Learn a Foreign Language Before It’s Too Late

AI translators may seem wondrous but they also erode a major part of what it is to be human.

An image of the word "hello: translated into Spanish, French, and Italian, which have been obscured with the word "hello" written in English.
Ben Kothe / The Atlantic

To me, AI’s scariest aspect is the so-called singularity—the threat of a runaway intelligence explosion leaving humanity in the dust. But today’s state of the art in artificial intelligence is already auguring smaller but still shattering scenarios.

Some people so deeply yearn to climb Mount Everest that they prepare for years, spend vast sums of money, exhaust themselves for weeks in the climb itself, and repeatedly put their lives at risk. Does that sound like you? Or would you rather just land on its summit in a helicopter and feast yourself on the great view? And what about scaling the metaphorical Everest of a foreign language? Two small episodes in my life in the past month led me to serious musings along these lines.

Two weeks ago, I watched, for the first time ever, a video of myself back in 2018 in Hangzhou, China, as I struggled mightily for three minutes to make a few off-the-cuff remarks to about 20 young people from Shanghai who belonged to a club of AI enthusiasts. They had traveled 200 miles to Hangzhou to meet me for dinner, and for two and a half hours we had spoken only English together, but toward the evening’s end they asked me if I wouldn’t mind saying something very brief in Chinese for those club members who hadn’t been able to make the trip. Uh-oh! Panic city! Even though I had devoted many arduous years to the study of Chinese (always thinking of the daunting phrase “Learning Chinese is a five-year lesson in humility” as a ridiculous understatement), and had worked like the devil during the previous three months in Hangzhou, I was caught way off guard by their request and, although in the end I obliged them, I felt super jittery while doing so. Shortly thereafter they sent me the video, but for all these years I hadn’t dared to look at even the opening few seconds of it, so scared was I of seeing myself linguistically stumble all over the place.

But what I saw, when I finally dared to watch myself very recently, was surprising. I saw a person who was not just struggling hard to express himself in a very difficult alien tongue, but who was actually doing a fairly decent job of it, while at the same time coming across as insecure and vulnerable, yet courageously willing to take the bull by the horns. In short, today’s me felt proud of my 2018 self! Since that day five years ago, sadly, my once-okay Chinese has gone to the dogs, and today I couldn’t give a three-minute talk in Chinese to save my life, so I’m thrilled to have proof that at one time in my life, I was actually able to wing it, actually able to give a tiny “talk” in Chinese, even if only a three-minute one.

So that’s the first episode; here’s the other. A few weeks ago, my very dear Italian friend Benedetto Scimemi passed away, and I spent hours writing heartfelt emails of condolence to all the members of his family. It happens that I lived in Italy for nearly three years and, on top of that, my two children and I have spoken Italian for 30 years as our family language, so my Italian is very fluent and comfortable—but, even so, it is not the Italian of a native speaker. In writing those difficult and emotional emails, I was constantly adjusting my words and phrases, lovingly remembering Benedetto and all the wonderful things we had done together, and pushing my Italian to its very limits. It took me perhaps two or three times as long as it would have taken me in English, but I did it with all my heart. I looked up lots of words in the big, heavy dictionary that I always keep right by my computer, and I felt my words were really me; my caring concentration on each and every turn of phrase made them mirror my feelings of love for my late friend in the most intensely personal way. Once again I was proud of myself and of the manner in which, over decades, I had come to be able to express myself clearly, strongly, and with a deeply felt voice in a tongue that was not my mother tongue.

Over the course of my life, I have studied lots of languages to various degrees, and I jokingly call myself “pilingual,” meaning that if you were to add up the fractional levels of mastery of all the languages I’ve tackled, you’d get a number a bit over three, counting English as one, French as 0.8, Italian as 0.7, and going down from there, with Chinese as maybe 0.3, at its apex (probably just 0.1 today).

Leaving aside my native tongue, I have devoted many thousands of hours of my life to seven languages (French, Italian, German, Swedish, Russian, Polish, and Chinese)—sometimes flailing desperately and sometimes finding enormous gratification. But through thick and thin, I have relentlessly bashed my head against each of those languages for years, because I love each one’s sounds, words, intonation patterns, idioms, proverbs, poetry, songs, and so on. It’s hard to think of anything else, in the world of the mind, that has pulled me as intensely as my craving to internalize the magic logic of an alien tongue from a faraway place.

But today we have Google Translate. Today we have DeepL. Today we have ChatGPT—and so on. There’s no need for me to list all the powerful technologies that allow anyone today—a monolingual American, say, who has never devoted a single moment to learning, say, Chinese—to write fluent passages in Chinese. Today it’s a piece of cake to send an email in a tongue you don’t know a word of. You just click on “Translate” and presto! There it is! Or at least, there it is, in a certain sense. Assuming that there are no egregious translational blunders (which there often still are), what you are sending off is slick but soulless text.

