Start Here

IThe story you were told00:00

no music. Dry voice, close to the microphone.

You were taught that English comes from Latin and Greek. That democracy is Greek and justice is Latin, that the doctor’s words are Greek and the lawyer’s are Latin, and that when you have traced a word back to Athens or Rome you have reached the beginning. The story stops there. It stops there in the dictionary, it stops there in school, and it stops there in almost every book about words ever written for people who are not linguists.

I want to show you where it does not stop.

Take the Greek alphabet — the letters that Plato wrote in, the letters of the New Testament, the letters that became the Roman letters you are hearing this sentence spelled in. Alpha. Beta. Gamma. Delta. Those are not Greek words. They mean nothing in Greek. They are Semitic words, carried across the Mediterranean by Phoenician traders somewhere around the ninth century before Christ, and they mean things in Hebrew and in Aramaic and in Arabic to this day. Aleph is an ox — the letter was once a drawing of an ox’s head, and if you turn a capital A upside down you can still see the horns. Beth is a house; in Hebrew Bethlehem is the house of bread, and Bethel is the house of God. Gimel is a camel. Daleth is a door. The Greeks took the whole list, ox, house, camel, door, and kept the sounds and forgot the meanings, and a thousand years later the Romans took it from them.

Every letter you have ever read is a Semitic word with its meaning worn off.

And it is not only the letters. Sack is Hebrew. Cinnamon is Hebrew. Sapphire and jasper are Hebrew. Camel is gamal. Jubilee is yōbēl, the ram’s horn. Sugar is Sanskrit by way of Persian and Arabic; cotton and coffee and alcohol and algebra and zero and zenith are Arabic; abbot and messiah and sabbath and amen are Aramaic and Hebrew. The Mediterranean was not a wall between Europe and the East. It was a conversation. And the languages on both shores of it were talking to each other long before anyone wrote any of them down.

That is the first hidden thing. Here is the second, and it is the one this show is about.

The textbooks say that Latin and Greek belong to one family, called Indo-European, and that Hebrew and Arabic and Aramaic belong to another, called Semitic — and that the two families are not related. Not cousins. Not anything. They say that the resemblances between them are borrowings, or coincidence, and that nobody can show a common ancestor because the common ancestor, if there was one, is too far back to see.

This show says there was one. This show says every language on earth — Indo-European and Semitic and the great families of Africa and Asia and the Americas — descends from one tongue, spoken once, before we scattered; that the trail to it has not gone cold; and that the proof is still in our mouths, if we learn to read a word by its sound.

I am a physician. My name is Azka Afzal. This is The Terminology Tales, and this first episode is the one to start with, because it is where I tell you what I am arguing, how I intend to prove it, and who tried before me.

title music. Let it run, then under.

IIThe hypothesis04:10

The hypothesis has a name, and the name is the shape of the hypothesis: Proto-World. Proto — first. World — all of it. The first language of all of us.

Let me say what it is, and then what it is not.

It is the hypothesis that language began once. That somewhere — most likely in Africa, most likely more than fifty thousand years ago, though nobody knows — a population of modern humans began to speak, and that every language spoken since descends from that speech the way every person alive descends from those people. One tongue at the root of the tree.

Here is what surprises people: the textbooks mostly agree with that part. Ask a professor of linguistics whether language probably arose once, and most will say yes, probably; the alternative, that it was invented separately in different places, is harder to believe, not easier. What the textbooks deny is the second half of the case: that any trace of that first tongue survives. They say that words wear out — that after eight or ten thousand years of sound changes and replacements, a word has been overwritten so many times that its original form cannot be recovered, and so the families of the textbooks — Indo-European, Semitic, Niger–Congo, Sino-Tibetan and the rest — are as far back as anyone can see. Beyond them, they say, there is only fog.

