Copyright (c) 2001 Georgetown Law Journal
Georgetown Law Journal
89 Geo. L.J. 605
Law, and Liberty: The Case of Required Rescue
* Professor of Law and Professor of Philosophy, New York University. I have
benefited from the comments of members of workshops at Columbia Law School and
the University of California, Berkeley (Boalt Hall), and from discussion with
Helen Hershkoff. Paul Chevigny and Mark Geistfeld provided extremely helpful
written comments on an early draft. The support of the Filomen D'Agostino and
Max E. Greenberg Research Fund of New York University School of Law is
... After all, the specific legal duty that has been at the heart of this
debate is a duty to rescue a person in distress when this can be done at little
cost or danger, something that most of these same lawyers assume that almost
all people would in any case do. ... Moore is thus content to employ the
traditional idea of a
"presumption of liberty" in this context, which is simply that infringement of negative liberty is
always regrettable just because freedom from coercion is always valuable, and
so any criminal duty must have enough to be said in favor of it to outweigh the
loss of liberty it brings. ... The paradigm agent-neutral duty would be a
(positive) duty of
paradigm agent-relative duties would be the (negative) duties not to violate
rights. ... I think it fair to assume that Macaulay operated with a legal
moralist view of the criminal law. ... Therefore, though there are, as I have
said, good utilitarian reasons to limit any legal duty of
to the kind of minimal duty found in Rhode Island, that does not mean that a
utilitarian supporter of such provisions can ignore Macaulay's problems. ... It
appears then that if we accept the optimizing principle of
as part of our account of the aim of the criminal law, we will nevertheless
restrict criminal duties of
to duties to rescue with minimal demands and minimal sanctions for violation.
English-speaking lawyers continue to be resistant to the idea of positive legal
obligation, whether civil or criminal, and in this they differ from lawyers in
Continental Europe and Latin America.
n1 This is in one way surprising, because English is the language not only of the
common law but also of utilitarianism, a normative theory that explains all of
morality and political justice in terms of positive obligation--the single
positive obligation to benefit people as much as possible.
n2 But perhaps it is precisely this greater exposure to utilitarian thought that
has made English-speaking lawyers more sensitive to the uncomfortable
implications of the legal enforcement of positive duties. The main normative
ground offered for the rejection of positive legal duties has been that they
constitute excessive interference with individual liberty, and perhaps only
those who take seriously the idea of truly extreme positive moral or political
duties, such as are implied by utilitarianism, would make this rather dramatic
objection. After all, the specific legal duty that has been at the heart of
this debate is a duty to rescue a person in distress when this can be done at
little cost or danger,
n3 something that most of these same lawyers assume that almost all people would
[*606] in any case do. If this were the most intrusive duty that came to mind when
thinking about positive duties it seems unlikely that great revulsion at the
infringement of individual liberty would naturally arise in one's breast. But
if much more extensive positive duties come to mind, such as Lord Macaulay's
imagined duty to travel across India to provide medical services,
n4 then a concern with liberty is easier to understand. It is noteworthy, indeed,
that in explaining his rejection of a criminal duty to render assistance for
the Indian Penal Code of 1837, Macaulay paid special attention to the fact that
Edward Livingston had included such a duty in his 1833 draft penal code for
n5 In including such a provision, Livingston was following Bentham.
This negative connection with utilitarianism is partly speculative, and there
are of course other aspects of the history and structure of the common law that
would figure in an explanation of its traditional antagonism to positive
n7 But my focus in this Article is normative; I concentrate on what has been
offered by way of normative defense of the traditional view by judges and legal
theorists, and in this connection the influence of utilitarian normative
thought is not unlikely to have been important.
The main argument in defense of the traditional view, as I have said, has been
that positive legal duties threaten the common law's traditional deference to
n9 I will argue that this avowed concern with individual liberty
[*607] is disingenuous, or at any rate mistaken. Positive duties as such do not raise
a significant concern about liberty in particular. What they do raise, for some
of us at least, is the potential for serious material cost--serious diminution
of our welfare or well-being.
n10 This was Macaulay's main objection to a legal duty to rescue, and none of the
many defenses of such duties that have followed Macaulay have adequately
This is not to say that the common law's aversion to positive legal obligations
is justified, but rather to say that the possibility of very costly legal
duties presents the most compelling ground for such aversion. As there are no
[*608] solutions to the problem of cost, it is not surprising that we have not
managed to improve on Macaulay's discussion.
Insofar as legal theory is concerned with the normative justification of the
imposition of civil or criminal liability, it can neither ignore the underlying
problem of cost, nor avoid it in an unprincipled way by drawing arbitrary
lines. Moreover, it will not be sufficient for legal theory to offer purely
practical explanations of why costly positive legal duties would not make
sense. A full defense of positive legal duties must ultimately appeal to a
fundamental normative principle that is itself plausible; and plausible in
general, not just as a criterion for torts and the criminal law, but for all
areas of government policy and indeed personal conduct apart from the law as
well. For example, it will not do to defend minimal positive criminal duties by
way of utilitarian argument, noting that there are good practical utilitarian
reasons for keeping the costs of such duties low, without addressing the
problem of the apparently unacceptable levels of sacrifice utilitarianism
requires in the realm of personal conduct.
Though legal theory must address the problem of the costs of positive duties as
it arises generally in moral and political theory, this does not mean that it
must embrace a particular solution to that problem. I will argue that all
plausible views on the underlying normative problem converge in supporting the
legitimacy of the kind of minimal criminal duty to rescue that is typical in
Continental Europe, and is becoming quite popular in the United States. This
"overlapping consensus" among plausible normative views is, I argue, sufficient to justify legal
duties to rescue.
It may seem that, despite the attention it has received from legal theorists,
the issue of legal duties to rescue is in practical terms rather trivial. Most
opponents of legal duties to rescue are quick to insist that the person who
fails to perform an easy rescue is a
"moral monster," but they also note, as I have said, that such monsters are rare. This low
for a legal duty to rescue is sometimes said to strengthen the case against
n13 But it could also be thought to show that the whole debate is somewhat
pointless, because so little turns on it in practice.
[*609] It is a mistake to assume that only those legal questions that promise
significant immediate practical payoff are important. A good reason for the
traditional common law aversion to legal duties to rescue may also be a good
reason to resist positive legal duties of all kinds.
n14 The implications of such a conclusion are potentially of enormous practical
significance, especially when we move from the issue of the positive duties of
individuals to the positive (constitutional) duties of government.
n15 One writer has even taken the extreme step of claiming that the general
absence of personal liability for failure to rescue is a ground for rejecting
of transfer payments and
So it is the broader significance of antagonism to positive legal duties that
largely justifies our interest in the special case of duties to rescue.
n17 Here we have a simple example of a highly contested positive legal duty where
the case in favor is clear. We can learn a lot about the idea that positive
legal duties are in principle undesirable by focusing on this example. More
generally, legal duties to rescue provide an excellent case study for thinking
about the appropriate
[*610] methodology for moral and political argument about what the substantive
content of the law should be.
In any case, it is simply not true that the issue of the legal duty to rescue
is unimportant in practice. It is thought to be so because people
"don't need the law" to incline them to provide emergency assistance where it can be done at low
n18 But even if it is right to assume that it is a rare person who lacks the
minimal level of benevolence needed to move her to perform an easy rescue, this
line of thought ignores the fact that strong contrary inclinations will often
be present in such situations. For example, in the well-known Kitty Genovese
case, it is asserted that a fear of
"getting involved" kept people from calling the police.
n19 The point is clearest in the internationally most famous instance of alleged
failure to rescue--that of the Princess of Wales and the paparazzi. One does
not have to believe that if the photographers had, as it was initially alleged
(the charges were later dropped), failed to call for the emergency services
when the Princess of Wales was dying in front of them, they would have shown
themselves to have been totally devoid of any benevolent impulses.
n20 These photographers had an enormous incentive not to waste any time calling
for assistance: As they knew very well indeed, photographs of the Princess were
worth extraordinary amounts of money.
With the possibility of strong self-interested reasons pushing in the opposite
direction, it is clear that an assumption of widespread minimal benevolence
does not establish the practical irrelevance of a duty to rescue. Now, of
course, it is true that not even France's very stringent criminal duty--which,
as all the world must now know, provides for a sanction of up to five years in
prison and a fine of around $ 66,000 (500,000 FF)
n21 --can eradicate failures to rescue from
[*611] French soil.
n22 But so too, of course, are legal duties not to murder ineffective in very many
cases. The point is not that no one in a jurisdiction imposing a legal duty to
rescue will ever fail to perform an easy rescue, but rather that the case for
such a duty does not turn on the prevalence of imaginary moral monsters who are
entirely indifferent to human life or who like to see people die just for the
fun of it. All too familiar human motivations can rather easily overcome
whatever degree of concern for human life we would like to think is distributed
among people generally.
I begin my argument by introducing the traditional debate about the duty to
rescue and canvassing the structural and doctrinal issues that such duties
raise for criminal law in Part I.A. and tort law in Part I.B. I argue that in
the criminal context there is no serious structural problem raised by the
enactment of a duty to rescue. Matters are less straightforward in tort law,
but the structure of tort doctrine may also be sufficiently flexible to
accommodate such a duty. The real problems in this area are therefore
normative. I discuss a variety of normative objections to the duty in Part II,
and argue that the traditional focus on individual liberty is misplaced and
misleading. This leaves the problem of setting principled limits to the
interference with individual well-being by way of legal institutions. Part III
explains the scope of this problem for legal theory; Part IV argues that any
plausible view on the matter will justify the kinds of minimally sanctioned
criminal duties that have been enacted in a small number of U.S. states.
n23 The desirability of civil liability for failures to rescue, it turns out, is
I. THE DUTY TO RESCUE AND THE COMMON LAW
In common-law countries, liability for failing to assist a person in immediate
peril, even though one could do so at little danger or cost, is exceptional
both at tort and at criminal law. Indeed, liability for any kind of omission or
failure to act is exceptional. As an initial question, we must examine whether
this legal preference for commission over omission--misfeasance over
[*612] made necessary by central structural features of legal doctrine. This Part
begins by noting the exceptions to the general rule of no liability for
omissions in criminal law.
A. CRIMINAL LAW
1. Commission by Omission
In the celebrated case of
People v. Beardsley
n24 a man had left for dead a woman, not his wife, whom he had observed taking
morphine and camphor tablets before she fell into a coma. The Supreme Court of
Michigan overturned his conviction for manslaughter, observing:
"The fact that this woman was in his house created no such duty as exists in law
and is due from a husband towards his wife . . . . Such an inference would be
very repugnant to our moral sense."
n25 After quoting these same sentences, Graham Hughes has this to say:
To be temperate about such a decision is difficult. In its savage proclamation
that the wages of sin is death, it ignores any impulse of charity and
compassion. It proclaims a morality which is smug, ignorant and vindictive. In
a civilized society, a man who finds himself with a helplessly ill person who
has no other source of aid should be under a duty to summon help, whether the
person is his wife, his mistress, a prostitute or a Chief Justice. The
Beardsley decision deserves emphatic repudiation by the jurisdiction which was
The State of Michigan has found it possible to resist Professor Hughes's
eloquence these past forty years--Beardsley
has not been repudiated.
The facts of
raise the issue of crimes of commission by omission.
n27 The contrast here is with genuine crimes of omission, discussed in the next
section, in which a statute directly imposes a positive duty to act--such as
the duty to stop to render assistance when involved in an automobile accident.
n28 In a crime of commission by omission, the charge is not that the accused
violated an explicit duty to act in a particular way, but that an omission
attracts liability for a crime of commission. The doctrinal link that renders
an omission sufficient for liability for a crime of commission is the existence
of some independent duty to perform the omitted act. The trial court in
erred, however, in holding that the defendant's commonsense moral duty to
render assistance was
[*613] sufficient for these purposes; as the Michigan Supreme Court noted, the
independent duty in question must be a legal duty.
This remains the standard account of the doctrine,
n30 but as Professor Hall long ago pointed out, it is not terribly informative.
n31 Not just any independent legal duty will be sufficient to attract liability
for any given offense.
n32 Instead, a series of particular legal duties has been recognized by courts as
sufficient to trigger liability for a failure to act. Such duties may be found
in statute or contract, or they may be entirely creatures of the courts. Thus,
one basis for liability for commission by omission is failure to perform some
statutory (not necessarily criminal) duty to act. If a person breaches a
statutory duty to render assistance at the scene of an accident and the victim
dies, this failure to perform the legal duty to assist may be the basis of a
prosecution for manslaughter by way of the doctrine of commission by omission.
n33 An example of commission by way of failure to perform a contractual duty is a
railroad gateman omitting to lower the gate to warn motorists of an approaching
n34 The range of duties to act that have been judicially recognized--without the
involvement of legislatures or any explicit undertaking by the defendant--is
wide, but most of them are based on the existence of some special relationship
between the defendant and the victim, such as the relationship of husband and
wife, parent and child, and others.
The emptiness of the slogan that commission by omission requires an independent
legal duty and not just a moral duty should be evident. The duty of husbands to
render assistance to wives became an independent legal duty sufficient to
ground prosecution for homicide when the courts began to treat it as such
sometime in the nineteenth century. As Hughes explains, the slogan
is a relic of the principle worked out in the nineteenth-century manslaughter
cases that the common-law duty to care for children could be used as a base on
which to found a homicide conviction if death ensued through neglect, but
[*614] that this duty could not be extended into a more general duty to aid those in
distress where the status relationship of paternal support was absent.
In its current usage, the slogan simply means that a duty will be deemed merely
"moral" when a court finds insufficient grounds for commission by omission liability
to attach. This aspect of the jurisprudence of criminal omissions has provided
courts with what Hughes calls
"a facade behind which to hide the progress of the law."
n37 This is unfortunate from Hughes's perspective because,
"as often happens with such protective devices, the refusal to acknowledge the
reality of the progress has impeded the advance."
Though I share Hughes's perspective, I find more troubling the rather neglected
fact that the control of judges over the expansion of liability for commission
by omission raises serious problems of legality.
n39 Even when the independent positive duty is imposed by statute, its sufficiency
for liability for commission by omission is typically a decision for the
courts. As Professor Fletcher notes, most discussions of liability for
omissions are either, in the pro-liability camp,
"concerned about condemning the injustice of not punishing immoral omission," or, in the anti-liability camp, concerned about protecting individual liberty;
hardly anyone is concerned about
"the issue of legality in letting courts roam freely about our moral sentiments."
I will not dwell on the issue of legality in this article. I raise it here
because it is one reason why criminal liability for omissions would be better
based on explicit provisions to that effect, rather than by piggybacking on
crimes of commission. Repudiation of
should take the form of an explicit statutory provision requiring the positive
act of rendering assistance, rather than further extension of liability for
commission by omission.
Though provisions criminalizing failure to assist are the norm in Continental
Europe and Latin America, in the common-law world they are currently found in
only a few U.S. states and Australia's Northern Territory.
n41 Rhode Island's provision is exemplary:
[*615] Any person at the scene of an emergency who knows that another person is
exposed to, or has suffered, grave physical harm shall, to the extent that he
or she can do so without danger or peril to himself or herself or to others,
give reasonable assistance to the exposed person. Any person violating the
provisions of this section shall be guilty of a petty misdemeanor and shall be
subject to imprisonment for a term not exceeding six (6) months or by a fine of
not more than five hundred dollars ($ 500), or both.
Similar provisions are found in Minnesota and Vermont.
n43 Though the Rhode Island
"duty to assist" provision appears innocuous enough, it is this kind of minimal duty to render
aid that has been at the center of the debate about positive legal duties. Some
of the reasons offered for opposition to such duties have been based in
features of the structure of Anglo-American criminal law doctrine. I consider
those objections here, leaving the normative objections to Part II.
That there should be structural/doctrinal objections to duties to rescue is
prima facie somewhat puzzling, as such duties are hardly unique in providing
for liability for omissions. I have mentioned the ubiquitous duty imposed on
motorists that they stop when involved in an accident; equally ubiquitous are
duties imposed on parents to provide for the minimal needs of their children
n44 and on income earners to file tax returns.
n45 In one enlightened common-law country, Australia, all resident citizens have
the duty to enroll with the electoral office and to vote in federal, state, and
n46 But despite the familiarity of these types of positive criminal duty, it is
undeniable that their existence sits somewhat uncomfortably with some general
doctrinal principles of criminal law.
n47 As criminal liability for omissions came to the common law slowly and rather
recently, this appearance of conflict is perhaps not so surprising after all.
We have already considered the legality problems raised by crimes of commission
by omission. Statutory provisions imposing liability for omissions directly
obviously raise no such problems. But these provisions do raise concerns, among
scholars at least, with respect to at least three other general requirements of
the criminal law: the act requirement, causation, and mens rea. Furthermore,
[*616] potential structural concerns arise in situations of multiple victims and
multiple potential rescuers.
Criminal liability is said to require an (voluntary) act.
n49 What then do we say about omissions? It is not helpful to defend the
conceptual propriety of regarding omissions as kinds of acts.
n50 It is true that the word
"omission" seems out of place in the context of some failures to act, such as my current
failure to do some act that has not even crossed my mind; there is thus some
plausibility to the view that omissions are a distinct subcategory of
non-actions. Nevertheless, the concept of omission is clearly not robust enough
in our legal and moral practice to carry much normative weight on its own. Even
if we were comfortable thinking of omissions as kinds of acts, they would
clearly be very special kinds of acts, and the substantive question would
immediately arise whether these special kinds of acts were appropriately
criminalized. Conceptual intuitions about what counts as an omission will not
be sufficiently reliable to help us. We will need to
omissions as non-actions with certain characteristics, and then we will have
to evaluate the significance of those characteristics.
The most sensible approach is to interpret the act requirement as aiming not so
much at the distinction between doing and not doing, but at the distinction
between thought and action.
n51 Thus it is the criminalization of thought alone that the act requirement rules
out. A crime must consist in either an act or a failure to act. This is the
approach of the
Model Penal Code
n52 Michael Moore takes a slightly different approach, defending a strict
interpretation of the act requirement--what is required is indeed an act, in
the usual sense of a willed bodily movement.
n53 He then treats omissions as exceptions to the act requirement.
[*617] This reflects Moore's view that punishment for action is the standard case
while punishment for omission is non-standard and requires special
justification. Because the difference between Professor Moore's approach and
that of the
Model Penal Code
turns on a particular normative view about the proper purpose of criminal law,
n54 I postpone further discussion of Moore's view until the next Part.
The best way to think about the causation requirement mirrors the best way to
think about the act requirement. Thus it is pointless to attempt to defend a
philosophical account of causation such that we can say that, for example, my
omitting to save a drowning person caused his death.
n55 Causation is an ancient and fascinating philosophical topic, but it is
unlikely to be relevant to substantive criminal law. This is fortunate, as the
issue is so puzzling, and disagreement about it so persistent, that we would be
in poor shape if the substantive question of liability for omissions required
prior conclusions on the correct philosophical theory of causation.
n56 Luckily, there is no need for lawyers to settle on one particular
philosophical understanding of causation; this is evident enough from the vague
and avowedly non-metaphysical notion of proximate causation that is at the
heart of the legal causation requirement. At the doctrinal level, indeed, there
is no problem with saying that omissions are causes: They are
"actual" or but-for causes, and whether they pass the proximate causation test is a
question not so much of the theory of causation but of the proper limits of
n57 --the very substantive question that we are investigating for the case of
omissions. If a legal theorist proposes that the limits of criminal liability
should be set in part by asking whether the defendant caused the harm according
to the correct philosophical--as opposed to legal--theory of causation, then
the appropriate response is to examine that philosophical theory of causation
and see whether its claimed normative significance is plausible. It should be
[*618] remembered, however, that if omissions turn out not to be causes, the upshot
will not be conservative--the well-established crimes of commission by omission
will turn out to be outside the scope of legitimate criminalization.
The only genuine doctrinal puzzle raised by criminal liability for omissions
lies with the requirement of mens rea. Professor Hughes argues that
satisfaction of the mens rea requirement for crimes of omission puts pressure
on the traditional maxim that
"ignorance of the law is no excuse." The maxim, Hughes writes,
"ought to have no application in the field of criminal omissions, for the mind
of the offender has no relationship to the prescribed conduct if he has no
knowledge of the relevant regulation. The strictest liability that makes any
sense is a liability for culpable ignorance."
n59 Though this seems to overstate the problem, Hughes has identified an important
distinction between crimes of commission and crimes of omission with respect to
Professor Hughes criticizes Glanville Williams for offering this test of the
intentionality of an omission:
"If the defendant had been asked at any time, while the omission was continuing,
'Are you doing so-and-so?' (which the statute makes it his duty to do), would
the true answer based on the facts . . . be: 'I am not'? If so, the omission is
n60 Hughes points out that although he will say that he is not climbing Mount
Everest if asked while at home eating an orange, it does not follow that he is,
at the moment asked, intentionally not so climbing.
n61 But the problem here raised seems not to be unique to omissions.
I would prefer to cast Hughes's point as follows. To conclude that an action or
its omission was intentional we require that the agent was aware that the
description made relevant by the criminal provision applied to her action or
omission; for a conclusion that the action or its omission was reckless or
negligent, we require that the agent knew or ought to have known that it was
likely that the relevant description applied to her act or omission.
n63 Where the agent performed an action, we can typically apply these tests
without inquiring into the state of mind of the agent with respect to the
existence of a legal duty. Where the agent omitted to do something, by
contrast, we may often be unable to apply these tests without such an inquiry.
When I break the window of the house and crawl through the resulting hole, I am
aware that the description
"breaking and entering" applies to my act even if that phrase isn't literally before my conscious
mind. I don't need to know that I have a duty not to break and enter before I
know that is what I am doing--as opposed to, say, removing immediate obstacles
to my progress in a northerly direction. And if, upon leaving the restaurant, I
pick an umbrella from the basket at random, I can be said to consciously
disregard a substantial and unjustifiable risk that I am taking someone else's
umbrella, and thus said to be reckless with respect to the relevant
description, even if I have no knowledge of the prohibition on theft.
