Part IV · Chapter 12

The Mirage

Faces in the clouds

It's happened to you. You look at a cloud and see a dog. You look at an outlet and see a scared little face. The front of a car seems to be smiling at you. The same number shows up three times in one day and you feel the universe is trying to tell you something.

This tendency to see connections and figures where there aren't any has a name: apophenia. It isn't an illness, and it isn't something only other people have. It's the price of having a mind that looks for patterns, which is to say, the price of having a mind. Our ancestors who saw a tiger in every shadow survived more often than those who saw none; we inherited from them a mind that would rather see too much than too little.

And here I have to be very honest with you: apophenia isn't the opposite of what this manual teaches. It's its engine without brakes. The same ability that lets you see that father and padre are relatives can lead you to see relatives where there's nothing. This chapter is about the brakes.

The best lesson I know

I promised you this at the beginning. Here it is.

Same meaning. Same skeleton. Same code. If there were ever two words in this whole manual that seemed to shout "we're related!", it's these two.

And they aren't.

Much comes from Old English micel, which meant "big." Mucho comes from Latin multus, from a completely different Proto-Indo-European root. The two words traveled roads that never touch, and still they ended up sounding almost the same and meaning the same thing.

But the story has one more twist, and it's the best part. If much isn't related to mucho, who is it related to? To something you use all the time without thinking: Greek mégas, "great," the one in mega-. Megaphone, megabyte, megastar: there's the cousin of much. It's also related to Latin magnus, the one in magnitude and magnificent. And does mucho have cousins in English? Yes: Latin multus lives in multi-. Multitude, multiple, multimedia.

See what happened? The eye saw much = mucho and thought it saw a family tie. That's where it went wrong, and that's the mirage: history says the real relatives were hiding where nobody would look for them, in a tech prefix and a math word.

There are more cases like this:

A mirage, yes. But the resemblance is still there

So far history has answered one question: are they related? No, they aren't. But there's a second question history doesn't answer: then why are they so alike? It's very easy to mix up the two, and you can go wrong in either direction.

1. Believing they're related. That's the mirage, and it really is one. We checked it with historical linguistics, and there's no way around it.
2. Believing that, since they aren't related, there's nothing there. This mistake is just as big, and it's the one almost nobody sees. Much and mucho are still practically the same thing: the same bone, the same code, the same meaning. So are day and día: a single tooth sound, and both name the day. That isn't an illusion. It's a fact anyone can check.

The illusion wasn't in the resemblance. It was in the explanation the eye attached to it: kinship. When history takes that explanation away, the resemblance doesn't vanish. It stays right there, unexplained.

And here endolinguistics does something historical linguistics, by its own method, doesn't do: it records the coincidence as a relation. When two words with different histories arrive at the same code and the same field of meaning, endolinguistics calls it a convergence. It isn't kinship, and it's recorded separately from kinship so the two are never confused. But it isn't thrown away either. You've already seen others in this manual: Italian caldo ("hot") and English cold, which come from unrelated families, have the same code Χ·Λ·Θ, and sit at the two ends of the same axis, temperature (chapter 4); terrain and rotate, with the same nucleus in reverse order.

Why do convergences happen? Nobody fully knows. There may be some chance in it: consonants are few and words are many, so some encounters are to be expected, and that has to be counted before claiming anything. There may be something about how the body makes and feels sounds, as you saw with vowels in chapter 10. There may be something endolinguistics doesn't yet know how to name. Until we know, the honest place for much and mucho is an open case: a relation you can see, that you can check, and that doesn't have a complete explanation yet. You don't invent a kinship that doesn't exist, and you don't dismiss the resemblance with "it's pure coincidence," which closes the question without answering it.

That's the mystery of coincidence, and it's one of the most beautiful in this whole manual.

The referee

So what separates a finding from a mirage? Not the operation: both consist of linking what was separate. What separates them is the referee.

