The Tanner Lectures on Human Values

Is Liberty Possible?

CHARLES FRIED

THE TANNER LECTURES ON H UMAN VALUES

Delivered at
Stanford University
May 14 and l8, l981

CHARLES FRIED was born in Czechoslovakia in 1935, received his A.B. degree from Princeton University and M.A. in Jurisprudence from the University of Oxford. His J.D. degree is from Columbia Law School. After serving as law clerk to Justice John Marshall Harlan of the U.S. Supreme Court in 1960–61, he began teaching at Harvard Law School, where he has been ever since. His books are An Anatomy of Values (1970); Medical Experimentation: Personal Integrity and Social Policy (1974); Right and Wrong (1978); and Contract as Promise: A Theory of Contractual Obligation (1981). Charles Fried is Carter Professor of General Jurisprudence at Harvard Law School.

I am grateful to the Trustees of the Tanner Lectures on Human

Values and the University of Utah Press for allowing me to combine in this one essay what were originally two lectures, both delivered at Stanf ord University in May 1981. Tbe second of these, appearing as Part II here, is part of a continuing project on the foundations of private law, assisted by a summer research grant from the Harvard Law School and a grant from the Olin Foundation. In addition to those wbo commented on these lectures in Stanford — I think especially of Patrick Atiyah, Wayne Barnett, Mitchell Polinsky, Thomas Scanlon, Alan Stone, and Steven Strasnik — I received valuable comments from Sissela Bok, John Ely, Frank Michelman, Richard Parker, William Ewald, Judith Jarvis Thomson, and the members of SELF to wbom I presented a preliminary draft, Hilary Putman was particularly generous in sharing his time, idea, and an early draft of his Reason, Truth and History (Cambridge University Press, 1981).

PART I, FAIR SHARES — THE SEARCH

FOR A STANDARD

1. Is liberty possible? The question is intended both in a practical and philosophical sense. The practical question is can we devise institutions which express a commitment to individual liberty while fulfilling the other imperatives of social ethics, particularly the claims which the unequal enjoyment of resources or outcomes make on the better favored and thus on the community as a whole. The related philosophical question asks whether there is a coherent concept of liberty at all. The questions tie together in this way. If the situation of individuals depends on collective claims for contribution to the well-being of others, that situation

is insecure and their liberty is threatened. If the very definition or conception of what a person´s situation is depends on collective claims directed at collective goals, including collective goals about fair distribution, the distinctness of the concept of liberty is undermined. To put the issues crudely at the outset, if everything about an individual - his person or his product - is available for redistribution, then individuals are not free. And if there is not even any stable way to define what a person´s entitlements or rights are before we proceed to consider redistributing, if the very conception of what are my resources such that I must make contribution from them inevitably touches on issues of collective and redistributive goals, then liberty is not possible in the deep, philosophical sense. It is not an independent concept at all.

The first of these questions, the question of fair shares or distributive justice is a staple of political philosophy. The most complete recent attempt to argue for a conception of fair shares, John Rawls´s A Theory of Justice, sets the standard for debate on this subject. In an earlier work1 I objected that Rawls´s celebrated maximin proposal is ambiguous about the extent to which a person´s talent, ambition, character are or are not social assets only provisionally assigned by the morally irrelevant hazard of what he calls the natural lottery. To the extent that maximin allows the better endowed to hold out for higher income in exchange for their contribution to the situation of the worst-off, it seems to recognize a moral title in those endowments after all. But if there is such a moral title, then one wonders how an obligation to contribute what is one´s own arises at all. Rawls´s suggestion that the better endowed are indeed entitled to enjoy the fruits of their greater endowments but only to the extent that this improves the situation of the least well-off hardly lays these doubts to rest, but rather just restates the initial proposition.

1 Right and Wrong, 161n (Cambridge, Mass., 1978); C. Fried, review of A Theory of Justice, 85 Harv. L. Rev. (1972).

The second of my questions, about the coherence of the concept of liberty, arises in the midst of the first, for both Rawls and a critic of Rawls such as Nozick assume that there is some clear sense to the notions of what is attributable to a person, what are his endowments.2 They assume a conception of property in one´s person, one´s talents and efforts, and a conception of property in the fruits of those talents and efforts. Indeed, as I intend to show in detail in Part II, what both Rawls and a critic such as Nozick need are stable conceptions of tort and contract - that is, notions about how and what interests deserve recognition and protection and how far one may dispose of or exchange what is one´s own. Such a stable basis is necessary so that we can make sense out of the question of how much of what is attributable to a person should be left with him and how much exacted by way of contribution to the well-being of others. But there is a critique which holds that there is no defensible way to resolve that subsequent question except by attending explicitly to distributive and other collective goals.3 For instance, a transfer induced by fraud should not be allowed to stand and the victim of the fraud should have his property returned to him. This seems like a judgment which has nothing to do with distribution - it is pre-distributive, a part of a system of rights on which the distributive judgment works. Yet what is a fraud is a question which itself seems to implicate distributive judgments: how far may one profit from greater shrewdness, how far must one share information which bears on the desirability of a bargain ? If liberty is threatened by the spectre of redistribution, which subjects me and what is mine to the needs of others, how much more is liberty threatened if we cannot even tell what is mine without considering those claims.

2 John Rawls, A Theory of Justice, ch. 5 (Cambridge, Mass., 1971); Robert Nozick, Anarchy, State and Utopia, ch. 7 (New York, 1974).

3 See, e.g., Kennedy and Michelman, “Are Property and Contract Efficient ?” 8 Hofstra L. Rev. 711 (1980); Kronman, “Contract Law and Distributive Justice,” 89 Yale L. J. 472 (1980).

Thus the two themes of fair shares and private rights connect. The first belongs to political philosophy, the second to legal theory, but both are necessary to a conception of liberty. And they seem to chase each other in a circle: distribution presupposes a system of private rights, and yet no such system seems available without settling distributive issues. So there is not only a set of substantive issues to be resolved, but an epistemological or methodological issue as well - how to cut into this circle. That should not be a surprise, for it has always been objected that liberty is a meaningless concept without the specification of the very social background which liberty is supposed to judge and constrain. I shall proceed by asking first in a general way why we should be concerned about liberty and about distribution. Then I shall propose and define a standard of distribution, ignoring provisionally the problems of legal theory which threaten the coherence of the terms in that proposal. In Part II I confront these, showing how we may hope for a theory of private rights. Along the way I hope to illuminate the methodological concern that any conception of liberty must try to jump out of its own conceptual skin, which of course it cannot do.

2. It is well to say in a simple way at the outset why I care about liberty and what I fear about redistributive and other collective claims that seem to threaten liberty. Though it may seem an irony in a statement which announces a desire to remain simple, the shortest way is to describe my orientation as Kantian. I do not know of a standard of value beyond man, and I know of nothing about man more valuable than his capacity to reflect about how his life should be lived, and to act on the conclusion of those reflections. This is a simple rendering of what Kant meant when he called man a moral being and defined moral nature as free and rational. Each person´s judgment finally is his own - there can be no conclusions about truth or right for him unless he attains conviction about them. From this follows a sense of responsibility. A man is responsible for his own judgments because they express

his moral personality, the exercise of his rational capacities - they are his. He is responsible for what he becomes because he chooses a conception of the good and lives according to it. And he is responsible for what he does in the world and for what he does to others, because the life he chooses and is responsible for is lived in the world among other people. And so in this most basic way we are separate, even lonely beings, choosing alone and responsible as we choose.

Kant speaks of rational beings, but the only moral beings we know are embodied human beings, with physical needs and capacities which correlate with or support the moral capacities. The happiness a man or woman seeks is the happiness which can be attained by a rational animal. The biological individual is the atomic unit of this system, just as the discrete consciousness is the atomic unit of the more abstract system of rational morality. Liberty is just the recognition of this moral status of flesh and blood men and women - each separate, each responsible to himself for the judgments he makes and the life he chooses.

Collective claims are a threat to this conception of liberty. Collective claims by definition are claims by the many, so that a lone individual lacks the brute power to resist them. At worst they can overwhelm an individual´s efforts to live his life according to his judgment and choices. At best he may be fortunate enough to belong to a coalition which shares his vision. Most likely he will have to compromise that vision in order to be part of a successful coalition. Now the prospect of joining successful coalitions improves one´s prospects of attaining his goals, but only by imposing those goals on others - in Kant´s phrase, by using others as a means to one´s ends. Thus coalitions are compromises in which we are partially overwhelmed and partially seek to overwhelm others. The principle of liberty resolves these conflicts by allowing each individual to choose his own life, neither imposed on nor imposing on others. Cooperation in achieving one´s ends must be completely voluntary on the part of each collaborator. The nonimposition or conservative aspect of liberty is expressed in the law of torts; the cooperative or creative aspect in the law of contracts.