With omissions, on the other hand, it is often the case that the only thing
that could connect the mind of the agent to the relevant description is her
state of mind with regard to the legal duty to perform the omitted act. Hughes
gives the example of the pharmacist who does not know about the legal duty to
register the sale of poisons. In such a case, Hughes says,
"liability should depend upon the culpability of his ignorance."
n64 The point here is that whether the pharmacist's failure to register the sale
of a poison was reckless or negligent depends entirely on whether the
pharmacist knew or ought to have known that there was a good chance that there
to register sales. There is simply no other way that the description,
"failing to register the sale of a poison," could become linked in the appropriate way to the agent's state of mind. No
doubt the pharmacist also failed to dance a jig in celebration of the sale; for
"failing to dance a jig" or
"failing to register" to be relevant descriptions, something is required to pick them out from the
crowd of possibilities. Thus also, if a convicted felon fails to register as
required by a city ordinance, while having no inkling of any such regulation or
of the likelihood of its existence, it is implausible to say that the failure
was intentional or reckless--though it might be arguable that it was negligent
on the ground that felons ought to consider the possibility of special
regulations governing their movements.
The asymmetry between actions and omissions here is clear, and it is due to the
simple fact that at any one time there are so many potentially legally relevant
things that I am not doing. But the asymmetry is not as stark as Hughes
suggests. Clearly, some omitted actions are sufficiently salient to justify
saying that the omission was intentional without any knowledge of the duty to
act: If I fail to rescue a child I know to be drowning in front of me, for no
reason other than that I cannot be bothered to get out of my chair, it is not
at all unnatural to say that I intentionally failed to prevent the drowning.
n66 Furthermore, some prohibited
are not of sufficient salience, under the relevant description,
[*620] to count as intentional absent knowledge of the prohibition. If a city were to
pass an ordinance (without debate or advertisement) prohibiting people
strolling in the city park from intentionally coming within fifteen feet of the
river (for some reason, perhaps having to do with hatching birds, not apparent
to the lay person), and a person were to stroll within the proscribed area, the
only way that we could resolve whether the person did so with the required mens
rea would be to inquire whether she was aware of the ordinance--and this
remains the case even if, had the person been asked whether she was within
fifteen feet of the river, the stroller would have replied that, as it
happened, she was.
An inquiry into the agent's state of mind with respect to the existence of the
legal duty is typically not necessary to establish mens rea when a person has
acted in some way, because the descriptions of that act that are likely to be
legally relevant will be familiar enough in ordinary life--the issue of
knowledge, recklessness, or negligence with respect to the description can be
solved without such an inquiry. With certain highly unusual descriptions,
however, the inquiry may be necessary. So the problem Hughes discusses, though
much more significant in cases of omission, is not unique to them. The moral to
draw from his insight is that mistakes of law, in the sense of ignorance of the
existence of the legal duty,
n67 will negate the required mental state much more commonly in the case of
offenses of omission. But we do not need to go further and say that the entire
requirement of mens rea is inapplicable in the case of omissions.
A final pair of possible structural objections to duty to rescue provisions are
more practical than doctrinal; they concern the possibility of situations of
multiple victims or multiple potential rescuers. But in fact neither situation
is problematic. In the case of multiple victims, a rescuer complying with our
exemplary Rhode Island provision would continue to render assistance to as many
people as she could, so long as this involved no peril to her and so long as
the cost sustained remained reasonable; beyond that, she is acting above and
beyond the call of legal duty. Where there are multiple potential rescuers, it
is not absurd to think that if none of them does anything to help the victim,
each of them could be held criminally liable.
n69 It may not be practical to prosecute all such people, but selective
apprehension and prosecution, so long as the selection criterion is legitimate,
should raise no greater problem than does selective
[*621] apprehension and prosecution of speeding motorists.
n70 And if one person out of the crowd does act, there is no problem with the
liability of the others--the victim no longer needs any help, and so they are
no longer under any legal duty.
n71 It may be thought that there is a puzzle about who should make the first move
in this kind of case,
n72 but there is not. Utilitarian wisdom, but also simple common sense, tells us
that the person best placed to help (in terms of effectiveness and cost) should
be the one. Of course, if the best-placed person does not act, the legal duty
does not exonerate the others. What is required is that, taking into account
everything that is going on around you (including the behavior of others) you
render reasonable assistance if you can. Needless to say, in many cases people
lack information about exactly who is best placed to help, who next best, and
other factors that would help them decide whether to act. What reasonable
people do in such circumstances is react to events as they unfold; one takes
the first steps to render assistance, ready to back off if it becomes apparent
that there is a superior rescuer at hand.
It could be said that this response to what is essentially a practical
objection is speculative. To the contrary, the objection itself ignores a
rather long history of experience with legal duties to rescue. The population
and the courts of France, to take just one example, have managed to cope with
criminal and civil duties to rescue for some fifty years. If the problem of
multiple potential rescuers were serious enough to warrant rejection of the
duty to rescue, one would expect some supporting evidence to have emerged by
n73 It is somewhat astonishing to read practical objections such as these that
take no account of the existence of a world in which the impossible has come to
n74 In any case, such objections also apply to current common-law doctrine. If
twenty school teachers all failed to throw a rope to a drowning elementary
school pupil at the school rowing regatta, would our courts be unable to find a
way to impose liability?
Outdoing the criminal law's
, tort law offers us
Osterlind v. Hill,
n76 Buch v. Armory Manufacturing Co.
n77 and many more
[*622] besides. In
, the stock example of failing to effect the easy rescue of a drowning person
came to life. The villain of the piece had rented a canoe to the evidently
drunk victims. As the victims drowned, the villain, boat and rope close at
hand, sat smoking on the dock. In
, a boy injured himself by putting his hands in the defendants' machinery.
Chief Justice Carpenter had this to say:
The defendants are not liable unless they owed to the plaintiff a legal duty
which they neglected to perform. With purely moral obligations the law does not
deal. For example, the priest and Levite who passed by on the other side were
not, it is supposed, liable at law for the continued suffering of the man who
fell among thieves, which they might, and morally ought to have, prevented or
relieved. Suppose A., standing close by a railroad, sees a two year old babe on
the track, and a car approaching. He can easily rescue the child, with entire
safety to himself, and the instincts of humanity require him to do so. If he
does not, he may, perhaps, justly be styled a ruthless savage and a moral
monster; but he is not liable in damages for the child's injury, or indictable
under the statute for its death.
have effectively been overruled, but on narrow grounds that apply only to the
duties of landlords.
n80 Chief Justice Carpenter's example of the child on the tracks remains apt. His
rhetoric about law and morality is, however, even less helpful than the similar
slogan used in the context of criminal offenses of commission by omission. The
question before him was whether a legal duty to protect the boy existed; it is
obviously empty to say that if a duty is
moral, and thus not also legal, it is not legal. Perhaps he meant to say that
moral duties are not ipso facto legal duties, but this hardly seems worth
A person may be liable in tort for failing to provide aid where that failure is
part of a positive course of action already undertaken, such as the failure to
blow the whistle on the part of a train driver;
n81 here the failure to act is counted as part of a wider pattern of
misfeasance--negligent or otherwise tortious operation of the train. Liability
for genuine nonfeasance depends on the existence of some relationship between
the agent and the victim; landholders, carriers, hosts, and various others may
be liable for not taking positive steps to prevent harm or provide aid.
n82 The authors of
Prosser and Keeton on Torts
announce that decisions such as
"revolting to any moral
n83 They write that the sorry state of the law is due to the
"difficulties of setting any standards of unselfish service to fellow men, and
of making any workable rule to cover possible situations where fifty people
might rescue one."
n84 The first of these difficulties does indeed go to the heart of the problem,
but the second difficulty seems illusory. If all fifty fail to rescue the one,
then, as in the criminal case, they are all liable (though in this case we
would add that the plaintiff can recover only once and that the
enforced-against tortfeasor may seek contribution from the others).
n85 Likewise, the concern that, where there are multiple potential rescuers, no
one will know what is required of her can be answered, as in the criminal case,
by appeal to commonsense norms.
The situation of multiple potential victims is also tractable in the same way
as for criminal liability: Any tort duty would presumably be one to take
reasonable steps to render assistance--those duties to aid that are currently
recognized take this form--and at a certain point the costs of continuing to
rescue may become unreasonable.
n86 A very particular kind of situation involving multiple victims does, however,
seem difficult under tort doctrine. This is the case of multiple victims and a
potential rescuer who fails to act, but with the added twist that if the
rescuer had acted, she would have been able to rescue only some of the victims.
We may suppose that a ferry has capsized in the harbor and fifty people are
floundering in the water. The villain has a motorboat that could save as many
as ten people before they drown, though not more; yet the villain sits smoking
on the dock. The potential rescuer has failed in her duty, but to whom? Who has
suffered the damage necessary for tort liability? In such a case the victims'
survivors could argue that the villain's failure removed a chance of their
decedent's survival, thus drawing an analogy with cases where conduct that
fails a but-for test for causation of harm nevertheless increased the risk of
n87 Once more, such a situation could certainly arise under existing
[*624] doctrine that imposes a duty to act on, say, the captain of a ship in respect
of her passengers; the problem is not unique to a general duty to rescue.
We can conclude then that there is nothing in the structure of common-law tort
doctrine to block adoption of a general duty to rescue.
n88 And indeed, a recent path-breaking decision of the New South Wales Court of
Appeals, upholding a damages award of over three million Australian dollars
against a physician who had refused to come to the assistance of a stranger,
resulting in that person's permanent quadriplegia, proves that it can be done.
n89 Whether this is a development to be welcomed, let alone extended beyond the
[*625] category of medical practitioners, is another matter. Like several other
commentators, I believe that the case for civil liability for failure to rescue
is weaker than the case for criminal liability.
n90 The reasons for this will be discussed in Part IV.
II. LAW AND LIBERTY
Though all agree that the person who fails to effect an easy rescue is a moral
monster, there is a good deal of disagreement about why. At the level of moral
theory, there should be no room for disagreement: The duty to rescue is a duty
a positive duty requiring people to benefit others, even total strangers.
n91 Most defenders of legal duties to rescue are anxious to avoid this natural
n92 The reason for this will emerge as we begin to consider the normative
objections that have been raised against a legal duty to rescue.
All normative objections to a legal duty to rescue depend upon at least one,
but sometimes two, kinds of normative premises. One premise of any argument
that there is a general reason of political morality to resist legal duties to
rescue must be some view about the appropriate aims of the law. More
specifically, whether explicitly stated or not, there will be a premise about
the positive aims of the particular branch of the law under discussion. The
second kind of premise, which may or may not be present in any given argument,
is a claim about the negative constraints that apply to the pursuit of the
n93 Thus, for example, one might hold that the aim of the criminal law is to
promote the social good and also accept, as a constraint on legitimate pursuit
of that aim by government, a certain set of individual rights.
My aim in this Part is not to critique the various premises of the arguments
discussed--that would be too great a task--but rather to show that these
premises do not lead to the conclusion that legal duties to rescue are
[*626] or obviously unwise. I will begin with the criminal law, and with a theorist
who is gratifyingly explicit about the premises of his arguments.
Harm to Others
, the first volume of his four-volume work,
The Moral Limits of the Criminal Law
n94 Joel Feinberg presents an elaboration and defense of what he sees as the
central thrust of John Stuart Mill's essay,
n95 As a utilitarian, Mill held that the aim of the law was to promote happiness,
n96 As a constraint on all attempts to promote general well-being by means of
coercion--either by the state or by informal social pressure--Mill presented
his harm principle, stating that
"the only purpose for which power can be rightfully exercised over any member of
a civilized community, against his will, is to prevent harm to others."
n97 The harm principle was not, for Mill, a departure from utilitarianism, but
rather a specification of what utilitarianism, properly understood, implied
about interferences with people's liberty.
As H.L.A. Hart emphasizes in his discussion and defense of Mill's position,
there are two strands to the harm principle: the rejection of paternalism and
the rejection of what Hart calls
"legal moralism," which is the view that a legitimate aim of the law is the enforcement of
morality for its own sake.
n99 Both strands are evident in Mill's sentences immediately following the one
His own good, either physical or moral, is not a sufficient warrant. He cannot
rightfully be compelled to do or forbear because it will be better for him to
do so, because it will make him happier, because, in the opinions of others, to
do so would be wise, or even right.
As Hart points out, the two strands of Mill's position do not stand or fall
together; Hart is himself less concerned about resisting paternalism than legal
n101 For us, however, the crucial point is that these are the main two strands of
Mill's position and that the use of the word
"harm" is not meant to mark out an important distinction between harm and
. This seems clear enough from the fact that Mill, like Bentham before him,
explicitly endorses a
[*627] legal duty to rescue.
n102 Mill states that
"in all things which regard the external relations of the individual, he is
amenable to those whose interests are concerned, and if need be, to society as
n103 He goes on to say that the decision whether to enforce a person's positive
responsibility to other persons depends largely on
"expediences," and his discussion makes clear his view that the enforcement of duties of
is in general less expedient than the enforcement of negative duties not to
It is therefore rather surprising that Feinberg, who is an enthusiastic
supporter of a criminal duty of easy rescue, should find it so important to
cast the duty to rescue as a duty to prevent harm--as opposed to
a duty to confer a benefit.
n105 The key contrast for Feinberg here is not that between a positive and a
negative duty, for he is willing to construe the harm principle as allowing for
the enforcement of positive duties to prevent harm rather than just negative
duties not to cause harm.
n106 Instead, he thinks it important for the legitimacy of a duty to rescue that it
be understood as the positive duty to prevent harm rather than the positive
duty to confer a benefit.
One of Feinberg's concerns is that if the moral duty to rescue is understood as
a duty of
then rescuees are merely beneficiaries and they cannot be said to have a
that the rescuer take action.
n107 In this connection he discusses Mill's version of the traditional distinction
between perfect and imperfect duties. For Mill, a perfect duty is one that
correlates with some particular person's claim, or right.
n108 Thus my duty not to kill correlates with each person's right not to be killed.
A duty of
by contrast, is open-ended--on any plausible account, I can fulfill my duty
while many people remain unbenefited. These unbenefited people are not wronged
by me--I do not violate their rights--for it is in the nature of imperfect
duties that I must make some choices between equally worthy beneficiaries.
Furthermore, if I fail entirely in my duty and assist no one, I violate no
one's rights even though I have acted wrongly. This feature of duties of
does cause some difficulties in the tort context, as we saw.
n109 But Feinberg is concerned about it because on his account of the harm
principle, harms are violations of rights.
n110 On Feinberg's relaxed view of rights, just about any setback to a person's
[*628] interest can count as a violation of a right, so his insistence that criminal
prohibitions are legitimate only when they protect individual rights is clearly
not motivated by some sense that rights protect especially important interests.
His concern is rather that for each crime there must be an identifiable victim,
lest we countenance illiberal
n111 But this is strange; as Feinberg himself acknowledges, there are many crimes
for which there may be no identifiable victim--violation of a criminal
pollution prohibition is one obvious example.
n112 The phrase
"victimless crime" is most naturally understood to refer to crimes condemned as paternalistic,
such as sexual conduct between consenting adults or the use of narcotics, but
this is not the problem with the duty to rescue.
The more important source of Feinberg's desire to understand a duty to rescue
as a duty to prevent harm rather than as a duty to benefit is his apparent
concern that the latter understanding opens the door to legitimate legal duties
beyond the rescue context.
n113 This too could seem a strange concern, because it could seem that rescue
situations are distinctive enough that commitment to the duty to rescue need
imply nothing about what other legal duties of
may be legitimate. However, it is extremely difficult to find a plausible
normative basis for distinguishing rescue contexts from other contexts in which
one person can benefit another. We can seek to define a rescue situation, or at
least offer paradigm examples, but it is very hard to justify treating these
situations as distinctive in any normatively relevant way. This issue is
discussed in some detail in the next Part; I mention it here to show that one
apparently easy way out is not available to Feinberg.
If Feinberg is right, then the significance of the issue of legal duties to
rescue is much less than many of us, especially opponents of the duty, have
supposed. Feinberg wants to present the moral duty to rescue as
. This would mean that we can support its legal enforcement without signing on
to the idea that it is in general legitimate to coerce one citizen to benefit
another. But opponents of the legal duty to rescue are right on this issue--one
cannot support the enforcement of the duty to rescue without admitting the
prima facie legitimacy of legally-enforced
generally. As I said in the introduction, it is this wider implication of the
legitimacy of legal duties to rescue that largely explains the importance of
On an ordinary understanding,
fails to benefit
could, but does not, perform an action that would make
better off. On this ordinary understanding, a failure to rescue is a failure
to benefit. Feinberg insists, however, that our ordinary understanding of the
"to benefit" is ambiguous in an important way. If
fails to halt some deterioration in
well-being and bring
back up to
"normal baseline," then, Feinberg says, this is not really a case of failing
[*629] to benefit, but rather a case of harming, or at least failing to prevent harm.
n114 Genuine--that is,
--benefits are those that
"advance another's interest to a point beyond his normal baseline."
n115 If someone fails to render a genuine benefit, that person has not harmed or
failed to prevent harm, she has merely failed to benefit. The point of all this
is to give an account of
"mere benefit" that marks off rescue contexts as involving something morally more serious
than the mere opportunity for
n116 The idea is that if
life was proceeding just fine prior to some sudden turn for the worse, then
failure to help
back to normal life is not a mere failure to benefit, but rather a failure to
prevent harm (which is morally much more serious). By contrast, we must assume,
has suffered a slow and steady decline, but is about to die (perhaps of
starvation), any benefit from
would count as a genuine--that is, mere-benefit; for in that case
has no normal baseline to be returned to. The discontinuity that appears in
the classic rescue case, and that Feinberg appeals to in his account of failing
to prevent harm, is lacking.
Feinberg is of course free to define terms as he likes, but his idea that the
failure to aid is morally much less serious in the latter case than in the
former strikes me as very implausible. Why are the transient and unexpected
crises of the moderately well-off more worthy of our attention than the chronic
and predictable dangers faced by a badly-off person--even a badly-off person I
may confront with my own eyes?
n117 I doubt that Feinberg would be happy to have the point put this way, but this
is the position to which his attempt to take the duty to rescue out of the
realm of controversy leads.
[*630] We can conclude that a duty to rescue is what it appears to be: a duty of
We can further conclude that this fact should cause a Millian liberal no
concern. Legal enforcement of a duty to rescue is clearly not paternalistic.
Nor need it be motivated by a concern to enforce morality for its own sake--the
case for a criminal duty to rescue does not rest on legal moralism. To be clear
about this point, we must not confuse the fact that a legal duty to rescue
would be the enforcement of a moral duty with the legal-moralist claim that the
aim of the enforcement of the duty to rescue is to enforce morality. It is
obvious enough that most of the criminal law enforces moral duties; the Millian
liberal insists that the purpose of this enforcement is to benefit society, not
to enforce morality for its own sake.
One might wonder at this point where liberty enters into Mill's liberalism as I
(following Hart) have described it. The answer is that for Mill, the value of
liberty, or more precisely of
"negative liberty" (the absence of coercive interference)
n120 lies in its contribution to a person's well-being. Human beings who make their
own decisions about how to live, who (as Joseph Raz puts it) make their own
lives, can be said to be autonomous. Mill values negative liberty as a
contribution to autonomy, which he believes to be a central aspect of people's
Negative liberty contributes to autonomy just because coercive interference
reduces autonomy. The most obvious way in which coercion reduces autonomy is by
restricting those basic requirements of an autonomous life which we may refer
to collectively as a person's positive freedom--the availability of a full
range of options and the ability to make reasonable choices among them.
[*631] Thus we can say that interference with my negative freedom matters because it
reduces my positive freedom, which in turn impairs my autonomy and thus reduces
To bring these points to bear on Mill's harm principle, we can see that
paternalistic coercion may be said both to reduce my positive freedom and to
undermine my autonomy more directly by expressing contempt for my ability to
create my own life. As Mill puts it,
"He who lets the world, or his own portion of it, choose his plan of life for
him, has no need of any other faculty than the ape-like one of imitation."
n123 Mill's opposition to legal moralism is also connected to the value of autonomy
insofar as autonomy is valued as an aspect of human well-being. As we have
said, on Mill's account the purpose of the law is to promote human well-being.
Human well-being is not promoted by the enforcement of morality for its own
sake, unless acting morally is intrinsically valuable for a person. Mill
obviously rejects this view.
n124 Most straightforwardly, we can see that coercion contrary to the harm
principle is bad because it reduces a person's negative liberty (and thus
potentially reduces her autonomy)
while doing nothing to promote the interests of anyone else
. But a duty to rescue can legitimately be enforced precisely for the reason
that, though it does diminish the negative liberty of the person coerced, it
promotes the interests of others. Of course, if the promotion of the interests
of others actually achieved by such a provision were outweighed by the loss in
autonomy caused by the interference with negative liberty, such a provision
would make no sense for a Millian liberal. But as a legal duty of easy rescue
would interfere with liberty only minimally, it is not surprising that Mill
felt it unnecessary to defend his support for it.
Given that there is no reason whatsoever for a Millian liberal to question the
legitimacy of the legal enforcement of a duty of easy rescue, what is
surprising is that so many people, in a Mill-dominated tradition, have thought
that such a duty would indeed involve an unacceptable interference with
liberty. One plausible explanation, mentioned at the start of this Article, is
that the utilitarianism of our tradition has alerted many people to the
possibility that legal duties of
a good deal more stringent than a duty of easy rescue will also survive the
balancing test of liberty lost versus interests of others served.
n127 Given that there are many people in any given country who are capable of being
[*632] benefited to a significant degree, it may turn out that this balancing test is
satisfied even when the losses to the people coerced--if not of liberty, then
at least of money--are very great. I return to this issue in the next Part.
