Skip to content
mentua.org

What Is the Sapir-Whorf Hypothesis?

Line drawing of a filled square above two empty squares that sit on either side of a dotted line, which also divides a row of eight small squares belowLine drawing of a filled square above two empty squares that sit on either side of a dotted line, which also divides a row of eight small squares below

5 min read≈ 1,200 words3 questions · up to 3 points

What this is about

Russian has no single everyday word for blue, and an Aboriginal language in Australia does without left and right. You learn what the Sapir-Whorf hypothesis claims, how one experiment measured the effect of two words for blue, and which famous examples turned out to be myths.

In a nutshell· skip ahead

The Sapir-Whorf hypothesis, also called linguistic relativity, is the idea that the language you speak influences how you think and perceive. The strong version, that language limits what you can think, is rejected today. A weaker version has good evidence: the categories a language forces you to use can make some distinctions faster and more habitual.

In Russian there is no single everyday word for blue. A Russian speaker has to choose between goluboy for lighter blues and siniy for darker ones, in the way an English speaker has to choose between pink and red. So when a Russian and an American look at the same two shades of blue, does the language each of them grew up with change what they see?

What does the Sapir-Whorf hypothesis claim?

The Sapir-Whorf hypothesis, also called linguistic relativity, is the idea that the language you speak influences how you think and perceive. It comes in two versions. The strong version, linguistic determinism, says that language sets the limits of thought, so that you cannot think what your language has no words for. The weak version says only that a language makes some distinctions easier, faster or more habitual than others.

We dissect nature along lines laid down by our native languages.

Benjamin Lee Whorf, “Science and Linguistics”, Technology Review, 1940

Whorf meant this strongly: in the same essay he wrote that speakers of very different languages are not led by the same evidence “to the same picture of the universe”. Today almost no researcher defends the strong version, so the real question is how far the weak one reaches.

How can you test whether a word changes what you see?

The clearest test so far is a 2007 study by Jonathan Winawer, Lera Boroditsky and four colleagues, who compared Russian and English speakers in four steps.

  1. Twenty shades. The researchers used twenty shades of blue, from light to dark. At the end, each person sorted them, which showed where that person’s line between goluboy and siniy lay.
  2. A simple matching task. Each trial showed three squares, one on top and two below, and people pressed a button for the bottom square that matched the top one. All three squares stayed on the screen, so nobody had to remember a color.
  3. The result. When the two bottom squares were close shades on opposite sides of the person’s boundary, Russian speakers answered in 1,164 milliseconds on average. When the two shades were equally close but on the same side, they needed 1,288 milliseconds, so crossing the line made them 124 milliseconds faster, about a tenth of the time. English speakers, who call all these shades blue, showed no such advantage.
  4. The interference test. When people silently kept an eight-digit number in mind while matching, the Russian advantage disappeared, but when they kept a pattern of squares in mind, it stayed.

The fourth step matters most. If Russian eyes saw blue differently, a number held in memory would not change anything, so the most likely explanation is that the words goluboy and siniy were active in their minds while they looked. The effect is also small: Russians can see that two dark blues differ, and English speakers can tell light from dark blue, but one group is a little faster at one particular line.

Space gives a stronger example. Guugu Yimithirr, an Aboriginal language spoken at Hopevale in northern Queensland, does not use left and right to say where things are. Its speakers use four words similar to north, south, east and west, as the anthropologist John Haviland documented. In memory tests published by the linguist Stephen Levinson in 1997, Guugu Yimithirr speakers mostly remembered arrangements of objects by compass direction, while Dutch speakers remembered them by left and right. In Pormpuraaw, another community on Cape York, Boroditsky and Alice Gaby found in 2010 that people laid out a sequence of events from east to west, whichever way they were facing.

What the Sapir-Whorf hypothesis teaches you

The hypothesis teaches you to ask which distinctions your words make easy, and which ones you pass over because you have no ready word for them.

  • Learning a language or a field. The words of a new language or a new subject bring new categories, and practising them until they come without effort is part of learning to notice what experts notice. What Is Spaced Repetition? shows how to make new words stick.
  • Words that change behavior. Whorf worked as a fire prevention engineer, and he described workers who were careful around “gasoline drums” but smoked near “empty gasoline drums”, although the empty ones contain explosive vapor. Psychologists found a related effect in the same language, which “Thinking, Fast and Slow” by Daniel Kahneman: Summary and Review calls framing: people choose differently when the same outcome is described as lives saved or as lives lost.
  • Claims about other cultures. When you read that some people “cannot think” about something because their language lacks a word, ask whether anyone measured a difference in a task or only counted words.

Where does the hypothesis go wrong?

The strong version fails on a basic fact: people can think about things their language has no word for. They translate, borrow and invent words when they need them, and babies reason about objects long before they can speak.