Now may not an individual´s effort to live his life and realize his conception of the good be overwhelmed as well by natural circumstances as by the claims of others, and does it not hurt as much one way as the other? Implicit in my position is the assumption that it hurts more and in a different way to be consumed by the state than to be consumed by a tiger. This assumption is captured by the usual definition of coercion as subjection to the power of another person, and liberty as the absence of coercion. The law of gravity does not coerce us. Coercion and its correlative concept, liberty, define a relation between persons, a relation specified by the principles persons may be taken to adopt to justify the way they treat each other. Thus, for instance, a claim to compel contribution of a man´s above-average talents implies an appeal to the principle that talents are social property. The tiger, however, takes no stand on the moral worth or defining characteristics of the person he devours. His action has no maxim. And thus a man´s liberty is not threatened - though his welfare may be - by drought, sickness, or the prevalence of tigers. A thief, extortionist, or demagogue threatens liberty directly.

3. I shall now offer as simple a ground for redistribution - for the claim that others make on us for our sharing and sacrifice. As other persons share our moral worth, as they have lives which it is of consuming importance to them that they be able to live out according to their conception, as they are for this reason beyond price (or in an older language, as they have souls), it is wrong for us to be indifferent to them. The success and happiness of my fellow men and women cannot be indifferent to me unless I would deny the moral worth of my own projects and my freedom to pursue them. Human misery is but the helpless sense that what you value above all is slipping irreversibly out of your grasp. And to proclaim indifference as one´s principle in the face of the misery of others is inconsistent with proclaiming the moral worth of one´s

own happiness. This is, I think, the true meaning of Kant´s argument for a duty of beneficence in the fourth example in the Grundlegung.4 And just as law and a formal system of property are the ways in which the principles of liberty and of man´s imperium over nature find expression in the world, so systems of taxation and redistribution express in modern social circumstances a part of our duty of beneficence, a part of our duty to have regard for the misery of others.

What standard of distribution does this conception entail? In Right and Wrong I sought to avoid that question by positing a generalized right to a fair share of the community´s scarce resources, leaving open what the specification of that share should be: an equal share, or a maximin share, or one where the top did not exceed some multiple of the bottom, or the bottom did not fall below some fraction of the mean.5 My concern there was to argue what the fair share was a share of: scarce resources as measured by their market price, that is, demand for them relative to their scarcity. I was anxious to exclude persons, their talents and abilities from that levy. I sought to make plain that the attempt was not to arrive at some fair proportion (equality or whatever) in the attainment of relative happiness, which rather it was the individual´s responsibility to construct out of his fair share. Nor yet was it an attempt to satisfy needs as distinguished from wants, since these might be met by what I called the insurance principle, the premiums being paid out of one´s fair share. But I doubt today if one can separate the incidence of the tax, what it falls on, from its rate, how progressive it should be. What the rate is will make a crucial difference to how significant an intrusion is authorized. A low rate levied on everything, even talents, would be much less significant than a high rate levied just on external goods.

4 Foundations of the Metaphysics of Morals 41[424], Lewis White Beck trans. (Indianapolis, 1959).

5 Supra, note 1, at chs. 5, 6.

And the concept of fair shares itself implies a prima facie standard of distribution: equality.

In my earlier treatment I sought to avoid this inevitable pressure toward equality by limiting the incidence of the tax to external resources, thus affirmatively excluding a man´s earning potential. This point was further made by following Rawls in suggesting that the tax be a consumption tax. Of course one may ask whether leisure should be classed as a form of consumption, or whether working at an occupation returning less than one´s highest available wage is a form of leisure, or whether and when personal services constitute a form of leisure/consumption. And thus arise again all the questions about measuring a man´s share of scarce resources versus his own personal characteristics. The talk of fair share of scarce (external) resources assumes a picture, as Robert Nozick has argued, of people coming upon a stock of goods (manna from heaven) which must be shared out - fairly.6 But as Nozick´s own discussion demonstrated, the stock of goods in the actual world is so inextricably bound up with the efforts of those who have identified or transformed them that there seems little hope of finding a way of separating out what portion of a man´s income or consumption is attributable to his efforts and what to the unworked store of natural external goods.

The same conclusion of hopelessness is forced on us from another angle. If equality is indeed the standard - perhaps only equality of external scarce resources - how are we to avoid constant interference? For as men and women of varying talents, dispositions, interests, and luck consumed, wasted, gave away, exchanged, ventured, and invested their resources and efforts significant departures from equality would soon arise. These departures would either have to be tolerated and equality therefore abandoned or frequent corrections made, which would, however, render impossible the investment and effort which the conception

6 Supra, note 2, at 198, 219.

of liberty and responsibility for one´s own life requires. Ronald Dworkin has argued recently that: (1) Equality is compatible with the kind of ex post inequality which results from the varying outcomes of individual risk-taking, provided there was equality initially; and (2) bad luck in risk-taking and bad luck in failing to have talents and personal characteristics which command a high rent in a particular social situation may both be assimilated to the bad luck of handicap or illness. As we can imagine individuals insuring against the bad luck of illness so also they may be imagined as insuring against bad luck in investments and more strikingly against the bad luck of being less talented than the mean.7 Thus is ex post inequality squared with ex ante equality. The proposal is ingenious since it allows a considerable degree of actual inequality while proclaiming continued allegiance to the standard of equality which fair shares seems to require. (In this it resembles Rawls´s maximin.) Unfortunately the application of the standard to actual social circumstances must proceed by way of a rather elaborate thought experiment about what level of premium individuals ignorant of their own economic rent would be willing to pay to obtain up to what level of insurance. As in the case of Rawls´s original position, this thought experiment may be used to conceal or to display substantive moral judgments necessary to its operation. And again as with the original position the question arises, once the substantive moral judgments have been displayed, whether these would not have sufficed to answer our practical moral concerns directly.

I would prefer to return directly to the Kantian basis for distribution which I have announced: a sympathetic concern by each for each, a moral imperative to avoid indifference in the face of a fellow man´s misery- as I have defined misery. This basis (unlike talk of fair shares) in no way suggests equality as a prima facie

7 “What Is Equality? Part 2: Equality of Resources,” 10 Philosophy and Public Affairs 283 (1981).

standard of distribution. Indeed distribution and fair shares are themselves not implied by it, for there is no commitment to a picture of things being shared out at all. My basis is quite compatible with the argument that there is no way of distinguishing between external good and those which have been transformed by individual and collective effort and appropriation. There is no need to distinguish between good luck in investments and good luck in the economic rent of one´s talents and characteristics. It is the need, the misery itself which makes its claim, and that misery and the steps toward its alleviation can be evaluated from the actual situation in which we find ourselves. There is no necessity of appealing to any hypothetical lotteries, auctions, contracts, or insurance policies.

4. If the claims of others are to accommodate liberty, those claims must be clearly defined and limited, and those limits must leave the individual a significant private sphere to live his life. Does the basis for redistribution I have just identified - sympathetic response to the misery of another - do more than announce a sentimentality which lacks any definite promise of substantive help or for that matter any limits to the claim for help? I shall suggest that it may do both to a sufficient degree. And by the way my proposal should explain as well why a proper conception of liberty and distributive justice, given suitable background conditions, requires no demonstration that the original historical acquisition of holdings was itself legitimate. It is sufficient that we have a just system of private law - that is, one which recognizes property and personal rights in the law of tort and contract - a democratic government, and appropriate redistributive mechanisms. What is a man due? I would say these things: First, so much of the community´s resources that he has a chance to live decently and to make a life for himself - by his own efforts if possible, by the community´s aid if those efforts are insufficient. Beyond that he should demand nothing. Beyond that to use political power to

demand more is to violate the liberty of his fellows. How and in terms of what is this social minimum to be determined?

This is my proposal: A person has a claim on his fellows to a standard package of basic or essential goods — housing, education, health care, food: i.e., the social (or decent) minimum - if by reasonable efforts be cannot earn enough to procure this minimum for himself.

To fix the level of enjoyment of the components of the social minimum, suppose: a state which for a substantial period has enjoyed normal democratic institutions: free expression, legislatures, elections - in short, a standard liberal democratic regime. In that state certain levels of essential goods and services will have been achieved. A long period of wide political participation and reasonably regular and stable institutions are hardly conceivable otherwise. This general distribution of basic goods is both a necessary condition for and an inevitable result of a reasonably long tenure of the kinds of institutions I am positing. Of course this leaves much open about the details of the pattern of distribution. There may be a small group with very large wealth. The lower income classes, say some unskilled workers, may exist in conditions of hardship and uncertainty. Also, there may be pockets of real misery. These may implicate the victims of past or continuing practices which under any conception of democratic theory must be taken as unjust. Or they may be the victims of dislocations, misfortunes, or circumstances not created by the government itself. Examples would be workers in sectors depressed by changing economic circumstances, recent immigrants lacking necessary skills, or the victims of natural calamities. Finally, there are neglected groups which while not actively exploited are less able to compete. Examples are the handicapped or aged. Together these segments may constitute a substantial underclass - perhaps as much as a fifth of the population. Although I do not insist on the precise proportion, I do assert that the proportion cannot be much larger on the assumption that the society has enjoyed liberal democratic institutions for a substantial period of time and has not suffered from war or natural calamity. The reason is that such institutions cannot survive unless a large segment of the population sees themselves as living in at least tolerable circumstances. Moreover, those institutions would tend to deliver to the remaining 80 percent at least some acceptable package of benefits and long-run expectations for the same reason. To fail to do so would create political opportunities which would, in the long run I have posited, certainly have been exploited.