Just imagine if the Shanghai AI club had asked me to say a few words for the club’s absent members not in Chinese but in English, and then, while I was speaking, they ran my English words through a speech-transcribing app, then a translation app, then a speech-producing app, so that my English words came out, in real time, in Chinese. (In fact, if this were happening today, the speech-producing app could even use my very own voice, speaking with a perfect Mandarin accent!) Had the club gone that techie-type route, which they might well have liked to do, we could have bypassed any need for me to struggle and strain to express myself in their tongue. For both me and the club members, it would have been effortless.

However, in this scenario, the video watchers would be deprived of coming to know key aspects of the very human personality of their invitee. They would not see Douglas Hofstadter (known in Chinese as “Hou Daoren”) groping for Chinese words, would not witness his insecurity, his vulnerability, or, for that matter, his dogged determination; they would merely see an American casually speaking in his native tongue (though what they would hear is perfect Chinese); they would get no sense for the real me, who had devoted thousands of hours, spread out over many years, to grappling with their native tongue. My ideas would come across, more or less, but not those hidden aspects of my self.

But let me play devil’s advocate for a moment. Today’s AI technology allows people of different cultures to communicate instantly and effortlessly with one another. Wow! Isn’t that a centuries-long dream come true, weaving the world ever more tightly together? Isn’t it a wonderful miracle? Isn’t the soon-to-arrive world where everyone can effortlessly speak every language just glorious?

Some readers will certainly say “yes,” but I would say “no.” In fact, I see this looming scenario as a great tragedy. I see it as the beginning of the end of the age-old tradition of learning foreign languages—not only here in America, but even in lands like Holland and Sweden, fabled for their citizens’ near-universal mastery of several tongues. The problem is that people of all cultures instinctively follow the path of least resistance.

Why would anyone want to devote thousands of hours to learning a foreign language if, by contrast, they could simply talk into their cellphone and it would instantly spit out “the same message” in any language of their choice, in their own voice, and with a perfect accent to boot? Who wouldn’t want to be able to have complex conversations with anyone they wish, in any country, no matter what language it involves? Why bother to take countless courses in Chinese and still feel deeply inadequate in it when, in a flash, you can communicate not only in Chinese but also in French, Hungarian, Swahili, and so on?

Suppose I had composed my condolences to Benedetto’s family in English and had then run them through a translation program such as DeepL. The words would have come out very differently from what I wrote in Italian. When I was writing in Italian, I was thinking in Italian, not in English. I was using words and phrases that I have made my own over decades, by having countless intimate conversations with close Italian friends (such as Benedetto himself), by reading hundreds of children’s books in Italian to my kids when they were little tykes, by listening hundreds of times to CDs of lilting Italian songs from the 1930s, by devouring Italian newspapers, by giving untold dozens of lectures in Italian, by watching scores of old Italian movies, by memorizing a few Italian poems, and so on. All that unique flavor, reflecting the myriad idiosyncratic pathways by which I lovingly internalized the Italian language, would be missing from an email that I composed in English and that was instantly converted into Italian by a machine.

You might say that such a loss is a small price to pay—a teeny price to pay!—for the amazing luxury of being able to produce flawless, flowing emails in a hundred different languages, the luxury of being able to give lectures in real time in a hundred different languages, and so forth and so on. Well, I would reply that the “you” who is “writing” or “speaking” so fluently in all these different languages is not you at all. It is, rather, a deepfake version (or a set of deepfake versions) of you.

When I was in the roughest times in my endless battles with the Chinese language, I often wished that I could just get an injection that would make me perfectly fluent in Chinese in a flash. How wonderful it would be to be able, at last, to understand everyone around me, to say anything I wanted to say, and so on! But when I thought about it for only a few seconds, I realized that after getting such an injection, I would not feel proud of having learned Chinese by struggling for many years. My instant fluency in Chinese would, in that case, be a trivial acquisition rather than a precious goal obtained thanks to immense hard work. It would mean nothing to me, emotionally. It would be like arriving at the summit of Everest in a helicopter. It would be like taking a new wonder drug that hugely boosted my muscles and hugely sped up my reflexes, making me (even at age 78!) suddenly able to run faster than anyone else in the world. Next thing you know, this old geezer would be winning a gold medal in the Olympic 400 meters. But big deal! “My” gold medal would be a hollow victory proving nothing about my athletic abilities. It would be purely the result of technological cheating. Likewise, my Chinese-fluency injection would be a hollow victory, because “my” Chinese would not in any way represent my very human, very fallible, but also very determined mind and spirit.

When, in my teenage years, I was striving so passionately to learn French, I sometimes wished that I had just grown up in France with my American parents, so that both French and English were 100 percent native to me. But when I thought about it more carefully, I realized that the reason I was so in love with French was precisely that it was not my mother tongue, and that if it had been, then I wouldn’t be able to hear it in anything like the same way I heard it as an outsider.

