On the Meaning of ‘Ought’1

I begin here in §1 with some prima facie linguistic difficulties had by two prominent metaethical accounts of the meaning of normative terms – basically the accounts implicit in realism and expressivism – when these are applied in a straightforward albeit naïve way to the word ‘ought’. In §2, I explain what I take to be the standard framework in theoretical semantics for accounting for the semantic value of modal auxiliaries and discuss its (not completely straightforward) application to the word ‘ought’. Then, in §3 I evaluate this account in comparison to the more traditional metaethical views of the meaning of ‘ought’. The upshot of these initial sections of this paper is that, although the standard framework raises some issues that require further refinement, there is strong reason to view the word ‘ought’ as some kind of univocal sentential operator whose semantic function is not, in the first instance, to refer to obligations nor to express moral attitudes – as some metaethicists seem to think – but to affect the modality of a sentence. In my view, this syntactico-semantic hypothesis easily overcomes the linguistic difficulties besetting naïve applications of realism and expressivism to the word ‘ought’, and it constitutes a systematic application of a more general semantic framework that has the resources to cover all modal auxiliaries (and potentially much more). However, rather than taking that to settle metaethical questions about the meaning of ‘ought’, which are at issue between realists and expressivists, I want to argue in the later sections of this paper that it points us in a new direction for thinking about these questions. I explore this first in §4, where I offer a constructive reinterpretation of the debate about the meaning of normative terms like ‘ought’ between realists and expressivists. That is, I seek an interpretation on