In the game of learning languages, the referee is history. The kinship of father and padre isn't an impression: it's documented link by link. The kinship of much and mucho doesn't exist, and an etymological dictionary tells you so in a minute. The whole method fits in a double gesture: see the connection, and ask history what kind it is. Seeing without asking is methodical delusion. Asking without seeing is not playing at all.

But look carefully at what that referee decides. History tells you whether there's kinship. It doesn't tell you whether there's a relation. A convergence like much and mucho goes past the referee without kinship, and it's still a relation waiting to be explained.

Three registers

Things aren't as simple as "history yes = true, history no = false." There's a middle ground, and keeping quiet about it would be dishonest.

Endolinguistics distinguishes three registers of sense, three ways a connection between words can exist:

1. Found sense. The connection is already made; history guarantees it. Father and padre. It's a fact.
2. Constructed sense. History doesn't guarantee the connection, but it isn't a delusion either: someone builds it through an interpretation, declares it as an interpretation, and stands behind it.
3. Projected sense. Someone asserts a connection as if it were a fact, where nothing supports it. That's apophenia.

Only the third one is the mistake.

With much and mucho you can see all three registers at once. That they match in code and meaning is a fact: that's found. Saying they're related would be projected: history denies it. And any explanation someone offers for why they're so alike is, for now, constructed: it's worth something if it's declared as what it is.

Here's an example of the second register, constructed sense. There's a group of Spanish words that many people in this tradition are fond of: casa (house), caer (to fall), caso (case), causa (cause). You can read a thread running through them: the house as the place where you finally fall into rest; the case as what falls in front of you; the cause as the reason things fall one way and not the other. Is this a historical fact? No, and nobody serious presents it as one. It's a reading: a sense that is built, and that's worth exactly as much as its power to shed light on something.

The difference between a reading and a mirage isn't in the operation. It's in the honesty of the gesture. The mirage says "this means that." The reading says "this can be read this way," and takes responsibility for it.

Almost all poetry lives in the second register. The poet brings together words that language left close to each other and draws a new sense out of them. The poet's operation and the operation of someone who sees faces in clouds are the same. What sets them apart is that the poet knows they're building, and signs it as a construction.

When the mirage becomes history

One last twist, because reality is more interesting than rules.

English has a word, crayfish, for a small freshwater crustacean, something like a little lobster. It looks like a compound word: cray + fish. But the animal isn't a fish, and the word wasn't built that way. It comes from Old French crevice, which has nothing to do with fish. English speakers heard that ending, -vice, and it sounded to them like fish. They saw a connection that wasn't there. And since everyone saw it, the word changed: today we say crayfish.

This is called folk etymology. From the point of view of history, it was a mistake. But that mistake, shared by a whole community, remade the word. The apophenia of many people, once it becomes habit, becomes history.

This teaches us something subtle. Saying "that's apophenia" is also a judgment, and every judgment has a scope. It's apophenia with respect to the word's history. It isn't with respect to how its speakers use it today. When you say something is a mirage, also say for whom, and at what scale.

Six questions before you believe a connection

Keep this list. It will serve you with languages, and beyond them.

1. Did I check? Did I look up the etymology in a dictionary, or am I trusting the resemblance?
2. Is it the same family of languages? If I'm comparing languages from different macrosystems, the connection isn't inheritance: it's a loan or a convergence.
3. Am I adding up classes? If my explanation says "this consonant means this and that one means that," I've left the method.
4. Did I count the ones that don't fit? Two pretty examples prove nothing. How many words with that code have nothing to do with what I'm saying?
5. Am I presenting it as a finding or as a reading? If it's a reading, I say so. That's fine: readings are valuable, as long as they're presented as what they are.
6. Does it survive someone who doesn't think like me? If I explain it to a skeptic, what do they say?

That leaves the biggest question of all, the one you may have asked yourself at some point in this manual. If codes aren't meanings, if they aren't just history, if they aren't pure coincidence… what are they? The last chapter doesn't have the answer. But it tells you where to look.