodates Griffin's claim, suitably reinterpreted, that what matters morally must have some negative impact or effect on a person's life. Equally, it does justice to Kumar's claim, suitably reinterpreted, that risking can be wrong in virtue of taking another's life into one's own hands without justification, whatever the outcome. Yet the two positions purport implicitly to be at odds with one another. They appear to be staking out opposing ground in the well-worn debate between moral theories that revolve around either “the act itself,” which are typically deontological theories, or “the consequences,” which are typically teleological theories. I attempt to negotiate a kind of détente between these camps by locating the moral significance of risking in what it is that happens in the act itself of imposing risk. Can this diplomacy really succeed?
Consider Thomson's claim that “[w]e do not think that the permissibility of acting under uncertainty is to be settled only later, when uncertainty has yielded to certainty.”Footnote 34 What Thomson means is that the permissibility of risk impositions cannot depend upon whether the risks imposed ripen into the harms that they threaten. Returning to the example of drunk driving, Thomson's claim would be that such conduct is wrong whether or not the drunk driver crashes into anyone. Thomson and Kumar thus agree: risky action alone provides basis enough for an assessment of its permissibility—risky action itself is morally significant in this way. This, I believe, is correct. But it also seems to me a mistake to think, as both Thomson and Kumar seem to, that no adverse effect on the victim of the risk imposition need be demonstrated to explain fully the moral significance of risking and, more specifically, how imposing risk can be impermissible and not just culpable.
Of drunk driving, Kumar for one argues that “[a]n adequate analysis of being wronged ought to be able to make good sense of our intuitions in this kind of case, rather than identify them as suspect because they do not involve anyone being left worse-off, or harmed.”Footnote 35 On his view, “[a] person can be wronged . . . simply in virtue of how she figures, or does not figure, in how one is rationally disposed to relate to her.”Footnote 36 This, I think, is mistaken. Such an account as stated is worth having, but it is an account of culpability or blameworthiness, not of permissibility, as it claims to be. For it revolves around the reasoning of the person imposing the risk, not the risk imposition itself. Especially in the first quoted passage, Kumar overlooks the possibility that a harm-based account of the moral significance of risking need not “avoid cashing out the basis of the victim's complaint against the wrongdoer in counter-factual terms”Footnote 37 in order to avoid cashing it out in terms of an actual material setback. Focusing on the nonmaterial harm that a risk imposition can inflict on a person's autonomy avoids doing that.
So it is true that one need not wait and see whether the drunk driver hits anyone to deem his behavior morally significant and indeed wrong, but the harm that I am arguing risk impositions effect is not the kind of harm that one waits on in observing a drunk driver. It is not material harm. The harm is instead imposed by virtue of the removal of safe options available to those in the drunk driver's vicinity, and that is an aspect of the act itself of driving drunk. By making other people less safe, the drunk driver harms. Thomson and Kumar are best interpreted, then, as holding that we need not wait for the material consequences of risky actions to play out before judging their permissibility, and that claim is fully compatible with the view advanced here.
The image of taking another's life into one's own hands, which Kumar invokes in avowing (to repeat) that “one has been wronged by the drunk driver . . . simply in virtue of his having, without justification, taken your life in his hands by exposing you, even briefly, to so serious a risk,”Footnote 38 in fact illuminates and is illuminated by my account of risking's moral significance. I conclude this discussion by explaining how that is so. The idea, at bottom, is that imposing risk on others is morally significant in these terms because risk impositions amount to claims of authority over others’ lives and, more specifically, over the range of options that constitutes their autonomy.
The idea of taking someone else's life into one's own hands is a moral one, nicely capturing the moral significance of risking. In a modern society permeated by risk, one cannot help but to have the lives of others in one's hands. This is not merely to say that people can die or otherwise get injured as a consequence of our risky conduct, but that their wider normative lives can be affected by what we do. Especially relevant here is that others’ autonomy is in our hands. This state of affairs is not just inescapable, moreover, but often appropriate: to the extent that people in complex industrialized societies elect to lead the kinds of lives that they do, which are inherently risky, they cede to each other some (and only some—they could not cede all) of that element of their autonomy that consists in having normative authority over themselves. Upon doing so, they are owed a duty of care, and the content of that duty must be responsive to the normative authority over themselves that they retain.Footnote 39
In this sense, then, by putting someone at risk and foreclosing options otherwise available to another person, one takes another's life into one's own hands and implicitly claims authority over that person. This in turn prompts an open question of whether wielding such authority is rightful or whether it is instead an abuse and thus wrong. The open question, in other words, marks the risky action as morally significant—the action becomes a candidate wrong. It may be that the risk imposition is permissible, all told, of course, but it would nonetheless stand in need of that justification. Thus a drunk driver has the lives of others in his hands in virtue of driving; in virtue of being drunk while driving, he breaches his duty of care to them. It is the fact that any driver has the lives of others in his hands, such that their lives are vulnerable to what he or she does, so that it makes sense to hold drivers to a (here unspecified) duty of care. It is when a driver does not heed that duty, as a drunk driver does not, that the risk imposed by the driver becomes wrong, for no one's life should be in the hands of a drunk driver.
The claims advanced here, in sum, are that risky conduct itself can be impermissible in virtue of diminishing the autonomy of those who are put at risk and thus nonmaterially harming them. Only if imposing risk involves a form of harm that diminishes autonomy and only if we recognize a normative conception of life wherein one's interests are wider than one's material interests can the imposition of risk itself be morally significant in the core sense of being potentially impermissible. When we take each others’ lives into our hands by subjecting each other to risk, then, more hangs in the balance and matters morally than is commonly recognized.
