Jingzhi Chen

Timothy Williamson’s influential objection to the understanding-assent link (UAl) challenges the idea that understanding necessarily requires assent to particular sentences, thereby casting doubt on the possibility of a litmus test for understanding. In response, some philosophers argue that Williamson’s objection is limited in scope and UAl still holds in some cases, while others retreat to a weaker constitutive connection between understanding and assent. I examine both strategies and argue that both fail to preserve UAl. I then defend a normativist version of UAl, which holds that understanding requires not semantic assent—assent to purported meaning-constitutive sentences—but normative assent: assent to the social rules that prescribe such semantic assent. This view, I argue, vindicates a litmus test for understanding while avoiding Williamson’s challenge. More importantly, it better reconciles the two competing concerns that motivate both sides of the debate. It imposes weaker conditions on understanding than traditional UAl, thereby recognizing a broader range of speakers, including many radically deviant ones, as competent. At the same time, it preserves a constitutive connection between understanding and assent, thus identifying the point at which radical deviance amounts to a failure of understanding.