- 1 See Priel 2014a: 326–28 and Priel 2018.
- 2 See, e.g., Perry 2014: 38.
1One of the merits of Gregory Keating’s thoughtful and learned essay is showing the wide range of legal areas which deal with similar questions regarding law’s proper response to risk. Contrary to a view currently popular among many legal philosophers that sees a categorical distinction between private law and public law, and between corrective and distributive justice, Keating’s essay is commendable for eliding these divides. His examples come from tort law and environmental regulation, from army rescue practices to health and safety legislation. As someone who thinks too much time has been wasted on supposedly conceptual arguments about what tort law “is,” or in efforts to carve out a distinct domain for private law,1 I find this view refreshing. Keating’s approach also provides an important counterbalance to much of the current literature in tort law that sees the theoretical debate as exhausted by corrective justice and economic analysis of law.2 Where I disagree with Keating is on the justifiability of cost-benefit analysis and its role both in tort law and in other forms of regulation. Our disagreements aside, I must acknowledge that Keating’s arguments have forced me to think afresh about my views, and made me see more clearly the place of tort law within a broader scheme of different legal techniques of regulating human behavior.
2Keating begins his essay by identifying three precaution standards—cost-benefit analysis, feasibility, and safety. Against the current, he argues that the latter two reflect identifiable and defensible standards. Feasibility and safety are neither irrational nor empty; they reflect conscious, and at times morally required, levels of precaution. They incorporate into the law a familiar moral tenet: the priority of preventing harm. Of the various possible questions Keating’s essay raises, I have chosen to raise two sets of concerns. I begin by contesting Keating’s claims about harm prevention. I argue that despite the apparent appeal of the priority of harm prevention, it is in fact not a standard that societies seem to follow in practice. I then argue that the principle of harm prevention is in many circumstances unappealing because if taken seriously it will lead to more harm.
3These points are quite different in nature and I think largely independent of each other, but a theme running through both sets of comments is the place interpersonal moral attitudes should have in legal regulation. I consider these attitudes in light of recent psychological work on human morality. Such studies lead me to a relatively skeptical attitude about the role of interpersonal morality should play in justifying legal rules. Given the brevity of this essay my discussion of these issues will inevitably be suggestive.
- 3 See Keating 2012. He writes there, for example, that “[w]orkers’ compensation and the common law of (...)
4I begin my discussion with tort law. In part, I do so because this is an area of law with which I am more familiar, in part because Keating himself has applied his ideas on the priority of harm prevention.3 And though his latest contribution is not specifically concerned with tort law, it includes an extended discussion of a tort case, which Keating argues illustrates tort law’s commitment to the safety precaution standard, and by extension to the priority of harm prevention.
- 4 See, for example, Abraham 2012: 288; Ehrenzweig 1966 [1951]: 1431–32, 1435, 1444–45; Gregory 1951: (...)
5I think Keating is correct when he points out that courts in practice don’t always adopt a cost-justified precautions as the standard of due care. As early as the 1950s commentators have noted that even in negligence claims (which ostensibly reflects the cost-justified standard), courts sometimes adopt a standard of care that is effectively indistinguishable from strict liability.4 But unlike Keating, I think the doctrine does not reflect a commitment to harm prevention. I read the doctrine as partly based on different reasons which I hope to explain below, and partly reflecting certain psychological biases.
- 5 See Priel 2014b, for a more detailed demonstration of the issues discussed in this section. For an (...)
- 6 Keating 2018: online §10.
6It is important to note first that it is simply not the case that in general tort doctrine reflects a normative commitment to the priority of preventing harm.5 This mismatch between Keating’s normative views and tort doctrine weakens his claims both to have given an accurate description of both “our law” and “our moral intuitions.”6 Defamation law is a straightforward counterexample to the claim that tort law is premised on the priority of the prevention of harm. In the United States it is almost impossible to get an injunction to stop a defamatory publication, which implies that the law allows harm to reputation to happen, even when it is possible to prevent it. In other common law countries, though less absolutist in their attitude toward free speech protection, injunctions preventing potentially libelous speech are still very rare.
7Defamation deals with harms to reputation, which may seem less significant than physical harm. What defamation shows, however, is that tort law does not have a general policy of preferring ex ante prevention to ex post compensation. More significantly, it shows that prevention is not treated as having some kind of logical or normative priority, or that the centrality of ex post damages to tort law is merely a concession to the limits of our ability to prevent certain harms. At least in some cases, the preference for damages over injunctions (liability-rule protection instead of a property-rule protection) reflects a conscious choice.
- 7 Geistefeld 2012: 158.
- 8 Keating 2012: 313.
8It is tempting to reply at this point that defamation reflects on exception rather than the rule, and that at least as far as physical harm is concerned, “for centuries the common law has recognized that the prevention of such an injury is better than compensation via the damages remedy.”7 But, to begin, the example of defamation challenges a general claim Keating makes about tort law, namely that before any obligation to repair, there is a right—in the case of defamation to one’s reputation—to which there is a corresponding “primary” duty not to breach them. If we cannot make sense of tort law’s remedial rights unless we conceive of them as “derivative of substantive ones,”8 this must be true across the board. If we can make sense of them in the case of defamation, then the conceptual argument for the priority of harm prevention seems to founder.
