University of Chicago Law School

Chicago Unbound

Journal Articles Faculty Scholarship

1986

Waiting for Rescue: An Essay on the Evolution and Incentive

Structure of the Law of Affirmative Obligations
Saul Levmore

Follow this and additional works at: https://chicagounbound.uchicago.edu/journal_articles

Part of the Law Commons

Recommended Citation
Saul Levmore, "Waiting for Rescue: An Essay on the Evolution and Incentive Structure of the Law of
Affirmative Obligations," 72 Virginia Law Review 879 (1986).

This Article is brought to you for free and open access by the Faculty Scholarship at Chicago Unbound. It has been
accepted for inclusion in Journal Articles by an authorized administrator of Chicago Unbound. For more
information, please contact [email protected].
 VIRGINIA LAW REVIEW
VOLUME 72 AUGUST 1986 NUMBER 5

WAITING FOR RESCUE: AN ESSAY ON THE EVOLUTION

AND INCENTIVE STRUCTURE OF THE LAW OF
AFFIRMATIVE OBLIGATIONS

Saul Levmore*

T HE legal distinction between omissions and commissions is

useful but vulnerable. It is easy to point out that commissions
are indistinguishable from omissions because both may be excused
in a similar fashion. Every law student learns that omissions some-
times lead to liability, as when physicians may be liable for failing
to employ certain procedures, and that negligent commissions oc-
casionally do not lead to liability, as may be the case when a land-
owner's commissions injure a trespasser.' The clever law student is
even able to turn commissions into omissions by arguing, for exam-
ple, that negligent driving is nothing more than the failure to
brake. Despite its vulnerability, however, the omission-commission
distinction survives as a useful descriptive tool. The common law
will, for example, neither penalize one who fails to rescue a stran-
ger from drowning nor normally impose liability on someone who
fails to disclose information as regularly as it will penalize one who
affirmatively misdiscloses. 2 The existence or non-existence of a

* Professor of Law, University of Virginia. I am indebted to Glen Robinson, Roberta Ro-

mano, Robert Scott, Helen Snyder, and participants at a Law and Economics Workshop at
Harvard and a Legal Studies Workshop at the University of Virginia for many constructive
suggestions and conversations.
' See generally W. Prosser & W. Keeton, The Law of Torts § 56 (5th ed. 1984) (discussing
the omission-commission distinction in American tort law).
I See id. at 373-74 (liability for omission often based on the finding of a special relation-
ship); see also Weinrib, The Case for a Duty to Rescue, 90 Yale L.J. 247 n.1 (1980) (listing
several special relationships).

HeinOnline -- 72 Va. L. Rev. 879 1986

880 Virginia Law Review [Vol. 72:879

"duty" is said to explain the imposition of liability in some of

these cases, but this duty appears inseparable from the question of
whether a commission or an omission is at issue.3 Although new
duties have been imposed over time,4 the overall impression has
remained relatively constant; our legal system is seen as one that
regularly deters antisocial commissions, often compensates victims
of commissions, rarely deters antisocial omissions, and virtually
never rewards rescuers. 5 It is sometimes said that the law's absolu-
tion of nonrescuers reflects a disinclination to legislate morality.6
But this is an inadequate explanation of the law because wrongful
commissions are the stuff of both legislation and moral outlooks.
What, then, does explain the often criticized disinclination to im-
pose a duty to rescue, a duty to disclose information, a duty to
return (rather than ignore) lost property, and other affirmative ob-
ligations? Although the treatment of rescue is not necessarily re-
lated to that of other matters easily tied to the omission-commis-
sion distinction, it is useful to focus mainly on rescue and to expect
that a better understanding of this "nonobligation" will inform
others. It is especially intriguing to focus on the duty to rescue in
light of the fact that so many other legal systems do penalize one
who fails to rescue.7
In this essay I suggest three themes, or arguments, that inform
the omission-commission distinction in general, and the treatment
of rescue in particular. Although these arguments are interactive,

3 See Handiboe v. McCarthy, 114 Ga. App. 541, 151 S.E.2d 905 (1966) (child drowning in
pool); Osterlind v. Hill, 263 Mass. 73, 160 N.E. 301 (1928) (drowning); Yania v. Bigan, 397
Pa. 316, 155 A.2d 343 (1959) (drowning); W. Prosser & W. Keeton, supra note 1, § 106, at
737-38 (collecting nondisclosure cases).
' See infra notes 46-65 and accompanying text.
'See Dawson, Rewards for Rescue of Human Life?, in The Good Samaritan and the Law
63 (J. Ratcliffe ed. 1966)
6 See 1 F. Lawson & B. Markesinis, Tortious Liability for Unintentional Harm in the
Common Law and the Civil Law 72 (1982); Gregory, The Good Samaritan and the Bad: The
Anglo-American Law, in The Good Samaritan and the Law, supra note 5, at 23.
1 See Feldbrugge, Good and Bad Samaritans: A Comparative Survey of Criminal Law
Provisions Concerning Failure to Rescue, 14 Am. J. Comp. L. 630 (1966); Rudzinski, The
Duty to Rescue: A Comparative Analysis, in The Good Samaritan and the Law, 91, 108-15,
supra note 5, at 23. German courts have allowed recovery of a rescuer's expenses. The Aus-
trian, Polish, and Czech Civil Codes explicitly allow a rescuer to recover his expenses. Portu-
gal, Holland, Italy, Norway, Soviet Union, Turkey, Denmark, Poland, Germany, Rumania,
France, Hungary, and Czechoslovakia have statutes that create a general duty to rescue. Id.
at 91 n.3.

HeinOnline -- 72 Va. L. Rev. 880 1986

1986] Rescue

they are best developed separately. Part I explores the behavioral

effects of rules affecting potential rescuers. The discussion consid-
ers the rewards and penalties that may be imposed to encourage
rescues and seeks to explain why, in fact, we find an interesting
symmetry in most legal systems: "sticks" (penalties for nonrescue)
accompany "carrots" (rewards for rescue) and the absence of sticks
seems to be correlated with the denial of carrots. Part II applies a
theme developed elsewhere8 and argues that the variety that is
found among different legal systems in the treatment of rescue and
in our own system's treatment of finders, disclosers, and salvors at
sea is largely within a predictable range. This variety may reflect
different conjectures about various behavioral effects as much as it
is a product of historical accidents or diverse cultural and political
attitudes." The parallel analyses of the rules pertaining to finders,
disclosers, salvors, and more familiar rescuers-all of whom must
decide whether to take affirmative steps-highlights the usefulness
of thinking about the behavioral effects of legal tools.
The discussion in Part III considers the likely evolution of the
treatment of rescue in American law and suggests that experimen-
tation with or evolution of the use of penalties may have been im-
peded by our legal system's need to develop first its approach to
"multiple commission" problems. Now that a number of doctrinal
steps have been taken to allocate liability among multiple
tortfeasors, or committers, there is some reason to think that the
allocation among potential omitters, and therefore the develop-
ment of a more general duty to rescue, will not be far behind. I do
not go so far as to predict a developing obligation to donate money
to the poor; indeed, it is easy to predict that there will continue to
be no obligation to do that which a legislature or society as a whole
could easily do. But I do conclude that there will continue to de-
velop an obligation on the part of persons especially well situated
to effect easy rescues. 10

' Levmore, Rethinking Comparative Law: Variety and Uniformity in Ancient and Modem
Tort Law 1-6 (1986) (unpublished manuscript on file with Virginia Law Review
Association).
See infra text accompanying note 44.
10 Although I do not undertake to examine the parameters of easy rescue, it is clear that
as such an obligation develops the law will need to define the level of risk that a potential
rescuer must absorb, and this definition will surely reflect process-cost considerations. As
will become clear, many obligations already required by the law impose such process costs.

HeinOnline -- 72 Va. L. Rev. 881 1986

 Virginia Law Review [Vol. 72:879

I. THE INCENTIVE STRUCTURE FOR RESCUE

Rewards and penalties will motivate a potential rescuer only if

he is aware of these incentives and then only if in a crucial mo-
ment he is able to be influenced by this knowledge. Although there
is some reason to believe that citizens are aware of the rules re-
garding rescue and that many rescues require prepositioning or af-
ford time for contemplation, arguments for and against various
rules must surely rest on a sense that incentives work on the mar-
gin.11 I do not argue in this essay that one set of rules will best
promote socially desirable rescues or, indeed, that these penalties
and rewards have substantial behavioral-as opposed to sym-
bolic-effects. I imagine, instead, that lawmakers will try to pro-
mote certain kinds of behavior with legal rules and will be espe-
cially eager to enact rules that either work on the margin or are, at
worst, harmless.1 2 This is a positive and not a normative exercise.
The law could, of course, simply give a successful rescuer a re-
ward for his deed. This "carrot" might be paid by the rescued vic-
tim, by the state, or by some third party. If the rescue is of prop-
erty rather than life, as when a rescuer brings a fire under control
and prevents further loss of property, then this reward could easily
be varied according to the value of the property saved. Rewards for
the rescue of life or property could be very substantial, sufficient
to promote the equivalent of a competitive industry of rescuers, or
barely enough to cover the rescuer's expenses. It is easy to quibble
with this classification and insist that mere reimbursement of ex-
penses simply erases a penalty otherwise incurred by a rescuer, but
both because these expenses are not imposed by the law and be-
cause the rescuer may well regard a package of reimbursement,
public acclaim, and private gratitude as a substantial carrot, it is
convenient to think of a legal right to reimbursement as a reward.
Apart from any other rewards, the law could exempt rescuers from
liability for some or all of the injuries they happen to impose while
engaged in their rescue."' Finally, rescuers who are themselves in-

" Zeisel, An International Experiment on the Effects of a Good Samaritan Law, in The
Good Samaritan and the Law, supra note 5, at 209, 210 (German, Austrian, and American
residents displayed fair knowledge of their respective countries' rescue laws).
12 Some rewards may be harmless in the sense that even if they do not stimulate desirable

rescues it is unlikely that they would suppress such rescues.

13 Every state has adopted some form of legislation that restricts the liability of partici-

HeinOnline -- 72 Va. L. Rev. 882 1986

1986] Rescue 883

jured while attempting to render assistance could be allowed to

collect from the victim (whether or not actually rescued) or some
14
other party.
The available penalties are even simpler to sketch. The law
must, if it seeks to penalize nonrescuers, first identify a nonrescuer
and then provide some penalty, or "stick." Both because penalties
are often contained within the criminal law, where doubts are usu-
ally resolved in favor of the accused, and because it would defeat
the instrumental purpose of the law to punish nonrescue too seri-
ously,15 the identification of nonrescuers is usually limited to per-
sons who should realize that they could easily take some step to
save life or valuable property. 6 It is very unlikely, for instance,
that any legal system would penalize someone who only called the
police-even though a more direct step would have been more use-
ful-or who genuinely believed that some other onlooker had al-
ready called for help. 17 Once a nonrescuer is identified, the penalty

pants in rescue operations. See Comment, Good Samaritan Laws-The Legal Placebo: A
Current Analysis, 17 Akron L. Rev. 303 (1983).
'4 See Wagner v. International Ry. Co., 232 N.Y. 176, 133 N.E. 437 (1921) (allowing an
unsuccessful rescuer to recover for his injuries from a negligent third party).
See infra text accompanying notes 31-33.
16 For example, the Vermont and West German statutes are illustrative of the emphasis
in duty-to-rescue laws on encouraging easy or simple rescues. Vt. Stat. Ann. tit. 12, § 519
(1973):
a) A person who knows that another is exposed to grave physical harm shall, to the
extent that the same can be rendered without danger or peril to himself or without
interference with important duties owed to others, give reasonable assistance to the
exposed person unless that assistance or care is being provided by others.
b) A person who provides reasonable assistance in compliance with subsection (a)
of this section shall not be liable in civil damages unless his acts constitute gross
negligence or unless he will receive or expects to receive remuneration. Nothing con-
tained in this subsection shall alter existing law with respect to tort liability of a
practitioner of the healing arts for acts committed in the ordinary course of his
practice.
c) A person who willfully violates subsection (a) of this section shall be fined not
more than $100.00.
German Criminal Code Art. 330c:
Anyone who does not render aid in an accident or common danger or in an emer-
gency situation, although aid is needed and under the circumstances can be expected
of him, especially if he would not subject himself thereby to any considerable danger,
or if he would not thereby violate other important duties, shall be punishable by
deprivation of liberty for a term not to exceed one year or a fine.
Library of Congress Law Library, Good Samaritan Laws in France, the Federal Republic of
Germany, and USSR 11 (1976) (footnotes omitted).
'7 Some legal systems specifically offer a potential rescuer the choice of informing the
proper authority or undertaking a rescue. See Rudzinski, supra note 7, at 108 (reviewing

HeinOnline -- 72 Va. L. Rev. 883 1986

884 Virginia Law Review [Vol. 72:879

that is imposed can vary over a familiar range; nonrescuers can be

fined or incarcerated, held civilly liable for the injury that befalls
the unassisted victim, or both."'

A. Landes and Posner's Arguments

Any discussion of the likely behavioral effects of the carrots and

sticks that might be offered to rescuers best begins with three
points made by Professors Landes and Posner in their well-known
work on the law and economics of rescue. 19 They argue that: (1) an
excessively large reward will sometimes lead to inefficient conse-
quences because victims or owners will take excessive precautions
to avoid paying such rewards; 20 (2) a rescuer may be motivated by
altruism or by the possibility of being regarded as a hero, and that
legal inducements can impede this motivation; 2' and (3) excessive
penalties can decrease the number of rescues because potential res-
cuers may avoid potential "rescue spots," such as beaches, in order
to avoid the reach of these penalties.22 The discussion that follows
emphasizes the importance of the first of these arguments, dis-
counts the second, and then explores in greater depth the likely
effects of various legal rules governing rescue.
The first of Landes and Posner's points is not controversial and
is easy to illustrate. If A can hire a watchman at a cost of $10,000
per year but a passerby (or even a professional rescuer) can save
A's $100,000 property from fire at an average cost of $500 per
event-and such an event is likely to occur once a year-then A
will inefficiently (at excessive expense) hire the guard if he knows
the law will require him to pay strangers who come to the rescue a
commission of, say, one-third or one-half of the value of the prop-
erty they save. Landes and Posner point out that under admiralty

Italian, Turkish, Rumanian, and Russian law). But the point in the text is simply that it is
difficult to imagine (and I have found no) prosecutions of persons who called for help but
did not take other actions themselves.
" See id. at 110-14; infra notes 114-17 and accompanying text; infra notes 121-26 and
accompanying text.
" Landes & Posner, Salvors, Finders, Good Samaritans, and Other Rescuers: An Eco-
nomic Study of Law and Altruism, 7 J. Legal Stud. 83, 91-92 (1978).
20 Id. at 91-93.
21 Id. at 94, 124.
Id. at 119-24.

HeinOnline -- 72 Va. L. Rev. 884 1986

1986] Rescue

law salvors at sea are indeed not rewarded so greatly as to en-

courage excessive precautions by shipowners; instead, they are
probably rewarded at a rate that gives them only a modest return
on their investment. 23 Admiralty law could provide for simple re-
imbursement of all salvage efforts, instead of granting expenses
plus a premium to successful rescuers, but this approach would re-
quire factfinders to separate out those who truly tried to rescue the
distressed at sea from those who simply announce that they ex-
pended effort while the rescue was accomplished by another party
or not at all.24 In short, admiralty law seems to contain just the
sort of incentives for "rescue," or voluntary action, that one would
predict on the basis of Landes and Posner's insight that excessive
rewards will lead to excessive precautions.
It is more difficult to evaluate the argument that legal incentives
may interfere with the utility of altruism as an incentive. In some
sense, the same can be said of most tort and criminal law rules;
their enactment may counterproductively lead to misbehavior on
the part of persons who respond better to untainted, extralegal
stimulants.2 5 It is, of course, difficult to measure or intuit the
strength of altruism as an incentive in various contexts. I think it
safe to intuit as little as possible and to imagine that legal sanc-
tions might interfere with whatever it is that already motivates
some rescuers. Large penalties for nonrescue, while stimulating
some rescue efforts, might decrease the total number of rescue at-
tempts if rescuers motivated by altruism realize that they will

13 Id. at 100-06.
" Id. at 104.
25 It is arguable that Landes and Posner's altruism argument is too broad for its own good
because it comes close to suggesting that all behavior might better be channelled by inter-
personal appeals than by legal sanctions. But most citizens would hardly think that horrible
acts should generally be decriminalized (in order to encourage potential criminals to behave
well) or that those who fail to pay taxes should be encouraged to do otherwise by extralegal
inducements alone. On the other hand, an important difference between rescue and tort or
criminal cases is that in tort and criminal settings it is hard for a well-behaved actor to feel
like a hero, for so many citizens have refrained from misdeeds. Potential rescuers may feel
more unique and, therefore, better about doing the good thing. This notion, that altruism
may be stifled by legal regulation, is not quite the same as the idea of "overjustification" in
the psychology literature. See Lepper, Green & Nisbett, Undermining Children's Intrinsic
Interests with Extrinsic Rewards: A Test of the Overjustification Hypothesis, 28 J. Pers.
Soc. Psychology 129 (1973) (children not initially rewarded for an activity more likely to
engage later in same activity). There the focus is on a given person's repeating certain be-
havior, whereas in rescue it is unlikely that one individual will be a repeat player.

HeinOnline -- 72 Va. L. Rev. 885 1986

 Virginia Law Review [Vol. 72:879

sometimes be perceived as having acted only to avoid legal sanc-

tions. On the other hand, rewards should not generate such an ef-
fect. A rescuer can always decline a reward and enjoy pure hero
status.
Finally, Landes and Posner's second and third arguments can
usefully be combined. In considering packages of carrots and sticks
that might be aimed at potential rescuers it seems sensible to re-
ject a "large sticks and no carrots" package. Not only would a large
penalty interfere with the incentive to be altruistic (or to be re-
garded as altruistic) but also, when unaccompanied by a large car-
rot, it might discourage potential rescuers from frequenting poten-
tial rescue spots if such spots can indeed be identified in advance.
With these points in mind, it is now possible to generalize about
those legal packages of rewards and penalties that might sensibly
promote desirable rescues and then to explore further the notion
that penalties may discourage potential rescuers.

