2.3 How Well Justified Are Our Principles of Justice?
We might press a more fundamental worry: can we really be so confident of any derivation of principles of justice that we are warranted in insisting that a society that is miserable or deeply dysfunctional, but yet perfectly conforms to them, is the ideally just social world?80 When we search for a plausible basis for such a radical insulation of the justification of our judgments of justice from their social realizations, we seem to be led to something like a strongly foundationalist account of justification. Roughly, on a strongly foundationalist account, some belief β1 (i) is the basis of the justification of other beliefs {β2… βn} such as beliefs about just institutions; (ii) β1 is not justified by any other beliefs; and (iii) given the degree of self-evident justification of β1, there is no other belief β* that could possibly defeat (i.e., undermine) the justification of β1. In terms more familiar to moral and political philosophy, β1 may be among our “deep intuitions”81 that simply cannot be defeated by pointing out other conflicting values, including those associated with “facts” about the social realizations of the principles expressing β1.82 So on this view, if people who conform perfectly to the correct principle of justice, which expresses β1 in social world S, are deeply alienated and cynical, or lead impoverished lives, this might lead us to think that people in S simply are not up to justice, but it couldn’t lead us to seriously question whether we are correct about the principles of justice. The correct principle is so near self-evident that no mere facts about the disaster of implementing it, or that reflective and goodwilled people do not will such a world, could defeat its justification.83
Strongly foundationalist accounts of justified belief have a number of intractable problems, the most important of which is their central claim that “deep intuitions” have very high (approaching unity) levels of credence; it is this that insulates them from defeat by other values, including values that are realized in facts concerning social realizations. Although on some accounts perceptual beliefs may approach this (some are tempted to say: “If I believe I perceive red, I cannot be wrong, and this could not be defeated by other beliefs”),84 it is very hard to accept that intuitions about social justice could possibly have this status. On any plausible view, the level of self-justification of beliefs about justice should be set low enough to admit defeat. Note that this does not mean that basic beliefs about justice cannot be self-justified and foundational (and so in this sense they still can have “fact independent” justification), but simply that whatever its degree of self-justification, it is sufficiently less than unity such that its (self-) justification can be defeated by other considerations.85 Once we allow that regardless of how deep it is, an intuition about justice can be defeated, we can press that the dastardly character of the social world that best realizes our intuitive principles could, at least in principle, defeat the claim that a world that meets these principles is just (in Kantian terms, we could not will that these principles be a law of nature in that social world).
I am not claiming that our moral judgments must be “coherentist” in any formal sense, nor must we embrace “reflective equilibrium.” I do, however, believe that we have very good grounds for (i) denying that any intuition that a substantive principle of justice is correct has such a high credence that it could not be defeated by other considerations; (ii) that when specifying a just social world, the realizations of universal or widespread acting on a principle are relevant defeaters for the claim that it is the most just principle in this social world. And (iii) so too for the ideally just society: to say that universal conformity to a principle is an aspect of the ideally just society can be defeated by the social realizations of acting on it in the ideal social world. None of this is to say that any and all adverse social realizations defeat the claim that a principle correctly identifies justice (if, say, we have a social world of evil agents, it may still be the correct principle), but only that social realizations are always of potential relevance, and so cannot be simply “bracketed” on the grounds that since “we know the truth about justice, no mere facts could ever change this.”
I believe that this is an important point about the justification of principles of justice and their social realizations. It is the deep reason why we should reject what Sen might call “principle fundamentalism”: that is to say, so long as the principles are satisfied, justice doesn’t concern itself with the way people’s lives go. However, the core analysis of this and the following chapters does not depend on accepting this point. All that is strictly required is that, for whatever reason, the ideal theorist acknowledges that determining the comparative justice of social worlds partly depends on the worlds’ institutional structures and their dynamics, given the population operating, and subject to, them.
3 MODELING THE IDEAL (AND NONIDEAL)
3.1 Setting the Constraints Regulating Coherent Social Worlds (One Sense of Feasibility)
We now can see that our initial distinction between the orienting role of ideal theory and political philosophy as recommending was too stark (§I.1.5). While we might have initially viewed the mere orientation function of ideal theory as one of “pure moral philosophy,” which is unconcerned about recommending moves to new social structures, and thus also unconcerned with predictive models of social worlds, we now see that even the most ideal of theories of justice must engage in modeling the ideal social world, as we need to make some effort to inquire about the social realizations of our ideal principles. Plato’s Republic is a good example of ideal theorizing, for it seeks to model the ideal state based on true justice: if we accept Plato’s account, we have some idea as to how ideal justice would pan out, at least under ideal circumstances. Notice that Plato does not say: ideally, acting on the principles of ideal justice must work out well—after all, they are ideal! Rather, he specifies some conditions (the nature of economic life, human psychology, international relations, and so on), and then engages in a long modeling exercise to show that under these conditions the social realizations of his understanding of justice are acceptable.
