We Are All Born Polyglots — Alfred Tomatis

“We only reproduce the sounds we know how to listen to. That is the fundamental law.”

In brief — Why does a German speak French so well, and a Frenchman English so badly? Why does English have 360,000 words and Spanish barely 45,000? For Alfred Tomatis, everything comes down to the ear. His thesis is at once provocative and tender: we are all born polyglots, capable of every language — because they were all born from one and the same ear, simply steeped in different sound “baths.” If the adult stumbles, it is neither from laziness nor from a lack of gift: it is that the ear has tightened around the music of its mother tongue. The task is to reopen it. In support of this: a colossus whose voice was recovered — Gérard Depardieu — and students who describe English that “comes in all by itself.”

Key points

We are all born polyglots: every language derives from one and the same ear, then shaped by the acoustic “environment” in which it is steeped.
The fundamental law: “we only reproduce the sounds we know how to listen to” — we can only pronounce what we truly hear.
Each language has its frequency band: French, narrow and “sclerosed” between 1000 and 2000 Hz; English, high-pitched, whistled, diphthongized (hence its 360,000 words); Spanish, pure; German, open across several octaves — hence its ease.
The ear is the brain’s dynamo: its great function is to supply the brain with stimulation.
Relearning through “the ear of the womb”: by filtering a language as a fetus would hear it, only the music of it remains — and the brain integrates it even before understanding the words.
Proof through the voice: Tomatis “brought out” the voice of a young colossus muted by shyness — Gérard Depardieu — before his roles as Cyrano and as Christopher Columbus.

One ear at the origin

The thesis fits in a single sentence: “We were all born polyglots, and what makes me say so is that every language — which has nonetheless degenerated since — was born from one and the same ear at the start.” This single ear, “steeped in different environments — different impedances, different resistances,” gave rise to the diversity of languages. But, Tomatis adds, “an ear is always adaptable, able to recover its potential.” In other words: closing is not a fate.

Along the way, a spark of Tomatisian anthropology: it was not standing upright that made us speak, but the reverse. “I believe it is because he was compelled to speak that man stood upright.” The human being, he says, “moves from eating to speaking — which is no small thing.”

“We only reproduce the sounds we know how to listen to”

This is the keystone, which he calls the fundamental law. We can only emit what we hear: the voice is a prisoner of the ear. From this, two corollaries he lays down like a theorem: if you modify hearing, you modify the voice; and conversely, “if the voice becomes good, the ear is modified.” Voice and ear form a loop — to touch one is to act on the other.

The example is delicious: transplant to Quebec people whose language ignores nasality; before long, “they speak through the nose,” like the air of the country. “Canada vibrates at 1500 Hz” — and the ear conforms to it.

Each language, a frequency band

This is where Tomatis becomes most concrete — and most amusing. Each language is said to occupy its own sound window, which sculpts its sounds, its vocabulary, right down to the bodies of its speakers.

French “vibrates” in a narrow band, between 1000 and 2000 Hz, which “scleroses” the French ear: comfortable, but closed — hence the difficulty of widening one’s listening and therefore of learning. With 65,000 words, “that’s enough for it.” English, for its part, “whistles”: open from 2000 to 15,000 Hz, everything in it is diphthongized, stretched, far removed from the written form — hence its 360,000 words. Spanish, “the Spanish of Cervantes,” without diphthong or distortion, remains pure with its 45,000 words: “a language is deformed all the more as it is pulled toward the high frequencies.” As for German, “much more firmly established,” open across several octaves and equipped with a longer analysis time, it makes its speakers “more upright, more open” — and more gifted at languages. “When you go to Germany, you are surprised to find that all Germans already speak French very well.”

And behind the frequencies, always the same idea: “the great function of the ear is to give the brain plenty of stimulation.”

Relearning through the ear of the womb

How does one reopen a closed ear? Tomatis had an intuition: the ear “hears perfectly in the womb” — so why not let language learning “benefit from this uterine course”? By filtering a language as a fetus would hear it, “only the music of the language remains.” And this music, he says, is already “neurologically” the whole system: “the subject, even understanding nothing, already has the field of the language.”

The students’ testimonies in the film describe the same experience in simple words: the English of a film that “comes in all by itself, without paying attention to it”; the need to turn up the volume that fades away; original-version films one “feels” even without understanding everything. One sums it up: “like a color-blind person who, all at once, would see every color.” Many also note the role of posture — holding oneself “more upright” to enter English — and the passive nature of the method, which one follows while reading one’s mail.

