The languages we speak help construct our inner worlds as well as describe them. A century-long scientific debate has begun to settle on an answer: language nudges habitual thought without imprisoning it.
There is a feeling you have probably had. You are remembering something — a place you loved, a version of yourself that no longer exists, a person who is gone. The memory carries warmth, but the warmth is inseparable from the ache of distance. You don’t want the feeling to stop. You want to lean into it, the way you might lean into a piece of music that makes you sad in a way that feels almost like joy. You sit with it.
Portuguese speakers call this saudade. English has no single word for it: “nostalgia” implies a wish to return, and “longing” is too thin. In everyday use saudade means fondly missing someone or something absent, together with the sorrow the lack brings; at its fullest it is the bittersweet pleasure of loving something in its absence, knowing the absence may be permanent. Fado, Portugal’s best-known song tradition, is built around it.
The lack of this word in English does not stop English speakers from feeling what Portuguese speakers feel. You likely recognized the feeling in the opening paragraph before you read the word for it.
The question that has occupied linguists and cognitive scientists for the better part of a century is whether the difference between a language that has this word and one that doesn’t is merely a matter of translation efficiency, a label missing from one toolkit, or something deeper. Does the word shape the feeling itself? Does the vocabulary you were handed at birth structure what you notice and what you feel?
The Vocabulary That Doesn’t Travel
German gives us Schadenfreude, pleasure at another’s misfortune. English borrowed the term because it needed it; there was no native equivalent for something English speakers demonstrably feel.
What the word adds is the ease of packaging the feeling, calling it up in conversation and sharing it with a knowing look. It makes the feeling common currency in a way it isn’t when you have to construct it from scratch each time.
Then there is mamihlapinatapai, from Yagán, the language of Tierra del Fuego whose last native speaker, Cristina Calderón, died in 2022. The missionary Thomas Bridges, who compiled the first Yagán–English dictionary, is credited with glossing it as two people looking at each other, each hoping the other will offer to do something that both want done but neither is willing to do. The word compresses into a single unit a whole social standoff of shared wanting and shared hesitation. English speakers have this experience. But English needs a whole sentence to describe what Yagán packs into a single word. The word is contested too: it never made it into Bridges’ own dictionary, and linguists who have analyzed its parts read it more prosaically, as something closer to two people getting awkward with each other. How Yagán speakers actually used it is hard to recover.
From Russian, toska: Vladimir Nabokov, translating for English readers, described it at its deepest as a great spiritual anguish, often without any specific cause; a level down, it is a dull ache of the soul, a longing with nothing to long for. Its shades run on through restless yearning and nostalgia to plain boredom, and Nabokov needed a whole string of English phrases, from “sick pining” to “vague restlessness”, just to gesture at it.
And from Japanese, mono no aware: the gentle, bittersweet awareness of impermanence. The phrase is both a feeling and an aesthetic stance, so central to Japanese aesthetic tradition that poetry, fiction and film have been built around it, from The Tale of Genji to the films of Yasujirō Ozu. Whether having such a concept trains attention itself, orienting speakers habitually toward endings and transience, has never been tested directly. It is exactly the kind of question the experiments described below try to answer in more measurable domains.
The four words show how languages carve up experience: each draws a precise boundary around a particular configuration of feeling. An English speaker can understand saudade once it’s explained. The open question is whether, without the word, they identify the feeling as a distinct thing worth having: whether they notice it at the edge of consciousness and cultivate it rather than letting it dissolve into undifferentiated background melancholy.
The observation that languages carve experience differently is ancient. What is modern, and still contested, is the claim that the carving shapes cognition as well as communication.
Whorf, Hopi, and the Fall of a Hypothesis
Benjamin Lee Whorf was, professionally, a fire-prevention engineer for the Hartford Fire Insurance Company, trained as a chemical engineer at MIT. Linguistics was his second career: he studied under the linguist Edward Sapir at Yale, though never for a degree, and kept his insurance job until his death in 1941. In essays from his last decade, Whorf argued that the grammar of a language is “itself the shaper of ideas” and that speakers of different languages are led to different pictures of the world. Later writers hardened this into the “strong” Sapir–Whorf hypothesis: that your native language determines the structure of your thought, constituting it rather than merely reflecting it.