This show says the fog is in the method, not in the languages. It says that words do wear out — but sounds do not, not in the same way, because sounds are made by a body that has not changed in fifty thousand years. Every human being who has ever spoken has done it with the same tongue, the same teeth, the same soft palate and the same two folds of tissue in the throat. And this show’s hypothesis, in one sentence, is this:

The sounds of a word carry its meaning; they carry it because of what the mouth is doing when it makes them; and because the mouth has not changed, the meanings have not changed either — so that the same sounds carry the same meanings in every language on earth, and the reason is that they came from one place.

That is the hypothesis. Everything else in this show is the proof of it, and the way to evaluate the evidence is a system of ten rules that I am going to lay out for you now, one at a time, because I want you to be able to check my work.

IIIThe ten rules07:00

Rule one. Sound, not spelling.

Writing is about five thousand years old. Speech is at least ten times older, and every word on this show lived most of its life in mouths, not on pages. So the system reads a word by its sound and only afterwards by its spelling. A K, a hard C, a Q and a Greek chi are one sound to the ear; the fact that four alphabets write them four ways tells you about the alphabets. Kayak and cake and quick and chaos all begin the same way in the mouth. When two words look different on the page, the system asks what they sound like. When they sound alike, the page is the thing to doubt.

The people who kept the scriptures knew this better than we do, and I will come back to them.

Rule two. The skeleton carries the meaning.

In Arabic and in Hebrew, a word is a skeleton of consonants, and the vowels are poured in around the skeleton to make it do a job. The skeleton k-t-b carries the idea of writing: kataba, he wrote; kitāb, a book; kātib, a writer; maktab, a desk; maktūb, written, and so, fated. Every Arabic speaker hears the bones inside every one of those words. The system reads every language that way. Strip a word to its consonants and you have its bones, and the bones are what survive the centuries. The vowels are breath; they change with every accent. The consonants are the frame, and the frame is what this show compares.

Rule three. Place is meaning.

Every consonant is made somewhere between your lips and your vocal folds, and where it is made is what it means. The sounds at the front of the mouth — s, sh, z, t, d, th — are made with the tongue at the teeth or just behind them, and they are the sounds of breath going out: motion, flow, the push outward. The sounds at the back — k, g, q, the hard c — are made by the back of the tongue sealing against the roof of the mouth and then breaking the seal: they are the stop, the wall, the container. And so a word is a small drama of the mouth. S-K is breath meeting a boundary. F-T is breath let out through the lips and then tapped shut at the teeth. This is the rule that a physician can prove with your own fingers, and there is an instrument on the website, The Throat, that shows every place, from the folds to the lips, in the order that the first Arabic dictionary used.

Rule four. Letters trade places.

Certain letters turn into one another, and they do it in two ways. Some are equivalents: the same sound written two ways, or a letter that became another letter across a whole language — J and H, as in Spanish, where José is Ho-say; C, K and Q; I, J and Y; U, V and W. And some are shades: one sound at another setting of the mouth. Put your fingers on your throat and say sss, then zzz. Same tongue, same air; only the vocal folds switch on. Say sss, then shhh: same hiss, with the tongue drawn back. Say k, then g: the same seal, with the voice on. The dictionaries treat s and sh and z as three letters. The mouth treats them as one letter under three pressures, and the pressure shades the meaning. I will lay the whole table out for you in a few minutes, because it is the rule the season leans on hardest. For now: a word read through its trades is still the word.

Rule five. A shift is a fingerprint, not a divorce.

In 1822 Jacob Grimm — yes, that Grimm — showed that every k in the ancestor of English became an h on its way into the Germanic languages. Latin centum, English hundred. Latin cornu, English horn. Latin cor, English heart. Latin canis, English hound. And the mainstream, rightly, calls that the proof that Latin and English are one family: a change that runs by rule is a fingerprint. The system agrees, and takes the border off. When sh becomes s becomes z along the road that sugar travelled from Sanskrit to Spain, the word did not stop being the word. When Latin scāla grew a crutch-vowel and became Spanish escala and then French échelle, the ladder did not stop being the ladder. A word that has shifted is still the word.

Rule six. Every root holds its opposite.