Another important explanation is that our tradition is not just Millian; it is
also, to a significant degree, libertarian or
n128 And the libertarian understanding of the right to liberty raises entirely
different problems for the duty to rescue. Before turning to that other strand
n129 however, we need to discuss the views of a prominent opponent of criminal
duties to rescue who, though he disagrees with Mill on the proper aims of the
criminal law, is broadly Millian in his understanding of the value of liberty.
This will also allow us to investigate a third explanation for the otherwise
surprising opposition to legal duties to rescue--many people believe, contrary
to what I have asserted, that positive duties in their very nature pose rather
significant threats to individual liberty understood in Millian terms.
In his recent book,
, Michael Moore defends the legal moralist view of the criminal law: The
purpose of the criminal law is to enforce morality. He reaches this position by
way of his defense of a retributivist theory of punishment.
n130 He thus departs radically from Mill's utilitarian account of the proper aim of
the criminal law as well as from the broader textbook account of that aim as
the prevention of social harm--or, as I would rather put it, the promotion of
the social good.
n131 But Moore does believe that there are countervailing considerations that
should temper the aim of enforcing morality. One important consideration of
this kind is the general value of negative liberty.
Moore's idea is not that individuals have a
to be free of legal coercion. Moore accepts Ronald Dworkin's argument that any
such right, so often overridden
[*633] by legitimate law as it obviously is, is too weak to be worthy of the name.
n133 In so doing, Moore also embraces Dworkin's understanding of the term
"right," an understanding that is very different from that of Feinberg mentioned above.
"right" in the stronger and more standard sense of a constraint on what may be done to
a person even for the sake of promoting general welfare.
n134 Rights in this strong sense thus correlate with negative duties whose moral
importance cannot be reduced to a concern with the promotion of human
well-being, or indeed even with the minimization of rights violations. As
Robert Nozick illustrates, if you have the right not to be assaulted, I am duty
bound not to assault you even for the sake of preventing several other assaults
to other people.
n135 In discussions of rights that follow, I too will have in mind this strong
sense of the term.
So Moore accepts that there is no moral right to negative liberty. The conflict
between liberty and the aim of the criminal law is rather a conflict between
two values; the value of enforcing morality and the value of negative liberty.
Moore is thus content to employ the traditional idea of a
"presumption of liberty" in this context, which is simply that infringement of negative liberty is
always regrettable just because freedom from coercion is always valuable, and
so any criminal duty must have enough to be said in favor of it to outweigh the
loss of liberty it brings.
n137 Freedom from coercion is valuable, according to Moore, because positive
liberty--by which he means strictly the range of options available to a
person--has intrinsic value.
n138 Moore does not go on to defend the value of positive liberty in terms of its
contribution to autonomy, but in holding that the value of negative liberty
lies in its instrumental contribution to positive liberty, he remains squarely
in the Millian tradition.
[*634] Moore believes that negative criminal duties will generally have a better
chance of overcoming the presumption of liberty than positive duties. This, for
Moore, explains and justifies the traditional position on the duty to rescue in
common-law countries and the reluctance to embrace positive criminal duties of
any kind. There are two mutually enhancing aspects to the argument. First,
Moore claims that negative moral duties are generally more important, or
weighty, than positive moral duties, and thus the legal moralist case in favor
of their criminalization is stronger. Second, he claims that positive duties
generally interfere with liberty more than do negative duties.
Neither of these claims seems at all plausible to me. It is a well-known
assumption of commonsense morality that it is worse to cause some harm rather
than allow it to happen; thus it is worse to drown a baby than to allow it to
n141 But it does not follow that all negative duties are morally more stringent
than all positive duties. The duty not to lie is obviously less stringent than
the duty to save a drowning child at low cost, or, indeed, to take care of
one's own children, even at high cost.
Responding to this point--mindful especially of the strong positive duties
people are commonly taken to have toward their children--Moore suggests that
the relevant distinction may be that between agent-relative and agent-neutral
duties, the former being more stringent than the latter. This widely-used
distinction, originally formulated by Thomas Nagel, can be drawn in various
n143 I find the following account congenial: agent-neutral duties give all of us
the same aim; agent-relative (or
"deontological") duties give different agents different aims.
n144 The paradigm agent-neutral duty would be a (positive) duty of
paradigm agent-relative duties would be the (negative) duties not to violate
n145 But some positive duties, such as the duty of a parent to take care of his own
children, are also agent-relative. Moore's suggestion is that while the
negative/positive distinction roughly marks out the more stringent from the
less stringent duties, the distinction between agent-relative and agent-neutral
duties does so more accurately.
There are several problems with Moore's shift to the distinction between
agent-neutral and agent-relative duties. In the first place, the moral duty to
rescue (which, as a duty of
is an agent-neutral duty) is widely
[*635] thought by the critics of legal duties to rescue to be very stringent indeed
(recall the moral monsters). Second, the shift of focus to the contrast between
agent-relative and agent-neutral duties makes it quite clear that the
negative/positive distinction in itself carries no weight on the question of
the stringency of moral duties. This is an uncomfortable result for a theorist
who has put so much weight on the importance of the act requirement in criminal
law. Indeed, because the distinction between agent-relative and agent-neutral
duties does not track action and inaction, we are left with no account of
we should believe the former to be more weighty than the latter--Moore simply
asserts that this is the case.
n146 Without some account of why we should take this view, legal moralists must
rely on case-by-case commonsense intuitions about which particular moral duties
are, in fact, more important.
Moore is actually prepared to grant that the positive (and agent-neutral) moral
duty to perform an easy rescue is more weighty than some negative (and
agent-relative) duties which are nevertheless criminalized. This is not fatal
to his position, he believes, because there is always the other aspect to his
argument--that concerning the impact on liberty. In discussing Samuel Freeman's
claim that there is much greater reason for the legal moralist to criminalize
failing to rescue a child than stealing her purse,
n147 Moore grants that the moral duty to save may be more important than the moral
duty not to steal, yet insists that the
"liberty differential is still there."
n148 But the claim that
"it diminishes liberty less to prohibit theft than it does to require
n149 is preposterous. Recall that for Moore the importance of negative liberty lies
in its contribution to positive liberty, or the range of available options for
choice. As the opportunity to effect an easy rescue never arises in the vast
majority of people's lives, the thought that there is a serious interference
with people's opportunities for choice here is so evidently bizarre
n150 that it calls for some explanation. One possibility, once again, is that Moore
and others who make this claim must have other, more demanding, duties of
in mind--the utilitarian duty always to perform that act which will maximize
expected aggregate well-being does indeed appear to be very invasive of
positive liberty. Be that as it may, the fact remains that it is not credible
to say that the Rhode Island provision constitutes more than a de minimis
interference with the liberty of Rhode Islanders.
The claim that a duty not to steal is in general
very invasive of liberty is also clearly false. Only well-off people who
rarely seriously contemplate the option of theft would ever be tempted to think
as much. There is, indeed, an obvious and simple political point to be made
about the claim that negative
[*636] duties as a class interfere with liberty less than positive duties as a class:
The richer one is, the more plausible this will seem; the poorer one is, the
less plausible it will seem. If we focus our minds on the options of a
destitute and uneducated person in the United States, and think of the entire
range of negative criminal duties lined up against her, the suggestion that she
remains free of significant interference with her liberty, because, thank
goodness, she has no positive duties imposed on her, should strike us as absurd.
It is important to remember that I make these responses to the second aspect of
Moore's argument without challenging his view that negative liberty is always
valuable because its infringement leaves us with a smaller range of options for
choice. Thus, as against two of his critics, I agree with Moore that the value
of liberty is not dependent on my desiring the options in question, and agree
also that there is value in being able to choose to do things that may be wrong
n151 Furthermore, though I am inclined to think that positive liberty is
nevertheless in some sense
valuable the more the relevant options matter (either subjectively or
n152 we can for the sake of argument grant Moore that this is not so. The rejection
of the idea that positive duties generally interfere with liberty less than
negative duties does not at all depend on disagreement with Moore's conception
of liberty; it depends rather on two simple points about the way that different
duties impact on a person's opportunities for choice.
To repeat and expand on these points, the first is that the extent to which a
positive duty will interfere with liberty--diminish the number of options for
choice--depends crucially on the pervasiveness of the duty's application to a
person's life. The popular claim that positive duties are terribly detrimental
to liberty, and much more so than negative duties, seems to be based on the
thought that while a negative duty merely cuts off one option, a positive duty
cuts off all options
one. But this abstract way of putting the point is very misleading precisely
because it focuses on the single moment when the duty must be performed, rather
than a person's life as a whole. Life is full of positive duties that hardly
interfere at all with our total range of options for choice. A positive duty to
take out the garbage once a week is very different from the positive duty to
become a garbage collector. Likewise, a duty to perform an easy rescue would
reduce most people's opportunity sets not at all; a duty to spend every day
standing by the river watching for trouble would of course be very detrimental
to liberty, but we shouldn't confuse the two duties.
The second point is the more important of the two, as it reveals in stark form
the ideological nature of the traditional aversion to positive duties
as opposed to
negative duties. One of the things that most determines a person's set of
possible options is the amount of resources at her disposal. The familiar
[*637] negative duties prohibiting attacks on person and property are extremely
detrimental to the positive liberty of those who lack resources; indeed, the
coercive prohibitions of theft and violence are far more restrictive of the
liberty of the destitute than any positive legal duty anyone has ever seriously
The rejection of the claim that negative duties are less detrimental to liberty
than positive duties obviously has force beyond the context of Moore's legal
moralist theory of the criminal law. It applies to any account of the criminal
law that includes, as a constraint on whatever aim the criminal law is taken to
have, a concern for the value of negative liberty conceived of as a means to
To find some general reason of political morality to oppose positive legal
duties, we therefore need to turn from Millian liberalism to the libertarian
version of rights-based liberalism that in the English-speaking world derives
primarily from standard interpretations of Locke.
n153 The central claim of libertarian political theory that touches on our topic is
that all persons have a right to negative liberty. Here, freedom from coercion
is not understood as a value, much less an instrumental value in the service of
positive liberty, but rather, simply, as a natural right. Indeed, many writers
in this tradition emphasize that in their view the commitment to negative
liberty implies no commitment at all to any form of positive liberty.
n154 What makes freedom from interference at the hands of other human beings so
important? Why is it prima facie so much worse if I am prohibited from
proceeding along a road by rocks that you put there rather than by ones that
fell? The idea is simply that a special wrong is done when people coerce other
people. Freedom is a right, and rights, in the strong sense of the term that we
are using, are all about constraints on what people can do to one another.
Thus the heart of the libertarian account of law's legitimate content is the
constraint that individuals' rights must not be violated by the legal
regime--especially not individuals' rights to be free of coercion. Of course,
it is obvious to everyone that coercion is, as Hayek says,
"unavoidable," for the very reason that coercion or its threat is necessary to prevent
n156 This leads straight to the idea that it is no violation of my right not to be
coerced for me to be coerced into not violating someone else's right not to be
coerced. But libertarians usually go much further and allow that coercion is
legitimate for the sake of protecting the whole range of rights that are
generally taken for granted in a
[*638] market economy--notably a right against all forms of aggression, a right to
private property (as somehow identified via principles of acquisition and
n157 contractual rights, and the right to be free of fraud.
n158 Thus libertarians believe that coercion, and rights infringement generally, is
justified for (and only for) the sake of protecting a certain familiar list of
individual rights. When it is recalled that on the account of rights we have
been employing, and which is employed by contemporary libertarians, there is no
such thing as a positive right to aid, it is not hard to see how libertarians
reach the conclusion that
"the state may not use its coercive apparatus for the purpose of getting some
citizens to aid others."
The most prominent opponent of a tort duty of easy rescue among contemporary
legal theorists has been Richard Epstein. In his early writings on the topic,
the main thrust of his opposition was that affirmative legal obligations
constituted an illegitimate interference with individual liberty--understood
along libertarian lines.
n160 As such, the rejection of duties of easy rescue came with radical
implications. Because one of the rights on the libertarians' standard list is
the right to private property, redistributive taxation turns out to be
illegitimate because it infringes on the right to private property while
protecting no other right. In his book
, Epstein does not shrink from this conclusion.
n161 Nevertheless, Epstein has long believed that there are areas of the law that
cannot be made compatible with libertarianism and must instead be subjected to
[*639] a kind of supplementary utilitarian analysis.
n162 Furthermore, in
, Epstein explicitly rejects Nozick's thoroughgoing libertarianism on the
ground that he cannot see how the state can be justified if all forced
exchanges are ruled out.
n163 Once rights can be infringed for the sake of the general good, however, the
cat is out of the bag, and it not surprising that shortly after the publication
Epstein announced that the traditional natural rights view he had been
espousing was defensible on utilitarian grounds.
Even though Epstein no longer holds to libertarianism at the foundational
level, an inchoate form of the libertarian view would appear to be widespread
among opponents of legal duties to rescue
n165 and thus it is important to discuss it a little further. One point has already
been alluded to: If one argues against legal duties to rescue on libertarian
grounds, one had better be prepared to reject as illegitimate whole swathes of
familiar legal regulation. A possible way of containing these radical
implications would be to tamper with the libertarian's typical list of rights.
Many of the most radical implications of the doctrine flow from the supposed
absolute moral right to property; if this right were removed from the list,
redistributive taxation would come out as legitimate while forced labor and
legal duties to rescue would remain illegitimate, for violating the right to
negative liberty. Given the difficulties facing accounts of an absolute moral
right to private property, and given the overwhelming normative implausibility
of the libertarian's traditional opposition to any form of redistribution away
from market outcomes, this would seem a promising way to save the core idea of
this form of liberalism. However, it is not so easy to deradicalize
libertarianism. Under this view, noncontractual positive duties to act will
remain illegitimate; the duty to file a tax return will be rejected even if
[*640] for the sake of redistribution will not. And there are of course all the other
noncontractual positive duties we have mentioned along the way--to register for
the draft, to
"rescue" one's aged parent, among others; all these legal duties will fall along with
the duty to rescue. Libertarianism remains radical so long as it insists that
the only permissible ground for coercion is to prevent the coercee from
infringing the negative rights of another.
The fact that libertarianism is a radical view does not disprove it, of course;
it merely makes clear the high stakes of any libertarian argument against the
duty to rescue. Whether libertarianism is a plausible political theory is
obviously not an issue we can settle here. What is important to note, however,
is that such a discussion would focus on, first, the plausibility of the
libertarians' list of rights, and, second, the claim that these rights can
justifiably be infringed only in order to prevent the coercee from infringing
the rights of others.
n166 This description makes it clear that despite libertarian rhetoric, the
evaluation of libertarianism has little to do with the question of whether one
is in favor of more or less liberty; it has rather to do with a very particular
focus on the moral force of certain rights. Whether it is therefore appropriate
to call libertarians
"rights-fetishists" depends upon your point of view, but it does seem right to say that
libertarianism is a misleading label.
To see this more clearly, consider the libertarian slogan that liberty may be
limited only for the sake of liberty; this suggests that the whole of political
morality can be understood in terms of negative liberty.
n168 As Sidgwick characterized the view one hundred years ago (commenting
optimistically that it was
"now perhaps somewhat antiquated"),
"all natural Rights, on this view, may be
[*641] summed up in the Right to Freedom; so that the complete and universal
establishment of this Right would be the complete realization of Justice,--the
Equality at which Justice is thought to aim being interpreted as Equality of
n169 The trouble is simply that so much of what the libertarian list of rights
protects has nothing to do with freedom--especially not on the negative
conception of freedom which libertarians insist is the only conception they
mean to employ.
n170 The most obvious, and politically most important, illustration of this once
again concerns the right to private property. In what sense is the prohibition
of theft a case of coercion for the sake of protecting liberty? But the point
holds even for the right against aggression, which cannot be equated with the
right to be free from coercion.
n171 For example, it is not in the least clear why prohibiting
water so as to cause
occasional headaches is a case of coercing
for the sake of protecting
liberty; it is not, indeed, even clear why prohibiting
is a case of limiting liberty for the sake of liberty.
n172 If there is a single conception of freedom animating libertarian political
morality at all, it is obviously a strongly moralized conception.
n173 The idea must either be that any violation of
rights is by definition a restriction of
liberty, and/or that whenever
is prohibited from violating
liberty is by definition not thereby restricted.
n174 But such definitions
[*642] are clearly ideological because the resulting notion of freedom bears almost
no relation to either negative freedom or any version of positive freedom.
This concludes our survey of the various versions of the claim that duties of
easy rescue constitute an excessive degree of interference with liberty. We
have found no version of the claim to be plausible. It is true that libertarian
political morality, unlike Millian liberalism, does provide
reason to reject legal duties to rescue (along with all noncontractual
positive duties)--such duties are said to violate our rights. But evaluation of
libertarian theories of political morality is, as I have said, beyond our
scope. As such theories have implications far beyond the issue of the duty to
rescue and yet are also--to my mind at least--deeply implausible, this is not
such a great loss to our evaluation of the case against legal duties to rescue.
Nevertheless, it must be acknowledged that the discussion that follows will be
beside the point for committed and consistent libertarians. What even they must
admit, however, is that their opposition to legal duties to rescue is not
grounded in a concern with individual liberty.
C. CORRECTIVE JUSTICE
With the exception of Epstein's libertarian opposition to tort duties to
rescue, the arguments so far discussed in this Part have concerned the criminal
law. Arguments in the tort context fit within the structure we have been using:
There must be some premise about the aim of tort law and perhaps also a premise
about the constraints on legitimate pursuit of this aim. In a general sense, we
might expect the case for criminal duties to rescue to face stronger
opposition, simply because criminal conviction is a more serious matter than a
judgment of civil liability. This is not necessarily so, however. Indeed, it is
the potential severity of a finding of civil liability for a failure to rescue
that leads some people, myself included, to question the wisdom of a tort duty
Furthermore, there are a range of views about the proper
of tort law, and not all of them allow for a duty to rescue. If the aim of
tort law is to promote general well-being, then the desirability of a duty to
rescue depends entirely on the question of whether, all things considered, such
a duty would promote general well-being. One might also adopt a legal moralist
conception of tort law according to which civil liability is simply a secondary
way of enforcing
[*643] morality; if so, there should also be no principled bar to the duty.
n176 Matters are otherwise if the preferred account of the aim of tort law appeals
to some notion of corrective justice--here understood as an independent,
normative account that is meant to be independently appealing, rather than as
an interpretation of the normative structure of tort law as it now is. Most
theories of corrective justice do place special emphasis on misfeasance as
opposed to nonfeasance and so would reject civil liability for failure to
n177 As with the libertarian theory of political morality, evaluation of the
various possible accounts of corrective justice is beyond my scope here.
n178 Therefore all I can do is acknowledge that there are specific normative views
concerning individuals' responsibility for each others' fates that rule out
requiring a non-rescuer to pay damages to the victim.
Like adherents of libertarian theories of political morality, adherents of
various theories of corrective justice rule out legal duties of
as a matter of principle. Such views therefore do not reach the issues that
will occupy us for the remainder of this paper, which concern the difficulties
that arise once legal duties of
get a foot in the door. Very many of us, however, think that the aim of tort
law should be to promote the social good, and most of us take this same view
(or else the legal moralist view) of the criminal law. Thus I think most of us
need to take seriously the following problems.
[*644] III. THE (UNAVOIDABLE) PROBLEM OF DEMANDS
In a few pages in his notes to the Indian Penal Code, Macaulay presents in
lucid form the most serious problem raised by duties to rescue. The problem has
two parts. The first is that it is very difficult to find defensible normative
grounds on which to distinguish rescue situations from nonrescue situations. As
a descriptive matter, rescue situations are sudden, unexpected emergencies
involving severe suffering or threat to life. But Macaulay rightly wondered
how, once duties to render aid in such situations are accepted, it could be
consistently maintained that the rich have no duty to save beggars in Calcutta
from slow but certain death by starvation.
n180 The second problem is that in rescue and nonrescue situations alike, there is
the potential for enormous loss to the benefactor. This second problem is
sometimes referred to, following Macaulay's own words, as the
n181 I call it the problem of demands. A duty to render aid is, as I have said, a
The duty of
embraced by utilitarians imposes demands with no limit: We must go on
benefiting others until the point where further benefits will burden the donor
as much as they will benefit the donee. For even moderately well-off people,
compliance with this duty would make for a
change of lifestyle. But if this optimizing duty of
is to be rejected as obviously absurd in virtue of the demands it makes, what
more reasonable duty of
should we embrace?
Before expanding on these problems, we need to answer a crucial question. As
presented, Macaulay's problems are problems of moral/political theory. What is
their relevance to the issue of the legitimacy or desirability of legal duties
to rescue? Why do we need to be able to have a complete account of the morality
before we can conclude that there is no objection to at least the kind of
minimal legal duty of
found in Rhode Island?
The first point to make is that even if these two problems of moral theory were
irrelevant to legal theory, strictly analogous problems would be relevant.
Suppose a legal theorist were to hold that the criminal law aims to promote the
social good and that criminal duties such as the Rhode Island provision do this
and therefore should be enacted. Such a theorist must still answer this
question: Why stop with minimal duties to rescue? Why not enact more demanding
legal duties of
as well? It will not do to reply that more demanding duties may or may not be
called for, but because they seem to be controversial, we can ignore them.