His most famous example was the Hopi people of the American Southwest. Whorf argued that the Hopi language had no words referring directly to time and no tenses like those of European languages, sorting events instead by whether they were reported or expected, and that as a result Hopi speakers did not treat time as a line running from past to future. If language constitutes thought, then speakers of very different languages live in different cognitive worlds, perceiving different realities.
The problem was that Whorf had worked largely with a single Hopi speaker living in New York, Ernest Naquayouma, and had little direct field contact. In 1983, linguist Ekkehart Malotki published Hopi Time, a study of more than 600 pages based on years of fieldwork. It documented hundreds of Hopi words and constructions for talking about time, from units of the day and year to a grammatical distinction between future and non-future. Whorf had been badly wrong on the specifics.
By then the hypothesis was already in retreat. “Strong” Sapir–Whorf had become a cautionary example of what happens when romantic ideas about linguistic difference outrun evidence. From the 1960s, Noam Chomsky’s universal grammar, which holds that all human languages share deep innate structures, pushed linguistics toward universalism, and Brent Berlin and Paul Kay’s 1969 survey of color terms seemed to show that even color vocabulary follows universal patterns. By the 1970s, linguistic relativity had become professionally unfashionable, and research on how language might influence thought largely stalled until a revival in the early 1990s. Malotki’s book landed as the final blow.
The evidence that killed the hypothesis showed the claimed effects to be overstated and poorly documented; it did not show that language has no effect on thought. The question itself survived, and while the strong version of Sapir–Whorf lay discredited, a more careful group of researchers was beginning to test it in the laboratory.
The Amazon Case
In 1977, an American linguist named Daniel Everett arrived among the Pirahã, a small Indigenous people who live by fishing, gathering and small garden plots along the Maici River in the Brazilian Amazon. He had come as a Christian missionary and stayed as a linguist. What he documented over decades of fieldwork offers an unusually stark natural experiment in the relationship between language and cognition.
Everett reports that Pirahã has no words for exact numbers, only terms for a small amount, a somewhat larger amount and “many”, and later experiments found that even the word nearest to “one” does not reliably mean one. He also reports that it has no basic color terms, only descriptive phrases, and only a minimal tense system with no relative tenses such as the perfect. It reportedly lacks recursion, the ability to embed one phrase or sentence inside another, which Chomsky and his colleagues had proposed as the core of the human language faculty. And, Everett argues, its speakers restrict talk almost entirely to events within their immediate experience.
The cognitive consequences were documented independently by the psychologist Peter Gordon, who ran a series of controlled experiments with Pirahã participants. Pirahã speakers struggle with tasks requiring exact quantity matching for numbers above three. Performance deteriorated steadily as quantities grew, in a pattern consistent with rough analog estimation rather than counting and miscounting: the way we all estimate when we’re not counting. Gordon concluded that numerical cognition is clearly affected by the absence of a counting system in the language.
We don’t generally think of counting as a linguistic skill. It feels like pure perception: surely you can see that there are five objects rather than four? But the Pirahã evidence suggests that exact quantity, the idea that the difference between seven and eight is always exactly one, needs linguistic scaffolding to be tracked reliably. When a team led from MIT, including Everett, re-ran Gordon’s tasks in 2008, the Pirahã matched large sets almost perfectly when they could line the objects up one-for-one; they faltered when an exact number had to be carried across space or held in memory. Number words, the researchers concluded, do not change how people perceive quantity; they work as a cognitive technology, a tool for carrying exact quantities across time and space.
