How Much Does Our Language Shape Our Thinking?

In 2010, a new goddess, about two feet tall and cast in bronze, was set to appear in a village within the Indian state of Uttar Pradesh. She looked nothing like the deities of Hindu mythology. In lieu of Durga’s bright saris or Lakshmi’s opulent jewels, she wore a wide-brimmed hat and the robes of the Statue of Liberty. She wasn’t riding a lion or a swan; she stood on a desktop computer. Instead of a sword or a spear, she held a pen in one hand and the Indian constitution—with its promise of legalized equality—in the other. Her name was Angrezi Devi, the Goddess English, and she was intended for India’s Dalits, or “untouchables.”

“The Goddess English can empower Dalits, giving them a chance to break free from centuries of oppression,” her creator, the prominent Dalit writer Chandra Bhan Prasad, declared. He saw English as an immensely valuable resource for the Dalit. “Will English-speaking Dalits be expected to clean gutters and roads?” he asked. “Will English-speaking Dalits be content to work as menials at landlords’ farms?” An atheist, he designed the goddess in order to infuse English into the Dalit identity, propelling his people from a feudal subaltern standing to the ranks of the modern and independent. “Learning English has become the greatest mass movement the world has ever seen,” he wrote.

He had a point. An estimated 1.5 billion people—roughly one in every five human beings—speak English, making it the most widely used language in the history of humanity. With an official status in the U.N., NATO, the W.T.O., and the E.U., it reigns as the dominant “lingua franca of the world,” Rosemary Salomone writes in “The Rise of English: Global Politics and the Power of Language” (Oxford). Like other colonial tongues, it spread first through “conquest, conversion, and commerce,” she notes, but its spread today is powered by a fourth process, what Salomone calls “collusion.” Around the globe, people pursue English and the opportunities it promises. “Korean mothers move their children to anglophone countries to learn in English,” Salomone observes. “Dutch universities teach in it. ASEAN countries collaborate in it. Political activists tweet in it.”

The expansion of English naturally evokes angst and opposition. Salomone, a law professor at St. John’s University, focusses on the political and legal tensions that accompany the diffusion of English. France, for example, fought for decades against English’s dominance in the European Economic Community, and then in the E.U. “If, with the arrival of the English, French no longer were the first working language in the Community,” the French President Georges Pompidou warned, in 1971, “then Europe would never be totally European.” Nearly half a century later, in 2018, President Emmanuel Macron declared English to be “too dominant in Brussels” and vowed to ramp up efforts to “teach French to European officials.”

In part, such apprehension reflects unease about the erosion of various cultural identities. Yet many researchers find another reason to worry about the spread of English: the prospect of cognitive hegemony. Languages, they argue, influence how we perceive and respond to the world. The idiosyncrasies of English—its grammar, its concepts, its connection to Western culture—can jointly produce an arbitrary construction of reality.

Speculation about these effects is widespread. Prasad, for instance, thinks there’s a kind of egalitarianism that’s inherent in English and missing from its Indian alternatives. “Hindi is full of caste biases,” he told me. “Idioms, phrases, sayings, jokes, songs belittle Dalits. How can any Dalit take pride in the so-called native tongue?” Other intellectuals, such as the Kenyan writer Ngũgĩ wa Thiong’o, maintain that English serves as yet another tool to entrench British and American culture. Pierre Bourdieu, the celebrated French sociologist, voiced a common concern when, in 2001, he wondered if “it is possible to accept the use of English without the risk of one’s mental structures being anglicized, without being brainwashed by linguistic patterns.” Was he right to worry?

Everyone can agree that language affects thought. If I told you that I have a pet badger and twenty-two canaries, you’d have new thoughts about my home life. The real question is whether a language itself has features that affect how its speakers think: Does conversing in Spanish for a month make objects seem more gendered? Does speaking English rather than Hindi make you less casteist, and maybe more capitalist?