The colossus whose voice was recovered

The finest proof Tomatis saves for the end, and it is a true story. A young man “of 17, that colossus you all know,” so blocked that “he could say nothing.” Tomatis “brought out his voice” — and the shy mute became one of the greatest voices of French cinema: Gérard Depardieu, soon to be Cyrano (“’tis a rock, a peak, a cape…”), then Christopher Columbus. “To speak several languages,” Tomatis concludes, “is a benefit: it is several mindsets.” And he dreams, with the linguist Troubetzkoy, of speaking one hundred and twenty: “then we’d be at peace.”

Today: what science says

As is often the case with Tomatis, one must separate the premise from the solution. The premise — we are born capable of every language, and the ear then closes around the mother tongue — is today one of the best-established ideas in the language sciences. The solution — reopening the adult ear through the passive listening of a “filtered” language — remains, for its part, unproven. To state one without the other would be dishonest; to hold them together is to do justice to the intuition while remaining clear-eyed.

“Born polyglots” — confirmed. The infant is indeed a universal listener: at 6–8 months, a baby distinguishes consonants of languages it has never heard; around 10–12 months, this capacity fades in favor of the sounds of its own language (the foundational work of Werker & Tees, then Patricia Kuhl’s “phonemic magnet,” with the famous r/l that Japanese babies stop discriminating). Tomatis called this an ear that “scleroses”; today we speak of perceptual reorganization. A finer nuance — and rather favorable to his optimism: this is not a definitive loss but a reversible specialization, sensitive to social bonds. American babies re-exposed to Mandarin by a live speaker (and not by a video) recover the lost discrimination.

Everything begins before birth — confirmed. Tomatis insisted on “the ear of the womb,” which hears only the music of the language. Science proves him right: French newborns cry with a rising melody, German newborns with a falling one — the prosody heard in utero already shapes the first cries. As early as four days old, the baby distinguishes its mother tongue from a foreign language. An important qualification: what is learned before birth is mainly the rhythm and the melody (the womb acts as a low-pass filter — a detail that, ironically, chimes with Tomatis’s interest in filtering frequencies); the phonemes, words, and grammar are acquired afterward.

“Deaf” to English — confirmed, and it can be repaired. The adult assimilates foreign sounds to native categories: this is the “phonological sieve” that Tomatis described in his own way. But — a crucial point — the current models (Flege, Best) show that the learning mechanisms remain intact for life: the adult ear is not a closed window, it mainly lacks training. And we now know what re-educates it: not the passive listening of filtered music, but active perceptual training with multiple voices (High-Variability Phonetic Training), whose meta-analyses confirm its effectiveness — to the point of improving even pronunciation without ever working on it. Tomatis’s intuition (“we learn through the ear, and the ear can be re-educated”) was correct; it is the means that has changed.

The “bilingual brain” — to be qualified. Tomatis praised multilingualism as “several mindsets.” On the cultural and human plane, nothing to object to. But the famous “executive advantage” of the bilingual, popularized by the press, is today widely contested: after correcting for publication bias, meta-analyses do not find it in healthy adults, and no robust cerebral difference is measured. What holds up better: a finer metalinguistic awareness in children, and above all a delay in the onset of dementia symptoms (of the order of 4 to 5 years) — a delay, not a prevention. Multilingualism remains precious; the bilingual “super-brain,” however, belongs to myth.

Musical ear and languages. The link Tomatis sensed (between the musician’s ear and a gift for languages) does exist — but it is mainly a correlation, strong for sounds (accent, perception of tones), weak for grammar. By contrast, the idea that “making music” would mechanically improve language learning is not established: a musical ear and a talent for foreign sounds go together without either causing the other.

And the method? Here, caution is required. Independent evaluations are negative or neutral (Cochrane review; Corbett 2008 randomized trial showing no benefit; ASHA’s position), and for language learning specifically, there is no robust controlled evidence. The “Electronic Ear” that would reopen the adult ear through filtering remains an unvalidated hypothesis — Tomatis himself, it is said, regretted not having provided more proof. The film is therefore valuable first as thought: a correct intuition of the ear as the gateway to language.