The point of this hypothesis is that it permits a criterion by which to determine the level of the social minimum. Take the educational opportunities, health benefits, housing standards enjoyed by unionized unskilled workers, which I shall call the reference group. I choose this category since, as they are lacking skills, they are unlikely to be in a strong position to demand a market rent, but being unionized they will be able to assure themselves against systematic exploitation and will have an organizational structure sufficient to define long-term goals and to discipline its membership to persevere in their pursuit. The social minimum is whatever package of essential opportunities, goods, and services (as I shall explain the term essential) is enjoyed by the reference group. Note that this does not state that the social minimum is whatever income is enjoyed by this reference group, since presumably most members of that group will have further income and engage in further consumption of goods and services not deemed essential.

This standard is lower than the average consumption of those essential goods for two reasons: (1) To allow substantial play for the differentiating effects of individual choice and market forces, that is to avoid the kind of constant interference, regimentation, and aggressive leveling down that a more exigent standard of need would entail. (2) To recognize the fact that in respect to most of these goods there is a discretionary or non-essential component which is hard to separate out. This last is most obvious in respect to such goods as housing and diet, but applies as well to education, health, and perhaps legal services. In any event in the case of the United States the implementation of this standard would have a significant effect on the welfare of the most disadvantaged groups.

5. The reference group´s consumption of essential goods is a function of the distribution of wealth in the particular society. If I am right, liberal democratic institutions will guarantee a certain substantial share to a group defined as I have. It may be objected that these reference levels change as the society changes, so that what constitutes justice at one time would be deemed unjust at another. If we had cut into the process in 1931 to make all the adjustments needed to insure a social minimum - as we probably did not - then the subsequent history of the system would have been different and so the adjustments required by the general criterion would be different in 1981 from what they are now in the absence of such earlier adjustments. In fact the objection might be pushed further. Since liberal democratic societies emerged from illiberal and unjust ones, often by at least partially evolutionary processes rather than by the sharp break of a total revolution, the citizens embark upon liberal democracy freighted with tastes and preferences found in an earlier distributionally and politically unhallowed past. Thus even if my distributional criterion were enacted at once as a part of the institution of democracy, still both the initial distribution and all its successors would seem to be tainted. There seems to be a kind of original sin which stains all subsequent attempts at virtue. I believe this objection is irrelevant.8 All I must show is that if the demands of distributive justice - the social minimum - had been implemented at some time in the past, say fifty or seventy-five years ago, what the social minimum requires today

8 For valuable general arguments for the proposition that an ancient wrong, if not persisted in or renewed, has a relatively short half-life, see Nozick, supra, note 2, at 152-53; Scher, “Ancient Wrongs and Modern Rights,” 19 Philosophy and Public Affairs 3 (1981).

would not be so different that my criterion for the social minimum must fall as circular. I concede right off that had a social minimum been instituted earlier many people would have been spared deprivations along the way. My argument is for a standard of just distribution now. It does not pretend that there have not been in the past and are not now unjust deprivations. Nor is there any reason to believe that the lowest group systematically able today to plan and provide for its essential goods (the reference group) would today enjoy a higher level of essential goods had the economic groups beneath it been assured a long time ago of these same essential goods - i.e., had the social minimum been assured in prior generations. If, for instance, the tax burdens necessary to provide the social minimum for all at an earlier date had lowered the rate of capital investment and entrepreneurial initiative of higher income groups - a point I do not here argue - then it might be that the income and therefore the level of spending for essential goods by the reference group today might actually be lower in a society which had been juster earlier, and the social minimum accordingly less exigent.

6. There is another similar objection to my proposal, and in answering it I can point to a general theme in my defense of a conception of liberty: the extent to which the standards which specify that conception can be and have to be independent of history, of culture and of politics. Since liberty is an ideal which claims to judge history, culture, and politics, it would seem to be undermined by any dependence on them, and yet some dependence seems inevitable. I propose a definite but limited and determinate standard of distribution: definite, because indifference to others is an offense to moral humanity; limited, because liberty requires a determinate domain of self-determination, free of the claims of others. I propose to measure the claim to redistribution by the level of essential goods enjoyed by unskilled, unionized workers. But it is the case that the situation of the reference class at any particular time is in part the product of politics, including redis-

tributive politics: Is free medical care available; is there subsidized housing; what is the level of spending on public education? In an egalitarian social democracy the situation of the reference group may be different from what it would be in a country with a more laissez-faire temperament. Yet — and now the objection closes in - my proposal is supposed to judge, rectify, and limit policies of redistribution by reference to a standard which must now be conceded to depend on the very variable and controversial policies it seeks to judge. If politics have won the reference group a certain level of mandated benefits at the moment we cut into the process, why should that level be frozen thenceforward as the absolute measure and limit of what it is just to accomplish by politics? In short, the objection states that by seeking determinacy I have attained only arbitrariness.

The objection misses the point, first technically and second in its basic conception. Technically, it must be remembered that I call for a redistributive floor not in terms of the total income of the reference class, but only in terms of their enjoyment of essential goods. The group I have designated is unlikely in any developed economy to rest content with consumption merely of essentials. Some important part of the group´s activism, therefore, would have been directed at raising its level of discretionary (non-essential) income. Given the group´s numerical importance their concern with discretionary income must necessarily impose a discipline on the demands for provision of essential services. Any increase in the consumption of essential goods by so significant a group will necessarily withdraw a large quantum of resources from the discretionary realm, and it is implausible that such a large burden could be shifted entirely to other groups. Thus, though it must be conceded that the reference group´s overall situation may vary considerably from time to time and from society to society, and that those variations will depend on politics, nevertheless this group´s enjoyment of essential goods may be said to approximate what a particular society at a particular state of development and prosperity considers a decent minimum.

7. I turn to a subtler, more insidious objection. First, it asks where do I get my list of essential goods from anyway. Second, it argues that the conception of what things are worthwhile, not to say essential, and a fortiori the decision of how essential these things are, how much to spend on them is internal to a socioeconomic system and therefore cannot provide a criterion for the justice of fundamental features of a social system. In a capitalist society, it is said for instance, a private space, a home, distinctive clothing, and other items of personal consumption loom far larger than they would in a more egalitarian, socialist society which would value more highly attainments and opportunities to contribute to the common good. That´s the story at any rate. But why should we believe it? Everything turns on the possibility of identifying the contents (rather than the level) of the social minimum with sufficient confidence that the designation of its contents would survive the kind of social transformation which might accompany the shift from a political democracy to a completely just society. The skepticism on this score is one of the more plausible applications of the Kuhnian argument that conceptual terms do not retain their meaning across basic theories. Though philosophers of science dispute the validity and even the coherence of Kuhn´s c laim,9 it has greatly impressed amateurs in law and the social sciences. Here the skeptical objection is clear enough: what constitutes a basic or essential or primary good is so obviously a cultural artifact that it cannot provide a criterion for judging alternative social arrangements. The essential goods may be thought to be the same as the primary goods, which Rawls defines as those things a man would

9 See, e.g., Hilary Putnam, Mathematics, Matter and Method - Philosophical Papers, vol. 1, 260 (Cambridge, 1975); Israel Scheffler, Science and Subjectivity (Indianapolis, 1967); Shapere, “ 'The Structure of Scientific Resolution,´ ” 73 Phil. Rev. 383 ( 1964).

want whatever else he wants: rights and liberties, powers and opportunities, income and wealth.10 If income and wealth are defined in terms of money alone, the socially dependent nature of the price system which gives wealth its value shows that it is necessary to be more concrete than he has been about the components of wealth.11 But I need only assume open democratic institutions and ask what material goods constitute a social minimum in such a society. The argument for the actual goods is then easy enough. I take education first.

Education is necessary to allow the members of the reference class to understand the functioning of the political system they participate in as citizens. Further, they must be able to obtain so much training as will enable them to take advantage of and compete for the opportunities available in an open economic system insofar as their natural abilities permit. And I suppose the entitlement extends to such education as will enable a person to participate in the culture of his society as well. The conditions of established democracy and of the political and economic awareness of the reference class explain the importance of this component of the social minimum: Education is just a precondition for intelligent choice, judgment about one´s circumstances - including the choice about whether to include further education in one´s vision of the good life. After all, the point of identifying the social minimum is to set the stage for individual free and fair choice. Health, housing, diet are only somewhat more problematical since their lack does not vitiate so definitively the idea of rational choice. Yet we must add them to the list. First, because rational choice is the rational choice of concrete human beings who require these things for the good functioning of their rational capacities. Second, because we are seeking to establish a reasonable measure for a

10 Supra, note 2, at 62. 11 See, e.g., Kennedy, “Cost Benefit Analysis of Entitlement Problems: A Critique,” 33 Stanford L. Rev. 387, 422–23 (1981).

fair share of scarce resources, and those things necessary to sustain life and vigorous function are an obvious minimum in any system which recognizes need and community as aspects of a fair share.

