Metaphysicians, ethical theorists and philosophers of law squabble endlessly about what it is for a person to act — or perhaps even to ‘will’ — more or less freely. A vital issue in this controversy is how we should analyse two obvious but surprisingly problematical contrasts. The first antithesis is between things we do because we are forced, and deeds we perform because we want to — sometimes after having discovered preponderant reasons in their favour. The other polarity is more (...) general. In most situations, if I act on my desire, I act more freely than if I had not had the desire. But what if my attitude is the product of childhood conditioning — or later brainwashing, brain surgery, hypnosis, behaviour modification, alcoholisim, narcotics addiction, neurosis, psychosis or worse? Then isn't my autonomy diminished? What is it about these latter desires, or their origin, that differentiates them from their unthreatening congeners? (shrink)

This paper is concerned with selecting an appropriate perspective from which to understand and evaluate social-decision procedures. Distinguishing between "agent rationality" and "option-rationality", the author argues that a rational agent may choose a social-decision procedure that is not itself agent-rational (but merely option-rational). The argument puts the voter's paradox in a context allowing evaluation of (a) its general import and (b) practical proposals for avoiding it in particular cases. Arrow's four conditions for a social-decision procedure are shown to have little (...) relevance to the understanding or evaluation of constitutions. The author concludes that the more fruitful perspective for discussing social-decision procedures is that of option-rationality rather than (as Arrow, Wolff, and others have supposed) that of agent-rationality. (shrink)