B. Rewards
Large carrots seem unwise for a number of reasons. First, there
is the potential inefficiency, just noted, of swimmers or other po-
tential rescuees' taking excessive precautions to avoid liability.
Carrots that are not financed by rescuees will, of course, not lead
to precautions on their part. Moreover, private carrots may need to
be somewhat more generous than publicly financed carrots because
potential rescuers may fear that rescuees will be unable to pay
whatever reward is required by law. This first argument against
large carrots is thus limited to those that are privately financed.
Second, there is a moral hazard that potential rescuers will cre-
ate the demand for their own services. This moral hazard is surely
more serious in some settings than in others. It is, for example,
easier to imagine that carrot-seekers might put holes in strangers'
boats and then later rush out to the rescue and reward than it is to
see just how such a wrongdoer might encourage swimming acci-
dents. Even when the moral hazard seems quite plausible, it might
be avoided if the carrots that are offered to rescuers are not terri-
bly large or, more interestingly, either not in a generally useful cur-
rency or accompanied by sticks.26 Thus, Good Samaritan statutes

26 These sticks must be aimed at the moral hazard rather than at the failure to rescue.
There must, for example, be a real deterrent to putting holes in strangers' boats.

HeinOnline -- 72 Va. L. Rev. 886 1986

1986] Rescue

do not generate moral hazard problems because they concern car-

rots that are in a sufficiently peculiar currency (enhanced reputa-
tion) to ensure that mischievous intermeddlers cannot profit mon-
etarily from a planned emergency." ' The moral hazard problem is
also more serious if the carrots are publicly financed, for parties
may be tempted to stage emergencies and rescues at the expense of
the fisc.
There is yet another reason why it might be sensible for the law
not to provide large rewards to rescuers. In developing this argu-
ment it will be useful to define a subset of "customized carrots":
these are rewards, often tailor made, offered when specific rescues
are needed. Thus, while successful salvors at sea normally earn un-
customized and longstanding carrots as provided by admiralty
law,28 one whoh returns a lost bracelet to its owner may be entitled
to a reward if the owner has advertised and thereby offered a cus-
tomized carrot. 29 That jewelry owners in general do not promise
rewards to be paid in the event of future losses, and do not press
for legislation that promises such rewards, may reflect the moral
hazard or even the excessive precautions that such promises might
generate. Standing rewards may encourage theft followed by re-
quests for rewards and excessive precautions against loss in the
first place.
But consider, now, those settings in which excessive precautions
are no longer at issue; something has already been lost or is other-
wise at risk and the party who is in need of "rescue," such as the
owner of a lost bracelet, offers a customized carrot. That such a
party may have taken suboptimal, optimal, or even excessive care
is now largely irrelevant. On the other hand, the moral hazard that
may be generated by carrots is quite relevant and suggests the
need for legal intervention. After all, while it is understandable
that one who has property in peril will assess the danger that a
reward will cause more of his property to be perilized as relatively
small, this party is unlikely to consider the hazard that other citi-
zens' property is put at risk. Inasmuch as a reward might en-

27 Additionally, these statutes are often limited to professional health care providers who
have no great incentive to cause roadside emergencies in order to earn fees. See infra text
following note 141.
2s See infra notes 104 and accompanying text; infra note 112 and accompanying text.
21 See infra note 191 and accompanying text.

HeinOnline -- 72 Va. L. Rev. 887 1986

 Virginia Law Review [Vol. 72:879

courage potential "rescuers" to stimulate demand for their own

services by other private parties who might be expected to offer
and finance large rewards, the law might forbid any party in need
of help to offer a carrot. It may, indeed, come to pass one day that
laws are effected that prohibit victims from meeting the demands
of kidnappers, blackmailers, and extortionists. On the other hand,
some combination of empathy for the immediate victims of such
demands, concern that kidnappers (for instance) are actually cap-
tured precisely because we allow ransoms to be paid (and fol-
lowed), and suspicion that such prohibitions would be impossible
to enforce may overcome any hope that the banning of carrots will
lead to a decrease in the crimes that create much of the demand
for rescues.
I do not mean to predict the evolution of the law in this regard.
The utility of customized carrots is, after all, limited to situations
in which there is time and manner for those in need of assistance
to communicate with potential rescuers.30 My point is simply that
the possibility of customized carrots is yet another reason why one
should not expect to see longstanding offers of large (uncus-
tomized) carrots. So long as citizens are permitted and the govern-
ment itself is able to offer customized carrots, it is understandable
that the law is less inclined to put in place standing offers of car-
rots to rescuers. Such carrots may generate excessive precautions
where voluntary customized carrots will not. Moreover, customized
carrots may create less of a moral hazard problem than carrots
that are announced in advance of any specific emergency. To be
sure, because potential rescuers may not always learn about the
availability of customized carrots, such carrots may not always mo-
tivate rescue efforts as forcefully as would carrots that are more
generally announced and available. But occasional customized car-
rots are surely better than no carrots at all.
In short, large carrots may best be reserved for customized offer-
ings in settings in which the probability of successful rescue is low,
the investment required of rescuers is great, or the temptation to
misbehave (rather than return lost valuables, for example) may be

3o On the other hand, it may be that when time is short incentives are generally not terri-

bly useful in promoting desirable rescues. Potential rescuers may best respond to incentives
when there is time to consider the offer, the potential risks, and the likelihood of success.
When such time is available, customized carrots can usually be offered.

HeinOnline -- 72 Va. L. Rev. 888 1986

1986] Rescue

serious. Carrots might be forbidden when they threaten to promote

an undesirable activity, such as kidnapping or extortion, especially
if the enforceability of such a prohibition could somehow be en-
sured. But whether society has chosen to allow victims to offer pri-
vate customized carrots because the plight of those in need of res-
cue is especially compelling, because any prohibition would be too
difficult to enforce, or even because private carrots often lead to
the capture of criminals, this choice militates against the regular
provision of large carrots to all rescuers.

C. Penalties
The imposition of large penalties for the failure to rescue may
also be unwise. Landes and Posner point out that rescuers may be
motivated by glory (or by nothing at all) and that sanctions for
nonfeasance may actually destroy this useful incentive. We do, af-
ter all, rarely regard an act as heroic when it was encouraged by
the threat of punishment for inaction. Unfortunately, it is difficult
to know when, if ever, legal sanctions are for this reason counter-
productive. It is surely unlikely that fewer car accidents or
murders would take place if we would withdraw the legal rules gov-
erning such events and rely on altruism to control behavior. On the
other hand, it is conceivable that those persons who are most likely
to undertake rescues respond differently to similar stimuli than
those who are most likely to commit murders and torts, so that it
is only the potential rescuers who behave worse when facing large
penalties.
A more compelling reason to think that penalties for nonrescue
may be counterproductive derives from the fact that criminal or
civil liability for nonrescue can hardly extend to someone who was
far from the scene of an emergency and in no position at all to act.
The threat of a big stick or any stick at all may, as Landes and
Posner argue, discourage potential rescuers from nearing potential
rescue spots.3 1 This second argument posits an "activity-level ef-
fect"; potential rescuers must care about the low probability that
an emergency will occur, imagine a significant probability that
factfinders will err after the fact and find that the defendant
should have known that a rescue was necessary and easy, and cor-

" See Landes & Posner, supra note 19, at 119-24.

HeinOnline -- 72 Va. L. Rev. 889 1986

890 Virginia Law Review [Vol. 72:879

rectly believe that rescues are more likely to be needed in some

locations than others.
It may be useful to see that the argument that penalties for
nonrescue are unwise because they threaten altruistic rescues pre-
supposes a substantial activity-level effect. Consider, for example,
a member of the Girl Scouts who is altruistically motivated to sell
cookies every year for the greater good of Scouting. If a penalty
such as suspension from the troop is imposed for subpar perform-
ance, various effects can be imagined. The scout may certainly ig-
nore the sanction and continue to be motivated by altruistic im-
pulses or may take note of the sanction and increase her efforts
because of a fear that she will be judged a shirker. She may, alter-
natively, be put off by the sanction and resign from the troop, thus
demonstrating the existence of an activity-level effect. But apart
from such an activity-level effect it is most unlikely that cookie
sales or other efforts will decrease. It is difficult to imagine a scout
who will continue in the troop but decrease her sales efforts, unless
there are steps to be taken that will fool the monitor but really
detract from sales.3 2 Indeed, the only appealing example of identi-
fiable rescue spots and, therefore, of a counterproductive activity-
level effect concerns beaches, but one can barely imagine evening
strolls on the beach decreasing if potential strollers feared that in
the event of a drowning they would be identified as having passed
within earshot of the victim. It is difficult to imagine potential res-
cuers identifying and avoiding other such prime rescue spots and
more difficult still to imagine their continuing to frequent such
spots but somehow putting less effort into rescue missions that
arise because of the imposition of penalties for nonrescue. In short,
both arguments concerning the counterproductive aspects of sticks
for ionrescue build on the questionable proposition that there can
be a substantial activity-level effect, or that rescue spots can be
33
identified and will be avoided.

32 But in the case of professional health care providers, defensive medicine does come to
mind. A physician who would on her own have undertaken rescue efforts but resents the
imposition of a duty to rescue may, one can imagine, do a poor job of the required rescue.
Even this possibility seems safe to ignore for the purposes of this article.
3 My intuition is that the problem of legal sanctions' leading to a decrease in rescue
efforts is an insignificant one. It is difficult, however, to find everyday examples that sup-
port or cast doubt on this intuition. I might argue, for instance, that if the activity-level
effect on rescuers were serious, then witnesses to both' criminal and civil matters who volun-

HeinOnline -- 72 Va. L. Rev. 890 1986

1986] Rescue
Finally, note that any activity-level effect that is generated by
penalties can almost surely be offset by coexisting rewards. Scouts
can be encouraged to remain in or even join the troop by carrots
for high sales or sales efforts. And it is most unlikely that the pres-
ence of these carrots will so destroy the atmosphere of altruism as
to lead to a decrease in the desired efforts, for nothing stops the
altruistic scout or beach stroller from declining or donating her re-
ward and feeling doubly altruistic.

D. The Balance Between Carrots and Sticks

American law does not penalize nonrescuers s4 and seems to pro-
vide virtually no carrots for rescuers.3 5 In contrast, many civil law
systems provide substantial sticks, in the form of criminal (and
sometimes accompanying civil) liability, to discourage nonrescue.
These systems, however, tend also to impute a "mandate", or
agency relationship, between rescuer and rescuee and to require
the latter to reimburse the former for expenses that the "princi-
pal" would surely have authorized.3 6 There is, in short, some evi-
dence of symmetry, or of balance, between carrots and sticks-the
systems that penalize nonrescue offer at least reimbursement to
rescuers. I explore this balance in some detail in this article. In
Part II the durability of this balance is investigated in a discussion

tarily come forward should be exempt from perjury charges. After all, potential witnesses
might avoid the very low probability of perjury charges arising out of their testimony much
as it is suggested that potential rescuers will avoid the beaches. But this analogy is open to
two objections. First, after-the-fact error (by fact-finders or prosecutors) may be perceived
as far less likely when intentional wrongs (like perjury) are involved. And, second, it might
be argued that the number of perjury prosecutions (of voluntary witnesses) is so low that
there is virtually no fear at all of this penalty and therefore no activity-level effect to be
analogized to walking on the beach.
A counterproductive activity-level effect may be a bit easier to imagine if the behavior of
risk-averse potential rescuers is considered. Even the smallest probability of the smallest
stick may discourage a very risk-averse person from abandoning the security of his own
home. Inasmuch as my own experience with people contradicts such behavioral assump-
tions, my inclination in this and in other inquiries is simply to ignore both the conveniences
and complications offered by the assumption of substantial risk aversion. A reader with
different inclinations must adjust various aspects of the analysis that follows and emerge
with a different "balance" between sticks and carrots than that which is developed in this
article.
11 See, e.g., Yania v. Bigan, 397 Pa. 316, 155 A.2d 343 (1959) (no duty to rescue drowning
companion).
11 See Dawson, supra note 5, at 63.
36 See Dawson, supra note 5, at 63; Rudzinski, supra note 7, at 115-16.

HeinOnline -- 72 Va. L. Rev. 891 1986

892 Virginia Law Review [Vol. 72:879

of the carrots and sticks that are provided in a variety of legal sys-
tems. But first, the discussion in this section explores at a theoreti-
cal level the mix of carrots and sticks. I suggest that through a
process of elimination one is inclined to expect at least a rough
balance in the design of incentives for rescue.
The disadvantages of large carrots alone, especially in the con-
text of a legal system that permits and presupposes customized
carrots, have already been discussed.3 7 I have also argued that
small sticks are unlikely to generate important activity-level ef-
fects."8 Very large sticks, such as long-term incarcerations, are, of
course, very costly to the society that imposes them. In any event,
until there is evidence that some society increases the penalties im-
posed on nonrescuers because of dissatisfaction with the effects of
small penalties, it seems sensible to limit the present discussion to
the availability of small sticks, such as fines or the threat of civil
liability. These penalties need not always be imposed; occasional
judgments against nonrescuers may, like a rule of partial liability,
be large enough to stimulate some rescues and small enough to
avoid substantial activity-level effects. And even if sticks inevita-
bly generate counterproductive activity-level effects, it seems likely
that modest carrots can offset these sticks by raising the activity
level that sticks may depress. To be sure, these carrots may gener-
ate their own problems, but the point is that some balance be-
tween the dangers of carrots and the activity-level effects of sticks
might be sought.
It appears, then, that large carrots alone are unwise and that
sticks alone are inferior to sticks accompanied by activity-inducing
carrots. By "large carrots" I mean only to point out that the disad-
vantages of carrots are almost surely a function of their size; in-
deed, I will now refer to "carrots" alone and mean to imply a pre-
sumption that they not be large. One is left with the following
alternatives: (1) "sticks and carrots"; (2) "no sticks and no car-
rots"; and (3) "carrots and no sticks." The first package is symmet-
rical, or balanced, and is close to what is found in most civil law
countries.3 9 The second package is also balanced and fits the usual
description of American law. One might dispute this description

See supra text following note 27.

'8 See supra text accompanying notes 31-33.
See infra notes 114-30 and accompanying text.

HeinOnline -- 72 Va. L. Rev. 892 1986

1986] Rescue

and argue that because American culture exalts rescuers, and be-
cause state statutes in the United States generally exempt some
rescuers from tort liability, it is the third package that most accu-
rately characterizes American law. But because I mean to focus on
carrots that are larger than those provided through Good Samari-
tan statutes and cultural attitudes toward heroism, I ask the
reader who believes that Good Samaritan statutes contain impor-
tant or substantial carrots to rephrase the third stick-and-carrot
package in a way that permits one to pose what I regard as an
important unanswered question: Why do we appear to find (1)
"sticks and carrots" and (2) "no sticks and virtually no carrots"
when it would seem that (3) "carrots and no sticks" is a package
that is to be preferred to package (2)? Again, the power of altruism
as a motivation to rescue is not something that argues strongly in
favor of package (2) over package (3) because so long as carrots can
be declined they should not defeat this power. Only one unbal-
anced package, (3), that of "carrots and no sticks," remains to be
promoted or explained away; if it generates no counterproductive
behavior and seems superior to package (2), then perhaps it should
be hailed as an appropriate goal of law reform.
One might insist that the carrots contained in package (3), even
though they are not large, would stimulate excessive precautions
by potential rescuees. But it is difficult to imagine the excessive
precautions that might be undertaken by swimmers or homeown-
ers who knew that those who one day rescued them from water or
fire would be entitled to a reward of, say, two hundred dollars. And
if excessive precautions by such potential rescuees were imagina-
ble, then it would even seem worthwhile for the state to absorb the
cost of carrots in package (3) so that potential rescuees would not
so fear the cost of being rescued. 40 Alternatively, it might be ar-
gued that the carrots in package (3) would, as discussed earlier,
generate the moral hazard of intermeddlers' creating the demand
for their own services. But this argument should at most call for a
modification of package (3) to contain "no sticks and skeleton 1

'0 See infra text accompanying notes 131-32. Note that the government could tax the
activity in question, such as boat-owning, in order to finance these carrots.
4' I use the term "skeleton" as opposed to say, "nominal," because the type of carrot to
which I am referring is bare, but not necessarily set at a small amount. A skeleton reward
can be large so long as it is adequate only to cover expenses or the rescuer's inconvenience.