Brought up to date. What remains is the essential, and it is of burning relevance. We learn languages in the age of apps and AI, and research confirms there is no “cliff”: aptitude stays high until around age 17 then declines gently — in other words, it is never too late, but starting early helps. And above all: preserving the ear has never been so urgent. The WHO estimates that more than a billion young people aged 12 to 35 risk an avoidable hearing loss from risky listening, and forecasts 2.5 billion people affected by 2050. “We are all born polyglots” — provided we still preserve the instrument that made it possible.

Sources

The universal listener & specialization — Cross-language speech perception: perceptual reorganization during the first year of life, Werker & Tees, 1984: pubmed · A new view of language acquisition, Kuhl, PNAS 2000: pubmed · Brain mechanisms in early language acquisition, Kuhl, Neuron 2010: pubmed · reversibility through a live speaker, Kuhl, Tsao & Liu, PNAS 2003: pubmed
Prenatal learning & prosody — Newborns’ cry melody is shaped by their native language, Mampe et al., Current Biology 2009: pubmed · A precursor of language acquisition in young infants, Mehler et al., Cognition 1988: pubmed · Statistical learning by 8-month-old infants, Saffran et al., Science 1996: pubmed
The adult ear can be re-educated — The revised Speech Learning Model, Flege & Bohn, 2021 · HVPT: a meta-analysis, Uchihara et al., Studies in SLA 2025: cambridge.org · /r/-/l/ training, Zhang et al., NeuroImage 2009: pmc
Bilingualism — a qualification — Cognitive advantage in bilingualism: publication bias?, de Bruin et al., Psychological Science 2015: pubmed · meta-analysis Lehtonen et al., Psychological Bulletin 2018 · cognitive reserve/dementia, Alladi et al., Neurology 2013: neurology.org
Musical ear & languages — Is musical ability related to second-language acquisition? A meta-analysis, Thompson, Salig & Slevc, Royal Society Open Science 2025: pmc
Optimal period — A critical period for second language acquisition, Hartshorne, Tenenbaum & Pinker, Cognition 2018: pubmed
Tomatis method — level of evidence — Cochrane review Auditory integration training… for autism: cochrane.org · ASHA position: asha.org
Hearing health — WHO, Deafness and hearing loss: who.int · WHO, World Report on Hearing (2021): who.int

Full transcription

Automatic transcription, proofread; some approximations possible.

We were all born polyglots, and what makes me say so is that every language — which has nonetheless degenerated since — was born from one and the same ear at the start. But this ear was steeped in different environments, different impedances, different resistances, and that is where the differences then arise. But an ear is always adaptable, able to recover its potential. Man invented language with the organs he had at his disposal. And was it the upright posture that prompted him to speak? I believe it is because he was compelled to speak that he stood upright.

And man moves from eating to speaking, which is no small thing. Many remain forever at the first stage. We only reproduce the sounds we know how to listen to. That, then, is the fundamental law. Two corollaries follow: that if we modify hearing, we modify the foundation, and, establishing a second corollary, if the voice becomes good, the ear is modified. In Canada, it is interesting to see that all the people whose language has no nasalization, even those who recoil from it, when you put them in Canada, they speak with the intonation of the Amerindians, they speak through the nose.

Because Canada vibrates at 1500 Hz with disconcerting ease. And once you move from 1500 Hz anywhere, you can hear it a kilometer away. And in France, do we vibrate across everything? We vibrate on a wider passband, between 1000 and 2000, but one that scleroses the French ear between 1000 and 2000, so that the Frenchman, when he uses this band to the maximum, is incapable of widening his ear and learns nothing else. It is the language that whistles, because it opens from 2000 Hz up to 15,000 Hz, and so everything is going to be not only whistled but diphthongized. So much so that an Englishman will always be obliged to take a great many words from elsewhere, which he will pronounce to the point of the unpronounceable, and which drifts further and further from the written form.

That is why they have 360,000 words in their language. French, for its part, has a band limited to the vowel range, and it has 65,000 words, which is enough for it. But if we took, for example, Spanish, which is very close to plain emission, there is no diphthong. There is no distortion, there is no diphthongization, and the Spanish of Cervantes can be said today to have only 45,000 words. A language is deformed all the more as it is pulled toward the high frequencies. Another problem is the fact that the production…

German is a much more firmly established language than ours, hence its aptitude for learning languages rather well, notably French and English. German has several octaves; on top of that it has a latency time for analyzing languages that is much longer. And this entails in the German that posture we know in him, he is much more… He is more upright than we are, and he is more open. And it is also the locomotive type, at a given moment, a colossal energy, which is linked to this passband, to this energy, for the ear — the great function of the ear is to give the brain plenty of stimulation. Mozart, who is somewhat external, so he is a little dog, he will touch only, at a given moment, the body’s internal rhythms.