Today, questions like these tend to be associated with Benjamin Lee Whorf, a fire-insurance analyst who studied linguistics at Yale in the nineteen-thirties. History has been both kind and unkind to him. On the one hand, his name has become synonymous with a theory about how language affects thought, though it predated him by at least a century. On the other hand, the version of the theory often attributed to him is so radical that few modern scholars would want the honor, anyway.

Whorf laid out his views in an essay titled “The Relation of Habitual Thought and Behavior to Language.” Contrasting the way time is discussed by English speakers (as an object that can be quantified and divided) and by Hopi speakers (as a more continuous process, or so Whorf believed), he suggested that linguistic differences contributed to differences in how each group understands temporal flow. Despite the boldness of his claims, he was also cautious, proposing merely “traceable affinities” between language and behavior, nothing ironclad, and stressing that he was “the last to pretend that there is anything so definite as ‘a correlation.’ ”

Unfortunately, that nuance has usually been forgotten. Whorf has since become the mascot of linguistic determinism—the position that language is the ultimate arbiter of thought. Whorfianism, as it’s sometimes called, quickly dissolves into absurdities: if your language lacks a proper future tense, tomorrow will be inconceivable; if your language lacks certain emotion words, you will never feel them. Preverbal infants, orangutans, and all other creatures incapable of language are, by implication, powerless to perform many basic mental operations.

Whorfianism has been the target of relentless discrediting. Some of the most striking counterexamples involve individuals unable to produce or comprehend language. Take the case of Brother John, a fifty-year-old French Canadian who suffered from spells of aphasia. Even during periods when he had lost the faculty of language, he mostly got along fine, according to a 1980 study published in Brain and Language. He could manipulate complex tools, follow instructions he’d been given beforehand, and sometimes succeed in hiding his impairment from others. The Harvard cognitive scientist Steven Pinker has had much to say about Whorfian fallacies. He has shown how common experiences—like searching for the right word or inventing a new term for an existing intuition—invalidate the idea that language always precedes thought. Writing in “The Language Instinct” (1994), he concludes that Whorfianism is “wrong, all wrong.”

That’s a fair assessment if we’re talking about the strongest interpretation of Whorf’s arguments. Yet the picture emerging from the latest research is more complicated. Whorfianism is wrong—but it isn’t all wrong.

“Each of my languages comes not only with its own patterns of sound and methods for arranging words but also with its social habits and its judgments about what to forgive, what to condemn, and what to revere,” Julie Sedivy writes in “Linguaphile: A Life of Language Love” (Farrar, Straus & Giroux). Born in what was then Czechoslovakia, Sedivy grew up in a “linguistic bedlam,” hopping between Austria and Italy before settling down in Montreal. She was acquainted with five languages by kindergarten and went on to study how people learn and process language. Science suffuses her book, yet, as a way of knowing, it coexists with experience; the resulting volume isn’t so much a standard pop-sci book as it is a rhapsodic meditation on loving, taming, and forgetting words. She senses that distinct cognitive styles are tied to the different languages she speaks, comparing them to personalities bickering for the spotlight. “I am a cacophony of voices, influencing each other, at times assisting each other, at times getting in each other’s way, always vying for turf,” she writes.

Testimony from polyglots like her has invited a more sophisticated take on Whorf’s ideas. What if language is less like a yoke than like a wind, nudging us in various directions? This moderate approach, which is more in line with Whorf’s original perspective, is known as “weak Whorfianism” or, paradoxically, neo-Whorfianism.

Some neo-Whorfian studies have already become classics. One led by the psychologist Jonathan Winawer and published in 2007 took advantage of the fact that, where English has the word “blue,” Russian has two basic color terms: goluboy (lighter blue) and siniy (darker blue). Russian speakers in the experiment proved faster than English speakers at distinguishing shades that corresponded to that lexical distinction.

