Mischief and Misfortune
(Annual McGill Lecture in Jurisprudence and
Public Policy)
Jules Coleman and Arthur Ripstein”
This article argues that the problems of corrective and dis-
tributive justice are, at bottom, the same. The authors argue that
both can be understood as responses to the question: who owns
which of life’s misfortunes? Two extreme but unattractive posi-
tions set the range of possibilities. All misfortunes could be left
where they fall, or all could be held in common. Neither extreme
is attractive, because neither has room for the intuitive idea of re-
sponsibility, that is, that people should bear the costs of their ac-
tivities. Libertarians try to incorporate that idea by adding a rule
of strict liability for injuries as an exception to a general rule that
injuries should lie where they fall. Liberal egalitarians seek to
make room for responsibility by supposing that all misfortunes
should be held in common except those to which people will-
ingly expose themselves. The authors argue that the libertarian
and the egalitarian employ parallel strategies, neither of which
can succeed, because, on their own, the concepts of causation
and choice are indeterminate. The idea that people should bear
the costs of their activities to others depends on an objective
measure of the value of activities.
In place of the attempt to base an account of responsibility
on a neutral conception of agency, the authors articulate the form
of reasoning that is implicit in the fault system in tort. Fault as-
signs liability to injurers only when they have failed to exercise
reasonable care for the security of others. How much care counts
as “reasonable” depends on the liberty and security interests that
are at stake, which in turn depends on the importance of the ac-
tivity in which the defendant is engaged and the significance of
the plaintiff’s interest that might be injured. Thus, the idea that
people should bear the costs of their activities can only be ren-
dered determinate in relation to views about the importance of
various activities.
The authors argue that parallel positions show that issues
of distributive justice depend on views about the importance of
various goods in enabling people to pursue their own ends. The
authors concede that any such views will be controversial but ar-
gue that the idea that people should bear the costs of their activi-
ties requires it.
Cet article soudent que ls probtmes de justice corrective et
distributive sont, au fond, identiques. Les autears soutiennent que ls
deux concepts peuvent re compris comme des ulponses A ls question :
qui subit ls infortunes de la vie ? Deux positions exrmes mais pea
6tablissent l’Mendue des possibilit6s. D’un core, routes ls
attrayantes
inforttses pourraient re prises telles quelles, ou, de ‘autre, dies pour-
rmient ere considr6es ensemble. Aucun des extemes nest anrayant,
puisqu’aucun ne laisse place A l’id&e intuitive de la responsabilit6 qUe
ls individas devraient assumer ls cons&lUesces de lurs acres. D’une
part, ls libertaires essaient
d’incorporer ete ide en ajoutant ue regle
de responsabilit6 stricte pour ls prejudices. Cette dernire constitue
use
exception ft Ia fgle gdnrae voulant que ls prejudices soient consid&
rns comme teis. D’autre part, ls dgalitaristes lib&aux cherchent Zt faire
une place A la respousabilit en supposant que routes ls infortunes doi-
vent ure consid&6s ensemble, sauf cells auxquelles ls individus
s’exposent volontairement. Les auteurs soutiesnent que le libertaire et
r’6galitariste emploiet des strategies pa’allles, aucune d’elles ne pou-
vant rdussir puisque ls concepts de cause et de choix sont ind&ermi-
ns. L’idfe voulant que ls individus assument les consequences de
lurs acres envers ls autres depend d’une 6valhation objective de la
valeur de ces acres.
Au lieu de tenter de baser rdvaluation
de la responsabiitW sur
tne conception neutre d’agency, les auteurs y substituent le type de rai-
sonnemet qui est implicite au systme de faute en responsabilit6 civile
extra-contractuele. Selon cc demier raisonnement, in faute atteibue la
responsabilit A ceux qui causent le prejudice, uniquement lorsque cos
demiers ofht manqu6 A leur devoir d’agir en personne raisonnable t
l’6gatd de la sdurit6 des autres. A quel point le sudi pour la personne
considdr risonnable ? La rfponse A cette question dpend des
est-il
de libert6 et de s6atuit6 qui sont enjeu qui, t leur tour, ddpen-
int6rSts
dent de limportanc de r’activit6 dens laquelle
le dffesdeur s’est enga-
g6 et de l’int&f. du demandeur auquel on pourrait nuire. Ainsi, I’ide
selon laquelle les individus doivent assumer les consequences de louts
acres ne peut ae d&ferrln qu’en prenant en consid ration ls inter-
prdations portant sur l’importance de divers acres.
‘
Les autears soutienrat que des argurentations parallles d6-
montrent que les questions de justice distributive dipendest des inter-
pretations de l’rimportance de plusirars moyens qi pemsettest aux per-
sonrs de poursuivre lurs propres fins. Solon ls auteurs,
cs points de
vue seront controversfs, mais ils estiment que l’id6e que ls individus
doivent assurner ls consfquences de lous acres les rend nfcessaires.
. We are grateful to audiences at Chicago-Kent School of Law, McGill University, Faculty of Law,
The University of Manitoba, U.C.L.A., The University of Southern California, the University of To-
ronto, Western New England School of Law, and Yale Law School. Many people have given us help-
ful written comments. We are especially grateful to Gregory Keating for convincing us we did not
mean some of the things that we had said about strict liability in earlier versions.
O McGill Law Journal 1995
Revue de droit de McGill
To be cited as: (1995) 41 McGill LI. 91
Mode de r6fdrence : (1995) 41 R.D. McGill 91
MCGILL LAw JOURNAL/REVUE DE DROITDE MCGILL
[Vol. 41
Introduction: The Interdependence of Corrective and Distributive Justice
Synopsis
I. The Principle of Fairness
A. An Objective Standard of Costs
11. Misfortune
A. Misfortune and the Hobbesian World of Perfect Agency
B. Misfortune and the World of Perfect Community
C. Misfortune and the Principle of Fairness
I. Libertarianism: Causation as a Measure of Misfortune
A. The Inadequacy of Causation as a Measure of Agency
B. The Inadequacy of Strict Liability Based on Causal Paradigms
IV. Fault as a Measure of Misfortune
A. Boundary Crossing
B. Establishing the Boundary Line: Balancing Liberty and Security Inter-
ests
C. An Objective Measure of Fault
D. Limiting Liability
E. A Nuanced Objective Standard
E Strict Liability as a Measure of Misfortune in Rights Violations
G. Fault as a Measure – Conclusion
V. Liberal Egalitarianism: Choice as a Measure of Misfortune
A. The Inadequacy of Option Luck
B. The Inadequacy of Insurance as a Measure
VI. Primary Goods as a Measure of Misfortune
Conclusion
19951
J. COLEMAN & A. RIPSTEIN – MISCHIEF AND MISFORTUNE
Introduction: The Interdependence of Corrective and Distributive Justice
Questions of distributive justice may seem far removed from those of tort the-
ory. Corrective justice concerns the rectification of losses owing to private wrongs.
In contrast, distributive justice concerns the general allocation of resources, bene-
fits, opportunities, and the like. The duty to repair under corrective justice is agent-
specific – only wrongdoers need make up the losses of others. The duties imposed
by distributive justice are, in contrast, agent-general –
everyone has a duty to cre-
ate and sustain just distributions. Aristotle, the first to distinguish the two forms of
justice, argued that corrective and distributive justice represent distinct and mutu-
ally irreducible modes of moral reasoning.’
In spite of these analytic or conceptual distinctions, some advocates of eco-
nomic analysis have tried to reduce both forms of justice to a common denomina-
tor, typically either overall utility or economic efficiency. But, one does not have to
believe that corrective and distributive justice are reducible to some other moral
ideal in order to see important connections between them. First, in order to sustain a
just distribution, the effects of wrongful transfers must be annulled. In this way, the
institutions of distributive justice require institutions of corrective justice. Second, a
practice of corrective justice has the effect of sustaining the prevailing distribution
of resources by annulling the effects of certain changes in it. Because corrective
justice imposes a moral reason on agents to make good others’ losses, in doing so it
sustains the prevailing distribution. It follows that not every distribution will be
sufficiently just to warrant protection by such a practice. Only some distributions
can be sustainable by a practice of corrective justice; a requirement that unjust
losses be rectified only makes sense provided holdings are not entirely unjust.
Thus, corrective justice imposes constraints on the existing institutions of distribu-
tion. To the extent that tort law helps shape our conception of corrective justice, this
means that it is impossible to entirely separate tort law from the institutions and
demands of both distributive and corrective justice
Corrective and distributive justice are connected in another way as well. How-
ever distributive shares are to be fixed, their value will depend, in part, on the tort
regime that is in effect. Depending on which transfers are considered wrongful and
which not, the particular bundles of goods that individuals hold may turn out to
have very different values. Just as corrective justice sustains existing distributions,
so the value of distributive shares depends on ways of correcting wrongs.
‘ See Aristotle, Ethics, trans. J.A.K. Thomson (London: Penguin, 1984) at 176-79.
2 This much we take to have shown in our (separate) previous works (see: J.L. Coleman, Risks and
Wrongs (Cambridge: Cambridge University Press, 1992); A. Ripstein, “Equality, Luck, and Respon-
sibility” (1994) 23 Phil. & Pub. Aff. 3). The current project builds on our previous work, but it aims
to make a much more radical point. The thesis of this paper is that in important and underappreciated
ways the problems of distributive and tort law are the same.
MCGILL LAW JOURNAL/REVUE DE DROITDE MCGiLL
[Vol. 41
We hope to show a further connection between distributive and corrective jus-
tice. Life is full of unplanned and unanticipated events, and even its best, planned
aspects build on the result of earlier chance. Chance has a downside, of course. Its
name is misfortune. We want to suggest that both tort law and the institutions of
distributive justice can be understood as responses to the question: who owns which
of life’s misfortunes? We can distinguish between liberal and non-liberal answers to
this question, and between liberal and non-liberal institutions of tort law and distri-
bution. Part of what it means to refer to institutions as “liberal” is that they aspire to
express a conception of equality. In the case of both tort law and distributive insti-
tutions, that conception of equality aspires to allocate misfortune and its burdens
fairly. What does fairness require in the allocation of misfortune?
If we focus on misfortune’s costs, one plausible suggestion is represented by
the principle that each of us should bear the costs of our own activities. In eco-
nomic vocabulary, no one should be allowed to displace the costs of their activities.
So put, the principle has the ring of a tautology; if someone else ends up bearing
the costs of one’s activity, they turn out not to be one’s own costs after all. But the
point is normative as well as conceptual: however we might fix which costs belong
to whom, it is surely unfair that some should end up bearing costs imposed by oth-
ers.
I. The Principle of Fairness
1. Let us refer to the claim that each person should bear the costs of her activities
as the principle of fairness. That principle requires a conception of the costs of an
activity. Any form of liberalism must answer the question of where misfortunes
properly lie in a way that satisfies the principle of fairness. Different forms of lib-
eralism will provide different interpretations of the principle, and what it requires
by the way of institutions of tort law and distributive justice. In part, different con-
ceptions of what the principle requires result from the fact that competing interpre-
tations of liberalism offer conflicting conceptions of an activity’s costs. In fact, we
will argue that the two poles of broadly liberal thought –
libertarianism and egali-
tarianism –
adopt the same general conception of an activity’s costs. This concep-
tion identifies the costs of an activity with its causal upshots.
If we identify the cost of an activity with its causal upshots, it may seem that
the principle of fairness would require that we annul any losses our activities cause.
That in turn suggests a system of generalized strict liability in torts, which would
make injurers liable for any and all injuries they cause. Each of us should bear the
costs our conduct imposes on others; otherwise, we force others to bear costs which
are properly ours –
the costs resulting from the expression of our agency through
our decisions and choices. Put differently, strict liability looks like it is simply the
idea that one should internalize all the externalities of one’s conduct. If strict liabil-
ity is necessary to prevent individuals from displacing costs onto others, then fault
1995]
J. COLEMAN & A. RIPSTEN – MISCHIEF AND MISFORTUNE
liability would allow individuals to displace some of their costs, presumably by im-
posing those costs an injurer faultlessly causes his victims. Faultlessly caused costs
are, thus, displaced from injurer to victim.
In the same way, and for much the same reasons, some defenders of a tort sys-
tem based exclusively on strict liability have maintained that the same principle,
where one should bear the costs of one’s activities, also lends support to non-
redistributive economic institutions. Individuals have no duty to share in the misfor-
tune of others unless they are causally responsible for having brought it about. Un-
fortunate individuals who manage to shift the burdens of misfortune to others who
did not bring them about would otherwise impose external costs on them, no less
than injurers do. In the same way that causal responsibility is enough to justify im-
posing the costs of misfortune on an injurer in tort, so the absence of causal re-
sponsibility is enough to block imposing the burdens of misfortune on others more
generally. Because redistributive policies aim to have individuals shoulder the costs
of misfortunes they did not cause, such policies are morally impermissible. For the
libertarian, the principle of fairness requires a system of strict liability and the ab-
sence of coercive redistribution; strict liability for accidents and leaving life’s
standing misfortunes where they fall.
2. The libertarian is not alone in attempting to claim the principle that people
should bear the costs of their choices. Where the libertarian emphasizes personal
responsibility for misfortune, the liberal egalitarian emphasizes a communal re-
sponsibility to remedy the misfortunes of birth and status. According to the liberal
egalitarian, we should all be made to share in all of life’s misfortunes except for
those we make for ourselves by our own choices. Only the misfortunes our choices
bring to us are ours to bear alone. Where the libertarian emphasizes the misfortunes
we cause others through our choices, the egalitarian liberal emphasizes those we
create for ourselves though our choices, preference, and tastes.