German, for its part, is going to touch the internal rhythms, plus the movement, and more — at a given moment, when you listen to Wagner, you wander through nature, it is enough to hear David Carey to carry on. English is an intermediate zone. But when you go to Germany, you are surprised to find that all Germans already speak French very well, they have been prepared for a long time, and we, with our faculty for believing we’ll be able to do everything and start whenever we want, we speak nothing at the moment. No more English than German. So German bothers us a little, because it is somewhat more complex grammatically, English is a facilitation, but it is still a minor English that we use. As things stand, it is better to get this junction moving for now, but later on, if there is to be a Europe, one must be polylingual, there’s no way around it.

We realized that the ear, in order to develop, is obliged to go through a whole course. It hears perfectly in the womb, and the idea came to me to see whether I could let language learning benefit from this uterine course as well. And it is interesting: when you skim off everything that is language, when you put it through a uterine state, only the music of the language remains. And there, you realize how much the music of this language is already, neurologically, the neurology of the system, everything is involved. There is therefore a greater integration. The subject, even understanding nothing, already has the field of the language.

Ah, little prince, little by little I came to understand thus your melancholy little life. For a long time your only distraction had been the gentleness of the sunsets. How is it? Deeper, deeper, deeper, deeper, those are real brushes. Very good, that’s done. This great drop from 750 to 8000, which has been completely reduced inasmuch as the diaphragm is opened completely up to 8 billion today.

I am pleasantly surprised that, after a year’s interruption, my ear has retained, let’s say, the gains made through the Tomatis method. If I may offer a visual comparison, for me the final result is… You can imagine a color-blind person who, all at once, would see every color. I have managed to listen to things, to hear things that I didn’t hear before. For example, I would listen to a program in English or a film in English. Before, I had a lot of trouble truly understanding the meaning of the sentences.

Whereas now, sometimes, without really paying attention to it, it would come in all by itself. I’d say to myself, well now, I understood. On the telephone especially, first of all through an exchange that was easier. We understood what was being said to us and we could express what we had to say. I do indeed have a tendency to need to hear the sounds and to turn the volume up quite loud. And I noticed, indeed, right at the very beginning, and I think it is still true today, that I can lower the volume and still hear the sounds anyway.

I greatly enjoy films in the original version. I can’t say I understand everything that is said, but I hear, and I clearly feel that I hear, everything that is said. Posture plays a great role. So, for example, we often tend to speak like this. And when we speak French, yes, or even a Frenchman, when he speaks English, he takes up his position or any other position. Whereas the position of the English language, which is often adopted by the English, is to be more upright, to have a much straighter back at that level, a slightly straighter head as well, which means there is more volume and more sonic richness in the timbre of the voice one can emit.

I think I listen to others better and I hear them better. It is a passive method, that is, one can keep on working. One can read one’s mail, prepare meetings, independently of undergoing the treatment. Ah no, that’s rather short, young man. One could say, oh Lord, a good many things, by varying the tone. For example, here — aggressive.

Why, sir, if I had such a nose, I should have it amputated on the spot. Friendly: but it must dip into your cup; to drink, have a special tankard made. Descriptive: ’tis a rock, a peak, a cape. What am I saying — a cape? ‘Tis a peninsula. Well, by God, I’ve had it.

And so, I have already brought out his voice. Because before, he wouldn’t speak to me at all. When he was 17, he had that musculature, that colossus you all know. But he was all the more dangerous because I couldn’t get a word in edgewise. He improved as soon as I brought out his voice. We all know it. He was, after all, the great fellow, by God. But he could say nothing.

And then afterward, his reputation took hold. He was obliged to integrate languages, notably English. And even now, he is poring over a script to try to play Columbus, which he is about to perform. To speak several languages is a benefit, to speak several languages, because it is several mindsets, that’s all. One ought to speak, like Troubetzkoy, 120 languages. Then we’d be at peace.