This is the rule people find strangest, so let me give you the evidence first. A thousand years ago the grammarian Ibn al-Anbārī wrote a book called Kitāb al-Aḍdād, the Book of Contraries, and it is a catalogue of Arabic words that mean a thing and its opposite: jawn, which is black and is white; qurʾ, which is the period of impurity and the period of purity. In 1884 a German named Karl Abel published an essay on the antithetical meaning of primal words, showing the same thing in the oldest layer of Egyptian, and in 1910 Freud wrote an essay about Abel. English has it too: cleave splits and cleave clings; to sanction is to permit and to punish; fast is fixed in place and fast is moving quickly. In this system a root is a charge, and the charge can run either way. Shukr is thanks and shakk is doubt. The soul is what lifts out of the body and the soil is what it lifts out of. And the difference between Allah and a lie is a slip of the tongue — which is the whole point. The mouth makes the charge; the heart decides its direction. An opposite meaning is not a failure of the match. It is the match’s second half.

Rule seven. Contact is family.

The textbooks file a borrowed word under contact, not descent, as if the two were strangers. But sack came into Greek from a Semitic ship, and sugar came into Europe from Sanskrit through Persian and Arabic mouths, and the letters of the alphabet came from Phoenicia. A borrowed word proves that two peoples stood close enough to speak to each other — and that is the story this show is telling, of a world that was never as divided as its family trees. Every loan is a cousin who came home.

Rule eight. There is no expiry date on kinship.

The comparative method’s defenders say that after eight or ten thousand years the evidence has been erased and nothing can be shown. I will answer that at length in a moment, because it is the objection on which everything else turns. For now, the rule: a limit on what a tool can see is not a limit on what exists. Every language alive is exactly as old as every other. The families are where the scholars stopped, not where kinship did.

Rule nine. Names remember.

A word for bread is replaced when the bread changes. A name is not. Personal names, place names and the names of God are the slowest words in any language, and they keep roots that the everyday vocabulary has worn away. Salem and Solomon and Jerusalem keep s-l-m, wholeness, peace. Shakur and Shakira keep sh-k-r, thanks. Hadid keeps the iron and Haddish keeps the new. The season reads names as the clues they are.

Rule ten. Convergence is evidence.

One resemblance proves nothing, and I will never tell you it does. Two words that sound alike and mean alike could be chance; the world is full of chance. But ten, on the same thread, in one family, are a root — the textbooks say so. And a hundred, on the same thread, across families that never met, are a pattern, and then the only question left is which is likelier: that the pattern arose a hundred separate times by accident, or once. The system bets on once. And it lays out the hundred, on a page of the website called The Register, so that you can bet for yourself.

Ten rules. Sound, skeleton, place, trades, shift, opposite, contact, no expiry, names, convergence. That is the whole system — and before I show you that it is not new, I owe you the fourth rule in full.

IVThe trades16:40

The fourth rule is the one that makes people say you can’t just swap letters. You can. Languages do it constantly, and the textbooks agree; they simply stop counting at the border. So here is the rule laid out in its two halves. The full table, with every example, is on the website, on the page called The System, under The Trades. Keep it open while you listen.

The first half: equivalents. An equivalent is the same sound written two ways, or a letter that turned into another letter across a whole language, every time, so that the two are one letter to the ear.

J and H. Spanish writes the sound that English spells with an H with the letter J. José is Ho-say. A jalapeño is a hala-peño. And the same word wanders between the two letters as it wanders between countries: Don Quixote with an X became Quijote with a J, and the mouth said the same thing all along. J equals H.

C, K and Q. The Romans wrote C. They borrowed K for a handful of words and kept Q for the sound before a U. The Arabs have two letters for it, a shallow kāf and a deep qāf, and so the book is Koran in one spelling and Qur’an in another and one book in the mouth. Kayak, cake, quick and chaos begin with one letter. C equals K equals Q.

PH and F. The Greek letter phi was once a P with a puff of breath; by the time the Romans borrowed it, it was an F, and they wrote it PH to show where it had come from. Philosophy is filosofía in Spanish and falsafa in Arabic. Philip is Felipe. PH equals F.