Libertarians like Richard Epstein oppose redistribution,’ whereas such liberal
egalitarians as Ronald Dworkin’ and John Roemer’ favor substantial redistribution.
Yet both positions draw support from the same general principle of fairness. Even
more importantly, in spite of very different views about what the principle requires,
they share a conception of the costs of activities. In both cases, the underlying
thought is that responsibility is fixed by such concepts as cause, agency, and espe-
cially, control. For the libertarian, the costs of my activities include those costs that
my activities cause others; for the liberal egalitarian, the costs of my activities are
the results of the choices I make regarding my fair share of resources. If I choose to
3 See R. Epstein, “Luck” (1988) 6 Soc. Phil. & Pol’y 17.
‘ See R. Dworkin, “What is Equality? Part 2: Equality of Resources” (1981) 10 Phil. & Pub. Aff.
283 [hereinafter “Equality of Resources”].
5 See J.E. Roemer, “A Pragmatic Theory of Responsibility for the Egalitarian Planner” (1993) 22
Phil. & Pub. Aff. 146. See also G.A. Cohen, “On the Currency of Egalitarian Justice” (1989) 99 Eth-
ics 906.
MCGILL LAW JOURNAL/REVUEDE DROITDE MCGILL
[Vol. 41
cultivate a taste for philosophy, the costs of so doing are mine alone to bear. No one
else has a responsibility to make good my loss.
3. How can libertarians and liberal egalitarians accept the same general principle
of fairness, interpret it in light of the same general conception of an activity’s costs
yet support radically different views about what that principle requires in the way
of political institutions? The difference between the libertarian and the liberal
egalitarian depends on the fact that they adopt different default rules. The libertar-
ian claims that in the absence of causal responsibility, the costs of misfortune
should lie where they fall – on victims. The liberal egalitarian holds the opposite
default rule: misfortunes that are nobody’s causal responsibility are to be held in
common. Like the libertarian and the liberal egalitarian, we accept the principle of
fairness in the allocation of the costs of misfortune. We reject, however, their
shared conception of an activity’s costs. Any theory that allows individuals to fix
the costs of their own activities involves a subjective measure of those costs. The
costs of my activities are fixed entirely by some fact about me: my control over
them, my causal responsibility for them, my intentions, the risks I mean to take or
impose. On any such account, those features would matter in the same way, regard-
less of the activity in which I was engaged. Any such conception of the costs of an
activity not only fails to honour the principle that one should bear the costs of one’s
activities but has the rather perverse consequence of allowing individuals to de-
termine what those costs are.
In contrast, we will argue that only an objective measure of costs can honour
the liberal principle of fairness. In explicating this idea, we defend the view that
any measure of the costs of an activity is normative and depends on an analytically
prior account of what each of us owes one another. Only when we know what each
of us owes the other can we determine who owns the costs of misfortunes that arise
in the course of our interactions. Whether the loss you suffer when my business
succeeds in taking clients away from you is yours or mine does not depend on
whether my business activity is the cause of your loss; instead, it depends on
whether I owe you a duty not to interfere with your business interests, whether
constraints of fairness limit competition. It depends, in other words, on what we
owe one another with respect to competitive business practices, and whether in
competing with you I violate or comply with those obligations.
We will argue that the only way to honour the principle of fairness is by em-
ploying a normative conception of what we owe each other. To the extent that such
an account is supposed to honour the principle of fairness, those duties must be
specified objectively. We will argue that an objective measure of the costs of an ac-
tivity needs more than activity-neutral concepts like choice, cause, and agency. In-
stead, any plausible account will depend on the value of the activities in which we
are engaged and the ways in which those activities figure in our lives –
how they
matter to us and why. Having shown this, we will demonstrate why the search for
an appropriate default rule for dealing with all misfortunes is misconceived.
1995]
J. COLEMAN & A. RIPsTEN – MiSCHIEF AND MISFORTUNE
A. An Objective Standard of Costs
4. At this point, we offer no general theory of the duties we owe one another. In-
deed, we may not even agree with each other about the specifics of the theory that
our argument claims to be necessary. We share, however, the view that the principle
of fairness requires an objective measure of costs and an account, therefore, of our
responsibilities to one another. We also share a commitment to the central idea of
this essay, namely, that there is a right way of thinking about how to formulate and
construct the general theory of obligations, which is implicit in the idea of bearing
the costs of one’s choices. In tort law, it is represented by a particular conception of
liability. That conception is more nearly explicit in the fault standard than in the
strict liability rules that govern particular areas of tort liability but is at work in
both. In each case, liberty and security interests, which everyone can be supposed
to share, determine the occasions for liability. In institutions of distributive justice,
it is represented by the strategy of what John Rawls calls “primary goods”.’ Rawls
describes primary goods as all-purpose means –
things that anyone can be pre-
sumed to want in pursuit of whatever conception of the good they might have. Both
rely on an interpersonal measure of value, of what is important in an individual’s
life and why. These concerns are often controversial. But controversy is unavoid-
able.
Our motives for defending these ways of looking at both liability and distribu-
tive justice and for emphasizing the connection between the two are not narrowly
analytical. One great attraction of the family of views we reject is that it seems to
give expression to an ideal of neutrality. Each person is free to cultivate her various
tastes (liberal egalitarian) or engage in whatever non-criminal activities she wishes
(libertarian). The state maintains its neutrality with respect to those differing con-
ceptions of the good and, in this way at least, purports to respect each individual’s
moral independence. All the state is called on to do is to protect fairness by ensur-
ing that each person bear the costs of whichever activities he or she happens to
choose.
Our view, in contrast, is that any measure of the costs of an activity presup-
poses substantive and contestable conceptions of what is necessary or important to
living a life in a liberal political culture. As a result, it is impossible for liberalism,
or any other regime insisting on fairness as its central virtue, to be neutral in this
way. We cannot be neutral with respect to all subjective accounts of what the ele-
ments of a good or meaningful life are. If we are to provide a defensible distinction
between the misfortunes that are mine to bear and those that fall on others or should
be held in common, while honouring the principle of equality, then we must appeal
to objective measures of the value of various activities and their importance in our
lives –
how they matter and why – not by retreat to the seemingly neutral high
ground of the concepts of choice, control, or human agehicy.
6 J. Rawls, A Theory of Justice (Cambridge, Mass.: Belknap Press, 1971) at 90.
MCGILL LAW JOURNAL/REVUE DE DROiTDE MCGILL
[Vol. 41
If a standard is objective, it has to take a stand on which subjectively-felt harms
count and, therefore, cannot be entirely neutral about what matters in life. Such
non-neutrality does not require that it take a stand on every disputed question. In-
deed, a plausible liberal view of what, matters in life maintains that part of a valu-
able life is being able to make one’s own decisions about certain important ques-
tions. More generally, a liberal commitment to toleration is honoured by describing
those interests in general terms, so that they are protected however people decide to
exercise them. In addition, of course, non-neutrality need not translate into a policy
of actively promoting particular activities. By letting the importance of activities
figure into determining the costs of choices, various goods can be recognized with-
out being openly promoted.
The questions that we claim are required to allocate misfortune fairly are sub-
stantive questions that many would like to avoid. Many, doubtless, wish to avoid
the further political issue of who gets to answer such questions. Indeed, one of the
great attractions to many of libertarianism and the liberal egalitarianism of Dworkin
and Roemer is that they seem to substitute personal choice and control for public
judgement. But, difficult questions cannot be avoided if we want to allocate misfor-
tunes while honouring the idea that people should bear the costs of their choices.
The questions raised by that appealing, if elusive, idea are normative all the way
down.
H. Misfortune
5. We will explore these issues in legal and political philosophy through the lens
of misfortune. We do not offer an analysis or general unifying conception of mis-
fortune but simply look at particular misfortunes. Some misfortunes are thought of
as injustices, in need of some form of repair; others are thought of as “mere” mis-
fortunes. A variety of political institutions and strategies deal with misfortunes, in-
cluding regulation, tort liability, and criminal sanction. Some try to prevent misfor-
tunes; others simply attempt to shift their costs from one person to another. By re-
flecting on our practices for dealing with misfortunes, we can see what ideals and
commitments they express. Perhaps, we can also gain sufficient critical distance
from those practices to determine whether they are reflectively acceptable.
While we can always ask whose bad luck any particular misfortune is, analyti-
cally, only three answers seem to be available: the misfortune either belongs to the
person it befell, to some other particular person(s),’ or to the community as a whole.
Tort law, traditionally, concerns itself only with shifting losses from one person to
another, while distributive justice typically looks to questions of which misfortunes
7 Losses may appropriately lie with some group smaller than the entire community. For analytical
purposes (depending on the details) we can think of such situations either in terms of losses lying with
several individuals or with some community appropriate to the loss.
1995]
J. COLEMAN & A. RIPSTEIN – MISCHIEF AND MISFORTUNE
are to be held in common. Thus, we might say that tort law is concerned primarily
with the “event” misfortunes of human agency, whereas the institutions of distribu-
tive justice focus primarily on the “standing” misfortunes of birth and place.
A. Misfortune and the Hobbesian World of Perfect Agency
6. It may be helpful to begin by identifying two extreme positions one might
adopt in dealing with misfortune. At one extreme lies a Hobbesian state of nature,’
in which all misfortunes are simply allowed to lie where they fall. Fortune is al-
lowed full sway at every level. Holmes recommends this as a default position, re-
marking that the cumbersome machinery of the law should only be used to move a
loss from one person to another when there is a compelling reason to do so.9
In the Hobbesian state of nature, there is no state, law, or other cumbersome
machinery. As a result, there is no agency to shift losses. Of course, people might
displace various misfortunes from themselves onto others – displacing their hun-
ger by taking someone else’s food or injuring another while attending to their own
wants. If they succeed in any such endeavours, the misfortunes now lie where they
most recently fell. Even deliberate shifting can be thought of as the full reign of
fortune, because whether someone is in a position to shift their losses to another is
at least from the perspective of the new owner of the
itself a matter of fortune –
misfortune. They just happened to be in the wrong place at the wrong time.
Hobbes himself describes the state of nature as a sort of cautionary tale, an ex-
planation of why there must be limits on the misfortunes people are allowed to
cause each other. Nobody has seriously recommended, however, that the state of
nature is a normative ideal. It purports to be a world of pure agency with no room
for ideas of concern or respect for others, let alone for even the thinnest idea of
community or of a common fate.
B. Misfortune and the World of Perfect Community
7. At the other extreme lies the ideal of a perfect community in which all misfor-
tunes are held in common. Whatever goes wrong and for whatever reason, all share
in the costs. Hence the effects on particular persons of fortune are cancelled en-
tirely. Where the Hobbesian state of nature is a world of pure agency with no com-
munity, this world is one in which the search for perfect community abandons the
idea of agency altogether.
Here too it is difficult to think of anyone who fully endorses such a view. Per-
haps some readings of utilitarianism hold out an ideal of perfect community, insofar
as they suppose that gains are to be maximized and losses minimized, quite apart
‘See T. Hobbes, Leviathan (Oxford: Basil Blackwell, 1955) at 80.
9 See O.W. Holmes Jr., The Common Lmv (New York: Dover, 1991) at 96.
MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL
[Vol. 41
from who they might happen to belong to.'” The other candidates are probably some
of those theories that are considered and rejected by Robert Nozick under the name
“patterned” theories of justice. Nozick hopes to tar all egalitarian and redistributive
schemes by criticizing patterned distributions, but no theory anyone else has put
forward is fully patterned.” Again, no-fault compensation schemes spread misfor-
tunes on a limited scale but typically wait for accidents to happen before compen-
sating. They are limited to particular misfortunes and maintain a distinction be-
tween those misfortunes which suddenly interrupt life, and those accidents of birth
that form the background of some peoples’ lives. The Soviet jurist Evygeny Pashu-
kanis comes closest to actually endorsing such a view when he suggests that the
idea of individual responsibility is a relic of a deformed social system, which will
give way to a system in which problems are solved and losses made good as they
come up.2
C. Misfortune and the Principle of Fairness
8. Neither the world of perfect agency nor the world of perfect community would
be much of a place to live. For present purposes, the important point is that they
share a single failing; neither has room for the idea that people should bear the full
costs of their choices. Hobbes describes the state of nature in a way that makes the
lack of security its most salient and troubling feature. But life in such a state would
not only be solitary, poor, nasty, brutish, and short. It would also be terribly unfair.
If I injure you in the Hobbesian state of nature, it is your problem not mine.
Whether through malice or indifference, people could take advantage of the efforts
of others and displace the costs of their activities onto them. The world of perfect
community is unfair in similar ways. If I injure you or fritter away my share of re-
sources, I don’t lose my place at the common trough. Costs imposed by the lazy,
the vicious, and the self-indulgent must be spread no less than those that fall on the
disadvantaged, the needy, or the injured.
Neither the world of agency nor the world of community seems capable of liv-
ing up to its own ideals. The state of nature presents itself as a world of perfect
agency. Yet, it fails to protect agency because it does not leave only mere misfor-
tunes where they lie; it does the same for misfortunes that one person creates for
‘0 This is the feature of utilitarianism that leads Rawls to conclude that it “does not take seriously the
distinction between persons” (Rawls, supra note 6 at 27).