HeinOnline -- 72 Va. L. Rev. 893 1986

894 Virginia Law Review [Vol. 72:879

carrots"-that is, carrots that will make rescue somewhat more at-
tractive but not induce the strategic creation of emergencies. The
legal right to reimbursement for the actual expense of a rescue
would, of course, be just such a skeleton carrot. 42 To be sure, such
compensatory or skeleton carrots might be difficult and relatively
expensive to administer. One might, therefore, expect to see some
mix of package (3) with a balanced package depending on the valu-
ation and administrative problems of particular settings.
One must always wonder whether carrots, skeleton or otherwise,
should and would also be available to unsuccessful rescuers. I
would expect such availability to be rare. In the first place, rescu-
ers' claims would be hard to confirm or disprove. Second, it might
seem wise to avoid the problem of someone beginning a rescue
with little ambition and then turning back but claiming a reward.
Finally, if rewards are privately financed, it might seem cruel to
ask a grieving family or owner of recently destroyed property to
pay for services that proved useless. The rewards received by suc-
cessful rescuers could, of course, be adjusted so that, ex ante, po-
tential rescuers are as motivated as they would be if they were to
receive some carrots for all rescue attempts. In fact, most of the
carrots described later on in this essay are available only to suc-
cessful rescuers.
In sum, there are reasons to expect a rough balance between the
rewards and penalties for rescues. Greater sticks must almost
surely be accompanied by greater carrots, unless there is no activ-
ity-level effect at all, and large carrots may be counterproductive
even when accompanied by sticks. But a slightly unbalanced pack-
age ("no sticks and skeleton carrots") seems, at least theoretically,
to be superior to the balanced package of "no sticks and no car-
rots." That the latter package appears to describe the strategy ex-
pressed in American law is, therefore, somewhat puzzling. The dis-
cussion in Part II will suggest that American law is, in fact, much
less puzzling than it first seems.43

" These actual expenses may even include the value of one's time. The text ignores the

moral hazard of an intermeddler's wanting to be a hero and creating an emergency to stage

a rescue. This moral hazard seems fairly constant in size regardless of the legal rules affect-
ing rescue.
43 See infra text accompanying notes 51-81, 184-95. It is useful to note that this question
of balance between rewards and penalties cannot simply be translated into a matter of
counting and comparing instruments and goals. There are a fair number of available instru-

HeinOnline -- 72 Va. L. Rev. 894 1986

1986] Rescue

II. VARIETY AND UNIFORMITY IN THE TREATMENT OF RESCUERS

I argued in Part I that "sticks and (not too large) carrots" and
"no sticks and skeleton carrots" would be sensible incentive pack-
ages for the purpose of promoting desirable rescues. I now turn to
the incentives actually offered to potential rescuers and develop
two points. First, American law may actually offer "no sticks and
skeleton carrots" and, second, although different legal systems may
provide dissimilar solutions to given problems-such as the promo-
tion of desirable rescues-their treatments may be constrained by
expectations about behavioral responses that may be quite alike in
various societies. Indeed, I argue in this part that the balance be-
tween sticks and carrots, described in Part I as a product of behav-
ioral considerations, is so durable that it is found not only in vari-
ous areas of American law but also in civil law systems and in such
ancient rules as those found in the Code of Hammurabi.
This second point is part of a larger positivist argument that I
have developed elsewhere about "uniformity" and "variety" among
legal systems. My suggestion is that different legal systems can be
expected to display uniformity when a particularly necessary be-
havioral effect can be accomplished only with a certain rule. Vari-
ety, on the other hand, is to be expected either over a range in
which rules do not greatly affect behavior or when reasonable peo-
ple could disagree over which rule best accomplishes a desired ef-
fect. 44 Although this uniformity-variety thesis cannot be demon-
strated in any deductive or inductive way, its plausibility is surely
enhanced by demonstrations of its predictive power in different ar-
eas of the law. The discussion that follows thus not only shows
that the incentive packages described as sensible in Part I are in
fact those that are found in various legal systems but also lends

ments: rewards financed by rescuees, rewards financed by the government (or other third
party), and sticks are three broad categories that have been discussed. But there is only one
central goal-effecting socially desirable rescues-because there is, for example, no need to
have people milling about on the beach if rescues are already taken care of in the calculus.
Adding another goal, like the prevention of hazards (such as the drilling of holes in boats),
hardly increases the number of goals beyond that of the available instruments. Generally
speaking, we might expect a legal system to utilize all the available instruments in order to
accomplish its goal(s) with greatest flexibility. And to some extent, American law, as we will
see, does not ignore any available tool. But administrative difficulties and side effects-such
as the danger that large carrots will in some settings lead to excessive precautions-suggest
that all instruments will not be used in all settings.
4' See Levmore, supra note 8, at 1-6.

HeinOnline -- 72 Va. L. Rev. 895 1986

896 Virginia Law Review [Vol. 72:879

support to the larger argument about the uniformity and variety of

legal rules.

A. Incentives for Rescuers in American Law

1. Life Rescuers
Medals and media attention aside, "no sticks and no carrots" is
said to be the Anglo-American norm for those not contractually
bound to undertake emergency rescue efforts.4 5 There are some
well-known exceptions to this norm. Physicians and hospitals may
be liable for ignoring emergencies and can certainly collect compet-
itive carrots for any rescue efforts that they do undertake. 46 Such
exceptions are not at all surprising for, as discussed in Part I,
"sticks and carrots" is in some settings quite a sensible package. In
particular, it is not likely that serious moral hazard and excessive
precaution problems would accompany the use of carrots for emer-
gency professional medical services. A second well-known excep-
tion to the "no sticks and no carrots" rule is the statutory develop-
ment of small sticks for nonrescue in some jurisdictions. Vermont
and Minnesota have both legislated a "duty to rescue" when such
rescue is riskless and have provided for a fine up to one hundred
dollars for violations of this duty; Massachusetts and Rhode Island
have legislated a duty to report a crime and have provided for sub-

See generally R. Goff & G. Jones, The Law of Restitution 263-67 (2d ed. 1978) (the
common law denies restitution to one who intervenes on behalf of another); Feldbrugge,
supra note 7, at 14 (Anglo-American legal system contains a notable exception to the general
duty to offer aid to another where possible that is found in other legal systems); Gregory,
supra note 6, at 23 (the common law has not enforced the moral duty to rescue); Weinrib,
supra note 2, at 247 (no common law duty to rescue); Note, The Failure to Rescue: A Com-
parative Study, 52 Colum. L. Rev. 631 (1952) (no general duty to rescue in Anglo-American
law).
48 Theoretically, physicians and hospitals can refuse to see patients just as a layperson
can ignore stranded motorists. On the other hand, courts may be quick to decide that physi-
cians and hospitals have undertaken to help a patient-and thus created a duty-and then
wrongfully abandoned the patient. If this abandonment, or negligence, is the cause of fur-
ther injury, the patient will be able to recover. See cases collected in W. Wadlington, J.
Waltz, & R. Dworkin, Cases and Materials on Law and Medicine 470-72 (1980); Note, Must
a Private Hospital Be a Good Samaritan?, 18 U. Fla. L. Rev. 475, 477-78 (1965).
Although there is some disagreement regarding the precise magnitude of recovery allowed
physicians and hospitals who have rescued patients in emergencies, some recovery is clearly
available. See cases collected in J. Dawson & G. Palmer, Cases on Restitution 32-46 (2d ed.
1969).

HeinOnline -- 72 Va. L. Rev. 896 1986

1986] Rescue

stantial penalties for failure to report.47 Inasmuch as the activity-

level effect of penalties is quite small, 48 this development is hardly
surprising. Still, I would expect these states to develop at least
skeleton carrots to accompany these small sticks. Meanwhile, it is
noteworthy that there have been no reported prosecutions under
these statutes.
The most curious thing about American law-even outside of
Vermont, Minnesota, Massachusetts, and Rhode Island-is the ap-
parent absence of even skeleton carrots, for such carrots may pro-
mote desirable rescues at low cost.4 9 This curiosity might be dealt
with in a number of ways. First, it could be argued that the costs
of legal intervention must figure prominently in uniformity-variety
50
predictions; American law may simply opt for little intervention.
Inasmuch as American law encourages physicians to provide emer-
gency medical services and thus intervenes in at least some set-
tings, a more sophisticated form of this argument would be that
the costs of legal intervention may, in the eyes of American
lawmakers, overcome whatever advantages carrots generate. Simi-
larly, it might be argued that the rule does not much matter so
that no level of carrots would be particularly surprising. Large car-
rots may generate large problems but skeleton carrots may offer
such small incentives that variety ought to be expected between

4, Mass. Gen. Laws Ann. ch. 268, § 40 (West Supp. 1985) (duty to report, when no danger
is thereby created, to appropriate official as soon as reasonably practicable when another
person is victim of aggravated rape, rape, murder, manslaughter, or armed robbery; violator
fined not less than $500 nor more than $2500); Minn. Stat. Ann. 88 604.05, 609.02 (West
Supp. 1986); R.I. Gen. Laws §§ 11-37-3.1, 3.3 (Supp. 1985) (duty to report sexual assault (of
another person) punishable by up to one-year imprisonment or fine not to exceed $500); Vt.
Stat. Ann. tit. 12, § 519 (1973). Until there are prosecutions to examine, I think it fair not to
dwell on the penalties for failure to report, for given that they were passed following a
widely publicized public sexual assault, see Prentice, Expanding the Duty to Rescue, 19
Suffolk U.L. Rev. 15, 33 (1985), they may have been meant to be used as tools with which
prosecutors can encourage members of a group to come forward and testify against wrong-
doers or other group members. They may even have been meant as a minimum (and easy to
prove) charge for group crimes where major criminal convictions can be difficult to obtain.
In any event, it is interesting to note that these duty-to-report statutes, in the style of Good
Samaritan statutes and in the direction of symmetry, absolve one who reports from civil
liability.
4 See supra text preceding note 32.
', See supra text accompanying notes 40-42.

80 Other legal systems may simply regard legal intervention as less costly. This is espe-

cially likely if private litigants in these systems have limited procedural options so that they
are unable to impose substantial costs.

HeinOnline -- 72 Va. L. Rev. 897 1986

898 Virginia Law Review [Vol. 72:879

"no carrots" and "skeleton carrots." But one need not choose be-
tween these explanations of American law, for the law and the cu-
riosity just noted are generally misstated. In fact, American law
does often provide at least skeleton carrots. The discussion in this
section considers the ways in which American law rewards not only
physicians but also other persons who voluntarily rescue life and
property. In the sections that follow the analysis is extended to
include other rescuers, including finders of lost property, salvors at
sea, and "disclosers" of valuable information.
An important, if skeleton, carrot in American law consists of a
promise to potential rescuers that they or their survivors will be
able to recover for injuries suffered in a rescue attempt. It is not
particularly interesting to find that such recovery is generally
available from a negligent third party who creates the need for the
rescue so long as the rescuer was not also a cause of the emer-
gency. 51 Courts may struggle a bit with causality doctrines, 52 but it
is rather clear that if without a rescue attempt of A by B, C would
be liable for injuries suffered by A, C should and will be liable for
53
injuries suffered by B while undertaking a reasonable rescue. It
is, however, much more interesting to note that a rescuer stands an
excellent chance of recovering for his injuries when the victim's
negligence created the need for rescue. 4 It is difficult to see why
the presence or absence of negligence by the rescuee himself
should affect the rescuer's chances of recovery. The rescuer acts to
save the rescuee the burden of his own injury-but this is so
whether or not the emergency arose because of the rescuee's negli-
gence. It is thus arguable that recovery by the rescuer in these
cases not only represents an important subset of rescue cases in
which carrots are available but also anticipates or opens up the
possibility of recovery by rescuers in future cases in which there is

51 See, e.g., Roanoke Hsp. Assoc. v. Hayes, 204 Va. 703, 133 S.E.2d 559 (1963); Andrews

v. Appalachian Elec. Power Co., 192 Va. 150, 63 S.E.2d 750 (1951); Southern Ry. Co. v.
Baptist, 114 Va. 723, 77 S.E. 477 (1913). For a case denying recovery where the rescuer was
at least somewhat negligent, see Nelson v. Pendleton, 214 Va. 139, 198 S.E.2d 593 (1973).
52 See Eckert v. Long Island R.R., 43 N.Y. 502 (1871).
53 This may even be so when the rescuer has sought to save property rather than life. See
W. Prosser & W. Keeton, supra note 1, § 44, at 307-09.
See e.g., Lynch v. Fisher, 34 So. 2d 513 (La. App. 1947); Brugh v. Bigelow, 310 Mich.
74, 16 N.W.2d 668 (1944); Talbert v. Talbert, 22 Misc. 2d 782, 199 N.Y.S.2d 212 (N.Y. Sup.
Ct. 1960). But see Saylor v. Parsons, 122 Iowa 679, 98 N.W. 500 (1904) (a lone case denying
recovery when only the victim was negligent).

HeinOnline -- 72 Va. L. Rev. 898 1986

1986] Rescue

no negligence at all. It is possible that judges exploit the presence

of negligence, however unconnected it is to arguments for recovery,
as a means of protection against criticism that there will be no
limit to their intermeddling. Over time, however, it may become
easier to build on the many precedents in favor of recovery and
argue that there is no reason why the absence of negligence should
require a decision against the rescuer. Indeed, the dicta of the
cases involving negligence by the rescuee himself have long con-
tained the seeds of such evolution, for they speak of the rescuee as
having created the need for rescue without insisting that the crea-
tion of the rescue-inviting situation was negligent. 55
Another relevant development in American law may point to the
development of sticks more than it does to that of carrots. Al-
though the general rule of no sticks, or no duty to rescue, is solidly
entrenched in most jurisdictions, many exceptions to this rule have
been discovered in the form of "special relationships" out of which
affirmative duties are said to arise. These exceptions have obvi-
ously not appeared in all jurisdictions but the overall appearance
of these exceptions signals expansion. Although a finding of a spe-
cial relationship normally precedes a finding of tort liability for
nonrescue, it is not unreasonable to expect that carrots may grow
next to these new sticks. After all, just as a physician's ability to
recover for noncontractual emergency services seems to accompany
the physician's duty to assist those in need of his services, perhaps
because sticks alone would create an undesirable activity-level ef-
fect, so too might carrots accompany the sticks that are extended
to "special relationships." The expanding set of exceptional special
relationships to the "no duty" rule includes common carrier-pas-
senger; 56 innkeeper-guest; 57 innkeeper-stranger (a duty to protect a
stranger from injury by a guest);5 8 employer-employee; 59 ship-crew-

5 See, e.g., Brugh v. Bigelow, 310 Mich. 74, 80-81, 16 N.W.2d 668, 671 (1944) (holding
defendant who shouted for help from his overturned car liable for breach of duty not to
make himself the object of rescue); Carney v. Buyea, 271 A.D. 338, 344, 65 N.Y.S.2d 902,
908-09 (1946) (holding that defendant who left her car out of gear on an incline breached
legal duty not to create undue risk to herself).
" See Yu v. New York, New Haven, & Hartford R.R. Co., 145 Conn. 451, 144 A.2d 56
(1958).
51 See West v. Spratling, 204 Ala. 478, 86 So. 32 (1920) (duty of innkeeper to provide and
maintain sufficient fire escapes for benefit of patrons).
" See Connolly v. Nicollet Hotel, 254 Minn. 373, 95 N.W.2d 657 (1959).
"g See Rival v. Atchison, Topeka & Santa Fe Ry., 62 N.M. 159, 306 P.2d 648 (1957).

HeinOnline -- 72 Va. L. Rev. 899 1986

900 Virginia Law Review [Vol. 72:879

man;60 shopkeeper-business visitor; 1 host-social guest; 62 jailer-pris-

oner; 3 school-pupil;6 4 drinking companions;6 5 landlord-trapped
trespasser;66 safety engineer-laborer;67 physician-patient;68 psychol-
ogist-stranger (a duty to protect a stranger from harm at the hands
of the psychologist's patient);69 manufacturer-consumer;7 0 land-
lord-tenant;7 ' parole board-stranger (a duty to protect strangers
from a released prisoner); 2 husband-wife;7 3 parent-child;7 4 and
tavern keeper-patron.7 5 In all these settings it appears increasingly
likely that duties will again be pronounced and that liability for
nonrescue will follow.
The discussion in Part III of this article suggests a positive the-
ory to explain the emergence of these special relationships and du-
ties and also suggests the likely evolution of these duties and of
rescue rules in the future. One feature of this list of special rela-
tionships is, however, of immediate relevance. Apart from the fa-
miliar case of the physician rendering emergency services, four of
the special relationships just cited do not involve preexisting rela-
tionships, although this is a concept that is often said to be related

'0 See Abbott v. United States Lines, Inc., 512 F.2d 118 (4th Cir. 1975).
1 See L.S. Ayres & Co. v. Hicks, 220 Ind. 86, 40 N.E.2d 334 (1942) (while shopkeeper
may not be liable for failure to prevent unforeseeable injuries to invitee, he may be liable for
failure to ameliorate the injuries via last clear chance doctrine).
62 See Tubbs v. Argus, 140 Ind. App. 695, 225 N.E.2d 841 (1967) (driver may not invoke

guest statute as a shield from liability for injuries not resulting from wanton and wilful acts
where car was stopped and driver could have prevented further injuries).
63 See Farmer v. State, 224 Miss. 96, 79 So. 2d 528 (1955).
64 See Pirkle v. Oakdale Union Grammar School Dist., 40 Cal. 2d 207, 253 P.2d 1 (1953).
65See Farwell v. Keaton, 396 Mich. 281, 240 N.W.2d 217 (1976).
66See Pridgen v. Boston Hous. Auth., 364 Mass. 696, 308 N.E.2d 467 (1974).
67 See Caldwell v. Bechtel, Inc., 631 F.2d 989 (D.C. Cir. 1980).
68 See Davis v. Potter, 51 Idaho 81, 2 P.2d 318 (1931).
66 See Tarasoff v. Regents of Univ. of Cal., 17 Cal. 3d 425, 551 P.2d 334, 131 Cal. Rptr. 14

(1976).
70 See MacPherson v. Buick Motor Co., 217 N.Y. 382. 111 N.E. 1050 (1916).
7! See Kline v. 1500 Mass. Ave. Apartment Corp., 439 F.2d 477 (D.C. Cir. 1970).

72 See Grimm v. Arizona Bd. of Pardons and Paroles, 115 Ariz. 260, 564 P.2d 1227 (1977).
7 See Territory v. Manton, 8 Mont. 95, 19 P. 387 (1888).
7 See Commonwealth v. Breth, 44 Pa. C. 56 (Clearfield County Ct. 1915).

75 See McFarlin v. Hall, 127 Ariz. 220, 619 P.2d 729 (1980) (tavern owner may be held
liable for injuries to plaintiff-patron from acts of other patrons where tavern owner was
negligent in not warning plaintiff of other patron's condition); see also cases cited in W.
Prosser & W. Keeton, supra note 1, § 56, at 376-77 nn.33-40 (discussing instances in which
courts have justified the creation of an affirmitive duty upon the finding of "some special
relation between the parties").

