Long ago I tried to rescue No-Ought-From-Is from Prior’s counterexamples by reformulating it as the thesis that you can’t get a non-vacuous Ought from an Is. But (replies Nelson) once we help ourselves to the notion of truth we can construct logically valid arguments from non-moral premises to (non-vacuous) moral conclusions. Is Nelson’s counterexample logically valid? This depends on the nature of truth. If truth is transparent then his inference is valid but it is not a counterexample to No-Non-Vacuous-Ought-From-Is. The ‘ought’ in his premises appears non-vacuously and is used as well as mentioned. Suppose we adopt a non-transparent theory of truth according to which we don’t officially know what the quoted statements mean? Taking my cue from Ramsey and Buridan I develop a formal theory of truth that relies on the notion of representing that. On this conception, Nelson’s argument is invalid but can be restored to validity by adding an extra premise; a premise, however in which ‘ought’ appears non-vacuously. So whether we adopt a transparent or a non-transparent theory of truth, No-Non-Vacuous-Ought-From-Is still stands which means that you can’t use truth to break down the barrier between substantively non-X premises and substantively X-conclusions.
