While some scholars argue that the Whorf hypothesis dates back to Aristotle’s Rhetoric or to German philosopher Gottfried Leibniz’s writings on language, we can safely start with Wilhelm von Humboldt,11 an early 19th-century German linguist and political theorist. Before and during his fieldwork, Humboldt often wrote about the relationship between language and thought. To him, language was not merely the means through which we conveyed what was going on in our minds. Instead, the language established a worldview: languages were the means through which we understood ourselves and the world.11
Humboldt’s ideas became influential in the late-19th century through the work of the German-trained Franz Boas:11 a professor at Columbia and founder of the American Anthropological Association.12 Boas’s work in linguistic anthropology (mainly on what we now call “Inuit languages”) followed Humboldt in arguing that different languages classify how we experience the world in different, subconscious ways.11,13 Crucially, though, he did not think that language determines how we view the world. Instead, he thought that our languages’ grammatical categories reflect the ways our culture classifies the world.13
Moving on to the early 20th century, one of Boas’s own students, Edward Sapir, would also be one of the main contributors to the development of the Whorf hypothesis. (This is why it is sometimes called the “Sapir-Whorf” hypothesis.) Sapir followed Boas in arguing that different languages classify how we experience the world, but he stressed that languages are complete systems, often untranslatable between each other.13 He also pushed further than Boas: he thought that language was necessary for us to fully develop the ability to think because our ability to think arises from our ability to interpret the language we speak.13 Different languages yield different interpretations, and those different interpretations place constraints on what we can think.13
This progressive strengthening of Humboldt’s original idea was finalized by one of Sapir’s students, Benjamin Lee Whorf.11 13 While not a professional linguist, Whorf was interested in documenting previous and current forms of the indigenous languages of North America, especially Nahuatl and Hopi. Whorf’s main contribution to the hypothesis was to point out that not all linguistic categories are overt; sometimes, a language encodes information implicitly. Whorf also accepted Sapir’s claim that languages place constraints on what we can think, based on the interpretations we give them. But because languages also marked things implicitly, these interpretations were widespread and pervasive — we didn’t have to actively use our language for us to be interpreting things through our language. As Whorf would put it:
“[…]users of markedly different grammars are pointed by their grammars toward different types of observations and different evaluations of externally similar acts of observation, and hence are not equivalent as observers but must arrive at somewhat different views of the world.”15
In the 1960s and onwards, with the rise of nativism in linguistics — especially Chomsky’s theory of Universal Grammar — the Whorf hypothesis began to come under scrutiny. It was believed that languages were just too similar to yield the kinds of effects on the thought that Sapir and Whorf hypothesized. Further empirical work also showed that the Whorf hypothesis, in its strong form, was shown to be flawed: humans and other primates display the ability to think without language5, refuting Sapir and Whorf’s claim that we needed to interpret the language to be able to think.
However, researchers in the 1990s started studying whether language still influenced thought in any interesting ways. Among other things, behavioral scientists began looking at language’s effect on color perception, spatial cognition, and more. Many studies suggest that language does have some effect on which kinds of processing are easier for a speaker.16, 17 The research on these weaker versions of the Whorf hypothesis is still ongoing, but many behavioral scientists— even ones who reject the stronger forms— accept one version or another.14