HeinOnline -- 72 Va. L. Rev. 900 1986

1986] Rescue

to liability for nonrescue.7 ' The liability of an innkeeper for not

preventing injuries to strangers77 does not, for example, seem far
from the question of liability for not rescuing one's neighbor. And,
once again, carrots for rescuing one's neighbor seem related to
sticks for not rescuing one's neighbor.
Finally, there is yet another subtle carrot (of unclear size) held
out to potential rescuers by American law. While it is true that if A
is injured while rescuing a nonnegligent B, A is unlikly to be able
to recover, what if B or a relative of B in a tender moment
promises to pay A? Such promises are much more likely to be held
enforceable than other promises for which there is no present or
future consideration. 8 I do not mean to insist that such promises
will always be enforced. Indeed, good arguments can be advanced
that A will only recover when it is B's estate that is sued after B
has in fact made some payments as promised 9-- so that enforce-
ability may be an outgrowth of the law of trusts and estates more
than contracts-or when it is plausible that A could have recov-

7, See supra, notes 48, 57, 59, 62 and accompanying text. A landlord and a trespasser on
his property also can be said to involve no preexisting relationship. See supra note 66.
77 These are not cases in which the psychologist, for example, or his agent injure a patient
who might be said to enjoy a preexisting relationship with the psychologist. Instead, the
case concerns liability to a third person who had an unfortunate relationship with the pa-
tient. Similarly, the parole board, innkeeper, safety engineer, and landlord in the cases
noted above were found to have duties not (merely) to the parties with whom they dealt,
but to strangers.
7s See E. Farnsworth, Contracts § 2.8, at 55-59 (1982). It is not clear that promises would
be enforced if the earlier rescue had been of the promisor's property rather than life or if
the rescuer's injury were of loss of body or life and not of property alone. Indeed, I have
chosen to concentrate on life rescue in this section not only because restitution for property
rescue is something I have considered elsewhere, see Levmore, Explaining Restitution, 71
Va. L. Rev. 65 (1985), but also because the special relationships and other developments
discussed in this section are probably not terribly indicative of the likely treatment of prop-
erty cases. A psychologist may be held liable for injury suffered by a stranger at the hands
of a patient who might have been known to be dangerous, but one might hesitate to predict
that the same legal result would be forthcoming in a case in which the patient carried out a
threat to damage the stranger's property. I do think that the evolution of property rescue is
likely to track that of life rescue. But because the developments in present law that most
influence my argument may be limited to cases involving life rather than property, the dis-
cussion in this section is limited to life rescue.
7' The best known of these cases involve claims against the estates of promisors who,
while alive, gave no indication that they wished to revoke their promises. See Webb v. Mc-
Gowin, 27 Ala. App. 82, 168 So. 196 (1935), cert. denied, 232 Ala. 374, 168 So. 199 (1936);
Estate of Hatten, 233 Wis. 199, 288 N.W. 278 (1940). The principle may thus have more to
do with carrying out decedents' wishes than with subtly supplying carrots to rescuers. Both
cases are discussed in E. Farnsworth, supra note 78, § 2.8, at 56-57.

HeinOnline -- 72 Va. L. Rev. 901 1986

 Virginia Law Review [Vol. 72:879

ered in tort from a negligent B so that the promise is in return for

A's forbearance from litigation." But it is noteworthy that the suc-
cess of promisees such as A is an exception to, rather than a well-
explained part of, contract law. Courts and commentators may in-
sist that B's moral obligation makes a difference, but inasmuch as
a promisor can always be said to be morally obliged to keep his
word and yet, as we know, gratutious promises are normally not
enforceable, it seems most reasonable to look elsewhere for an ex-
planation of A's success. It is, in short, possible that any special
success of promisees such as A is best understood as part of the
law of rescue. Although in some formal sense a potential rescuer
does not face sticks and carrots, he is subject to penalties (and con-
ceivably entitled to carrots) when a "special relationship" is found,
he is able to collect when some negligent party created the need for
rescue, and he may be able to collect on any promise that is made
by one who benefits from his rescue efforts even though the prom-
isor later seeks to repudiate his grant. Unfortunately, I have no
way of knowing how often rescuees voluntarily reward their rescu-
ers and how often rescuers sue on promises once made by their
rescuees.
These developments and the preceding arguments lead one to
believe that American law already grants some rewards to rescuers
and may not be far from granting many more. I suspect that a
rescuer who seeks to recover out-of-pocket expenses or other costs
that are easily valued is more likely than other rescuers to secure
recovery.8 1 But it is difficult to assess the precise degree to which
carrots are already available. If, for example, judges feel bound to
the doctrinal steppingstone of negligence as a requirement for re-
covery by an injured rescuer, then it may take some time before
sufficient precedents are created by more aggressive judges who
overlook this apparent doctrinal requirement or create new "spe-
cial relationships" that trigger both sticks and carrots.

11 See Palmer v. Dehn, 29 Tenn. App. 597, 198 S.W.2d 827 (1946) (forebearance of suit
sufficient consideration to support promise of compensation).
81 It is arguable that professional rescuers, such as physicians and hospitals, have an eas-
ier time gaining carrots for their work because their services are often purchased and, there-
fore, relatively easy to value. See Landes & Posner, supra note 19, at 109-10. This argument
suggests that nonprofessional rescuers might be most successful when bringing claims for
(easily valued) out-of-pocket expenses.

HeinOnline -- 72 Va. L. Rev. 902 1986

19861 Rescue

2. Finders and Disclosers

One who stumbles across lost property or useful information and

one who invests in discovering such property or information are
similar in important ways to more recognizable potential rescuers
who come across victims in distress or more purposefully set out on
rescue missions. There are, of course, many other actors that are
encouraged by sticks and carrots-soldiers and taxpayers come
quickly to mind-but because the incentives offered these actors
do not normally generate moral hazard problems, activity-level ef-
fects, and excessive precaution-taking, I think it is fair to say that
it is finders and disclosers who are sufficiently like more familiar
rescuers to merit parallel discussion.
Under American law a finder is perfectly free to pass by and
ignore the property he sights;82 in this sense there are no sticks. To
be sure, a finder must not simply appropriate for himself property
that he discovers, much as one must not swim out to a drowning
stranger who is calling for help and strike him on the head. There
is, however, no legal obligation to secure and return lost propertys3
just as there is no obligation under American law to go out and
save a drowning swimmer.8 ' If a finder does affirmatively take
charge of lost property and return it or seek out its owner, then a
variety of carrots are available. First, any reasonable expenses in-
curred in caring for the lost property will usually be recoverable
from the owner, or implied bailor.8 5 The finder, or involuntary
bailee, may be held liable for losses his negligence causes but such
negligence is not quickly imputed.8 6 He is given loose rein to sell
the property and hold the proceeds for the owner's arrival when

" Not only are there no statutory provisions or judicial decisions requiring someone to
"find" lost property, R. Brown, The Law of Personal Property § 3.5, at 30 (3d ed. 1975), but
also I know of no case in which a plaintiff even sought to have such a duty imposed. I do not
think that those who have lost property have been unwise in failing to bring such claims.
83 Id.
W. Prosser & W. Keeton, supra note 1, § 56, at 378.
,5 See Chase v. Corcoran, 106 Mass. 286 (1871); R. Brown, supra note 82, § 3.5, at 30-31.
But see Watts v. Ward, 1 Or. 86 (1854) (finder unable to recover expense of returning lost
horses to owner). Some states now have statutes requiring the compensation of finders for
their expenses. See, e.g., Or. Rev. Stat. § 98.025 (1984) ("all the costs and charges incurred
in the finding, giving of notice, care and custody"); Wis. Stat. Ann. § 170.09 (West 1974)
("all the costs and charges . including a reasonable compensation to the finder for his
trouble").
"6 See R. Brown, supra note 82, § 12.2, at 322.

HeinOnline -- 72 Va. L. Rev. 903 1986

 Virginia Law Review [Vol. 72:879

the ongoing expenses of care seem likely to overwhelm the value of

the property,8 7 and, unlike other bailees, he will not be held liable
for nonnegligent misdelivery.8 s Second, and more important, with
the exception of some rare cases in which there is a terrific moral
hazard that the "finder" has simply taken possession of another's
goods, 9 finders are allowed to keep property that is unclaimed af-
ter the passage of some specified amount of time."0 The finder thus
gains title superior to all the world but for the original owner-and
eventually even the original owner loses out because of a statute of
limitations. The finder's most important carrot is, therefore, the
expected value of the owner's inability or unwillingness to claim
the lost property despite any rules that require the finder to adver-
tise the find,9 ' discounted by the possibility that a claim by the
owner of the locus in quo will be treated sympathetically 92 and by
the possibility that the property will be regarded as legally mis-
placed rather than lost.9 3 Third, the finder may hope that the
property will sometimes be treated by the law as abandoned rather
than lost or misplaced and, again, that no owner of the locus in
quo will successfully claim the property. In these circumstances
the finder will gain title to the property regardless of whether its
former owner materializes.9 4 Fourth, a finder may hope that the

87 Thus, a landlord is often allowed to sell goods left behind by a tenant. Id. § 12.1, at
320-21 & n.4.
" See id. § 21.3, at 327-28.
89 In Iowa, found vessels, rafts, logs, or lumber that are unclaimed and exceed twenty
dollars in value become the property of the county (after expenses, such as that of publica-
tion are paid). Iowa Code Ann. §§ 644.1, 644.4 (West Supp. 1986). Inasmuch as other un-
claimed lost property becomes the property of the finder, it seems that the point is to de-
crease the moral hazard of "finding" lost property that is afloat and really under the control
(but not surveillance) of its owner. The same is true of "treasure troves" (caches of gold and
silver of unknown origin) in English law, where a similar moral hazard may exist. R. Brown,
supra note 82, § 3.3. Wisconsin does, however, appear to require all finders of (eventually)
unclaimed property to share their luck with the government. Wis. Stat. Ann. § 170.10 (West
1974). Such a rule might be regarded as an income tax on receipts that are likely to go
unreported (if only because finders might not realize their reporting obligations) in both the
state (if any) and federal tax systems.
9o See, e.g., Mont. Code Ann. § 70-5-206 (1985).
9' Montana has particularly elaborate requirements. Mont. Code Ann. §§ 70-5-203 to -206
(1985) (requiring appearance of finder before justice of the peace, swearing of affidavit, ap-
praisal, recordation, notice, and advertisement).
92 See R. Brown, supra note 82, § 3.2, at 26.

93 Id. § 3.4, at 29-30.

9, Id. § 3.1, at 25 n.1.

HeinOnline -- 72 Va. L. Rev. 904 1986

1986] Rescue 905

owner will provide a reward for the return of his property, or the
finder may know of such a reward and simply expect (quite cor-
rectly) that such a promise will be enforced. 5 Fifth, and finally, a
finder may be able to extract a reward from the owner who re-
claims his property because such a reward is provided by statute.
Iowa, in particular, provides for a ten percent commission in addi-
tion to the expenses incurred for one who finds and takes up boats,
money, bank notes, lost goods, and "other things." 96
That these carrots are, ex ante at least, relatively substantial and
yet not balanced by accompanying sticks is quite explicable. Here,
it is an analysis of the behavioral effects of carrots and sticks more
than an application of the uniformity-variety thesis that allows one
to predict the range of legal rules. Two minor problems might be
generated by these carrots. It is possible that owners will take ex-
cessive precautions against losing property because their liability
to finders may seem great in light of the likelihood that somebody
will return lost property even if no reward is available. It is also
possible that potential finders will actually cause losses and then
hide the items that they find from their true owners. The notion
that "misplaced," as opposed to lost, property remains in the cus-
tody of the owner of the locus in quo and not of the finder (who is
also then unable to claim a reward offered for finding)9 7 appears to
respond to this concern. The behavior of someone who "finds"
property that has been intentionally placed down by another is
simply suspect, for it is too easy to "admire" a coat or jewel set
down by its owner and then, if opportunity strikes and the law
rewards, claim to have found it. The owner who traces his steps
will also have an easier time locating his property if it remains
where the owner has actually been.
But a more important moral hazard dominates any consideration
of carrots: Finders may simply appropriate their discoveries for
themselves and not publicize their finds. Lawmakers must thus
balance the degree to which moral behavior, altruism, probabalistic
rewards, and the fear of sticks for affirmative misdeeds such as

95 See Wentworth v. Day, 44 Mass. 352 (1841); R. Brown, supra note 82, § 3.5, at 31.
The
finder may, however, need to have learned of the reward before completing his performance.
See E. Farnsworth, supra note 78, § 2.10, at 65.
" Iowa Code Ann. § 644.13 (West 1979).
97 See R. Brown, supra note 82, § 3.4.

HeinOnline -- 72 Va. L. Rev. 905 1986

906 Virginia Law Review [Vol. 72:879

hiding the find or causing the loss itself will offset the finder's self-
interest in simply taking or ignoring the property. Put differently,
since affirmatively misbehaving finders run a low chance of detec-
tion and prosecution, their behavior may be thought best modified
by rewards and penalties rather than by penalties alone. This pol-
icy alternative is surely made more feasible by the fact that the
law's target is identifiable-it would be easier to give carrots to
finders than to give carrots to all those who do not shoplift, for
example-and more attractive by the fact that the rewards to find-
ers need not be financed by the state itself. The state could claim
for itself property that no owner claims and in this sense the
state-by allowing finders to keep unclaimed property-does fi-
nance the most important carrots, but it is easy to imagine that
lawmakers regard as costless the rule that finders gain title when
original owners fail to appear. In short, "no sticks and no carrots"
would not be a terribly surprising package because lawmakers
might think that the moral hazard that potential "finders" will
cause losses or just not announce their finds combined with the
problem of excessive precautions by owners overcomes the advan-
tages of carrots both in encouraging finds and in discouraging post-
find misbehavior. Yet "carrots and no sticks" is surely no less sur-
prising; reasonable people could disagree over the magnitudes of
the various behavioral effects, and experience or experimental data
seem unlikely to clarify the choices."" In sum, an analysis of the
law pertaining to finders does not demonstrate a perfectly durable
balance between rewards and penalties. Instead, it demonstrates
the usefulness of beginning the inquiry with such a balance in
mind and then understanding the deviations from it as relating to
particular circumstances. The "carrots and no sticks" pattern
found in the law of finders is surely responsive to the moral hazard
of finders' taking and not reporting their finds.
Disclosers are best compared with finders in a way that antici-
pates the analysis in Part III. It is difficult, as a law enforcemant
matter, to identify potential finders who simply ignore lost prop-
erty (as opposed to misbehaving finders who convert the property
for their own use) but easy to single out persons who do not dis-

98 It is difficult to collect data on the ignoring of obviously lost property and on the pock-
eting of such property. The effect of a reward may be to encourage finders to report that
which they have gathered in.

HeinOnline -- 72 Va. L. Rev. 906 1986

1986] Rescue

close material information when bargaining, if only because those

who bargain or transact are a subset of the larger group of those
who have information. Hundreds of people in a neighborhood may
know that termites have infested the area, but if the law wishes for
a buyer to bargain for a home with this information in mind, it is
much easier to impose a duty on the seller rather than on all
neighbors with information.
The familiar problem with sticks in this context is that invest-
ment in discovering information may be discouraged if the investor
is not permitted to profit from his information. Inasmuch as I have
elsewhere expressed my own positive view that the law generally
protects those investors by not requiring disclosure except in those
circumstances, such as termite infestations and situations in which
the uninformed party reveals idiosyncratic concerns by asking spe-
cific questions, in which non-disclosure allows net losses to occur,9 9
I will not review it or the intricacies of the substantive law here. I
think a fair summary of the law is that "rare sticks and negotiated
carrots" are available to potential disclosers of information. Sticks
are imposed only when the cost of nondisclosure, such as continued
termite damage, exceeds the social loss that will result from deny-
ing some "investors" the full value of their information. Homeown-
ers, for example, have a fair incentive to inspect regularly for ter-
mites, especially if buyers will as a matter of course ask specifically
to inspect. Similarly, managers of business enterprises have good
reason to explore for corporate opportunities, even if they must
then abstain from or reveal information when trading shares of the
enterprise. On the other hand, if sticks were instead imposed for
all nondisclosure by corporate insiders, so that all who traded igno-
rantly on the stock exchange could recover if material information
had been withheld (by insiders who did not themselves trade),
then surely there would be much less exploration in the first
place. 0 0
It may seem at first that these "rare sticks" could only be ac-
companied by "no carrots." In the corporate sphere, a non-insider

"9 Levmore, Securities and Secrets: Insider Trading and the Law of Contracts, 68 Va. L.

Rev. 117, 133-42 (1982).

100 This stick for nondisclosure is especially interesting because without it one might be
tempted to suggest that sticks are used only when life (as opposed to property) rescue is at
stake.

HeinOnline -- 72 Va. L. Rev. 907 1986

908 Virginia Law Review [Vol. 72:879

who develops information need not disclose such information when

he trades with a stranger (no sticks) and it is hard to see how the
law could possibly reward him for generously revealing informa-
tion. Valuation, moral hazard, and precaution problems abound.
But in some sense the discloser can extract carrots by contracting
for them with the party with whom he trades. If he has developed
a reputation as someone who often possesses useful information,
then a stranger may pay him for his whispers. And, in general, it
may be possible to give unrevealing but convincing clues so that a
stranger will believe that valuable information is available for
purchase. In many settings it will be possible to package informa-
tion with products or services so that carrots can be bargained for
in return for information. An automobile mechanic need not tell
me that he can hear that my carburetor needs adjusting (sticks
would discourage investment in his training), but if I regularly buy
gasoline at his service station he may make a practice of so inform-
ing me. These "negotiated carrots" encourage the production of
useful information and do not normally create moral hazard
problems. When such problems may be generated, as when corpo-
rate insiders may intentionally cause financial losses in order to
profit personally from inside information, sticks (for "tipping," for
example) are available to combat the hazard. In contrast, the law
will not enforce "bargains" for carrots made between potential res-
cuers and drowning victims. 101 And although anticipatory contracts
made long in advance between these parties may well be enforced,
affirmative misbehavior (such as setting fire to a neighbor's house
after he agrees to pay for a rescue) is severely penalized, and the
parties themselves know to be wary of the moral hazard created by
these contracts. 102 I may, for example, contract with a reliable
friend for some future activity in a way that I would not contract
with a stranger.

101 The statement in the text is not based on real case authority for there appears to be
none. Still, it is almost inconceivable that a profitable rescue contract, "bargained" for dur-
ing an emergency, would be enforced. Courts would surely put terms like "duress" and
"public policy" to good use. A contract made under duress is voidable at the insistence of
the victim. See E. Farnsworth, supra note 78, §§ 4.9, 4.16-4.20 (discussing the requirements
for a showing of duress).
102 A and B may agree to rescue one another in the event of a future fire. But A is un-

likely to agree with stranger C that if A's house catches fire C will save A and earn a good
deal of money. C, after all, may be tempted by these terms to start a fire.

HeinOnline -- 72 Va. L. Rev. 908 1986

1986] Rescue 909

"Rare sticks and some negotiated carrots" is thus quite explica-

ble, given the now familiar behavioral effects involved, but by no
means entirely predictable or stable. Variety may extend to pack-
ages with a few more or a few less sticks, for example, but it is
surely unlikely that a legal system will offer "sticks and (allow) no
negotiated carrots" or "no sticks and large carrots" because such
packages would, respectively, stifle the production of desirable in-
formation and generate serious moral hazard and excessive precau-
tion problems.