” See R. Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974) at 156. The position
that Dworkin considers, and rejects, under the name “equality of Welfare” might count as fully pat-
terned and as representing the idea that all misfortunes are held in common. Since welfare is equal-
ized under such a scheme, misfortune, as well as everything else, is held in common (see R. Dworkin,
“What is Equality? Part 1: Equality of Welfare” (1981) 10 Phil. & Pub. Aff. 185 [hereinafter
“Equality of Welfare”]). See also T. Scanlon, “Nozick on Rights, Liberty, and Property” in J. Paul,
ed., Reading Nozick: Essays on Anarchy, State, and Utopia (Towtowa, N.J.: Rowman & Littlefield,
1981) 107.
12 See E.B. Pashukanis, Lmv and Marxism: A General Theory, trans. B. Einhom (London: Pluto
Press, 1983) at 109-33.
19951
J. COLEMAN & A. RIPSTEIN- MISCHIEF AND MISFORTUNE
another, whether wilfully or carelessly. That is, which problems belong to which
persons bear an entirely contingent relation to anything they have ever done. Each
person’s agency is left entirely at the mercy of the agency of others. In short, it does
not adequately tie the ownership of misfortune to agency. The converse is true of
the world of perfect community. This time, the problem is that, rather than simply
rectifying the unfortunate effects of circumstances, it does the same for misfortunes
that people create for themselves, whether wilfully of carelessly. Agency only
makes sense if we can draw a line between it and mere fortune. Neither the world
of perfect agency nor of perfect community has the resources to draw such a dis-
tinction between what I do and what happens to me or to others. The two extremes
would use such a distinction very differently, of course, but both need it. Without it,
a way of understanding misfortunes becomes a way of understanding all of life.
Thus, we might say that both the Hobbesian world of perfect agency and the
world of perfect community violate fairness in the same way; they allow costs to be
displaced. The state of nature allows one to impose costs on others, which are the
result of one’s agency, and the state of perfect community allows one to impose
costs that are the result of one’s choices and indulgences. But this is too simple, be-
cause putting the problem in this way presupposes that we already have a concep-
tion of what the costs of a person’s activities are. We might put the point differently,
however, in a way that is more consistent with the theme of the argument. Instead
of saying that the state of nature and that of perfect community unfairly allow in-
dividuals to displace costs that are in fact theirs, we might say that both the state of
nature and perfect community are ways of specifying what the costs of various ac-
tivities are. The Hobbesian state of nature says, in effect, that the costs of misfor-
tune are costs of the victim’s activity, not the injurer’s –
no matter what activities
each is engaged in. In contrast, the state of perfect community says, in effect, that
the costs of misfortune are costs neither of the injurer’s nor of the victim’s activi-
ties; they are everyone’s to bear. Thus, rather than saying that the state of nature and
perfect community violate the principle that one ought to bear the costs of one’s
activities, we might say that they represent particularly unattractive conceptions of
what those costs are. Turning to our central claim that any measure of costs ulti-
mately depends on a theory of obligation – what each of us owes one another –
perhaps the most accurate characterization of the problem with both the state of
nature and perfect community is that they represent normatively indefensible con-
ceptions of individual and collective obligations.
m. Libertarianism: Causation as a Measure of Misfortune
9.
In spite of its emphasis on agency, the problem with the Hobbesian state of na-
ture lies in its inability to distinguish between misfortunes one person creates for
another, and those that are merely bad luck. The state of nature elides the distinc-
tion between mischief and misfortune. If this is the problem, the solution should be
apparent. Some line must be drawn between those misfortunes that lie where they
fall, and those that properly belong to some other person. The most obvious place
to look for that distinction, consistent with the emphasis on agency, is in a distinc-
MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL
[Vol. 41
tion between what a person does and what (merely) happens. That is, what is
needed is a line between the sufferings that one person causes another, and those
that are merely the result of chance. The former must be compensated, the latter left
where they have fallen. The resulting position is a regime of generalized strict li-
ability for accidents, which allows no further non-voluntary wealth transfers. Basi-
cally, this is the libertarian solution.
In this way, libertarianism is one step removed from the Hobbesian state of na-
ture. It seeks to capture the essential importance of agency in a way that is consis-
tent with the liberal picture of fairness, something the state of nature cannot do. In
the libertarian picture, the Hobbesian position serves as a default rule. Misfortunes
do not simply lie where they fall. Instead, they lie where they have fallen unless
they result from the causal agency of some other individual –
in which case they
are shifted to that person. So understood, the libertarian approach might be thought
of as more accurately and adequately representing the ideal of pure agency.
Libertarians tend to focus on the particular misfortunes that are the standard
province of tort law – manglings by defective bottles, chainsaws, drugs, and fork-
lifts. But, that focus is not (just) the result of a morbid fascination with manglings.
It is of a piece with their opposition to- redistribution. A night-watchman state
charged solely with protecting property rights, and a regime of generalized strict li-
ability in tort law come as a package.
The underlying view is this: what I do is mine, what is mine I own; what is not
mine, I do not own. The costs my conduct imposes (that is, causes) are the result of
what I do and thus mine; I own them. I have to take back the costs my conduct has
imposed on you, not so much to rectify your loss but to give me what is mine. In
contrast, if the costs that have befallen you are not the result of what I do, then they
are not mine to bear. That does not automatically mean they are yours to bear, instead;
they belong to whoever’s conduct caused them. If the misfortune that befalls you is
your doing, you own it. If it is no one’s doing, it belongs to no one; it is the result of
chance, not agency. At this juncture, the default rule comes in. If a loss is no one’s
property but just the result of chance, then it must lie where it has fallen. The default
rule is itself justified by a further appeal to agency: if, by chance, misfortune befalls
you, it is yours, not because you deserve it, but because nobody else does. Shifting a
loss to someone who did not cause it violates their agency. So it is yours just because
it happened to you, just as your own body is yours. Chance plays an important role in
both directions: whether or not my activity injures you may well depend on factors
outside anyone’s control, as will all of those natural facts not attributable to any hu-
man agency. For the libertarian, this way of dividing agency and chance is essential to
the idea that people should bear the costs of their activities.
Put slightly differently, the libertarian view is that only the consequences of
agency can be the objects of justice; mere misfortunes –
the upshots of chance –
are not the proper object of justice. If someone causes an injury to another, it is the
result of agency, not chance. Thus, justice requires that the loss be returned to its
1995]
J. COLEMAN & A. RIPSTEIN – MISCHIEF AND MISFORTUNE
proper owner. Sharing in the burdens of misfortune, in contrast, may be a matter of
charity; it is not a matter of justice. Doing so may be laudatory, but it is not obliga-
tory.
It is easy enough to see how the same line of argument might be deployed in
support of an anti-redistributive political theory. No human agent is typically re-
sponsible for such misfortunes of birth as poverty or disability. If these are genuine
misfortunes, that simply means that they are no one’s doing. If we think of the mis-
fortunes of birth as no one’s doing, then they have no owner. If no one owns them,
there are no grounds for coercively imposing them on anyone. Again, the default
rule comes into play, and the losses must lie where they fall – on those disadvan-
taged by birth and chance.”
Recall the basic, impetus for the libertarian position on misfortune. The problem
with the Hobbesian state of nature is that, while it properly emphasizes the role of
agency in determining the locus of misfortune, it lacks the resources to draw a
meaningful distinction between persons, between what someone does and what
merely happens to him or her. The libertarian provides a solution to this problem by
tying agency to causation and both to ownership. Agency provides a way to express
one’s personality by making a mark on the world. One presumably owns one’s own
body and whatever one does with it. If I injure myself, I own my own injury; if I
injure someone else, it is as though I injured myself. In acting I create things, and I
own my own creations.” In short, causation links agency with misfortune. If my
body figures in a valid causal explanation of what happened, then I am responsible.
A. The Inadequacy of Causation as a Measure of Agency
10. If causation is a good starting point, it does not get us far; for causation is eve-
rywhere. Ronald Coase and Guido Calabresi independently invented law and eco-
, Probably the best known and most articulate proponent of strict liability in torts is Richard Ep-
stein. Epstein has more recently argued that it is a mistake for the state to suppose that it is able or
competent to cancel the pervasive effects of luck by correcting the standing misfortunes of birth and
status (see “Luck”, supra note 3). To be sure, his more recent arguments are presented in utilitarian
terms and rest on empirical assertions about the incompetence of public institutions. Other libertarians
(including an earlier Epstein) have been less willing to embrace utilitarianism, and have defended the
same position in terms of everything from natural rights to social contracts. Our present concern is not
with the details of those deeper arguments but with the structure of the regime in support of which
they are offered (see R. Epstein, A Theory of Strict Liability: Toward a Reformulation of Tort Law
(San Francisco: Cato Institute, 1988) [hereinafter Strict Liability]).
“This seems to be the reasoning underlying Epstein’s discussion of Vincent v. Lake Erie Transpor-
tation, 124 N.W. 221, 109 Minn. 456 (1910) [hereinafter Vincent]. In Vincent, a ship owner was held
liable for damages to a dock to which its boat had been tied in a storm, despite the court’s concession
that it was reasonable for the shipowner to stay at the dock despite the dockowner’s protests. Epstein
suggests that liability can be explained by recognizing that if both ship and dock had been owned by
the shipowner, the shipowner would have decided whether to risk the ship or the dock, and been li-
able for the entire loss whichever decision had been reached. Because the shipowner acted, he is li-
able for losses, wherever they fall (see Strict Liability, ibid).
MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL
[Vol. 41
nomics when they realized that both injurer and victim cause any injury.” The point
is, of course, older than that; in the seventeenth century, Pascal made essentially the
same point: “I have often commented that the sole cause of man’s unhappiness is
that he does not know how to sit quietly in his room.”‘” The injurer’s role is obvi-
ous, for the reasons that libertarian advocates of generalized strict liability focus on.
But, the victim’s role is no less real. In all but the most bizarre cases,” the accident
could have been prevented had the victim stayed home, taken a different route, or
whatever. Thus, any injury is always a joint product, which must somehow be di-
vided between the parties. If causation is made the basis for shifting losses, two
questions can be asked in every case of injury. First, was the defendant the cause of
the injury? Second, was the plaintiff? The trouble is that the answer to both ques-
tions seems to always be “yes”. If we understand the causal question in terms of
what would have happened but for the act of the defendant (plaintiff), we are likely
to get the same result for every accident: both the injurer and the plaintiff cause the
injury. Thus, we seem to find everyone liable for everything.
A tempting response to the problem of too many causes is to draw a distinction
between causes and conditions. Thus, the libertarian might hope to claim that the
injurer’s act is the cause of the injury, while the victim’s is merely a condition. Both
are necessary conditions, and so but-for causes. Only one is the cause, because only
one seems to have made a crucial difference. In ordinary parlance, we have no dif-
ficulty distinguishing between a match as a cause of a fire and the presence of oxy-
gen as a mere condition. In the same way, the libertarian might hope to keep strict
liability general by assigning losses to those who cause them, rather than to the by-
standers who are the mere conditions of them.
The distinction between causes and conditions can be of no real help to the de-
fender of generalized strict liability. Any account of causation that will be adequate
to account for the distinction between causes and mere conditions in scientific and
common-sense explanations will, in virtue of that very adequacy, fail to solve the
libertarian’s problem with causation. The legal account of causation requires
uniqueness, while ordinary causal explanations do not. Any account of causation in
non-legal contexts needs to accommodate the fact that, in ordinary discourse, the
distinction between causes and conditions is contextual and malleable. For exam-
ple, in some circumstances we might say that a match was caused to ignite by being
struck, taking the presence of oxygen to be merely a background condition. In
other, less familiar, circumstances we might say that the presence of oxygen was
the cause. Therefore, we can give different answers to causal questions about the
” See: R.H. Coase, “The Problem of Social Cost” (1960) 3 J. L. & Econ. 1; G. Calabresi, “Some
Thoughts on Risk Distribution and the Law of Torts” (1961) 70 Yale L.J. 499.
‘6 B. Pascal, Pensies, trans. A.J. Krailsheimer (London: Penguin, 1976) at 67.
” See e.g. Tomaki Juda v. United States, 6 Cl. Ct. 441 (1984) in which the plaintiff was forcibly re-
moved from his home in bikini atoll to allow nuclear weapons testing and then returned when radia-
tion levels were dangerously high. In that case, presumably nothing the plaintiff could have done
would have prevented the injury.
1995]
J. COLEMAN & A. RIPSTEIN – MISCHIEF AND MISFORTUNE
same event depending on our interests. The malleability of the distinction reflects
the deeper fact that the two causal explanations are not incompatible. The person
who claims that striking caused the match to ignite, and the one who claims it was
the presence of oxygen do not disagree about causation. At most, they disagree
about what is interesting about the situation both seek to make sense of, and neither
the very fact that en-
would declare the other’s purported explanation irrelevant –
ables Pascal and Coase to claim that causation is reciprocal.
Some accounts of causation have sought to narrow its scope by reserving the
name “cause” for those conditions which are not normal or typical.” Putting aside
the difficulties any such account has in making sense of such banal claims as
“evaporation causes cooling” or “eating causes weight gain”, it is of no use to the
libertarian. What counts as unusual will often depend on the circumstances and the
purposes of the inquiry. The unusual event may change depending on whether our
interests in the causal inquiry are explanatory, predictive, or engineering. If we
want to control or engineer the future, we might look to conditions that would be of
less interest if we sought, instead, to understand the past in a detailed way. Until we
know what we are doing, we cannot identify causes.