This is of even more importance for Rawls, who stresses not only that the ideal of justice must suppose a certain background social world, but that the postulated basic features of the social world must not stretch credulity:
Some philosophers have thought that ethical first principles should be independent of all contingent assumptions, that they should take for granted no truths except those of logic and others that follow from these by an analysis of concepts. Moral conceptions should hold for all possible worlds. Now this view makes moral philosophy the study of the ethics of creation: an examination of the reflections an omnipotent deity might entertain in determining which is the best of all worlds. Even the general facts of nature are to be chosen. Certainly we have a natural religious interest in the ethics of creation. But it would appear to outrun human comprehension.86
Rawls abjures the ethics of creation, where just about all parameters of possible social worlds, including the laws of nature, are open to specification; instead, he seeks to develop “the conception of a perfectly just basic structure … under the fixed conditions of human life,” assuming that given these fixed conditions the social world is characterized by “reasonably ideal” circumstances.87
The last decade has witnessed intense debate among political philosophers whether a theory of “ideal distributive justice” should, as does Rawls’s, postulate social facts as, for instance, that without incentives people will not voluntarily work at the most socially useful job, or that they will not voluntarily pay taxes without any coercive threats.88 This is often understood as a debate about feasibility, in one of the many senses of this complex concept: what constitutes a “feasible ideal” in the sense of one that identifies an ideal social world that is fixed by credible parameters of the socially, economically, and politically possible.89 For our present purposes, the important point is not whether Rawls has correctly identified the parameters, but that any account
of the ideal must settle on some parameters or constraints in modeling the social realization of the principles of justice if its ideal is not to “outrun human comprehension.” What Rawls calls the “ethics of creation” allows, as it were, social philosophers to treat almost every parameter as open to setting at any value they see fit: manna falls at will from heaven so all resources are available in infinite quantities, genders can change at will, one can be a critical critic and fisherman at the same moment since time-space variables allow simultaneous activity by the same person in any number of time-space coordinates, one lives forever, we have worlds where actions have no opportunity costs, or perhaps one is reincarnated in various lives such that by the end of time all bad luck will be equalized, or whatever. If the principles of justice hold in all possible worlds, they hold in all these creations, for the idea of “possibility” is inherently one that is set against a set of constraints. There is no such thing as “absolutely impossible”: every claim of possibility is of the form “X is possible (is not precluded by) the set of constraints (including laws) L.”90
Of course no theory of ideal justice allows that many variables to remain open. A plausible current proposal is to distinguish “hard” constraints, such as the laws of nature and the natural sciences, and “soft” constraints, such as those concerning present human motivations.91 The idea is that the hard constraints entirely rule out some sorts of social realizations as infeasible, while the soft constraints can be seen in more probabilistic terms, yielding judgments of more-or-less feasible. It is also supposed that hard constraints cannot be removed by humans, while soft constraints describe “that some particular fact could be changed, now or at some point in the future.”92 While no doubt there is something to this, when we begin to model social worlds and predict social realizations, the distinction pretty much evaporates for at least three reasons.
(i) In determining whether a social world, u, is feasible, we need to know not only that constraint C cannot be violated, but what social states C rules out. Once we add technological innovations, for example, whether C is or is not a constraint on whether social state u can be realized can radically vary and, indeed, be manipulated, often more easily than “soft” constraints, such as the degree to which humans are public-spirited. Consider Thomas Malthus’s two “hard” constraints in his Essay on Population:
First, That food is necessary to the existence of man.
Secondly, That the passion between the sexes is necessary and will remain nearly in its present state.93
For Malthus these “laws, ever since we have had any knowledge of mankind, appear to have been fixed laws of our nature; and, as we have not hitherto seen any alteration in them, we have no right to conclude that they will ever cease to be what they now are, without an immediate act of power in that Being who first arranged the system of the universe; and for the advantage of his creatures, still executes, according to fixed laws, all its various operations.”94 In contrast, William Godwin’s utopian anarchism (to which Malthus was partly replying) saw the second as a “soft” constraint, which could be overcome by intellectual progress.95 Godwin accepted that his anarchist-socialist ideal, which resisted the struggle for existence, needed some way to control population.96 So what is hard and what is soft is by no means uncontroversial. Leaving aside this dispute, suppose one does accept that enjoyment of, and consequently the having of, sex is a hard constraint. To cope with this constraint a “socialist” utopia such as Thomas More’s included institutions such as colonization to control population.97 However, as John Stuart Mill recognized, one could accept Malthus’s “hard constraint” (that people will continue to enjoy sex), but deny that it precluded a more socialist ideal by technological change—the diaphragm.98 Thus under some technologies the hard constraint that people will continue to enjoy sex does not constrain utopian social worlds.
