The World Philosophy Made

by Scott Soames

  Since those in the original position could not in good faith agree to a social contract based on average utility, though they could agree to one based on his principles (i) and (ii), Rawls concludes that those principles would be chosen. He can’t demonstrate this because the original position, which includes knowledge of the sciences, can’t be stated precisely enough to provide proof. But the informal case he makes is plausible enough to be worth subjecting to closer scrutiny.

  What makes principle (ii) seem reasonable is the vast ignorance of the parties in the original position. Since severe misery and extreme poverty for some aren’t ruled out in societies under consideration, it may seem reasonable that those in the original position would try to minimize their plight if they turn out to find themselves at the bottom. Whether or not rationality dictates this risk-averse strategy is debatable; at least it isn’t obviously unreasonable. But it’s also not unreasonable to sacrifice principle (i), ensuring maximum individual liberty for all, if doing so would mitigate the misery one would suffer if one were badly enough off. Thus, it’s hard to make the case that those in the original position would choose both (i) and (ii). Perhaps they would choose (ii) while swapping (i) for a principle that merely outlawed slavery. This would violate the priority that Rawls assigns to (i), but it is hard to see how the choice he sets up in the original position would exclude it.30

  However, the worst problem with Rawls’s discussion is his supposition that any principles chosen in the original position must be just because that process is fair. Rawls designed the process to reflect his unargued assumption that all socially desirable characteristics, including “even the willingness to make an effort, to try, and so to be deserving in the ordinary sense,” are ultimately undeserved, from which he concluded that there is no morally significant notion of desert or entitlement antecedent to an all-encompassing theory of political justice to which that theory must conform.31 Instead, he took it for granted that our ordinary notions of entitlement and desert are conceptually dependent for their justification on a comprehensive and antecedently justified theory of political justice. This, I believe, is back to front.

  Presumably, Rawls thought that all socially desirable characteristics are ultimately undeserved because he took their causal antecedents to include biological and environmental factors outside of one’s control. If to deserve x one must do something to bring x about, then, of course, we don’t deserve our genetic inheritance or our early environment, which together determine who we are. Since we did nothing to bring it about that our parents loved and supported us, we didn’t, in the relevant sense, deserve their love and support, or anything else stemming from them, including their genes. But the trivial sense in which this is so has little, if any, moral significance.

  Did our parents have a duty to love and support us, independent of their duties to society as a whole? Yes. Were they entitled to give us their love, time, energy, guidance, and some of what they had saved? Of course they were. Since they were entitled to do what they did, we are entitled to use what they gave—first to develop the socially desirable skills and traits they hoped to instill in us, and then to put those skills and traits to work to benefit ourselves and others without harming or coercing anyone. Because we live in a society with others with whom we are not acquainted, we may be morally constrained to devote some of our efforts to assist them. But this is a far cry from Rawls’s claim that one’s natural gifts and socially desirable traits are the common property of all.32 They are parts of oneself to which one has a prior, though not unlimited, right to profit from as one sees fit.

  Because Rawls misses the primacy of such pre-political entitlements, his argumentative strategy threatens to self-destruct. Fairness, his supreme evaluative notion, is far from our only way of valuing others. We also recognize a duty not to harm innocent others simply because they are human beings (whether members of our society or not). In many cases, we are prepared to go further—to assist strangers when doing so may be important to them without being onerous for us. In short, we recognize the presumptive moral value of innocent others, while presuming that their welfare deserves some respect. However, when we move beyond this broad but mild benevolence, the desirable or undesirable characteristics of other people, though ultimately “undeserved” in Rawls’s extraordinary sense, always matter to our morally significant relationships with them.

  Although real fairness is one component of our moral outlook, Rawlsian fairness isn’t. As ordinarily understood, fairness involves cooperative activities, relationships, and agreements that give rise to reciprocal expectations, which, in turn, generate duties. Not all agreements do. Under certain conditions, implacable enemies—imagine human beings versus space invaders—could reach agreements that each would find advantageous, without in any way valuing the welfare of those on the other side. If such agreements were vigorously monitored, each side might have purely self-interested reasons to abide by them simply to avoid retaliation. But since genuine moral duties aren’t generated, questions of fairness wouldn’t arise.

  They do arise when the parties are presumed to be moral equals in the modest sense I have sketched. Then, one is morally obligated to bargain in good faith, not to lie or mislead, not to agree to terms one is unwilling to honor, or to break agreements one has made, even if one’s breach were to go undetected. Although Rawls recognizes this, his conception of fairness goes much further. Though it is part of his conception of fairness, it is no part of our normal conception that parties to a voluntary agreement aren’t allowed, should they so choose, to condition their participation in a cooperative, rule-governed, activity or relationship on receiving benefits roughly proportional to the value they bring to it. Nor is it part of the ordinary conception of fairness, though it is part of his, that participants in such an activity or relationship should be committed to maximizing the advantages of some individuals, no matter how little they have, or are willing, to offer others.

