Matt makes the case that utilitarians have to count other regarding preferences, such as the squeamishness of some heteros about open discussion of homosexuality, in assessing public policy. In a stripped down sense, this claim is clearly right: these kinds of preferences can’t be per se left out of a utilitarian calculus at the highest level. But there are plenty of perfectly good utilitarian reasons for a general presumption against taking them into account for the purposes of on-the-ground policy formation.
For one, these kinds of preferences are pretty clearly not entirely independent of the law. It is, for example, hard to preserve a strong “ick” response to something that’s always been routinely and openly discussed. Most Americans have an “ick” response to the thought of eating insects; people who grow up in cultures that munch on bugs routinely don’t.
More generally, the presumption built into the legal regime can affect whole classes of preference formation. That is, when private behavior is taken off the table of legal control, you’re more likely to get a background cultural presumption of tolerance, such that people are encouraged from a young age not to concern themselves overly with other people’s private behavior (the attitude of Jefferson; I’ve always loved his line: “It does me no injury for my neighbor to say there are twenty gods or no God. It neither picks my pocket nor breaks my leg.”) Let policy be shaped by other-regarding attitudes, and I suspect people will feel more justified in nurturing their indignation and disgust. In fact, there’s a sort of perverse incentive to overstate one’s own level of revulsion and to attempt to persuade others to share it, so as to get more “points” in the felicific calculus. To the extent that it’s coherent to do a second-order utilitarian evaluation of which preference sets are desirable (it is if you’re a hedonistic utilitarian; it probably isn’t if you’re a preference-satisfaction utilitarian, and I don’t know where Matt falls) then it seems pretty clear that, for instance, sadistic preferences are suboptimal. You’d rather that people not tend to develop zero-sum preferences if it can be avoided.
There’s a separate stability argument to be made, the general outline of which I assume is pretty obvious: attitudes like this change or reemerge over time, and there are disadvantages to leaving open the possibility that some practice you’ve made central to your life plan could be banned at any time if the winds of popular sentiment begin blowing the wrong way.
Of course, I’m not a utilitarian, so I tend to think these considerations are beside the point—they give you the right answer for the wrong reason. But just as Mill discovered that trying to be too much of a direct utilitarian in your private life ends up being suboptimal from a utilitarian perspective, it seems highly probable that a straight utilitarian calculation taking other-regarding attitudes as given would lower aggregate utility… well, to the extent that you think the concept of “aggregate utility” makes a whole lot of sense, which I usually don’t either.