(ii) Because technological variables are difficult to determine, the ideal theorist seldom is in the position to assert with anything like a probability of 1 that some hard constraint definitely rules out desired social states. Moreover, it is often difficult to determine whether some hard constraint is “redundant” in the sense that, while it is a constraint on possible social realizations, it in fact does not constrain the set of relevant options in an ideal theory. Assuming that it is a hard constraint that we cannot go faster than the speed of light, this may well not impact any utopian scheme and so is to be ignored. But, perhaps, it could affect social realization, if instantaneous communications was part of paradise. As Giandomenico Majone concludes, because “it is often difficult to determine a priori which limitations are binding … feasibility statements must then be interpreted in probabilistic terms.”99 This, then, means that, as far as theoretical understanding of the ways a utopia might function, the distinction between probabilistic “soft constraints” that may be overcome, and “hard” constraints that definitely rule out possibilities, is not of great import.
(iii) Lastly, because social realizations depend on the interaction of a number of variables, what appears as a number of soft constraints can set into a much harder one. For example, suppose that a theory posits five soft constraints each of which yields a 20 percent chance that the utopia will arise, and all come to bear. Given the joint probabilities, there is only about a .00032 chance utopia will come about.
In modeling whether some social world is feasible given a set of constraints (C) that govern its behavior, then, pretty much all constraints should be understood probabilistically. Given this, for any set C of constraints there is a probability distribution of possible social worlds that might emerge on a given model. Let us call this probability distribution the ideal theory’s predictive modeling of the social realizations of meeting C. It is fundamental that we should not fall into the trap of thinking that for every set of constraints C an ideal theory could predict a unique social world arising out of it. Every ideal involves specifying constraints or parameters (I shall not distinguish these here), taking some as very likely to affect social outcomes, and others as less apt to.
3.2 The Aim of Ideal Theory
On a naive view—which, alas, seems widely embraced in political philosophy—the aim of an ideal political theory is uncovering “the truth,” as if, in making decisions between rival theories, we decide on the basis of a clear and unidimensional goal—truth. Not even natural science is well understood as having a monistic goal of “the truth.” As Thomas Kuhn has shown, a true scientific theory aims at maximizing a set of values: accuracy of fit with the data (and of course that requires a decision about what the relevant data are), simplicity (this can mean different things, such as ease of computation or axiomatic parsimony), consistency (not simply internal, but also with related theories), scope (is the theory a comprehensive explanation of a range of phenomena?), and fecundity (does the theory open up fruitful lines of research?).100 And in some fields, perhaps social utility is a desideratum.101 Some think accuracy is all that is necessary, but as Kuhn stresses, on accuracy grounds alone, Copernicus’s system was not, until revised by Kepler, more accurate than Ptolemy’s, yet Kepler had already made a theory choice in favor of Copernicus’s system before he began to work on revising it.102 In some cases a theory may rank higher on all these criteria, and so dominate competitors, but often one theory will rank high on some, another theory on others. And as Kuhn effectively argued, there simply is no algorithm for combining these values into a single value to be maximized.103 Different scientists will trade off these values in different ways, sometimes leading to different theory choice, even among those who share the same values and agree on the data.
Theory choice in political philosophy too has multiple desiderata. According to Rawls, a political philosophy should aim (i) “to focus on deeply disputed questions and to see whether, despite appearances, some underlying basis of philosophical and moral agreement can be found”;104 (ii) “to orient us in the (conceptual) space, say, of all possible ends, individual and associational, politica
l and social”;105 (iii) at “reconciliation … to calm our frustration and rage against society and history,” leading us “to accept and affirm our social world positively, not merely be resigned to it”;106 and (iv), as a variation of reconciliation, to propose a realistic utopia, “probing the limits of practical possibility.107 Our hope for the future of our society rests on the belief that the social world allows at least a decent political order, so that a reasonably just, though not perfect, democratic regime is possible.”108 Without claiming that these are canonical desiderata, we can readily see that, depending on how they are weighted, different theorists will endorse different theories, and so model the ideal in different ways. Importantly, a political philosopher who gives significant weight to the third and fourth aims is unlikely to decide that a political philosophy of hopeless realism is the way to go (§I.1.4). Estlund’s theory of the hopeless ideal is hardly apt to “calm our frustration and rage against society and history,” by inducing us “to accept and affirm our social world positively, not merely be resigned to it.” Given Rawls’s understanding of the aims of political philosophy, he is committed to an ideal that fixes parameters within, but at the limits of, some plausible range. In contrast, a theory that stresses orienting, or one that focuses on the importance of conformity with strong intuitions about justice, is apt to fix less of the parameters in the modeled social worlds, allowing for possible but hopeless ideals.109 What must be stressed is that to invoke “possibility” as a constraint is itself to fix some parameters; it is the area (conceptual, empirical, etc.) defined by the fixed parameters that define the space of possibility or, as I shall say, the option space of an ideal theory. So it is not as if Rawls fixed parameters in his ideal social world whereas Estlund’s utopia does not: the difference is in the parameters that are identified, and the range of variance that is allowed in others, and that in turn will depend on the criteria employed in theory choice.
The Tyranny of the Ideal Page 5