  In short, fairness is a pre-political social notion that must be understood before the basic principles for organizing the justice of a complex society can be established. To understand fairness, we need to understand how it is connected to the value we place on others and our relationships with them. We don’t interact with Kantian ciphers. When we interact with people, the strength, the extent, and the nature of our desire to be fair—indeed, our very conception of what being fair, in the circumstance, consists in—is conditioned by what we value, by those with whom we are dealing, and by what each of us may reasonably expect from the other. What fairness requires in a given situation depends on the nature of the interaction, the social connections of the parties, their needs, and the antecedent entitlements generated by their assets and advantages. If a more abstract notion of fairness is to be included in a theory of justice for society, it must somehow incorporate these features of ordinary instances of fair dealing; it must factor them into the calculation of the overall justice of a society. Since the original position was designed not to do this, the Rawlsian decision procedure based on it is not an ideally fair procedure for allocating social benefits and burdens.

  If this is right, then Rawls’s principles (i) and (ii) wouldn’t be justified by showing that they would be chosen by agents in the original position (even if that could be shown). What about the principles themselves? What, independent of the original position, can be said for or against them? Principle (i), which stipulates maximum liberty for all, recognizes the fundamental value of liberty in human life, while implicitly incorporating the wisdom that wealth creation–-along with medical, scientific, technological, and intellectual advance—requires relatively free markets, the rejection of a command economy, and the broad freedom to inquire and experiment. Hayek could not agree more. Though principle (ii) raises some problems, one aspect of it—the idea that some sacrifices by the more advantaged are required to aid the less advantaged—is not unreasonable, and indeed, if properly implemented, may enhance social solidarity in a way that benefits the
society as a whole. To this extent, there is a good deal to be said on behalf of Rawls’s vision.

  Nevertheless, it isn’t the social or political ideal. The flawed conception of fairness that provides the rationale for principle (ii) is one which would, I think, be widely rejected by many who clearly understood it—thereby increasing discontent and instability in a society organized around it. Because of this, implementing it would be contentious, while also presenting daunting, and perhaps insoluble, practical and theoretical problems. How, for example, are the “least advantaged” to be identified? Are they those with the lowest income in a given period, the lowest income plus wealth, the lowest income plus wealth corrected (somehow) for health, age, location, marital status, and other family obligations, the lowest income plus wealth, even further corrected by one’s occupation (including its inherent danger, desirability, or opportunity for future advancement), the lowest income plus wealth, still further corrected by available free time and free access to facilities provided by other private parties, and so on? Since each of these criteria could be expected to gain its own political constituency, the practical problems posed by implementation of principle (ii) might well lead to social conflict.

  There is also a theoretical problem: we don’t know how to compare, weigh, and aggregate the different goods that make up total well-being. As we saw in chapter 8, we can measure not only what alternatives a single individual prefers to other alternatives, but also how much more the individual prefers them to others. But, as Rawls recognizes, there is no known solution to the problem of determining how much more I prefer A to B than you do.33 Because of this, there is no precise way of measuring and comparing the total welfare (utility) of different individuals, and so of identifying the least well off. Although Rawls’s use of interpersonal comparisons of access to primary goods (particularly wealth and income) is a defensible rule of thumb, it doesn’t solve the justification problem. Because we know that much more than wealth and income go into happiness and well-being, we may resist taking wealth and income from some (who don’t fare well on other dimensions) to give to others (who do fare well).

  Finally, Rawls cannot escape Hayek’s dilemma: any attempt to improve the material standard of living of the least well off by limiting the liberty of the most productive members of society is likely, in the long run, to lower the standard of living for everyone. To see this, first consider Society 1. It adopts Rawls’s principles (i) and (ii), redistributing money from the most to the least well off, until the latter reach an income of X dollars, which cannot be increased because further redistribution would disincentivize those at higher levels enough to reduce the gross domestic product (GDP) below the level needed to ensure X dollars for those at the bottom. Society 2 is just like Society 1 in all respects except that its redistribution scheme provides an income of 90% of X dollars to the bottom group, allowing the more productive groups to keep slightly more income than their counterparts in Society 2. Being allowed to retain more, they spend a bit more, invest a bit more, and work a bit more, thereby increasing GDP by some percentage Z.

  Imagine these policies continuing for decades. At the end of year 1 the GDP of Society 2 will be slightly greater than that of Society 1. Starting from a higher level in year 2, this increase will be compounded; at the end of year 2 the gap between the two GDPs will be a little greater than it was before. Since the compounding continues, the gap will continue to grow. At some point Society 2 will be more than twice as rich as Society 1, and the worst off in Society 2 will have nearly twice the material standard of living as their counterparts in Society 1. (A 1% increase in the wealth of Society 2 over that of Society 1, compounded annually, will lead in 70 years to Society 2 being twice as wealthy as Society 1.) Since health, longevity, education, and culture tend to increase with GDP, Society 2 is likely to provide a much better life for all than Society 1. It hardly seems just that those born into Society 1 should be deprived in this way of the better life they could have had.

