Utility is a technical concept and a theoretical one, undiscussible in isolation from the theories in which it has its being. No more can the possibility of interpersonal comparison of utilities be discussed in isolation from theories in which such comparisons play a role. Thus, the problem is undiscussible where the theoretical context consists solely of the von Neumann–Morgenstern theory of personal preference, for nothing there depends on the possibility of interpersonal comparison of utilities. Nor does anything in the von Neumann–Morgenstern theory exclude the possibility of such comparisons. The contrary impression, which is widespread, is based on the unfounded assumption that the theory of personal decision making is the definitive context for the utility concept.

In this paper I consider the problem of interpersonal comparison of utilities in the context of a theory (the New Utilitarianism) which is obtained by substituting the von Neumann–Morgenstern utility concept into (say) Bentham's normative political theory. In that context, interpersonal comparisons of utilities become interpersonal comparisons of preferences: “Individual i's preference for A over B exceeds (or is exceeded by, or is the same as) individual j's preference for C over D.”

It is my contention that, commonly enough, we do make such intercomparisons, and that the theoretical apparatus of the New Utilitarianism can serve to test and to supplement them by bringing them into confrontation with each other and with judgments about even-handedness of compromises. But this is no brief for the New Utilitarianism as a political or economic norm.