9Beyond this, the case of defamation highlights a point that is true of cases of physical injury just as much as it is for cases of reputational harms: There are always costs (and not just monetary costs) to preventing harm, and such costs may make it rational for a society to opt for less than the highest degree of prevention possible. One can easily imagine legal regimes that would do more than American law currently does to prevent tortious harms to reputation. American law allows for more harm to reputation than the lowest realistically possible, because of the perceived harms from restricting speech.
- 9 See the statistics collected in List of Motor Vehicle Deaths in U.S. by Year, in Wikipedia, availab (...)
- 10 See Epidemiology of Motor Vehicle collisions, in Wikipedia, available at https://en.wikipedia.org/w (...)
- 11 Keating 2005: 176.
- 12 Cf. Daborn v. Bath Tramways Motor Co. Ltd., [1946] 2 All E.R. 333, 336 (“if all the trains in this (...)
10Once this point is recognized with respect to reputational harms, it is not difficult to show that a similar choice can be—and in fact, is—made with respect to other types of harm, including physical harm. That damages is by far the most dominant remedy for torts dealing with personal injury is, if you pardon the pun, not an accident. If the ex ante priority of prevention had been a central goal of tort law, it would have looked rather different. It would use injunctions much more broadly to limit activities that have a known, significant potential to cause harm. Even after significant advances in car safety, there were over 37,000 vehicle-related fatalities in the United States in 2016 alone (with millions more injured).9 It is estimated that in the course of the twentieth century over sixty million people died in automobile-related accidents around the world.10 Despite Keating’s claim that “[n]o amount of inconvenience distributed across a large number of distinct persons sums to the loss of a single life,”11 tort law (or for that matter any other law) has not responded to these harms by prohibiting the use of cars. If a ban on the use of cars sounds too radical consider a legal determination that driving at any speed over 5mph (8 km/h) constitutes prima facie (or even per se) carelessness.12 That this is not the case is at least implicit acknowledgement of the costs involved in doing so.
- 13 The statement in the text would be false if all regulation would reflect private interests masquera (...)
11Despite familiar rhetoric about the pricelessness of each human life, our tort practices simply do not bear out a real commitment to this idea. Below I will suggest a reason why; here, I wish to make a narrower point: It is true that human societies spend considerable amounts of money on harm prevention, including the prevention of harm from automobile accidents, but almost invariably these actions do not emerge as a result of tort litigation. For instance, whatever one thinks of the ideas mentioned in the previous paragraph about reducing the risk of harm from motor vehicles, they do not sound like the kind of actions that a court deciding a tort case would take. This shows that as a society, we do not turn to tort law (as opposed to other areas of law) for the sake of preventing harm. Perhaps the best evidence of tort law’s limited ability to prevent harm can be found by examining not tort law itself, but looking outside of it. The growth of government regulation of potentially risky activities in the course of the twentieth century demonstrates both society’s concern with harm prevention,13 and lawmakers’ recognition that tort law (alone) is not up to the task of doing so.
- 14 629 N.W.2d 727 (Wis. 2010). For a remarkably similar case see Grant v. Australian Knitting Mills, [ (...)
- 15 Keating 2018: online §25.
12If I am right in saying that tort doctrine on its own does not reflect a commitment to harm prevention, it behooves me to provide an alternative explanation to decisions that look like they reflect such a commitment. I will attempt to provide such an explanation using Green v. Smith & Nephew AHP, Inc.,14 which Keating argues shows tort law’s (occasional) commitment to the safety standard. Green involved a nurse who suffered serious adverse medical conditions as a result of her exposure to gloves that contained an excessive amount of latex proteins. The manufacturer was found liable, even though the risks of latex allergy were unknown at the time. Keating argues that a cost-benefit standard should have led to a dismissal of the claim, but that the decision can be justified because with certain products, such as clothing, we “normally expect [them] to be safe […] The question of whether this expectation is cost-justified never arises.”15
- 16 See Green, 629 N.W.2d at 731.
- 17 Green, 629 N.W.2d at 734.
13The case was decided on the basis of a strict liability standard for defective products. It is worth mentioning, however, that there is evidence that strongly suggests the defendant could have been found liable even on the basis of a negligence standard. The court in Green reported that the rate of allergy among health care professionals exposed to high-protein latex gloves was between 5% and 17%.16 An allergic reaction that affects as many as one in six users of a product is one that probably could have been discovered at a reasonable cost prior to the launch of the product. Furthermore, the court mentioned the testimony of expert witnesses who stated that the defendant’s gloves contained more latex proteins than almost all competing products.17 While possibly a coincidence, this fact suggests that information about the risks of protein allergy was already available and that other manufacturers had already taken active measures to reduce those risks.
14Setting these factual concerns to one side, let us assume with Keating that the costs of eliminating the risk were not cost-justified. I still think there is a more convincing explanation for the outcome of the case than a normative commitment to the priority of harm prevention. The alternative to ex ante prevention is ex post compensation. Tort liability comes after the fact and has limited ability to directly change behavior has significant practical consequences. To see them, I think it is important to keep separate the following two questions:
-
What risks should be eliminated and to what degree (ex ante)?
-
Who should bear the cost of non-eliminated materialized risks (ex post)?
15By relying on three public law precautionary standards which are unquestionably concerned with question (1) as a means for explaining tort doctrine, Keating invites us to think of tort doctrine through the lens of levels of harm prevention. By contrast, I think it is better to explain the outcome of Green in terms of considerations that are more relevant to question (2). In plainer terms, cases like Green address the question, “Given that harm has already occurred, who should bear it?” That the two questions are different is evident from the fact that even if we judge (by whatever standard) that a risk was one that should not have been eliminated, it remains an open question who should bear the loss in case that risk materializes.
