Cognitive science is a study of human universals. This assumption, which we will refer to as the Newtonian principle (NP), explicitly or implicitly pervades the theory, methods, and prose of most cognitive research. This is despite at least half a century of sustained critique by cross-cultural and anthropologically oriented researchers and glaring counterexamples such as the study of literacy. We argue that a key reason for this intransigence is that the NP solves the boundary problem of cognitive science. Since studying the idiosyncratic cognitive features of an individual is not a generalizable scientific enterprise, what scale of generalization in cognitive science is legitimate and interesting? The NP solution is a priori-only findings generalizing to all humans are legitimate. This approach is clearly flawed; however, critiques of the NP fail to provide any alternative solution. In fact, some anti-NP branches of research have abandoned generalizability altogether. Sailing between the scylla and charybdis of NP and hermeneutics, we propose an explicit, alternative solution to the boundary problem. Namely, building on many previous efforts, we combine cultural-evolutionary theory with a newly defined principle of articulation. This framework requires work on any given cognitive feature to explicitly hypothesize the universal or group-specific environments in which it emerges. Doing so shifts the question of legitimate generalizability from flawed, a priori assumptions to being a target of explicit claims and theorizing. Moreover, the articulation framework allows us to integrate existing findings across research traditions and motivates a range of future directions. (PsycInfo Database Record (c) 2024 APA, all rights reserved).