3. Admiralty Law

The set of incentives available under American law to salvors at

sea is arguably a bit less consistent with the analysis begun in Part
I than is the treatment of finders, disclosers, and rescuers on shore.
The general rule for property salvage in admiralty law is "carrots
and no sticks;" successful salvors are compensated and yet one
need not attempt the rescue of a sinking ship any more than one
must volunteer to fight fires on land. 10 3 Commentators seem to
agree that compensation for salvage, while somewhat arbitrarily
determined, does more than reimburse salvors for their expenses
and may be generous enough, in fact, just to succeed in promoting
a competitive market in such rescues.0 So long as salvage awards
are not too great, it is possible that the unbalanced package just
described does not contradict the analysis in Part I. The absence
of sticks ensures that there will be no activity-level effect and the
presence of moderate carrots encourages rescue without encourag-
ing excessive precautions. My impression is that most seagoing ves-
sels are usually guarded at all hours so that there is not much of a
moral hazard problem generated by the offer of carrots to salvors.
To be sure, there is quite a history of land-based pirates' sending
out false signals in order to induce ships to take courses that cause
them to crash on rocks. But given that these pirates have sought to
help themselves to the spoils rather than to collect lawful salvage
claims, it is arguable that carrots have little to do with the criminal

See G. Gilmore & C. Black, The Law of Admiralty §§ 8.1-8.4 (2d ed. 1975).
' See Landes & Posner, supra note 19, at 100-05 (the goal of salvage law is to recon-
struct the contract that the parties would have made ex ante if a competitive market trans-
action had been available).

HeinOnline -- 72 Va. L. Rev. 909 1986

 Virginia Law Review [Vol. 72:879

behavior just described. 10 5 In any event, sticks seem inappropriate

for potential salvors, for there are few passers-by at sea and it is
hard to see how, for instance, everyone in Boston could be penal-
ized for nonrescue if a radio station announced that a ship in Bos-
ton Harbor was in distress and needed help. The activity at issue is
not beach strolling or sudden heroism but long term investment in
tugboats and other salvage equipment. Carrots that provide a com-
petitive rate of return seem both necessary and sufficient to pro-
mote this end. Sticks are, of course, necessary to deter criminal
activities such as piracy; but the magnitude of these sticks and the
enforcement strategies that accompany them are topics that can be
separated from the questions of the ideal magnitudes of sticks and
carrots held out to potential rescuers.
If the moral hazard problem accompanying these carrots is a se-
rious one, perhaps because potential salvors collude with crew
members, then the carrots for salvors could be smaller, if not
downright skeleton-that is, simple reimbursement for ex-
penses-by awarding smaller carrots for all rescue efforts rather
than larger ones only for successful rescues. But such a scheme is
almost surely more trouble than it is worth, for unemployed sal-
vors might flock to the scene of a hopeless disaster and insist later
that they had genuinely tried to save the day and should all be
compensated. The problem may be no more than a factfinding
nightmare 0 6 in which errors are costly because they generate ex-
cessive precautions by owners. But this is probably enough to lead
one to expect, from the perspective of the uniformity-variety
theme, that "no sticks and skeleton carrots (to all who attempt
rescue)" will not be found interchangeable with "carrots and no
sticks."
The "carrots and no sticks" rule is abandoned in the law's treat-
ment of pure life salvage, when the salvor saves human life and is
unable to attach his claim to one for property salvage. When both
life and property are saved in one incident (but even, or especially
if, by different salvors) the life salvor does recover, so there is little

105 Criminals who sought salvage awards would increase the chance of their own identifi-
cation and conviction. It is, therefore, hard to imagine that salvage rewards had anything to
do with piracy.
106 See Landes & Posner, supra note 19, at 104 (noting that it is easier to monitor
outputs).

HeinOnline -- 72 Va. L. Rev. 910 1986

1986] Rescue

chance that a strategic salvor will ignore life and save property. 10 7
When life alone is at stake, however, no carrots are offered. More-
over, there is both a commonly accepted moral duty and a penalty-
reinforced statutory duty "to render assistance to every person
who is found at sea in danger of being lost." 10 8 Variety thus seems
to appear just where it might have been expected because this
"sticks and no carrots" package can be viewed as an alternative
method of encouraging certain behavior. The discussion in Part I
suggested that this sort of package was attractive because the de-
nial of carrots eliminates problems of moral hazard and excessive
precautions (although it is questionable whether these problems
are present in the admiralty context). Carrots may also be inferior
to sticks, especially in this setting, if the potential rescuers are con-
cerned that rescuees will be judgment proof and if the state has
not promised to step in and pay deserving rescuers in such circum-
stances. Moreover, the presence of sticks was said to generate an
activity-level problem only where there were indentifiable rescue
spots that potentially liable parties might avoid. It is arguable that
there is an especially insignificant activity-level problem generated
by sticks for nonrescue at sea. To the extent that familiar sea lanes
present indentifiable rescue spots, potential rescuers will be in
these lanes as a matter of course-so that no substantial detour is
required to accomplish a rescue-and the advantages of using
charted sea lanes almost surely outweigh the cost of rescuing lives
that happen to be in need of rescue in these very lanes.
Even if the "sticks and no carrots" package does create an unfor-
tunate activity-level effect, there is little room for variety. It has,
for example, been alleged that vessels have steered clear of normal
sea lanes in the South China Sea to avoid coming across "boat peo-
ple" in great need of rescue.109 This activity-level effect may be

1*7 For the detailed rules, see G. Gilmore & C. Black, supra note 103, § 8.12; see also

Landes & Posner, supra note 19, at 105 & n.48 (noting that law must be careful not to
increase opportunity costs of altruistic life-saving and citing fascinating and supporting
cases).
108 See generally G. Gilmore & C. Black, supra note 103, §§ 8.1, at 534 n.7, 8.12 (discuss-
ing duty to rescue in American admiralty law).
109 See, e.g., N.Y. Times, July 15, 1979, at 1, col. 3 (reporting on allegations that Japanese
vessels avoided refugees). The overwhelming impression one gets from reading other news-
paper articles in this period supports the contrary conclusion discussed in the text; potential
rescuers will, as a matter of course, be in the sea lanes to which victims will aim so that
actual detour costs of such rescues will be low.

HeinOnline -- 72 Va. L. Rev. 911 1986

912 Virginia Law Review [Vol. 72:879

quite striking because people in need of rescuers would surely try

to manuever their inadequate boats to spots where rescue would be
most likely. But even if the "sticks and no carrots" package gener-
ates this behavior it is hardly the case that "carrots and no sticks"
would be superior. First, these victims are not in the financial posi-
tion to guarantee the carrots. Second, in other circumstances not
involving boat people victims in need of rescue at sea must often
have been aboard other vessels whose owners will be liable in tort
(for the negligent navigation of the vessel, for example) for all inju-
ries suffered. If the owners of these vessels must because of subro-
gation arguments pay life salvage awards, then it is possible that
other tort claimants will simply go uncompensated because total
liability is as a practical matter often limited to the value of the
vessel and of insurance policies carried. Carrots may thus accom-
plish questionable redistributions and encourage one form of desir-
able behavior only at the expense of another. 110 Third, there is no
reason to think that penalties decrease the total number of rescues;
some ships may, at a minor cost, avoid the usual sea lanes, others
will continue as before, but it is hard to imagine that fewer ships
will venture out to sea or that the withdrawal of penalties will sud-
denly cause more ships to undertake the rescue of persons they
encounter.
"Sticks and no carrots" may thus be superior to "carrots and no
sticks" in the context of life salvage. Of course, as noted earlier,
carrots could be guaranteed by a government or by a relief organi-
zation so that "sticks and carrots" would be even better than
"sticks and no carrots." These carrots must not be too rewarding
and there must be some policing of salvage claims because there is
the moral hazard of "salvors" colluding with "victims," but it is
likely that some positive level of carrots will do more good than
harm. Interestingly, English law does provide for a special fund for
life salvors."' In short, it would not be suprising to see a move
towards "sticks and carrots," but the prevalence of "sticks and no
carrots" for pure life salvage is especially interesting because it dis-
places the "carrots and no sticks" norm in salvage situations pre-
cisely in the circumstances in which sticks are feasible (or at least

'o The point is discussed in Friedell, Compensation and Reward for Saving Life at Sea,
77 Mich. L. Rev. 1218, 1254-55 & n.131 (1979).
"I Id. at 1232.

HeinOnline -- 72 Va. L. Rev. 912 1986

1986] Rescue 913

not completely unworkable) and in which carrots are either unbe-

lievable, ultimately financed only by other sympathetic victims, or
the cause of a potential moral hazard. This displacement is thus
quite consistent with the notion that variety among legal rules will
be found where such variety does not lead to terribly different or
destructive behavior.
The preceding analysis may, however, be entirely premature.
There is some reason to think that the traditional "no carrots" rule
for life salvage has been slowly eroded and that recovery by pure
life salvors is increasingly likely.11 2 Such a development, combined
with the obligation to save life at sea, would be an obvious step
towards the symmetry introduced in Part I and demonstrated in
this Part. It may, moreover, signify that at the present time carrots
are thought not likely to generate moral hazard problems and not
too frequently to be at the expense of sympathetic tort claimants.
But there is no reason not to expect variety over time and across
jurisdictions between "sticks and carrots" and "sticks and no car-
rots." There is, on the other hand, every reason to think that uni-
formity will extend so far as to hold off a "no sticks" approach
because, as discussed earlier, it is hard to see how such sticks could
be counterproductive when minor detours at sea to rescue lives are
concerned. On the other hand, "no sticks" is quite predictable in
the context of property or property and life salvage, because it is in
this context that investment in equipment and prepositioning are
normally required. Such investment and prepositioning are pre-
cisely the sort of activities that might be discouraged by sticks for
nonrescue.

B. European Civil Law Systems

The balance between the sticks and carrots offered to potential
rescuers is hardly unique to American law. Although its durability
is perhaps increasingly impressive the further one looks from
American law-for a society's legal rules then appear predictable

'2 See id. at 1271-87. The discussion in this article excludes the possibility that the Coast
Guard and other "public rescuers" might collect for their successful (or all) efforts. Essen-
tially, this is best seen as a contractual question. My taxes support the local fire depart-
ment, but the government could presumably declare that free fire protection is only availa-
ble to nonnegligent homeowners. Such a government's success in recovering for
noncontractual emergency services would then depend on many of the considerations sug-
gested earlier.

HeinOnline -- 72 Va. L. Rev. 913 1986

914 Virginia Law Review [Vol. 72:879

even without much knowledge of its cultural norms and sociologi-

cal features-it is useful to move away from American law rather
113
slowly and to consider first modern European law.
The most striking aspect of the treatment of rescue in modern
European civil law systems is that sticks are often available for and
applied to nonrescuers. These sticks usually take the form of crim-
inal penalties, consisting of fines, imprisonment, or both, and may
11 4
be accompanied or extended by exposure to civil liability.
French law, for example, provides a fine and jail sentence for
someone who does not undertake a riskless rescue of an endan-
gered person.'1 The statute applies only when a person-and not
property alone-is in danger.11 6 The nonrescuer is also subject to
civil liability, and such liability is triggered by both intentional and
negligent omissions.117
In practice, French law provides skeleton carrots to go along
with the sticks just described. If a rescuer has suffered damage or
incurred expenses, recovery will usually be possible from the
rescuee who will be said to have impliedly contracted for the res-
cue effort, from a tortfeasor who will be said to have caused the
rescue attempt, from the rescuee or his insurer who may be identi-
fied as unjustly enriched, or simply from the rescuee by an action
of negotiorum gestio, under which one who acts for the benefit of
someone who is unable to take care of himself is able to recover,
especially when the unilaterally created "agency" turns out to be
useful to the principal.1 1 8 Inasmuch as all but the very first of
these four routes to recovery by the rescuer promise rewards suffi-

113 The text does not focus on the treatment of rescue in Roman law. It appears that

although Roman jurists, through the labels of actio utilis and actio in factum, sought to
sanction some wrongful omissions, there was no clear move towards an obligation to rescue
unrelated persons or property. See B. Nicholas, An Introduction to Roman Law 220-21
(1962). Nevertheless, one might recover for the expenses of such a rescue, within the con-
fines of the principle of negotiorum gestio. Id. at 231-33. I have avoided more discussion of
Roman law precisely because there is such limited evidence of its practices and rules relat-
ing to rescue and nonrescue.
114 See Rudzinski, supra note 7, at 108-15.
'" See generally Tunc, The Volunteer and the Good Samaritan, in The Good Samaritan
and the Law, supra note 5, at 43, 45-48 (discussing article 63 of the French Penal Code and
its provision that one who abstains from a riskless rescue incurs a punishment of three
months to five years in jail in addition to a fine).
116 Id. at 47.
1" Id. at 49-50 (civil liability for negligent omissions even when no crime).
18 Id. at 52-54 & n.17.

HeinOnline -- 72 Va. L. Rev. 914 1986

1986] Rescue 915

cient only to reimburse the rescuer for expenses, the package pro-
vided by French law is best described as "sticks and skeleton car-
rots." Unless factfinding errors are great and are biased in the
rescuer's favor, these carrots are surely insufficient to generate
moral hazard problems and most unlikely to lead to excessive pre-
caution by potential rescuees. Moreover, these carrots can be de-
clined by rescuers who wish no material rewards. Although the
sticks offered in French law could conceivably lead to an undesir-
able activity-level effect if potential rescue spots are identifiable,
reasonable people can disagree whether French-style "sticks and
skeleton carrots" or "no sticks and skeleton carrots" is prefera-
ble." 9 The point is that the "sticks and skeleton carrots" package
offered in French law is an illustration of the sort of variety one
might expect after considering the likely behavioral effects of pen-
alties and rewards of various size. 120 "Sticks and no carrots" would,
in contrast, be quite a surprising package, for the behavioral effect
of not reimbursing the expenses of rescue (that is, the denial of
skeleton carrots) and the activity-level effect of penalties, however
mild, would together be enough to convince lawmakers that such a
package would discourage rather than encourage desirable rescues.
Other European legal systems also provide "sticks and skeleton
carrots." Nonrescue is a criminal offense in most of Europe; penal-
ties range from a fine only to imprisonment of up to five years and
fine. 12' In some systems the penalty is increased if the victim dies,
in one legal system there is no penalty unless death actually results
from the failure to rescue, and in at least one legal system the
nonrescuer is punishable even if, because of fortunate subsequent
events, no harm at all befalls the victim. 122 About one half of the
countries with sticks for nonrescue limit their applicability to situ-

119 There could also be disagreement over whether the carrots should be more than skele-
ton in order to better overcome the activity-level effect of the sticks.
120 The text does not mean to suggest that this package is predictable, but rather that it

is one of a group that might have been expected. Again, "sticks and carrots" would also have
been possible.
321 See Rudzinski, supra note 7, at 110 (reporting penalties that "range from a fine only

(Turkey) through a fine or simple detention (Denmark), fine or imprisonment (Norway),

prison or fine (Germany), correctional labor or social censure or applying measures of social
influence (Russia), arrest or fine (Italy), detention and fine (Netherlands), imprisonment or
fine or both (France, Belgium), imprisonment or arrest (Poland), to deprivation of liberty
only (Rumania, Czechoslovakia, Hungary)" (footnote omitted)).
122 Id. at 108-10.

HeinOnline -- 72 Va. L. Rev. 915 1986

 Virginia Law Review [Vol. 72:879

ations involving sudden and imminent danger to human life, so

that there is no liability for failing to rescue someone from nonfa-
tal injury. 123 In a few European countries there is a clear private
right of action against the nonrescuer; it is uncertain to what ex-
tent such civil suits would be successful in other countries with
statutes that impose criminal sanctions for nonrescue but are silent
regarding civil liability.1 24 In some countries the potential for crim-
inal liability is limited to persons who actually witness or find a
victim in distress, 25 in at least one there is criminal liability for
nonrescue even if there is serious risk involved in a rescue at-
tempt, 126 and in some, all risk can be avoided by obtaining the help
of professionals 7or by informing the proper authorities of the need
12
for assistance.
It appears that the very modest carrots offered to rescuers in
French law are also available in other European systems. German
courts, for example, have explicitly extended the implied agency
notion of negotiorum gestio to allow the rescuer not only to re-
cover the expenses of "managing the affairs" (the rescue) of the
"principal" in need of assistance but also to recover for injuries
suffered as a result of the rescue effort. 128 Whether courts in other
countries use this means of granting recovery, or instead, extend
the notion of general average contribution in admiralty law to
emergencies on land 12 , or simply follow explicit legislative direc-
tions to compensate rescuers,' there is every reason to think that
"sticks and at least skeleton carrots" is the rule in a great many
European legal systems.
The balance I have suggested between sticks and carrots would
appear even more durable-and the uniformity-variety thesis itself
especially attractive-if some legal system offered carrots (but not
terribly unbalanced ones) that were financed by the state. Reason-
able lawmakers might, after all, think that excessive precautions

1' Id. at 96 (Netherlands, Norway, Denmark, Poland, Czechoslovakia, and Germany limit
duty to situations involving danger to human life).
124 See id. at 112-15.
125 Id. at 101.

126 Id. at 105-06 (only danger to life of rescuer relieves responsibility in Rumania).

127 Id. at 108 (in Italy, Turkey, Rumania, and Soviet Union law formulates alternative

duty to render help or inform immediately the proper authority).