I, J and Y. Latin had no J at all; the letter I did both jobs, and Iulius was said Yulius. J was invented in the Renaissance to mark the I that starts a syllable, and English then hardened it into the sound we know. Hebrew yod and Arabic yāʾ are one letter, and so Yūsuf is Joseph and José, Yaʿqūb is Jacob and Iago and James, and Yaḥyā is John. In German a J is still a Y: Jahr, year; ja, yes. I equals J equals Y.

U, V and W. Latin V was U, and it was said as a W: vinum was wee-num, which is why English has wine. W is, in its own name, a double U. Germans say their W as a V — Wien is Vienna — and Spanish makes no difference between B and V at all, so Havana is Habana. Arabic wāw becomes Persian V: a walī in Mecca is a valī in Tehran. U equals V equals W, and B is standing right beside them.

W and G. The Romance languages could not say a Germanic W at the start of a word, so they put the back of the tongue behind it — gw — and then dropped the W. Every English word with a W that has a French cousin with a GU is one word twice. War and guerre. Ward and guard. Warranty and guarantee. Wise and guise. William and Guillaume and Guillermo. W equals G.

F and H. Every Latin word that began with an F begins with an H in Spanish, and the H has since gone silent. Filius, son, is hijo. Farina, flour, is harina. Ferrum, iron, is hierro. Fernando is Hernando. The breath moved from the lips back to the throat and then went out altogether. Japanese did the same the other way round: its H is said F in front of a U, which is why Fuji and Huji are one mountain. F equals H.

S and H. This one the textbooks love, so listen. Greek and Persian both turned every old S at the start of a word into an H. Sanskrit sapta and Latin septem are Greek hepta: seven. Latin sal is Greek hals: salt. Semi and hemi, super and hyper: one prefix in two alphabets. And Sanskrit Sindhu, the great river, is Persian Hind — which is where India and Hindu come from: the people of the Sindh, with the S turned to H by a Persian mouth. S equals H.

And Grimm. I gave you the K that became H: centum, hundred; cornu, horn. The same law turned every P into an F — pater, father; pes, foot — and every T into a TH — tres, three; tu, thou. Arabic, which has no P, does its own version: every Persian P becomes an F in an Arabic mouth, so Pārs became Fārs and the language of Persia is Farsi, and Plato is Aflāṭūn. K equals H, P equals F, T equals TH.

Three more that the ear will thank you for. L and R: in Japanese and Korean they are one letter, and Latin arbor is Spanish árbol, marmor is marble, peregrinus is pilgrim — the peregrine falcon is the pilgrim by another road. D and L: the oldest Latin said dingua and dacruma where classical Latin said lingua and lacrima, the tongue and the tear, and Greek kept the D in dakru; Odysseus is Ulysses. And G, J and Y: a G in front of a thin vowel slides forward, so Latin gentilis is gentle and genteel and gentile, three arrivals of one word; Arabic jīm is a G in Cairo, so Jamal is Gamal and Jīza is Giza; and the Old English G softened all the way to a Y, so that garden and yard are one word, and gelb is yellow, and gestern is yesterday.

That is the first half. Every one of those trades is in the textbooks. The system’s only addition is to say: if this is how letters behave inside a family, it is how they behave across one.

The second half: shades. A shade is one sound at another setting of the mouth — the same place, the same skeleton, a different pressure — and the pressure shades the sense.

Put your fingers on your throat again and say sss, then zzz. Same tongue, same air; the folds switch on. Now say sss, then shhh: the same hiss, with the tongue drawn back. S, Z and SH are three letters in the dictionary and one letter at three settings in the mouth. Salām and shalom are one peace. Sibboleth and shibboleth are one ear of grain — and the Book of Judges says that forty-two thousand men died at the fords of the Jordan for the difference. Persian shakar, Arabic sukkar, Spanish azúcar are one sugar. And the S-K root of the third episode is seek at its thinnest, shake blown harder, and zikr — remembrance — voiced and turned inward.