The libertarian had hoped to embrace a general regime of strict liability as an
interpretation of the idea that people should bear the costs of their activities. But,
any attempt to retrieve that idea by narrowing the range of causation faces a di-
lemma. If we concede that the explanatory interest we are pursuing in identifying
causes is tied to our interest in assigning liability, we must first interpret the idea
that people should bear the costs of their choices in order to know our explanatory
purposes in establishing liability. Of course, if we need an interpretation of the
principle in order to distinguish causes from conditions, causation cannot itself
provide the basis for an interpretation. Thus, causation provides no leverage in dis-
tinguishing plaintiff from defendant, unless we already have some other way to
distinguish them.’9 Alternatively, the libertarian might insist that liability be as-
signed on the basis of what would seem to be the most natural causal explanation
,” See J.L. Mackie, The Cement of the Universe: A Study of Causation (Oxford: Clarendon Press,
1973).
“‘ More embarrassing still, if we tie liability to causation and causation to what is atypical, the
boundaries of a person’s activity turn out to depend largely on how many other people engage in it.
When white students at the University of Georgia rioted to protest integration, a number of them
screamed insults at the lone black student in a dormitory: “[D]oes she realize she’s causing all this
trouble?” (C. Trillin, “An Education in Georgia!’ The New Yorker (16 July 1963) 30 at 67, quoted in J.
Feinberg, “Action and Responsibility” in A.R. White, ed., The Philosophy of Action (London: Oxford
University Press, 1968) 95 at 115-16). While many libertarians oppose anti-discrimination laws, pre-
sumably few would want to rest any judgements of liability on the claim that the black student caused
the riot.
More generally, tying causation, and thus, liability to statistical infrequency not only does badly
as an interpretation of the idea that people should bear the costs of their choices, it also betrays an-
other idea that libertarians claim to embrace, namely, that of attaching priority to individual liberty to
do novel and interesting things.
MCGILL LAW JOURNAL/REVUE DE DROITDEMCGILL
[Vol. 41
quite apart from concerns about liability. This time, the problem is slightly differ-
ent. If the libertarian appeals to some merely explanatory, predictive, or engineering
interests, then all connection between agency and liability is lost. The problem is
not that we cannot distinguish among causes; it is that we cannot do so in a way
consistent with the idea of agency and the simple idea underlying libertarianism,
namely, that there is a conceptual connection between what I do and the costs of my
activities. Instead, agency must give way to predictability or, worse yet from the
libertarian’s point of view, (social) engineering.
B. The Inadequacy of Strict Liability Based on Causal Paradigms
11. Epstein, for one, is aware that there are problems in distinguishing various
causes and suggests that we must narrow our notion of causation to include some-
thing less than all cases of “but-for” causation. Instead, Epstein insists, we can look
to the clear paradigms of causation which the law recognizes. He focuses on force,
fright, compulsion, and the setting of dangerous conditions as the paradigms of
causal connection.” If it is true that not every causal connection can suffice to
ground ownership or liability, the question is, why pick these as our paradigms?
Why not use Bob Dylan’s more comprehensive list? I might compete with you, beat
you, cheat you, or mistreat you, simplify you, classify, deny, defy or crucify you,
select you, dissect you, inspect you, or reject you. All of these things I might do
are ways in which I could, and often do, cause harm to you. Some of these are even
plausible grounds of liability, others not. Without an account of the obviousness of
Epstein’s paradigms, we are in danger of losing sight of what makes them obvious.
Again, the point is not that causation’s reach cannot be appropriately narrowed,
but that attempts like Epstein’s turn out to be unprincipled: there is no rationale for
them that does not beg important questions. Although the law may have no choice
but to work outward from paradigm cases, Epstein’s paradigms seem an ad hoc
collection designed to get his preferred result. Why are not other omissions, besides
the setting of dangerous conditions, the paradigmatic cases of causes that create a
basis for liability? Again, why not focus on other, equally paradigmatic actions: I
might embarrass you into spilling you coffee, goad you into hitting someone else,
or humiliate you. In each of these cases, we need far more detail in order to assess
whether your injury would be compensable. With each of these examples, certain
questions cry out for answers in establishing liability: Did you deserve humiliation?
Are you too easily embarrassed or frightened? Should you have stood up in the face
of my goading? The same applies to any number of the other examples of action
that might be thought to be paradigmatic. We seem to need a conception of the
boundaries between persons before we can determine ownership. But if we do, the
appeal of strict liability as a way of drawing a line between what I do and what
happens evaporates. For it is no longer the fact that I caused you harm that makes it
20 See Strict Liability, supra note 13 at 166-80.
2 See B. Dylan, “All I Really Want to Do” on B. Dylan’s Greatest Hits, vol. 2 (New York: Colum-
bia Records, 1970).
1995]
J. COLEMAN & A. RIPSTEIN- MISCHIEF AND MISFORTUNE
mine but something else. It seems liability can only be assessed once we have an-
swered questions about things other than causation.
Worse yet, Epstein’s paradigms look suspiciously like they avoid the difficult
questions by presupposing answers to them. We do not need to ask whether I
frightened you carefully, or if you yielded to my force too readily. That is not be-
cause these idioms are particularly causal, but because it is so difficult to figure out
what would count as exercising care in forcing someone to do something or fright-
ening them. In the same way, any distinction between the setting of dangerous
conditions and the mere setting of conditions, seems to turn on what those condi-
tions are, rather than on any causal notions. The line between dangerous and safe
conditions cannot depend on whether the defendant could have made them safer, on
pain of falling back into Pascal’s world. The defendant always could have done
something, and the plaintiff virtually always could have avoided them; but without
an account of dangerous conditions, the paradigm of force includes too much.22 The
line can only be drawn in terms of some idea of how dangerous conditions are al-
lowed to get.
Our claim is not that there are no principled ways of drawing distinctions be-
tween but-for causes. Our point is that any principled distinctions are also norma-
tive. What counts as a cause will depend on what turns on the answer.
We are not the first to point to the pervasive problems of indeterminacy that
plague Epstein’s defense of strict liability. Indeed, no one has been as powerful and
sustained a critic of it as has Stephen Perry.’ One might wonder why its appeal
persists despite what strike many as obvious problems. One answer is that the liber-
tarian’s root concern is with the idea of self-ownership. Since (it is supposed) we
own ourselves, any harm that one person does another must be thought of on the
model of a person (or piece of property) straying onto another person’s property. If
I trespass on your land, it does not matter whether or not I was being careful –
I
still need to leave. The libertarian wants to generalize this idea to cover all of tort
law –
I am liable if I harm you, regardless of how careful I was being. If I cross
your boundaries, I am liable, regardless. Hence the appeal of strict liability; but this
is also its fatal flaw. Even if we stick to metaphors of borders and boundary cross-
ing, we need some account of where the boundaries between persons are to be
drawn. Causation cannot do the job. It merely presupposes that the job has already
been done.
One other way in which the reach of this interpretation of strict liability might
be narrowed, while keeping it general, is by holding people responsible not for all
they have caused but only for the intended consequences of their actions. This ap-
proach looks promising, since those who intend to frighten, force, compel, or set
dangerous conditions plainly differ from those who are frightened, forced, com-
‘ See S.R. Perry, “The Impossibility of General Strict Liability” (1988) 1 Can. J. Law & Jur. 147.
‘See Perry, ibid
MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL
[Vol. 41
pelled, or unwittingly walk into dangerous conditions. It also has a certain natural
attractiveness, as evidenced by the way children often defend themselves against
charges of carelessness by saying, “But I didn’t mean to.” It is difficult to think of
anyone who has actually put this view forward, though one might expect it of
someone who came to tort law after thinking about the criminal law.
The naturalness of such a suggestion aside, its application would be counter-
intuitive at best: if I intend to hit you on the head but instead hit you on the shoul-
der, surely I should not be entirely relieved of liability, or your injury treated no dif-
ferently than it would be were you hit by a bolt of lightning.
The more serious problem with the suggestion is deeper: limiting liability to
intended consequences undermines the idea that it is supposed to express, namely,
that people should bear the costs of their choices. For it effectively says that each
person gets to decide what will and will not count as a cost of her activity. I deter-
mine the intended consequences of my conduct: after all, they are the consequences
I intend. I can be liable only for them. Therefore, I determine the costs of my activ-
ity, so long as I do not want to harm you. Rather than being justified on the grounds
that it expresses the idea that one should bear the costs of one’s activities, this sug-
gestion is, instead, a way of determining what those costs are; and it is a norma-
tively unattractive way at that. For the intuitive idea is surely that people should
bear the costs that their activities impose on others; it is not the idea that they are
free to determine what those costs are. There can be no measurement of such costs
unless we have some independent criteria for measuring them.
IV. Fault as a Measure of Misfortune
12. How can the costs of activities be measured? The most plausible illustration of
the answer can be found in the fault system. We offer the fault system for purposes
of illustration because its analytical structure is clear. However, our discussion of it
is not meant to suggest that there is no room in a legal system for particular areas of
strict liability. As we shall show, where strict liability is defensible, it is best under-
stood as employing the same type of reasoning that the fault system exemplifies. In
defensible areas of strict liability, the boundaries between what a person does and
what merely happens are given by moral boundaries and protected interests.
Instead of supposing that the duty to repair is a matter of causal relations or
crossing prior natural boundaries, the fault system recognizes that the boundary
between persons can only be understood in terms of our duties to each other. Thus,
its inquiries are normative, not scientific or even folk-scientific. It narrows the
scope of liability for harms caused to others by imposing a requirement of reason-
able care. So long as I am careful, I am not liable for harms that I cause you. Those
misfortunes are yours, as though they had been caused by some natural event. If I
am careless, then I am liable for whatever harms I cause you –
those misfortunes
are mine, costs of my activity which I must bear. The whole point of the fault sys-
1995]
J. COLEMAN & A. RIPSTEIN – MISCHIEF AND MISFORTUNE
tem is that when I am careless I may own more than the intended consequences of
my actions; hence, the duty to repair.
Drawing the boundary between us in terms of the fault standard solves the
problem of determining what counts as the cost of my activity. The fault system ac-
knowledges the importance of concepts like control, agency, and choice. But, those
concepts only work in concert with substantive judgements about why various ac-
tivities matter to us and about the ways in which they do. That is, the fault system
requires a substantive political theory. Indeed, because there is no natural feature to
mark such boundaries, the fault system cannot help but make judgements about the
importance of various activities.
In a liberal regime, the fault system aims to treat parties as equals, protecting
each equally from the other and granting each a like liberty. In order to protect both
liberty and equality, it must suppose plaintiff and defendant to have both liberty in-
terests in going about their affairs and security interests in avoiding losses due to
the activities of others. In so supposing, it does not look to how strongly either
plaintiff or defendant feels about either liberty or security. For to do so would re-
produce the problems with limiting liability to the intended consequences of ac-
tions, making each person’s security depend on the particular priorities of his or her
injurer. Those difficulties would be doubled because each person’s liberty would
depend on the particular sensibilities of his or her potential victims. It also does not
(because it cannot) recognize either a generic interest in either doing what you
want, or in being secure against all injuries. Instead, the amount of risk you can
faultlessly expose me to depends on what you are doing and on the kind of harm
that might eventuate.
A. Boundary Crossing
13. The fault standard invokes the idea that you and I have duties to one another.
We can, if we like, put this in the libertarian’s vocabulary of boundary crossings as
the claim that each has a duty to repair should they cross the other’s boundaries.
Two caveats are required. First, such a duty presupposes independently established
boundaries, whereas the whole point of fault is to help fix those borders. The
boundaries between us are normative, fixed not by space but by the scope of the
duties we owe one another. Second, I might violate my duty and thereby “cross”
my own boundary, within which I am free of liability for harms I might cause –
for
example, by failing to exercise reasonable care – yet fail, through the good fortune
of us both, to cross your boundary, if my lack of care causes no harm to you. That
is, there is a moral space between our boundaries where fortune lurks. Crossing my
own boundary does not entail that I thereby cross yours.
Because my crossing your border requires not only that I fail to meet my duty
to you, but that in doing so I harm you as well, invoking the importance of fault
does not require giving up on agency or causation. On the contrary, the fault system
lets us see why agency matters at all. Liability depends on the fact that I am an
McGILL LAW JOURNAL/REVUE DE DROITDE MCGILL
[Vol. 41
agent, capable of acting in keeping with my duties. Were I not an agent, I could
neither discharge nor fail to discharge any duty to you. The fault standard also ex-
plains why causation matters. Causation turns my failure to discharge my duty to
you into a crossing of your boundary. Without fault, causation does not signify a
boundary crossing; without causation, crossing my boundary does not mean I have
crossed yours.
Third, and most importantly, while the concept of fault is understood in terms
of reasonable care, the concept of reasonable care is to be understood as taking into
account not just the relevant liberty and security interests of the parties but the rela-
tive value of the activities in which they are engaged. The measure cannot just be
their evaluation of those activities but must, in some sense, be the value those ac-
tivities actually have. This last point requires further elaboration.
B. Establishing the Boundary Line: Balancing Liberty and Security Interests
strict liability, in other words –
14. To see the way the importance of various activities enter into the fault system,
think of the fault standard as dividing the risks of injury between plaintiff and de-
fendant.” When an accident occurs, the fault system assigns the misfortune to the
defendant if negligent, otherwise to the plaintiff. If all of the risk were to lie with
the defendant –
the defendant would always act at
his peril. If all the risk were to lie with the plaintiff – holding the defendant liable
only for intended consequences –
the plaintiff’s security would be hostage to the
defendant’s action. Since neither extreme treats liberty or security interests equally
or fairly, neither alternative is acceptable as a way of giving expression to the lib-
eral idea of fairness or equality. The fault system solves this problem by supposing
that the dividing line must be somewhere in between. Establishing the line depends
on the particular liberty interests and security interests that are at stake.