Now all this is vague enough, but not fatally so, and certainly does not fall under the strictures of Kuhnian solipsism. Further, it is quite immune to the issue of the potentially voracious claims of handicapped persons who can only approach these educational or other standards by dint of enormous expenditures. I do not say they do not present a problem, but they do not present one for this theory. The criterion is constructed by reference to the natural capacities of the average member of the reference group. That there are those who require more attention is not a problem of distributive justice between income groups. It is, as it were, a random chance which affects citizens or families of all income classes.

8. It is natural to ask whether those entitled to benefits under the provision for the social minimum might be entitled instead to claim the cash value of those benefits, spending that cash on whatever they want. Economic theory teaches that cash payments are superior to in-kind distributions, since they (by hypothesis) cost the donor no more while leaving the beneficiary with enlarged options. This theorem fails to hold only if we posit some kind of market failure, on which I do not rely, or if we adopt toward the beneficiaries a paternalistic attitude, which I find distasteful. Yet I do conclude that there is an entitlement only to the essential goods themselves, not their cash equivalent. (I leave open - indeed I prefer - the possibility of giving a kind of restricted cash to beneficiaries, which can be used to purchase, say, health care on the market.) The reason derives from the basis of the entitlement: a particular kind of human need (misery), which we are morally bound not to ignore. It is the fact, for instance, that another person is, say, suffering from appendicitis which could be cured by hospitalization but which if uncured would lead to a painful death, it is that fact and not the fact that this same person

is so poor that he can´t pay for the hospitalization which makes the claim upon us. If we imagine that we do give this person the money for an operation and he gives it to his lover to spend on liquor, the claim on us for help would still be there - we would not (and should not) allow him to die of a ruptured appendix and peritonitis. But this means that it is the medical care to which he is entitled and not anything else. I suppose he can decline to accept medical care; to force that on him would be paternalistic indeed. He has no right, however, to dictate the terms of the provision: if he does not wish medical care at all (or housing, or food, or other essential goods) then he just waives his claim on us altogether. This is not to say that under certain circumstances it may not be more practical and effective to give poor people, or certain classes of poor people, cash instead of goods or restricted money. That will depend on details of administration and politics. The principle, however, should be clear.

9. This brings me to the final question for this part and the bridge to the next: how important is it to a defense of liberty or to normative political philosophy generally that its points of reference themselves be fixed and neutral? My standard for a decent minimum seems to be neither. It is all mixed up with the changing political and economic arrangements of the subject society, and so seems ineligible as a standard to judge those arrangements. But it is a serious mistake to be disturbed by this. By demanding an impossible level of objectivity or neutrality such solidity as one is capable of is ignored or devalued. After all, the aim is to give meaning to the idea of a decent minimum so that it can stand as the compromise between our right to what we are and what we make and our duty to respond to the need and humanity of others. What is fixed, independent of, and can judge particular politics is just the wish to find such a compromise. So also the concept of a decent minimum as the further specification of that compromise is a concept independent of the particular politics it judges. The focus on a particular reference group in a democratic society and

11 0

on a particular part of that group´s situation or budget makes the standard sufficiently independent of the politics it judges to provide discipline and control in that judgment. What we have is a changing standard, but one disciplined by a prescribed procedure (liberal democracy) and a prescribed concern (the identification of a decent minimum of essential goods). More than this, I think, it is unnecessary to ask, and at any rate I believe it is impossible to get.

PART II. MEANING AND CAUSE

1. A social minimum has been assured. There is fair opportunity for all and even those who fail, whose luck is bad, need not fear catastrophe. Given those conditions, liberty requires that we accept the outcome of individual transactions as not only efficient, but fair. If the choices made in the market are freely made, there is no sufficient reason to reverse or mitigate their effects, and whether those choices have been freely made is not a political, certainly not a distributive, judgment. (After all, the claims of distributive justice have been sufficiently met through the provision of the social minimum.) Rather it is a judgment of principle, legal principle, principle about such things as when an assurance makes a binding contract, the obligation of good faith in carrying out a contractual arrangement, when a person is liable to compensate for harm which he has caused but did not intend, whether a harm which would have befallen anyway is caused bY an agent just because the agency would have been sufficient to cause the harm in other circumstances, whether someone who in good faith receives and spends money given him by mistake is bound to return it. It is a tenet of liberalism shared by Ronald Dworkin12 and positivists like H. L. A. Hart13 that legal disputes between citizens

12 Taking Rights Seriously, ch. 4 ( Cambridge, Mass., 1987 ed.). 13 The Concept of L aw, chs. 5, 7 (Oxford, 1961).

are at least in large part decided as a matter of right. The parties are entitled to a decision by virtue of the preexisting law which it is the duty of the judge to apply. The differences between Dworkin and the positivists break out only in what Dworkin calls hard cases and in what Hart calls the penumbra of settled rules, where discretion is required. Dworkin carries the duty of judges to decide disputes according to law all the way through, so that even in hard cases (in which the positivists concede to the judge the role of policy-maker) the judge must apply principles rather than make policy furthering the collective goals of the community as a whole. Judgment, balance, principle are applicable in all cases even those where the positivists claim a role for policy-making discretion. For Dworkin this exercise of judgment is the search for answers truly inhering in the legal and moral materials, principles the judge is bound to apply. It is crucial that this concept of law hold - either in its positivist or in its (to my mind richer and more coherent) Dworkinian version - if men are to be free, if liberty is to be possible. If judges were not (at least in large part) constrained by the law then the distinction between the right and the good would be gone, and all men would be servants of some authoritative view of collective policies and goals.

A liberal theory of law must accomplish at least these things:

• It must allow persons to make agreements and it must enforce those agreements when made.
• It must establish the boundaries around a person and his property so that incursions across those boundaries constitute compensable wrong.14

Tort law defines our rights. It draws the boundaries of what Hayek calls our private sphere,15 within which we make the life we choose. The social minimum assures both that this private

14 Cf. David Hume, A Treatise of Human Nature 526 (Selby-Bigge ed., Oxford, 1888).

I5 Friedrich Hayek, The Constitution of Liberty 21 (Chicago, 1960).

sphere will have certain minimum dimensions and that contributions to assure that minimum for others are themselves limited and fairly exacted. The notion of a private sphere is precisely correlated with the main aspects of liberty I set out in Part I: an individual´s opportunity to choose her own life according to her own judgment, and responsibility for the consequences of her choices. If the boundaries of that sphere are fixed by the judgment of others about what are good outcomes of personal choices, then the individual is deprived of responsibility for her choices and thus of liberty of choice. She does not choose and act on her conception of the good; it is chosen for her. By contrast, when natural forces impinge on an individual´s projects this does not derogate from her moral authority to determine those projects. It is only when the judgments of others impinge that liberty is at stake. Yet how can these boundaries be drawn other than as a result of a collective judgment? And, as we shall see, even so neutral-seeming a concept as cause, which may seem a natural candidate for identifying when another´s conduct imping es on me or mine, is freighted with normative content.

Contract law is another indispensable aspect of liberty. The worth of my liberty is critically diminished if I cannot treat across my boundary with willing others. One of the things which I certainly want to do with what is assured to my discretion is to give it to others or trade it. When others impede my gifts or exchanges they assert an authority not just over me but over my partner as well, an authority therefore doubly offensive to liberty. And promises are simply gifts or exchanges projected into the future. By recognizing and enforcing promises society enlarges liberty.16 By selectively refusing to enforce some promises because of how the promisors use their liberty, or because of who ends up better off because of them, or who the promisor is, the collectivity is again

16 See Immanuel Kant, The Metaphysical Elements of Justice, Ladd trans., 54—55 [248] (Indianapolis, 1965).

asserting an authority over individuals which is incompatible with their liberty. Of course only those obligations freely contracted warrant this recognition. Yet the notion of what is freely undertaken, and how the content of those undertakings can be understood, is controversial. What is fraud, what duress - these are normative questions. So also may be the fixing of the meaning of general terms.

Liberty requires conceptions of private law based on principles, not on episodic collective judgments of policy. But given these controversies about tort and contract, are such conceptions available ?