Despite what some defenders of legal duties to rescue say, this
[*645] problem is not one that can be solved by careful drafting.
n182 In the absence of any principled basis for drawing the line of legally
at a particular point, the legitimacy of any actual rescue provision, no
matter how minimal, will remain in doubt, and concerns about later extensions
of the legal duty of
So at the very least problems analogous to Macaulay's must be dealt with in any
principled legal-theoretic defense of legal duties to rescue. But I want to
make a stronger claim than that. Legal theory itself must confront Macaulay's
as problems of moral theory
. This claim is significant, because it renders inadequate one possible
principled basis for a strict limit on the extent of legally required
As we will see in Part IV, there are very important
reasons why it would most likely not make sense, even for a utilitarian, to
support general (that is, not limited to rescue contexts) and highly demanding
legal duties of
Thus, a utilitarian defender of Rhode Island-style duties to rescue could
claim to have the very best kind of reason to support the enactment of only
minimally demanding and rescue-specific legal duties of
Minimal duties, focused in particular on emergencies, would do more good than
harm, but general and more demanding legal duties of
would do more harm than good. A practical explanation for limiting duties to
rescue, however, will not be adequate if a legal theorist must
be able to explain why, on her view, people do not face extremely demanding
moral duties of
The practical reasons that recommend only limited legal duties of
do not apply in the moral sphere.
Why does a defense of legal duties of
require a plausible account of moral duties of
as well? First, suppose we take the legal moralist view of the criminal law
(for the remainder of this section it will not be necessary to distinguish
between the criminal law question and the tort question). On that view, the
purpose of the law is to enforce morality, so it is abundantly clear that the
question of moral theory is a question for lawmakers as well. I think it fair
to assume that Macaulay operated with a legal moralist view of the criminal
law. And it was his lack of comfort with the idea of enforcing one particular
while not being able to say why other more stringent duties of
should not, on pain of inconsistency, also be enforced, that led him to
exclude the duty to rescue from the code. So any legal moralist who wishes to
be able to give a principled defense of legal duties to rescue needs to take
Macaulay's problems seriously.
The more important and interesting point, however, is that the theorist who
holds that the purpose of the law is to promote the social good must also take
Macaulay's problems seriously. Legal theorists working in a normative vein
provide, naturally enough, normative arguments about what the content of the
[*646] law should be. But in reaching their conclusions about the preferred content
of certain specific legal rules, they often do not ask whether the fundamental
principles to which their arguments appeal are plausible
. In particular, they do not ask whether those fundamental principles, used in
some specific context of institutional design, are plausible in other areas of
institutional design; nor do they ask whether those principles are plausible as
principles for personal conduct. The most obvious example of this phenomenon is
found in the frequent avowal of utilitarianism as the correct account of the
aim of promoting the social good, and thus the correct principle to guide
institutional design, by legal academics who would certainly reject the idea
that they must live their lives by the extremely demanding utilitarian
criterion of personal conduct.
There is, I believe, no justification for the assumption that principles that
govern specific questions of institutional design need not be plausible
generally, including at the level of personal conduct. The issue of the
connection between norms for institutional design and norms for personal
conduct is important and controversial, and I can do no more than briefly set
out my own view.
Traditionally, moral and political philosophers believed that the very same
fundamental normative principles governed both institutional design and
personal conduct. Thus, for example, though classical utilitarians such as
Bentham were primarily interested in institutional design, this focus was due
more to a belief in this issue's practical importance and to a skepticism about
human motivation than to some sense that the principle of utility did not also
apply directly to people.
n186 The idea that different principles govern on the one hand institutional design
(the subject of justice) and on the other hand personal conduct (the subject of
morality) is due to Rawls. This dualist view of political morality, as I call
it, is a central theme of both
A Theory of Justice
. But I see no good reason to follow Rawls here, and one very good reason not
n187 The main problem with dualism emerges when we ask what connects just
institutions to actual people. There must be such a connection:
[*647] Even if different normative principles govern institutional design and
personal conduct, some normative principle must nevertheless connect
institutions to people--institutions do not get to be just all of their own
accord. Rawls posits a
"natural duty of justice"--people have a duty to sustain and promote just institutions.
n188 This provides the necessary normative link between people and institutions,
but it also brings out what I believe to be the deep implausibility of the
dualist approach. On this view, the obligations people have with respect to
matters of justice are always and only to perfect institutions, even when the
of such institutions could be better promoted in some other way. But I believe
we have little interest in institutions for their own sake; we are interested
in them primarily for what they can do. If, in a given situation, trying to
make institutions more closely correspond to some criterion of justice will
achieve less than some direct attempt by people to, say, promote well-being (if
that is the criterion for institutional design), then people should pass
institutions by. The overwhelming practical importance of institutions in
achieving the aims of justice in the typical case should not blind us to the
fact that what matters to us, ultimately, is not whether our institutions are
just, in the sense that they achieve our aims, but rather simply the extent to
which those aims are achieved, however that might be done.
Thus, whatever fundamental normative principles we use to evaluate institutions
must also be used to evaluate ourselves. We cannot breezily evaluate legal
institutions such as tort law or the criminal law with the utilitarian
criterion without thinking about the implications of that criterion in the
realm of personal conduct. If the utilitarian criterion is correct, it is
correct generally; we do not have some fetishistic attachment to utilitarian
institutions and no direct concern with the promotion of human well-being
independently of institutional design. Therefore, though there are, as I have
said, good utilitarian reasons to limit any legal duty of
to the kind of minimal duty found in Rhode Island, that does not mean that a
utilitarian supporter of such provisions can ignore Macaulay's problems. If the
only way to justify a legal duty to rescue is by appeal to a fundamental
principle that we reject as unacceptably demanding as a principle of personal
conduct, the case for the legal duty must fail.
This first reason illustrates why a legal theorist who holds that the aim of
the law is the promotion of the social good cannot ignore Macaulay's two
problems. It is rather fundamental, and does not turn on any particular
commitments of political morality. A second reason why the problems may be
unavoidable is more specific and substantive. A theorist who holds that the aim
of the law is the promotion of the social good might hold that there is
nevertheless an important constraint on the pursuit of that aim: The law cannot
promote the social good by requiring people to sustain sacrifice beyond what
they are morally required to
[*648] sacrifice voluntarily.
n189 This would be a genuine constraint on the law, so that, if it is accepted,
legal duties of
that are more demanding than the duties of
found in the best view of morality would be illegitimate.
Having shown the importance of Macaulay's problems for the issue of legally
I return now to setting out those problems in greater detail. The first
problem is that of explaining why
should be required in rescue situations but not for the sake of meeting
desperate needs that are mundane and chronic.
n190 None of the criteria that one might suggest to mark out rescue cases as
generating special and especially stringent obligations of
seems on reflection to carry any weight. In his recent book
Living High and Letting Die
, Peter Unger devotes considerable attention to this issue. He considers nine
possible factors that might mark out rescue cases as having special normative
significance, and rejects all of them.
n191 The force of Unger's discussion comes not just from direct reflection on the
plausibility of the various possible criteria, but also from his ability to
construct, for each factor, a case that seems to prompt the wrong intuitive
response. For example, to those who suppose that physical proximity is what
underlies the stronger obligation to help in a rescue case, Unger offers the
case of a motorist who, rather than confronting a bleeding victim on the side
of the road, hears over his CB radio that a twenty-mile detour will take him to
a bleeding victim. Unger notes that people he has asked tend to feel that it
would be just as wrong to fail to rescue in both cases.
n192 But if a rescue situation is marked out by the factor of proximity, this is
the wrong answer.
Unger's eventual conclusion is that in most of the cases where we readily
accept a stringent obligation to meet the needs of others, those needs are very
conspicuous to the person in question.
n193 He further claims that while
"our basic moral values" support a stringent duty of
for rescue and nonrescue cases alike, our
"futility thinking"--roughly the thought that anything we can do to help will be a mere drop in
the ocean--blinds us to this obligation in most cases. In cases like the
typical rescue case, however, the needs of the victim are
[*649] sufficiently conspicuous to us to break the hold of the futility thinking.
n194 The upshot of this is that there is nothing morally special about rescue
cases; the widespread assumption to the contrary is due entirely to the
psychological effect that a dramatic confrontation with a particular person's
needs has on our futility thinking.
I find Unger's exhaustive argument against the existence of a distinct
normative category comprising rescue cases convincing.
n195 At least, we can be confident that there is no simple or obvious way to mark
out a distinct principle of
that concerns rescue situations only.
In any case, even if we could explain the special moral significance of rescue
cases (thus disposing of Macaulay's first problem), Macaulay's second problem,
the problem of demands, would remain. This problem is most obvious for a
general requirement of
(that is, one not limited to rescue contexts), but an open-ended duty to
rescue could also generate extreme demands. We can ignore any duty to rescue
that requires an agent to sustain expected burdens greater than the expected
benefit of her act. But even a duty that requires a person to act so long as
the expected benefit is substantially greater than the expected burden to her
can yield extreme demands. Thus a boat owner who would not normally take a
fifty-fifty risk on her life by going out into the storm would nonetheless be
required to do so by such a duty if there is the same chance of rescuing all of
the ten people from the capsized yacht.
Rescue situations can also generate extreme demands of other kinds, not just
threats to the rescuer's life. Suppose that a group of astronomers plan an
[*650] observation that must take place at sea at a particular date and time, and
spend many hundreds of thousands of dollars in preparation. Just as the
astronomers are about to take their readings, they receive a Mayday signal; if
they rescue the sailors in distress, the project will come to nothing.
Clearly, an unlimited duty to rescue can generate extreme demands.
Nevertheless, even more dramatic demands of
emerge with principles of
that are not restricted to rescue contexts. The utilitarian's optimizing
requires of ordinarily well-off people such enormous sacrifice that it is
widely regarded as absurd.
n198 There is no need to dwell on the details of the sacrifices that the world's
moderately well-off must sustain before further beneficial efforts would do no
more good than harm. But one point must be emphasized. It is not plausible to
say that the losses that would flow from compliance with the optimizing
are all, or even most importantly, losses in liberty.
n199 The problem of demands is simple: Compliance with the optimizing principle of
would have a terrible impact on the agent's resources (including time) and
thus on her well-being.
n200 It is true that if one were required to give all one's money away, this would
greatly reduce one's positive freedom; but we must remember that if positive
freedom matters, it matters just because it is an aspect of human well-being.
The demands of utilitarianism fall heavily on all aspects of well-being,
including the more mundane, but nevertheless central, goods of leisure and
pleasure. Thus a claim that utilitarianism must be rejected because it
interferes so heavily with freedom in particular must depend on the idea that
the most important element of human well-being is the extent of the range of
choices for action--either as an intrinsic good, or as a means to autonomy.
Because this idea is so controversial, it is misleading to describe the costs
to complying agents solely in the lofty language of liberty. The degree to
which one makes one's own life surely does matter; but it is not the only thing
What, then, is to be done? If commitment to a duty to rescue brings with it a
commitment to a general moral requirement of
and if the most
[*651] straightforward general moral requirement of
is the optimizing requirement of the utilitarians, it would seem that
commitment to legal duties to rescue comes at the price of embracing the
allegedly absurd demands of that requirement. What is needed, obviously, is an
account of a general principle of
that does not impose absurd demands. This, it turns out, is an extremely
complicated matter. All I will do here is lay out some possibilities. If one of
these approaches to the morality of
is plausible, it may be possible to defend legal duties of
without committing oneself to an implausible personal morality.
The first possibility is to hold tight with the utilitarians' optimizing
extreme demands and all; with this approach we simply deny that a moral
principle making such extreme demands is absurd and must be rejected. This
extremist view is not unrepresented among philosophers.
n201 A second (and more popular) view responds to the extreme demands of the
by embracing a
principle--one that gives out at a certain point when the demands get too
great (we can call this a
"limited principle of
n202 Though natural, this response faces severe problems.
n203 In the first place, Macaulay's worry about the basis on which the line is
drawn remains. It is hard to come up with criteria for the limit to required
that do not seem simply to track our natural tendency to think that the way we
live now cannot be too bad. More fundamentally, it is very hard to know how we
are to assess the demands of a moral requirement. What baseline should be used?
In particular, how are we take into account the fact that, though I might be
worse off if I take some beneficent action than if I do not, I am still very
much better off, overall, because of the
of my community that is expressed in the form of social and legal institutions
that do me so much good? Once we take into account how much good comes to us
from the compliance of other people it is hard to see any moral principle as
demanding, in an overall sense, at all.
In the face of these problems, it is worth exploring a third possibility: The
problem with the demands of the optimizing principle of
is not that they are extreme, but that they are, in a special way, unfair.
n204 For it is
[*652] characteristic of the optimizing principle that in situations of partial
compliance, where not everyone is acting as a good optimizer, those who
complying must take up the slack left by the others. Thus, a complying person
may have to do more than she otherwise would not just because of natural
disasters, for example, but also because others are not doing what they ought
to do. This requirement that one shoulder not only one's own share of what
should be a collective effort but also the shares of others can be regarded as
unfair. If this is right, what we can call a
"collective principle of
" according to which people must sustain only that amount of sacrifice they
would be required to sustain if everyone were doing their part, becomes
The collective principle of
faces its own problems, of course. Most simply but also most fundamentally,
not everyone agrees that the reason why the demands of the optimizing principle
are absurd is that they are unfair in situations of partial compliance. Even
under full compliance, the optimizing principle can make very high demands, and
this can still seem objectionable. At the same time, in some contexts doing
one's fair share may not seem to be enough.
I have introduced the three responses to the problem of demands--the hold-tight
utilitarian response, the response of the limited principle of
and the response of the collective principle of
--not to defend one of them over the others, but to lay out our options.
Fortunately, when we turn back to the question of the desirability of legal
duties to rescue, each of the three moral views yields the same verdict: Rhode
Island-style duties to rescue may be desirable but more extensive legal duties
are very unlikely to be.
IV. AN OVERLAPPING LEGAL CONSENSUS AMONG PLAUSIBLE MORAL VIEWS
A. A CRIMINAL DUTY
I favor the solution to Macaulay's problem of demands that is offered by the
collective principle of
But this difficult issue of moral theory does not need to be finally resolved
before we can justify some legal duties to rescue. As I have argued, legal
theory must address the problem of demands as it emerges in moral theory
because the fundamental principles we use to guide institutional design must
also be plausible in the realm of personal conduct. Thus if there were no
plausible alternative to the utilitarian's optimizing principle of
defense of Rhode Island's criminal duty to rescue would require the acceptance
of that principle's truly radical implications in the personal realm. But there
are alternatives to the optimizing principle, and at
[*653] least one of them, the collective principle, seems to me to be plausible. And
both of the two alternatives agree with the utilitarian verdict that legally
should be restricted to criminal duties to rescue with strictly limited
demands. As luck would have it then, legal theory can to a certain extent
side-step Macaulay's problems after all.
It is not hard to see that a general criminal duty of
(one not limited to rescue contexts) would make no sense, even for
n206 A general duty of
requires the promotion of overall well-being. It is obvious enough that if the
state has this aim, the best way to discharge it is by setting up institutional
structures that transfer resources, provide education and health care, and
other social services--not by criminally enforcing
"a thousand points of light." Of course if the state does
promote general well-being, a moral principle of
will require (better-off)
to do so.
n207 But our question is now not what people should do, morally speaking; our
question is whether there should be a legal requirement of general
On the view that the aim of the criminal law is to promote the social good,
the state must obviously take a holistic view of its responsibilities--it
should not try to achieve through the criminal law what can be better achieved
in other ways. And on the legal moralist view too, if the state can in effect
discharge people's responsibilities of
for them via the tax and transfer system or other mechanisms, and indeed do a
better job of it than people could themselves, then there is no call for
enforcing individuals' duties of
n208 The general point is that overall well-being is best promoted through the
organized and coordinated efforts of well-trained people.
n209 This holds true for most emergency situations as well--that is why we have
police forces and fire departments. The only kind of situation in which the
involvement of the ordinary civilian is indispensable, even under an ideal set
of institutions, is one calling for immediate attention by whomever happens to
be both aware of the problem and able to act.
n210 Often, all that is required in such cases is that the bystander summon trained
experts, but because we cannot and do not want to have policemen on every
corner, this summoning function is very important indeed.
These considerations suggest that criminal duties of
should, like the Rhode Island provision, be limited to emergency scenes--to
[*654] where action by a particular private citizen will make all the difference. But
though emergencies are the only
of case where the involvement of private individuals is indispensable, it is
surely true that individuals can do a lot of good in other contexts too, even
under ideal institutions. Should not a utilitarian therefore favor a mopping-up
legal duty of
--with the defense available that no liability attaches where a given
beneficent course of action would not have been necessary if the state had been
doing its beneficent job as well as is possible? Such a legal duty does not
require people to do what could better be done by the state, but it is
objectionable nevertheless for the way in which it would operate in practice. A
mopping-up duty would put an inordinate informational burden on individuals,
who would constantly need to be assessing whether there is good to be done that
could not, ideally, be better done by the state. Restricting the application of
the duty to cases in which a person has actual knowledge that some
cost-justified benefit can be conveyed would reduce these burdens, but the
determination of whether the benefit could, ideally, have been better handled
by the state would still be extremely burden-some.
n211 And a rule that is costly for each individual to comply with is also costly
for the state to administer. These costs of deliberation and investigation, not
to mention the costs of people's apprehension about possible failure to fulfill
such an obtuse duty, must be taken into account by a utilitarian.
n212 In addition, insofar as positive freedom contributes to a person's well-being,
we must remember the presumption of liberty: All criminal duties come at the
basic cost of curtailing people's range of choices. If these costs are
outweighed by the benefits achieved, there is no utilitarian objection to the
mopping-up duty. It seems very doubtful, however, that the aggregate costs
would be less than the aggregate good such a criminal duty might achieve.
We thus have two considerations that govern the shape of desirable criminal
Such a duty should apply only to situations where the involvement of
individuals is indispensable; and it should specify such situations in advance,
rather than requiring people to determine the matter for themselves. The duty
to rescue is a criminal duty of
that fits this bill. But is it the only such duty? Granting the need to limit
the duty to emergency situations, what is the reason for further limiting it to
situations that involve grave suffering or the threat of death? From a
[*655] there is no magic in the magnitude of the benefit conveyed on each occasion.
Whether the legal duty of
should be extended to cover emergency cases involving moderate need, or even
to situations where the destruction of valuable property can be prevented,
depends only on whether the extra benefits of such an extension outweigh any
additional burdens it would bring. It seems to me that this test would not be
met because the extension would yield a great benefit only if it applied to a
large number of cases.
n214 But even if it did apply to a large number of cases, the additional
administrative costs and costs to positive liberty would likewise be large.
Thus, the net benefits of extending the duty seem likely to be small at best.
n215 Any assessment of the benefits and burdens of a particular legal rule is in
large part speculative; a sensible rule of thumb would be that the costs of
legal regulation should be incurred only where the expected net benefit is
substantial. The reason for moderate confidence in the overall net benefits of
the duty to rescue is that it has the potential to bring out a very great net
improvement in overall well-being in each situation to which it applies.
So far we have seen that practical considerations support limiting legal duties
to rescue contexts even if we accept the optimizing principle of
These considerations apply
to our two alternative principles of
Matters are less straightforward when it comes to the issue of demands. As we
saw, rescue cases can give rise to extreme demands even where the expected
benefit of the rescue is very great. Should the criminal law require a person
to take a fifty-percent chance of losing her life for the sake of a
fifty-percent chance of saving twenty? To address this question, we need to
take each of our three principles of
in turn. If we take a legal moralist approach to the criminal law or accept
the constraint against legally required sacrifice beyond what a person is in
any case morally required to sustain, then the upper bound of the legitimate
demands of legally required
will lie wherever the relevant moral principle of
[*656] Let us again start, then, with the implications of the optimizing principle of
Like most other actual legal duties to render aid, the Rhode Island provision
"reasonable" assistance, and specifies that this will never involve danger or peril to the
n217 As a matter of statutory interpretation, it is not plausible to read this
latter limitation such that a rescue attempt is required only if it would
involve no risk whatever of physical harm. Even walking to the edge of the dock
to toss the life preserver entails
risk to my life (I could slip and fall off, hitting my head on the rocks). In
any event, a sensible legal duty to rescue will require rescue attempts so long
as only a minimal risk of serious harm is involved.
n218 This means that the risk that a good swimmer will get a cramp and drown if she
swims out into the calm lake to rescue a floundering child would not be
sufficient to justify inaction. But an average swimmer (as opposed to an
off-duty, professional surf lifeguard) should not be required to head off into
rough surf to attempt to rescue someone caught in a rip tide.
From a utilitarian perspective, it is clear enough that people who have no
special training, and thus cannot count on any substantial probability of
succeeding in a rescue attempt, should not be encouraged, let alone required by
the criminal law, to take substantial risks with their own lives.
n219 Of course, it may be that a particular boat owner can count on roughly a
fifty-percent chance of success, and so, if there are twenty people to be
saved, the optimizing principle of
would, as a moral matter, require her to take the chance. The utilitarian
legislator's first-best result would be a provision that requires such a person
to act without encouraging dangerous rescue attempts by the rest of us. It is
not clear, however, that this can be done. A provision could specify that
rescue attempts are required so long as
"the effort, risk, or cost of acting is disproportionately less than the harm or
n220 Whether a particular person is liable for nonrescue would then turn on an
investigation of her actual chances of success and (perhaps) her awareness of
those chances. But though courts in tort cases deal routinely with the idea
that more can be expected of some people than of others, such an approach seems
very unwise in the current context. As noted, it is not for the best to
encourage people with no particular skills to take unknown (to them) levels of
risk with their own lives; yet, the suggested language would do just that. It
would be possible to require particular classes of rescuers to sustain a higher
level of risk, but this also seems
[*657] unnecessary because the relevant classes of people are likely already to be on
the job when the opportunity to attempt a rescue arises.
