Topics Ethics

Question I recently read the following argument on a blog, and I was wondering what the panelists might say about it. It is a well known philosophical principle that one cannot infer normative facts from empirical ones (this is the is-ought problem). But if, as it is often supposed, "ought implies can," then cannot implies ought not ("ought not" in the sense of "not obligatory"). In that case, we can infer normative facts from facts about empirical facts about what people cannot do.

Accepted:
December 9, 2008

Comments

Right. So either the philosophical principle is incorrect or the "ought implies can" is incorrect.

It's an interesting question whether the existence of an is-ought (or ought-is) gap is inconsistent with the voluntarist principle that ought implies can. I think the answer depends on how we understand the gap. You frame it in terms of inference -- you can't infer an ought from an is. Others frame it in terms of validity, entailment, or logical consequence -- on one formulation, no is statement entails an ought statement. If we focus on entailment, we may need to distinguish different conceptions of entailment. On a metaphysical or modal conception, one statement entails another just in case there is no possible world in which the former is true and the latter is not. But sometimes we understand entailment as this modal relation obtaining by virtue of the logical form of the argument (If P, then Q; P; therefore, Q). This requires a syntactic or epistemic conception of entailment. With this distinction in mind, we can see that whether voluntarism is compatible with the gap may depend on which conception of entailment we have in mind. Suppose that if voluntarism is true, then it is necessarily true (true in all possible worlds). On the metaphysical or modal conception of entailment, voluntarism implies that an ought does entail an is (because oughts entail cans). But if we have in mind the syntactic or epistemic conception of entailment, then voluntarism is arguably consistent with the gap. On this conception, the gap says that you can't validly infer an ought statement from a set of exclusively is statements and that you can't validly infer an is statement from a set of exclusively ought statements. Because even if voluntarism is (necessarily) true, there is no valid argument form directly from the ought claim to the possibility claim. Any such argument would be enthymematic, with the missing premise being the voluntarist principle itself. But then a defender of the syntactic or epistemic conception of the gap will say that such arguments confirm, rather than disprove, the gap, because they are valid only once the missing voluntarist principle is supplied, and that principle, they will claim, bridges is and ought. So the valid inference to an is comes from (bridge) premises that mix ought and is. Anyway, that's a more complicated take on your question about the is-ought gap and voluntarism.

What a fun question!

Suppose we agree that if X is something we ought to do, then X is also something we can do. Suppose further that X is not something we can do. Then as your blogger points out, it follows that X is not something we ougt to do. But that's perfectly consistent with there being nothing we ought to do. It's perfectly consistent with saying that there's no such thing as moral obligation.

Compare: if there is such a thing as a necessary being, then that being can't be an ordinary space-time being. Since my desk is an ordinary space-time thing, it follows that my desk isn't a necessary being. (This is a point that gos back at least to Anselm.) But that's consistent with saying that at the end of the day, the idea of a necessary being is incoherent and that nothing is or even could be a necessary being. Likewise, saying that if there's anything I'm morally obliged to do, it must be something I can do is consistent with saying that on further analysis, the idea of moral obligation is incoherent.

And so we don't really get from a non-normative fact (my inability to do X) to a normative fact. The fact that I'm not morally obliged to do X needn't be understood as a normative fact.