Everett argues that if Pirahã truly lacks recursion, it refutes the influential proposal by Chomsky and his colleagues that recursion is the core of the human language faculty — and with it a central pillar of universal grammar. The linguists Andrew Nevins, David Pesetsky and Cilene Rodrigues reply that a language that happens not to use recursion says nothing about the human capacity for it. Either way, the number findings suggest that language can shape whole cognitive domains rather than merely carrying pre-formed thought.
The evidence splits by claim. The number data are fairly robust: the follow-up confirmed both the missing number words and the pattern of failure, and speakers of Mundurukú, another Amazonian language with number words only up to about five, show the same split, handling large approximate quantities well beyond their vocabulary while failing at exact arithmetic above four or five. The recursion claim remains sharply disputed. Nevins, Pesetsky and Rodrigues argued that Everett had misanalyzed the grammar and that Pirahã does have embedded clauses; a later analysis of Pirahã texts co-authored by Everett found no unambiguous evidence of embedding, without settling the question. Pirahã is probably an extraordinary case of language and culture co-evolving in ways that demonstrably affect cognition, even if some of the most dramatic claims remain contested.
Even if Everett overstated some conclusions, Pirahã raises the question of cognitive minimums: what a language makes possible, and what a language without certain words forecloses. The number results put exact counting among the things language makes possible.
What the Experiments Show
Pirahã is extreme. But the evidence for language’s influence on thought doesn’t require extreme cases. It shows up in the ordinary cognition of speakers of Russian, Mandarin, and English, in carefully controlled laboratory settings designed to isolate specifically linguistic effects.
The color of language. Russian makes an obligatory distinction between lighter blues (goluboy) and darker blues (siniy), two separate basic color words where English uses one. A team at MIT and Stanford that included the cognitive scientist Lera Boroditsky ran a simple experiment. Russian and English speakers were shown three color swatches and asked to identify, as fast as possible, which of two bottom swatches matched a target at the top. Russian speakers were faster when the non-matching swatch fell on the other side of the goluboy/siniy boundary from the target than when both were the same kind of blue. The linguistic categories provided a cognitive shortcut. English speakers, tested on identical colors, showed no such advantage.
The critical control was verbal interference. When participants silently rehearsed an unrelated eight-digit number during the task, occupying the verbal processing channel, the Russian speakers’ advantage disappeared entirely. A spatial interference task, which occupied spatial memory rather than language, left the advantage intact. Only interference with language erased the advantage, which marks the effect as specifically linguistic. A later re-run of the same tasks, however, found no speed advantage at the light/dark blue boundary.
The shape of time. Languages differ in how they talk about time. English speakers mostly use horizontal spatial metaphors — the meeting was moved forward, the deadline is coming up. Mandarin also uses vertical ones: last month is shàng ge yuè, literally “the up month”, and next month is “the down month”. Boroditsky showed people objects arranged vertically or horizontally just before they judged statements such as “March comes earlier than April”. Mandarin speakers were faster after vertical primes, English speakers after horizontal ones, and English speakers briefly taught to talk about time vertically began to show the Mandarin pattern. The result has not held up: two independent teams, in ten attempts between them, failed to replicate the original result, and one found that Mandarin speakers use horizontal time metaphors more often than vertical ones. Later experiments with a different task, in which people arranged pictures of events in space, did find Mandarin speakers laying time out vertically more often than English speakers, so the two groups may still picture time differently even though the priming result failed. The time effect is far less secure than the color effect.
The orientation of a life. Perhaps the most vivid evidence involves space. Speakers of Guugu Yimithirr and Kuuk Thaayorre, Aboriginal Australian languages, navigate using absolute cardinal directions rather than the egocentric left/right that English speakers default to. In these languages, directions come in north, south, east and west rather than left and right; there is no “turn left at the corner.” Stephen Levinson and the anthropologist John Haviland documented the cognitive consequence: speakers of these languages stay oriented even in unfamiliar landscapes and inside unfamiliar buildings, and in Levinson’s experiments Guugu Yimithirr speakers remembered arrangements of objects in compass terms, where Dutch speakers remembered them relative to their own bodies. Boroditsky likes to contrast a five-year-old Kuuk Thaayorre girl in Pormpuraaw, who pointed north instantly and correctly, with lecture halls of distinguished scholars at Stanford, many of whom could not. Nothing in English requires its speakers to keep track.