So much for the risk of physical harm. Equally difficult is the issue of all
other kinds of loss. The drafters of the Rhode Island provision take the
standard approach of not specifying what
"reasonable" non-physical cost might amount to. Is this what utilitarianism recommends? A
utilitarian rescuer must weigh the value of the chance of saving a life or
preventing or alleviating severe suffering against some very different kind of
value, such as a simple loss of money. A utilitarian lawmaker understands that
this kind of cost-benefit analysis cannot reliably be done on the spot. Because
the criminal law should not encourage wasteful rescue expenditures, the
utilitarian lawmaker aims to encourage potential rescuers to attempt a rescue
only in those cases where the expected benefit clearly outweighs the expected
loss. Because it is doubtful that any useful specific criteria could be
extracted from the range of possible cases and written into the law, this aim
could not be better served than by asking the trier of fact to determine
"reasonable" efforts were used.
n222 Even if further criteria were available, making the duty more complicated
would greatly weaken its educative/expressive role, which may be more important
than its role as a deterrent. The presence of such a duty in the criminal law
reminds us of our responsibilities toward other people, even total strangers,
and the more qualified the duty, the less powerful this message.
Triers of fact can get things badly wrong, and potential rescuers can know
this. The loose reasonableness standard, coupled with ex ante mistrust of
triers of fact, could thus seem to encourage rescue efforts that are not
cost-justified. This potential problem provides a further reason to limit the
duty to rescue to cases of severe suffering or threat to life, because in these
cases it is hard for an individual to expend too much in the way of material
resources. Moreover, in the criminal context the prosecution's high burden of
proof should allow people to be less mistrustful of triers of fact than may be
warranted in the civil context.
It is clear that even an ideal trier of fact might decide that a person ought
reasonably to permit the ruin of his million dollar yacht in order to rescue
some drowning people, and utilitarians would have no grounds for regret.
[*658] even if I am right that utilitarianism would recommend a legal duty to make
only reasonable efforts, this may still leave us with a legal duty of
that some will find too demanding.
As it happens, there is a utilitarian answer to this problem. Saul Levmore has
plausibly argued that the most efficient incentive package for legal duties to
rescue combines a (small) sanction for failure to rescue with a small reward
for rescuing as required.
n226 By a small reward, Levmore means one that does no more than compensate the
rescuer for her expenses.
n227 Such is the arrangement in France and Germany.
n228 The upshot of this result is that even in those cases where reasonable rescue
is very costly for the rescuer, an optimal legal regime would reimburse those
costs. Most states' crime victims' compensation statutes already provide for
compensation for rescuers in the context of violent crimes.
n229 And in the non-crime context, existing principles of restitution may already
provide for such compensation where the rescue is carried out pursuant to a
legal duty such as the Rhode Island provision.
n230 If so, providing for compensation would require no further legal
innovation--unless it was thought, as seems plausible, that the compensation
should flow not from the person rescued, but from the community at large by way
of a statutory compensation scheme.
[*659] I will not rehearse the details of Levmore's argument about the optimal size
of the rescue
"carrot"; but we should briefly address the issue of the rescue
"stick"--the penalty for violation. On a legal moralist approach, the penalty for
violation should depend on our judgment of the severity of the moral wrong done
by the person who fails to assist.
n232 One of the problems with the commission by omission approach to failure to
save a life is that it makes it impossible for the law to be sensitive to the
difference between the moral wrongness of killing versus letting die--a
commonsense distinction that many moral philosophers take very seriously.
n233 A statutory duty to rescue, however, can treat letting die as far less serious
than homicide, and existing provisions do just that. The maximum penalty for
failure to assist in Vermont is a fine of $ 100, Rhode Island's maximum of $
500 and six months imprisonment is at the high end in the common-law world,
n234 and the norm in Europe and Latin America is for a penalty of either a fine or
no more than a year's imprisonment.
But we do not have to be retributivists to have a reason to distinguish between
the sanctions appropriate for positive harming as opposed to failure to aid.
Though the incentives for failing to aid
be great, as in the case of the Princess of Wales, it is generally true that
the payoff available from doing
[*660] positive harm will be much greater--simply because one can do harm all on
one's own and thus choose to harm where it is most profitable. Similarly, a
person motivated by great hatred and anger will rarely be able to satisfy his
desire by merely letting harm befall someone. Thus, in a deterrence analysis,
the sanction for failing to harm should generally not be as severe as the
sanction for harming. Indeed rather minimal sanctions would seem to be
adequate, and this is what we observe.
It appears then that if we accept the optimizing principle of
as part of our account of the aim of the criminal law, we will nevertheless
restrict criminal duties of
to duties to rescue with minimal demands and minimal sanctions for violation.
I have argued, however, that such a principle of
is very implausible in the realm of personal conduct. Thus we must examine
whether our two alternative principles of
the limited principle and the collective principle, also support this kind of
minimal legal duty to rescue. Of course neither principle will support a legal
duty to rescue that is
demanding than the utilitarian duty. What needs to be shown, to satisfy legal
moralists and those who embrace a constraint against imposing sacrifice beyond
what a person is morally required to make independently of law, is that this
minimal legal duty to rescue is not more demanding than is compatible with the
collective and limited principles of
As a matter of personal morality, the operation of the limited principle and
the collective principle in rescue contexts turns out to be a very complicated
n237 But the main point is simple: Because rescue contexts are not normatively
special, and because under both principles the sacrifice required for the sake
of benefiting others is not unlimited, there will be some situations in which
these principles will not require a person to perform even a very easy and
cheap rescue that is certain to succeed. This may seem to ruin the case for
legal duties to rescue; the only principles of
that seem to have a hope of being plausible generally do not in the end
support a blanket legal requirement to rescue, no matter how minimal.
Here again the option of compensation makes all the difference. Remembering
that more-than-minimal risks of physical harm would not be required, even by a
legal duty based on the optimizing principle of
the provision for compensation of material expenses means that those who
comply with the legal duty to rescue sustain no significant net sacrifice at
all; the problem of demands is, in the end, sidestepped at the level of
[*661] Thus plausible principles of
do appear to converge in recommending legal duties to rescue that require
minimal risk-taking and that provide for both compensation and a moderate
sanction for violation.
n239 Needless to say, this conclusion is not irresistible. For one thing, I have
not considered the possibility that there are ancillary bad effects of criminal
duties to rescue that, together with the costs already discussed, outweigh the
good these duties would do. This does not seem to me to be likely, but I will
not investigate the issue here.
n240 As already noted, I do not pretend to have conclusively established that the
benefits of a legal duty to rescue substantially outweigh even the costs so far
considered. All these arguments involve speculation; their conclusions can only
be regarded as more or less plausible. My main aim has been to uncover the
structure of the legal, moral/political, and practical issues that underlie the
debate about the duty to rescue and to outline some plausible options available
[*662] to us. And what I do claim to have shown is that if there is a reason to
oppose all criminal duties to rescue, it will not be because such duties are
offensive to plausible considerations of political morality. So long as
compensation is provided, all plausible principles of
support such duties and no plausible arguments grounded in individual liberty
count against them. Rather, it will be because an exhaustive evaluation of the
benefits and burdens of criminal duties to rescue leaves us insufficiently
confident that such duties yield a clear net benefit. That almost all the
countries of Continental Europe and Latin America should be wrong about this
practical issue, and the handful of common-law countries right, seems to me
prima facie dubious.
n241 But it could be so.
B. NO TORT DUTY
The reasons for thinking that tort liability for failure to aid is less
desirable than some form of criminal liability can be quickly stated.
n242 The advantage of criminal liability lies in the flexibility of the sanction.
If we could assume that no one would ever violate it, a tort duty to make
cost-justified rescue attempts would be unobjectionable from the point of view
of social costs.
n243 But of course, some people would violate such a civil duty to rescue,
irrational though that may be.
n244 And such tortfeasors would be liable for full compensatory damages. There are
two problems with this.
Just as one does not have to be a full-fledged retributivist to believe that
some criminal sanctions that would be appropriate on a pure deterrence
rationale must be rejected as too severe (no matter how the deterrence
calculations come out, drawing and quartering would not be an acceptable
punishment for tax dodgers), one need not deny that the aim of tort law is the
promotion of social good in order to believe that legitimate tort liability has
its limits. Indeed, whatever the correct account of the normative basis of tort
law might be, the prospect of a
[*663] person who failed to respond to an emergency situation being liable for
damages in full compensation for the death of the victim seems likely to be
n245 The recent Australian decision of
Lowns v. Woods
n246 awarding three million Australian dollars in damages against a doctor whose
nonfeasance led to a child (not his patient) becoming permanently disabled may
seem tolerable only because the defendant was a physician--protected, no doubt,
by liability insurance.
Apart from considerations of proportionality, there is the simple fact that,
assuming that the aim of tort law is to promote general well-being, full
compensatory liability may well be excessive from an instrumental perspective.
First, the availability of either rescue or (at least a chance of) full
compensation may encourage foolish and highly risky activities by potential
n247 Second, as full compensatory liability is almost certainly more than is
required to deter nonrescue for most people, the costs incurred in bringing
about the transfer of the loss from a victim to a nonrescuer are costs without
benefit. Which brings out the main point: Unlike civil liability, criminal
sanctions can be adjusted to achieve the desired deterrence effect. Once
criminal liability is in place, there is simply no need for additional civil
n248 In the absence of any grounds of corrective justice for requiring the
nonrescuer to compensate the victim--an absence that is pronounced in
contemporary theories of corrective justice
n249 --the only possible ground for imposing tort liability would be that criminal
liability has special disadvantages that outweigh its immeasurably greater
It is notable that the conclusion we have reached, in favor of limited criminal
duties to rescue but wary of civil liability, has some support from the
situation in Continental Europe, where civil liability for failure to rescue is
much less common than criminal liability.
CONCLUSION: LAW, MORALITY, AND STATE
Criminal duties to render aid of the kind that are now increasingly popular in
the United States face no objection on the ground of their interference with
liberty, and moreover may be desirable, all things considered. The problem of
finding a plausible moral principle of
must be addressed by defenders
[*664] of legally required
but it need not be finally solved, because it turns out that all the plausible
options would support appropriately limited criminal duties to rescue. The
reasons why more extensive legal duties of
are not appropriate on any of the plausible views about the morality of
are practical: State institutions can do a better job of promoting well-being
than can individuals in all but emergency contexts, and even in those contexts
there are good reasons not to legally require people to sustain a substantial
risk of death or injury attempting to aid others.
In conclusion, I would like to make two observations that turn on the fact that
the reasons why only strictly constrained legal duties of
are defensible and are, indeed, essentially practical. First, we may have here
one of those areas of normative thought where the law misleads us about
political morality. The fact that it would be absurd to enact a criminal duty
to promote social welfare generally may lead us to conclude, wrongly, that the
duties people have in respect of
are minimal or optional, especially outside rescue contexts. This would be a
n252 It is true that when we think about what kind of criminal duty of
would make sense we must consider how the state might better promote welfare
via some other institutional means: In an ideal society, very little may end up
being the responsibility of individuals, especially not on pain of criminal
sanction. But it is also true that if the state
to set up the appropriate institutions, then moral requirements of
apply directly to people, regardless of the law. The fact that an ideal state
would make it unnecessary for people to be that much concerned, in their daily
lives, about social welfare does not mean that we need not be concerned about
social well-being in our actual, nonideal situation. And so, as I have said,
the fundamental normative principles to which we appeal when trying to describe
ideal legal institutions had better be ones we can live with in our actual,
The second observation returns us to the issue of the positive responsibilities
of government. If I am right, there is no fundamental objection of political
morality to the very idea of positive legal obligations. But notice the
difference between the case of people and that of government. In the case of
people, we saw that the problem of finding a reasonable moral principle of
did not have to be finally resolved by legal theory because all the candidate
principles yielded the same recommendation on the issue of legally required
When we turn to states, however, the practical reasons for accepting only
strictly constrained legal requirements of
disappear. At least, the nature of the practical arguments changes
dramatically, and the practical potential for the state to do good is obviously
vastly greater than is the case for individuals. Thus, when we are addressing
the responsibilities of states, the question of which principle of
we accept as a matter of political morality cannot be avoided. Whether the
state's duties of
to its members should be constitutionalized is another matter. But it must not
be pretended that the fundamental question of what a decent community must do
to promote the well-being of its members need not be reached--either because
there is some intrinsic incompatibility between the idea of positive state
obligation and the basic values of our legal culture, or because, practically
into the law would never make any sense.
And so we see that acceptance of legal duties to rescue, and thus of
does indeed open the can of worms that some commentators have been so keen to
n1 For the common law, see discussion
Part I. For Continental Europe and Latin America, see Alberto Cadoppi,
Failure to Rescue and the Continental Criminal Law, in
THE DUTY TO RESCUE: THE JURISPRUDENCE OF AID 93 (Michael A. Menlowe
& Alexander McCall Smith eds., 1993);
at 97-100 (providing a brief history of legal duties to rescue in ancient
Mediterranean and European cultures and in modern Europe); F.J.M. Feldbrugge,
Good and Bad Samaritans: A Comparative Survey of Criminal Law Provisions
Concerning Failure to Rescue
14 AM. J. COMP. L. 630, 632-34 (1966); Gilbert Geis,
Sanctioning the Selfish: The Operation of Portugal's New 'Bad Samaritan' Statute
, 1 INT'L REV. VICTIMOLOGY 297, 297-300, 301-02 (1991); Aleksander W.
The Duty to Rescue: A Comparative Analysis, in
THE GOOD SAMARITAN AND THE LAW 91 (James Ratcliffe ed., 1966); Martin Vranken,
Duty to Rescue in Civil Law and Common Law
47 INT'L & COMP. L.Q. 934, 938-39 (1998).
n2 Utilitarianism, which holds that people are required to promote overall
well-being (either of humans or of all sentient life) is thus one form of
consequentialism, the theory that holds that people are required to optimize
the overall good (on some account of the good, which may include more than
well-being). The version of utilitarianism popular among economists is narrower
still, as it understands human well-being exclusively in terms of preference
satisfaction. This preference-satisfaction version of utilitarianism should
not, in turn, be confused with either the Pareto or Kaldor-Hicks criterion of
efficiency; these efficiency criteria are employed by economic analysts of law
in an attempt to guide institutional design without making the interpersonal
comparisons of utility required by any version of utilitarianism.
, JULES COLEMAN, MARKETS, MORALS AND THE LAW 95 (1988).
n4 THOMAS BABBINGTON MACAULAY,
Notes to the Indian Penal Code, in
THE WORKS OF LORD MACAULAY 314-15 (Lady Trevelyan ed., 1900).
at 315-16; 2 EDWARD LIVINGSTON,
Code of Crimes and Punishments, in
COMPLETE WORKS 126-27 (1873).
"In cases where the person is in danger, why should it not be made the duty of
every man to save another from mischief, when it can be done without
prejudicing himself, as well as to abstain from bringing it on him?" JEREMY BENTHAM, THE PRINCIPLES OF MORALS AND LEGISLATION ch. XVII,
§ 1.XIX (1789). In a footnote, Bentham offers these excellent examples:
A woman's head-dress catches fire: water is at hand: a man, instead of
assisting to quench the fire, looks on, and laughs at it. A drunken man,
falling with his face downwards into a puddle, is in danger of suffocation:
lifting his head a little on one side would save him: another man sees this and
lets him lie. A quantity of gunpowder lies scattered about a room: a man is
going into it with a lighted candle: another, knowing this, lets him go in
without warning. Who is there that in any of these cases would think punishment
at 323 n.1.
Part I. Cadoppi very plausibly suggests that a significant factor is the lack
of large scale codification. Cadoppi,
note 1, at 115-16.
n8 The strange juxtaposition of the popularity of utilitarianism and the
unpopularity of positive legal duties is illustrated by James Barr Ames in his
well-known 1908 lecture.
James B. Ames,
Law and Morals
22 HARV. L. REV. 92 (1908),
THE GOOD SAMARITAN AND THE LAW,
note 1, at _. Ames, having declared roundly and without argument that
"the law is utilitarian,"
at 17, and having noted that a duty to rescue a stranger from peril, even at
little cost, is unknown in the common law, has no trouble concluding that the
law ought not
"remain in this condition,"
Robert L. Hale,
Prima Facie Torts, Combination, and Non-Feasance
46 COLUM. L. REV. 196 (1946). Hale writes:
Perhaps judicial reluctance to recognize affirmative duties is based on one or
both of two inarticulate assumptions. One of these is that a rugged,
independent individual needs no help from others, save such as they may be
disposed to render him out of kindness, or such as he can induce them to render
by the ordinary process of bargaining, without having the government step in to
make them help. All he is supposed to ask of the government is that it
interfere to prevent others from doing him positive harm. The other assumption
is that when a government
requires a person to act, it is necessarily interfering more seriously with his liberty
than when it places limits on his freedom to act--to make a man serve another
is to make him a slave, while to forbid him to commit affirmative wrongs is to
leave him still essentially a free man.
at 214. A recent articulation of the second of these assumptions, from high common-law
authority, is found in Lord Hoffmann's speech in
Stovin v. Wise
There are sound reasons why omissions require different treatment from positive
conduct. It is one thing for the law to say that a person who undertakes some
activity shall take reasonable care not to cause damage to others. It is
another thing for the law to require that a person who is doing nothing in
particular shall take steps to prevent another from suffering harm from the
acts of third parties . . . or natural causes. One can put the matter in
political, moral or economic terms. In political terms it is less of an
invasion of an individual's freedom for the law to require him to consider the
safety of others in his actions than to impose upon him a duty to rescue or
1996 C.A. 923, 941 (appeal taken from Court of Appeal). For Lord Hoffmann's
"moral" point, see
A nonjudicial source on the same issue is Sheldon Richman:
The American philosophy of jurisprudence doesn't permit the government to
impose positive obligations on citizens. America was founded on a bedrock of
inalienable individual rights. Under that theory, each person is the owner of
his life and has no positive legal obligations to others that are enforceable
by government except those that are voluntarily accepted. All that government
can require of you is that you abstain from violating the rights of others by
subjecting them to force or fraud. Thus, even if you believe that morality
requires you to help an accident victim, it's not enforceable by government . .
. . People might find Good Samaritan laws reasonable because they believe
people of good will should help others in distress. But where individual rights
are respected and government power is limited, good will cannot be enshrined in
the law. It would undermine freedom.
You Can't Legislate Goodwill
, CHRISTIAN SCI. MONITOR, Oct. 2, 1997, at 19 (author is identified as
"vice president of policy affairs for the libertarian Future of Freedom
Foundation in Fairfax, Virginia");
Bad Samaritanism and the Causation of Harm
, 9 PHIL.
& PUB. AFF. 230 (1980) ("In our law, there's no duty to rescue someone or save someone's life . . . .
Our society is based on the right and sanctity of the individual." (quoting Allegheny Court, Pennsylvania Judge John P. Flaherty in
McFall v. Shimp
, 10 Pa. D. & C.3d 90 (Allegheny County 1978))). For further discussion, see
n10 In seeing this issue of costs or demands as the central normative problem
associated with legal duties to rescue, I am in agreement with Michael Menlowe.
Michael A. Menlowe,
The Philosophical Foundations of a Duty to Rescue, in
THE DUTY TO RESCUE: THE JURISPRUDENCE OF AID,
note 1, at 5. Another discussion similar in approach to my own is Samuel
Criminal Liability and the Duty to Aid the Distressed
142 U. PA. L. REV. 1455 (1994).
n12 The expression
"overlapping consensus" is John Rawls's.
JOHN RAWLS, POLITICAL LIBERALISM (1993).
Richard A. Epstein,
105 HARV. L. REV. 1106, 1118 (1992) ("Imposing affirmative legal duties necessarily increases the total level of
coercion, public and private, in society. The risk of excessive government
power should not be run because, in the good Samaritan context, self-help
remedies are usually effective."); Marc A. Franklin
& Matthew Ploeger,
Of Rescue and Report: Should Tort Law Impose a Duty to Help Endangered Persons
or Abused Children?
40 SANTA CLARA L. REV. 991, 1004 (2000); William M. Landes
& Richard A. Posner,
Salvors, Finders, Good Samaritans
7 J. LEGAL STUD. 83, 127 (1978); Melody J. Stewart,
How Making the Failure to Assist Illegal Fails to Assist: An Observation of
Expanding Criminal Omission Liability
25 AM. J. CRIM. L. 385, 432 (1998).
It is true that reported cases are rare, to put it mildly. The only reported
case concerning a prosecution solely under a duty to assist provision I am
aware of is
State v. La Plante
, 521 N.W.2d 448, 452 (Wis. Ct. App. 1994) (upholding constitutionality of provision imposing duty to assist crime
n14 It is, of course, not the case that there are no positive legal duties of any
kind in common law jurisdictions.
Part I. The question is whether there is a general normative reason to treat
their introduction with caution.
Jackson v. City of Joliet, 715 F.2d 1200 (7th Cir. 1987) (Posner, J.);
DeShaney v. Winnebago County Dep't of Soc. Servs., 489 U.S. 189 (1989). For critical discussion, see Susan Bandes,
The Negative Constitution: A Critique
88 MICH. L. REV. 2271 (1990); Frank I. Michelman,
The Supreme Court, 1968 Term--Foreword: On Protecting the Poor Through the
83 HARV. L. REV. 7, 13-14 (1969) (arguing that the government has an affirmative obligation under the
Constitution to provide assistance to the badly-off). For discussion of
positive duties under state constitutions, see Helen Hershkoff,
Positive Rights and State Constitutions
112 HARV. L. REV. 1131 (1999). I do not in this Article attempt to engage with the further legal issues
raised when we move from the case of individuals to that of government. For
discussion of some of those issues, see generally Barbara Armacost,
Affirmative Duties, Systemic Harms, and the Due Process Clause
94 MICH. L. REV. 982 (1996); David Currie,
Positive and Negative Constitutional Rights
53 U. CHI. L. REV. 864 (1986); Hershkoff,
n16 Richard Epstein writes:
The fundamental problem in a system of welfare is that it conflicts with the
theory of private rights that lies behind any system of representative
government . . . . The legal theory that recognizes no obligation to rescue a
stranger in imminent peril cannot generate, let alone nourish, a system of
transfer payments and welfare obligations.
