The Tyranny of the Ideal

by Gerald Gaus

  Proponents of the harm principle, if they are interested at all in accommodating such “harms,” typically seek to depict them as psychological harms or offensive actions.35 No one truly has a basic interest in living in a society without homosexuality, or one free of ungodly rituals, since such religious worldviews are inadmissible in the normalized world of the secular harm principle. However, if people get really upset, that at least is real, and perhaps something can be done about that. This may be enough for the secular, normalized view, but it misses the point of the religious complaint.36 The complaint is not that one is getting upset; whether or not one knew about it, one’s interests would be set back by such behavior.37 Perhaps one of the most striking ways in which debates involving religious categorizations have been normalized (and in Foucault’s more sinister sense)38 in much contemporary liberal discourse is the redescription of religious-based disapproval of homosexuality as a disease, “homophobia”—a manifest effort to control a deviant categorization by labeling it as illness.39

  Many of the same points apply to environmentalism and various forms of support for animal rights. Again, some of the claims of environmentalists and animal rights advocates can be translated into, say, a standard Western normalized world, where humans are the sole persons, but pain to all sentient creatures is bad.40 But many cannot. Those who see nature as an entity to be respected, or who hold that ecosystems have basic rights, do not live in this normalized world. Other examples abound. Even those who embrace almost all the details of the Rawlsian analysis of justice fundamentally disagree on its application to international or global justice, and surely one of the important reasons for this is a fundamental dispute whether “peoples” are entities with moral status.41 Here we have a stark reminder that agreeing on a theory of justice, without agreeing about the social world to which it applies, by no means guarantees agreement in judging social realizations.

  None of this is to say that in a society where there is great disagreement about the social world we inhabit any perspective may legitimately insist that its social world be imposed on others. That in your social world ecosystems are persons, that God forbids homosexuality, or that jokes about sex are not funny42 does not itself tell me anything about what I am to do, or how I am to live with you. It is to insist however, that, unless our thinking about diversity takes seriously these extraordinarily deep disagreements about the nature of the social world (the categories used to characterize it), we will fail to understand what is required to live in a free and open society.

  1.3 Muldoon’s Nonnormalized Contract

  1.3.1 When Worlds Overlap. The most innovative and sophisticated work rethinking political philosophy in light of deep perspectival differences is that of Ryan Muldoon, who proposes a social contract free of normalization assumptions. Muldoon’s key insight is that, even if your and my social worlds do not categorize objects in the same way, we may still have common social “objects,” or, we might say, partial projections of objects, about which we can negotiate. Taking our cue from Muldoon, let us start with a simple case,43 as in figure 4-2. Alf, Betty, and Charlie have three perspectives on the same object; all can see the world in three-dimensional space. On Alf’s perspective the object is a two-dimensional circle, as it is on Charlie’s, though on Charlie’s perspective it is a different circle (it does not share the same spatial location as the one Alf sees). Betty sees the shape as a cylinder, but not as a closed one. Compare: to a research scientist, a fetus may be a source of stem cells; to a Catholic, a person with a soul; to a natural rights philosopher, a fetus may be a trespasser violating the rights of the mother-owner. That there is no normalized characterization of the real object that all must endorse does not imply that these non-normalized perspectives share nothing; they share partial characterizations of the object (or, we might say, they see different projections of the object). That, after all, is why the three can argue about the proper treatment of “the fetus” even though they ascribe very different properties to it—and so in a fundamental sense do not have the same object in mind at all. Notice that, in contrast to Sen’s social choice approach, which founders on disagreements as to the objects being evaluated, Muldoon makes such disagreement the center of his analysis.44

  The core of Muldoon’s insightful—indeed revolutionary—proposal is that we can model social contracts as bargains about who has rights over an “object” while simultaneously acknowledging that the parties do not agree precisely about the nature of the object over which they are bargaining. Writes Muldoon:

  If two agents have different categorizations, this means that they will not divide state space up in the same way. This is simply because they categorize the world differently, and so they have different conceptions of salience. … They are seeing the same thing, but just interpret it differently. As there is no neutral [or normalized] representation of the state space, we cannot take one partition as better than another, and so we cannot just pick one and require everyone to use it. This much is clear. But we also cannot just take the union of everyone’s partition sets and expect everyone to share this larger partition set. This is because there is no assurance that people can see the distinctions that are in other perspectives if they do not themselves hold them. So just as we do not have the same accounts of value (and hence utility functions), we do not have the same partition sets. But even though we cannot expect that individuals can adopt this union of all partition sets, we can describe the overall bargain by making reference to this union set, so long as we do not assume that the agents in the bargain themselves are aware of it.45

  Figure 4-2. Common projections of different perspectives

  The overall bargain really concerns the union of Alf’s, Betty’s, and Charlie’s perspectives, for depending on how the bargain comes out, it will affect, as it were, the way the entire cylinder is used; but none of them see a closed cylinder, so none will think the bargain is about that (it will, from their individual perspectives, be a bargain about who has rights over which circle, or who has rights over an open cylinder).

