Several arguments in the literature suggest that the following three theses are incompatible: robust moral realism, epistemicism about vagueness, and the claim that basic moral terms are vague. This paper explores the viability of resolving this trilemma by rejecting moral vagueness.

Such a rejection is no easy feat because vagueness is seemingly infectious: in many cases, the vagueness of a descriptive statement threatens to directly entail the vagueness of a materially equivalent moral statement. For example, suppose I promise to give my friend ten pounds if they date a bald person. Following the promise, I am required to give my friend ten pounds if and only their date is bald. The worry is that the vagueness of the descriptive statement (‘The date is bald’) infects the moral statement (‘I am required to give my friend ten pounds’), and thus the moral statement, and consequently the moral term ‘required’, must be vague as well.

However, I argue that epistemicism has the resources to respond to this problem of infectious vagueness. In previous work, I argued that epistemicists should reject a key principle in the logic of vagueness: Distribution, which states that if a biconditional is definitely true, then if one side of it is borderline, then so is this other. Here, I expand that argument to show how epistemicists can coherently reject not only the principle in general, but the particular instances of it that connect the moral and descriptive statements. This allows epistemicists to maintain that even when the relevant descriptive statement is borderline, the materially equivalent moral statement can nevertheless be definite and non-vague, thus allowing the epistemicists to reject moral vagueness and uphold moral realism.