RICHARD EPSTEIN, TAKINGS 318-19 (1985). The connection with the takings clause
is explained by Epstein.
at 322 ("The basic rules of private property are inconsistent with any form of welfare
benefits."). Epstein has recently recast his earlier, rights-based opposition to duties
to rescue in utilitarian terms.
text accompanying note 163. This wholly undermines the direct argument from
the lack of a duty to rescue to the rejection of transfer payments because a
utilitarian will have to evaluate each issue on its merits.
Richard Epstein's Takings: Private Property
& the Power of Eminent Domain
41 U. MIAMI L. REV. 49, 179-91 (1986) [hereinafter
Richard Epstein's Takings
]. Epstein does, however, make broadly utilitarian points against the public
welfare system in
, at 320-23.
Theodore M. Benditt,
Liability for Failing to Rescue
, 1 LAW
& PHIL. 391, 396 (1982). In addition to the effect on thinking about the legal
duties of the State, there is the effect law-school discussion of the issue of
the duty to rescue may have on the attitudes of public officials.
MARY ANN GLENDON, RIGHTS TALK: THE IMPOVERISHMENT OF POLITICAL DISCOURSE 80
note 13, at 1118.
The recent admitted actions and subsequent statements of Mr. David Cash may
make such remarks less common in the future. David Cash admitted seeing his
friend Jeremy Strohmeyer begin the assault on a seven year old girl that
eventually led to her death in a Nevada casino on May 25, 1997. Mr. Cash is
reported to have said:
"I'm not going to get upset over somebody else's life. I just worry about myself
first." Don Terry,
Mother Rages Against Indifference
, N.Y. TIMES, Aug. 24, 1998, at A1. This case provided the motivation for
Nevada's duty-to-report provision referred to
A.M. ROSENTHAL, THIRTY-EIGHT WITNESSES (1964).
, Marlise Simons,
French Magistrates Clear Photographers in Death of Diana
, N.Y. TIMES, Sept. 4, 1999, at A2. Mohammed al-Fayed lost his first appeal of
the decision to drop the charges but has announced his intention to take his
appeal to the Superior Court.
Al-Fayed to Appeal Ruling over Death of Diana and Dodi
, THE HERALD (Glasgow, Scotland), Nov. 1, 2000, at 14.
n21 C. PEN. art. 223-6. The sanction provided for in this provision is a good deal
harsher than similar provisions in Continental Europe and Latin America--the
norm is a fine and not more than one year's imprisonment. The reason for the
exception is that the French penal code is unique in having no separate
provision providing for offenses of commission by omission.
note 1, at 109.
See generally infra
Part I.A. The French provision was part of a proposed reform of the French
penal code in 1934 but was first enacted by the Vichy regime in 1941; it was
included in the Penal Code in 1945.
& Eva Steiner,
Criminal Omissions and Public Duties: The French Experience
, 10 LEGAL STUD. 153, 156-57 (1990). Ashworth and Steiner provide a helpful
summary of the main types of cases in which the provision has been invoked.
"It is obvious that Diana, Princess of Wales, Dodi Al Fayed, Henri Paul, and
Trevor Rees Jones were no safer from their peril in Paris with its
duty-to-assist law than they would have been in any place without such a law." Stewart,
note 13, at 435.
n23 Currently, duty-to-aid provisions are in force in Vermont, VT. STAT. ANN. tit.
§ 519 (1973), Minnesota, MINN. STAT.
§ 604.01(a) (1998), and Rhode Island, R.I. GEN. LAWS
§ 11.56-1 (1994). Hawaii, HAW. REV. STAT. ANN.
§ 663-1.6 (Michie 1995), Massachusetts, MASS. GEN. LAWS ANN. ch. 268,
§ 40 (West 1990), Washington, WASH. REV. CODE ANN.
§ 9-69.100 (West 1998), and Wisconsin, WIS. STAT. ANN.
§ 940.34 (West 1996
& Supp. 1997), have provisions imposing duties either to aid crime victims or
report an ongoing crime to the authorities. Florida, FLA. STAT. ANN.
§ 794.027 (West 1992
& Supp. 1998), and Nevada, NEV. REV. STAT.
§ 202-882 (1999), have duty-to-report provisions restricted to certain sexual
offenses. In recent years, duty-to-aid bills have been considered in Hawaii,
H.R. 885, 20th Leg. (Haw. 1999) and Illinois, H.R. 2612, 91st Gen. Ass. (Ill.
1999). A bill that would enact a duty to report the witnessed sexual assault of
a child is still, at the time of writing, under consideration in New York,
Senate 12, 223d Leg. (N.Y. 2001). For an account of the incident that led to
the Washington bill, see Stewart,
113 N.W. 1128 (Mich. 1907).
n26 Graham Hughes,
67 YALE L.J. 590, 624 (1958).
n27 Hughes takes this expression from the French,
at 598; it is now widely used,
, GEORGE FLETCHER, RETHINKING CRIMINAL LAW 422 (1978).
, N.Y. VEH.
& TRAF. LAW
§ 600 (McKinney 1996) ("hit-and-run" statute).
See generally infra
notes 45-46 and accompanying text.
, 113 N.W. at 1129. The Model Penal Code provides for liability for omissions in two
circumstances: (1) where the law defining an offense so provides; or (2) where
the duty to act is
"otherwise imposed by law." MODEL PENAL CODE
JOSHUA DRESSLER, UNDERSTANDING CRIMINAL LAW 91 (1995); WAYNE R. LaFAVE
& AUSTIN SCOTT, JR., CRIMINAL LAW 202 (1986).
note 29, at 203.
n31 JEROME HALL, THE GENERAL PRINCIPLES OF THE CRIMINAL LAW 193-94 (1960).
note 26, at 620.
note 29, at 204-05.
at 205. George Fletcher makes the plausible point that in many cases it is not
so much the contract (in the technical sense of an agreement enforceable under
contract law) as the undertaking to provide assistance that creates the
relevant duty. FLETCHER,
note 27, at 615-16.
State v. Miranda, 715 A.2d 680 (Conn. 1998) (upholding conviction for assault in the first degree on basis of defendant's
omission to protect child of live-in girlfriend from abuse).
Note that not all the duties--omission of which may give rise to criminal
liability for a crime of commission--are in themselves sufficient to ground
W. PAGE KEETON ET AL., PROSSER AND KEETON ON THE LAW OF TORTS 376 (5th ed.
note 26, at 620.
note 27, at 628-31; PAUL H. ROBINSON, CRIMINAL LAW 195-99 (1997); George P.
On the Moral Irrelevance of Bodily Movements
142 U. PA. L. REV. 1443, 1449 (1994) [hereinafter Fletcher,
On the Moral Irrelevance
Does Criminal Liability Require an Act?, in
PHILOSOPHY AND THE CRIMINAL LAW 60 (Antony Duff ed., 1998).
On the Moral Irrelevance, supra
note 39, at 1449. The problem of legality is said to be one reason why French
law allows liability for commission by omission in only a very narrow range of
note 21, at 155.
n41 See Cadoppi,
note 1, for a recent list of European and Latin American provisions, along
with translations. For U.S. provisions, see
note 23. For Australia's Northern Territory, see Criminal Code Act
§ 155 (N. Terr. Austl.), which provides for a sanction of up to seven years
imprisonment for a
"callous" failure to rescue.
See also infra
Steven Heyman argues that in addition to the offense of
"misprision of felony" (failing to report a felony), the early common law imposed a criminal duty to
Steven J. Heyman,
Foundations of the Duty to Rescue
47 VAND. L. REV. 673, 685-90 (1994).
n42 R.I. GEN. LAWS
§ 11-56-1 (1994).
n43 MINN. STAT.
§ 604.01(a) (1998); VT. STAT. ANN. tit. 12,
§ 519 (1973).
, N.Y. FAM. CT. ACT
§ 413 (McKinney 1998).
§ 6012 (West Supp. 1999).
n46 Commonwealth Electoral Act, 1918,
§ 101 (Austl.).
note 29, at 193.
note 26, at 590-97 (history of criminal omissions);
note 41, at 682-90.
MICHAEL S. MOORE, ACT AND CRIME: THE PHILOSOPHY OF ACTION AND ITS IMPLICATIONS
FOR CRIMINAL LAW (1993); Husak,
In what follows, I treat the act requirement as distinct from the requirement
note 29, at 71-77.
Criminal Liability for Failures to Act
, 49 LAW
& CONTEMP. PROBS. 161, 162-67 (1986).
Christopher H. Schroeder,
Two Methods for Evaluating Duty to Rescue Proposals
, 49 LAW
& CONTEMP. PROBS. 181, 183-84 (1986).
MICHAEL MOORE, PLACING BLAME 262-77 (1997) (listing references).
n51 For a different approach that turns on the notion of control, see Husak,
n52 MODEL PENAL CODE
§ 2.01(1). For discussion see DRESSLER,
note 29, at 69-71 (though the tone of Dressler's discussion of omissions
suggests that he is attracted to Moore's approach, discussed
notes 53-55 and accompanying text); LaFAVE
note 29, at 195; ROBINSON,
note 39, at 175-87.
n53 See MOORE,
note 49, and his replies to critics in
note 50, chapter six and accompanying references.
One of Moore's critics, Bernard Williams, doubts that much of substantial
relevance to the criminal law could turn on the philosophy of action. Bernard
The Actus Reus of Dr. Caligari
142 U. PA. L. REV. 1661 (1994). I agree. For the related discussion of causation, see
text accompanying notes 56-59. Part of Moore's response to Williams in
note 50, at 253, is that his argument starts with the moral claim that acts
are central to criminal liability, and only then proceeds to the investigation
of the essence of acts. But it is a strange view that asserts the moral
significance of acts--whatever they on further investigation may turn out to
be. How do we know whether they are morally significant if we don't know what
they are? Note that Moore's argument is different from Derek Parfit's use of
metaphysics in normative theory. Parfit believes that his discussion of the
metaphysics of personal identity has relevance for ethical theory, but here the
argument proceeds by showing that the question of personal identity (what makes
me, me, and you, you), which is central to a range of ordinary moral views--and
which in ordinary life we think we understand well--turns out to be badly
understood, resulting in the revision of some of our ethical views.
DEREK PARFIT, REASONS AND PERSONS pt. 3 (1984). Moore's argument is different
in that we do not, in ordinary life, operate with an uncontroversial account of
what is an act, and (not surprisingly given the first point) the normative
relevance of acts to criminal liability and punishment is itself not
note 49, at 34 ("Perhaps . . . [the] criminal law is mistaken.").
, John Harris,
The Marxist Conception of Violence
, 3 PHIL.
& PUB. AFF. 192 (1974); John Kleinig,
, 5 PHIL.
& PUB. AFF. 382 (1976); Arthur Leavens,
A Causation Approach to Criminal Omissions
76 CAL. L. REV. 547 (1988). Alison McIntyre provides an excellent critical discussion of this approach.
Guilty Bystanders?: On the Legitimacy of Duty to Rescue Statutes
, 23 PHIL.
& PUB. AFF. 157 (1999). Christopher Schroeder succinctly states the essential
"If failures to act can be causes, the issue of assigning legal responsibility
causes is not yet resolved." Schroeder,
note 50, at 190 (emphasis added).
n56 See also the related discussion of Williams's critique of Moore,
note 29, at 167; LaFAVE
note 29, at 278; Hughes,
note 26, at 627-31. The
Model Penal Code
does away with the traditional doctrine of proximate causation and explicitly
replaces it with tests of culpability.
MODEL PENAL CODE
§§ 2.03(2)(b), 2.03(3)(b); DRESSLER,
note 29, at 174-75.
note 31, at 195 (noting that the issue of causation is never considered to be
a difficulty in the well-recognized cases of commission by omission); Husak,
note 39 (arguing that the notion of control would do a better job of capturing
the underlying substantive point of the causation requirement).
note 26, at 602.
at 604 (quoting GLANVILLE WILLIAMS, CRIMINAL LAW 40 (1953)).
note 26, at 604.
note 31, at 200. Hall's argument for this point is somewhat different from
that which follows in the text.
note 27, at 424-25.
n63 What I am calling the
"relevant description" is, in the terms of section 1.13(9) of the
Model Penal Code
, a combination of specified
"attendant circumstances"; for relevant discussion, see ROBINSON,
note 39, at 232-38.
note 26, at 602.
n65 These are the facts of
Lambert v. California
, 355 U.S. 225 (1957), a case which turned on the constitutional issue of the notice required by due
process; the majority judgment held that in the specific circumstances of the
case, a strict liability application of the requirement to register violated
due process. It is not at all clear what principle motivated the decision in
note 26, at 618-19. The case is rarely discussed and has not, in federal
courts, been the basis of the constitutional invalidation of any other
note 29, at 152-54.
note 29, at 208.
note 29, at 405-10.
note 26, at 605-06. Jerome Hall argues that the relevant distinction is
whether the crime is petty or not: Ignorance of the law may (should be able to)
undermine mens rea in the case of petty offenses of both omission and
note 31, at 200-01, 403. But the seriousness of the offense, though correlated
with the salience of the relevant description to the agent, is not the same
thing. Consider an offense of failing to remove exposed asbestos from one's
building: If the duty was established before there was public knowledge of the
hazards of asbestos, the failure to act may lack mens rea and yet be a very
n69 This is the approach taken in Continental European jurisdictions with criminal
duties to rescue.
note 1, at 641.
note 26, at 628; A.D. Woozley,
A Duty to Rescue: Some Thoughts on Criminal Liability
69 VA. L. REV. 1273, 1290-91 (1983).
note 1, at 641 (citations of European case law on this point).
n72 See the comment by Richard Epstein in
Richard Epstein's Takings, supra
note 16, at 184.
note 21; Andre Tunc,
The Volunteer and the Good Samaritan, in
THE GOOD SAMARITAN AND THE LAW,
note 1, at 43.
"The short and simple answer to the it-won't-work objection is to point out
where and for how long it has been working." Woozley,
note 70, at 1290.
Pirkle v. Oakdale Union Grammar Sch. Dist., 253 P.2d 1 (Cal. 1953) (stating standard of care required of an officer or employee of a public
160 N.E. 301 (Mass. 1928)
44 A. 809 (N.H. 1898).
n78 KEETON ET AL.,
note 35, at 375 n.22. Another favorite of commentators is
Yania v. Bigan
, 155 A.2d 343 (Pa. 1959), where the defendant somehow incited the victim, a visitor on defendant's land,
to jump into a deep trench with water at the bottom; the defendant then did
nothing to help the victim, who drowned.
, 44 A. at 810.
Pridgen v. Boston Hous. Auth., 308 N.E.2d 467, 477 (Mass. 1974) (overruling
Castonguay v. Acme Knitting Mach. & Needle Co., 136 A. 702, 705 (N.H. 1927) (overruling
Southern Railway v. Grizzle, 53 S.E. 244 (Ga. 1906).
KEETON ET AL.,
The quotation from Lord Hoffmann, mentioned
note 9, continues:
A moral version of this point may be called the
"why pick on me?" argument. A duty to prevent harm to others or to render assistance to a person
in danger or distress may apply to a large and indeterminate class of people
who happen to be able to do something. Why should one be held liable rather
Stovin v. Wise, 1996 C.A. 923, 941.
Ernest J. Weinrib,
The Case for a Duty to Rescue
90 YALE L.J. 247, 262 (1980) (arguing in support of a tort duty to rescue; Weinrib later came to believe
that such a duty is incompatible with the fundamental normative structure of
Waiting for Rescue: An Essay on the Evolution and Incentive Structure of the
Law of Affirmative Obligation
72 VA. L. REV. 879, 936 (1986); Richard W. Wright,
The Standards of Care in Negligence Law, in
PHILOSOPHICAL FOUNDATIONS OF TORT LAW 274 (David G. Owen ed., 1995). Levmore
also discusses a different (and rather unreal) problem of what to do where only
one person fell under the duty to rescue and failed to act as required, but it
is not known which person that is. Levmore,
, at 934-35. On contribution generally, see KEETON ET AL.,
KEETON ET AL.,
note 35, at 377.
Herskovits v. Group Health Coop. of Puget Sound, 664 P.2d 474 (Wash. 1983) (finding a fourteen percent reduction, from thirty-nine percent to twenty-five
percent, in decedent's chance for survival sufficient evidence of causation to
allow jury to consider possibility that physician's failure to provide timely
diagnosis of decedent's illness was the proximate cause of his death). This
doctrinal maneuver has so far been limited to cases of medical malpractice. The
more general issue of casting omissions as causes is not substantially
different--that is, equally unproblematic--in torts than in the criminal law.
text accompanying notes 55-58;
KEETON ET AL.,
note 35, at 265.
n88 By this I merely mean that the existing conceptual and doctrinal structure of
tort law is flexible enough to accommodate such an innovation. Having once
note 85, Ernest Weinrib now opposes a tort duty to rescue on the ground that
this would be false to the essential features of tort law as a normative
system, one of which is that there is no liability for nonfeasance,
, Ernest J. Weinrib,
Understanding Tort Law
23 VAL. U. L. REV. 485 (1989). In his later writing, Weinrib has noted:
The misfeasance requirement is essential if tort law is to represent the
ordering intrinsic to doing and suffering. The requirement signals that
consequences do not attract obligations except as they arise out of the risk
inherent in a particular act. Suffering by the plaintiff that is independent of
the defendant's doing has no significance for tort law. Accordingly, no
liability lies for failure to prevent or alleviate suffering. The defendant's
duty reflects the correlative morality of doing and suffering, and thus arises
only from the potency of injury in the defendant's act. The plaintiff's
unilateral need for assistance, no matter how urgent, falls outside the
relationship of doing and suffering. The exclusion of liability for nonfeasance
ensures that the normative implications of a tortious act depend not on the
undesirability of plaintiff's suffering as such but on the suffering being the
materialization of a risk inherent in the defendant's action.
We can grant that the misfeasance requirement is essential
"if tort law is to represent the ordering intrinsic to doing and suffering." We can even grant that tort law, as it has in fact been developed in
Anglo-American law, reflects that ordering. But why should tort law continue to
represent that ordering?
Let us first leave aside the issue of whether a general duty to rescue could
legitimately be introduced by way of judicial innovation. It may be that on the
best theory of adjudication this would be illegitimate because too great a
departure from precedent. But we would still need to ask whether the fact that
tort law has represented the ordering intrinsic to doing and suffering gives us
any reason to resist legislative enactment of a general tort duty to rescue.
There are two possible reasons. The first is that we accept a normative theory
of corrective justice that expresses the ordering of doing and suffering, and
we believe that the aim of tort law is to enforce corrective justice. This is
clearly one part of Weinrib's motivation for his rejection of tort duties to
rescue. I will return briefly to such arguments
Part II.C. But Weinrib's main concern would seem to be that introducing
liability for nonfeasance into tort law would render it incoherent, which would
in itself undermine its justification as a coercive regime.
ERNEST J. WEINRIB, THE IDEA OF PRIVATE LAW 38-45 (1995) [hereinafter WEINRIB,
PRIVATE LAW]. I cannot discuss this aspect of Weinrib's position here.
Lowns v. Woods, AUSTL. TORTS REP. P81-376 (1996). For critical discussion of
the decision, see Les Haberfield, Lowns v. Woods
and the Duty to Rescue
, 6 TORT L. REV. 56 (1998). For a more positive view, emphasizing the special
social role of physicians, see Natalie Gray
& James Edelman,
Developing the Law of Omissions: A Common Law Duty to Rescue?
, 6 TORTS L.J. 240, 243 (1998).
note 17, ch. 4; Benditt,
note 17, at 410-18; Anthony D'Amato,
"Bad Samaritan" Paradigm
70 NW. U. L. REV. 798 (1976); Franklin
note 13, at 1008-09.
n91 Leslie Bender argues for recognition of a tort duty to rescue as an expression
"ethic of care."
A Lawyer's Primer on Feminist Theory and Tort
38 J. LEGAL EDUC. 3, 33-36 (1988) (citing the feminist approach in CAROL GILLIGAN, IN DIFFERENT VOICE (1982)).
Linda C. McLain,
"Atomistic Man" Revisited: Liberalism, Connection, and Feminist Jurisprudence
65 S. CAL. L. REV. 1171, 1238-42 (1992) (criticizing Bender's approach, and noting that the ethic of care and
are not the same thing).
JOEL FEINBERG, HARM TO OTHERS ch. 4 (1984); McIntyre,
note 55, at 177; Woozley,
note 70, at 1293-99. Weinrib's early defense of a tort duty to rescue provides
an important exception.
note 85, at 266;
PATRICIA SMITH, LIBERALISM AND AFFIRMATIVE OBLIGATION 33-43 (1998).
n93 A different methodology can be found in Weinrib's work. Both his early defense
of a tort duty to rescue and his later position that rejects such a duty turn
on accounts of the overall coherence of private law.
WEINRIB, PRIVATE LAW,
note 88, at 46; Weinrib,
note 85, at 268-79.
note 92, at 15.
n95 JOHN STUART MILL, ON LIBERTY (1859) [hereinafter MILL, ON LIBERTY],
JOHN STUART MILL, THREE ESSAYS (Richard Wollheim ed., 1975).
JOHN STUART MILL, UTILITARIANISM (George Sher ed., 1979) (1861).
N97 MILL, ON LIBERTY,
note 95, at 15.
