Linguistic relativity—the hypothesis that the language one speaks influences how one perceives and categorizes reality—has undergone dramatic swings in scholarly acceptance. Benjamin Lee Whorf's mid-twentieth-century formulation of the hypothesis, often called the Sapir-Whorf hypothesis, proposed a strong version: language does not merely color perception but actually determines the limits of thought. A speaker of a language lacking a word for a particular color, Whorf argued, would be literally unable to distinguish that color from adjacent ones. This strong version fell sharply out of favor after the 1960s as cognitive scientists demonstrated that speakers of languages without dedicated color terms could nonetheless reliably perceive distinctions among those colors when given nonlinguistic tasks.

Yet a weakened version of linguistic relativity has experienced a renaissance in recent decades. Researchers have shown that the presence or absence of grammatical categories—such as obligatory tense marking or grammatical gender—do correlate with measurable differences in how speakers organize memory and make inferences. In one widely cited set of experiments, speakers of languages that mark spatial relations primarily in absolute terms (north/south/east/west) performed differently from speakers of languages using egocentric terms (left/right/front/back) on nonlinguistic spatial-reasoning tasks, even when both groups had equal sensorimotor exposure to the task environment. The researchers concluded not that one linguistic community lacked spatial cognition, but that language primes particular cognitive strategies that then persist in nonlinguistic contexts.

Critics counter that correlation does not establish causation: communities that differ linguistically often differ in ecology, practice, and cultural routine, and any of these factors could independently account for the observed cognitive differences. Proponents respond that controlled studies isolating linguistic variables have replicated the core effects, though the magnitude remains debated. The field now broadly accepts a moderate position: language shapes, but does not wholly determine, the architecture of thought.