  The argument works no matter what nonzero amount X over any fixed period of time T is selected to implement Rawls’s principle (ii). Surely, however, one is inclined to think that there must be some policy that would, over an indefinite future, maximize the position of the worst off. Hayek and Nozick would probably agree, setting X equal to 0. That is, they would say, a policy of no redistribution for the purpose of reducing economic inequality will, over time, redound to the advantage of the least well off. Nevertheless, they would, I think, find it reasonable to provide some social minimum. Crucially, however, their social minimum would not be defined in terms of its relation to the levels of wealth or income of those at higher levels. The proper principle would be one that provides the minimum needed for living a decent human life under contemporary circumstances, without requiring present citizens to make sacrifices below this level for the sake of future citizens. However, it would require beneficiaries to make reasonable contributions to the common good, consistent with their abilities to do so.

  Although this is clearly not what Rawls had in mind, it is, I think, a reasonable blend of what he gets right with the indispensable insights of Hayek and Nozick. Whether or not it is the blueprint for an ideal society is a very different matter. Indeed, the presupposition that there is such a thing as an ideal society, which must follow some determinable blueprint, is not obviously correct. In recent years, the efficacy and feasibility of the philosophical quest for such a society have been subjected to critical scrutiny. The contemporary philosopher Gerald Gaus is one of the leading skeptics.

  GERALD GAUS: THE TYRANNY OF THE IDEAL

  Unlike Rawls, who used an imagined consensus on first principles to sketch an ideal system which, he hoped, might help us improve the societies in which we live, Gaus argues that there is no idealized consensus to be had, and that such ideal approaches should be replaced by more realistic models for extracting more limited agreements. Recognizing that complex societies are made up of many groups with different social and political ideals, he recognizes that improvements depend on substantial overlap of the beliefs and ideals of various groups, as well their willingness to converge on what all take to be better, even if none take it to be best. For Gaus, all social and political perspectives are partial. Because none comprehends the full potential for human social interaction, there is room for each to grow and change by interacting with others. When this leads to limited agreements between proponents of different conceptions of the good, new institutions arise. When these are seen to be beneficial, the resulting new knowledge may expand the original perspectives, making possible new agreements generating further progress. This process by which not consensus, but disagreement about the good (against a background of overlapping agreement about many things), leads to progress may, in principle, be unending. Thus, he suggests, there may be no realistic need to search for, and probably no realistic possibility of finding, a comprehensive political ideal to guide us.

  His book examines how we ought to reason about sociopolitical norms. Like Rawls, Gaus uses the notion of a model consisting of (i) criteria for evaluating when one society is better than another, (ii) a description of the features of a society to which the evaluative criteria are to be applied, and (iii) the range of possible societies evaluated and the outcomes of the evaluations. For Rawls, (i) measured how desirable different societies were to self-interested rational agents in the original position, (ii) applied that criterion to the distribution of liberty and of primary goods (wealth and income) in the society, and (iii) measured how well any possible society fared in comparison with all the others.34

  Unlike Rawls, Gaus doesn’t favor one model above all others; nor does he demand that all possible societies be modeled. Since his goal is to understand the logic for using models to reason about social and political reality, he allows a model to use any evaluative standards it wishes to rank the total goodness of a society in terms of whatever social and political institutions it takes to be relevant to determining that ranking—e.g., a
society’s social, political, economic, educational, cultural, and religious institutions or practices, its taxation and national service policies, the extent of its debt and borrowing against the future, its uses of natural and other resources, etc. Applying the evaluative standards to the relevant social institutions and practices results in an “inherent justice score” (really an inherent goodness score) for each society. For simplicity, we may imagine extracting a partial linear ordering in which, for every pair of distinct societies modeled, either one is ranked higher than the other, or they are ranked as equal. If our society is among those modeled, we can use its position in the ordering to determine which possible changes would allow us to march upward in the ordering, only stopping if and when we reach the point at which further changes significantly reduce its score.

  It is useful to begin by thinking about an actual society @, e.g., our own, about which we already know a great deal, identifying institutions and practices that have changed plus the characteristics of those changes we have found to be good or bad. These can be used to construct a simplified model that includes a description of @ along relevant dimensions, and a set of criteria for evaluating modifications that would result if so far unexperienced changes were to occur. The goal is to map possible futures of @—@1, @2, etc.—that we might use in our plans for improving @.

  The model might be expanded by adding descriptions of actual societies other than our own, as well as possible variations on them, expanding the evaluative criteria in our model to apply to them. Models could then be tested against reality by noting and evaluating what happens to one of our real societies when it changes. To the extent that our model accurately evaluates the new change, we take it to be confirmed; to the extent that it doesn’t, we revise it. This, very roughly put, is the kind of model-based investigation that Gaus presupposes.