For example, the liberty to drive without brakes is not sufficiently important to
allow others to be exposed to injury from a car that is unable to stop in a timely
fashion. In contrast, the person with an eggshell skull’s interest in being able to go
out is too important to make them bear the full risk. Yet, other examples, empiri-
cally like the eggshell skull, get treated differently. A rule of contributory or com-
parative negligence limits the plaintiff’s recovery if his or her activity made the in-
jury more likely or more costly, despite the fact that there would have been no in-
jury at all had the defendant exercised appropriate care. By ignoring Pascal’s advice
and going out, the eggshell skull makes the activities of negligent people more
costly, just as the person who walks on railway tracks does. Yet the defendant is li-
able for the full extent of the eggshell’s injuries but not the trackwalker’s. Although
each would be fine if nobody else was careless, they get treated very differently. In
the trackwalker’s case, a rule of contributory or comparative negligence is applied
which limits the plaintiff’s recovery if his or her activity made the injury more
likely or more costly, despite the fact that there would have been no injury at all had
24The risks being divided are always risks of particular injuries (see Part IV.!)., below).
1995]
J. COLEMAN & A. RIPSTEIN- MISCHIEF AND MISFORTUNE
the defendant exercised appropriate care. These are standard examples, so familiar
that their presuppositions are hidden. They depend on substantive, though familiar,
judgements about the importance of various liberty and security interests. They also
provide models through which other activities get classified.
Does carrying out one’s perceived religious duties get treated as the equivalent
of having an eggshell skull or as the equivalent of walking on railroad tracks?’ The
only way to answer such a question is to defend a substantive, if controversial, con-
ception of the role of religion in our society. A general rule of contributory negli-
gence requires plaintiffs to mitigate damages. Yet if someone fails to mitigate on
recognized religious grounds, the defendant may well be liable for the full amount.’
This is a reflection of the idea that the reasonable person would give more weight
to certain of her religious convictions than to saving a negligent defendant money.
Yet if the plaintiff fails to mitigate for other reasons, however deeply felt, damages
are reduced accordingly. In other words, it is not how deeply religious convictions
are felt that matters, for that would not distinguish religious conviction from other
beliefs as a defense against failure to mitigate. Rather, what matters about religious
beliefs is connected with a view of the ways religion matters to people.” Not all
sincerely felt religious beliefs can be accommodated within such a model; tort rules
that seek to accommodate religion in this way cannot pretend to be neutral.
Again, rescuers are allowed to expose themselves to risks that would normally
preclude recovery from their injurers. Although the special treatment accorded res-
cuers is sometimes described in terms of foreseeability,2″ a more honest way of
looking at the special treatment accorded rescuers supposes their activity to be im-
portant enough to make them bear proportionately less risk. 9
The claim here is not that these cases invariably strike the balance correctly.
Rather, it is that implicit in the fault system is the need to come to an understanding
of the substantive value of various activities, their importance to us in a liberal so-
” See: Meistrich v. Casino Arena Attractions Inc., 155 A.2d 90, 31 N.J. 29 (1959); Lange v. Hoyt,
159 A. 575, 82 A.L.R. 486 (Conn. 1932).
26 See G. Calabresi, Ideals, Beliefs, Attitudes, and the Law: Private Law Perspectives on a Public
Law Problem (Syracuse: Syracuse University Press, 1985).
7 Such an account would presumably offer some explanation of why religious beliefs are typically
deeply felt. But, it is not the depth of the feeling that counts.
2 As Justice Cardozo put it: “Danger invites rescue” (Wagner v. International Railway Co., 133
N.E. 437,232 N.Y. 176 (Ct. App. 1923).
29 For example:
When one risks his life, or places himself in a position of great danger, in an effort to
save the life of another, or to protect another, who is exposed to a sudden peril, or in
danger of great bodily harm, it is held that such exposure and risk is not negligent. The
law has so high a regard for human life that it will not impute negligence to an effort to
preserve it, unless made under circumstances as to constitute rashness in the judgment
of prudent persons (Peyton v. Texas & P Railway Co., 6 So. 690 at 691, 41 La.Ann.
861 (1889)).
MCGLL LAWJOURNAL/REVUEDE DROITDE MCGILL
[Vol. 41
ciety, and the ways they matter to us individually and collectively. Of course, nei-
ther judges nor juries normally think of themselves as engaged in an inquiry into
the importance of various activities. Instead, they ask what a reasonable person
would do in such circumstances –
a person who had enough foresight to see what
was likely to happen and appropriate regard for the interests of others. But, the
construct of the reasonable person incorporates answers to questions about which
things matter and how they matter – we might even say that the reasonable person
embodies them. Judges sometimes also appeal to what is customary or conven-
tional. This, too, reflects implicit judgements about the importance of various ac-
tivities. Provided that background conditions are otherwise fair, they enable people
to take the costs of their activities into account by making those costs visible. In
circumstances in which almost everyone engages in the same activities, and all are
equally likely to be both plaintiff and defendant, a customary-activity rule may be a
reasonable proxy for a substantive examination of the interests that are at stake.
That is simply because in such circumstances (if they ever existed), everyone is en-
gaged in the same activities and so has the same liberty and security interests in re-
lation to them?
C. An Objective Measure of Fault
15. The same considerations that support this interpretation of the fault system and
argue against the intended consequences thesis explain why the fault required for
liability in torts must be objective. Because fault is supposed to measure the costs of
activities fairly and across individuals, it cannot be understood subjectively in terms
of good faith efforts at care. To do so would create, yet again, the problem that
faced strict liability for the intended consequences of actions: it would make each
injurer the determiner of his or her exposure to liability. Thus, what counted as the
costs of an activity would vary wildly. As a result, the scope of each person’s right
to security would depend on the decisions of others. The idea that people should
bear the costs of their activities would be replaced with the very different and less
appealing idea that people should do the best they can –
at whatever they choose to
do.
We have defended the fault system as a way of establishing the boundary be-
tween those misfortunes that are the injurer’s to bear, and those that are to remain
the misfortune of the victim. Such a boundary is necessary if we are to understand
and give content to the principle that one ought to bear the costs of one’s activities.
The fault system accomplishes this, moreover, only by determining the relative im-
portance of various activities, the ways they matter in peoples’s lives, and why.
3 Both liberty and security interests are typically described broadly. For example, a defendant has
an interest in walking, not in walking to some particular place. As we shall explain below, at text ac-
companying notes 56ff, broadly described interests are a hallmark of liberal views about toleration
but should not be mistaken for neutrality.
1995]
J. COLEMAN & A. RIPsTEIN- MISCHIEF AND MISFORTUNE
We might put the connection between morally defined boundaries and the re-
quirement of causation in Kantian terms: an agent can either follow the moral law
or fail to do so. For the agent who follows the moral law, his agency –
for which
consequences of his or her actions he or she is responsible or owns – pretty much
ends with the intended consequences of his or her action. The situation is very dif-
ferent for the person who acts contrary to the moral law. In so doing, he or she
takes authority upon himself or herself that he or she does not rightly have, and
opens himself or herself up to the causal upshots of his or her conduct in ways in
which the person who follows the moral law does not. Consequences of his or her
wrong are his or hers to own, whether intended, or even foreseen, or not, including
some consequences that would not have occurred but for the conduct of others. The
scope of bad luck expands as a result of violating the moral law.3′
Simply substitute fault for the moral law. The person who is at fault opens him-
self or herself up to liability for unintended consequences of his conduct, including
some that would not have occurred but for the conduct of others. The conduct of
others may be a but-for cause of the harm, but the agent’s fault distinguishes his or
her conduct from that of others. Of course, if the plaintiff is also at fault, both have
a similar claim to own the costs of misfortune. Fault, far from rendering causation
and agency otiose, actually defines the scope of their relevance.
D. Limiting Liability
16. Still, fault need not open one up to unlimited consequences of one’s conduct.
So long as we understand the boundaries between persons in terms of the libertar-
ian’s spatial metaphors, it seems as though any boundary crossing must open up
unlimited responsibility. The result is a sort of “outlaw” theory – having done
wrong, one becomes liable without limit for an infinity of consequences. However,
if we understand the boundaries between persons as existing in moral rather than
geometric space, we see that the fault system can coherently hold someone liable
only for the losses that are within the risk implicit in the violated standard of care.
The standard of care divides particular risks between the parties, and if one fails to
exercise appropriate care, then he or she is liable for any harms following from that
particular risk. Because different degrees of care are called for in relation to differ-
ent activities and different risks, violating a standard of care is not enough to estab-
lish liability for consequences that result from unforeseeable causal chains. To es-
tablish liability in such circumstances would be to let unlimited liability in through
the back door.2
There are (at least) two different ways the fault system can be understood: we
can distinguish between the activities one chooses and one’s execution of them.
son, trans. L.W. Beck (Chicago: University of Chicago Press, 1949) 346.
“‘ See I. Kant, “On a Supposed Right to Lie from Altruistic Motives” in Critique of Practical Rea-
2 Even in those areas in which liability is strict –
not thereby become unlimited (see infra note 43).
that is, no showing of fault is required-
it does
MCGILL LAW JOURNAL/REVUE DE DROiTDE MCGILL
[Vol. 41
The first, more standard interpretation of fault holds that fault marks a failure in
execution: the failure to render reasonable care, given the activity one has chosen.
On this view, the standard of care owed varies with the activities in which one is
engaged. Each agent is presumed to have an absolute liberty with respect to the
choice of activities.” An agent is free of fault provided that the care taken is
deemed appropriate to the activity.’
We are explicating an importantly different view of the fault principle. On this
view, it requires that, in acting, each person takes into account the costs his or her
action may impose on others. Thus, the question is not whether I am being careful
by the standards of what I am doing, but whether I am being appropriately careful
in light of my neighbour’s interests in security and mine in liberty. The importance
of my particular activities enters into defining the appropriate degree of care, not by
holding me to a standard appropriate to the activity but in fixing the degree of lib-
erty appropriate to those engaged in that sort of activity.
Only this conception of fault can provide an objective measure of the costs of
my activity that will enable us to honour the liberal principle of fairness that one
should bear the costs of one’s own activities. To hold me only to taking such pre-
cautions as are reasonable or cost-justified for those engaging in that activity would
mark a move back to a subjective standard, and hold others’ security hostage to my
choice of activities. Rather than protecting each equally from the others, such a rule
would surrender security to an unlimited liberty interest in choosing activities,
however dangerous, provided one took precautions customary to those activities.
Thus, it would be a minor variant on limiting liability to intended consequences.
Because the standard of care protects both liberty and security and is supposed to
respect equally our interests in both, we cannot avoid looking at the activities them-
selves. The same logic that requires that we look to degrees of care, at all, requires
us to look to the values of those activities. Judgements of what conduct is at fault or
unreasonable involve both.
E. A Nuanced Objective Standard
17. Talk about objective standards makes some people uneasy. The idea of objec-
tivity may suggest that such standards are somehow eternal and exist quite apart
from questions about which interests people have and how important they are. Any
such conception of objectivity might well be suspected of being little more than a
3′ On a variant of this view, there is presumed to be something like an activity-neutral criterion of
what degree of care would have been reasonable or due. So, in this view, we might say that fault is in
the manner of execution, and the standard of care is independent of the activity chosen. This view
represents a logical possibility, rather than a position anyone has seriously put forward.
T’ This is the view of fault that underlies the Learned Hand formula, at least in its non-economic in-
terpretations. The root idea is that the benefits to someone participating in an activity serve to fix the
degree of care that should be attached to the security interests of others.
19951
J. COLEMAN & A. RIPSTEIN- MISCHIEF AND MISFORTUNE
smoke screen for interests that are already well-entrenched.” But we mean some-
thing considerably more modest.” Precisely because the fault standard turns on
substantive views about the importance of various activities, its contours will al-
ways be open to debate. It is objective in a negative sense, inasmuch as it is not
subjective, that is, the limits of liability are not fixed by the views, interests, or
abilities of either of the parties to a tort action. Instead, it protects the interests in
both liberty and security that everyone is assumed to have. On the basis of those
interests, it asks whether a reasonable person is entitled to have a particular interest
protected. The importance, and even existence, of particular interests is often con-
troversial, and the common law has sometimes been indifferent to what now seem
significant interests and concerned about insignificant ones. Clear examples of such
indifference can be found in the absence, until recently, of any legal recognition of
the interest that women have in being free of sexual harassment. The appropriate
response to such indifference, however, is to move to a more nuanced objective
standard, not a subjective one. New causes of action for sexual harassment have
developed in just this way.3 Rather than allowing the victim to be the sole judge of
the costs of the harasser’s activities, such torts sometimes import a “reasonable
woman” standard, which aims to recognize both the (potential) harasser’s interest
in self-expression and the (potential) victim’s interest in being free of behaviour she
perceives as inappropriate. While such a standard may carry risks of stereotyping,
when properly deployed it works in the same way as traditional objective tests are
supposed to. The standard is objective in that it divides the risk between the parties,
allowing both the possibility that the plaintiff is being too sensitive, and that the de-
fendant is behaving inappropriately though in good faith. Exactly where the line is
drawn depends on the importance of those two interests –
a matter that is sure to
be hotly contested. Recent feminist scholarship and advocacy have changed our
views about the boundaries of both sensitivity and good faith in such circum-
stances. What they have changed, however incompletely, are views about the costs
of various activities and who should bear them.”