  Muldoon’s nonnormalized social contract, then, is an n-person bargain over the allocation of rights that, in one respect, is about rights over overlapping bits of the world, yet determines rights over the “full objects” in each perspective (when these full objects include the bits that are jointly seen). Note that this contract does not require that we share a normalized, or even partially normalized, perspective on the social world. Yet the allocation of rights applies to each nonnormalized social world and—if we accept the Nash bargaining solution on which Muldoon relies—each perspective can endorse the overall social distribution of rights.46

  Muldoon’s contract yields public justification without normalization. The ultimate terms of association are justified to all relevant evaluative perspectives insofar as each perspective deems the terms ones that advance its concerns as far as reasonably can be expected, given the perspectives of others.47 Although Muldoon often sees his proposal as an alternative to public reason political theories, that is only because the dominant public reason views have been committed to strong normalization: public reason has typically been identified with the reason of the normalized public perspective—the liberal perspective.48 Muldoon, though, is correct that such traditional public reason views, rather than helping us devise a theory of justice for all, valorize one perspective and its understanding of justice. In response, those excluded either oppose public reason liberalism as a sectarian project, or draw back, alienated, into their own perspectives. The promise of public reason as providing a justified conception of justice in the midst of diversity goes unfulfilled. More than that: it becomes a screen for the dominance of a controversial secular social world.49 In the end, such public reason views are very similar to traditional moral theories, which are based on a normalized perspective claiming to be the uniquely correct way to identify the morally relevant properties of social worlds.50 Muldoon’s project promises to revive public reason as a doctrine that trul
y responds to our deep and enduring differences.

  1.3.2 Dynamism and the Open Society. Muldoon insists that his understanding of the social contract is dynamic, stressing constant adaptation. As new perspectives enter and old ones die away, the social contract must adjust.51 Not surprisingly, he endorses John Stuart Mill’s stress on experiments in living, and the ongoing discovery of better ways to live together;52 conversely, he is especially critical of Rawls’s claim that in modeling a social contract, we should suppose the parties will be bound through an unpredictable future, and cannot renegotiate as they proceed.53 Constant changes in the perspectives of the population (for example, the immigration of Muslim groups), changes in environmental conditions, and discovery of better ways of living together all require that we conceive of the social contract as a set of dynamic, evolving bargains for mutual benefit, not a set of principles fixed “once and for all.” Because there is no normalized, fixed perspective from which to reason about fixed principles, the social contract is necessarily tightly coupled to the perspectives presently populating a society.

  In a fundamental sense Muldoon is absolutely right: an adequate social contract (or theory of justice that can be justified to diverse perspectives) must have the resources to adapt to drastically changed conditions, both internally and externally. We must incorporate the Millian-Hayekean-Popperian insight that we are always refining our ideas about justice and our analysis of the conditions for its social realization.54 The notion of ideal justice as a fixed point not subject to revision is perhaps the most implausible assumption of what Popper called “utopian” theory (§II.3.4). On all this, we should follow Muldoon. However, Rawls’s supposition that we should conceive of the parties as choosing principles “once and for all” need not be interpreted as a statement of the fixed point view but, rather, as reminding us about two important facts about the very idea of rules of justice as a basis for social cooperation.

  First, to endorse a rule of justice requires that the endorsement is “reversible”: if rule R specifies the way that roles 1 and 2 must relate to each other, for Alf, basing his judgments on perspective Σ, to truly endorse a rule requires that Alf endorses it whether he occupies position 1 or 2.55 If Alf endorses a rule because he believes he will solely occupy position 1 (say, he knows that he is only an employer and never an employee), in an important sense he does not endorse it as a requirement of justice (but as, say, a rule of self-interest). Even though our goal in a nonnormalized social contract is to uncover rules of justice that do not presuppose a normalized perspective (or small groups of such perspectives), we still seek terms of association that conform to some ideal of impartiality in the sense of reversibility. To say that a rule is not open to continual renegotiation is to stress that parties are bound by it, even on those occasions when it does not promote what their perspective deems important (including their interests). More pragmatically, if Alf’s endorsement of the rule presupposes that he will never occupy role 2, expected changes in circumstances (i.e., he becomes an employee) will lead him to abandon the rule, showing his endorsement of it to be unstable. In his early versions of his social contract theory, Rawls thus argued that the parties must suppose the agreement cannot be renegotiated should it turn out that they did not occupy the more advantaged positions. “They each understand further that the principles proposed and acknowledged on this occasion are binding on all future occasions. Thus each will be wary of proposing a principle that would give him peculiar advantage, in the present circumstances, supposing it is accepted. Each person knows that he will be bound by it in future circumstances the peculiarities of which cannot be known, and which might well be such that the principles are to his disadvantage.”56 Because one is bound by the choice of the principles over an extended period of time regulating circumstances that one cannot predict, a rational person will not seek to tailor the principles so that he gains advantages given his present circumstances, since these may unpredictably change, and he may end up on the losing end of rigged principles.57 In this case resisting calls for renegotiation because the facts have changed is fundamental to the very idea of what it means to endorse rules of justice.