On the next page, Mill writes:
"I forego any advantage which could be derived to my argument from the idea of
at 16. Whether Mill is right that utilitarianism sets the limits of
interference with liberty at the prevention of actions that harm others is
dubious. Feinberg himself believes that the prevention of actions that cause
serious offense to others is a legitimate aim of the criminal law, and that the
harm principle and the
"offense principle" together make up the
"liberal position" on the moral limits of the criminal law. FEINBERG,
note 92, at 26.
n99 H.L.A. HART, LAW, LIBERTY AND MORALITY 6 (1963).
n100 MILL, ON LIBERTY,
note 95, at 15.
note 99, at 30-34.
n102 MILL, ON LIBERTY,
note 95, at 16-17.
note 92, at 130-50;
note 70, at 1293-99 (a similar view). Feinberg's view is consistent with David
Lyons's interpretation of Mill.
Liberty and Harm to Others, in
RIGHTS, WELFARE, AND MILL'S MORAL THEORY 102-03 (1994).
note 92, at 165-71. Feinberg also argues, in the alternative, that failures to
prevent harm can be cast as causes of the harm.
at 171-85. I find this a futile strategy.
text accompanying notes 55-58.
note 92, at 131-34.
note 96, at 48-49.
text accompanying note 87.
note 92, at 109-14, 144.
see also id.
at 117-18 (discussing Feinberg's concept of a victim).
note 92, at 46-74.
note 92, at 136-39.
n116 This is not quite accurate. There are cases where
back to his normal baseline that Feinberg would want to count as a genuine
benefit--for example, when
hands over to
the amount of money he has just lost at the races. In that case, we measure
benefit against a baseline of
post-loss condition. Why? Because
has no duty to benefit
in this case. Feinberg in effect simply defines all those cases where
"normal" baseline and where, as it intuitively seems,
a duty, as failures to prevent harm rather than failures to benefit. For
effective criticism of this aspect of Feinberg's view, including the point that
it may take Feinberg close to legal moralism, see McIntyre,
note 55, at 174-81.
n117 For very similar criticism of Feinberg on this point, see 2 F.M. KAMM,
MORALITY, MORTALITY: RIGHTS, DUTIES, AND STATUS 24 (1996). Kamm asks us to
suppose, in the drowning child case, that
the child had been born to its mother while she was in the pool, so that it had
never been on dry land, and it was drowning from the time it was born. Would
the fact that we would not be bringing it back to status quo ante be relevant
in deciding whether it had a human right to aid?
n118 For related criticism, though not specifically directed at Feinberg, see PETER
UNGER, LIVING HIGH AND LETTING DIE 42-45 (1996).
There is another way, more straightforward than Feinberg's, in which failure to
rescue could be seen as a failure to prevent harm--we could cast it as a
failure to prevent a deterioration in a person's well-being. On a broad sense
of benefit, such a failure is also a failure to benefit; we could, however,
draw a distinction between failure to prevent deterioration (call that failure
to prevent harm) and failure to improve a person's well-being (call that
failure to benefit). The relevant baseline is simply the victim's condition at
the time the potential rescuer first has an opportunity to act. This
distinction is more natural than Feinberg's and is congenial to tort law, but
it would not suit Feinberg's purpose. On this account, I could walk by an
accident victim suffering extreme pain that, we can imagine, will not get any
worse (though it may take a long time to get better), and fail to rescue the
victim. The proposed distinction would capture only those
"rescue cases" where intervention could prevent further disaster. Note that the Rhode Island
provision refers to a person who
"is exposed to, or has suffered, grave physical harm." R.I. GEN. LAWS
§ 11-56-1 (1994). The suggested alternative account would also include too much
in that a general duty to prevent all deteriorations in well-being (harms)
would require us to assist all people on a downward slope, not just those
involved in emergencies.
Legal Enforcement of Morality
85 J. CRIM. L. & CRIMINOLOGY 710 (1995). Greenawalt discusses criminal duties to rescue.
at 713-15. For a non-Millian liberal discussion of the enforcement of morality, see
RONALD DWORKIN, TAKING RIGHTS SERIOUSLY 240-57 (1977).
n120 For the distinction between negative and positive liberty, see ISAIAH BERLIN,
Two Concepts of Liberty, in
FOUR ESSAYS ON LIBERTY (1969). Many different accounts of both positive and
negative liberty have been proposed since Berlin's essay was first published in
1959. Furthermore, the notion of coercion that the definition of negative
liberty in the text employs is itself capable of various importantly different
ROBERT NOZICK, SOCRATIC PUZZLES 15-45 (1997); ALAN WERTHEIMER, COERCION
(1990). For our purposes we can leave this complication to one side.
MILL, ON LIBERTY,
note 95; JOSEPH RAZ, THE MORALITY OF FREEDOM (1986) (especially chapters one,
fourteen, and fifteen). Mill's word for what I, following Raz, call autonomy,
"individuality." Raz's book develops this aspect of Mill's view, while nevertheless departing
from Mill in several important ways.
n122 I follow Raz's account.
note 121, at 408. Obviously, if one values negative liberty for its
contribution to autonomy, then one will also value positive liberty directly.
at 148-57 (discussing the various different ways in which coercion interferes
with human autonomy). Raz's discussion makes it clear that this question is
more complicated than many have supposed.
n123 MILL, ON LIBERTY,
note 121, at 155. However, Raz's own version of liberalism is not rigidly
n124 Note that Raz, for reasons largely turning on his conception of political
authority, does not resist the idea of legal moralism.
note 121, at 414-15.
n125 Raz rightly points out that not all interferences with negative liberty reduce
text accompanying note 102.
A Theory of Strict Liability
, 2 J. OF LEGAL STUD. 151, 198-99 (1973); Schroeder,
note 50, at 192-93.
n128 I use quotation marks because the tendency of contemporary libertarians to
appeal to Locke for historical support,
, ROBERT NOZICK, ANARCHY, STATE, AND UTOPIA (1974), seems tendentious at best,
A. JOHN SIMMONS, THE LOCKEAN THEORY OF RIGHTS (1992) (especially chapter six).
n129 I do not mean to suggest that the two kinds of liberal theory that I discuss
here, the Millian and the libertarian, exhaust the liberal possibilities. Most
importantly, my discussion ignores the contemporary egalitarian liberalism most
significantly developed by John Rawls in
A Theory of Justice
(1993). This is because there is nothing in the Rawlsian account of liberty
that raises problems for the duty to rescue.
n130 Moore writes that:
Retributivism, when combined both with the principle of legality and the
insight that law as law does not even prima facie obligate citizen obedience,
yields the legal moralist theory of proper legislative aim: all and only moral
wrongs should be prohibited by the criminal law, for the reason that such
actions (or mental states) are wrongful (or culpable) and deserve punishment.
note 50, at 754;
see also id.
pt.1 (Moore's defense of retributivism);
at 661-65 (argument from retributivism to the legal moralist account of the
appropriate legislative aim in criminal law).
note 29, at 69; LaFAVE
note 29, at 10. We can take
"social good" to include a variety of different goals, such as the promotion of human
well-being, the protection of rights, and others; what it excludes, at least in
the usage I will adopt, is the goal of enforcing morality for its own sake.
note 50, at 746-50.
note 119, at 266-78; MOORE,
note 50, at 749. Moore does believe that there is
right to liberty, roughly corresponding to the right to privacy that has been
recognized under the U.S. Constitution.
note 49, at 763-77. As this narrower right is not relevant to the issue of
positive criminal duties, I will not discuss it here.
note 119; RONALD DWORKIN, A MATTER OF PRINCIPLE (1985). Dworkin's view falls
with Rawls's into the category of liberal theory I described,
note 129, as
"liberal egalitarian." Neither theory recognizes a general right to
"liberty," and neither offers a Millian account of the value of liberty in terms of human
well-being; both are rights-based liberal theories that reject a blanket right
to liberty and as such provide no liberty-based grounds to resist legal duties
to rescue. For Rawls, see especially JOHN RAWLS,
The Basic Liberties and Their Priority, Lecture VIII, in
POLITICAL LIBERALISM 289 (1993).
note 128, at 28-33. This makes it clear that, on this understanding of rights,
we cannot say that they protect especially important aspects of well-being,
such as, in this example, the value of not being assaulted by another person.
For this weaker understanding of rights, which is compatible with
utilitarianism, see generally Amartya Sen,
Rights and Agency
, 11 PHIL.
& PUB. AFF. 1 (1982).
n136 This understanding of the concept of rights seems to me the most useful for
philosophical discussion. It is obviously not the only possibility.
note 121, at 165-216, 245-66.
note 50, at 749.
n139 Moore believes that negative liberty is also a means to certain other goods,
in addition to positive liberty.
n141 This distinction has been embraced by the U.S. Supreme Court in cases
concerning physician-assisted suicide.
Vacco v. Quill, 521 U.S. 793, 793-94 (1997). For a broadly sympathetic discussion of Moore's appeal to the greater weight
of negative duties, see F.M. Kamm,
Action, Omission, and the Stringency of Duties
142 U. PA. L. REV. 1493 (1994).
note 10, at 1463-64.
THOMAS NAGEL, THE POSSIBILITY OF ALTRUISM (1970); THOMAS NAGEL, THE VIEW FROM
NOWHERE 152-53 (1986) [hereinafter NAGEL, THE VIEW].
note 53, at 27, 143.
n145 The duty not to violate rights is agent-relative precisely because it operates
as a side-constraint.
text accompanying note 135. If the duty were to minimize rights violations,
rather than not to violate rights, it would be an agent-neutral duty.
note 50, at 280.
note 10, at 1463.
note 50, at 279.
note 92, at 43; Freeman,
note 10, at 1478-79; McIntyre,
note 55, at 182.
n151 For Moore's discussion of the criticisms of Samuel Freeman and George
Fletcher, see MOORE,
note 50, at 281-82.
note 121, at 13.
n153 For criticism of the standard libertarian interpretation of Locke, see
note 128; Freeman,
note 10, at 1465-67; Heyman,
note 41, at 699-703.
, F.A. HAYEK, THE CONSTITUTION OF LIBERTY 16-19 (1960).
n155 The first sentence of
Anarchy, State, and Utopia
"Individuals have rights, and there are things no person or group may do to them
(without violating their rights)." NOZICK,
note 128, at ix.
note 154, at 21.
note 128, at 150-82.
n158 On fraud, see HAYEK,
note 154, at 143-44.
note 128, at ix. It is quite clear that in rejecting positive rights to aid,
"charity," contemporary libertarians depart from Locke.
note 128, at 307-52.
The core of the libertarian position, as I have said, is a constraint on the
legitimate content of law: Individuals' rights must not be violated. This
leaves open the (positive) aim of either criminal law or torts. Nozick,
following Locke, holds that at least part of the positive aim of law is the
enforcement of morality.
note 128, at 10-11. Locke puts the point by saying that individuals have the
right to enforce the law of nature.
2 JOHN LOCKE, TWO TREATISES OF GOVERNMENT ch. 2 (1698). For discussion, see
note 128, at 121-66. Libertarians need not, however, embrace this legal
moralist aim. They could hold that the point of law is solely to promote the
overall social good. For libertarians, this might be understood in terms of the
protection of rights (although that is not an especially plausible goal--it
does not seem to be worse to die in a fire than to be murdered). Be that as it
may, government pursuit of any conception of the social good is constrained, on
the libertarian view, by the prohibition on violating the rights of citizens.
The only occasion on which coercion does not violate rights is when it is for
the sake of preventing the coerced person from violating rights. Thus, we reach
the same result without invoking legal moralism.
note 127, at 198. Epstein's libertarianism is more fully set out in a later
Causation and Corrective Justice: A Reply to Two Critics
, 8 J.L. STUD. 477 488-89 (1979) [hereinafter Epstein,
Causation and Corrective Justice
"I attach a good deal of importance to the 'natural' set of entitlements that I
think are generated by a concern with individual liberty and private property
rights. In most cases I think that these rights are deserving of absolute
protection and vindication."
note 16, at 306-29 (chapter entitled
"Transfer Payments and Welfare Rights").
"The basic rules of private property are inconsistent with any form of welfare
Richard A. Epstein,
Nuisance Law: Corrective Justice and Its Utilitarian Constraints
8 J. LEGAL STUD. 49, 74-79 (1979) [hereinafter Epstein,
Causation and Corrective Justice, supra
note 160, at 487-90.
note 16, at 331-38.
See Richard Epstein's Takings, supra
note 16; Richard A. Epstein,
The Utilitarian Foundations of Natural Law
12 HARV. J.L. & PUB. POL'Y 713 (1989). Epstein had listed mutual benefit as a condition for deviation from the
libertarian constraint on law-making:
"Presence of implicit in-kind compensation from all to all that precludes any
systematic redistribution of wealth among the interested parties." Epstein,
Nuisance Law, supra
note 162, at 79. This does not make the departure from libertarianism any less
real, because (paternalistic) coercion for the sake of benefiting the coerced
is clearly a violation of individual rights on the libertarian account, and the
coercion is not itself for the sake of preventing the coerced person from
violating rights. What it shows is that Epstein's bedrock moral conviction is
that compulsory redistribution against a baseline of market outcomes is wrong;
all else flows from that.
In this Article, I will not discuss the outlandish claim that utilitarian moral
theory somehow dovetails with libertarianism. For criticism from a libertarian,
see Eric Mack, Comment,
A Costly Road to Natural Law
12 HARV. J.L. & PUB. POL'Y 753 (1989). Epstein replied to Mack's criticism.
Postscript: Subjective Utilitarianism
12 HARV. J.L. & PUB. POL'Y 769 (1989);
& Maimon Schwarzschild,
The Uncertain Relationship Between Libertarianism and Utilitarianism
19 QUINNIPIAC L. REV. 657 (2000).
n165 See the quotation from Sheldon Richman,
note 9. The libertarian opposition to legal duties to rescue is, in my
experience, very popular among law students.
Libertarianism Without Foundations
85 YALE L.J. 136 (1975). There are, in fact, several different arguments to the libertarian conclusion,
each of which requires separate evaluation. For a very useful survey, see WILL
KYMLICKA, CONTEMPORARY POLITICAL PHILOSOPHY 95-160 (1990).
PHILLIPE VAN PARIJS, REAL FREEDOM FOR ALL 15 (1995). Van Parijs offers this
example to make his point:
Think of an island which happens to be owned . . . by one of its inhabitants.
Providing it is difficult or expensive enough to leave the island, the owner
can impose on the other inhabitants any condition she fancies. If they are to
be allowed to earn their livelihood, they may have to work abysmally long
hours, for example, or give up their religion, or wear scarlet underwear. On a
libertarian account . . . such a society would not cease to be free.
, JAN NARVESON, THE LIBERTARIAN IDEA (1988). The first sentence of chapter one
of his book reads:
At the outset let us say that 'Libertarianism,' as the term is used in current
moral and political philosophy . . ., is the doctrine that the only relevant
consideration in political matters is individual liberty: that there is a
delimitable sphere of action for each person, the person's 'rightful liberty,'
such that one may be forced to do or refrain from what one wants to do only if
what one would do or not do would violate, or at least infringe, the rightful
liberty of some other person(s).
at 7. Friedrich Hayek writes as the first sentence of chapter one of his book:
"We are concerned in this book with that condition of men in which coercion of
some by others is reduced as much as possible in society." HAYEK,
note 154, at 11.
n169 HENRY SIDGWICK, THE METHODS OF ETHICS 274 (7th ed. 1907).
n170 For a very clear account of the equivocation between different senses of
liberty in the libertarian slogan, see MOORE,
note 50, at 744.
n171 Nozick calls the constraint against aggression
"the libertarian side constraint,"
note 128, at 33, whereas I have said the core libertarian right was the right
to be free of coercion. It is important to note that Nozick himself does not
trade in the misleading rhetoric about liberty I discuss in this paragraph.
Rawls on Liberty and Its Priority
40 U. CHI. L. REV. 534 (1973),
READING RAWLS 244 (Norman Daniels ed., 1975).
n173 See generally three articles by G.A. Cohen:
Capitalism, Freedom and the Proletariat, in
THE IDEA OF FREEDOM: ESSAYS IN HONOUR OF ISAIAH BERLIN 9 (Alan Ryan ed.,
Illusions About Private Property and Freedom, in
ISSUES IN MARXIST PHILOSOPHY 223 (J. Mepham
& D. Ruben eds., 1981);
Are Freedom and Equality Compatible?, in
ALTERNATIVES TO CAPITALISM 113 (J. Elster
& K.O. Moene eds., 1989);
The Argument from Liberty, in
IN HARM'S WAY: ESSAYS IN HONOR OF JOEL FEINBERG 16 (Jules L. Coleman
& Allen Buchanan eds., 1994).
It is not the case, of course, that only libertarians favor a moralized
conception of liberty. Indeed, Jeremy Waldron defends a legal duty to rescue
against the liberty objection by appealing to such a conception.
On the Road: Good Samaritans and Compelling Duties
40 SANTA CLARA L. REV. 1053, 1082-84 (2000).
n174 A fine illustration of the moralized conception of liberty and its dangers for
the champion of negative liberty is provided by this passage from Richard
A person does not exercise his liberties when he kills or enslaves another; he
does not vindicate his property rights when he steals from another. If he is
restrained from these actions by another, he cannot claim a loss of liberty,
but only the loss of an ability to act to which he was never entitled. Liberty
is best understood as freedom from force and falsehood, not as a maximization
of the things which are under one's disposition and control. And the principle
(at least as a matter of corrective justice) is the same whether the
infringements of the rights of another are great or small.
Causation and Corrective Justice, supra
note 160, at 489. The first two sentences make it clear that it is
individuals' rights, not any recognizable notion of freedom, that is at the
center of the account. In the next sentence, however, Epstein reverts to the
familiar assertion that it is negative and not positive liberty that matters.
In doing so, however, he all but contradicts the first two sentences. Though
the idea that veracity promotes liberty requires a moralized conception of
freedom, Epstein undermines the moralized view when he says that liberty should
be understood as freedom from force or falsehood
--as he has just stated that, on his view of liberty, the person who is forced
not to steal has
had his liberty infringed. On the moralized conception of freedom Epstein
announces in the first two sentences, he can at most claim that liberty is
force not, obviously enough, a particularly informative or uplifting idea. The
moralized conception of freedom returns to center stage with the focus on
rights in the final sentence.
n176 This may be the correct way to understand Weinrib's view in
The Case for a Duty to Rescue. See
note 88 (discussing Weinrib's later view);
JULES COLEMAN, RISKS AND WRONGS (1992); Stephen Perry,
Risk, Harm, and Responsibility, in
PHILOSOPHICAL FOUNDATIONS OF TORT LAW,
note 85, at 321. The corrective justice approach could be cast either in legal
moralist terms (the aim of tort law is to make sure that tortfeasors fulfill
their duties of compensation), in terms of the social good (the aim of tort law
is to vindicate the rights of tort victims), or a combination of the two.
n178 My own view is very close to that expressed by Jeremy Waldron.
Moments of Carelessness and Massive Loss, in
PHILOSOPHICAL FOUNDATIONS OF TORT LAW,
note 85, at 387.
n179 One alleged normative problem with legal duties to rescue that I have not
mentioned in this section is that such duties would reduce opportunities for
morally worthy acts. There are at least three different claims that have been
made along these lines.
First, Epstein makes a point that appears to derive from a mistaken
understanding of Kant's idea that actions done from inclination have no moral
worth. On Epstein's version of this idea, if inclination (to avoid legal
sanction) is present, an act in accordance with duty lacks moral worth, whereas
on Kant's account a person could still act from duty in such a case.
note 127, at 200; Weinrib,
note 85, at 266. Second, the claim is sometimes made that it is important to
preserve the realm of the morally supererogatory and that what is legally
required cannot be supererogatory. This claim seems to depend on some
(controversial) assumption of a standing moral obligation to obey the law. In
any case, legal duties to rescue would hardly exhaust the realm of the
supererogatory. Third, there is Landes and Posner's claim that enacting a legal
duty to rescue would lead to less acts of altruism because the possibility of
legal liability would lead potential altruists to avoid sites where they may
have the opportunity to rescue. Landes
note 13, at 120-22. The claimed substitution effect that this point depends on
note 243 and accompanying text. Apart from that, the point has some
plausibility--Aristotle remarked in
that one reason to favor the institution of private property is that it
affords greater scope for the virtue of generosity or liberality--so long as
one accepts what I do not: that acting morally or displaying virtue has value
note 70, at 1292-93.
note 4, at 316. In the relevant passage, Macaulay is specifically responding
to the provision in Livingston's code that required rescue when this could be
"without personal danger or pecuniary loss."
note 92, at 47 (referring to the problem as the
"slippery slope argument").
note 92, at 156; Woozley,
note 70, at 1299.
note 127, at 197-201; Schroeder,
note 50, at 193-94;
note 85, at 267-68.
, Louis Kaplow,
A Fundamental Objection to Tax Equity Norms: A Call for Utilitarianism
48 NAT'L TAX J. 497 (1995). For a critique along the same lines as that sketched in the text, see Ronald
Dworkin's discussion of the economic analysis of tort law. RONALD DWORKIN,
LAW'S EMPIRE 285-301 (1986). Though I agree with Dworkin's criticism of the
utilitarian argument he constructs in those pages, I disagree with his own
view, which turns on a non-instrumental division of private and public
at 299, and which is therefore a version of the dualist view I criticize.
text accompanying notes 185-89.
n185 For the full argument, see Liam Murphy,
Institutions and the Demands of Justice
, 27 PHIL.
& PUB. AFF. 251 (1998).
n186 For a canonical statement of the utilitarian position on this issue, see
note 169, at 457-59;
see also id.
at 87-88 (a note on Bentham).
n187 Note that my argument is against dualistic theories of
only, which are distinct from (what Waldron aptly calls) theories of politics,
"the nature and principled basis of political
on matters of justice and right" in the face of disagreement about these matters; democratic theory is largely
a theory of politics in Waldron's sense and it is, indeed, all about
institutional design. JEREMY WALDRON, LAW AND DISAGREEMENT 1-4 (1999).
n188 JOHN RAWLS, A THEORY OF JUSTICE 115, 333-37 (1971).