2. Contract requires a theory of language which assumes the stable and common identification of the subject matter of commitment. Given the controversies about the indeterminacy of translation, we see that even about present exchanges there may be problems.17 But how much worse are the problems for promising, which makes commitments into the future and thus requires a theory of the identity of singular objects over time in addition to a theory of the identity of reference by two persons referring simultaneously. There is the famous case of Rose II of Aberlone, the prize breeding cow sold off at the price of beef because her owner thought her sterile, while in fact she was not only fertile but pregnant and worth a hundred times that price.18 Or the case of the mysterious stone sold for a dollar which turned out to be an uncut diamond?19 Or the rooms rented for a day at fancy prices to view a coronation procession which never took place.20 Much ink has been spilled applying theories about the identity of objects to such cases and debating whether changed attributes imply a change

I7 Cf. W. V. Quine, Word and Object, ch. 2 (Cambridge, Mass., 1960). 18 Sherwood v. Walker, 66 Mich. 568, 33 N.W. 919 (1887). 19 Wood v. Boynton, 64 Wis. 265, 25 N.W. 42 (1885). 20 Griffith v. Brymer, 19 T. L. R. 434 (K. B. 1903) ; Krell v. Henry, 2 K. B. 740 (1903).

in subject matter.21 These cases have engendered metaphysical speculation which would delight Saul Kripke22 and Hilary Putnam:23 Does not Rose II of Aberlone designate a particular cow irrespective of mistaken attributes?

Even such problems as that may be dismissed as of no practical significance. What cannot be so easily dismissed, however, is the problem of general terms. Both H. L. A. Hart24 and John Austin before him25 noted that law — by practical necessity — must use general terms lest it be relegated to a system of simple singular commands. The same is, of course, true of promises. To accomplish regimes of coordination and cooperation of any complexity, promises/contracts must be cast in general terms as well. And that's where the difficulties begin. They break out in problems of interpretation, of course, but most obviously in the cases collected under the rubrics of mistake, impossibility, and frustration.

Whether this issue is one of the designation of singular terms or the meaning of general terms, for the purposes of contract law the question should be: had the parties come to an actual agreement about the relevant risks such that the promise principle is properly invoked? And this is a question at least in principle susceptible of an answer. The appropriate inquiry must relate to the intentions of the parties rather than the canons of fairness that take over when it appears that the parties had no convergent intention in the premises. Agreement - convergent substantive intentions (about risks and values of stones and cows) embodied in

21 For reviews of the literature, see Bronaugh, “The Place of Identity in Contract Formation,” 18 U. W. Ontario L. Rev. 185 (1980); Francis de Zulueta, The Roman Law of Sale: Introduction and Select Texts 28—30 (Oxford, 1945).

22 Saul Kripke, Naming and Necessity (Cambridge, Mass., 1980); “Identity and Necessity” in Stephen Schwartz, ed., Naming, Necessity and Natural Kinds (Ithaca, NY., 1977).

23 “Meaning and Reference,” in Schwartz, s upra. 24 Supra, note 13, at 2 1. 25 The Province of Jurisprudence Determined, Lecture I (H. L. A. Hart ed., London, 1954).

the intentional invocation of the promissory form - is one category, and fairness in picking up the pieces after it becomes evident that the parties had sailed out together in a leaky boat (a defective agreement) quite another. I make this distinction for the sake of liberty. If the law's, the courts' substantive grounds of fairness threaten to come in each time to tell what the agreement is, then it is the judges who make the contract and not the parties.

3. There is an argument for the proposition that the courts do in fact and inevitably make contracts for the parties which has gained currency among some legal writers. The argument takes as its point of departure the philosophical problem about general terms. General terms, as the argument goes, are attempts to sweep a large set of particulars under a single conceptual rug; but it is an attempt doomed to failure. General terms have no objective correlates; they are merely compendia of particulars collected together for our convenience in terms of similarities we note for our purposes. To say that general terms have no objective reality is to say first that general terms do not of themselves identify the particulars subsumed under them, and second that therefore the process of subsumption is a value-laden process, one which refers to human goals and purposes. Since agreements which work into the future are cast by a practical necessity in general terms, then the process of interpreting the agreement is a value-laden process, not an impartial search for the intention of the parties.26 And therefore the courts cannot escape making contracts for the parties and liberty is pro tanto impossible. This argument might be refined somewhat - perhaps along these lines: Fried says that a contract is really about a set of risks - about the value of a stone, a cow, or a view from a window. And, the argument continues, as soon as the dispute relates to matters which have not just precisely and

26 See Grant Gilmore, The Death of Contract (Columbus, Ohio, 1974; Morton Horwitz, The Transformation of American Law, 184—85 (Cambridge, Mass., 1966); Kennedy, “Form and Substance in Private Law Adjudication,” 89 Harv. L. Rev, 1685 (1965).

explicitly been included in that set, as soon as one must extrapolate, as soon as one must ask the meaning or the intention of that set, one has been cut loose from the solid shore of the parties' intention and is adrift on a sea of indeterminacy.27 And in that case the court may as well impose its values, since someone must.

Exactly the same point comes up in the case of good faith. In the recent case of Fortune v. National Cash Register,28 Fortune had for over twenty years worked for NCR under a contract allowing him to quit on sixty days´ notice and to be fired on ninety days´ notice. His employer exercised this option shortly after Fortune procured a five-million-dollar order for NCR, an order which would have required the payment of large bonuses to Fortune over several years if he had been kept on. The court balked at this bad faith on the employer´s part. The skeptical objection to the liberal conception of contract would say that here plainly the parties had not thought of just this eventuality. Nor will the objection allow the court´s solution to be described as the determination of the extension of the general conception of the arrangement by an intuition of its intention, i.e., its meaning, as one might say that a man who used the term “all the primes between 0 and 1000” meant to include 719 though that number had not entered his head. Not a bit of it - the critic insists that the court is simply imposing its conception of fairness on the parties.

Now one will recognize here analogues to Wittgenstein´s discussion of “continuing the series.”29 It is that discussion which sinks this skeptical argument. For we have learned that an account of our capacity to understand general terms, to follow rules, is not like the description of a train moving along tracks - there is not some necessity wholly apart from our purposes, apart from the

27 Cf. Kennedy, “Legal Formality,” 2 J. Legal Studies 351 (1973). 28 373 Mass. 96, 364 N. E. 2d 1251 (1977). 29 Philosophical Investigations, §§136—242 (Anscombe trans., Oxford, 1953). See also Crispin Wright, Wittgenstein on the Foundations of Mathematics, ch. 2 (Cambridge, Mass., 1980).

way we are which draws us on. The referents of general terms are not objectively, metaphysically real (to use Hilary Putnam´s term)30 in that sense. Indeed, once we confront the idea squarely one doesn´t even know what that kind of objectivity would look like if it were the correct description. Yet rules and the referents of general terms are quite objective enough - they are (to paraphrase Putnam again) internally real. They are real - they draw us on — inside the world of purposes and experiences we inhabit, the only world we know. The skeptical critic may pounce on this as proving his supposed point - for how, he will ask, can we be sure we inhabit the same worlds: promisors, promisees and judges ? But it is too obvious that if we did not inhabit the same world communication would be impossible, including the communication of the skeptic´s proposals and refutations. To be sure, we do not have the same purposes: trivially, because my purposes are always mine and so at best our purposes run parallel to each other; more substantively, because our purposes often pull in quite different directions. But there is no more problem due to this than there is due to the evident fact that our perceptions also - trivially and substantively - are not the same. Our successful communications show that we share purposes and perceptions to a sufficient degree. That they also diverge gives us something to talk about.

One might ask how the judge at one and the same time can share the values and perspective of the contracting parties so that she is able to interpret their will and yet stand ready to disagree with what the parties have accomplished, so that she leaves them enough space to do what they want, not what she wants. The answer that common experience proposes distinguishes between that sharing of goals which is necessary to make general terms commonly comprehensible and the closer sharing which comes when we are in agreement on particular goals. Wittgenstein´s

30 Reason, Truth and History (Cambridge, England, 1981); “Realism and Reason,” in Meaning and the Moral Sciences (London, 1978).

11 8

talk of “forms of life” and Putnam´s insistence that general terms—natural kinds—are relative to our needs and ends31 all point to this same phenomenon which common sense too identifies: To understand each other we must at least be able to imagine what it would be like to have the other person´s goals. And the practice of sociology, anthropology, philology, far from supporting the crude skepticism of the cultural relativist is wholly encouraging on this score, as it shows how far we can indeed go in making transparent what seemed opaque. The possibilities of understanding demonstrated by these disciplines - as by the evident fact of the translatability of languages quite remote from our own in time and cultural distance - are remarkable.

It makes sense, I think, to speak of our common humanity as largely coextensive with the kind of shared form of life necessary to attain understanding. This is not to say that we always understand each other at once and without a sometimes considerable imaginative effort. And of course we run the risk that we not only will understand initially alien conceptions but come to share them. Similarly our understanding of another person´s goals may convince us that we have a better grasp of some of the entailments of those goals than their original proponent. This is the kernel of truth in the objections we have been considering. This is the idea which Fuller captured in the phrase that the common law works itself pure.32 But it is pure sophistry to conclude from this that therefore it is the judge´s goals which every time control after all, when it is the task of a judge to determine what the goals and commitments of others may have been. And it is a further kernel of truth that contracting parties (or any users of general language) often seek to enlist the cooperation of a third party in drawing out the entailments of their own arrangements. In such cases they may also want the third party to use his own particular ends and values in performing

31 Supra, notes 23 and 30. 32 Lon L. Fuller, The Law in Quest of Itself (Chicago, 1940).

this operation. All this is true, but it still does not lead to the conclusion that there is not a distinct operation in which it is the other person´s goals we are trying to understand. And so as to understanding and interpretation, we might say that the shared goals - way of life - are the ground which makes diverse goals perspicuous to us.

