Cardinal preferences: Harsanyi's rule utilitarianism
Harsanyi's rule utilitarianism (RU) combines Benthamism with the Humean view of preferences. It allows interpersonal comparisons of utilities. According to Harsanyi (1976, p. 50), 'the ultimate logical basis for interpersonal utility comparisons lies in the postulate that the preferences and utility functions of all human individuals are governed by the same basic psychological laws'. Nevertheless, RU does not identify utility with pleasure, and claims that a moral value judgement is a judgement of preference. Since preferences are assumed to be cardinally measurable, utility is identified with cardinal preferences.
The ultimate moral value is the welfare of individual members of a society. Each agent expresses his or her own preferences not only in his or her personal or individual utility function (actual preferences), but also in his or her personal social welfare function (moral preferences). Individual preferences are defined in terms of personal and partial criteria, while moral preferences are identified as the hypothetical preferences that each agent 'would entertain if he forced himself to judge the world ... from an impersonal and impartial point of view' (Harsanyi, 1976, p. ix)3. In this way the social good is not an objective quantity, but is fully determined by subjective preferences. More specifically, Harsanyi makes reference to the choice of a moral code which has the social function 'to enjoin people to do certain things and not to do some other things'. Therefore, individual rights and special obligations may be established without requiring the SRC principle to be observed. For example, it can satisfy the demand of justice (Harsanyi, 1976, p. 74).
Harsanyi's idea of moral or ethical preferences is used by the ethical social choice theory, according to which ethical preferences are included in sustainable growth models as moral duties (Asheim, 1996; Asheim, Buchholtz and Tungodden, 2001; Marzetti, 2007). Nevertheless, as regards practical applicability, Harsanyi (1986, p. 60) himself maintains that 'rule utilitarianism . is not a criterion always easy to apply in practice'. In the real world, value judgements concerning social welfare may not be of the moral preference kind: in practice an agent is unlikely to choose a particular action in complete ignorance of his or her personal position.