16In most tort cases we assume these two questions go together: whoever is “responsible” for taking care of a risk ex ante, is also “liable” for the materialized risk (loss) ex post. More concretely, if responsibility for a risk is allocated to A and the risk materializes in a loss that B suffers, A is liable to pay compensation to B; if responsibility for the risk is allocated to A, who also suffers the loss, in effect the same principle applies: A ends up “liable” for that loss. Put in a different terminology, losses lie where they fall, unless someone else was responsible to address the risk that “created” them and failed to do so. This formulation is itself silent on what makes one “responsible” for a risk, but if this notion is expanded widely enough, one can end up with something that looks like Keating’s safety standard. So understood the debate on the standard of precaution is limited to question (1). Against this view, I will try to show that the choice of a more “expansive” principle of responsibility may be the result of different views on question (2).
17In focusing on question (1) Keating echoes much tort law scholarship. In an uncommon alliance between legal economists and corrective justice scholars, both advanced interpretations of tort law in which both questions were aligned. By conceiving of negligence as a problem of externalities, Posner’s article “A Theory of Negligence” both provided an answer on the optimal level of activity (and thereby the optimal level of risk) (question (1)), and explained tort liability as an internalization mechanism thereby providing an answer to the question of the limits of liability for materialized risk (question (2)).18 Corrective justice scholars have explained that liability for certain risks as “substantial” or “non-reciprocal” and as such reflecting a degree of disregard of others (question (1)) that merits the imposition of materialized risks on the person responsible for the risk (question (2)).
- 19 Posner (1972: 34) substantiated his views on a (contested) reading of late nineteenth-century Ameri (...)
- 20 See Restatement (Third) of the Law of Torts: Products Liability § 16 cmt. b.
- 21 See Henderson 1973: 1552–53.
18I believe both approaches reflect a possible reconstruction of nineteenth-century tort law, one that existed against a relatively limited regulatory apparatus.19 But one of the most important developments in twentieth-century tort law has been the willingness to break the link between these two questions. There is no question that the twentieth century has seen an active state endeavor to prevent harm. What is notable about that effort is just how little tort law has contributed to it. Continuing with the example of motor vehicle accidents, a long series of cases shows courts’ general unwillingness to impose liability on manufacturers for the design of their cars. I know of one tort doctrine, that was accepted by some American courts (but not, to my knowledge, elsewhere in the common law world), that could be understood as imposing liability for car designs that are deemed not sufficiently “crashworthy.” More accurately, this doctrine imposes liability on car manufacturers for an aggravation of a harm (compared to what it could have been) because of a faulty design.20 Generally, however, courts have not contributed much to improving car safety, because they do not have the knowledge of how to do so.21 The main legal mechanisms used for reducing motor-vehicle accident harm involved traffic regulation, car-safety regulation, car and driver licensing, and perhaps also other, less direct means (such as tax deductions for users of public transportation). In short, even if the law in general prioritizes harm prevention, it does not follow that tort law is entrusted with this task.
19Interestingly, it was legal economists, who in their effort to explain tort law in terms of deterrence, have squarely attempted to explain tort law as an area law designed to prevent some accidents from happening, but their analysis focused on knowable risks. Assume, as Keating does, that the defendant in Green could not discover at a reasonable (cost-justified) expense the allergy risk from high protein latex gloves. What lesson should it take from losing the case in terms of future actions? The answer cannot simply be “invest more in making sure the product you are selling is safe,” because that immediately raises another question: “How much more?” By hypothesis, the allergy risk in this case was unknown and unknowable by expending cost-justified resources, so it is impossible to know how much more to invest before it could be certain that the product is “safe.” If the safety standard reflects the expectation that certain items will never cause harm (if properly used), then the safety standard gives no guidance to a manufacturer as to the amount it should invest in additional product safety. Even if the risk can be somehow estimated, when the elimination of the risk is perceived to be not cost justified, tort liability will not lead to action.
- 22 Incidentally, and contrary to popular perceptions, this is not what happened in the Ford Pinto case (...)
20This is not just the notoriously vexed empirical question about tort law’s deterrent effect, it also touches on an analytical point: To the extent that negligence law seeks to bring about deterrence, it usually does so by way of imposing damages after the fact. The practical implication of this mechanism is that it leaves the option for the potential wrongdoer to keep her activity at the cost-justified level, and pay compensation for additional harms. By hypothesis this will cost the manufacturer less than acting at the safety standard. In other words, even if we want to force potential injurers to adopt the safety standard, tort law looks like a poor mechanism for doing so. It is true that explicitly adopting such a policy may lead to punitive damages, but in most cases, this will be difficult to demonstrate.22 Since the investment in product safety is unlikely to be zero—various considerations, both legal (regulatory inspection of the product), and non-legal (industry standards, commercial reputation, self-commitment), will make sure of that—punitive damages are also unlikely. As a result, the safety standard will not lead to a change in behavior. Under such conditions, one can expect regulation to do a better job.
21My more general point is that to understand what tort law is, we should seek to understand what it does, and that to understand that one cannot look at tort law in isolation, for it requires examining tort law in the normative environment in which it is found. Tort law in a world with relatively little government regulation will have different tasks to perform than tort law in a world of extensive government regulation. In the latter, tort law can be largely relieved of the task of setting precaution standards (as well as the task of serving as a tool for attaching blame based on moral responsibility) and made to focus more on the question of distribution of the losses that have already occurred.
