Recently, the hypothesis that purely evaluative disputes are metalinguistic negotiations has gained traction. The purpose of this paper is to resist a strong version of that hypothesis, and defend that some of those disputes are not metalinguistic negotiations. To defend that claim, I argue that metalinguistic negotiations have three linguistic properties that some purely evaluative disputes lack. First, in a metalinguistic negotiation it is always felicitous to embed the dispute-initial statement under the subjective attitude verb consider. Second, in a metalinguistic negotiation a speaker can reply to that initial statement using a metalinguistic comparative. Thirdly, metalinguistic negotiations address metalinguistic questions under discussion. Some purely evaluative disputes lack these properties, and therefore there are reasons to think that they are not metalinguistic negotiations
More details here