ROBERT H. MYERS, SELF-GOVERNANCE AND COOPERATION 59-68 (1999); Larry
Scheffler on the Independence of Agent-Centered Prerogatives from
, 84 J. PHIL. 277 (1987); Robert H. Myers,
Prerogatives and Restrictions from the Cooperative Point of View
, 105 ETHICS 128 (1994).
n190 My account of this problem draws on LIAM MURPHY, MORAL DEMANDS IN NONIDEAL
THEORY 127-33 (2000).
note 118, ch. 2;
at 53-54 (listing the nine factors). Unger himself does not use the
note 118, at 34-35 ("The CB Radios");
note 70, at 1289 ("Sometimes it happens that a ship's Mayday signal is picked up only by a ham
radio operator halfway around the world; we expect him to report it as quickly
as he can . . . ."). For criticism of Unger's argument, see F.M. Kamm,
Rescue and Harm
, 5 LEGAL THEORY 1 (1999).
Waldron defends duties to rescue from
"general apprehensions about the law requiring abstract benevolence," by appeal to the factor of proximity. Waldron,
note 173, at 1075, 1097-1103.
note 118, at 28-29, 77-79.
n195 Unger's argument is not uncontroversial, and a full discussion cannot be
attempted here. For criticism, see F.M. Kamm,
Faminine Ethics: The Problem of Distance in Morality and Singer's Ethical
SINGER AND HIS CRITICS (Dale Jamieson ed., 1999) [hereinafter Kamm,
n196 For arguments in defense of a distinct normative category comprising rescue
cases, see SMITH,
note 92, ch. 3; Kamm,
Faminine Ethics, supra
note 195, at 198-202; Kamm,
note 192, at 19-20.
One common argument in favor of the special normative status of rescue
situations is that such situations could be the subject of a mutually
beneficial ex ante agreement.
note 168, at 242-44; D'Amato,
note 90, at 805-08; Freeman,
note 10, at 1480-81; Eric H. Grush,
The Inefficiency of the No-duty-to-rescue Rule and a Proposed
"Similar Risk" Alternative
146 U. PA. L. REV. 881 (1998); Richard L. Hasen,
The Efficient Duty to Rescue
, 15 INT'L REV. L.
& ECON. 141 (1995). But such an agreement could only be hypothetical, and it is
not clear why hypothetical bargains should have any moral force at all.
note 119, at 150-52. Furthermore, the argument requires strict unanimity, and
this seems dubious. Take a well-guarded billionaire who never swims or takes
personal risks; such a person's resources might well be called on in
emergencies (imagine some variation on Unger's case,
"Bob's Bugatti," UNGER,
note 118, at 136) while his chances of needing a stranger's assistance are
close to zero.
note 53, at 109. Lastly, we may note that if the argument from mutual benefit
succeed, it would surely be a wonderful coincidence that the agreed upon duty
was precisely the minimal duty to rescue and not one requiring significant
sacrifice in some circumstances. I take this point from some remarks attributed
to F.M. Kamm in Samuel Freeman's article,
note 10, at 1480 n.96; Freeman's response to Kamm's point (essentially that
demanding general duties of
are, indeed, too demanding) seems irrelevant to the mutual benefit argument.
n197 I owe this example to Shelly Kagan.
Liberalism, Bad Samaritan Law, and Legal Paternalism
, 106 ETHICS 4, 17-19 (1995). The costs of rescue may be especially high if the
rescuer's ethical sense requires her to take responsibility for the saved life.
BOUDU SAUVE DES EAUX (Sirius 1932) (film by Jean Renoir).
, RICHARD B. BRANDT, A THEORY OF THE GOOD AND THE RIGHT 276-77 (1979) ("Act utilitarianism makes extreme and oppressive demands on the individual, so
much so that it can hardly be taken seriously; like the Sermon on the Mount, it
is a morality only for saints.").
n199 I here disagree with Schroeder,
note 50, at 192-97.
note 190, ch. 2. Note that extreme demands can come from parts of morality
including such paradigm negative duties as the duty not to kill, but such
duties are never accused of being absurd because extemely demanding (no one
says that it is absurd to ask me to forego the million dollars I would gain if
I killed you; many say it is absurd to expect a millionaire to give a million
dollars to famine relief).
Does Consequentialism Demand Too Much?
, 13 PHIL.
& PUB. AFF. 239 (1984); F.M. Kamm,
Supererogation and Obligation
, 82 J. PHIL. 118 (1985). This factor raises significant complications that I
will simply ignore in this Article. For discussion, see MURPHY,
note 190, chs. 3, 5.
SHELLY KAGAN, THE LIMITS OF MORALITY (1989) (especially chapter ten,
"Extraordinary Morality"); UNGER,
note 118 (especially chapter six,
"Living High and Letting Die Reconsidered: On the Costs of a Morally Decent Life"). Some philosophers who officially embrace utilitarianism try to avoid the
problem of demands by arguing, in effect, that most people are not really in a
position to do much good for anyone else.
, Frank Jackson,
Decision-Theoretic Consequentialism and the Nearest and Dearest Objection
, 101 ETHICS 461 (1991).
THOMAS NAGEL, EQUALITY AND PARTIALITY (1991) [hereinafter NAGEL, EQUALITY];
NAGEL, THE VIEW,
note 143, at 189-208 (chapter entitled
"Living Right and Living Well"); SAMUEL SCHEFFLER, HUMAN MORALITY (1994); SAMUEL SCHEFFLER, THE REJECTION OF
CONSEQUENTIALISM (rev. ed. 1994).
note 190, chs. 3-4.
CHARLES FRIED, RIGHT AND WRONG 130 (1978); MURPHY,
note 190, ch. 5; PARFIT,
note 53, at 30-31; HENRY SHUE, BASIC RIGHTS: SUBSISTENCE, AFFLUENCE, AND U.S.
FOREIGN POLICY 114-19 (2d ed. 1996); Dan W. Brock,
Defending Moral Options
, 51 PHIL.
& PHENOMENOLOGICAL RES. 909, 912 (1991); L. Jonathan Cohen,
Who Is Starving Whom?
, 47 THEORIA 65 (1981); Carlos Rosenkrantz,
Igualitarismo y Libertarianismo: Politica no antropologia
, 7 REVISTA DEL CENTRO DE ESTUDIOS CONSTITUTIONALES 193 (1990).
n205 For details, see MURPHY,
note 190, chs. 5-7.
note 92, at 158-59; Freeman,
, note 10, at 1467-68; McIntyre,
note 55, at 181-82; Menlowe,
note 10, at 48; Weinrib,
note 85, at 272-73, 292.
n207 A crucial point not acknowledged by the authors cited in the previous note.
note 10, at 1467 ("There is in fact little that we as individuals can do to help the destitute
(except in emergencies, and then the relevant duty is the duty to aid the
distressed). The problem of destitution is largely one of background justice .
note 185, at 278-84.
n208 See Thomas Nagel's discussion of the
"moral division of labor." NAGEL, EQUALITY,
note 202, 53-62.
n209 This is not to deny the obvious fact that each of us is especially well-placed
to promote the interests of a certain small group of intimates.
n211 Stipulating, as an element of the offense, that the defendant must have known
that this was not the case would render the duty largely inert in practice.
Proof beyond reasonable doubt that a person knew that she could have conveyed a
cost-justified benefit to someone, the need for which, as she also knew, would
not have been eradicated under ideal institutions, would seem to be a very
difficult matter. The only cases in which it seems plausible to have no
reasonable doubt about the defendant's knowledge of these two facts are,
indeed, rescue cases.
n212 Though perhaps not by the utilitarian legal moralist. On such a view, these
kinds of costs could be an acceptable price to pay for the sake of enforcing
individuals' moral duties.
n213 It seems likely that such a duty would, in any case, be unconstitutionally
vague. See the discussion of vagueness in the context of Wisconsin's duty to
aid provision in
State v. La Plante
, 521 N.W.2d 448 (Wis. Ct. App. 1994).
n214 What about a specific duty to call the fire brigade if one knows that a major
art museum is burning down? This could potentially produce a great benefit even
if it only once applies to an individual's situation, so long as that
individual is the only person in a position to raise the alarm. The more
important the property, however, the less likely it is that this condition
would be met, especially under ideal institutions.
n215 There is also this further problem. One of the aspects of rescue situations
that makes them easy to identify is that they involve severe suffering or
peril. It is not as easy to know whether one is confronted with emergencies
more broadly conceived. Stipulating, as an element of the offense, knowledge
that one is in a situation of the relevant kind does not fully solve the
problem here, because a person may be apprehensive about the way in which this
mental state requirement would be applied in practice and thus find herself
wasting time and energy searching for emergencies in his neighborhood. For a
similar point supporting the limitation to cases of severe suffering or threat
to life, see
text accompanying note 224.
n216 If we are not legal moralists
we reject that constraint, then there would presumably be no limit to
legitimate legal demands on people, even if we accept either the limited or
collective principle of
as our principle for both personal conduct and institutional design. The prima
facie implausibility of such a view shows the prima facie plausibility of the
constraint against forcing sacrifice beyond what a person is morally required
to sustain voluntarily.
text accompanying note 42.
note 1, at 636-38 (discussing European approaches to this issue).
n219 Note that this point does not depend on the assumption that the aim of the
criminal law is to promote the social good, because a legal moralist
utilitarian holds that if it makes no utilitarian sense for a particular person
to act, there is no moral requirement that she act.
Wallace M. Rudolph,
The Duty to Act: A Proposed Rule, in
THE GOOD SAMARITAN AND THE LAW,
note 1, at 243.
n221 In this respect professional rescuers--surf lifeguards, firemen, search and
rescue personnel, and others--are unlike medical practitioners.
n222 For some worries about leaving such decisions to jurors, see Malm,
note 197, at 21-22 (arguing that it is illiberal to allow jurors to decide
which losses count as reasonable).
n223 One reason for opposition to criminal duties to rescue may be precisely an
awareness of the expressive effect of the lack of such a duty: In lacking such
a duty, our law expresses the idea that we have no responsibilities to mere
Jason S. Johnston,
Punitive Liability: A New Paradigm of Efficiency in Tort Law
87 COLUM. L. REV. 1385, 1395-96 (1987) (explaining how raising the burden of proof in civil liability cases reduces
the risk of inducing too much investment in care). We have here a further
reason--additional to those set out
Part IV.B.--for preferring a criminal to a tort duty to rescue.
n225 See Unger's example,
"The Yacht." UNGER,
note 118, at 63-67.
note 85, at 882-94. Levmore draws on the analysis of Landes and Posner.
n227 Larger rewards seem likely to provide incentives for potential rescuers to
create demand for their own services.
note 85, at 886-87.
John P. Dawson,
Rewards for the Rescue of Human Life?, in
THE GOOD SAMARITAN AND THE LAW,
note 1, at 63-87.
Desmond S. Greer,
A Transatlantic Perspective on the Compensation of Crime Victims in the United
85 J. CRIM. L. & CRIMINOLOGY 333, 350 (1994).
n230 Under current law, a nonprofessional rescuer is generally not entitled to
restitution even of his out of pocket expenses.
RESTATEMENT OF RESTITUTION
§ 116(a) cmt. a (1937) (referring to comment c of section 114;
"intent to charge" a requirement for recovery);
RESTATEMENT (SECOND) OF RESTITUTION
§ 3 cmt. c (Tent. Draft No. 1, 1983) ("It is virtually a precept that the merit of emergency salvage efforts is, and
should be, its own reward."). The existence of a criminal duty to act would presumably rebut this
presumption of gratuitousness.
2 GEORGE E. PALMER, THE LAW OF RESTITUTION
§ 10.3, at 369-70 (1978
& Supp. 1999); Ross A. Albert, Comment,
Restitutionary Recovery for Rescuers of Human Life
74 CALIF. L. REV. 85 (1986); John W. Wade,
Restitution for Benefits Conferred Without Request
19 VAND. L. REV. 1183 (1966).
For discussion of compensation for rescuers from a
"liberal-communitarian" perspective, see Heyman,
note 41, at 746-50.
n231 Adjudication of this question of where the compensation should come from would
require, inter alia, discussion of the normative basis of the law of
restitution; it is beyond our scope here. For general discussion of restitution
in the rescue context, including discussion of the civil law doctrine of
, see Hanoch Dagan,
In Defense of the Good Samaritan
97 MICH. L. REV. 1152 (1999). For discussion of the possibility of a state compensation scheme for rescuers,
see Norval Morris,
Compensation and the Good Samaritan, in
THE GOOD SAMARITAN AND THE LAW,
note 1, at 136-39. One reason why a state compensation scheme would be
preferable is that it would allow greater certainty and flexibility with
respect to the amount of compensation, a difficult issue under restitution
doctrine. Suppose that a reasonable rescue nevertheless turned out to be fatal
for the rescuer. Should the rescuee be liable for this loss of life? For
discussion of this problem, with focus on a German case on point, see Dawson,
n232 Note that a legal moralist
"utilitarian" would not object to the compensation of rescuers' expenses if that were for
the best in utilitarian terms: on such an approach what matters is that people
do their utilitarian duty and suffer retribution for not doing so, not that
they suffer costs
note 1, at 117 (stressing this advantage of duty-to-rescue provisions and the
early--early nineteenth century--appreciation in Continental Europe of the
distinction between liability for omission and liability for commission by
note 55, at 187-90. For philosophical discussion of the general significance
of the distinction between killing and letting die, see the articles collected
Killing and Letting Die
. KILLING AND LETTING DIE (Bonnie Steinbock
& Alistair Norcross eds., 2nd ed. 1994);
note 117, at 13-200.
In most Continental European jurisdictions, the duty-to-rescue statute cannot
be used to ground homicide by way of commission by omission.
note 1, at 95; Feldbrugge,
note 1, at 648-52. The very fact that U.S. duty-to-rescue provisions specify
specific and minimal sanctions should block their use in a
commission-by-omission prosecution as a matter of statutory interpretation, but
it would seem that an explicit provision to that effect would be wise; for a
suggestion along these lines, see Jay Silver,
The Duty to Rescue: A Reexamination and Proposal
26 WM. & MARY L. REV. 423, 436 (1985) (proposing rescue statute providing that
"nothing contained in this Act shall alter existing law with respect to
liability for criminal homicide"). In
State v. Cabral
, 810 P.2d 672, 677 (1991), the Hawaii Intermediate Court of Appeals prospectively approved the use on
retrial of Hawaii's duty-to-assist provision as the basis of a homicide charge.
This had not been done in the original trial; it was apparently not done in the
State v. Cabral, 883 P.2d 638, 639 (Haw. Ct. App. 1994).
I cannot in this Article evaluate the appropriateness of commission by omission
liability generally. One obvious point is that the existence of a relationship
is likely to increase culpability. For discussion, see generally FLETCHER,
note 27, at 581-634; Glanville Williams,
Criminal Omissions--The Conventional View
107 L.Q. REV. 86 (1991).
n234 The Australian Northern Territory's provision, Criminal Code Act section 155,
provides for up to seven years imprisonment, but requires that the failure to
"callous," which has been interpreted to mean
"more than normal" intent in
Salmon v. Chute
, 4 N.T.L.R. 149 (Austl. N. Terr. 1994). This still leaves matters rather
unclear; for a brief discussion see John T. Pardun,
Good Samaritan Laws: a Global Perspective
20 LOY. L.A. INT'L & COMP. L. REV. 591, 594-97 (1998). The geography of the Northern Territory (extremely sparsely populated with a
very harsh environment) greatly heightens the need for private efforts to
assist, for example, stranded motorists.
n236 This further reduces the seriousness of the problem of the possibility of
excessive expenditures due to mistrust of triers of fact.
note 224 and accompanying text.
note 190, ch. 7. I argue there that the strong commonsense moral condemnation
of a nonrescuer is perhaps based more on implicit judgments of bad character
than on--what seems to be the case--judgments of wrong action. The upshot is
that failure to rescue will suggest an undesirable set of motivations even
where it is not strictly wrong.
n238 If this is right, and I am also right that at the level of personal morality
there may not be a general requirement to perform easy rescues, but rather a
requirement to develop the kind of motivations that would lead one to be
strongly disposed always to perform easy rescues,
note 237, then the upshot would be that as a requirement on
, legal duties to rescue can be defended even though they are not in the end
part of the morality of personal conduct. This is exactly the reverse of what
is generally supposed.
n239 It is true that the provision of compensation renders the issue of the demands
of the legal duty to rescue moot at the level of individuals who must act on
the duty. But could not the limited nature of the demands of the limited and
collective principles of
undermine the case for the legal duty to rescue when they are applied at the
level of institutional design as accounts of the aim of promoting the social
good? I assume that the answer is no, though establishing this for our two
principles would be complicated and involve speculation. What needs to be
established is that legal duties to rescue could be part of an overall scheme,
support of which does not burden any member of the community (by way of the
cost of funding the scheme) more than is justified by each principle.
n240 It may be worth just mentioning some problems that have been raised. There is
the problem of the officious intermeddler: the busybody who believes, for
example, that the existence of a criminal duty to aid makes it his business to
supervise the neighbors' child-rearing practices.
note 85, at 281 (for the particular example I thank Christopher Eisgruber).
This does seem to be a possible problem, but it is hard to see why
problem could not be solved by careful drafting; the Rhode Island provision
specifies that there must be an emergency that exposes someone to grave
physical harm. It is not plausible to interpret
"grave physical harm" to include the effects of spanking or playing in the sun without a hat.
A more serious problem concerns the potential for prosecutorial abuse. In
brief, the concern is that prosecutors have been given further means for the
extraction of guilty pleas and cooperation with the prosecution.
Svetlana Kornfeind, Deputization and the Statutory Duty to Aid (May, 1997)
(unpublished paper, New York University Law School) (on file with author);
Daniel B. Yeager,
A Radical Community of Aid: A Rejoinder to Opponents of Affirmative Duties to
71 WASH. U. L.Q. 1, 36-38 (1993). I do not know how serious a problem this is, but to the extent that it is, it
also seems to be solvable by drafting. A duty-to-notify bill that was
considered (and then abandoned) by the Florida legislature in 1999 shows how it
might be done. Subsection (2) of that bill provided:
This section may not be construed to apply to a person either who is prosecuted
as a principal in the first degree or an accessory after the fact to an
offense, or who is prosecuted for attempting, soliciting, or conspiring to
commit an offense, when the prosecuted offense has occurred in the course of
the same criminal conduct, transaction, or episode as the criminal offense
which exposes the victim to bodily injury.
H.B. 69, 1999 Reg. Sess. (Fl. 1999).
Finally, a recent essay by Eugene Volokh raises the possibility that duty to
report provisions would have the bad effect of deterring witnesses who failed
to report at the time of the crime but who might otherwise have cooperated with
the police at a later date.
Eugene Volokh, Essay,
Duties to Rescue and the Anticooperative Effects of Law
88 GEO. L.J. 105 (1999).
n241 It is worth noting that the different countries of Continental Europe did not
simply inherit their legal duties to rescue en masse from some common
source--Roman Law or the Napoleonic Code. Rather, such duties were introduced
independently and at different times in different countries. The earliest was
Russia in 1845, but in most countries such duties were first enacted after
World War II. Feldbrugge,
note 1, at 630-31.
n242 There is a danger that courts will find a civil duty just in virtue of the
existence of the criminal duty.
KEETON ET AL.,
§ 36; Franklin
note 13, at 1010-17. Once again, one would hope that sensible statutory
interpretation would prevent this, but as it may not, an express provision to
the effect that mere violation of the criminal duty to rescue does not ground
civil liability would be desirable. Hawaii's provision requiring aid for crime
victims excludes civil liability for violators. HAW. REV. STAT. ANN.
§ 663-1.6(c) (Michie 1995).
n243 The doubts expressed by Professors Landes and Posner on this score seem
A Theoretical Fox Meets Empirical Hedgehogs: Competing Approaches to Accident
82 NW. U. L. REV. 837, 841 (1988); Richard L. Hasen,
The Efficient Duty to Rescue
, 15 INT'L REV. L.
& ECON. 141, 146 (1995); Levmore,
note 85, at 890-91 (discussing Landes and Posner,
note 13, at 119-27); Wright,
note 85, at 272;
note 127, at 190 (noting that the no-duty-to-rescue rule seems inefficient
according to the
formula for negligence).
n244 Under a cost-justified rescue rule, the burden of required rescue will always
be less than the loss to the victim that rescue would prevent, and thus it will
always be cheaper to sustain the burdens of rescue than to pay compensatory
note 13, at 1009; Wright,
note 85, at 274. There are many cases where full compensatory liability for
misfeasance can seem disproportionate to fault as well--as for example when a
mildly negligent moment leads to disastrous consequences. On this issue, see
note 178. The upshot is that my objection to tort liability for failure to
rescue may not be neutral with respect to existing features of tort law.
Nevertheless, the issues are distinct, as the objection in the text applies
even to an explicitly intentional tort of failure to rescue.
n246 AUSTL. TORTS REP. P81-376.
note 90, at 802-04.
note 240 and accompanying text.
note 1, at 111-15.
n252 One danger of Weinrib's methodology,
note 85, at 275, 293; sources cited
note 88, is that it encourages this mistake,
note 17, at 89-108 (an illuminating discussion of this kind of mistake).
Glendon seems to believe that the mistake is ineradicable in the United States,
note 17, at 102, and therefore suggests that authors of legal opinions,
especially on constitutional questions, should be careful to remind readers
that the legal question and the question of political morality are different.
See also Mark Kelman's discussion of
"conflating legal solubility with the existence of a problem." MARK KELMAN, A GUIDE TO CRITICAL LEGAL STUDIES 275-79 (1987). Also to the
point is the book's next section, on
"the synthetic individualist tradition."
note 16, at 318-19; Epstein,
note 127, at 199-203.