22Let me add that saying that tort law, at least nowadays, is far more concerned with the question of distribution of losses for materialized harms (question (2)) should not be read to imply a neat division between public law and private law, for the answer to question (2) has itself been influenced by regulatory changes, for instance the institution of compulsory liability insurance. In one of his judgments, Lord Denning gave a clear indication of both tort law’s shift toward disentangling the two questions, and of the fact that tort law’s attitude with respect to the second question has been influenced by such regulatory changes. The context was whether liability should be imposed on a learner driver who caused an accident and argued that she was not at fault, because she drove to the best of her (subjective) ability. Denning answered that she should:
- 23 Nettleship v. Weston, [1971] 2 Q.B. 691, 699–700 (C.A.) (emphases added).
Parliament requires every driver to be insured against third party risks. The reason is so that a person injured by a motor car should not be left to bear the loss on his own, but should be compensated out of the insurance fund. The fund is better able to bear it than can. But the injured person is only able to recover if the driver is liable in law. So the judges see to it that he is liable, unless he can prove care and skill of a high standard… Thus we are, in this branch of the law, moving away from the concept: “No liability without fault.” We are beginning to apply the test: “On whom should the risk fall?” Morally the learner driver is not at fault; but legally she is liable to be because she is insured and the risk should fall on her.23
23Without question, these ideas are not universally accepted, but I think Denning’s words capture something of great importance about the transformation of tort law in the course of the twentieth century. In the words just quoted Denning makes three important points: The first is that concern with the already-injured victim (question (2)) has led to the enactment of compulsory insurance. The second is that once this was in place judges have moved to “see to it” that the defendant is liable, with little regard to moral fault. Finally, Denning says that the question “On whom should the risk fall?” is answered by looking at which party is, and can more easily be, insured against the harm. The effect of this reasoning is to separate the question of the desired level of protection against risk from the question of the identity of the bearer of the materialized risk.
24None of this leads to the conclusion that the higher (safety) standard suggested by Keating is never justified, only that cases that appear to favor a high level of precaution do not necessarily do so because of commitment to the priority of harm prevention, because there is an alternative interpretation of the cases. Admittedly, however, what I said so far does not yet show that my alternative interpretation of the cases is superior. So how are we to tell? As in most interpretive questions, answers are unlikely to be clear cut. Some cases will better fit one account while others will better fit the other; and consequently the conclusion as to which account is overall superior is to some extent in the eye of the beholder. Nevertheless, I think there is one strong reason to prefer the account proposed here. If tort law showed relatively consistent commitment to the principle of harm prevention through its adoption of the three standards of precaution, we would expect to see the different standards to roughly match the perceived importance of the protected interest. Where the harm is to the most important interests, tort law would fairly consistently opt for the safety standard, with the lower standards adopted for increasingly less important interests. Similarly, we would expect a higher standard in cases of more frequent or probable harms. By contrast, on the alternative explanation suggested here, we will tend to see a shift toward stricter standards of liability in cases where defendants are perceived (for whatever reason) to be better bearers of the materialized risk.
25Though my remarks here are not definitive or exhaustive of the issues (in fact, in the next section I will present another, psychological, explanation for the diversity in precaution standards), they present competing explanations about the cases. My impression (and admittedly it is only that) is that my account provides a more accurate description of tort doctrine, although to reach more definitive conclusions, one will have to examine more than a handful of cases.
26I have provided an interpretation of existing tort doctrine and variations within it as reflecting a concern with the distribution of losses more than the prevention of harm. I have also argued that even when liability is imposed, the law gives an option to the defendant. My discussion so far has not directly addressed the question of the desirability of adopting precaution standards that are not cost justified. One may accept that tort law should be largely unconcerned with harm prevention, and consistently argue that other areas of law should.
- 24 Keating 2018: online §59.
27In this section I aim to address these question by pointing out some difficulties with operationalizing Keating’s ideas into workable legal doctrine. Suppose we were convinced by Keating’s ideas, what would they actually amount to? The question is not just a petty attempt to get Keating to provide a more precise designation of the appropriate level of precaution, given that he does not call for perfect prevention. The question is significant for two reasons. First, trying to articulate the ideas will likely show that the claims of incommensurability—“Aggregating harms and benefits does not make moral sense when the harms and benefits are not comparable”24—are weaker than Keating makes them out to be. For if harms really are incomparable with benefit or compensation, that remains true at any level of precaution short of complete prevention. Since Keating does not advocate for that standard, his position retains a degree of acceptability by not articulating his alternative more concretely. The difficulty here is that to turn an abstract stance into a more precise legal policy will force Keating to make the kind of comparison he claims is not just morally obtuse, but “does not make sense.” The view that (say) preventing harms is twice as important as avoiding a benefit is still a comparison.
