Does language influence the perception of emotions and colors ?

Listen to audio

The third lecture examined the influence of language on perception. There are two opposing theories. According to Whorf and a few other contemporary researchers, such as Lera Boroditsky, perception can be profoundly altered by the availability, or otherwise, of a word to describe it. On the contrary, according to Leila Gleitman and many other cognitive scientists, the initial stages of perception are modular, independent of language, and any linguistic effects appear only if participants rely on language to solve the required task. A key test distinguishes the two theories: according to the second, the possible influence of language must disappear if subjects are prevented from using their linguistic processes, for example by interfering with them.

We have examined the empirical literature in two areas: emotion perception and color perception. As far as emotions are concerned, Paul Ekman's work is the benchmark. Extending Charles Darwin's ideas in The Expression of Emotions in Man and Animals (1872), Ekman shows that all human groups express and perceive in the same way at least six basic emotions: anger, disgust, fear, joy, sadness, surprise - a list to which Ekman would later add a few other emotions such as contempt, shame, pride, guilt... Their expression on the face, and the ability to recognize them, would be universal. However, different cultures may modulate the expression of these emotions, or even repress it in certain contexts. The lecture examined in detail two recent articles (Jack et al., 2012; Jackson et al., 2019) that purport to challenge Ekman's work, but we saw that their results show only slight cross-cultural variations and, above all, provide absolutely no evidence that these differences are determined by the lexicon of the language rather than by other factors.

The situation is different in the field of color perception, which has been the subject of a great deal of rigorous experimentation since the seminal work of Berlin and Kay(Basic color terms: their universality and evolution, 1969). Berlin and Kay show that, while basic color terms vary from language to language, they form a universal hierarchy. Mathematical simulations, notably those by Regier and colleagues (Regier et al., 2007), indicate that color terms, far from varying arbitrarily from one culture to another, reflect the near-optimal subdivision of an irregular perceptual space, but one shared by all members of the human species. Depending on the number of words available, the division will be different, but it will respect the hierarchy of word introduction observed by Berlin and Kay: white and black, then red, green or yellow, blue, brown, and finally the less frequent words purple, pink, orange and grey. A strong universality therefore lies behind the apparent diversity of languages.

However, the central question of the Whorfian thesis remains: do these linguistic variations affect color perception? Early psychological experiments in this field showed that, in different tasks, participants' behavior varied according to the language they spoke. For example, if two colors have the same name in a given language, then speakers of that language tend to judge them as more similar, all other things being equal, than if the colors are given different names. They also tend to confuse them when asked to retain them in short-term memory. These differences appear quite early in childhood, when children acquire color terms. However, Leila Gleitman and, following her, Steven Pinker criticize these subjective or memory tasks. In fact, there is no evidence that the effect comes from perception, and not from later stages of linguistic recoding.

Take the example of Russian, which makes an obligatory distinction between goluboy (light blue) and siniy (dark blue). Recent work by Boroditsky and colleagues does indeed demonstrate differences between speakers of Russian and English in a perceptual judgment task (choosing which of two color tasks is identical to a third). Crucially, however, these differences disappear under verbal interference, when the subject is asked to recite numbers silently (Winawer et al., 2007). Fascinatingly, moreover, this effect only exists when colors are presented in the right hemifield, i.e. to the left, language-dominant hemisphere (Regier & Kay, 2009). Within the same person, color perception would therefore be " Whorfian" in the left, linguistic hemisphere, but not in the right!

Even if these experiments show that the linguistic level is involved in a perceptual decision, two interpretations remain possible: either there is an additional, purelybottom-up activation of the language areas, which contributes to decision-making; or, there is a genuinetop-down feedback of this linguistic activation on the perceptual stages - and only this second interpretation would be truly "Whorfian". The available data are limited and inconclusive. They do, however, show that top-down perceptual effects, if they exist, are minuscule: roughly speaking, we all perceive the world in the same way.

The main conclusion to emerge from this body of research is that color perception is highly universal across cultures. But beyond the perceptual level, when we need to categorize or retain perceptual information, we often rely on linguistic recoding to help or interfere, especially when the stimuli are very close to the perceptual threshold. Fundamentally, when the linguistic code is rendered unavailable (as in the case of an anomic patient, or after verbal interference in the normal subject), perception continues to be categorical, with only the boundaries becoming slightly blurred. Thus, the most plausible interpretation is that of Gleitman and Pinker: language appears as an additional cognitive tool which, contrary to Whorf's ideas, does not radically alter our perception.