For various reasons I have gotten pretty involved in the literature on endogenous preference change. My first push came from reading Rawls. As I see it, the key difference between Rawlsian contractarianism and Buchanan/Gauthier rational choice contractarianism is not just that Rawls posits a sense of justice, a capacity enabling agents to be motivated by considerations that nicely allow for the choice of non-Nash, Pareto-improving strategies (Gauthier’s “constrained maximization” gets you this, as does McClennan’s closely related “resolute choice”) but that Rawls has something of an account of endogenous preference change that accounts for the convergence of the right and the good and thus the stability of social ordered according to the principles of “justice as fairness.”
The trouble with theories of endogenous preference change is that they seem the ruination of neo-classical theories of efficiency. The usual Pareto or Kaldor-Hicks (or Marshallian, if you like),criteria for efficiency work only by holding preferences fixed or exogenous. We evaluate the desirability in changes by tracking their relation to people’s preferences. If a change makes someone better off and no one worse off, in terms of preference satisfaction, then it is worth doing. But if a change can modify individuals’ preference-profiles themselves, then our efficiency criterion becomes a moving target, and one becomes quickly mired in paradox.
Rawls’s argument for the stability of justice as fairness (JAF) in TJ is, in econo-speak, that individuals raised under a society well-ordered by JAF will develop preferences that will facilitate willing compliance with the principles of JAF. The structure will produce a distinctive set of preferences, which, in turn, serve to reinforce the structure. And so on. That accounts for the robust stability of JAF. Another way to think of it is: under JAF, the preferences for JAF are endogenously reinforced — “consuming” justice strengthens the demand for justice — so that the order stabilizes preferences just as the preferences stabilize the order. Once so stabilized, if one were to consider our preferences as exogenous, JAF would be efficient, although JAF is not efficient given present preferences.
One obvious reason not to use a traditional efficiency criterion to determine a conception of justice is that the conclusion will beg the question if we accept the possibility of endogenous preference change. Let’s call this the problem of specious stability. If our current preferences are conditioned by our current institutions – e.g., we might come to prefer what our present system provides – our welfare criterion may lead us simply to reaffirm our current system in most respects. This is a recipe for complacency, ill-suited for the tasks of a normative theory. One thing we want from our normative theory is to distinguish between what we do and what we ought to. Someone may wish to argue that we could have better preferences, and if we did, we would endorse some other kind of system. This possibility shouldn’t be ruled out by fiat.
However, it is quite unlikely that our preferences will in general so easily shape themselves to our opportunities. So the problem of specious stability is a not the only problem. I think it is more likely that that a system of principles efficient given present preferences will not prove robustly stable – will not have the self-equilibrating properties Rawls thinks are necessary for a stable, well-ordered, and just society. Call the problem of unstable efficiency. If you consider preferences exogenous, then it seems that a system that is efficient in a Pareto sense should be stable. By definition, nobody is worse off, and so everyone has a stake in system. But if you propose endogenous preference change, a change that is Pareto-nice given present preferences may alter preferences in subsequent periods and you could get all sorts of inconsistency, cycling (a kind of unattractive stability), arms-races, or just chaotic cascades.
I think whether you are more troubled by the problem of specious stability or the problem of unstable efficiency will reflect your psychological theory.
Radical nuture types will get quite worried about the problem of specious stability because they’re worried that people can be indoctrinated into liking anything. Girls will think they need to be mothers because they are given dolls. Low-wage workers may become grateful for their low wages, threatening their motivation to negotiate or organize. Slaves may become accustomed to their chains. Etc. I think something like this is the default leftist position.
Nativists, those of us who think there a lot of built-in and not-very-malleable high-order preferences, may worry more about the problem of unstable efficiency, especially if they also think that some of these preferences are incompossible (impossible to jointly realize).
I think Rawls gives hints of being a nativist in this sense, especially when he compares the sense of justice to universal grammar. In fact, I think Rawls’s project is impossible without taking the ultimate structure of the sense of justice as exogenous. This is the only way the theorist sitting in his armchair could intelligibly use his own considered moral judgments as the basis for an argument to the effect that people raised under institutions designed to accord with our present sense of justice would recognize and affirm these institutions as just. If the institutions we regard as just altered the underlying structure of their sense of justice, rather than only its expression, then we could not say how they would regard those institutions (unless we had a fairly comprehensive theory of contextual variability of the sense of justice), and thus could not say that they would be stable in the way Rawls argues they would be.
The argument from Rawls’s left to the effect that the method of reflective equilibrium is too conservative, and thus that a theory of justice based on it will too complacently reflect current institutional structure, flows, I think from an implicit arguments about the limits of endogenous preference change in Rawls’s theory. By holding the structure of the sense of justice (call it “universal moral grammar”) fixed (although allowing endogenous change in the expression of its parameters), Rawls places fairly firm limits on what we can be socialized into regarding as moral.
Of course, we have to have some fixed standard (whether it is preference-based or not) in order to avoid thoroughgoing normative vertigo. Rorty, for example, reads his relativist crypto-Marxism into Rawls and interprets him as allowing the sense of justice to vary with institutional structure. But he doesn’t worry so much about the problem of specious stability. Rorty claims to be a pragmatist and pragmatists start from where we are. Rorty seems to think there is enough play in the method of reflective equilibrium to move us incrementally toward justice. However, there is also enough play in the method to move us toward totalitarianism, or, worse yet, laissez faire. So, may the best rhetorician with the best arbitrary commitments win.
Those unwilling to go the way of Rorty must hold something fixed. That’s your normative standard. The trouble I’m having with folks who seem obsessed with the problem of specious stability is that I can’t quite make out what they’re using as the standard by which they wish to evaluate the quality of our present preferences.
Well, what do you think of that? This concludes my musings for today…
[Please note: This is public exploratory reflection, not scholarship. Take my interpretative claims with a grain of salt. I do.]