28Second, calling for a clearer articulation of promoting safety highlights the fact that the appeal of this proposal may rest on its vagueness. Such claims sound good as long as they are stated using verbal formulas. When considered in greater detail, I believe much of their seeming plausibility dissipates, for it then becomes clearer that the proposed policy will have a negative effect on harm prevention. To see this imagine a community allocates a $1,000,000 to investments in safety measures. To simplify, assume that there are only two manageable (reducible) risks of death, one from motor vehicles and another heart attacks. Investing in improved safety measures to reduce the risks of motor vehicle accidents could save lives according to the following schedule:
$0–$250,000
|
10
|
$250,001–$500,000
|
8
|
$500,001–$750,000
|
5
|
$750,001-$1,000,000
|
3
|
29Investing in better heart attack prevention technologies could save lives according to the following schedule:
$0–$250,000
|
7
|
$250,001–$500,000
|
7
|
$500,001–$750,000
|
4
|
$750,001-$1,000,000
|
2
|
30If the money allocated to saving lives is limited to $1,000,000, then investing $500,000 in motor vehicle accident prevention and $500,000 in heart attack prevention will save the largest number of lives (thirty-two). Imagine that in the simple numerical example given above, we decide to adopt a “safety” standard for motor vehicle accidents, and so instead of investing $500,000, we now invest $750,000. As long as the overall expenditure on safety remains the same, the additional expenditure on motor vehicle safety will come at the expense of preventing deaths from heart attacks. The overall lives saved will be lower (thirty).
- 25 This helps address another point Keating raises. He argues, correctly, that in fact it is often alm (...)
31The example so far did not really involve any analysis of costs against benefits, because no information was given on the benefits from saving a life. It assumed a constant amount allocated to precaution. Adding the benefit from a saved life makes it possible to decide how much to invest in safety. One possibility is to say that, because each life is “priceless,” the benefit accrued from every saved life is infinite. It follows that society should go on investing money until no more lives can (realistically) be saved. In the example above, that would mean doubling the investment in precaution (from $1,000,000 to $2,000,000), as that would yield fourteen additional saved lives.25
32If this is what Keating advocates, the implications of his view are far more radical than he presents them to be. Societies today spend considerable amounts on harm prevention; but they also spend on public parks and libraries, on holiday parades and art galleries. Such investments are hard to square with a thoroughgoing commitment to the priority of harm prevention, and yet Keating does not seem to make the claim that states should stop funding museums, sell off their Rembrandts and Picassos and redirect all saved resources to better accident prevention. More prosaically, if what Keating means by advocating for higher precaution standards is that societies should invest more than they currently do in harm prevention at the expense of their subsidizing leisure activities, his call may be justified, although in that case, his real thesis is about the need to reallocate public resources. Another possibility is that he calls for higher taxes, but as he undoubtedly knows, appetite for such measures is limited, even when guaranteed to save lives. That, again, shows the distance between rhetorical commitments to preventing harm and limited willingness to act on such slogans.
33Though Keating does not specify in detail what the priority of harm prevention would look like in practice, it does not look like Keating is calling for any of this, because he subjects the priority of harm prevention to a serious caveat. In earlier writings, Keating acknowledged that the private ownership of cars seems to conflict with commitment to the priority of harm prevention, and further stated that his views may call for “taking various steps to reduce these risks, even if those steps, come at some cost in convenience.”26 Overall, however, Keating was unwilling to challenge the practice of private car ownership, because
- 27 Keating 2005: 200. Notice how Keating aggregates the benefits from driving—“[c]ollectively,” he wri (...)
[w]hat is at stake in evaluating the reasonableness of the risks of driving is a general practice—the loose practice of private automobile use as it now exists in this country. Within the practice as we now conduct it, each of the innumerable risk impositions that put others at risk of devastating injury for trivial gain are essentially indistinguishable from each other… So if we judge any one automobile errand unacceptable because it wrongly risks devastating injury for trivial gain, we should judge all such errands unacceptable. When we shift our focus from particular trip to general practice, however, we encounter a benefit that is not trivial. Collectively, these mundane trips are an important part of a normal life in our society. Doing without a private automobile in contemporary Los Angeles, for instance, is a hardship—the kind of hardship that makes the lives of the working poor in Los Angeles so onerous.27
34Contrary to what this argument suggests, the choice here is not between modern life and the Stone Age; in many developed countries today people lead good lives without owning a car. A society seriously committed to the normative priority of harm prevention could have adopted some other solutions. Here is one: Banning the private ownership of cars and massively expanding public transit. Specially-trained drivers would reduce the incidence of error, mass-transit vehicles would significantly reduce road congestion and with it the opportunities for accidents. In addition, such a change will lead to a significant reduction in air pollution, itself a cause of many deaths. This proposal will probably be somewhat inconvenient, but it is without question feasible. The fact that modern life in cities like Los Angeles is difficult for the working poor who cannot afford a car would now become just a further reason for implementing this proposal, for the expansion of public transportation would significantly improve their lives.
- 28 See Stock et al. 2017. On the effects of “distracted walking” see Bomey 2018.
35Keating’s answer to the prevalence of private car ownership virtually guarantees that in practice lives will be lost to convenience. It is well known that regulation almost always lags behind new technologies, and this is especially true for technologies that make life more comfortable, as these very quickly become “an important part of a normal life in our society.” For a recent example, consider the fact that a growing body of evidence suggests that the use of smartphones while driving (or, apparently, while walking) has led to a spike in traffic accident fatalities, after years of secular decline.28 But as mobile phone use is now an important part of normal life, we allow lives to be spared for what is, for the most part, a mere convenience. In short, when qualified by what may be called the “normal life proviso,” the proclamation that life should never be traded for convenience seems rather hollow.
- 29 It is possible that there is something inherent to cost-benefit analysis that leads to the underval (...)