And so the doubt that promisors, promisees, and judges can determine general intentions which are transparent all around is just another case of sophomoric general skepticism, and not worth agonizing over. It is worth seeing that the form of this answer to the skeptical objection is related to what I argued earlier about the standard of distributive justice. There too we seemed to be stymied by the demand for a wholly objective measure of fair shares, until we saw that a standard could be defined as what emerged from the common experience of democratic institutions, and moreover that it was perfectly all right for that standard to change, that such changeability did not imply a fatal subjectivity in the standard. The standpoint of internal realism (as opposed to metaphysical realism), realistic though it is, does not promise immobility, a fixed right answer, from the perspective of some universal and uninvolved observer. And yet to the involved observer every difference of judgment is the occasion for argument in which each participant acknowledges the possibility of his or his interlocuter´s (or both) transcending a particular point of view and coming to a new truth.

4. A theory of gifts and exchanges is useless, however, without a theory of what is ours to give, what are our interests and against what they are protected. We need a theory of tort law. Two recent authors greatly concerned with liberty, Robert Nozick and Richard Epstein, see the matter this way: There is a system of property rights, including of course rights in the integrity of our own physical person. Anyone who impinges on those rights is prima facie liable to compensate the aggrieved party. Nozick char-

acterizes the impingements as boundary crossings.33 For Epstein the prima facie rule is one of strict liability: If you cause harm to another´s protected interest, you must pay.34 Thus the topological paradigm of a personal boundary to be secured against all incursions is carried forward. But of course the concept of cause must now bear a heavy burden in the account. If a man. stores a dangerous substance on his land and it escapes without his further intervention is he or is he not the cause of the harm?35 Is a person who remains or allows his property to remain in the way of a potentially damaging force a contributing cause? Epstein speaks of a person as being a cause when he releases a force which brings about the harm, or when he stores up a force which brings about the harm as a consequence of the application of some slight additional force.36 And, following Hart and Honore,37 he holds further that something is a cause when it is one of a jointly sufficient set of conditions for that result, and the result comes about without the intervention either of a voluntary human agency or some unusual circumstance? This conception of cause they claim is embedded in both ordinary usage and the common law.

One sees why Epstein puts it this way. The picture is of a status quo, a base line which the actor disturbs. Behind his idea is Nozick´s, Kant´s, Locke´s conception of a private domain which defines the individual´s discretionary space, within which he can work out his conception of happiness. It is the state´s, the law´s, duty to protect this privacy. Epstein´s conception of causality hopes to permit the simplest possible conception of boundary

33 Supra, note 2, at 57. 34 The corpus of Epstein´s writing is large. His views are summarized in a recent essay, “Nuisance Law: Corrective Justice and Its Utilitarian Constraints” (hereafter “Nuisance Law”), 8 J. Legal Studies 49 ( 1979). The primary article is “A Theory of Strict Liability,” 2 J. Legal Studies 151 ( 1973).

35 Epstein, “Nuisance Law,” supra, at 66-68. 36 “A Theory of Strict Liability,” supra, note 34, at 166-68, 177-78. 37 H. L. A. Hart and A. M. Honore, Causation in the Law (Oxford, 1959). 38 “Nuisance Law,” supra, note 34, at 56-60; “A Theory of Strict Liability,” supra, note 34, at 160–89.

crossing, and thus to implicate the sparest possible system of judgments in determining whether rights have been infringed. If they have been, then compensation is due. In other words, compensation is not compelled, rights protected for forward-looking reasons of the community, but rather just in recognition of those rights. This is what is classically called corrective justice,39 contrasted to distributive justice or to the pursuit of other policy goals of the community.

Thus the causal theory may be seen as seeking to do for tort law what I have tried to do for contract law — which is to show how legal rules and institutions do (or with some changes, can be made to) recognize individual rights by carrying out the determination of private wills. Epstein´s picture, which I linger over because it is so intuitively powerful, is intended to assure that tort law is moved not by considerations of economic efficiency, or distributive concerns, or any policy, but just by a concern to vindicate the rights of the parties to the dispute. Strict liability plus causality appear to do just that. Did the defendant´s act invade the plaintiff´s right ? If it did, the associated harm must be compensated.

5. Unfortunately, causation is a normative, not to say a controversial, concept. Hart and Honore´s proposal, for instance, depends on the absence of “unusual” intervening events. And that is a matter of evaluation. For Epstein the case is much worse, since for him (though not for Hart and Honore) causing harm is prima facie tortious. Would anything a man does which enters into a set of conditions sufficient to bring about the harm be culpable? Would planting a tree on your lawn which twenty years later is blown onto a passing car be culpable? Not so, Epstein would respond: a cause, as opposed to a condition, is active; it is the application of force, or the storing of force easily released. If defendant has walked across plaintiff´s land trampling his flower

39 See Aristotle, Nicomachean Ethics, Bk. V, iv–v; Epstein, “Nuisance Law,” supra, at 50. For a different view, see Richard Posner, The Economics of Justice, at 73–74 (Cambridge, Mass., 195 1).

1 22

garden, the causal paradigm seems unproblematic, but if that is the easy case, hard cases unfortunately are the rule.

First, every case with a moving plaintiff and a moving defendant (the typical road accident which makes up the vast bulk of tort cases) involves mutual impingement.

Second, it runs counter to at least 150 years of law to make the question of liability depend on whether the defendant was moving and the plaintiff still regardless whether the defendant made reasonable efforts to consider the plaintiff´s interest.40 Imagine the case of the careful motorist who injures plaintiff pedestrian as the result of a skid on an unnoticeable patch of ice.

Third, every effect is the resultant not of some single, isolatable factor but of an infinite manifold of circumstances, which is only carved down into discrete conceptual packets for an explanatory purpose. In the case of the skidding motorist one description of the event might focus on the force originating in the kinetic energy of the automobile´s engine. And to push inquiry back beyond this to the action of the motorist in turning the ignition key but not further still to the conduct of those who assembled the car or refined the gasoline is a selective judgment based on our explanatory or other social purpose.41 Similarly, the focus on forces is at times inappropriate. If the icy patch had been intentionally or negligently placed in the roadway, or (to modify an example of Epstein´s) if a defendant walking along an unlit corridor steps on a roller skate left there and blunders into plaintiff, whose arm he breaks, the force is all the defendant´s, but the cause and blame is his who created the dangerous condition. Epstein holds a man to be the cause of harm if he stores up potential energy in such a way that a small extraneous force will cause its harmful release. And so it is with conditions generally. The presence of oxygen in the air is not the cause of a fire set in the

40 See, e.g., Brown v. Kendall, 60 Mass. (6 Cush.) 291 (1850. 41 Hart and Honore, supra, note 37, at ch. 2.

open by defendant which spreads to plaintiff´s property. Neither is a light breeze which comes up after the fire is started. If, however, an unusual wind spreads the fire to plaintiff´s property, or oxygen is introduced into a chemical reaction begun by defendant which in the absence of oxygen would have progressed slowly and with no harm to anyone, then these intervening events may well usurp the sobriquet of cause.42

Fourth, at the other end of the physical process, what counts as harm is hardly a matter to be settled on purely physical or topological grounds: whether a force was imparted across a designated line.

Epstein sees heat, sound, concussion waves from explosions, odors and gases as instances of invading objects or transmitted energy.43 Here his physics is beyond reproach but misses the point that light may come in the form of a destructive laser beam or as a spotlight on an outdoor movie screen44 or as the dim but decipherable bearer of an offensive image. Sound may be shattering, or merely discordant, or the bearer of a disturbing message. Whether or not the intrusion is harmful will be a function of factors other than the physical magnitude of the force. Thus, for instance, a modicum of noise or light introduced onto another´s property will not be actionable, but the very idea that the house next to mine is being used as a funeral parlor might be. (Epstein says that the funeral parlor-as-nuisance cases are wrong because ideas are not physical forces.45 But images are, and they carry the idea which turns out to be noxious. A discordant noise becomes noxious because of its form, not because of its acoustical pressure, yet here Epstein does see a possible nuisance.) Indeed it is not clear why only positive intrusions should be actionable. Corre-

42 See, Hart and Honore, at 31–38. 43 “Nuisance Law,” supra, note 34, at 60. 44 See Amphitheaters, Inc. Portland Meadows, 184 Ore. 336, 198 P.2d 847 (1948).