  Second, as Hayek so effectively argued, just because we want a society that is dynamic and constantly adjusting to new conditions, we need relatively stable social and moral rules.58 A deep source of uncertainty in social life concerns the actions and expectations of our fellows; binding us to common rules allows us to reduce this uncertainty and so engage in fruitful cooperation. Unless we are bound to relatively stable rules we cannot reduce this uncertainty. The basic moral rules of society, and the basic moral norms governing the political, provide the background for social and economic experimentation. Unless this background sets relatively fixed parameters as to, say, what claims of property are to be respected and what personal rights are to be guaranteed, innovators have insufficient settled points from which to plan. Innovation is not typically a planless, random activity; it is based on decentralized individual plans, and these plans and their coordination require important fixed parameters.59

  To be somewhat less abstract, contrast two views of a dynamic society, in the sense of one that readily and continually admits new perspectives—say, immigration from ethnic or religious groups not presently represented or simply internal processes (such as recombination, §III.3.4) that regularly create new perspectives. On Muldoon’s proposal, such a society will acknowledge that the social contract will require continual revision, so the basic allocation of rights is being revised as new groups (with new perspectives) enter.60 But note that this can entail a high cost for current perspectives. While Muldoon rightly stresses that through the division of labor I can benefit from new groups and perspectives (see §IV.3.3), it also is the case that on his dynamic bargaining model I do not know the rules on which I shall interact with them, as the social contract will be renegotiated as they enter. So it will be harder to anticipate the costs and benefits. Supposing also, as seems to be the case, that most people are conventionalists and so endorse most of the rules they live under, they will certainly see one clear and significant cost in allowing immigration of culturally different groups: it is apt to significantly revise the current system of moral rules. Perhaps a worry about the nature of religious accommodation as it was negotiated in some Western European countries is that it was rather close to Muldoon’s model, being based on a bargain between different Christian sects, who all received their share of state influence and funding. Enter Muslim populations, and this basic bargain has to be renegotiated, and now with groups whose perspectives are less familiar. The result is apt to be great reluctance to accommodate Muslim groups: the costs are upfront and clear, the potential benefits in the future, and, given that we are unclear about the new allocation of rights, far less certain. In contrast, if our rules are able to secure significant stability in the face of new perspectives, although current residents will anticipate new interactions (some welcomed, some not) the basic allocation of rights and duties will not constantly change. And because the rules of the game will be relatively stable, there will be both a greater ability to plan how to interact with the newly arrived groups and far less cost—new arrivals do not mean new social contracts.

  1.4 Not All Liberal Justice Is Fit for the Open Society

  Let us draw together some of the lessons we have learned from our analyses of Rawls, Sen, and Muldoon. Under full normalization, the relation of a perspective Σ to judgments of justice is manifest; Σ draws on the five elements of its perspective61 to determine the justice of any social world. When we allow a theory of justice that itself seeks to accommodate multiple partially normalized perspectives on justice, so that we have fundamental disagreements about justice, we immediately confront the question of the relation of any given partially normalized perspective to the theory’s overall judgment of justice. As I argued (§IV.1.1), Rawls never really solved this problem: he allowed for a set of eligible perspectives based on different principles of justice
(the family of liberal views), but he was never clear how the diverse set of “reasonable” perspectives could all be encompassed in overall, consistent judgments of justice in his theory. His political liberalism project understood some perils of normalization but did not see how we might arrive at way to relate each of the eligible perspectives to the interperspectival judgments of justice.

  In §IV.1.2 we saw that Sen provides a method to integrate diverse evaluative standards into a theory of justice; each partially normalized view of justice (modeled as an impartial spectator) takes its evaluative standards and applies them to a common, normalized set of social worlds. If all impartial spectators agree that social world a is more just than b, then according to the full or complete theory of justice a is more just than b. As we saw (§IV.1.2.3), Sen’s procedure requires full normalization of the properties of the set of social worlds in the domain {X}; while Sen clearly makes progress in thinking about justice without full normalization, his social choice approach cannot accommodate significant differences about the justice-relevant features of the social worlds. This suppresses a great deal of perspectival diversity about the nature of the social world. But it also suppresses diversity of evaluative standards; some evaluative standards—such as religiously informed ones—make sense only if the world has certain features. If souls and knowledge of God’s will are not features of the social world, there is not much to say for a Divine Command theory (or, perhaps, even for a liberal egalitarian theory).62

  Muldoon’s analysis overcomes this limit. It is useful to think of his account as making justice a vector of all the eligible perspectives.63 For any given array of perspectives, interperspectival justice is the result or upshot of the combined effects of the perspectives in the array. To use more philosophically familiar language, justice is “constructed” out of the entire set of eligible perspectives. The worry is that this ties interperspectival justice too closely to the set of current eligible perspectives, and so Muldoon’s proposal, insightful as it is, leaves something to be desired as a framework for a diverse and open society.