36All this suggests that the changes Keating envisions are quite moderate. He does not literally think that the law should work on the assumption that life is priceless, only that life should count more. The question of the value allocated to life is not one I can contribute to, and I readily acknowledge that it is possible that current measures undervalue life. But unless it is shown that adopting the cost-benefit standard itself leads to the undervaluation of life, this alternative argument does not call for a change in the precaution standard. It calls for a different calculation within the cost-benefit evaluation. To illustrate this point assume, using the numerical example above, that we value life at $36,000. (If it is not clear already, all figures in this example, and especially the value attached to life, are completely made up.) The cost-benefit standard would still call for investing $500,000 in automobile accident prevention and $500,000 in heart attack prevention. Suppose that a safety standard would mean valuing life at $50,000 instead. If we adopted such a standard for automobile accidents and kept everything else the same, that would lead to increasing the expenditure on safety from $1,000,000 to $1,250,000. Society would now invest $750,000 in automobile accident prevention, and $500,000 in heart attack prevention. As a result the number of lives saved in the example above would rise from thirty-two to thirty-seven. This finding, however, is not a challenge to cost-benefit analysis, but an affirmation of it.29
- 30 Keating (2018: online §§26–29) uses this example. One might argue that in such cases there is an im (...)
- 31 Keating 2018: online §27.
37If all this is true, Keating’s paper still leaves us with an unanswered puzzle. Societies clearly spend much less than they can on preventing harm. At the same time, the rhetoric of the pricelessness of life is still widely used, and in certain circumstances it is not just talk. Societies seem willing to spend virtually unlimited amounts to save lives in certain contexts, amounts that far exceed what they are willing to expend on saving lives in other circumstances. In a case of known trapped miners, for example, societies have been known to spend enormous sums to save the miners’ lives with virtually no costs spared.30 Why do we behave this way? The first thing to notice is that such scenarios do not demonstrate society’s support for the safety or feasibility standards. Given that these standards do not assume perfect prevention, these standards countenance the possibility of accidents, and in the long run they will happen. And yet, I believe people would react in exactly the same way to a case of miners who were trapped in a mine that scrupulously followed a feasibility or safety standard. Similarly, and contrary to Keating’s claim, rescue operations of dead bodies left behind enemy lines do not in any way show that societies occasionally adopt the feasibility or safety standards. Keating himself explains this practice in terms of another value—solidarity.31
- 32 For a small sample of relevant research see Small & Loewenstein 2003; Loewenstein et al. 2001.
38Rather than demonstrating a consistent and justified commitment to certain levels of precaution, both examples show how moral judgments change in response to factors of questionable moral significance. I believe the best explanation for these seemingly inconsistent responses to risk have to do with the way humans respond to abstract versus concrete risks and harms, to risks involving unidentified versus known victims, and to familiar versus unusual risks and harms.32
- 33 Cf. Leonard 2018.
- 34 Green, 629 N.W.2d at 735.
39Several examples illustrate these phenomena. Start with the distinction between familiar and unusual risks. We are so used to cars, they are so much part of our lives, that we simply take it for granted that driving cars in and of itself is not morally problematic. They are the kind of “normal” background risk that is just part of life. People who live in areas where gun ownership is very common use precisely the same kind of arguments to oppose gun control: They see themselves as upright citizens and responsible gun owners, as are the people they know. To restrict gun ownership, they say, is to let the actions of a few individuals infringe on the rights of law-abiding citizens.33 Guns are as integral to their lives as cars are to people living in Los Angeles, so much so that the thought that gun ownership itself could be seen as dangerous (let alone immoral), strikes them as outrageous. This kind of thinking also works the other way: We are used to certain things (such as surgical gloves or clothes) being perfectly safe in normal usage. When that is what we expect, any departure from the expected standard is perceived as “wrong” as thus warrants liability. When the Wisconsin Supreme Court in Green said that the test for a defective product is based on what “would be contemplated by the ordinary user or consumer possessing the knowledge of the product’s characteristics which were common to the community,”34 it enshrined in legal doctrine a familiar psychological judgment (or bias).
- 35 Keating 2018: online §58; Keating 2005: 176.
- 36 I initially considered this a hypothetical thought experiment. I then found out this was no mere fl (...)
40The distinction between identifiable and abstract victim is also important. Keating mentions a well-known thought experiment invented by Scanlon of an individual who is made to suffer excruciating pain so that millions of others could go on enjoying a television broadcast of a popular sporting event.35 Scanlon has argued that this example shows that utilitarianism is false: It would be wrong to subject that person to pain, even if a utilitarian calculus would suggest that doing so would increase overall welfare. Keating uses this hypothetical as proof that losses and harms are categorically different from benefit, and that they cannot be evaluated on a single metric. Again, however, I see in this example something else, namely the way humans react to an easily visualized injury to a particular, identified, person. For consider a slight variation on Scanlon’s thought experiment: Suppose evidence incontrovertibly shows that each Super Bowl leads to one avoidable death. The death may be the result of (among others) increased traffic to and away from the game, accidents due to drunk driving of jubilant (or despondent) fans after the game, alcohol poisoning, injuries in post-game celebrations or from altercations among supporters of the rival teams.36 I find it hard to believe that any such information would lead to the abolition of the Super Bowl or to the adoption of other measures (no broadcast, crowd-free game, temporary ban on alcohol consumption) that might reduce the risks associated with the game. In other words, societies constantly make the kind of tradeoffs Keating says they should not (and will not) make as long as the victim cannot be identified in advance and the harm involved is not vividly described.