The texture of feeling. The most intimate domain is emotion itself. Lisa Feldman Barrett, a psychologist and neuroscientist at Northeastern University, has spent decades studying the relationship between emotion concepts, language, and emotional experience. Her core argument, developed across studies, is that emotional granularity, the ability to tell similar emotional states apart, depends on the emotion concepts a person has learned, and words are one of the main ways those concepts are acquired. On her account, people with finely grained emotional concepts both describe and construct their inner states with more precision. Barrett reports that people who could distinguish finely among their unpleasant feelings were 30 percent more flexible in regulating their emotions; in a set of diary studies, those who differentiated their emotions well were less prone to aggression when angry.
Saudade may work the same way. A culture with a precise word for this bittersweet longing may produce people who notice and cultivate the feeling more readily, since the word could make a feeling anyone can have salient and shareable.
What the Science Does Not Show
Some effects of language on cognition have held up across studies, with the firmest footing in spatial memory and exact number. Others have not: the Russian blues advantage did not reappear when the tasks were re-run, and the Mandarin time-priming result has not been reliably reproduced. Popular summaries tend to blur the two.
The strong version of Sapir–Whorf, which claims that language determines thought, traps you inside its categories, and prevents you from perceiving what your language cannot name, is not supported. The evidence runs the other way. Translation is possible, bilinguals switch between cognitive frames as they switch between languages, and people can learn new conceptual frameworks as adults, sometimes with surprising speed. Brief training can shift behavior too, though the clearest example, English speakers taught vertical time metaphors, comes from the same Mandarin study whose main result later failed to replicate.
Many cognitive universals are robust. Basic color categories appear cross-culturally in predictable patterns that suggest perceptual constraints independent of language. Small-number cognition and core spatial reasoning vary far less with language than the strong hypothesis would predict. (Even the supposedly universal reading of facial expressions is now contested: Barrett’s own team found that Himba participants in Namibia did not sort faces into the “universal” emotion categories without cues to Western emotion concepts.)
The effect sizes in the experimental literature are real but modest. What language provides, at the margin of cognition, is something more like a default path than an absolute constraint: habitual ways of perceiving and categorizing.
The Lens You’ve Never Looked At
The concept in the headline, one that shapes your entire life and for which everyday speech has no word (it is the effect linguists study under the name linguistic relativity), is linguistic framing itself: the set of cognitive habits installed by your native language before you were old enough to examine them. The words and grammar you learned as a toddler set habitual defaults for what you attend to, from where north lies to how exactly you count and which feelings you tell apart, with the evidence strongest for space and number.
The Kuuk Thaayorre five-year-old who points north without hesitation was shaped by a language whose speakers reach for north, south, east and west even to say where a cup sits on a table. Her skill is practiced perception, and her language demands the practice almost every time she says where something is, long before she is old enough to choose it.
You can learn the word saudade and, having learned it, begin to notice the state it names. The language you grew up with set its defaults long before you could ask.
Gen AI Disclaimer
Some contents of this page were generated and/or edited with the help of a Generative AI.