45 “Nuisance Law,” supra, note 34, at 64–65.

sponding to the light shone upon your movie screen would be the erection of a barrier blocking the sun from your bathing beach46 or greenhouse. Corresponding to the offensive images projected onto your property would be the erection of a barrier to a delightful view. Is it crucial whether you project a light or a shadow - or a combination of light and shadow forming an image? It should be said that Epstein resolutely hews to his physicalist line, finding the light actionable but not the shadow.47

Nor can one escape these difficulties by a more careful definition of the rights, the boundaries, in issue: Is there a right to “ancient lights,” to a view, to quiet? If so, then to interfere with that is tortious. We are then driven to ask why are these conceptual boundaries not normatively permeable to some but not to other forces - permeability being a function of probability of risk, motive, type of force. Perhaps, for instance, noise only crosses my boundary at nighttime, because the boundary, like Goodman´s grue emeralds, is defined in spatio-temporal terms. Or consider the right to one´s good name or to a purely residential neighborhood-free of funeral homes as neighbors. These considerations show that neither the concept of cause nor that of a spatial boundary which a defendant causes a force to transgress are as normatively neutral or even as coherent as one would wish for the purposes of a tort system which respects individual liberty.

6. The difficulties in Epstein´s conception of cause can be traced to his treatment of cause as an exclusively physical relation between events: the impact of a brick causing a window to shatter. As Mill pointed out, however, there is an infinity of conditions - negative and positive - which must concur for a given effect to obtain.48 Which of these is designated the cause seems not to be determined by something about the conditions themselves, nor yet

46 See, Fountainbleau Hotel Corp. v. Forty-Five-Twenty-Five, Inc., 114 So. 2d 35 (1959).

47 “Nuisance Law,” supra, note 34, at 61–62. 48 For a discussion of Mill´s views, see Hart and Honore, supra, note 36, at 12–21.

by the kind of physical relation in which a condition stands to the effect. Rather it is an explanatory relation. To use an example from Hart and Honore, different people may consider the speed of the train, the degree of the grade, or the displacement of a slightly bent rail as the cause of the train wreck, depending on their explanatory purposes.49 But explanation is not a physical relation.

On the other hand there is more reason to go along with the common-sense view that cause is a singular relation.50 Although scientific laws and regularities may stand behind the singular event (brick shattering window) we do not need to know them to know the cause. “Why did the window break?” “Because Johnny threw a brick through it.” This is a perfectly satisfactory explanation. Accounts in terms of the strength of the molecular bonds in glasses may be a further explanation, or an explanation of the explanation, but contrary to the covering law hypothesis51 it is not at all clear why we must have such further explanations before we are warranted in making the judgment that the brick shattered the glass52 One may venture the guess that it is this sense of cause being a singular relation that leads ordinary persons to think of it also as an exclusively physical relation. You feel it in your bones when the force of your body causes something to move, break, collapse, The very causal vocabulary is singular: cut, smash, break, slash, pierce, kill.

But why do we pass from the singularity of the causal relation to the conclusion that it is an exclusively physical, metaphysically real relation?” Because of our pervasive tendency to project our purposes and interests on the world in which we play them out. If you accept the conception of cause as explanation relative to a purpose or interest, and if certain purposes and interests are

49 Supra, note 41. 50 See, generally, G. E. M. Anscombe, Casuality and Determination (Cambridge, 1971); Davidson, “Causal Relations,” 64 J. Phil. 691 (1967).

51 Carl Hempel, “Aspects of Scientific Explanation,” in Aspects of Scientific Explanation and Other Essays in the Philosophy of Science (New York, 1965).

52 See Anscombe, supra, note 50 .

pervasive, constant, deep-rooted, such a projection of a singular relation on the outside world is altogether natural. It does not even seem wholly mistaken since after all we construct the world out of our experiences and interests. Causes are natural as kinds are natural — i.e., relative to the way we are in the world.53

We may, then, make an argument about causing boundary crossings analogous to the argument about the use of general terms to embody agreements. And thus we would put tort law on as secure a footing as contract law, and this part of the task of a theory of liberty would be accomplished. We must regretfully say goodbye to Epstein´s account in terms of imparting force or releasing forces. His account does no more than substitute a more general causal account for the more familiar particular causal accounts of breaking, kicking, turning on or off, without any gain in clarity or insight. It is just plausible to say that, prima facie, we wrong another when we kick him, break his arm or burn down his house. It is palpably implausible to utter Epstein´s generalization about the wrongness of doing harm by imparting or releasing forces. Either we stay at the level of concrete relations altogether or we move to quite a different set of categories.

7. As I have argued, Epstein´s account is defective in part because it misses the need to identify the boundaries which must not be crossed, the interests which may not be impinged upon. The boundaries of one´s personal space may be drawn variously and the lines need not be limited to the three dimensions - they may include time or risk and be permeable (ethically permeable) to some but not all intrusions. How are they to be drawn, then? An a priori answer seems beyond our grasp. Yet if the boundaries are drawn (as the law-and-economics, justice-as-efficiency analysis of tort law would have it) to accomplish extrinsic social ends,54

53 Cf. Putnam, Reason, Truth and History, supra, note 30; R. G. Collingwood, “On the So-Called Idea of Causation,” 38 Proceedings of the Aristotelian Society 85 (1938).

54 See, Epstein, “A Theory of Strict Liability,” supra, note 34, at 203–4; Charles Fried, Right and Wrong, ch. 4 (Cambridge, Mass., 1978).

then it is those ends and not the recognition of individual rights which stand behind the law of torts. Individual liberty is not secure if its contours depend on the convenience of others. Rights, it would then seem, are either natural, necessary, a priori, and so secure, and liberty secure with them, or merely conventional and our liberty at the mercy of collective judgments of policy. But this is just the same false dichotomy which we faced in respect to distribution and agreement. The boundaries of our persons and property cannot be anything other than conventional - in the sense that they are not given independently but are discerned in the context of human needs and concerns.

I propose that two sets of conventions (let us call them that for the moment) work together to identify wrongful invasions: conventions about what the boundaries are and conventions about when a boundary is wrongfully crossed. They work together, but they are not the same. (You could make them the same by defining the boundary and its permeability in highly abstract nontopological terms, so that for instance we might say that a nonnegligent entry onto a topologically defined space or an accidental blow to my person doesn´t cross the boundary of my normatively defined space while, say, a negligent intrusion or an intentional blow would.) I would not collapse the questions of boundary and of intrusion because our tort theory should capture the distinction between what´s mine or what´s me (boundaries) on one hand and what is a wrongful act on the other. Both are conventional. There is no natural way of dividing up moral space any more than there is of dividing up the objects in the world - yet the conventions go very deep. The boundaries of my physical person are as palpable and fundamental as the individuating segments of the manifold of experienced reality.55 As to objects and land, however, the conventions are often no firmer than the set of things we need and value in our activities, and so those conventions change with our interests.

55 Ibid., at ch. 2 .

All this must be pinned down and illustrated with specifics. At this point I want to refer to the parallel to the other component, the causal component of tort theory. Causal terms — break, burn, shatter - also embody our interests. They do so by reflecting the ways in which we intervene in natural processes to effect those interests. Indeed the very phrase “intervene in natural processes to . . .” is a studied equivocation, for it seems virtually impossible to discuss the interplay between ourselves and the world without invoking such causal terms or their analogues. An account of the world just in terms of sequenced sets of conditions would be impenetrably opaque. It could accommodate neither moral advice nor moral judgment. So why should we flee the greater richness of our causal vocabulary? Because it lacks neutrality vis-à-vis our goals and interests? But there is no reason to yearn for that.

And this points the parallel between the active (causal) and passive (boundary) portions of the tort question. The drive for neutrality comes, I am sure, from the same drive as the drive to define a stable and uniform criterion of distributive justice, and at a more general level to profess a theory of language and of truth which is secure and valid from a point of view outside of the world we occupy. The concepts of cause are palpable and real enough; they are just not metaphysically real. The crucial question is not - it never is - whether the boundaries of me and mine, whether the concept of cause are real in this impossible sense, but rather whether such contingency as they inevitably have makes them useless to define rights and wrongs and thus useless to serve in establishing the concept of liberty. As to cause, the concepts are stable or common enough. We can understand each other´s use of causal terms, just as we can understand each other´s use of general terms; we can translate. Translation is possible because of a common human nature. That human nature is not some third, deeper common language we both share and into which we translate our different languages when we understand each other. Just as we do not translate to this ghostly common language when we understand each others´ same surface language. Surface languages are all there are, and in understanding and translating we stay on that surface. So it is with cause. We may disagree about, change, or deepen our ideas about intervening in the world, about ourselves being active and producing results, but as we do, it is not because we are getting deeper in touch with some underlying causal notion: producing changes in the world. “Producing changes in the world” is a generalized causal notion which is only as good as the ways in which we produce changes in the world - by breaking, burning, pushing, ionizing, lasing - it is not itself a way of producing changes in the world.