Other such studies exploit variations in, for instance, how languages talk about temporal duration (long and short, as in English, or big and small, as in Greek) or the ordering of events (B following A, as in English, or B below A, as in Mandarin), testing whether those differences correlate with performance on experimental tasks. Many linguists remain unimpressed. In the anti-Whorfian polemic “The Language Hoax” (2014), John H. McWhorter, of Columbia University, describes this research as mostly showing “eensy-weensy differences” of the sort that “one might find in the cosseted context of a psychological experiment.”

Caleb Everett, an anthropologist and psychologist at the University of Miami, arrives at another conclusion in “A Myriad of Tongues” (Harvard). Like Sedivy, Everett was reared in a morass of languages. The son of the missionary turned linguist Daniel Everett, he spent much of his childhood in Brazil. His parents intermittently took him to the Amazon to stay with the Pirahã, a people legendary for speaking a language devoid of words for numbers and colors, which contributed to his lifelong fascination with linguistic and cognitive diversity.

Everett’s book is about the surprising ways that languages differ and about the significance these differences may have. He starts by covering Whorf’s favorite topic: time. English speakers instinctively split time into categories of past, present, and future, but many others don’t. Karitiâna, an Amazonian language Everett studied two decades ago, has two tenses, future and non-future, while another Amazonian language, Yagua, seems to have eight, including for events that occurred between a month and a year ago, for events that are about to happen, and for events expected to happen further into the future.

More relevant for Whorfianism are the metaphors people use to organize time. For English speakers, time is understood spatially, with the past typically “behind us” and the future “ahead.” Aymara, an Andean language spoken by millions of Indigenous Bolivians and Peruvians, likewise uses space to talk about time but favors a metaphor about sight. In Aymara, nayra, or last year, translates literally to something like “the year I can see.” The past, visible, thus stands in front of the speaker, while the future, unseeable, looms behind. Ancha nayra pachana, or a long time ago, can roughly be translated as “a time way in front of me.” When researchers analyzed videos of people chatting, they noticed that the metaphors inform gesture, with fluent Aymara speakers pointing backward to talk about the future and forward to talk about the past. Spanish speakers from the same region show the opposite patterns, suggesting that language configures how speakers map time onto space.

Aymara speakers are far from unique here. Speakers of Lisu, a Tibeto-Burman language, also talk about the future as lying behind them and the past as in front of them. Everett tells us that Yupno, a language spoken in the eastern New Guinea highlands, invokes a three-dimensional analogy. Like a gravity-defying river, the future is said to flow up mountains, while the past flows downhill. As with Aymara speakers, the metaphor manifests in gesture: how Yupno speakers point depends on the orientation of the nearest mountain range. Some cognitive scientists have assumed that all humans, whatever their local quirks, reason about time using spatial metaphors, yet at least one language, Tupi-Kawahíb, evidently lacks any mapping between time and space—not left to right, back to front, or downhill to uphill. When Tupi-Kawahíb speakers were asked to organize objects to chart out the seasons of a year, researchers struggled to understand the arrangements the speakers had created. More than communication tools, languages help concretize the abstract, providing frameworks for making sense of concepts as fundamental as time.

Of the many topics Everett covers—which include space, number, and object categorization—the most fascinating is probably sensory vocabulary. Western writers have long assumed that human beings have an inherently limited capacity to describe some senses, with olfaction ranking as the most elusive. We can speak abstractly about colors (red, blue, black) and sound (high, low, loud). With smell, though, we usually give “source-based” references (“like cut grass”). But the cognitive scientist Asifa Majid, now of Oxford, and the linguist Niclas Burenhult, of Lund University, in Sweden, have shown that this needn’t be the case. They discovered that the Jahai, hunter-gatherers living at the border of Malaysia and Thailand, have a rich vocabulary of abstract smell words. One Jahai term, itpit, refers to the “intense smell of durian, perfume, soap, Aquillaria wood, and bearcat,” Majid and Burenhult report. Another, cnes, applies to “the smell of petrol, smoke, bat droppings and bat caves, some species of millipede, root of wild ginger, leaf of gingerwort, wood of mango tree.” Subsequent research has found large olfactory lexicons in at least forty other languages, among them Fang, Khmer, Swahili, and Zapotec.