T, D and TH: the tongue taps the teeth; the folds are off for T, on for D, and the tap loosens to a breath for TH. Tin, din, thin. The Arabic letter thāʾ is said TH in Mecca, S in Tehran and Lahore, and T in Cairo, so that three is thalātha, salāsa and talāta in three cities — and a Castilian says his C and Z as TH, so that Barcelona is Barthelona.

K, G and Q: one seal at the back of the mouth, voiceless, voiced, and driven down into the throat. Cot, got, Qur’an. Arabic keeps K and Q as two letters, and the shade between them is a shade of meaning: kalb is a dog and qalb is a heart.

P, B and F: the lips close, with the folds off and then on, and then loosen to a breath. Pat, bat, fat. The root of opening in the next episode runs on exactly this shade: Arabic fataḥa opens, Hebrew pātaḥ opens, Latin patēre lies open.

V, W and B: the lower lip on the teeth; the lips rounded without touching; the lips closed. Latin vinum is English wine, and the vine is the plant that makes it — one word at two settings, the plant and the drink. The fourth and fifth episodes are built on this shade: V-S, the vessel and the voice; and W-S, one setting softer, the whisper and the wisp.

And M and N, the two hums: with the lips closed, and with the tongue at the ridge. Latin septem is seven; novem is nine. Hebrew plurals end in -īm and Arabic plurals in -īn. Solomon is Sulaymān.

So when this show puts shukr beside shake, or ḥadīd beside hide, or fataḥa beside fat, it is not swapping letters at random. It is walking the trades — every one of them a trade the textbooks already accept inside a family — and it is showing you the same skeleton on both sides of the line. Equivalents, and shades. By the end of the season you will not need the table.

Now let me show you that none of this is new.

VHow the ancients read a word27:00

The idea that a word means what it sounds like is not a modern idea, and it is not a fringe one. It is the oldest idea in the study of language, and the people who held it were the people who kept the oldest texts on earth — and they held it, I think, because those texts were spoken long before they were written, and a spoken text is guarded by the ear.

The Vedas of India are called śruti: what is heard. They were composed more than three thousand years ago and memorised, syllable by syllable, by families of reciters, with systems of chanting designed so that a single misplaced sound would be caught — the words recited forwards, then in pairs, then in reversing pairs, so that the sound of the text was checked against itself. They were not written down for something like two thousand years. In the fifth century before Christ a scholar named Yāska wrote the Nirukta, the first book of etymology in the world, and its founding principle is that every noun comes from a verb — that a thing is named for what it does — and that you can find the doing by listening to the name.

The Qur’an’s own name means the recitation. It was delivered aloud, memorised aloud, and transmitted for a generation by people who had heard it before it was gathered into a book; and the science of its correct sounding, tajwīd, is a science of the mouth — of where each letter is made, how long each vowel is held, which sounds may lean into which. In the eighth century, in Basra, a scholar named al-Khalīl ibn Aḥmad set out to make the first dictionary of Arabic, and he did not order it by the alphabet. He ordered it by the throat. He began with ʿayn, the sound made deepest in the pharynx, and worked outward to the lips, because he had decided that the true order of sounds was the order of the body. And within each entry, he listed every permutation of a root’s consonants together — k-t-b with b-t-k with t-k-b — as if they were one family. The book is called Kitāb al-ʿAyn, the Book of ʿAyn, after the letter it begins with.

Two centuries later, in Baghdad, Ibn Jinnī — one of the greatest grammarians who has ever lived — wrote a chapter in his al-Khaṣāʾiṣ with a title that could be the title of this show: on the sounds of words resembling their meanings. He gave examples: that a word for chewing something hard has a hard sound in it, and a word for chewing something soft has a soft one. And he proposed what he called the greater derivation — that all the permutations of a root’s letters share a single core meaning. His contemporary Ibn Fāris wrote a dictionary, the Maqāyīs, on the principle that every root has one or two core senses from which all its words descend. And Ibn al-Anbārī, whom I mentioned, catalogued the words that mean a thing and its contrary. Rule two, rule three, rule six: the Arabs of the tenth century had them in writing.

