Abstract
Epistemicism explains ignorance due to vagueness through semantic plasticity: the propensity of intensions of vague terms to shift across close linguistic communities. In the case of moral vagueness, e.g. when it's vague whether it's permissible to terminate a pregnancy after a certain number of days, epistemicism predicts that 'permissible' denotes distinct properties in different close linguistic communities. This epistemicist prediction has been pressured by arguments due to Miriam Schoenfield (Ethics 126: 257-282, 2016) as well as certain interpretations of the Moral Twin Earth cases. Schoenfield (Ethics 126: 257-282, 2016) argues that epistemicist account of moral vagueness leads to an unfeasible treatment of moral deliberation. A related worry comes from the Moral Twin Earth cases, which produce the intuition that the reference of moral terms such as 'permissible' remains stable across different linguistic communities. The problem for epistemicism is that metasemantic models that are meant to account for the Moral Twin Williams (Philosophical Review 127(1): 41-71, 2018) or Billy Dunaway and Tristram McPherson (Ergo 3(25): 239-279, 2016), predict that moral vocabulary is stable, which makes them incompatible with epistemicism. My aim is to make use of the inferentialist metasemantic framework presented by Robbie Williams (Philosophical Review 127(1): 41-71, 2018), and I refine it to give an epistemicist account of moral vagueness.