128 See id. at 115.
128 Id. at 115-16.
230 Id. at 116 (noting Austrian, Polish, and Czech legislation).

HeinOnline -- 72 Va. L. Rev. 916 1986

1986] Rescue

by potential victims constitute a more serious problem than the

cost of carrots and the special moral hazard created by tempting
"victims" to collude with "rescuers" in order for the latter to col-
lect their carrots from the fisc. Remarkably, Austrian law appears
to provide just such carrots: a cash award from a public fund is
available by law to someone who risks his life and saves another
from imminent death.1 31 It seems, however, that in practice the
awards are quite discretionary, more like occasional medals than
calculated carrots, and sufficiently unfamiliar that few potential
rescuers could possibly be motivated by their existence.' 32

C. The Treatment of Rescuers in Primitive Laws

The analytic tools developed in this article are even useful in
explaining the carrots and sticks held out to potential rescuers in
civilizations far removed from our own. Although there is little, if
anything, in any code of laws of which I am aware that contradicts
the thesis developed in this essay, there is also not a great deal in
many ancient or primitive legal systems that supports my thesis,
for in many cases there is simply very little information regarding
the treatment, if any, of finders, salvors, and other rescuers. There
are, however, significant and fascinating clues about the law of res-
cue in three primitive legal systems. In this section I report the
rules that contain these clues and consider the ways in which the
now-familiar balance between sticks and carrots may or may not
be reflected in these rules. I begin with the Code of Hammurabi,
thought to have been promulgated by the ruler of Babylon be-
tween the twenty-second and eighteenth centuries B.C., continue
with the laws of the ancient Hittites, thought to have been codified
between the fifteenth and thirteenth centuries B.C. in a civilization
located in modern-day Turkey and Syria, and conclude with the
customary laws of various Mongolian tribes, thought to have
emerged in the thirteenth century and evolved through the seven-
teenth and eighteenth centuries when many of these customs were
codified.' 3 Although these primitive provisions may at first seem

Id. at 117.
"' Interview with Herbert Hausmaninger, Professor, University of Vienna Law School
(Aug. 7, 1985).
133 See G. Driver & J. Miles, 1 The Babylonian Laws 3, 34 (1952); E. Neufeld,
The Hittite
Laws 70 (1951); V. Riasanaovsky, Customary Law of the Mongol Tribes 5-17, 20-21, 30, 39-

HeinOnline -- 72 Va. L. Rev. 917 1986

 Virginia Law Review [Vol. 72:879

quite unrelated to the very real questions posed by contemporary

rescue law, I warrant that the trusting reader will find these old
rules remarkable and informative. The problem of encouraging de-
sirable rescues, and of creating affirmative obligations in general,
is, as will soon be clear, hardly a new one. Unfortunately, the mate-
rial that survives from primitive legal systems is in a form that
requires careful (some would say impossible) textual exegesis. The
discussion that follows thus does not simply list various sticks and
carrots but instead explains why I think it fair to conclude that
certain incentives were offered to potential rescuers in these an-
cient or remote civilizations.

1. Hammurabi's Code
The Code of Hammurabi contains the earliest known rules that
deal with the question of affirmative obligations.3 After providing
for serious punishment for one who "has harboured in his house" a
slave he knows to be a fugitive,13 5 the Code provides in section 17
that: "If a man has seized in the field a fugitive slave, male or fe-
male, and has brought him back to his lord, the owner of the slave
shall pay him two shekels of silver." There is no provision that
affirmatively requires the capture and return of a fugitive slave or
of any other property; sticks were presumably limited to active
misdeeds such as the harboring-as opposed to the overlook-
ing-of a fugitive. The carrot provided in section 17 is, however, of
substantial size; two shekels is, for example, the compensation sug-

40, 52 (1981).
M The laws of the city-state of Eshnunna are thought to be the only older (surviving)
system of laws, and they contain no provisions that can safely be interpreted as dealing with
omissions. See G. Driver & J. Miles, supra note 133, at 6; R. Yaron, The Laws of Eshnunna
2 (1969). The laws found in the Old Testament are not thought to have been codified earlier
than Hammurabi's Code, but it may be useful to note that although Leviticus 19:16, for
example, provides that "neither shalt thou stand idly by the blood of thy neighbor," it is not
at all clear that such a provision carried with it a sanction or was otherwise something other
than a suggested moral standard. See 1 Pentateuch and Haftorahs 501 (J. Hertz ed. 1941).
Similarly, the Old Testament requires one to "find" and return lost property. See Exodus
23:4-5. But, again, no penalty is indicated for the failure to rescue and no reward is provided
to the rescuer. It may, of course, be quite significant that earlier codes announced such
standards, but my point is that one might just as well say that modern American society
condemns one who stands idly by while another is in danger and, yet, legislates no sticks for
so many nonrescuers.
135 Code of Hammurabi § 16. All translations of H1ammurabi's Code in this article are

from C. Edwards, The Hammurabi Code (1904).

HeinOnline -- 72 Va. L. Rev. 918 1986

1986] Rescue

gested for one who builds a sixty-ton boat (perhaps as a member of

a typical work crew) and is twelve times the daily rental price of
such a boat.1 36 On the other hand, the use of the word "field" in
section 17 seems to imply that the fugitive was seized not on city
streets but out in the country. It is therefore possible that the
lawmaker meant to supply only a modest or even a skeleton carrot
to the "rescuer," for the work and time required to seize a slave
and transport him to the city could have been substantial. It is
most unlikely that this effort would have equaled that required to
build (or help build) a boat, but because there was surely a fair
amount of risk involved in this sort of work, it is possible that po-
tential rescuers would have regarded two shekels more as reim-
bursement than as enticement.
It would not, however, contradict the analysis in Part I of this
article to discover that one who "rescued" lost slaves earned sub-
stantial rewards. It will be recalled that the drawbacks of a "no
sticks and large carrots" incentive system include the problem of
excessive precautions by potential victims, who will wish to avoid
the payout of large carrots, and the moral hazard that emergencies
will be created by profit seekers. But in the circumstances de-
scribed in section 17 of Hammurabi's Code neither of these draw-
backs is present. Slaveowners must have had a fair incentive to
guard against escapes even in the absence of section 17. And the
primary means of guarding against escapes would surely have been
to promise slaves that large penalties would accompany their re-
capture.13 7 Indeed, the reward provided in section 17 might simply
have worked to discourage escapes, for slaves would have known
that strangers were encouraged by the law to help recapture escap-
ees and strongly discouraged by the law from harboring them or
helping them escape.13 8

136 Code of Hammurabi § 277.

'37 See G. Driver & J. Miles, supra note 133, at 107 n.6 (presuming penalties by owner
and noting that § 282 of the Code provides for removal of the ear of a slave who denies his
master). It is difficult to imagine a flourishing slaveowning society in which most slaves were
constantly shackled or watched, for such precautions would be extremely costly. Work
farms, for example, are hardly thought to be profit centers. Instead, an economically suc-
cessful society such as Hammurabi's Babylon must surely have threatened potential escap-
ees with punishment more than it would have constrained their movement.
138 Sections 15 and 19 of the Code of Hammurabi provide for the slaying of someone who

accompanies an escaping slave or hides an escaped slave in his house.

HeinOnline -- 72 Va. L. Rev. 919 1986

920 Virginia Law Review [Vol. 72:879

It is also likely that the moral hazard that typically accompanies

large carrots would have been minimal in the circumstances ad-
dressed in Hammurabi's Code. A profit-seeking rescuer could, for
example, have taken steps to ensure that a boat would soon require
towing but it is hard to see how he could have increased the num-
ber of escaped slaves in order to profit from section 17. This poten-
tial "rescuer" was not in a position to promise a slave a share in
the profit-and without a slave's cooperation, or escape, no cap-
ture would have been possible.
In short, unless there is evidence that in Hammurabi's day re-
wards were generally offered to finders and other unobliged rescu-
ers, there is good reason to discount the likelihood that the "no
sticks and large carrots" package provided for the capture of slaves
was, found in other contexts. The punishment of slaves after their
return would have substituted for inefficient precaution-taking by
slaveowners and would have more than offset the moral hazard af-
fecting intermeddler-rescuers. There is, therefore, every reason to
think that the drafter of the Code supposed a norm of "no sticks
and no carrots" or "carrots and no sticks" but provided for "no
sticks and large carrots" precisely where other factors worked to
eliminate the usual drawbacks of such an unbalanced package.
The only other provisions in the Code that can possibly be said
to deal with rescuers and their rewards provide that physicians
earn extraordinarily large rewards for successful operations but
must, when unsuccessful, pay compensation to their patients."3 9 A
comparison of these provisions with those found today in American
jurisdictions illustrates the uniformity-variety argument rather
nicely. If the ancient Babylonians wished to impose liability on un-
successful surgeons-perhaps because inevitable or nonnegligent
fatalities were hard to separate out from those caused by negligent
surgeons 4-then in order not to discourage aspiring surgeons
(create a serious activity-level effect) it was surely necessary to al-
low surgeons to charge high fees or to legislate such fees for suc-
cessful operations. It would have been disastrous to encourage suc-

139 Code of Hammurabi § 215 provides: "If a doctor has treated a man with a metal knife
for a severe wound, and has cured the man, or has opened a man's tumour with a metal
knife, and cured a man's eye; then he shall receive ten shekels of silver." Code of Hammu-
rabi § 218 provides: "If a doctor has treated a man with a metal knife for a severe wound,
and has caused the man to die, or has opened a man's tumour with a metal knife, and
destroyed the man's eye; his hands shall be cut off."
140 See Levmore, supra note 8.

HeinOnline -- 72 Va. L. Rev. 920 1986

1986] Rescue

cess in surgical techniques with large carrots only, for the moral
hazard of unlicensed surgeons' encouraging all strangers onto the
operating table would have been too great. Carrots alone or sticks
alone would clearly have been unwise. As between the two possible
balanced packages, the lawmaker obviously rejected the laissez-
faire "no sticks and no carrots"-perhaps because patients could
not be expected to bargain wisely over the ramifications of their
own nonrecovery-in favor of "sticks and carrots."
American law, by contrast, has adopted the "carrots and no
sticks" package. Physicians are hardly held to a more grueling lia-
bility standard than other members of society, are entitled to bar-
gain for substantial compensation, and are almost uniquely able to
collect their usual fees when noncontractually undertaking res-
cues. 141 But, again, the unbalanced package is found precisely
where the usual drawbacks are absent. It is surely easier for a
profit-hungry physician to encourage elective surgery and excessive
office visits than it is to create a medical emergency on the road
(without incurring tort liability). Moreover, the larger moral haz-
ard is controlled through licensing boards and professional stan-
dards-constraints that were surely less effective in Hammurabi's
day. And potential victims already have every reason to take pre-
cautions against such emergencies; the potential cost of a doctor's
services on the roadside will, for example, hardly change driving
behavior. To be sure, moral hazard problems and excessive precau-
tions might well be generated by an extremely unbalanced package
that offered physicians very large carrots for their rescue efforts.
But this is precisely the message of the uniformity-variety thesis;
one should not expect to find a very unbalanced package in any
legal system, but among relatively balanced packages variety is un-
surprising for it is not obvious which package is best. In both
American and Babylonian law we find that a balanced package is
the norm but that physicians confront unique packages. In Ham-
murabi's Code it is a balanced package of carrots and sticks. In
American law the balance is closer to "no sticks and no carrots,"
but the package is somewhat unbalanced with respect to emer-
gency medical services-but this is precisely where an analysis of
the usual drawbacks to unbalanced packages suggests that less bal-
ance is needed.

" See Landes & Posner, supra note 19, at 109-11.

HeinOnline -- 72 Va. L. Rev. 921 1986

 Virginia Law Review [Vol. 72:879

2. The Hittite Laws

Assuming that section 17 of the Code of Hammurabi is limited
to the capture of escaped slaves, the ancient Hittite laws appear to
be the first to have designed carrots for finders, or rescuers in gen-
eral.14 2 Whereas earlier rules merely prescribe that a finder of lost
property should seek out its rightful owner, 41 3 the Hittite laws con-
tain the following provisions:
§ 22. If a slave escapes and anyone brings him back, if he finds him
in the vicinity, he (the owner) shall give him shoes; if (he is caught)
on this side of the river, he shall give him two shekels of silver; if
on the other side of the river he shall give him three shekels of
silver.
§ 23. If a slave escapes and goes to the country of Luiyas, to him
who brings him back he (the owner) shall give him six shekels of
silver. If a slave escapes and goes to an enemy country, he who
brings him back shall himself take him.

§ 45. If anyone finds utensils, he shall return them to the owner

and he shall reward him. But if he does not return them he is a
thief.

§ 71. If anyone finds an ox or a horse or a mule, he shall bring it to

the royal gate. But if he finds it in the country, it shall be brought
before the elders and he may continue to harness it. But if its
owner finds it and receives it as it was, he shall not hold him as a
thief. If he does not bring it before the elders, he is a thief.

§ 79. If oxen stray into a field and the owner of the field finds
them, he shall yoke them for one day, and when the stars come
out, he shall return them to the owner.
Here, as in Hammurabi's Code, punishment of the recaptured
slave was at the discretion of the slaveowner,'14 4 and carrots but not

142 See supra note 134.

The Hittite civilization flourished in the fourteenth and fifteenth centuries B.C. 14 Ency-
clopedia Americana (1975). The Hittite "laws" found between 1906 and 1912 are probably a
set of royal decrees and refer to an older set of laws. All references to and translations of
Hittite law are from E. Neufeld, supra note 133.
143Deuteronomy 22:1-3, for example, provides that a finder must seek out the rightful
owner of lost property but sets out no sanctions for finders who fail to carry out this obliga-
tion or injunction.
144 See supra note 137.

HeinOnline -- 72 Va. L. Rev. 922 1986

1986] Rescue

sticks are used to encourage the "rescuer." That these carrots were
graduated according to the distance the rescuer travelled simply
1 5
reinforces the idea that these carrots may not have been large. "
Once again, the incentives facing slaves surely overwhelmed the
problems of excessive precaution and moral hazard.
Section 45 extends this "carrots and no sticks" approach to
utensils-and thus almost certainly to identifiable lost property in
general.14 6 As the earlier analysis in Part I implies, we should ex-
pect to find "carrots and no sticks" in a number of legal systems;
indeed if the carrots are not in cash or otherwise not very likely to
generate moral hazard problems, this package seems superior to
"no sticks and no carrots." Moral hazard problems may be less se-
rious with regard to lost property than with regard to other rescues
unless evil intermeddlers can be expected often to steal property
and then come looking for a reward for its return. And even this
potential moral hazard may be avoided if property reported as
probably stolen by an owner does not later qualify as "lost" for
purposes of the reward of section 45.
The only difficulty in interpreting section 45 concerns its last
provision. The preceding discussion assumed that the finder was
labeled "a thief"-and would suffer some unstated punish-
ment147-not if he simply ignored the lost utensils but only if he
took possession of them. Similarly, the last provision in section 71
can be taken to mean that under Hittite law the finder is punished
only if he commits some wrongful affirmative act. The most likely
interpretation of the several sections then seems to be that a two-
pronged approach is legislated: (1) A carrot is offered to a finder
who returns lost property; in the case of lost animals, the finder's
reward is the use of the animal for one day (section 79), for the
cost of feeding the animal can hardly be equal to the value of one
day's work or animals would be worthless assets, and in the case of
utensils the reward is left for the parties to work out; (2) a filing
system is created for lost property that is expensive to transport
and not immediately returnable to its owners. Utensils would have
had family markings on them and, in any event, storage would

See supra text following note 136.

16 The ownership of a utensil could have been ascertained through family markings on
the utensils or through inquiries of local artisans. See E. Neufeld, supra note 133, at 163.
147 Id. at 161.

HeinOnline -- 72 Va. L. Rev. 923 1986

924 Virginia Law Review [Vol. 72:879

have been relatively easy. But one who found lost animals in the
country was encouraged by the Hittite laws not to bring them to
town where upkeep would be more expensive but instead to bring
word of their discovery to the elders who would match up finders
and losers.
This interpretation is not perfectly elegant; it would be more
convenient if section 79 preceded section 71 and if it, rather than
section 71, listed a variety of animals and thus signalled strongly
that its rule was illustrative. Still, I think the "carrots and no
sticks" package just described to be the best interpretation of
these sections.
It would not, however, do great violence to the uniformity-vari-
ety thesis to suppose that sections 45 and 71 do impose criminal
sanctions on one who ignores apparently lost property. Read this
way, the Hittite rules form a seemingly balanced "sticks and car-
rots" package, which may or may not be better than "no sticks and
no carrots." The important point is that it is virtually inconceiv-
able that the ancient document describes a "large carrots only,"
"sticks only," or "sticks and large (moral hazard-inducing) carrots"
approach.

3. Mongolian Tribal Law

As I have described elsewhere, 148 the customary law of
Mongolian tribes was full of "sharing" rules. A person who caused
an accident in nonnegligent fashion, for example, might have in-
curred partial liability so that he and the victim shared the burden
of the accident. 149 Although Mongolian tribal law did not always
provide for sharing, 150 the presence of some sharing rules may lead
one to expect the granting of recovery for rescue, for the rescuer
and rescuee would then share in the good fortune of a successful
outcome. But the treatment of rescue in these laws goes beyond
that which might be said to follow from whatever underlying sense
of justice generated the sharing rules. These laws are far better

M4Levmore, supra note 8. The Mongolian tribes referred to in the text were nomadic
cattle breeders and occasional farmers. Their customary laws were recorded between 1229
and 1709. See V. Riasanovsky, supra note 133, at 5, 17, 19.
14' Thus, if disease spread from A's cattle to B's cattle, A paid one-half the damages. V.
Riasanovsky, supra note 133, at 177.
150 For example, one who was injured while breaking another's horse was treated entirely
at the owner's expense. Id. at 158.

HeinOnline -- 72 Va. L. Rev. 924 1986

19861 Rescue

understood from the perspective of sensible packages of sticks and

carrots (in light of their likely behavioral effects) than from a per-
spective that simply admires or expects the repeated use of sharing
rules.
I have come across only two fragments of customary Mongolian
law that deal with rescuers other than finders. The first comes
from the Buriat Code of 1781: "Whoever does not come to the as-
sistance of a perishing person is liable to corporal punishment, but
whoever saves a perishing person is awarded a horse." 15 1 It is not
clear whether the horse was to be supplied by the rescuee or, some-
how, by public officials. The latter possibility would seem to create
real moral hazard problems, for A and B could conspire that one
will rescue the other in order to extract a horse from the "taxpay-
ers" at large. Both because the customary laws are elsewhere care-
ful not to create opportunities for mischievous strategic behavior
and because these laws generally set out the responsibilities of in-
dividuals and never explicitly speak of the expenditure of funds by
the government, I think it fair to assume that the rescuee was to
award a horse to the rescuer. In any event, the law clearly chooses
a "sticks and carrots" approach-with relatively large sticks and
carrots. I suspect-but I have no way of knowing-that evil inter-
meddlers were in some way successfully deterred from creating the
demand for their own services and thus from "earning" horses. Nor
is it possible to say whether in everyday life among the Buriats
excessive precautions by potential victims might have been a con-
cern or whether there were clear potential rescue spots, so that the
sanction imposed for nonrescue might have created a counter-
productive activity-level effect unless accompanied by a substan-
tial carrot. It is clear, however, that the package offered was a bal-
anced one.
The laws contain much more material regarding the rescuer we
call a finder. The Buriats treated finders much as they did more
traditional rescuers; sticks for nonreporting and a reward of ten
percent for finding. 1 52 Among the Kirghiz' 53 a finder was obliged to

' See V. Riasanovsky, Customary Law of the Nomadic Tribes of Siberia 55 (1965).
151 See V. Riasanovsky, supra note 133, at 185 (providing that the owner was bound to
give one-tenth to the finder and providing for corporal punishment for failing to report a
find).
"I The Kirghiz constituted a nomadic tribe in Siberia and Eastern Russia. They num-
bered more than three million at the beginning of the twentieth century. V. Riasanovsky,
supra note 151, at 7, 8, 13.