41My point is not that people should suffer pain for the sake of others’ entertainment, but rather that a serious discussion should be taken about what kinds of risks society is willing to take, fully recognizing that many activities will entail greater harm to others. My further point is that appeal to concrete moral intuitions about particular cases is often a bad guide for an informed discussion of such questions, because these are often influenced by cognitive biases. This is a loaded claim, for what I call “biases” are often seen by others as the very essence of (deontological) morality. I think, however, we can identify moral responses we should be suspicious of, if there are patterns in the moral choice cases that are in relevant ways similar to the choice in in non-moral cases, where we know the choice is the result of a cognitive bias.
- 37 Evidence that this is how our moral intuitions work can be seen in cases of moral dumbfounding, i.e (...)
42The reason for suspecting moral judgments based on cognitive heuristics comes from understanding their strengths and weaknesses. Cognitive heuristics are often very useful, as they allow us to act “without thinking,” but as we know from many studies of such judgments in the non-moral domains, mental shortcuts that rely on limited information tend to go wrong in systematic ways. In the case of non-moral intuitions we have a way of identifying the right answer and in this way show how the intuition can lead us astray. There is no similar way of identifying mistaken moral judgments, and this makes it difficult to judge moral intuitions as biases.37
43One way some have attempted to identify “mistaken” intuitions was to trace their evolutionary origins. Roughly the argument looks like this: If we can show how a particular moral intuition provided our ancestors with adaptive advantage in the African savannah, we can discount those intuitions now to the extent that we now live in very different conditions. An evolutionary story can explain why sugar is “intuitively” appealing to creatures living in an environment where sugar was scarce. Such an account can then explain why we should seek to “discount” this appeal in the world we live in today, where sugar is available in virtually unlimited quantities. If we can tell a similar story about moral intuitions, we may have on occasion reason to doubt some of our intuitions.
44Here, I propose a different way of identifying questionable moral intuitions. As mentioned, moral intuitions come with no clear standard of moral correctness, but we might be able to identify such errors indirectly by comparing decisions in moral and non-moral choice scenarios. If we can identify a heuristic as leading to an obvious error in a non-moral case, and then describe a similar choice in a comparable moral scenario, we have reason to suspect the decision in the moral case. This approach is obviously filled with difficulties, some of which are considered below. However, together with other arguments, I believe this approach can be used to question some moral intuitions.
45The subjection of moral attitudes to challenges may sound odd, or even fallacious (confusing empirical realities with normative judgments). This is not the place for a full-blown defense of this idea. All I can do here is present some reasons that may make it more palatable. In the non-moral domain we often rely on simple visual markers for quick judgments about the quality of objects; humans often employ similar techniques for drawing qualitative judgments about other humans. What in the former domain is a useful mental shortcut (albeit one that we recognize can lead us to error) is typically known as a “stereotype” in the latter domain, and is (rightly) condemned. Such stereotypical judgments often seem to rely on the same intuitive mechanisms we use in other domains; they may have their origins in the need to make snap distinctions between friend and foe. Many societies have been working, only partially successfully, to minimize the effects of such judgments, because they have come to recognize the limitations of the psychological mechanisms that lead to such judgments. I propose extending this idea to other moral intuitions.
- 38 The literature is extensive. Of particular relevance is Greene 2013, especially at ch. 5.
46For the comparisons suggested here to be valid, one must assume that the cases are similar in all relevant respects. But is it not obvious that the fact that one case involves a moral choice and the other a non-moral choice makes the two cases incomparable? To some extent the challenge is best addressed by looking at the proposed comparisons. If I say that moral stereotypes can be (at least partly) explained by analogy with non-moral decisions, it behooves the person who challenges this view to present a competing explanation. The advantage of the approach proposed here is that it does not treat moral discourse as mysteriously sui generis, but one that has many of the same features as non-moral discourse. Indeed, it supports the observation that the boundaries between moral and non-moral discourse are fluid: Are attitudes to risk “moral”? Does gossip have moral significance? It turns out that the answer to either question is not straightforward. Nevertheless, I recognize that the only way to adequately address this question without begging any question is to acknowledge that the proposed argument must be supplemented by a fuller account of the development of moral discourse. Here, all I can do is mention the fast-growing research from cognitive psychology, developmental psychology, neuroscience, and related areas that treats moral behavior and moral discourse as natural phenomena, and which seems to explain them using the same methods used for explaining other aspects of human behavior.38
47It is also important to note that not all moral intuitions will be discounted by this method. Some intuitions—call them “abstract moral intuitions”—such as the moral intuition that causing harm to others is prima facie wrong, that all people should be treated equally, do not seem like an application to the moral domain of intuitive judgments from other areas. I limit my argument here to different moral intuitions—call them “concrete moral intuitions”—that arise in response to particular events. Abstract moral intuitions (whether the result of innate tendencies, learning and socialization, or some combination of the two) can be defended as reflecting individual or (in aggregation) societal choices with respect to how to organize society, how to lead life. As such, while debatable (and constantly debated), there is no obvious standard against which they can be shown to be mistaken. My suggestion is that the analogy with non-moral intuitions (where showing how intuitive judgment can be mistaken) provides a possible basis for questioning some concrete moral intuitions.