Media
Key Sources and References
Boroditsky, L. How language shapes thought. Scientific American, 304(2), 62–65, 2011. https://www.scientificamerican.com/article/how-language-shapes-thought/
Winawer, J., Witthoft, N., Frank, M.C., Wu, L., Wade, A.R. & Boroditsky, L. Russian blues reveal effects of language on color discrimination. PNAS, 104(19), 7780–7785, 2007. https://pmc.ncbi.nlm.nih.gov/articles/PMC1876524/ · https://doi.org/10.1073/pnas.0701644104
Martinovic, J., Paramei, G.V. & MacInnes, W.J. Russian blues reveal the limits of language influencing colour discrimination. Cognition, 201, 104281, 2020. https://doi.org/10.1016/j.cognition.2020.104281 · https://hira.hope.ac.uk/id/eprint/3058/
Boroditsky, L. Does language shape thought? Mandarin and English speakers’ conceptions of time. Cognitive Psychology, 43(1), 1–22, 2001. https://doi.org/10.1006/cogp.2001.0748
Chen, J.-Y. Do Chinese and English speakers think about time differently? Failure of replicating Boroditsky (2001). Cognition, 104(2), 427–436, 2007. https://doi.org/10.1016/j.cognition.2006.09.012
January, D. & Kako, E. Re-evaluating evidence for linguistic relativity: Reply to Boroditsky (2001). Cognition, 104(2), 417–426, 2007. https://doi.org/10.1016/j.cognition.2006.07.008
Boroditsky, L., Fuhrman, O. & McCormick, K. Do English and Mandarin speakers think about time differently? Cognition, 118(1), 123–129, 2011. https://doi.org/10.1016/j.cognition.2010.09.010
Fuhrman, O., McCormick, K., Chen, E., Jiang, H., Shu, D., Mao, S. & Boroditsky, L. How linguistic and cultural forces shape conceptions of time: English and Mandarin time in 3D. Cognitive Science, 35(7), 1305–1328, 2011. https://doi.org/10.1111/j.1551-6709.2011.01193.x
Levinson, S.C. Language and cognition: The cognitive consequences of spatial description in Guugu Yimithirr. Journal of Linguistic Anthropology, 7(1), 98–131, 1997. https://www.mpi.nl/publications/item66562/language-and-cognition-cognitive-consequences-spatial-description-guugu
Gaby, A. The Thaayorre think of time like they talk of space. Frontiers in Psychology, 3, 300, 2012. https://doi.org/10.3389/fpsyg.2012.00300
Gordon, P. Numerical cognition without words: Evidence from Amazonia. Science, 306(5695), 496–499, 2004. https://doi.org/10.1126/science.1094492
Frank, M.C., Everett, D.L., Fedorenko, E. & Gibson, E. Number as a cognitive technology: Evidence from Pirahã language and cognition. Cognition, 108(3), 819–824, 2008. https://doi.org/10.1016/j.cognition.2008.04.007 · https://web.stanford.edu/~mcfrank/papers/old/FEFG-cognition.pdf
Pica, P., Lemer, C., Izard, V. & Dehaene, S. Exact and approximate arithmetic in an Amazonian indigene group. Science, 306(5695), 499–503, 2004. https://doi.org/10.1126/science.1102085
Everett, D.L. Cultural constraints on grammar and cognition in Pirahã: Another look at the design features of human language. Current Anthropology, 46(4), 621–646, 2005 (author’s copy). https://daneverettbooks.com/wp-content/uploads/2023/11/culturalgrammarWEB.pdf
Nevins, A., Pesetsky, D. & Rodrigues, C. Pirahã exceptionality: A reassessment. Language, 85(2), 355–404, 2009. https://dspace.mit.edu/handle/1721.1/94631 · https://www.cambridge.org/core/journals/language/article/abs/piraha-exceptionality-a-reassessment/BC4ED2BEB83C3F012F0ADEA31ABFE0E9
Everett, D.L. Pirahã culture and grammar: A response to some criticisms. Language, 85(2), 405–442, 2009 (author’s copy). https://daneverettbooks.com/wp-content/uploads/2023/11/everett_09_Pirah-Culture-.pdf
