Introduction to theories of linguistic and cultural relativism

Listen to audio

The idea that human thought varies according to the language we speak is frequently referred to as the Sapir-Whorf thesis. Edward Sapir (1884-1939), an American anthropologist and linguist, asserted that "human beings live not only in the objective world, nor only in the world of social activities as they are ordinarily understood, but they are at the mercy of the particular language that has become the medium of expression in their society." His pupil, the linguist Benjamin Lee Whorf (1897-1941), extended this thesis: "the linguistic system (...) of each language is not only an instrument of reproduction serving to vocalize ideas, but is itself a formator of ideas (...). We cut up nature according to the lines that our language has drawn on it"

This appealing thesis made its way into^{twentieth-century}Western culture. It plays a central role in George Orwell's novel 1984 (1949), in which a dictator shapes and imposes a new language, "Novlangue", with the aim of restricting freedom of thought. By gradually eliminating certain words from the vocabulary, Orwell writes, "the total climate of thought will be different. In fact, there will be no thinking as we now understand it." The Whorfian thesis also spreads to the general public through urban legends, such as that the Eskimo language has hundreds of words for snow, and that this is proof that the Eskimo people categorize the world idiosyncratically. This idea is a myth, as we explained in the lecture, based on Pullum's (1989) article.

The cognitive science community, however, remains considerably divided on the Sapir-Whorf thesis. For some researchers, such as Lera Boroditsky and Steve Levinson, language shapes or restructures thought: the words we use have a decisive effect on the way we conceptualize and interact with our environment. For many others, such as Leila Gleitman and Steven Pinker, the influence of language on non-linguistic processes is minimal or non-existent. Words are merely labels on non-verbal, universal concepts that we all share, with only the tiniest variations. Leila Gleitman, for example, argues that linguistic structures are only influential when the speaker chooses to adopt a verbal strategy - and of course, it's hardly surprising that language influences... linguistic processing in this way. In The language instinct, Steven Pinker takes up Jerry Fodor's thesis that there is a "language of thought" quite different from the language we use to communicate. According to this hypothesis, we don't think in French or Russian, but in an abstract and universal internal language, and it's only when it comes to communicating that we need to freeze it in a specific language.

In fact, we can schematically consider that there are three versions of the Sapir-Whorf hypothesis. According to the strong version, language intervenes directly to shape our mental representations. According to a weaker interpretation, language does not play a central role in thought, but provides an additional, more or less effective tool in our mental toolbox. Finally, according to a third interpretation, language would bias us only insofar as it provides a means of transmitting new ideas, including attentional or cultural biases: through the act of speaking, the speaker would draw the public's attention to this or that perceptual or conceptual distinction.

The rest of the first lecture was devoted to discussing this third interpretation, the reality of which is beyond doubt, but which is not really "Whorfian". First, let's remember that there are significant variations in vocabulary exposure from one person to another and that, from early childhood, these differences are reflected in changes in brain activity: in children aged between 4 and 6, activation of Broca's region depends directly on the number of conversational turns in which the child has been engaged (Romeo et al., 2018). Secondly, numerous experiments show that word choice can elicit significant cognitive and perceptual biases in the public. For example, in social psychology, the same cognitive activity (reproducing a figure from memory) may or may not induce gender biases against girls, depending on whether it is described as a "drawing" or "mathematics" test (Huguet et al., 2001). Several studies, notably those by Sandy Waxman, show that language is an amplifier of sexist or racial prejudices. In another field, the placebo and nocebo effects, or hypnosis, which are very real biological phenomena, illustrate the extent to which words can affect, for example, the subjective sensation of pain.

But are these "Whorfian" effects? Not really, at least not in the sense in which Whorf himself understood them. Firstly, they are universal effects, on a universal perceptual system - all speakers experience the placebo effect, whatever their language. Secondly, the transmission of ideas and knowledge is the strictly normal function of language, so it's hardly surprising that it changes our decisions. From birth, we often adhere closely to what others say, either because we think they know more than we do, or because we fear being wrong, or to preserve our integration into the social group. Brain imaging suggests that these linguistic transmission effects occur at a high level of cognitive, conscious processing, sensitive to attention and distraction, while low-level sensory areas are unaffected. Most low-level sensory computations are "cognitively impenetrable", i.e. they do not change according to the information we receive. Visual illusions, for example, continue to produce their effect even when we are informed of their nature.

The final pitfall of Whorf's hypothesis is that it is not always easy to separate the effects of language from those of culture. As Richard Nisbett's work illustrates, there seem to be subtle differences in cognitive style, for example between American and Chinese students, but it's not necessarily language itself that causes them, but rather cultural habits.