It makes a difference. In a study that Majid and Burenhult conducted a decade ago, Jahai and English speakers were asked to identify and name twelve smells, including cinnamon, turpentine, gasoline, and onion. English speakers, despite their greater familiarity with the odors, faltered. They mostly gave rambling source-based answers and showed almost no agreement among themselves. One English speaker presented with cinnamon responded, “I don’t know how to say that, sweet, yeah; I have tasted that gum like Big Red or something tastes like, what do I want to say? I can’t get the word. Jesus it’s that gum smell like something like Big Red. Can I say that? Ok. Big Red. Big Red gum.” But Jahai speakers named smells with relative ease. They used abstract terms and were much more likely to converge in their responses. In a follow-up study, wine and coffee experts performed just as badly as novices when given non-wine and non-coffee smells, suggesting the Jahai’s enhanced abilities aren’t simply a result of practice in attending to aromas. Rather, the regular exercise of sorting the olfactory world with abstract labels seems to change how the Jahai understand all smells, familiar and otherwise.

The work on olfaction is a tiny part of a large research program, much of it headed by Majid, that has overturned the scientific consensus on how humans talk about the senses. At least since Aristotle, many writers have posited a sensorial hierarchy: seeing and hearing are said to be the most salient to our minds and the easiest to verbalize, followed by taste, touch, and finally smell. Contesting that thesis, Majid and her colleagues have developed a measure called codability, which captures how easily a sense is expressed. Codability is high when members of a language community converge on one or two abstract labels to describe a stimulus; ask English speakers to tell you the color of a stop sign, and you’d expect high codability. It’s low, in contrast, when people provide diverse, protracted, and ad-hoc descriptions—as when, say, you ask English speakers to describe the smell of a rutabaga.

Majid and her team measured codability for the five senses in twenty far-flung languages, including three unrelated sign languages. English, the only spoken Western European tongue in the sample, was also the only one to exhibit high codability for sight and hearing and low codability for everything else. “Rampant variation” reigned, the researchers found. English speakers floundered when talking about touch (in response to sandpaper, felt, rubber, etc.), but speakers of certain other languages—such as Dogul Dom, in Mali, and Siwu, in Ghana—tended to agree in their descriptions. In many languages, including Lao, Farsi, Yucatec, and Cantonese, taste turned out to be the most expressible sense.

As researchers look beyond English, close relatives (like Spanish and German), and other so-called behemoth languages (like Mandarin and Arabic), they encounter differences long thought impossible. Twenty years ago, abstract smell vocabularies seemed ridiculous. Burenhult studied the Jahai language for a decade, even writing a doctoral dissertation on its grammar, before Majid asked him to run a battery of tasks that revealed Jahai speakers’ exceptional way of talking about smell. Other linguistic features once assumed to be universal—such as tenses, personal pronouns, and even, potentially, a distinction between nouns and verbs—have turned up missing when greater numbers of languages have been scrutinized. Likewise, we’ve enlarged our sense of the metaphors used to map concepts. English describes acoustic pitch using a verticality metaphor (high-low), but a study by experts in musical cognition found that people around the world use at least thirty-five other mappings, such as small-big, alert-sleepy, pretty-ugly, tense-relaxed, summer-winter, and—in the case of some traditional Zimbabwean instrumentalists—“crocodile” (low pitch) and “those who follow crocodiles” (high pitch).

Everett’s book revels in such discoveries, which multiply the conceivable differences separating languages. In a recent review of the research literature, the language scientist Damián E. Blasi, along with Majid and others, listed the many cognitive domains that English seems to affect, including memory, theory of mind, spatial reasoning, event processing, aesthetic preferences, and sensitivity to rhythm and melody. Languages help shape the worlds we inhabit less through a few grammatical rules than through countless subtle distinctions. John McWhorter might have been right that the effect of any single linguistic feature is minor. But, as Isaac Newton realized when developing calculus, innumerable tiny effects create large-scale patterns.