3 See e.g. C.A. MacKinnon, Toward a Feminist Theory of the State (Cambridge, Mass: Harvard
University Press, 1989). For a discussion of Professor MacKinnon’s views, see especially S.
Haslanger, “On Being Objective and Being Objectified” in L.M. Anthony & L. Witt, eds., A Mind of
One’s Own: Feminist Essays on Reason and Objectivity (Boulder, Colo.: Westview Press, 1993) 85.
” We have both discussed objectivity in our previous work (see especially: A. Ripstein,
“Questionable Objectivity” (1993) 27 Nous 355; J. Coleman & B. Leiter, “Determinacy, Objectivity,
and Authority” (1993) 142 Penn. L. Rev. 549).
” Although these are typically statutory rather than tort based, they turn on all of the same issues of
reasonableness.
” See e.g.: K. Abrams, “Gender Discrimination and the Transformation of Workplace Norms”
(1989) 42 Vand. L. Rev. 1183 at 1197-1214; S. Estrich, “Sex At Work” (1991) 43 Stan. L. Rev. 813 at
842 (advocating reasonableness standards that recognize the seriousness of women’s interests and the
limited importance of men’s interest in harassment). As Estrich points out, the danger is that the
wrong objective standard will be applied (ibid at 843). See also Ellison v. Brady, 924 E2d 872 (9th
Cir. 1991). Despite the court’s claim that the standard is “not fault based” (ibid at 879), its reasoning
plainly is. Robert S. Adler and Ellen RL Peirce argue that a reasonable woman standard must be unfair
because it may hold men liable even though they are trying their best (see R.S. Adler & E.R. Peirce,
“The Legal, Ethical, and Social Implications of the ‘Reasonable Woman’ Standard in Sexual Harass-
MCGILL LAW JOURNAL! REVUE DE DRO1TDE MCGILL
[Vol. 41
F Strict Liability as a Measure of Misfortune in Rights Violations
18. Having argued that we must look to the importance of various activities in or-
der to set boundaries between persons capable of honouring the principle that we
should bear the costs of our activities rather than displacing them, we return for a
moment to consider a final effort to do without such an inquiry. The goal of such an
argument, recall, is to justify shifting the costs of some misfortunes while avoiding
anything like a moralized notion of fault and relying in its place on agency, control,
and the like. The first two interpretations of generalized strict liability relied on nar-
rowing the causal condition: the first by introducing a distinction between causal
and background conditions, the second by introducing the notion of the intended
consequences of one’s conduct. We found both of these wanting: the first because,
in narrowing the range of causes, it lost contact with the principle it was supposed
to express, the second because it allowed individuals to determine what counts as
the cost of their activities, thus, violating the principle it was supposed to express.
Our argument so far has been that we cannot determine the costs of activities
by focusing on activity-neutral concepts like cause, control, and agency. Instead,
the costs of one’s activities depend on what one owes others, which, in turn, de-
pends on the activities on which individuals are engaged: why they matter and in
what ways. The concept of duties between parties is expressed by the fault princi-
ple. Once we have a fault principle, we have a concept of boundaries between par-
ties and, thus, a role for causation and agency. The question is whether we need to
introduce fault in order to create these boundaries. What is it about fault that allows
it to do this job? Can the same task be accomplished in a way that is consistent with
a general principle of strict liability?
Suppose we re-cast libertarianism as the view that the costs of misfortune lie on
those they happen to befall unless they result from the rights-invasive conduct of
another.’
In such a view, it is not enough that the harm result from another’s
agency. The harm must also consist in the invasion of a right. Such a view draws a
boundary between the misfortunes that are an injurer’s bad luck to bear, and those
that are to remain the victim’s responsibility. Thus, it creates a basis for determining
the costs of various activities. It does so, moreover, without introducing the concept
of fault.
ment Cases” (1993) 61 Fordham L. Rev. 773 at 826). On our interpretation of objective standards,
any standard that protects people equally from each other may make someone liable for harms they
unwittingly cause. The alternative is to make one person’s well-meaning indifference the measure of
another’s interests. Our account also avoids the alleged difficulties with reasonableness standards
considered by N.S. Ehrenreich, “Pluralist Myths and Powerless Men: The Ideology of Reasonable-
ness in Sexual Harassment Law” (1990) 99 Yale L.J 1177.
“‘ Judith Jarvis Thomson seems to advocate a structurally similar view (though not as an interpreta-
tion of libertarianism) (see J.l. Thomson, Realm of Rights (Cambridge, Mass.: Harvard University
Press, 1990)).
1995]
J. COLEMAN & A. RIPSTEIN – MISCHIEF AND MISFORTUNE
This modification does not introduce the concept of fault, but it does introduce
all of the same substantive normative considerations that figure in determining
one’s obligations of care to others. Moreover, rights can serve as premises in argu-
ments about who is liable to whom only because who has what right is itself the
conclusion of a prior argument of precisely the sort that we have been invoking.
Ascriptions of rights are the conclusions of arguments regarding the relative impor-
tance of various activities and the weight that the relative security and liberty inter-
ests ought to command. Thus, a libertarian who introduces the concept of a rights
invasion introduces precisely the same sort of normative considerations that under-
write the fault standard; and in so doing, the libertarian risks losing touch with the
connection between agency and ownership, which makes the principle intuitively
plausible in the first place.
A regime of strict liability for rights violations faces another problem as well:
strict liability for rights violations turns out to be little more than a requirement that
harm be caused if a boundary has been crossed. Not only do boundaries need to be
defined in ways that cannot avoid substantive moral judgements, but we also need
to be careful about how we understand those boundaries. If we think of them in the
libertarian’s standard spatial terms and suppose that one is responsible for the full
consequences of one’s action once those boundaries are crossed, any rights viola-
tion seems to open the violator up to all of the consequences of his or her action.
But as we have suggested, this is neither how tort law works, nor is it an appealing
account of corrective justice. Boundaries need to be understood in terms of liberty
interests and the risks to security interests. If they are, boundary crossing can be
thought of as opening the wrongdoer up to the risks that were ignored in the bound–
ary crossing, rather than as a sort of forfeiture of any ability to disown conse-
quences of his or her actions.’
So far, our target has been a general theory of strict liability, which is a norma-
tive view about how the costs of misfortune are to be generally allocated. Our dis-
cussion has focused on the normative thesis, not on the strict liability rules that
have found their way into a variety of areas of tort law. That focus reflects our con-
cern with the fairness of any general principle for allocating the costs of misfortune.
We began with the simple thought that fairness precludesindividuals from displac-
ing the costs of their activities onto others. We showed why such a principle cannot
be coherently understood as the requirement that people bear all of the costs that
they cause others. We have tried to defend two claims against this apparently plau-
sible position. First, the concept of the costs of an activity cannot be understood in
terms of causation alone. Instead, it requires an account of the duties between the
parties. Second, any account of the duties of parties to each other invokes inquiries
into the substantive value of various activities and the seriousness of various harms
in order to assess the relevant liberty and security interests of the parties.
‘ Of course, rights could be structured to incorporate these concerns. But if they are, any remaining
connection with strict liability is lost. Without the dogmas, there is no strict liability left.
MCGILL LAW JOURNAL/REVUE DE DROiTDE MCGILL
[Vol. 41
G. Fault as a Measure – Conclusion
19. We conclude this section by considering whether existing legal practices can be
thought of as giving expression to this understanding of fault and the liberal prin-
ciple of fairness on which it rests. Of course, in a system of decentralized courts in
numerous jurisdictions, there may be no principled path through all of the ration-
ales that have been offered in tort decisions. But, an apparently more serious prob-
lem is that the conventional view claims that a large class of cases are adjudicated
under a principle of strict liability. What can we say about these cases? We cannot
take this question up in any detail here, but our view is that important areas in
which strict liability is employed as a rule for deciding cases turn on precisely the
same sort of reasoning that we have illustrated in our discussion of the fault system.
Consider first Rylands v. Fletcher,” the case which has been taught to genera-
tions of torts students as the leading example of strict liability in the nineteenth
century. In that case, the defendant’s reservoir leaked into mine shafts under the
plaintiff’s property. The defendant was held liable despite the Court accepting that
he had exercised considerable care in constructing the reservoir. The decision was
cast in terms of a distinction between natural and non-natural uses and imposed
strict liability for non-customary uses. Liability for mining or agriculture requires
that the plaintiff establish the defendant’s fault; liability for unusual activities re-
quires a showing of causation only. Thus, the defendant can be liable even if he ex-
ercised reasonable or, even, exceptional care. Yet, the Court’s reasoning turns out to
be considerably more subtle. A landowner must answer for the “natural and antici-
pated consequences” of keeping something mischievous on his property.2 For cer-
tain types of activities, no amount of care will do. This is not the same as claiming
that the defendant’s conduct is wrong in a way that would justify prohibiting him
from engaging in it, nor even wrong in a way that justifies granting injunctive relief
to the plaintiff. Rather, engaging in the conduct is permissible, but any way of en-
gaging in it imposes risks, which the person engaging in it must bear. Thus, any
costs resulting from the activity must lie with the defendant. The defendant’s duty
to the plaintiff is defined accordingly. The rationale is to be found in the relatively
greater importance of the plaintiff’s security interest than the defendant’s liberty
interest. No amount of care is sufficient because engaging in the activity itself poses
too great a risk to others.
So understood, important areas of strict liability involve the same sort of
judgements about the importance of various liberty and security interests as does
the fault system as we have explained it. Imposing liability in such circumstances is
thus an instance of the general strategy we have been advocating. If the standard of
care is supposed to protect people equally from each other, it cannot leave one per-
son free to choose any activity and then exercise care in relation to that activity. The
4 (1866), L.R. I Ex. 265, 4 H. & C. 263, aff’d (1868), L.R. 3 H.L. 330, 19 L.T. 220 (H.L.)
[hereinafter cited to L.R. I Ex.].
42Ibid at 279-80.
1995]
J. COLEMAN & A. RIPSTEIN – MISCHIEF AND MISFORTUNE
choice of activities may itself expose others to undue risks. In certain circum-
stances, the normality or otherwise of the use is a reliable indicator for the riskiness
of such activities. In circumstances in which almost everyone engages in the same
activities, and all are equally likely to be both plaintiff and defendant, a customary
activity rule may be a reasonable proxy for a substantive examination of the inter-
ests that are at stake. But that is because in such circumstances, everyone is en-
gaged in the same activities and, so, has the same interests in relation to them.”3
V. Liberal Egalitarianism: Choice as a Measure of Misfortune
20. So far, we have focused on variations of the libertarian view of event misfor-
tunes. The libertarian is committed to the same view of standing misfortunes. As a
result, that part of his view is subject to the same objections –
demarcating the
class of losses that do not result from agency is no less difficult than specifying
whose agency is involved in other losses.” Rather than continuing to focus on the
libertarian, however, we want to take up the liberal-egalitarian view of the standing
misfortunes of birth and status.
43 Other strict liability areas can be understood in a similar way. For example, master-servant liabil-
ity reflects the view that employing people to engage in risky activities is itself a risky activity so that
no amount of care in selecting employees will do. Justice Cardozo says as much in Leonbruno v.
Champlain Silk Mills, 128 N.E. 711, 229 N.Y. 470 (Ct. App. 1920) [hereinafter Leonbruno cited to
N.E.]: “The claimant while engaged in the performance of his duties in the employer’s factory was
struck in an eye by an apple which one of his fellow-servants, a boy, was throwing in sport at an-
other”. The accident was one “arising out of and in the course of employment … since the claimant
was injured, not merely while he was in a factory, but because he was in a factory, in touch with as-
sociations and conditions inseparable from factory life. The risks of such associations and condi-
tions were risks of the employment” (Leonbruno, ibicL at 711). The basic assumption is that in
keeping a group of employees working together at a dull job, you do so knowing that there will be
horseplay that may injure someone. Thus, you are liable, as you would be if you brought together any
other set of dangerous forces you can expect to be unable to control. The reason is simple: the secu-
rity interest of those who may be injured is important enough and the employer’s interest in saving
expenses insignificant enough that the former outweighs the latter. This is not to say that anyone who
employs others is really “at fault’ for so doing but only that they must bear the costs of their risky ac-
tivity.
The same factors can limit the reach of strict liability in areas in which it is otherwise important.
In Williams v. RCA Co., 376 N.E.2d 37, 59 1. 229 (Ct. App. 1978) [hereinafter Williams], a security
guard was injured because his walkie-talkie failed when he was seeking assistance to apprehend an
intruder. In Klages v. General Ordnance Equipment Co. 367 A.2d 304, 240 Pa. S. 356 (1976), the
failure of a can of mace was grounds for liability because it was brought into the stream of commerce
as a crime prevention device. The court in Williams held that the manufacturer was not liable for a
risk for which its product was not designed. Once again, risks are allocated on the basis of various
liberty and security interests. In this case, the court held that the defendant’s interest in selling its
product did not require it to indemnify all users against all possible risks, despite the general rule of
strict liability.
“That is, if agency is everywhere but-for causation is, then virtually all misfortunes will be com-
pensable, while if agency’s reach is narrower, we need some way of specifying how to narrow it.
MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL
[Vol. 41
Conventional wisdom places libertarianism and liberal egalitarianism at differ-
ent ends of a continuum of liberalism. The libertarian emphasizes ownership,
agency, and human autonomy, whereas the egalitarian emphasizes community and
equality. Certainly, these differences are apparent if we think about their respective
positions on the default rule that ought to govern misfortunes. The libertarian opts
for the Hobbesian default rule that costs are to fall where they lie unless there is an
agency-related rule for shifting them. The liberal egalitarian opts for the opposite
rule that, in the absence of a reason for shifting the costs of misfortune, they ought
to be held in common.
