Against Conditional Obligation

Lennart Aqvist (1984) explains why most contemporary deontic logics use a primitive binary conditional obligation operator. He then points to an account of prima facie obligation as the primary outstanding problem facing these logics. Solving that problem, I hope to show, also solves the puzzles that motivated such theories in the first place. The crucial feature of obligation sentences to which the puzzles point is that such sentences, and evaluative sentences more generally, are defeasible. They may be warranted, given some information, only to be defeated by further information. A theory that recognizes this no longer needs to see conditional obligation as anything more than a simple combination of unary obligation and the conditional. My title may thus be overly provocative. I do not mean to deny that some obligations hold only if some condition is fulfilled. Nor do I mean to deny that sentences Kant would have called hypothetical imperatives, such as If you like Mexican food, you should try Jorge's, express such obligations. Finally, I do not mean to deny that there is anything wrong with the common strategy of using a primitive conditional obligation operator such as O(qlp) in a logical or philosophical theory of deontic concepts. It usefully avoids commitments about the relative roles of obligation and the conditional. What I do want to attack is the idea that a theory with a primitive conditional obligation operator-a binary theory, as I shall call it-can be afinal logical or philosophical theory about deontic concepts. At the very least, a theorist using a conditional obligation operator owes us an explanation of how the semantics of the operator depends on the semantics for obligation and the conditional simpliciter. Sentences expressing conditional obligations are intelligible to anyone understanding should (or ought to) and if. The combination of these words is no idiom. The meanings of such sentences, therefore, should be explicable in terms