2. 1. ETHICAL theories (in the sense of theories about the meaning and logical properties of moral terms) fall into two main classes: descriptivist and non-descriptivist ( SOE 3. 4). According to the former, moral words like 'good' and 'wrong' resemble ordinary descriptive words like 'red' and 'rectangular' in this respect, that their meaning and their application conditions are firmly linked, so that to change either is to change the other. Victims of the 'descriptive fallacy' ( Austin 1961: 234; 1962: 3) think that the same has to be true of all words; but this has been shown not to be the case. Imperatives are the obvious counterexample. Prescriptivists took their cue from an observation about these. The application conditions of imperatives clearly vary independently of their meaning. If I tell someone to shut the door and you tell him not to, we may still be meaning the same by the imperative 'Shut the door', which I affirm and you negate.
Prescriptivism holds that, in addition to any factual or descriptive meaning they may have, there is a prescriptive element in their meaning which is irreducible to any factual equivalent, but serves to prescribe or direct actions. This prescriptive meaning is often likened generically to that of imperatives, which also prescribe actions; but careful ethical prescriptivists emphasize also the differences, as well as the similarities, between moral statements and imperatives. Most prescriptivists did not completely assimilate moral judgements to imperatives (obviously they are very different in some respects); but in this respect they might be similar.
2. 2. The chief of these differences is that moral statements also have, unlike imperatives, a descriptive or factual content. This is linked with another feature of moral statements, known in the jargon as universalizability: the feature that a moral statement about one situation must apply to any other situation similar in all its universal properties____________________