An attractive way to characterize substantive disagreement as opposed to merely apparent disagreement is in terms of noncotenability. Noncotenability is standardly described in terms of exclusion of joint truth in a broadly Kaplanian framework. Against this standard model, it has recently been suggested that this kind of truth-based noncotenability – sometimes called “referential” noncotenability – is not sufficient for substantive disagreement. The reason for this is that substantive disagreement often seems to require a more fine-grained understanding of content. In this vein, Verdejo and de Donato-Rodríguez (2021) argue that disagreement is best characterized at the level of intension conceived of as Fregean sense, or mode of presentation. This paper agrees that referential noncotenability is not sufficient for disagreement. However, it argues against the conclusion that disagreement is best described at the level of Fregean sense. Contending that neither a referential nor a Fregean characterization covers all intuitive cases and thus fails to get the extension of “substantive disagreement” right –, the paper recommends a conceptual ethics approach to the notion of substantive disagreement. It argues that considerations stemming from the debate about peer disagreement in social epistemology should guide our delineation of this notion in philosophy of language.