Kingdom of Speech


There’s a tribe in the Amazon called the Pirahã that talks like birds. Like, they communicate using variations in pitch, stress, and rhythm, as opposed to consonants and vowels.

That’s not the weird part. Researchers who spent fifty years studying the Pirahã language say that it doesn’t adhere to Noam Chomsky’s theory of a universal grammar.

Tom Wolfe has a new book called The Kingdom of Speech that describes the rivalries between linguistics researchers over whether or not humans have an innate idea of grammar. Chomsky’s key idea is that human brains are Turing machines capable of producing infinite combinations of sentences from a finite number of words. The reason why we can make infinite use of finite means is because language has a recursion function, which is the ability to put a sentence inside another sentence. ←See?? I just did one right there!

The Pirahã language is non-recursive, so their language lacks relative clauses. Here’s an example: Ko Paitá, tapoá xigaboopaáti. Xoogiai hi goo tapoá xoáboi. Xaisigiai. Word for word, it translates to “Hey Paitá, bring back some nails. Dan bought those very nails. They are the same.” In English, we would say “Hey Paitá, bring back some nails that Dan bought.” Without recursion, they need extra statements to perform the same function.

Right, who cares. But when you lose recursion, you also lose numbers and counting. Like, any number greater than ten is a recursive word. “Twenty-five” is actually 2*10+5, which we can figure out in our Turing-complete heads. Recursion lets us make infinite numbers out of ten numerical words.

Without recursive numbers, the Pirahã only have three words to describe quantity: hói, hoí, and baagiso. This roughly translates to “small amount”, “somewhat larger amount”, and “many”. One researcher showed a group of Pirahã between one and ten spools of thread, and asked them to describe how many they saw. Responses varied, even when people saw the same quantity twice. The word used depended entirely on the last-seen quantity. If a Pirahã subject saw one spool, and a second one was added, the subject would describe the two spools as “somewhat larger amount”. If the subject saw three spools, and one was taken away, the subject described the two spools as “small amount” [1].


It wasn’t just a language thing. Without numbers or a writing system, there’s no way to track quantity. When Pirahã subjects were shown a group of 1-10 spools, and then asked to identify whether or not a second set had the same number of spools, they failed for quantities greater than three.

Researchers have observed this lack of numerical vocabulary in multiple indigenous groups. Possibly all the early humans were like this. But if primitive people had no concept of quantity, how did they exchange goods with other humans?

Adam Smith popularized the theory that money evolved from barter – people needed a store of value when there was no coincidence of wants. Socialists have a competing theory that money came from theft – people needed a unit of account to quantify how much they were owed after being robbed. The social anarchists might be right on this one.

In the Amazon, river traders come by daily looking for jungle things like Brazil nuts and sorva fruit. They get these from indigenous people in exchange for clothing, manioc flour, and booze.

As far as the Pirahã are concerned, fruit and nuts are all over the place; it just takes time to gather them. Their culture has no sense of time, so they’re not exchanging a scarce resource here. The stuff they really want is the sugarcane rum, which is quite cheap to the river traders. Both sides think they’re getting a good deal, when really the indigenous people are getting ripped off if you take market prices into consideration. It’s like how the Native Americans gave Manhattan to the Dutch for some stupid beads.

Wolfe says the Pirahã asked their resident researcher to teach them numbers because they thought the traders were cheating them. They couldn’t possibly know the market price for Brazil nuts versus rum, but at least they could figure out which traders were the stingy ones.

Anyway, after eight months of lessons not one Pirahã could count to ten [2]. If it turns out that counting, and hence recursion, is an artificial construct, then there’s no such thing as a universal grammar that lives in a language organ. Tom Wolfe managed to write a whole book about the controversies around this debate. As they say in academia, the fights are vicious because the stakes are so low.

1. M. Frank, et al. Number as a cognitive technology: Evidence from Pirahã language and cognition. Cognition, Volume 108, Issue 3, September 2008.
2. Daniel Everett. Don’t Sleep, There Are Snakes: Life and Language in the Amazonian Jungle. 2009
3. The Interpreter –New Yorker, April 16 2007