ABSTRACT
Max has vitiating implications, both logical and practical. So does its most natural generalization to nonlinear value structures: the principle that we may choose any unexcelled outcome. There is, however, is a broad class of alternatives to both principles that falls under the general heading of “satisficing.” All forms of satisficing overlay a value structure with an adequacy criterion that is foreign to it. These adequacy criteria are, strictly speaking, not axiological but decision-theoretic in character. They must have a definite “shape” when graphed in Cartesian coordinates, and that “shape” must be adjustable to the value structure to make the adequacy criterion more or less strict. The “shape” itself is constrained by the relative importance of the various dimensions of value in a decision—and, more problematically, by the indeterminacy of that relative importance. It can, however, be designed to reduce vulnerability to bias and self-interested manipulation (which are obvious desiderata for ethical decision-making) by two procedures: prior calibration of the units of measurement and making the “shape” hyperbolic—or, in many dimensions, hyper-hyperbolic. The resulting decision procedure, which I call hyperbolic satisficing, can be applied both to decisions under certainty and decisions under risk, and it is ethically superior to the looser procedures of contemporary multicriteria decision analysis.