And Plato had them earlier. In the Cratylus, the dialogue on the correctness of names, Socrates analyses the Greek alphabet by what the mouth does: rho, the trilled r, was given to motion, because the tongue is most agitated in pronouncing it; lambda, the l, to smoothness, because the tongue glides; and the breathy letters — phi, psi, sigma, zeta — to everything that blows and shivers and shakes. Socrates does not settle the argument; he leaves it open and tells Cratylus to keep investigating. This show is Cratylus, still investigating.

Even the scriptures do etymology. In Genesis, the tower is called Babel because there the Lord balal, confused, the language of the earth. Adam is named for adamah, the ground he was made from; Eve, Ḥawwah, for ḥay, life; Isaac, Yiṣḥāq, because Sarah laughed. The oldest texts we have explain names by sound. It is the modern dictionaries, not the ancients, that decided sound was an accident.

So when I say sound, not spelling, I am not proposing something new. I am proposing that we go back to reading the way everyone read before print — the reciters of the Veda, the guardians of the Qur’an, the grammarians of Basra and Baghdad, and Socrates.

VIThe same and the opposite32:10

I want to spend a few more minutes on the sixth rule, because when people first hear it they think it is a trick — a way of never being wrong. If a root can mean a thing or its opposite, what can it not mean?

Here is the answer. It can mean the thing, or its contrary, and nothing else. S-L can mean the whole and the broken, the peace and the surrender, the soul and the soil. It cannot mean yellow, or Tuesday, or to swim. The charge has a direction and a reverse; it does not have a random value. That is a constraint, and a strong one. And the reason the reverse exists is physical, and every physician knows it: the body’s responses come in pairs. The same shudder is fear and is ecstasy. The same catch of breath is a laugh and a sob. The same word — shake — is the tremor of terror and the tremor of the dance. A root names the event in the body; the heart supplies the sign.

Say Allah. Now say a lie. The tongue does almost the same thing — the breath, the l laid along the ridge, the open vowel — and one of them is the name of truth and the other is its absence, and the whole distance between them is a slip. That is not a coincidence I am exploiting. It is a rule I am stating: the sounds that name the highest thing are the sounds that, turned over, name its shadow, and the turning is the smallest motion the mouth can make.

The Arabs saw this and wrote it down. Abel saw it in Egyptian. Freud saw it in dreams, where a thing and its opposite are the same image. This show sees it in every root of the season, and it will show you the pair every time.

VIIEveryone who came before34:20

Now the history. Because the hypothesis I am proposing has been proposed before, more than once, by serious people, and you should know what happened to them.

It begins with a judge. In 1786, in Calcutta, Sir William Jones, a British judge who had learned Sanskrit to read the Hindu law codes, gave a lecture in which he said that Sanskrit, Greek and Latin bore a resemblance to each other so strong that no philologer could examine all three without believing them to have sprung from some common source — which, perhaps, no longer exists. That sentence founded a science. Within forty years Rask and Grimm had shown the sound shifts that made the resemblance a proof, and within a century the scholars had reconstructed the common source itself — Proto-Indo-European, a language nobody ever wrote down, recovered from its children. That is the thing to remember about the mainstream: its greatest achievement is the recovery of a lost tongue from the sounds of its descendants. The whole discipline is built on the belief that you can hear a dead language in living ones.

The question was always: how far back? And in 1903 a Danish linguist named Holger Pedersen gave a name to the further answer. Nostratic — from Latin nostrās, of our country: our languages. Indo-European, he proposed, was kin to Semitic, to Uralic, to Altaic, to more, in one super-family. Two years later, in Bologna, Alfredo Trombetti published L’unità d’origine del linguaggio — The Unity of Origin of Language — and argued the whole thing, monogenesis, in full, with comparisons across every family he could reach. He was a professor; he was learned; and he was laughed at for a hundred years.

Then Moscow. In the 1960s a young linguist named Vladislav Illich-Svitych began the Nostratic comparative dictionary — sound correspondences, reconstructions, the whole apparatus that Grimm had used for Indo-European, applied above it. He died in 1966, in a road accident, at thirty-one, with the first volume unfinished. His colleagues finished it. Aharon Dolgopolsky carried it for forty more years and published the Nostratic Dictionary in 2008, at Cambridge. Allan Bomhard, in America, published his own reconstruction the same year.