So if in tort law we are interested in rightful and wrongful conduct, we are interested in rightful and wrongful ways in which people produce changes in the world. The causal aspect, then, of tort law is no simpler but no more complicated than that. Epstein´s mistake is to seek an explication of cause in terms of some deeper, unified idea of producing changes in the world as such. And as we saw, he came up with some rather bizarre but not always unilluminating notions about releasing stored forces and so on. If you do things my way, then there is no reason to shrink from saying that the person who casts shadows is a causal agent every bit as much as one who shines a light. Or that one who poisons a reputation is a causal agent just as one who poisons a well. True, our repertoire of causal notions will change as our repertoire of causal means changes. If we cease to believe in witchcraft, putting a hex on somebody ceases to be a way of effecting changes in the world, but now altering their genetic structure may be. Though there is change there is total translatability, which is all we need. The causal concepts are quite universal: no one can hex or be hexed, and anyone might change or be changed in respect to genetic structure.

8. An analogous approach is in order when it comes to drawing the boundaries of our rights: the bounds within which I may act even if I do harm you, and the bounds within which you are

secure from my incursions. Even the bounds of my person are conventional but relative to the most pervasive conceptions. More contingent is the question of negligent impingements on my person. How much attention do you have to pay to my potential presence in going about your projects? The answer there will be closely related not to the idea of the person, but to property. If I am driving along the highway then I am far more open to being hit by you as you drive on that highway (morally open, of course) than if I am rocking on my porch at home. We are both going about our business on common property, in a common area. To ask me to be too attentive to the risk of hitting you is to allow you to take over what is common property. We must divide it up: not just by drawing a line down the topological middle, but also by drawing a notional line down the middle of the risks we suffer and impose there. Not so on my front porch, perhaps. There I should be quite secure - and it is not too much of an inhibition to say that if you blunder on me there, you must pay for the harm you do.

And what then are the boundaries of my property? Those lines are highly conventional, in the sense that they are changed frequently and deliberately. I suspect that it is this variability above all which leads to the conclusion that tort law, and thus the extent of private rights, is a creature of and subject to public policy. Of course if that is true, then as to property the force does go out of the concept of liberty. Liberty is correlative to rights - my liberty is the discretion I enjoy within the notional sphere of my rights. As to the boundaries of my person the principles are general, pervasive. Though these principles develop and are influenced by the social systems and material circumstances which evolve against the background of these rights, I hope I have made it clear that this should not undermine these principles as principles, nor tempt us to try to step outside of our conceptual skins. Such arguments are, however, progressively less convincing as we move from personal to property rights and from property rights closely associated with personality (e.g., privacy of one´s home or rights to choice of profession) to property rights involving largescale economic enterprise. All I want to say about even the most extreme end of that spectrum is that there is a large difference between the community pursuing its policies (economic development, protection of the environment, national prestige or defense) by creating rights and by adopting a directly managerial roIe.56 The regime of property rights represents a compromise between collective concerns (which determine the specific contours of property rights) and a respect for individuals, whose collaboration is enlisted on the understanding that they retain the measure of discretion that any right implies. Indeed, to withdraw all relations to things from the regime of rights would render largely nugatory the natural right to one´s own person and efforts, for those efforts are expended on the outside world. Nor is the changeability of property rights an insurmountable obstacle, so long as changes are gradual or prospective. A sharp, retrospective change in definition does, however, amount to confiscation.

9. How do we establish where the boundaries fall, what is included within our conception of cause, what are the fair implications of general terms, what does fairness require when contractual arrangements fail? In general the method for specifying rights and wrongs is the method of reflective equilibrium.57 At the specific level we may have to rely on the more intuitive method of argument by analogy. Analogistic reasoning, which is the staple of the law, is a poor cousin of reflective equilibrium. An analogy is an implicit invocation of an unstated principle, an inchoate theory, which it is believed or hoped will justify both the decision analogized from and the decision analogized to. Neither method pretends to deductive or a prioristic necessity; it is content to work with the convictions which develop within an existing situation,

56 See Charles Fried, Contract as Promise 99–103 (Cambridge, Mass., 1981); Dworkin, supra, note 12, at 297-311.

57 See John Rawls, A Theory of Justice §§1, 87.

conceding the necessity for their criticism and refinement, but never hoping in some way to step outside of the skin of the considered judgments of actual moral reasoners. It is my claim that citizens and judges working within a legal tradition are able to develop rules of property and tort law, rules about boundaries and causation, in this implicit way: that this development is sufficiently gradual and in sufficient touch with the common morality of the community that Dworkin´s and Hayek´s conception of the common law as (relatively) impersonal has sufficient validity.58

Both argument by analogy and Rawls´s reflective equilibrium have come under attack. Reflective equilibrium, as is well known, depends on considered judgments. These considered judgments, it is objected, are merely subjective intuitions, personal preference.59 The same criticism applied to argument by analogy urges that what will seem like a close or a distinct analogy is once again a mere matter of intuition at best, naked subjectivity at worst. A number of authors have converged on a somewhat half-hearted defense of rationality against these criticisms. The first was Edward Levi in his classic work, An Introduction to Legal Reasoning.60 Levi was ready to concede (quite unnecessarily in my view) that argument by analogy is imperfect reasoning (at best it is a form of implicit inductive reasoning) and to concede as well that in the event it only gives form to the changing policy judgments of judges. The one saving feature which he found in common law argument from precedent, common law argument by analogy, is that it takes place in a public forum, with the affected parties urging their competing analogies, and the judge being compelled to offer a justification for the conclusion he finally announces on the basis of these analogies. That same idea is repeated and becomes

58 Dworkin, supra, note 12; Hayek, supra, note 15, at Part II. 59 See, Noble, “Normative Ethical Theories,” 62 Monist 496, at note 5 (1979); Shaw, “Intuition and Moral Philosophy,” 17 Am. Phil. Q. 127 (1980).

60 Chicago, 1949, at 1–27.

central in Lon Fuller´s posthumous work “The Forms and Limits of Adjudication” 61 and receives its most elegant form in a chapter of Ronald Dworkin´s Taking Rights Seriously called “Justice and

Rights.” 62

Dworkin, commenting directly on Rawls´s method of reflective equilibrium, demurs to the proposition that this method will yield some kind of objective moral truth. The method has independent force and validity, in Dworkin´s view, as a method of public controversy and public decision-making. The necessity for judges to consider competing arguments, to propose principles standing behind their judgments, to seek in their opinions to make these principles acceptable to the community and to relate this acceptability to a demonstration that the newly announced decision is part of the community´s own moral sense and moral theory; this public posture, says Dworkin, gives the conclusions of judges a weight and a legitimacy, a title to the obedience and respect of the community, which is independent of any claims about the objective moral truth of the proposition announced. All three of these authors, Levi, Fuller, and Dworkin, thus make no claims for their preferred method (and mine) in terms of moral truth: their claims are rather to the reasonableness and legitimacy of the result so achieved.

Levi, Fuller, and Dworkin are quite correct in emphasizing the essentially public nature of moral argument and moral discourse. It would be a mistake, however, to conclude therefore that such public arguments are at best only legitimate, that they do not merit the sobriquet of validity or truth. This conclusion would be a mistake because it supposes that moral argument can ever be anything other than in principle public. To be sure, you or I may carry on an intense moral inquiry in total, silent privacy, preparatory to making some deep and important decision. It is a mistake

61 92 Harv. L. Rev. 353 (1978). 62 Supra, note 12, at ch. 6 .

to imagine that this process represents the true form of moral deliberation, and argument before a court, for instance, represents only a vulgarized, public version. When I deliberate in private, if it is true deliberation, I seek to reproduce in my own mind an analogue of a public debate. I try to imagine what analogies would be urged, what arguments made; I test the conclusion I am deliberating by imagining having to justify it to a skeptical but reasonable audience. My point, then, is that justification to and within a moral community is not some externalized version of an ideal of internal moral deliberation, but rather the ideal itself. This is not at all to say that the right or the good is what happens to be accepted in a particular moral community at a particular time, any more than the truth about matters of science is just what happens to be accepted by the scientific community - as if the way to find out what is good or true were to conduct an opinion survey. In a moral or a scientific community the good or the true are goals of inquiry. Were it not so, what would direct the inquiry of those within the community - surely not a search for their own consensus. The point is rather that justification to reasonable men and women pursuing the good or the true is a regulative ideal of the pursuit of moral or scientific truth.63 And this is what one would expect, given the fact that moral discourse is carried on in a language that is a public language, with concepts which are public concepts. When I think even about the most private, the most personal matters, I cannot do so other than in such a public language.

These general considerations spell the limits of my ambition in offering a theory of liberty. Absolute certainty and objectivity are not possible. What I have tried to offer instead has been a sense of confidence in a process of learning, of learning the truth about ourselves and the world. This confidence determines what at bottom we owe each other (a decent minimum) and what are

63 For an interesting discussion of this conception of scientific inquiry, see the account of C. S. Peirce´s views in Bernard Williams, Descartes — The Project of Pure Inquiry, 244–49 (Harmondsworth, 1978).

the bounds of our private sphere. If the answers to these questions are certain enough so is our liberty. I believe our liberty is certain enough. What I cannot assure you of is that what we are certain of today is what we will be certain of tomorrow. But why should this discourage us? In the end the belief in liberty, as in truth, on which it in part depends, is an act of faith.