Of course, over my six-plus decades of speaking French, I have become less and less of an outsider to it, but still I have somehow preserved the intense love that came from confronting the huge challenge of making French my own in my teenage years, as opposed to simply imbibing it like mother’s milk, as a small child. And I am oh so proud of myself if, after half an hour’s conversation, my native-French interlocutor is startled to learn that I did not grow up speaking French. By dint of intense concentration over decades, I’ve earned that supreme compliment, and knowing I’ve reached that long-dreamed-for level thanks to my years of really hard work is as great a feeling as any I have ever had.

Today’s young people (even in Holland and Sweden) who grow up with translation software, however, will not be lured in the same way that I, as a teenager, was lured by the fantastic, surrealistic goal of internalizing another language. They won’t feel the slightest temptation to devote a major fraction of their lives to slowly and arduously acquiring the sounds, vocabulary, grammar, and cultural richness of another language. To them, someone with my self-punishing attitude would seem hopelessly wedded to the past. Why on earth cling to riding a horse or a bicycle for transportation, when you can drive a car (not to mention flying in an airplane)? What’s the point of going super slowly when you can go superfast? Okay, okay, on a horse or bicycle you’ll see the scenery a bit better, but is it really worth it, when you can cross an entire continent in hours or days, instead of in weeks or months?

The question comes down to why we humans use language at all. Isn’t the purpose of language just the communication of facts? If so, then why not simply go for maximizing the number of facts transferred per second? Well, to me, this sounds like a shockingly utilitarian and pragmatic description of what I view as a perpetually astonishing and quasi-magical phenomenon that lies at the very core of conscious life.

When I speak any language, as all my friends know well, I am always searching for the most appropriate word or idiom, frequently hesitating, stumbling, or suddenly changing course midstream; constantly joking by playing with ambiguity; having fun by putting on droll accents and personas, not to mention coming out with puns (some lovely, some lousy); using alliterative phrases; concocting new words on the fly; making accidental mistakes and laughing at myself; committing deliberate grammatical errors; unconsciously blending idioms and thus creating delightful new turns of phrase; tossing in words from other languages left and right; citing proverbs and quoting snippets of poetry; mixing metaphors; etc., etc. Speaking any language, for me, is a living, dynamic process that is permeated by my own unique humanness, with all its frailties and strengths. How is all of this wildly bubbling richness in Language A going to be mirrored in real time in Language B by a mechanical device that has nothing of those qualities driving it, that has no sense of humor, that has no understanding of irony or self-mockery, that has no awareness of how phrases are unconsciously blended, and so on?

For me, using language is the very essence of being human. When I speak, I am communicating not only facts, but a way of being. Through my word choices and subtle intonations and tiny hesitations and droll puns and dumb errors (and so on), I am revealing who I am. I am not a persona, but a person.

Today, though, it strikes me as possible—in fact, quite likely—that humans are collectively going to knuckle under and throw in the towel as far as foreign languages are concerned. Are we language users going to obsequiously hand over all engagement with other tongues to chatbots? Will young people in the coming decades share my youthful ardent desire to tackle towering linguistic Everests demanding long years of dedication? Or will they opt for the helicopter/chatbot pathway, preferring their linguistic lives to be struggle-free? If everything we might ever wish for is just handed to us gratis on a silver platter, then what, I wonder, is the purpose of living?

As my friend David Moser put it, what may soon go down the drain forever, thanks to these new AI technologies, is the precious gift that one can gain only by immersing oneself deeply in another culture and thereby acquiring an entirely new set of ways of looking at the world. It’s a gift that can’t help but turn any human being into a far richer and broader one. But David fears that it may soon become as rare as hen’s teeth. And, I might add, David knows perfectly whereof he speaks, because in his 30s he recklessly threw himself into the bustling, boiling cauldron of China and its mysterious languages, and after long years of tenaciously clambering up its nearly vertical slopes (sorry for the mixed metaphor!), he emerged as a marvelously fluent speaker of Chinese, able to come out with breathtakingly witty puns on the fly and to do stand-up comedy on national television, not to mention hosting his own weekly TV show, in Chinese, about little-known facets of Beijing.

To Mo Dawei, as David is known in China, it’s incredibly depressing to contemplate the profound impoverishment of people’s mental and emotional lives that is looming just around every corner of the globe, thanks to the slick seductiveness of AI translation apps, insidiously creeping their way into ordinary people’s lives and sapping their desire to make other tongues their own.

When children first hear the sounds of another language, they can’t help but wonder: What in the world would it feel like to speak that language? Such eager childlike curiosity might seem universal and irrepressible. But what if that human curiosity is suddenly snuffed out forever by the onrushing tsunami of AI? When we collectively abandon the age-old challenge of learning the languages of other lands, when we relinquish that challenge to ultrarapid machines that have no inner life of their own but are able to give us fluent but fake facades in other languages, then we will have lost a major part of what it is to be human and alive.

Douglas Hofstadter is a professor of cognitive science and comparative literature at Indiana University at Bloomington. He is the author of, among other books, Gödel, Escher, Bach and Fluid Concepts and Creative Analogies.