Prasad’s Goddess English was meant to début on October 25, 2010. She was supposed to inhabit a temple of black granite, its walls engraved with scientific formulas and the names of great English writers. But the big day never came. The district government blocked the temple’s construction. Prasad recalls a local official hinting that the decision came from Kumari Mayawati, then the chief minister of Uttar Pradesh. Mayawati, who is herself Dalit and who had statues made in her honor while in office, enjoys an almost mythical status in the Dalit community. An Indian journalist reported rumors that Mayawati had blocked the temple because she wanted only “one Dalit goddess in the state.” After she left office, in 2012, Prasad hoped construction on the temple might resume. It didn’t. Five years later, Uttar Pradesh elected Yogi Adityanath, a monk and a Hindu nationalist from the Bharatiya Janata Party, as its new chief minister, and the prospect of a shrine to English may have dimmed further.

The Bharatiya Janata Party, or B.J.P., has ruled India since Narendra Modi took power, in 2014. The third of a tea seller’s six children, Modi is a departure from the pedigreed, English-speaking bureaucrats who governed India for most of its post-independence history. He rejects the secular, hyper-diverse vision of the country set forth by Gandhi and Nehru, and has been intent on reshaping India into an explicitly Hindu nation. Language has been central to that project. English is depicted as a symbol of the country’s imperialist past and disconnected élites, while Hindi, the Sanskrit-derived “language of the masses,” has been recast as an avatar of common identity—despite the fact that only about forty per cent of Indians speak Hindi or one of its many dialects.

Although Prasad disagrees with the ideologues of the B.J.P. about the proper place for English in India, they all agree that English is powerful. Putting aside the economic opportunities it unlocks, they see the language as a psychosocial force—one that is rooted in its foreign origins and that, rather like Hinduism’s divine triumvirate, can create, sustain, or destroy social orders. Bourdieu’s suggestion that the spread of English represents a form of cognitive hegemony is one that Prasad and the B.J.P. accept, whether as a blessing or as a curse.

Yet we can attend to the recent research into cultural psycholinguistics without ascribing so much authority to English. The language is far less unitary than we often suppose. The version of English that the B.J.P. seeks to weaken differs considerably from whatever the British first imposed some two centuries ago. It’s a language enamored with the progressive tense (as in “I am going to school every day”), whose pronunciation is transfigured by retroflex consonants (say “dog” but curl your tongue back until it resembles a fishhook) and unaspirated voiceless plosives (say “kick” without expelling air), and in which “Kindly do the needful and prepone our meeting to tomorrow itself, na?” is a typical request. Honorifics like “ji,” “sir,” and “sahib” are mobilized to accommodate local nuances of deference. After centuries spent fermenting in the subcontinent, English has turned Indian.

Such local reworking is ubiquitous. In Singapore, English bred with Malay, Tamil, Hokkien, and other Asian languages to produce that lah-filled tongue known as Singlish, a creole that nonlocal English speakers struggle to follow. Seven years ago, the BBC began offering news online in West African Pidgin English, which is spoken by more than a hundred million people and influenced so strongly by local African tongues that some linguists consider it tonal. There are countless examples like these, in which English scaffolding has adapted to the demands and the cultural heritage of its speakers, from Jamaican patois to Tok Pisin, of New Guinea. For that matter, few tongues exemplify the ever-evolving nature of language better than so-called Standard English, which, after millennia of conquest, conversion, and commerce, has acquired a vocabulary that is roughly seventy per cent non-Germanic and a simplified grammar that facilitates its spread among adults.

Bourdieu was right that linguistic patterns affect us. Yet, going by the best ethnographic and social-science research, his fear of brainwashing was overblown. If ways of speaking can alter ways of thinking, ways of thinking can alter ways of speaking as well. The dynamic interaction between the two is part of the ongoing story of how we try to make the world intelligible to us—and to make ourselves intelligible to one another. Talk about the human conversation. ♦