Futrell, R., Stearns, L., Everett, D.L., Piantadosi, S.T. & Gibson, E. A corpus investigation of syntactic embedding in Pirahã. PLoS ONE, 11(3), e0145289, 2016. https://doi.org/10.1371/journal.pone.0145289
Hauser, M.D., Chomsky, N. & Fitch, W.T. The faculty of language: What is it, who has it, and how did it evolve? Science, 298(5598), 1569–1579, 2002. https://doi.org/10.1126/science.298.5598.1569
Everett, D. Don’t Sleep, There Are Snakes: Life and Language in the Amazonian Jungle. Pantheon Books, 2008 (review: Science News). https://www.sciencenews.org/article/book-review-dont-sleep-there-are-snakes-life-and-language-amazonian-jungle-daniel-l-everett
Instituto Socioambiental. Pirahã. Povos Indígenas no Brasil. https://pib.socioambiental.org/en/Povo:Pirahã
Barrett, L.F. How Emotions Are Made: The Secret Life of the Brain. Houghton Mifflin Harcourt, 2017 (author’s notes on emotional granularity; publisher listing). https://how-emotions-are-made.com/notes/Negative-1 · https://www.publishersweekly.com/978-0-544-13331-0
Pond, R.S., Kashdan, T.B., DeWall, C.N., Savostyanova, A., Lambert, N.M. & Fincham, F.D. Emotion differentiation moderates aggressive tendencies in angry people: A daily diary analysis. Emotion, 12(2), 326–337, 2012. https://doi.org/10.1037/a0025762
Gendron, M., Roberson, D., van der Vyver, J.M. & Barrett, L.F. Perceptions of emotion from facial expressions are not culturally universal: Evidence from a remote culture. Emotion, 14(2), 251–262, 2014. https://doi.org/10.1037/a0036052
Whorf, B.L. Language, Thought, and Reality: Selected Writings of Benjamin Lee Whorf (J.B. Carroll, ed.). Technology Press of MIT, 1956. https://archive.org/details/languagethoughtr00whor
EBSCO Research Starters. Benjamin Lee Whorf. https://www.ebsco.com/research-starters/biography/benjamin-lee-whorf
Malotki, E. Hopi Time: A Linguistic Analysis of the Temporal Concepts in the Hopi Language. Mouton, 1983 (reviews, including Bernard Comrie’s). https://ac.nau.edu/~malotki/html/hopi_time_reviews.html
Endangered Language Alliance. Hopi (Ernest Naquayouma). Language Map NYC. https://languagemap.nyc/Explore/Language/Hopi
Regier, T. & Kay, P. Language, thought, and color: Whorf was half right. Trends in Cognitive Sciences, 13(10), 439–446, 2009. https://doi.org/10.1016/j.tics.2009.07.001
Kay, P. & Regier, T. Resolving the question of color naming universals. PNAS, 100(15), 9085–9089, 2003. https://doi.org/10.1073/pnas.1532837100
Wolff, P. & Holmes, K.J. Linguistic relativity. WIREs Cognitive Science, 2(3), 253–265, 2011. https://doi.org/10.1002/wcs.104
NPR. Cristina Calderón, Chile’s last known Yaghan speaker, dies at 93. February 18, 2022. https://www.npr.org/2022/02/18/1081584395/cristina-calderon-chiles-last-known-yaghan-speaker-dies-at-93
Language Hat. Mamihlapinatapai (with Yoram Meroz’s analysis); untranslatable.org. Mamihlapinatapai. https://languagehat.com/mamihlapinatapai/ · https://untranslatable.org/words/mamihlapinatapai
Nabokov, V. Commentary on Eugene Onegin (1964), entry on toska, as reproduced in the NABOKV-L archive, The Nabokovian. https://thenabokovian.org/sites/default/files/2018-01/NABOKV-L-0021083___body.html
Parkes, G. & Loughnane, A. Japanese Aesthetics (section on mono no aware). Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/japanese-aesthetics/
UNESCO Intangible Cultural Heritage. Fado, urban popular song of Portugal, 2011; Dicionário Priberam da Língua Portuguesa. Saudade. https://ich.unesco.org/en/RL/fado-urban-popular-song-of-portugal-00563 · https://dicionario.priberam.org/saudade
