A Critique of Preferentism

A previous essay suggested ways in which preference-satisfaction utilitarianism (“preferentism”) is superior to forms of utilitarianism that focus on promoting or maximizing desirable states of consciousness (such as pleasure). Preferentism is better equipped to handle cases where it’s intuitive that someone is wronged even if they’re never aware of it, such as when a comatose patient’s boundary is crossed.  

Preferentism starts its framework with equal respect for informed preferences–each person is counted equally, and no informed preference is deemed “better” than others on some independent grounds. Some people likely have more numerous and/or intense preferences than others, and calculating how to balance these brings complications I’ll leave aside, but the main point is that respect plays a vital role at a fundamental normative level. Interestingly, respect is a mode of responding to value not readily associated with moral views focused on promoting value, that is, bringing about more of that value as a goal. 

By emphasizing individual preferences over controversial (rankings of) substantive values, preferentism gives greater emphasis to respecting individual agency than do views which focus solely on promoting values or states of affairs themselves, independently of regard for the agent instantiating these values or states of affairs. 

However, respect for preferences may raise tensions with preferentism’s attention to consequences. Here Robert Nozick’s distinction between maximization of rights and “Kantian” side constraints is instructive. On any utilitarian view, even respect for preferences (or rights), if it is a thing, is something to be maximized all else equal. This means we can violate respect for preferences (rights) if we anticipate doing so will minimize further violations of preferences (rights) better than respecting them as a matter of principle.  By contrast, a side-constraint view of respect admits no exceptions even if disrespect sometimes maximizes overall respect. On a side-constraint view, morality sometimes involves something other than promoting a goal, such as following a rule to respect x regardless (largely) of outcomes from respecting x. 

Perhaps preferentists, and utilitarians generally, should abandon talk of respect and advocate for “promotion all the way down.” However, this comes at the cost of equality, as some people’s preferences will be deemed more worthy of promotion. (Who decides worth and by what authority? Must they use a standard of “better” other than preference-satisfaction?) Or perhaps they should favor the more informed, more identifiable, or more easily realizable preferences. (Do we presume preferences on behalf of the less-informed? Which way do we correct for any inconsistencies, and wouldn’t this correction risk substituting the interventionist’s judgment for that of the agent with the assumed preference?)

On the other hand, “respect preferences equally, then promote accordingly” is arbitrary. Why is respect the appropriate mode at this fundamental level but then replaced by promotion downstream? Equal respect for individual preferences is not itself called for by utilitarian considerations but is rather a formal element that motivates maximization on some views. But why have that formal constraint if value maximization is the endgame?  

More pressing, why respect an agent’s preferences, or the content of those preferences, but not the agent herself? Mustn’t we treat preferences as free-floating to some degree, as other forms of utilitarianism treat their maximands, in order to calculate how to aggregate them? If so, respecting only the preferences doesn't capture our conceptions of ourselves and other moral participants as inviolable.

The preferentist may respond that it would be really bad if coercive institutions focused directly on maximizing respect of whatever kind, so it’s better not to let them have that kind of power. Maximizing preference satisfaction should be a criterion for evaluating how well a society is doing morally, but it shouldn’t be an action guide for determining specific policies. Maybe our best bet is largely to disregard attention to preferentism because this is the best available way of promoting preference satisfaction! 

  

I’d argue that it’s generally wise not to presume that one’s ethical theory can and should directly inform specific coercive policies. Still, does the preferentist framework suffer from issues other forms of utilitarianism face–namely, paying insufficient attention to respect for persons and agency? Preferences may not simply be viewed as self-contained entities to be aggregated. Rather, they manifest a process–namely, the agent’s ongoing decisions and judgments (active or latent) whether to form, maintain, revise, or discard this or that preference based on their beliefs and values. In respecting preferences, we should thereby respect the sources of those preferences. But if we respect agents as the sources of these preferences, we acknowledge agents’ inviolability to this degree, in which case it becomes difficult to see why maximization–whether as an ethical ideal or policy goal–should comfortably enter the scene thereafter. Maybe it should, but a further case is needed.

Until then, respect for individual agency is too important to be maximized, paradoxical as that might sound in a utilitarian mindframe.   

All that said, the numbers may count sometimes. Respect for agents is the default stance, but it can be overridden in circumstances where respect becomes supererogatory. Rights respect though the heavens may fall suggests morality should be done for its own sake, regardless. Rights have a function, which in one plausible view is to allow agents jurisdiction. But Hume’s “circumstances of justice” also apply. Rights are non-starters in a world of extreme scarcity or dire emergency, for instance. Someone would likely find his sacrifice more easily justified to prevent the world from ending or some other “moral catastrophe.” However, it’s much harder for someone to find his sacrifice justified if doing so helps GDP net a few million more dollars, allowing a modest net increase in overall preference satisfaction.