In spite of this obvious difference, our view is that the liberal egalitarian and
libertarian face fundamentally the same problem: both claim allegiance to the lib-
eral principle of fairness that one should not be allowed to displace the costs of
one’s activities onto others; both are committed to understanding those costs in
terms of activity-neutral features, such as causation, agency, and choice. Both are
also committed to the idea that there must be some default rule against which the
activity-neutral concepts define an exception. We have seen where this approach
leads the libertarian astray. Our aim in this section of the paper is to develop a
parallel argument against the liberal egalitarian. In so doing, our sympathies are
closer to the liberal egalitarian’s conclusion than they were to the libertarian’s. Our
target for the present is the attempt to understand liberal equality in activity-neutral
ways.
Once we see the difficulty with any attempt to achieve neutrality, we will see a
further sense in which issues of distributive and corrective justice cannot be sepa-
rated. Both the libertarian and the liberal egalitarian define their views in terms of
default positions. An examination of the difficulties this creates for the liberal
egalitarian will enable us to see that thinking about misfortune in terms of a default
position reflects an untenable view of the boundaries between persons. Once we
think of those boundaries in terms of the factors that the fault standard requires us
to consider, we see that it is a mistake to think that there must be a single-default
position for dealing with misfortunes, and any departure from this must meet a
single test. Where losses appropriately lie depends on what sort of losses they are
as well as how they came about.
Where the libertarian focused primarily on event misfortunes of the sort that
occupy torts casebooks, the liberal egalitarian is typically concerned with standing
the misfortunes that Dworkin calls “brute luck”.”3
misfortunes of birth and status –
Where the libertarian tried to move one step in from the Hobbesian state of nature,
the liberal egalitarian moves one step in from the world of perfect community. All
misfortune is to be held in common unless it is the result of choice and control, the
result of what Dworkin labels “option luck”. If the relevant distinction for the liber-
,5 We will use Ronald Dworkin’s seminal works on equality (see: “Equality of Resources”, supra
note 4; “Equality of Welfare”, supra note 11) as our example here, noting along the way the variants
proposed by others concerned with fine-tuning Dworkin’s basic picture.
19951
J. COLEMAN & A. RIPSTEIN- MISCHIEF AND MISFORTUNE
tarian is between agency and chance, the relevant distinction for the liberal egalitar-
ian is between brute and option luck.
Where the libertarian would be happy with the state of nature but for its failure
to distinguish between those who keep to themselves and those who harm others,
the liberal egalitarian would be happy with the world of perfect community but for
the fact that it treats hardworking ants and lazy grasshoppers alike. The remedy for
the libertarian is to introduce a requirement of agency as a condition of liability.
The remedy for the liberal egalitarian is to give everyone a fair share of resources,
rectifying those misfortunes that nobody can help but allowing those who wish to
husband their share of resources to their own advantage.
All brute luck is held in common, its effects are cancelled. Only option luck,
the result of voluntary risk-taking, plays a role in Dworkin’s scheme. Thus, the lib-
eral egalitarian appeals to “a certain view of the distinction between a person and
his circumstances and assigns his tastes and ambitions to his person, and his physi-
cal and mental powers to his circumstances.”
If the libertarian could be said to be committed to the ideal that individuals
must bear the costs of their agency, the liberal egalitarian could be said to be
committed to a strikingly similar interpretation of the idea that people should bear
the costs of their choices. Where they differ is in their account of the conditions un-
der which this requirement makes sense. For the liberal egalitarian, the principle
that individuals should bear the costs of their choices –
and the misfortunes of op-
tion luck –
applies only if certain background conditions are met. These are repre-
sented by the idea of equality of resources: provided that people have fair shares of
resources to begin with, it makes sense to ask that they bear the costs of their
choices. If, however, initial holdings of resources are unfairly distributed, the idea
of bearing the costs of choices is under-defined. People cannot be said to bear the
costs of their choices, because the costs they end up bearing reflect initial holdings
as much as the choices themselves. Let us see how the argument goes.
21. Dworkin emphasizes the important role of markets in defining equality of re-
sources. When markets function efficiently, prices reflect the degree to which peo-
ple value various goods that are for sale. If something is scarce, and many people
want it, it is appropriate that its price should be high. That higher price reflects the
opportunity costs to others of any particular individual’s choices: if by consuming
something, I make things like it more scarce and others have to forego them, its
price reflects the costs of my consumption to others. Again, if I produce something
others want, it is appropriate that I should benefit from the resources created or
freed up by the opportunities I have provided them. Should I decide to do some-
thing doubly self-indulgent –
say taking a year off from being a philosophy profes-
sor to study law –
it is appropriate that I should have to forego other things I might
otherwise have been able to do. The costs of my decision appropriately lie with me,
6″Equality of Resources”, supra note 4 at 302.
MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL
[Vol. 41
not with any other particular person or with the larger community. They are mine
because they result from my choices. Of course, in making any such decision, I
take a chance: if things work out for me, great; if not, it is simply my own bad luck
– bad option luck.
The benefits and misfortunes of option luck are mine to bear because they re-
flect my choices. But those costs will not fully reflect my choices or their oppor-
tunity costs to others unless the effects of brute luck are somehow cancelled, If
people are to bear the costs of their choices, the irrelevant effects of circumstances
must be controlled for. If someone begins with a larger bundle of goods, their abil-
ity to outbid others for scarce resources (and the prices each must pay) may reflect
their initial good fortune, rather than the costs of their choice. Likewise, the person
who starts off poor or handicapped will, through no choice of her own, have fewer
resources to bid for the things she wants. If the effects of brute luck are left in
place, the costs of my choices will reflect its contingencies at least as much as they
reflect my choices. Dworkin’s root idea is that my fortune in life should reflect my
choices, not my circumstances.
Here we find both the similarity and the difference between the liberal egalitar-
ian and the libertarian. In both cases, the costs of one’s activities are fixed in terms
of activity-neutral concepts: cause in the case of the libertarian, choice in the case
of the liberal egalitarian. The difference is that causation for the libertarian requires
no special context within which to operate. Indeed, any principle for sharing the
costs of setting up any sort of corrected context would violate the libertarian’s un-
derlying principles of agency and ownership. In contrast, the point of the liberal
egalitarian position is to have one’s choices determine the costs of one’s activities,
and that requires setting the right context. In the absence of a context corrected to
remove the effects of brute luck, the costs of choices cannot even be defined. For
the liberal egalitarian, agency and ownership are only related in the appropriate
context, so no concerns about violations of agency that might arise in a different
context can even arise.
According to liberal egalitarianism, costs can only be properly defined in terms
of fair initial shares of resources. But the value of those resources, in turn, can only
be measured once the effects of brute luck are cancelled. Once the effects of brute
luck are cancelled, ordinary market transactions will ensure that people bear the
costs that their activities create for others. If I do something others value highly, I
will be rewarded; if I have expensive tastes, I will have to decide what else to
forego.
The liberal egalitarian’s view reflects a particular interpretation of the idea that
people should bear the costs of their choice, which makes the idea of a choice turn
on whether or not something is within a person’s control. The interpretation of fair-
ness might be summed up as follows: how well or badly a person’s life goes should
not depend on things they cannot control. Thus, the exception to perfect community
in cases of option luck gets some of its appeal from the twin ideas that people
19951
J. COLEMAN & A. RIPSTEIN- MISCHIEF AND MISFORTUNE
should not be made to bear the costs of things they cannot control, and that they
should bear the costs of the tastes they choose to develop. The distinction Dworkin
draws between expensive needs and expensive tastes rests on the idea that tastes are
the sorts of things that people can control. Again, the idea of option luck is just the
idea that someone could have contented themselves with their share of resources
but, having decided to take a chance, must bear its cost. In addition, behind the idea
of brute luck lies the idea that some things simply happen, beyond the reach of
anyone’s control.
22. Appealing as these intuitive ideas are, they turn out to be rather less clear when
applied to specific cases. Just as philosophers love to find causation everywhere, so
economists are fond of claiming that virtually any human action can be represented
as a calculated gamble. Often, the probabilities are sufficiently high or low that or-
dinary people overlook the risks; but they are no less real for that. Every action has
costs in terms of opportunities foregone, and so anything anyone does can be
treated as exposing oneself to risk: if I invest in Beta and other people choose
V.H.S., I took a risk and lost. Likewise, if I invest in tubes and the world chooses
transistors, or I invest in LP’s and the world chooses CDs. In each of these cases,
there may be some temptation to suppose that I did not need to take the risk, so that,
although the outcomes themselves were not within my control, my exposure to
them was. But tamer choices also have the same implications: if I work hard on the
family farm, the industrial age may overtake me. If I plant canola the same year
everyone else does, the drop in prices may overtake me, or the weather may turn
bad, hurting all of us. If I follow Pascal’s advice and stay in my room, I may get
bedsores.
Not every misfortune, however, should be treated as option luck to be borne by
the person it befell simply because they implicitly took a calculated risk in leaving
the house or getting out of bed in the morning. Part of the problem is that if we
construe risk so expansively, we lose the connection with control that first made it
appealing. But if instead we limit the idea of risk to those circumstances in which
the person could in fact control the outcome, the results are still stranger. If you and
I are both trying to build a better mousetrap, and I outsell you, society would have
to indemnify you. Indeed, any risk that did not work out would have to be treated
as brute luck, thus, leaving the category of option luck all but empty. Like causation
in the libertarian’s account, option luck seems to be everywhere and so provides no
way of deciding to whom particular misfortunes belong.
The libertarian is one step in from the state of nature and the liberal egalitarian
one in from the perfect community. Not surprisingly, their difficulties are parallel.
Each wants to carve out an exception to an otherwise attractive extreme position.
Each wants the exception to accommodate the idea that people should bear the
costs of their actions, and each tries to carve out the exception by looking to such
value-neutral features of action as causation, intention, and control. Any neutral
feature of human action, however, will also be generic and, as a result, too perva-
sive to neatly demarcate an exception.
MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL
[Vol. 41
A. The Inadequacy of Option Luck
23. One response to the way in which all action and inaction involve risk will
surely not work. Option luck cannot be modified to include only conscious, delib-
erate risk-taking. The problem is identical to that with our imagined example of
strict liability for the intended consequences of actions. Rather than honouring the
principle that people should bear the costs of their choices to others, limiting option
luck to intended exposure to risk effectively allows each person to decide what will
count as the costs of their choices. The person who carelessly took the same risks
would be relieved of the costs of his or her choices.’7 Yet, the whole appeal of a
market-based account of equality was that the notion of opportunity costs appeared
to provide a way of measuring the costs of each person’s choices to others.”
Another approach, parallel to the attempt to retrieve generalized strict liability
by thinking of causation in terms of what is normal or typical, might seek to distin-
guish brute from option luck by distinguishing choices that are unusually risky
from those that are not. If someone takes a risk that others in similar circumstances
typically avoid – pursuing the life of an artist, or making a risky investment –
the
good or bad luck that accrues belongs to that person alone. In contrast, if someone
does something that is normal for those in similar circumstances, the results might
be thought of as brute luck. 9 Like the attempt to narrow causation’s reach, however,
any such account will sever any intuitive connection between choice and agency.
Whether something counts as choice or circumstance will depend on the frequency
7 What about the benefits? The situation is not strictly symmetrical, because the Dworkinian hopes
to cancel the favourable as well as the unfavourable effects of brute luck. Thus, the benefits of brute
luck would be taxed away at a much higher rate than those of option luck, which would still be taxed
where necessary to indemnify others’ brute losses. Still, treating either benefits or burdens alike based
on what someone thought they were doing would just recreate the problem that faced the subjective
standard of fault and strict liability for the intended consequences of actions. Far from underwriting
equality, it would make each person’s fair share depend on others’ decisions about what counted as
their fair shares.
48 Dworkin sometimes comes close to this position when he defines tastes as those desires with
which people identify (see “Equality of Resources”, supra note 4).
49 There are at least two ways of reading “similar circumstances” here. One looks to the details of
each person’s natural endowments, considering the relative shares of various natural talents, and
makes comparisons only across persons who are similarly situated with respect to all, or nearly all of
them. Another possible approach, independently suggested to us by Bruce Ackerman and Robert
Bright, arrives at a measure of a person’s total circumstances by somehow aggregating various abili-
ties within each person’s circumstantial endowment. The latter approach would allow us to make
comparisons across a broader number of persons. Unfortunately, it probably faces deep problems of
definition, because the relative values of various natural and circumstantial endowments will depend
on how various misfortunes are treated. To return to an earlier example, whether an eggshell skull
gives a person a less valuable overall endowment will depend on who bears the increased risk of seri-
ous injury. Thus, it requires an account of risk-taking and so cannot provide one. The same structural
features stand in the way of applying a parallel account where the alternatives are between a cost be-
ing borne alone and costs being held in common.
1995]
J. COLEMAN & A. RIPSTEIN – MISCHIEF AND MISFORTUNE
with which others make similar choices. If many people take a risk, it will not
qualify as a choice. Therefore, if there is a wild frenzy of stock speculation, the los-
ers will need to be indemnified. But if a neighbourhood is so dangerous that few
people venture out at night, they will count as risk-takers and, so, not have their
losses made good. Like Epstein’s attempt to distinguish causes from conditions,
this approach ties responsibility to unusual activities. As a result, it severs the tie
between agency and responsibility.’