HeinOnline -- 72 Va. L. Rev. 925 1986

 Virginia Law Review [Vol. 72:879

spread news of his find or report it to the district authorities. He

received one-quarter of the value of the returned property (one-
''154
eighth in the case of cattle) "or whatever the master may give.
If he hid his discovery he was simply "without honor."' 155 Inasmuch
as our extralegal culture also looks unfavorably on one who fails to
act, it is fair to describe the Kirghiz rule as "carrots and no sticks."
If honor carried with it special privileges, then the Kirghiz' scheme
was more balanced. Among the Tunguses, a finder received one-
tenth but if he hid the thing found, then upon its discovery the
informer received the one-tenth and the first finder was subject to
corporal punishment. 58 It seems that a finder who simply ignored
the lost property and went on his way would not have been pun-
ished, although it may be that such behavior is simply not ad-
dressed in the reported laws. The approach taken thus seems to
have been of the "(substantial) carrots and no sticks" sort-with
the carrots extended to informers as well as to finders. Among the
Yakuts, a finder received a reward "in the measure of his labors
and expenses" up to one-quarter the value of what was found and
was treated as a thief if he concealed his find. 57 The language just
quoted implies that a one-tenth or even one-quarter commission
may have only reimbursed the finder for his time and expenses
and been viewed not at all as a large carrot. But either way, it is
reasonable to conclude that the Yakuts, Tunguses, and Kirghiz of-
fered finders "carrots and no sticks," while the Buriats opted for
"carrots and sticks."
Finally, among the Kalmucks a fine was imposed on one who did
not report the find of stray cattle. 5 8 No provision is made for the
reward of an honest finder. Inasmuch as the rule addresses stray
cattle that "join a herd,"' 1 9 it is, unfortunately, not clear whether it
might have been extended to include circumstances in which a
finder simply ignored and passed by apparently lost animals or

"I' Id. at 13.

155Id.
158 Id. at 75.
157 Id. at 88-89. The Yakuts inhabited North Central Asia and numbered about 250,000 at
the beginning of the twentieth century. Id. at 87.
"' The Kalmucks originated in western Mongolia and migrated first to Tibet and then to
the steppes of the lower Volga. V. Riasanovsky, supra note 133, at 244-45.
1I Id. at 276.

HeinOnline -- 72 Va. L. Rev. 926 1986

1986] Rescue 927

even other property. A most interesting twist in the law of the

Kalmucks concerns the disposition of found cattle that are turned
over to the authorities but then not claimed. The law provides that
after three years the cattle shall be sold and the proceeds spent on
the poor. 160 The law also calls for a reward "in proportion to the
value of the saved cattle" for the rescue of cattle from "wolves,
storms, blizzards or drowning in mud or wells." ' Inasmuch as
there is no reward to the honest finder, for apparently he gets
neither a commission from the owner nor the right to keep the
property if it be unclaimed, the system appears to be either of the
"sticks and no carrots" or the "no sticks and no carrots" variety.
Only when he rescues cattle from a serious emergency, such as
wolves or blizzards, is the rescuer treated to a carrot. Moreover,
since there is clearly no stick provided for someone who fails to
rescue cattle from blizzards, it appears that the Kalmucks
deployed a "carrots and no sticks" package for daring rescues. This
combination of rules is not terribly surprising. Even if mere finders
were penalized for ignoring lost property, it is very possible that
there would have been no activity-level effect under such a rule
because it is difficult to identify and avoid potential finding spots.
In contrast, it is easy to imagine that one could seek shelter when
wolves howl or blizzards rage so that the withdrawal of sticks in
the case of daring rescues may have worked to avoid counter-
productive activity-level effects precisely in the circumstances
where such effects might have been generated. It is possible that
the potential liability for carrots might have caused excessive pre-
cautions in general-but it is hard to imagine that precautions
against blizzards or wolves would have been any different if such
liability were imposed, for the probability of serious loss once bliz-
zards or wolves were on hand must have been very high. Similarly,
while it is barely possible that evil intermeddlers might have been
motivated by carrots to cause someone else's cattle to stray off and
thus create the need for their own rescue services, it would have
been much harder for such intermeddlers to stir up blizzards or
bring on the wolves. In short, the Kalmucks' use of carrots (and
possibly sticks) in some settings and not in others is neatly ex-
plained by an examination of the likely behavioral effects on own-
ers and potential rescuers of these carrots and sticks.

16oId.
'' Id. at 258.

HeinOnline -- 72 Va. L. Rev. 927 1986

928 Virginia Law Review [Vol. 72:879

Although the Kalmucks' rules reflect a concern about moral haz-

ard problems, a fair summary of the treatment of rescue in
Mongolian tribal laws in general is that potential moral hazards
influence the law much less than other factors. The discussion has
already exhibited the regular use of carrots' 6 2-and while these
carrots were surely designed to stimulate rescues and the return of
lost property in particular they could, theoretically, induce evil in-
termeddling. Perhaps the best illustration of this faith in the pre-
dominance of the positive aspects of carrots is found in the
Mongolian rule that a "witness is rewarded for helping justice: he
receives nine animals out of the (criminal) fine and a certain part
of the property in proportion to the quantity thereof judged by the
court."' 6 3 As one might expect, this carrot was accompanied by a
stick for false witness,' 4 but the interesting point is that variety is
found among legal systems precisely where it might be expected by
an aggressive application of theory. American law does not reward
witnesses and promises a stick (a low probability of a perjury pros-
ecution) to a dishonest witness. The stick seems unlikely to gener-
ate much of a marginal activity-level effect 6 5 and a carrot seems
unwise for moral hazard reasons, until one realizes that carrots can
encourage unknown witnesses to come forward and can, at least in
theory, be combined with sticks to discourage profit-seeking wit-
nesses. That just such a package was found in Mongolian tribal law
does not, of course, indicate that it is either culture-bound, supe-
rior, or inferior and intentionally abandoned in an evolutionary
way. It is, instead, one more example of the way in which variety is
found in a predictable range while uniformity-in the form of a
balance between rewards and penalties-outside of this range is a
rather durable phenomenon across legal systems.

"I2It is interesting to note that modern Mongolian law offers carrots to finders who re-
turn lost property. This carrot is equal to 10% of the value of the thing and is paid either by
the owner or by the state, if the latter gains ownership because no private owner claims the
lost property. W. Butler, The Mongolian Legal System 293 (1982) (translating article 100 of
the Civil Code of the Mongolian People's Republic).
1"3 V. Riasanovsky, supra note 133, at 93.

164 For example, for falsely accusing another man of theft a fine is imposed and any un-

just enrichment is extracted. Id. at 91. I assume that this rule (or an even harsher one)
applied to witnesses as well as plaintiffs.
165 See supra text accompanying note 32.

HeinOnline -- 72 Va. L. Rev. 928 1986

1986] Rescue 929

III. THE FUTURE OF RESCUE

There is no reason why the law regarding rescue could not re-
main just as it now is. There is no sudden outbreak of conflicting
decisions or hostile public opinion on the matter. Concern about
activity-level effects, excessive precautions, unwanted intermed-
dling, unnecessary judicial intervention in the private affairs of cit-
izens, and the monetization of moral norms may forestall any
changes in rescue law or even ensure that the clock is rolled back a
bit and that "no sticks and no carrots" is practiced as well as ad-
vertised. On the other hand, courts may continue to designate spe-
cial relationships"'6 and may often decide that rescuers shall re-
cover for their expenses when rescuees were negligent; if so, the
general rule will be "occasional sticks and infrequent carrots,"
rather than virtually none of each. Indeed, courts may even drop
these preconditions to their own intervention. My intuition is that
this second scenario, a steady evolution toward the European
model of "sticks and carrots," is more likely. I think that I am not
the only reader of these decisions who emerges with a strong sense
that the law regarding salvors at sea and finders of lost property is
much more stable than that dealing with disclosers of information
and noncontractual rescuers. 6 7 I would hardly argue that these
four problems are so alike that the latter two are destined to be
treated under the generous 6 8 "carrots and no sticks" approach

I'l See supra notes 56-75 and accompanying text.

1M This assertion is difficult to support with hard evidence. My impression derives both
from judicial decisions and from academic writings. No one suggests that salvors at sea re-
ceive carrots that are too large or too small or that sticks be used to encourage their efforts.
The little debate that has materialized is focused on the rewards for pure life salvage-a
tiny inlet, at most. Similarly, it is very difficult to find a judge or commentator suggesting
reform of the law affecting finders. There is some problem in distinguishing mislaid from
abandoned from lost property (compounded by a disinclination of courts to think in terms
of probabilities), some movement toward statutory carrots, see supra note 85, and some
discussion of the enforceability of rewards when it is criminal not to return found property
(but, interestingly, no discussion of the reward's role in causing the finder not to ignore the
property). See Note, Rewards for the Return of Lost Property. Are They Void in New
York?, 24 St. John's L. Rev. 287 (1950). But these issues are quite minor when compared to
the problem of motivating finders in general. In contrast, there is an enormous amount of
discomfort and discussion regarding the duty to rescue. For a tip of the iceberg, see Wooz-
ley, A Duty to Rescue: Some Thoughts on Criminal Liability, 69 Va. L. Rev. 1273 (1983)
(arguing for a criminal penalty for nonrescuers).
,18By "generous" I refer to the fact that salvage awards provide for more than reimburse-
ment and that finders are often allowed to keep unclaimed property. The expected value of

HeinOnline -- 72 Va. L. Rev. 929 1986

930 Virginia Law Review [Vol. 72:879

that seems so unshakable with respect to salvors and finders.

Sticks would surely create more of an activity-level effect on infor-
mation production than on findings and rescues. And customized
carrots, although used to encourage finders and disclosers, are un-
derstandably unenforceable when used by salvors and rescuers., 9
However, I do think that there is much more judicial and legisla-
tive discomfort with respect to the treatment of rescuers than
there is with regards to salvors and finders. In this discomfort may
lie the seeds of change.
The aim of the discussion in this part is not, however, to demon-
strate that the law of rescue is especially ripe for change. I argue
instead that an important theme that runs through the rescue
cases also appears in many important modern tort cases-and that
the evolution of rescue law may track that of tort law. Tort law can
be described in a manner that may first seem paradoxical as both
imposing liability on the single party that is the least-cost avoider
and as increasingly imposing some form of joint liability, rather
than no liability, when no such single avoider exists. And rescue
law can be described as expanding the obligations of potential res-
cuers precisely where a single party is the best or most obvious
rescuer. Finally, it is possible that rescue law will now follow tort
law and develop means with which to grapple with cases where no
single potential rescuer stands out. The recent evolution of tort law
may therefore give us a sense of the future of rescue law.

A. Single and Multiple Accident-Avoiders in Tort Law

The notion that much of tort law can be viewed as imposing lia-
bility on the party, or single avoider, best able to accomplish a so-
cially desirable or "would-be-bargain" result needs no rehearsal. 110
The search for a single and best avoider is easy to understand in
terms of transaction costs. If no liability is imposed, then the vic-
tim, assuming he is not the best avoider, must try to contract with
potential tortfeasors in order to avoid injury.""1 Such bargains with

the finder's residual claim almost surely exceeds the cost of picking up and announcing the
find.
169 Disclosers are "offered" their customized carrots in negotiations with other parties.
See supra text accompanying notes 101-04.
170 An excellent introduction (and summary) is found in R. Posner, Economic Analysis of

Law, §§ 6.1-6.17, at 119-159 (2d ed. 1977).

171 If A need not pay when he pollutes B's property, then B could bargain with A to avoid

HeinOnline -- 72 Va. L. Rev. 930 1986

1986] Rescue

many unknown parties will obviously be difficult to accomplish. If

the threat of liability hangs over multiple potential avoiders, then
they may undertake duplicative precautions or may each hope that
someone else avoids the accident so that no one in fact takes the
necessary precaution. In short, when transaction costs are substan-
tial, not only is the need for legal intervention clear but also the
advantages of aiming that intervention at a single avoider are ap-
parent. The single-avoider notion is somewhat contradicted by the
evolution of products liability law where a plaintiff can now sue
both the manufacturer and the dealer of the product he regards as
defective; 17 2 but it is fairly clear that in these cases the contractual
ties between the potential defendants must usually be thick
enough to overcome the potential for duplicative or inadequate
precautions that normally exists with multiple potential
17 3
avoiders.
In more complex settings the problems posed by multiple parties
are not so easily solved by imposing liability on a single avoider.
There is, I think, no paradox in suggesting that when single-
avoider solutions are feasible tort law evolves toward such solu-
tions, but that when multiple party problems are endemic tort law
will simply try to solve the deterrence problem as well as possible.
Broadly speaking, the multiple party problems are of two types,
interactive and horizontal. By interactive I refer to circumstances
in which harms can be prevented by multiple parties acting jointly
or alternatively. The manufacturer who can add more safety de-
vices to a machine and the consumer who uses this machine are
examples of such interactive parties. In such settings, the ideal be-
havior of one party depends on the behavior of another, whose own
ideal behavior depends in turn on expectations about the first
party's behavior. 1 74 Unless the parties that are involved and the
factfinder that reviews the situation later are extremely well-in-

some or all of such pollution.

172 See W. Prosser & W. Keeton, supra note 1, § 100, at 703-07.
171 Moreover, manufacturers are often such attractive defendants that plaintiffs may not
even bother to sue both the manufacturer and the retailer (with whom there was privity) of
the allegedly defective and injury-causing product.
171 Joint care refers to a situation in which A and B must take steps to avoid a harm (to
themselves or to some other party). Alternative care refers to a setting in which A or B can
take a step to avoid an injury. Some circumstances involve both, as when A can spend 30 to
avoid a harm (of 100, say) or B can spend 35 to avoid the same harm or A can spend 14 and
B can spend 10 to avoid the harm (jointly).

HeinOnline -- 72 Va. L. Rev. 931 1986

932 Virginia Law Review [Vol. 72:879

formed, these interactive (multiple party) problems are con-

founding. 1 5 Tort law has experimented with different kinds of lia-
bility rules including various contributory and comparative
negligence rules, but no rule is quite the match for the problem.176
I think it fair to say that the direction of change has been toward
partial recovery; comparative negligence rules, under which liabil-
ity is spread among parties who might have taken accident-avoid-
ing steps, have become increasingly popular.
"Horizontal" problems involve multiple misbehaving or poten-
tially misbehaving parties (whose ideal behavior depends very little
on that of other parties) and uncertainty regarding which of these
parties caused a particular harm. The law has had great difficulty
in settling on rules that best "match" injured plaintiffs with those
negligent defendants who truly caused their injuries.177 The law
dealing with these horizontal problems is hardly settled, but it is at
least arguable that recoveries are unlikely when there is a horizon-
tal matching problem with limited misbehavior, as when plaintiff
only knows that he was hit by a bus and more than one party
might have owned the bus that hit him,'7 8 but that probabilistic
recoveries are increasingly likely when horizontally aligned defend-
ants have all misbehaved, as when plaintiff's injury is caused by a
defective drug manufactured by one untraceable member of a
group of unrelated companies making the same product. 17 9 This re-

175 The factfinder must know the precise least-cost solution.

"I See supra note 175.
7 See generally Kaye, The Limits of the Preponderance of the Evidence Standard: Justi-
fiably Naked Statistical Evidence and Multiple Causation, 1982 Am. B. Found. Res. J. 487
(analyzing the use of statistical probabilities to apportion error among several defendants);
Robinson, Multiple Causation in Tort Law: Reflections on the DES Cases, 68 Va. L. Rev.
713 (1982) (damages should be apportioned among all defendants that created unreasonable
risks according to the magnitude of the risks created). The "matching" idea comes from
Sindell v. Abbott Laboratories, 26 Cal. 3d 588, 607 P.2d 924, 163 Cal. Rptr. 132 (1980).
178 See Smith v. Rapid Transit, 317 Mass. 469, 58 N.E.2d 754 (1945). But see Kaminsky v.

Hertz Corp. 94 Mich. App. 356, 288 N.W.2d 426 (Ct. App. 1979) (refusing to dismiss case
where 90% chance that truck was owned by defendant).
179 See Hardy v. Johns-Manville Corp., 509 F. Supp. 1353 (E.D.Tex. 1980) rev'd on
other
grounds, 681 F2d 334 (5th cir. 1982) (concluding that Texas courts would adopt the Sindell
approach to apportionment); Sindell v. Abbott Laboratories, 26 Cal. 3d 588, 607 P.2d 924,
163 Cal. Rptr. 132 (1980) (defendant drug manufacturers required to pay a proportional
share of the judgment as represented by their respective market shares). As a matter of
positive and even of normative theory, I think an important element of these cases is that
they involve defendants who are involved with numerous plaintiffs. As such, "averaging"
may seem more equitable and efficient because it becomes quite likely that tortfeasors and

HeinOnline -- 72 Va. L. Rev. 932 1986

1986] Rescue

view of single and multiple tortfeasors, or potential accident-

avoiders, sets the stage for a discussion of single and multiple
omitters, or potential rescuers. As we will see, the increasing will-
ingness of courts to accept probabilistic causation arguments in
cases involving multiple tortfeasors may signal-at least as a logi-
cal matter-a willingness to accept similar causation arguments
that must underlie claims that assert a duty to rescue.