Then Stanford. Joseph Greenberg, the most important classifier of languages of the twentieth century, proposed a method called mass comparison: instead of proving two languages related by reconstructing every sound law, lay many languages side by side and let the pattern show itself. He used it on Africa, and reduced the hundreds of African languages to four families — and the field dismissed it, and then, over twenty years, adopted it; his four families are the ones in every textbook now. He used it on the Americas, in 1987, and proposed that most of the languages of the New World are one family, Amerind — and the field dismissed that, and has not adopted it, and the fight is still on. And in 2000 he used it on Eurasia and proposed Eurasiatic, which is Nostratic by another route.

Then his student. Merritt Ruhlen published The Origin of Language in 1994, and with John Bengtson a paper called “Global Etymologies”: twenty-seven roots, they argued, that could be found in every family on earth — tik for finger, aqʼwa for water, kuna for woman, mena for to think. The fifth episode of this show’s first season was going to be about those twenty-seven; it will be, in its turn.

Then Moscow again. Sergei Starostin built a database called the Tower of Babel, with the reconstructions of every family he could gather, and proposed a super-family above Nostratic — he called it Borean, the northern — that would hold most of the languages of the Old World and the New in one tree. He died in 2005, at fifty-two, in the middle of it.

And there is a name I want you to know because it is closest to what this show does. Saul Levin was a professor of Hebrew and of Greek, at Binghamton, and in 1995 he published a book called Semitic and Indo-European: The Principal Etymologies — four hundred pages arguing that the two families the textbooks keep furthest apart share roots, and showing them. And in France, from the 1990s on, Georges Bohas built the theory I mentioned: that the three-letter roots of Arabic rest on two-letter cores whose sounds carry meaning — matrices and etymons, he calls them — which is rule two and rule three of this system, worked out inside a single language by a man who has spent his life on it.

What happened to all of them?

They were told the same thing. That the comparative method cannot see past eight or ten thousand years. That beyond that, resemblances are chance. That mass comparison finds patterns in noise. That Nostratic is unproven, Amerind is unproven, Proto-World is unprovable, and that the right answer to the question where did language come from is we cannot know. The objection is always the same objection, and it is a number, and I am going to take it apart now.

VIIIThe wall at ten thousand years40:00

Here is the objection at full strength, because you cannot answer an argument you have only heard mocked.

Words are replaced. Every language, over time, swaps out some of its vocabulary for new words — borrowed, invented, drifted. And every language changes its sounds — a k becomes an h, a p becomes an f, a vowel drops. Now imagine two languages that were one language ten thousand years ago. In ten thousand years each has replaced most of its words and changed most of its sounds, so that the few words they still share have been changed on both sides, in different ways. The mainstream says: at that distance, the words they still share look no more alike than words chosen at random would. The signal has dropped below the noise. And so — the argument goes — the comparative method has a horizon, somewhere around eight or ten thousand years, past which nothing can be shown, and every case for kinship beyond it is a case that cannot be tested.

That is the wall. It is quoted as if it were a law of nature. Here is why it is not.

First: it is a statement about a tool, not about the languages. We cannot see has been quietly turned into there is nothing there. But every language alive is the current form of an unbroken line of speakers reaching back to the first speakers; the kinship exists whether or not one method can resolve it. The wall tells you where the method’s eyesight fails. It does not tell you that the family ends at the edge of the light.

Second: the number is an average, and the average hides the range. It was worked out from how fast words are replaced, on the whole, in a few families with long written histories. But not all words are replaced at the same rate. In 2013 Mark Pagel and his colleagues at Reading published a paper in the Proceedings of the National Academy of Sciences showing that a small set of words — the pronouns, the low numerals, mother, fire, hand, hear, who, what — are replaced so slowly that their forms can be followed across the Eurasiatic families for fifteen thousand years and more. The wall moved, the moment somebody measured the slow words instead of the average ones. It was never a wall. It was a fog, and it thins where the words are old.

