In recent decades, philosophy of language and philosophy of science have had little interaction. This paper challenges that separation by exploring how research on natural kinds in philosophy of science can inform debates about semantics in philosophy of language. The starting point of my analysis is Häggqvist and Wikforss's (2018) critique of the ‘referentialist revolution' that took place in the 1970s in the philosophy of language. They argue that this revolution — marked by a shift from descriptivism to direct-reference semantics — was based on an outdated commitment to micro-essentialism. Since micro-essentialism has faced extensive criticism within philosophy of science (Bursten 2014, Chang 2012, Needham 2011), Häggqvist and Wikforss claim that the assumptions underlying the referentialist revolution are untenable and that we should return to some form of semantic descriptivism, such as cluster-descriptivism.  

In this paper, I present an alternative reading of this debate. While I agree with Häggqvist and Wikforss's diagnosis of the problematic essentialist assumptions underlying referentialism, I argue that their conclusion is too strong. In particular, I argue that they consider only two possible responses: refining micro-structure as a placeholder or adopting Homeostatic Property Cluster Theory. I propose an alternative: a contextualist approach to natural kinds. This approach, inspired by Recanati's (2019) distinction between reflexive and referential levels in language, reconciles essentialist intuitions while allowing for a role for context. The instantiation of a natural kind results from the interaction of these two levels. Then, to instantiate a concept such as ‘water' or ‘tiger,' one initially considers contingent factors — such as the speaker's intentions or the context, which can vary between scientific and ordinary discussions. Once these factors are settled, reference becomes fixed. If this argument holds, then direct reference in semantics can be preserved, provided we embrace a contextualist account of natural kinds.