Two famous examples also turned out to be wrong. Whorf claimed that Hopi, a language of Arizona, had no words or grammar that refer to time the way English does. In 1983 the linguist Ekkehart Malotki published Hopi Time, with hundreds of Hopi sentences that use exactly such expressions.

The second example is snow. The anthropologist Franz Boas mentioned four Eskimo word roots for snow in 1911, Whorf’s 1940 essay made the list longer, and later retellings grew the number into dozens and even hundreds. Laura Martin in 1986 and Geoffrey Pullum in 1989 traced this growth and pointed out that in languages that build long words from roots, the number of “words” for anything is almost unlimited. What Is the Golden Ratio, and Is It Really in Art? shows the same pattern with famous paintings.

Newer findings on grammatical gender have not held up well either. Studies in the early 2000s reported that German and Spanish speakers think about objects such as a bridge, feminine in German and masculine in Spanish, in line with the gender of the noun. Careful attempts to repeat two of these experiments, in 2014 and in a preregistered study in 2022, found no clear effect. The strongest objection to the whole field is that the surviving effects are small and depend on the task. That fits language as a tool used in the moment better than language as a lens that changes the world you see.

Learn more

The hypothesis has no single author, and Sapir and Whorf never stated it together; the name was given later. Edward Sapir, an American linguist and anthropologist, argued in 1929 that the language habits of a group shape how its members experience the world. Benjamin Lee Whorf, born in 1897, trained at MIT and joined the Hartford Fire Insurance Company in 1919. He studied linguistics with Sapir at Yale in the 1930s while keeping his job, and died in 1941.

For a readable book on the evidence, from color words to compass directions, try Through the Language Glass by Guy Deutscher (Affiliate Link).

Quiz

3 questions · up to 3 points

Open a question and pick one answer. Only your first answer counts: +1 if it is right, −1 if it is wrong.

  1. What is the difference between the strong and the weak version of the Sapir-Whorf hypothesis?
    Show answer

    The strong version says language limits what you can think; the weak version says it only makes some distinctions easier or faster. The strong version is called linguistic determinism and is rejected today, because people can think about things their language has no word for. The weak version has experimental support.

  2. In the Russian blues study, Russian speakers lost their speed advantage at the light-blue and dark-blue boundary in one condition. Which condition was it, and what does that suggest?
    Show answer

    While they silently kept an eight-digit number in mind, which suggests they were using the words for blue in the moment. A task that kept their verbal memory busy removed the advantage, while a spatial task did not. So the effect seems to come from words used during the task, not from a permanent change in the eyes.

  3. What is the best summary of the claim that Eskimo languages have dozens or hundreds of words for snow?
    Show answer

    It grew in retelling from Franz Boas's mention of four word roots, and counting words is unclear in languages that build long words from roots. The anthropologist Laura Martin and the linguist Geoffrey Pullum traced how the number grew from Boas's 1911 example. A language that adds endings to a root can form a very large number of words from almost any root.

The takeaway

The Sapir-Whorf hypothesis says that the language you speak influences how you think. Its strong version, that you cannot think what your language has no words for, is wrong, and the claims about Hopi time and Eskimo snow words did not survive checking. A weak version has support: Russian speakers, who must name light and dark blue differently, told such shades apart about a tenth of a second faster, but not while their verbal memory was busy. So learning words for a distinction helps you notice it.

References

  1. Russian blues reveal effects of language on color discrimination. Jonathan Winawer, Nathan Witthoft, Michael C. Frank, Lisa Wu, Alex R. Wade and Lera Boroditsky, Proceedings of the National Academy of Sciences 104 (19), 7780–7785, 2007

    The full study behind the worked example: the twenty shades of blue, the matching task, the reaction times and the verbal and spatial interference tasks.

  2. Science and Linguistics. Benjamin Lee Whorf, Technology Review 42 (6), reprinted copy hosted at MIT, 1940

    Whorf's own short essay for a general audience, with the line about dissecting nature, his principle of relativity and the snow example.

  3. Guugu Yimithirr Cardinal Directions. John B. Haviland, Ethos 26 (1), 25–47, 1998

    An anthropologist's detailed account of how speakers at Hopevale describe position and movement with four words similar to north, south, east and west.

  4. Language and cognition: The cognitive consequences of spatial description in Guugu Yimithirr. Stephen C. Levinson, Journal of Linguistic Anthropology 7 (1), Max Planck Institute for Psycholinguistics, 1997

    The experiments that compared Guugu Yimithirr and Dutch speakers on memory for arrangements, with a link to the full paper.

sapir-whorf-hypothesislinguistic-relativitylinguistic-determinismlanguage-and-thoughtbenjamin-lee-whorfedward-sapirlinguisticscognitive-sciencecolor-perceptionrussian-bluesguugu-yimithirrhopi-timeeskimo-words-for-snowgrammatical-genderreplicationmyths