Although philosophers have lavished attention on statements of the forms 'It ought to be (the case) that p' and 'This agent ought to do A', they have rarely considered the more general forms 'Xs ought to F' and 'An X ought to F.' The exceptions do little but prove the rule.1 This neglect is, I think, ill-advised. In this essay I wish to draw attention to some of the interesting features of such statements and point out how such a model offers an appealing third alternative to attempts to understand moral 'ought' judgments as either hypothetical or categorical imperatives.
Roles and 'ought'-statements
Anscombe writes: 2
The terms 'should' or 'ought' or 'needs' relate to good and bad: e.g. machinery needs off, or should or ought to be oiled, in that running without oil is bad for it, or it runs badly without oil.
Here, I think, she is trying to make the point that the fact that lacking a certain feature makes something do its work badly shows that it ought to have that feature. Unfortunately, she falls short of the mark by stopping at the claim that it ought to be made to have that feature by someone. This still seems to be a statement about what someone ought to do or, perhaps, about what it would be desirable for someone to do;it is not yet a claim about what features the thing itself ought to have. But the insight she almost expressed is important. Sharpness makes a knife cut well and dullness keeps it from doing so. Sharpness is therefore a virtue in knives by Aristotle's account, for it makes the knife and its work good. (N.E., 1106a). It is fitting and appropriate to the work of a knife that it be sharp; it is better (with respect to its work) that it be sharp than not be. If contemporary philosophers are right in thinking that we use 'ought'-statements to say that something is fitting, or appropriate, or better, 3 this re-enforces the conclusion that the features that make