B. Single and Multiple Potential Rescuers

I have not meant to use this article as a forum for the view that
the imposition of tort liability is always a good thing when a single
avoider can be identified or that it is not a good thing when inter-
active or horizontal multiple party problems are unavoidable. Sim-
ilarly, I do not wish to contend that no duty to rescue should be
imposed unless there is clearly a single potential rescuer to moti-
vate with sticks and carrots. Instead, I have only noted that as a
descriptive matter tort law can be characterized as searching for a
single best avoider, but increasingly imposing liability when there
are multiple avoiders and no single best avoider. Similarly, the dis-
cussion in this section now links the growth in liability for poten-
tial rescuers who have special relationships with victims to the no-
tion of the single rescuer. Although this link is promoted in this
article as descriptively useful, it is worthwhile-if only because
positive arguments are more appealing when they are coextensive
with arguments that are normatively plausible-to explore the rea-
sons why lawmakers might limit sticks (or sticks and carrots) to
circumstances in which there is a clear single potential rescuer.
Such an exploration is undertaken below, following the establish-
ment of the link between the duty to rescue and the notion of a
single potential rescuer.
When a victim's call for help goes unanswered, there may be
more than one potential rescuer within hearing range. Indeed, it
must often be the case that when B could have rescued A, B's pres-
ence at the scene only comes to the attention of the law because
other potential (and often unhelpful) rescuers, C and D, were also
at the scene of A's trouble. Such multiple potential rescuers pose a

victims will be properly "matched" over the entire set of cases. This is a topic that is, unfor-
tunately, quite beyond the territory of the present article.

HeinOnline -- 72 Va. L. Rev. 933 1986

934 Virginia Law Review [Vol. 72:879

doctrinal problem; if no rescue is attempted, it will be unclear

whether B, C, or D "caused" A's injury-and misbehavior without
causation of an injury is traditionally an insufficient basis for lia-
bility. The problem may not seem any different from one in which
two hunters negligently shoot in the direction of a third person in
their party and one (but it is unclear which) hits this unintended
mark.18 0 The hunters will, of course, be held jointly and severally
liable. 8 ' Apart from insisting that three potential causal agents
present a very different problem than do two, there are at least
three ways to distinguish the multiple nonrescuers from the multi-
ple shooters. First, one can simply follow the traditional line and
insist that the law distinguishes omissions from commissions. This
route will not be a terribly successful one both because some omis-
sions, such as draft evasion and tax evasion, are seriously and un-
derstandably penalized and because, as reviewed presently, rescue
is often obligatory.
Second, it is arguable that the causation problem in rescue goes
to the root of negligence so that rescue cases are more like the bus
case, 1 2 in which only one actor is negligent and none is held liable,
than like the shooting or defective drug cases, in which all actors
are negligent and at least partially liable. In rescue cases, after all,
if B rescues A, C and D's inaction is perfectly acceptable in both
legal and moral terms. As such, because in many cases each poten-
tial rescuer is probably unaware of the actions of other potential
rescuers, it is conceivable, even if horrible, that each behaves non-
negligently because each thinks that there is no need for his own
efforts. On the other hand, no hunter should shoot in the direction
of a camouflaged buddy and no drug manufacturer should turn out

180See Summers v. Tice, 33 Cal. 2d 80, 199 P.2d 1 (1948) (both hunters hit the unfortu-
nate victim, but the litigation concerns the more serious injury). It is arguable that the more
relevant analogy is to the cases in which both defendants do wrong and both injure plaintiff,
as when two negligently set fires consume plaintiff's property. See Kingston v. Chicago &
N.W. Ry. 191 Wis. 610, 211 N.W. 913 (1927) (when two "unnatural" fires cause damage,
both firestarters will be liable jointly and severally, and if only one firestarter is known it is
he who will be liable for the full amount of the damages). But as the text goes on to discuss,
neither analogy is quite perfect, for when there are multiple nonrescuers any one defendant
could have saved the day. The tort cases, on the other hand, concern defendants who might
-not have added to the harm at issue.
181 See Hall v. E.I. du Pont de Nemours & Co., 345 F. Supp. 353 (E.D.N.Y. 1972); Sum-
mers v. Tice, 33 Cal. 2d 80, 199 P.2d 1 (1948); Restatement (Second) of Torts § 433B, illus-
tration 9 (1965).
82 See supra note 178 and accompanying text.

HeinOnline -- 72 Va. L. Rev. 934 1986

1986] Rescue

defective drugs regardless of the actions of fellow hunters and

manufacturers. I must add that my own normative view is that this
distinction is unimportant. It may be sensible to deny recovery to a
plaintiff who offers only probabilistic evidence regarding which of
a number of bus companies caused his injury because in many
cases this rule might lead plaintiffs to investigate further and iden-
tify their tortfeasors more precisely. In contrast, the probabilistic
basis for believing that a nonrescuer may not have been negligent
is not capable of being altered after the fact. The problems of
probabilistic proof and probabilistic need for rescue are thus dif-
ferent. Still, I advance this second argument for distinguishing
multiple nonrescue from multiple tortious behavior because it is
certainly possible that it is this sort of argument that has caused
judges to hold multiple shooters and not multiple nonrescuers
liable.
A third difference between multiple nonrescue and negligence
may be one of time and not substance. Tort law has only recently
evolved to the point of assigning liability in probabilistic fashion
among numerous defendants. Two negligent hunters seemed to
present an easy case, perhaps because the probability of a "match"
was just one percent below that required in the usual more-likely-
than-not test. But multiple drug manufactuers would almost surely
not have been liable twenty or fifty years ago unless the
probability of a match between a plaintiff and a given manufac-
turer was also in the fifty-percent range. It is thus arguable that
multiple nonrescue and multiple affirmative negligence were
treated alike for many years and that evolution or experimentation
has recently affected the treatment of the latter only so that we are
now in a developmental period in which multiple negligent parties
but not multiple nonrescuers incur liability. This argument sug-
gests that the law may change in either area once again and multi-
ple nonrescuers and multiple negligent tortfeasors will then be
similarly treated.
A single identifiable nonrescuer, however, presents no matching
or deterrence difficulties and is very much like a paradigmatic sin-
gle tortfeasor. To be sure, it is possible to insist that nonrescue
always presents the problem of multiple parties. When, for exam-
ple, A's surviving family learns that B could have rescued A, it is
always possible that C and D could also have effected a rescue but
that their identities or presence at the scene of A's tragedy remains

HeinOnline -- 72 Va. L. Rev. 935 1986

936 Virginia Law Review [Vol. 72:879

unknown. But this uncertainty alone should not lead one to expect
that single nonrescuers will be excused from liability, because this
is the sort of uncertainty that the tort system frequently indulges.
When Z's negligently started fire sweeps across X's property, Z
may be liable even though Y may also have started a fire that was
on its way but is now unknown-perhaps because Z's fire con-
sumed the evidence of its origins and path. The case presents no
difficulty; X bears the burden of proving that there was at least
one negligently set fire, and Z can only escape full liability by iden-
tifying Y and then passing on the blame, sharing it, or seeking con-
tribution. 183 The remote possibility that a negligent Y was involved
but that Z is unable to identify Y and prove Y's involvement is
hardly allowed to interfere with Z's liability to X. Similarly, the
mere possibility of multiple nonrescuers is unlikely to interfere
with the liability of a single identifiable nonrescuer.
But the most important reason to think that the failure to rescue
is really not so different from the commission of a negligent act is
that, as indicated earlier,""' courts have discovered a surprising
number of special relationships as bases for the imposition of du-
ties to rescue. Most significantly, these special relationships have
one thing in common: when there is a special relationship there is
no multiple nonrescuer problem, for such a relationship is pro-
nounced only in circumstances in which there is one identifiable or
best-situated nonrescuer. This is not to say that whenever there is
but one nonrescuer, courts will insist that he and the victim had a
special relationship out of which a duty to rescue arose. The pres-
ence of a single nonrescuer is at present a necessary but not a suf-
ficient condition for liability; it is the growing number of special
relationships that indicates an increasing likelihood of liability on
the single nonrescuer. As indicated earlier, 185 the most remarkable
thing about these cases in which special relationships are found is
that some involve a single nonrescuer and a stranger.Liability has
been found appropriate for an innkeeper who could have protected

183 See Kingston v. Chicago & N.W. Ry., 191 Wis. 610, 211 N.W. 913 (1927). But see
Carpenter, Concurrent Causation, 83 U. Pa. L. Rev. 941 (1935) (discussing cases that reject
the Kingston principle and refuse to find liability unless defendant's negligence was a sub-
stantial factor in the cause of plaintiff's injury). On contribution, indemnity, and apportion-
ment, see W. Prosser & W. Keeton, supra note 1, §§ 50-52, at 336-55.
184 See supra text accompanying notes 56-77.
185 See supra note 77 and accompanying text.

HeinOnline -- 72 Va. L. Rev. 936 1986

1986] Rescue

a stranger from injury by one of the innkeeper's guests, 186 a safety

engineer who could have prevented an injury to a laborer he did
not employ,18 7 a psychologist who might have warned an identifi-
able stranger his patient was intent on harming,18 8 and, similarly, a
parole board, acting as a single entity, that might have warned
someone who was the target of a released convict.'"" In these cases,
there is of course no "relationship" at all. Instead, these cases con-
tain three elements. First, there is a single nonrescuer. Second, this
nonrescuer could with little effort have prevented a serious loss.
Third, this nonrescuer had no reason to think that someone else
would save the day. In short, although I think that a single
nonrescuer is not very different from a single negligent tortfeasor,
it is hardly arguable that the two are treated the same under cur-
rent law. But it is arguable that nonrescuers are increasingly incur-
ring liability and that through the concept of special relationships
this liability is most often imposed precisely where there is a clear,
single nonrescuer.
There is another clue that suggests in a striking way that the
major obstacle to holding nonrescuers liable is the historical disin-
clination to impose liability where multiple parties are at fault. It
has long been the case that if B begins to rescue A but then aban-
dons the rescue effort, then B will often be held liable for
nonrescue even though he is said to have had no duty to rescue in
the first place.' 90 Occasionally, B may have made A worse off by
leaving him in a place where other rescuers were less likely to help
or simply by using up time in which other rescuers might have ap-
peared. But it is rather clear that courts do not look carefully for
evidence of such causation but instead often hold B liable for what
might be regarded as active, or conscious, nonrescue.' 9' Clearly, B's
initial efforts single him out and "solve" any multiple party prob-
lem. B's false start allows the law to identify him-as opposed to

"84See supra note 57.

1S See supra note 67.
See supra note 69.
See supra note 72.
"o See W. Prosser & W. Keeton, supra note 1, § 56, at 378.
"o Id. § 56, at 378-82. For cases in which the volunteer made things worse, see id. § 56, at
378 n.55, 381 nn.83-85. To the extent that those who undertake (and then drop) a rescue
attempt move a victim or appear to passers-by to have the situation under control, I think it
plausible that their liability derives in part from a sense that the victim would have been
better off without such an attempt.

HeinOnline -- 72 Va. L. Rev. 937 1986

938 Virginia Law Review [Vol. 72:879

other potential rescuers-as the nonrescuer and to see him in a

light that is so similar to that which falls on a typical tortfeasor.
The inclination to hold the "withdrawing rescuer," B, liable is thus
a further indication that it is the problem of multiple nonrescuers
that is at the heart of any difference between omissions and com-
missions. 1 92 Similarly, the admiralty law rule that a vessel involved
in a collision must "stand by" to assist the other vessel or be
rebuttably presumed to have been at fault is, like attitudes to-
wards a hit-and-run driver on land, an example of the imposition
of an affirmitive obligation where an identifiable, single and obvi-
193
ously least-cost rescuer is on hand.
These arguments-that there is not a convincing analytic differ-
ence between the case involving two negligent hunters and that in-
volving two nonrescuers and that the expansion of special relation-
ships is no more or less than the gradual (if slow) imposition of a
duty to rescue on single identifiable potential rescuers-lead to
fairly specific predictions about the future of the use of sticks in
rescue law. First, there is every reason to think that nonrescuers
will increasingly be held liable, even in the absence of further stat-
utory innovation. Second, multiple nonrescuers may also be held
liable in the future, for now that tort law has grappled with multi-
ple parties there is less reason to think that rescue law will not do
the same. Finally, although this section has focused on sticks and
not on carrots, it is useful to repeat the thesis developed early in
Part II: American law may already provide at least skeleton carrots
to many rescuers. Whether or not the expansion of special relation-
ships will mean not only that an identifiable A might be liable for
not rescuing B but also that A could collect his expenses (or more)
from B is a difficult question. Many of the special relationships
that, at least at first, link A and B are contractual, so that A could
collect "negotiated carrots" from B. Landlords and psychologists,
for example, could surely raise rents and fees to finance the cost of
their increased precautions (or liability).19 4 It is, in short, possible

192 Similarly, "hit-and-run driver" statutes, which impose liability for failure to stop and

"rescue," see id. § 56, at 377, are easily described as imposing liability on the best positioned
single party who will realize that it is unlikely that someone else will save the day.
193 See G. Gilmore and C. Black, supra note 103, at 490-91.
194 Obviously these increased prices will decrease the demand for that which landlords

and psychiatrists offer. Still, in a competitive environment prices will eventually be related
to the average cost of production so that consumers will pay for the precautions they enjoy.

HeinOnline -- 72 Va. L. Rev. 938 1986

1986] Rescue

that American law will raise the rewards offered to potential rescu-
ers but also possible that it will, instead, simply leave these re-
wards to market forces. If liability for nonrescue extends one day
to all single or even multiple nonrescuers, then it is most likely
that skeleton or statutory carrots will continue to sprout. 195

IV. CONCLUSION
I have tried in this article to make a comparative claim and a
predictive claim and to base both on the likely behavioral effects of
legal rules affecting potential rescuers. The comparative, or histori-
cal, claim is that certain combinations of penalties and rewards are
more attractive than others-and that different legal systems
should be found to have selected from the most attractive sets.
The prediction is a bit more restrained: American jurisdictions
may increasingly impose liability for failure to rescue much as they
have increasingly imposed tort liability in general. There is also
every reason to believe that carrots will accompany these sticks.
My greatest hesitation in predicting these developments in Ameri-
can law concerns the evolution of the treatment of multiple com-
mitters. The decisions that have allowed probablistic claims
against multiple defendants have insisted that a large percentage
of such defendants be included in the claims and have concerned
cases in which the multiple defendants' behavior or products were
virtually identical. 96 In addition, the behavior of all the defend-
ants in these cases needs to be altered, while in multiple
nonrescuer cases only one rescuer is usually needed. Tort law does
not yet contain decisions that directly stretch interactive or hori-
zontal liability to cases in which multiple parties do different negli-
gent things-only one of which caused an injury-so that an in-
jured plaintiff could recover from all jointly and severally or from
each for a small percentage of the total claim.1 97 In the famous

1,& Such carrots have already sprouted in the form of Good Samaritan statutes. See supra
notes 45-81 and accompanying text.
' See Sindell v. Abbott Laboratories, 26 Cal. 3d 586, 612, 607 P.2d 924, 936, 163 Cal.
Rptr. 132, 145 (1980) (implying that when 90% included, "substantial percentage" require-
ment met).
'97 See Robinson, supra note 177, at 749-68. Of course, decisions that use the doctrines of
comparative negligence and proximate cause may accomplish such results in subtle ways.
The point in the text is simply that it would be regarded as quite pathbreaking for a judge
to spread liability in a setting in which A's injury had some small probability of being

HeinOnline -- 72 Va. L. Rev. 939 1986

 Virginia Law Review [Vol. 72:879

Kitty Genovese incident, for example, in which every one of thirty-

eight witnesses to a horrible assault failed to call the police or oth-
erwise assist the victim, 198 liability in the manner of tort law would
have required separate inquiries into the costs that each potential
rescuer would have needed to overcome, calculations regarding the
probability that a rescue would have been successful, and discount-
ing for the relative contribution of each of the nonrescuers. The
possibility that sticks will one day be imposed for such nonrescue
behavior must not be entirely dismissed, but it is not clear that
such probabilistically small commissions will be penalized and cer-
tainly not likely that penalties for omissions will develop ahead of
99
those for commissions.
There is, however, the possibility that the law will offer "carrots
only" rather than "sticks and carrots" or even "sticks only." These
carrots may need to be developed further in legislative rather than
judicial chambers 20 0 and would, presumably, only be developed by
lawmakers who thought that any moral hazard problems were rela-
tively small. The evolution of tort law for commissions is unhelpful
in predicting the further evolution of carrots, for while the identifi-
cation of a single rescuer or a limited number of rescuers is often
possible, the identification of a single non-tortfeasor is usually not
possible. It is obviously not feasible to reward all those who do not
negligently shoot guns or pollute but it is feasible to reward one
who first assists a victim in an emergency.2 0 I am left with the
sense that it would be rash to predict that sticks will be applied in
cases where there are many or many differently situated nonrescu-
ers. But it is far easier to predict that sticks will be imposed for
less complex cases of nonrescue, that carrots will accompany these

caused by each of ten different defendants in ten different ways.

"I See B. Latane & J. Darley, The Unresponsive Bystander: Why Doesn't He Help? 1-4
(1970); A. Rosenthal, Thirty-Eight Witnesses (1964) (describing the details of the assault in
which Kitty Genovese was stabbed to death over a forty-five minute period).
199For a more complete discussion of the ways in which comparative negligence might be
used to address this problem, see S. Levmore, The Tort System (forthcoming).
200 It is easy to imagine judges' holding that rescuers are entitled to reimbursement, but
very difficult to conceive of judges' allowing recovery for more than skeleton carrots. There
are, of course, analogies such as salvage to draw upon, but such a move seems more likely in
legislative chambers.
20I See generally Wittman, Liability for Harm or Restitution for Benefit, 13 J. Legal Stud.
57 (1984) (using an economic model to describe why governments tend to rely on coercion as
opposed to positive incentives).

HeinOnline -- 72 Va. L. Rev. 940 1986

19861 Rescue 941

sticks, and that carrots may be extended to rescuers (privately, ju-

dicially, or legislatively) even in circumstances that sticks never
reach.

HeinOnline -- 72 Va. L. Rev. 941 1986

HeinOnline -- 72 Va. L. Rev. 942 1986