Third: the tool was built for Indo-European, whose languages have been written down for three thousand years, so that every step of a sound change can be checked against a manuscript. Most of the languages of the world have no manuscripts. To judge them by a standard built for written languages is to judge the ear by the rules of the page — and rule one of this system says the ear came first. The reciters of the Veda held a text intact for two thousand years without writing. The method that says sound cannot survive ten thousand years has never had to account for them.

Fourth: the history of the field is a history of impossible connections becoming textbook ones. Nobody believed Hittite, dug out of the ground in Turkey in 1906, could be Indo-European; it was, and it confirmed a sound that had been predicted and never heard. Nobody believed Tocharian, found in the deserts of western China, could be Indo-European; it was. Greenberg’s Africa was dismissed and then adopted. And in 2010 Edward Vajda showed, with sound correspondences of exactly the kind the textbooks demand, that Ket — a language spoken by a few hundred people on the Yenisei river in Siberia — is kin to the Na-Dene languages of North America: to Navajo, to Apache, to Tlingit. Across the Bering Strait. Across more than ten thousand years. Many of the field’s most careful people have accepted it. The wall was leaned on, and it moved.

And fifth, the one I care about most: the mainstream has already measured the thing this show is built on. In 2016 a team led by Damián Blasi took a hundred basic words in word lists covering nearly two-thirds of the languages on earth and asked whether particular sounds turn up in particular meanings more often than chance. They do. Nose tends to hold an n. Round and red tend to hold an r. Tongue holds an l; small holds an i; full holds a p or a b. Across every continent, across families that no tree connects. The authors say the pattern must have arisen independently, over and over, because the same mouth keeps making the same choice. I say: or once. The finding is the same either way — sound carries meaning across every language on earth — and the textbooks published it and declined to name it.

So the wall is a convention, and I am setting it aside. Not because I think evidence does not matter. Because the evidence is on the far side of it.

IXWhat this season will do45:40

Here is how the season works, so that you can follow it and check it.

Each episode after this one takes one root — a skeleton of two consonants — and reads it by the ten rules. It begins with the sound: where the mouth makes it, what the mouth is doing. It gives you the root in the family you know best, and then it walks it across the borders, one family at a time, until you have heard the same sound carry the same sense in six or seven languages that the textbooks say never met. Every word goes on The Register on the website, language by language, with the dictionary it was checked against beside it. Every time a word means a thing and its opposite, the episode shows you the pair. And every episode ends with the oldest evidence of all: a name, or a line of scripture, that has kept the root intact while the everyday words wore away.

The second episode is Al-Fattāḥ — the Opener, one of the ninety-nine names of God — and the root F-T: the breath let out through the lips and tapped shut at the teeth, the sound of opening, from the first chapter of the Qur’an to fat and fed and fate and faith. The third is Shook Ones, S-K, the breath and the wall — Tupac, Sheck Wes, Shaquille and Shakira, and a hundred and twenty words in nine languages. The fourth is A Voice or a Vessel?, the root V-S — vox and vase, the Sanskrit goddess of speech and the Arabic name of the All-Embracing. The fifth is Wassap, W-S, what the vessel becomes when the lips barely touch: the whisper, the wisp, the wasp, the peace at the end of every greeting. The sixth is Salām, S-L, the sel that ends vessel: peace, wholeness, Salem and Solomon, the soul and the soil. The seventh is Hadid, H-D — iron and the new, the thing told and the thing hidden. And the eighth is Uncle Sam, S-M — the same, the name, the sky, and the question of whether the family of nations was ever more than one.

Seven roots. Seven demonstrations. And by the end of them I will not have asked you to believe anything. I will have asked you to listen, and to check, and to decide for yourself which is likelier: seven threads running through every language on earth by accident, or one tongue.

beat. Music begins, low.

The story you were told stops at Rome. This one does not stop.

I am Azka Afzal. This is The Terminology Tales. Start here — and then go to the Opener.

music up and out.

XSources49:00