B. The Inadequacy of Insurance as a Measure
24. Dworkin is aware that the distinction between brute luck and option luck is a
matter of degree.” He suggests that the two can always be linked by the amount of
insurance people would buy against various forms of brute misfortune. Dworkin
argues that there are some risks that it would be rational for people to insure
against, and others it would not. Equality requires treating the former risks as held
in common and thus compensating for their effects. That suggestion has generated a
substantial literature of its own, the central concern of which is whether particular
risks would or should be insured against. 2 The candidates range from limited tal-
ents to the propensity to develop expensive tastes. Without trying to cover this lit-
erature, let alone join it, a few comments are in order.
Although Dworkin himself never explicitly makes such a claim, the basic strat-
egy throughout the “equality of’ literature he fostered is to insist that some misfor-
tune or other is beyond people’s control, so that it is appropriate for them to insure
against it. The link between insurance and control, however, is not helpful. There is,
though, a more powerful way of employing the idea of rational insurance which
helps to make sense of equality and explains why something like an index of pri-
mary goods is needed. The rationale is parallel to that which leads to the fault stan-
dard. Where fault divides risks, insurance provides a way of pooling them. In real
life, people insure against risks based on their likelihood, the cost of insurance, and
” These examples may seem to invoke paradigmatic cases of agency to which responsibility clearly
does or does not attach. We do not mean to deny that there are paradigm cases of choices in which
people should be made to bear the costs of those choices because they have brought them upon them-
selves. Just as we can distinguish force, fright, and the setting of dangerous conditions from other
types of causation, so we can distinguish self-indulgence or expensive tastes from other cases of ex-
posure to fortune. The point is that we can only do so with an independent account of responsibility,
of which choices count which ways.
See “Equality of Resources”, supra note 4 at 293.
52Dworkin’s own answer is that people would only insure against the risks that they can imagine if
they materialized. Thus, some tastes or ambitions might be thought of as matters of circumstance
rather than choice, perhaps because those whose tastes they are experience them as unwanted crav-
ings and might wish they had insured against them. Dworkin says little about how to identify such
cases, though the central idea seems to be that such cravings may be a matter of brute luck. The ap-
proach seems to us unpromising, because unlike actual insurance, hypothetical insurance purchasers
cannot look to their personal idiosyncrasies. If anything, displacing the question of what people
would insure against onto the question of whether they would regard certain tastes as unwanted
cravings makes the problem less tractable.
MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL
[Vol. 41
the importance of the things that they stand to lose if things go wrong.3 Asking
what rational agents would insure against, while not knowing various things about
them, is simply a vivid way of representing the importance of various goods and
activities in a political theory.’ Many things lie beyond a person’s control, yet most
of us would bridle at the thought of providing insurance against them. People may
have limited control over whether or not they are charming or have various abilities
that are unusual but too common to command a market premium, like the ability to
juggle two balls in the air. But whether or not one has such an ability is not, we may
suppose, important enough to a minimally decent life that we want to pool the risk
of not having that attribute. Conversely, being able-bodied, or good-natured or de-
cently educated are not things that all actual people would, in fact, purchase insur-
ance to protect; rather, they are plausible candidates because we judge them to be
important to any reasonably decent life.
VI. Primary Goods as a Measure of Misfortune
25. With his idea of primary goods, Rawls offers us just such a strategy for decid-
ing which risks matter and how. Rawls introduces primary goods as the basis for
fair distributive shares. He offers them as things people can reasonably be expected
to want as much as possible of, and thus, any inequalities in their distribution must
be justified. Like the fault system, primary goods provide an objective standard for
measuring costs. A public measure of important interests makes interpersonal
judgements of fairness possible and lets us distinguish between those misfortunes
that lie where they fall, and those that are to be thought of as belonging to every-
one. Like the fault standard, in order to be up to the task, an index of primary goods
cannot adapt itself to individual idiosyncrasies. Some people might have different
values –
and
object to the choice to pool risks other than those on their preferred list. After all,
they are being asked to give up other things for benefits they do not actually want.
The person who wants other goods, however, does not have a claim in justice
against others. Nor does the person who is unhappy with his or her share, nor the
person who finds that, say, liberty makes life more difficult. None of these people
have a claim in justice because each effectively asks that their own view be allowed
to fix what does and does not count as a cost of their activity to others. If we are to
give sense to the idea of people bearing the costs of their activities, we need to de-
cide which things count as a person’s activities and which do not. Leaving it up to
each person to decide which risks to bear and which others must bear undoes the
idea of people bearing the costs of their activities to others, for each person both
thinking that being a good juggler is essential to any decent life –
losses.
Though they may underestimate the risks and have difficulty gauging the importance of various
” In a review of Rawls’s book (see Rawls, supra note 6), Dworkin makes essentially the same point
about the former’s use of the “original position” as an expository device (see R.M. Dworkin, “Justice
and Rights” in R.M. Dworkin, ed., Taking Rights Seriously (Cambridge, Mass.: Harvard University
Press, 1978) 150).
1995]
J. COLEMAN & A. RIPSTEIN- MISCHIEF AND MISFORTUNE
sets the costs of their activities to others and is subject to the choices of others.
Treating people as equals requires a common currency.
Control cannot be used to determine what things count as the costs of an activ-
ity for the same reasons that tort liability cannot be limited to the intended or actu-
ally foreseen consequences of actions. The difficulty with limiting tort liability to
intended or actually-foreseen consequences is that it leaves the determination of the
costs of various activities as a matter of idiosyncratic individual choices. These
were not choices about which activities to engage in (which do rightly enter into
determining the costs of various activities) but choices about what counted as those
costs. The only way to measure costs fairly is to measure them objectively, to
measure them in a way that makes it possible to say, “You should have been more
careful,” or “You should have taken the dangers to others into account.”
At the same time, we do not want to say such things about everything that the
person might have done in aid of another’s security. Only the important ones; hence
the fault system. In just the same way, if we limit the place of option luck to those
things of which a person knew the risks or, in fact, controlled, we lose any measure
of the costs of activities to others in favor of what the person took to be the costs of
those activities. In consequence, we lose the ability to make a point essential to the
idea of bearing the costs of one’s choices: “You ought to have realized there was a
risk involved,” or “You should have realized that was going to turn out to be so ex-
pensive that it would get in the way of the rest of your life.” But, of course, we do
not want to say such things about everything that someone might have done some-
thing about. Hence we need a solution parallel to the fault system.
Once we see that what is really at issue is not control but the importance of the
various aspects of a person’s life, some supposed problems for liberals disappear.
Libertarians are fond of taunting liberals with the charge that economic redistribu-
tion cannot be distinguished from such gruesome practices as redistributing body
parts to those who need them.5 So long as we suppose that we can get substantive
politics out of action theory, the charge is worrisome; but once we recognize the in-
escapability of substantive judgements about the importance of various activities
and security interests, the charge becomes less interesting. For we can plausibly
suppose bodily integrity to matter enough that body parts cannot be redistributed
simply because others find them useful.
We can also insist that property rights need not be thought of in the same way
as is bodily integrity. The libertarian might stick to his guns, hoping to show that
property is just as important as bodily integrity. But showing that would require a
55 This, no doubt, grows out of the idea that ownership begins with owning one’s body; hence, any
taking of property is no different from a taking of a body part (see e.g.: Nozick, supra note 11 at 206;
J. Narveson, “On Dworkinian Equality” (1983) 1 Soc. Phil. & Pol’y 1). In defending Dworkin’s
views, Will Kymlicka concedes that liberals should be concerned by this accusation (see W. Kym-
licka, Contemporary Political Philosophy: An Introduction (Oxford: Clarendon Press, 1990) at 154).
MCGILL LAW JOURNAL/REVUE DE DROIT DE McGILL
[Vol. 41
substantive argument about the importance of various activities, not conceptual leg-
erdemain about agency and the like.
So we must reject the view that we can settle questions about the costs of ac-
tivities by invoking concepts like control or choice, not because these concepts
make no sense, but rather, because their content cannot be fixed independently of
engaging in normative argument about the values of various activities to us and the
ways in which they matter in people’s lives. We need something like a public index
of goods.
Some candidates are easy: a modicum of wealth, health and education at least.
A particular person’s lack of these is a misfortune that we all have a responsibility
to rectify. This is not because such circumstances are necessarily beyond the per-
son’s control but, rather, because of the ways in which such goods are substantively
important in people’s lives. There is no liberal theory worthy of the name that is
neutral on such matters. Liberalism must face up to questions about misfortune and
bad luck that are political throughout. To answer these questions together is to be a
liberal-political community. Such a community need not share all misfortunes but
must share a vision of what is to count as the costs of various activities.’
‘6 Maybe, partly because we seem to be talking about everything at once here, and partly because
one of us (Coleman) talks about it in everything he writes in torts these days, a few words are in order
about generalized no-fault compensation schemes, such as the New Zealand model (see: Royal
Commission to Inquire into and Report Upon Workers’ Compensation, Compensation for Personal
Injury in New Zealand (Wellington, 1967); K.H. Marks, “A First in National No-Fault: The Accident
Compensation Act 1972 of New Zealand” (1973) 47 Austl. L.J. 516. For a discussion of implicit
subsidies for motorcyclists, and others engaging in risky activities, see G.T. Schwartz, “National
Health Care Reform On Trial: A National Health Care Program: What Its Effect Would Be On
American Tort Law and Malpractice Law” (1994) 79 Cornell L. Rev. 1339 at 1356). Anyone writing
about corrective justice and fault (with the possible exception of libertarians, who, as we have seen,
have no room for fault in their accounts) needs to explain why the eclipse of fault in favor of a system
of generalized, social insurance is not a loss of something morally significant. The general strategy of
objective standards requires a substantive, though thin, index of primary goods in determining what
counts as the cost of yarious activities. Narrow social insurance schemes, such as health-care pro-
grams, rest on the idea that the risks of certain misfortunes created by a host of everyday activities are
to be held in common. (In the case of health care, these range from eating and exercise habits to
methods of transportation and types of employment as well as sexual activities and the types of public
places one frequents.) The rationale for treating health in this way is that it is a primary good, impor-
tant to everyone’s life plans whatever they may be. A generalized social insurance scheme seeks to go
one step further, protecting against all harms occasioned by accidental interactions of two or more
parties. As such, it treats losses occasioned by accidents in the same way many health-care schemes
treat illness. So described, such a scheme may sound unattractive for any number of reasons, not the
least of which being that the class of accidents protected against does not seem like a morally signifi-
cant category. There may also be losses of other important goods in a ‘society that make accident
avoidance a central primary good, since any change in an index will have impacts elsewhere. But if
such a good is deemed important enough, there is no sacrifice to the general way of looking at misfor-
tune that we have put forward here.
19951
J. COLEMAN & A. RIPSTEIN – MISCHIEF AND MISFORTUNE
If neutrality is not a genuine option, however, toleration still is. Fairness and
toleration fit together if the interests protected by primary goods include liberty in-
terests in deciding what to do about important matters and security interests in the
wherewithal to pursue the choices one makes. But to protect this sort of toleration,
other putative security interests, including some that many people may think impor-
tant, may need to be ignored. Thus, any interest that some might claim in cultural
survival ‘ or a secure moral climate must be rejected. Here, too, any index of pri-
mary goods will be controversial. Some resolution to these questions is required,
however, if the idea that people should not displace costs onto others is to be hon-
oured. The liberal commitment to toleration is expressed in the same way by
broadly-defined liberty and security interests in tort law.
Conclusion
26. The libertarian and the egalitarian differ in their default rule for misfortunes but
not in their rule for the other cases. Both aspire to neutrality, so both think that the
default position itself can be given a neutral definition. But it cannot; it establishes
boundaries between persons, and must be negotiated in the same substantive way
that any other boundary is. Once we see that it must be negotiated, however, we
also see that there need be no single default rule either. Instead, we always face the
question of whether a misfortune should lie where it falls, be shifted to some other
particular person, or be held in common. Because the boundaries between persons
reflect judgements about the importance of various activities and security interests,
the question of where some particular cost lies –
like the related question of what
counts as taking a risk – must have a substantive answer. Looking for an appealing
default position only makes sense if a neutral rule is available to define exceptions
to it.
Both an objective standard in tort and primary goods for distributive justice re-
quire balancing very different things against each other; but it does not require that
they share some other, underlying feature that makes them commensurate. Indeed,
the point of fixing an objective standard is to provide a way of measuring the costs
of activities across persons. Doing so does not depend on a prior commensurability
but provides a public measure that makes it possible to measure costs across per-
sons. 8 Money provides a helpful, if potentially misleading, example. Prices do not
57 See e.g. C. Taylor, Multiculturalism and the Politics of Recognition: An Essay (Princeton: Prince-
ton University Press, 1992): “Political society is not neutral between those who value remaining true
to the culture of our ancestors and those who might want to cut loose in the name of some individual
goal of self-developmene'(ibid at 58).
” See e.g.: J. Raz, The Morality of Freedom (Oxford: Clarendon Press, 1986); A. Sen & B.
Williams, Utilitarianism and Beyond (Cambridge: Cambridge University Press, 1982) at 17-18.
Those who insist that incommensurability entails arbitrariness share a questionable view of rationality
with those who insist that choice entails commensurability. Both suppose that the only way something
can be rational is if it reflects the application of a quasi-mechanical procedure –
utterly free of
judgement or evaluation –
to some set of independently describable facts. Like the advocate of strict