48The view adumbrated here, if convincing, can be the basis for discounting some uses of moral intuitions in debates about tort law. Keating relies on abstract intuitions (e.g., “saving life is morally important” or even “saving lives is morally prior to promoting welfare”) which he supports by appeal to concrete moral intuitions (e.g., reactions to Scanlon’s thought experiment). This is a familiar strategy in philosophical argumentation which aims to give a more precise content to abstract intuitions, and especially their boundaries, with the aid of eliciting our intuitive responses to concrete situations. But comparing Keating’s concrete moral intuitions with non-moral cases suggests a possible problem with this practice: Focusing on individual cases may actually distort our conclusions about the general cases, because they introduce into our evaluation of the situation factors that we know from non-moral cases can lead to mistaken judgments. (The comparison with non-moral cases is important, because without it, one could argue in the opposite direction: The focus on concrete cases is important precisely because it makes us take into account considerations that are overlooked when analyzing questions in the abstract. The approach proposed here can be used to identify which of the two possibilities we should favor.) In this case, intuitions elicited by Scanlon’s thought experiment reflect a more general tendency to favor readily available, vivid scenarios, over abstract calculations.
49Although the scenario depicted in this thought experiment is odd, the clash of the abstract and the concrete is a familiar feature of the law. Tort law, like much private law litigation, typically involves concrete cases, and those often lead to similar intuitive reactions. As long as accidents happen, decisions will be made in circumstances that encourage reliance on concrete intuitions, and where it will be difficult to ignore them. There are at least two additional reasons why it will not be easy to rid ourselves of the influence of concrete moral intuitions in legal analysis. First, even if the distinction between abstract and concrete moral intuitions is accepted in principle, the classification of cases into one or the other will undoubtedly prove contentious. Second, and most important, in a democracy some fundamental choices are made by people, who may enshrine their moral intuitions, including ones derived from thinking about concrete scenarios, as law. All democracies rely on some division of labor in decision-making between the people (directly or through their representatives) and experts. This can be seen as a mechanism for limiting the influence of concrete intuitions by allocating the bigger-picture choices to democratic processes, while lower-level implementation to experts. In practice, however, the divisions are never neatly kept apart.
50The decision between legislation and common-law forms of rule-making is another institutional mechanism for allocating decision-making (and hence the considerations likely to be taken into account) to different individuals on the basis of their perceived different kinds of expertise. In other words, one can see the discussion about the role of concrete moral intuitions as a proxy for a discussion about the proper role of adjudication as a form of lawmaking. Notice, however, that thinking about concrete moral intuitions in this way places them in a different location within the overall argument behind a legal regime that addresses accidents. For there is no longer an assumption that the law should simply match our moral intuitions on any given question; rather, it is political considerations that take center stage. The choice to give concrete moral intuitions a role in the design of legal rules is now evaluated as part of a broader inquiry into their relevance to the functioning and legitimacy of state institutions. This is, in my view, the perspective from which they should be considered, and it is true for environmental regulation (public law) and tort law (private law) alike.
51And so we have now come full circle: I began this essay by commending Keating for recognizing that there is no fundamental divide between public and private law, contrary to what many philosophically-inclined private law scholars today think. But we now see more clearly where our differences lie: whereas Keating wants moral intuitions to play a more significant role in both private law (tort) and public law (environmental regulation), I wish to see both areas of law analyzed through the lens of legitimacy and comparative advantages of different political institutions. It is through this perspective that we should ask what role (if any) we should give to moral intuitions in decision-making.
- 39 MacCoun (2000: 1825) reported that “[m]any [of his] students roll their eyes, shake their heads, or (...)
52Prioritizing harm prevention seems both commonsensical (“better safe than sorry,” “a clever person solves a problem, a wise person avoids it”) and humane; challenging it looks crass and “instrumental” (“putting profits before people”). Even when defended in the name of increasing welfare, rejecting the priority of harm prevention suggests a willingness to sacrifice actual human lives at the altar of some nebulous notion of welfare; willing to do so for the sake of convenience looks even worse. Without question, something feels wrong about the transactional manner in which the discussion above considered the costs and benefits from saved lives, the casual way in which it weighed Old Masters against accident prevention.39 And when faced with a particular, identified injury or death from an accident, the thought that we could have done something to prevent it is inescapable, and in some sense invariably true. Keating’s essay reflects a thoughtful attempt to turn this feeling of unease into a theory of state action.
53My response consisted of two main points. First, that as a matter of fact our practices do not seem to match many of our immediate moral judgments. If we look at what societies do rather than what they say, we see they recognize the inevitability of tradeoffs, even when life is on the scales. And when making such tradeoffs, their actions (rather than rhetoric) shows their willingness to trade life for convenience. Second, I suggested that we should be careful about basing our policies on the analysis of individual cases. Doing so potentially skews our decisions by giving special weight to highly visible particular events and by giving insufficient weight to the implications of aggregated risks. I further argued that while this approach often corresponds to our intuitive judgments on particular cases, we have reason to be skeptical about some of these intuitions. Contrary to the view of many moral philosophers who see moral intuitions as (at least) highly reliable evidence of moral truth, I have argued that we think of them in the same way psychologists have argued we should think of intuitions in non-moral domains, namely as valuable but error-prone mental shortcuts. As such, despite their appeal, they may lead to mistaken actions with a fair degree of regularity. Ignoring those intuitions may be a better way of achieving the moral goals that lie behind them.
—Acknowledgments.— I thank two anonymous referees for their comments, which forced me to rethink and (I hope) improve my arguments. The limits imposed by the genre to which this essay belongs—a response to another article—have limited my ability to provide a fuller argument to some of the ideas presented, especially in the last section of this essay.