Part II · Chapter 9

When There Are No Relatives

Farther than you think

Before we talk about the languages that are not related to English, it's worth noticing how far the ones that are related reach. Look at this table:

Persian is spoken thousands of miles from Madrid or Mexico City. And yet its word for father has exactly the same skeleton as the Spanish one: p · d · r. Its mother, too: m · d · r. That's no accident and no loan. It's inheritance: Persian, Sanskrit, Russian, Spanish, and your own English are branches of the same Indo-European tree.

So the method in this manual doesn't stop at Spanish and German. It reaches, with more effort, to Russian, Hindi, Persian. The farther away the branch, the fewer co-derivatives there are and the more worn down they look, but the map is still the same map.

The edge of the map

Now the real question: what about Nahuatl, Arabic, Japanese, Chinese, Maya? None of these languages belongs to the Indo-European family. Each belongs to a different family, with thousands of years of history of its own.

Endolinguistics calls each of these great families, seen as a system, a macrosystem. Indo-European is one macrosystem. Semitic (the family of Arabic and Hebrew) is another. Uto-Aztecan (the family of Nahuatl) is another. Sino-Tibetan (the family of Chinese) is another.

And there's a rule that doesn't bend: macrosystems don't get mixed. You don't compare an English code with an Arabic one as if they belonged to the same language, because they don't. There are two reasons.

The first is history. Between English and Arabic there are no inherited co-derivatives. If you find an Arabic word that looks like an English one, either it's a loan (like sugar or algebra, which you'll meet in a moment), or it's a coincidence with no shared history: what chapter 12 will call a convergence. Inherited kinship there is none.

The second reason goes deeper. In chapter 3 I told you that classes belong to a family of languages: the regions of the mouth are the same all over the world, but the sensation a group of sounds produces, and how sounds group together, can change from one macrosystem to another. When this research program measured whether the field of meaning of an Indo-European code also shows up in other families, it matched very rarely: less than one time in ten. What is shared across families is something more basic: that languages organize meaning with patterns of consonants. How they do it, each family does in its own way.

Loans that crossed the edge

Words, however, do travel across the edge of the map, even when languages aren't related. They travel as loans, and English is full of them.

Take sugar. It started in India, in Sanskrit, as śárkarā. From there it passed into Middle Persian, šakar; from Persian into Arabic, sukkar; and from Arabic, through Italian and French, into English. Four language families' worth of hands, and the word traveled with the goods. Alcohol comes from the Arabic al-kuḥl, and algebra from the Arabic al-jabr, "the resetting of broken parts." Notice the al- at the start of those two: it's the Arabic word for "the," which came along stuck to the noun.

These are loans, not co-derivatives. They connect English to Arabic by a road, not by a family tree.

What does travel: the way of looking

So is this manual useless for these languages? No, but it works differently. What doesn't travel is the specific codes and their fields. What does travel is the way of looking: finding the skeleton, asking what is root and what is added on, suspecting that behind a list of words there's an order, and looking for that order inside the new language's own system, not inside yours.

Here are two examples.

Arabic: a language of skeletons

Arabic is the best proof that the idea of a skeleton isn't something this manual invented. In Arabic, many words are built from a root of three consonants, and the meaning is shaped by putting vowels and add-ons around them. Look at the root k-t-b, which has to do with writing:

They all carry k · t · b in the same order. The vowels and the ma- at the front aren't part of the root: they're molds, pieces of grammar, and each mold says something (the one who does it, the place where it's done). Anyone learning Arabic soon learns to see the three consonants underneath, and every new root opens up a whole family of words.

That's why Arabic and Hebrew are written mostly with consonants, as you saw in chapter 2. For someone who speaks the language, the skeleton carries the meaning, and grammar supplies the vowels.

Does this mean k · t · b has something to do with some English word with those consonants? No. Arabic is another macrosystem. What you take away from Arabic isn't a code, but confirmation that looking at the bone works there too, with the bones of that system.

Nahuatl: what came to you through Spanish

Nahuatl, the language of the Aztecs, still spoken by more than a million people in Mexico, isn't related to English or to Spanish. But it has lived next to Spanish for five hundred years, and through Spanish it left a gift in your kitchen:

These are loans, and they made a double journey: from Nahuatl into Spanish, and from Spanish into English. Look at the endings. The Nahuatl words end in -tl; the Spanish words end in -te. That's no accident: -tl is a very common noun ending in Nahuatl, and Spanish, which doesn't have that sound at the end of words, always adapted it the same way. It's a loan rule: a correspondence, like the ones in chapter 6, only between languages that aren't related. English then took the Spanish words and adapted them once more: -te became -to in tomato.

If you ever study Nahuatl, it's useful to know that the -tl isn't part of the root: it's an add-on, like the plural -s in English. Toma-, chocolā-, āhuaca- are what's left once you take it off. From there, what you have to learn are the structures of Nahuatl, from inside Nahuatl.

Learning a system from the inside

When there are no co-derivatives to anchor you, the work changes shape. It's no longer about recognizing your own words in another accent. It's about learning how the system thinks: how it builds words, which parts are root and which are molds, which sounds it groups together, which distinctions it cares about. In Arabic, three-consonant roots and molds. In Chinese, tone, which is part of the word (you'll see it in the next chapter). In Nahuatl, the endings and the habit of joining roots together into long words.

Those structures can also be discovered by studying that system's codes: its skeletons, its roots, its repetitions. But they're discovered inside it.

And here's a question with no easy answer, which I'll leave open for you. When an English speaker studies how a language from another macrosystem "thinks," how can they know they're discovering something about that language, and not projecting their own way of thinking onto it? Deciphering another people's thought and laying your own on top of it look far too much alike. The only thing that helps is humility: asking the people who have spoken that language since childhood, checking, and not confusing what seems true to you with what is.

In Part III we leave the skeleton for a while and go to the flesh: the vowels, the rhythm, the tone, everything the voice says besides the consonants.