Slashdot Mirror
Amazonian Tribe Has No Word To Express Numbers
In 2004 we discussed the Piraha, a tribe in the Amazon, when a study appeared characterizing their language as a "one, two, many" language. Now reader mu22le informs us of a new study of the Piraha pointing to the possibility that they use no number words at all. Instead they seem to use the word formerly thought to mean "two" to represent a quantity of 5 or 6, and the "one" word for anything from 1 to 4. The language has about 300 native speakers. "The study... offers evidence that number words are a concept invented by human cultures as they are needed, and not an inherent part of language, Gibson said."
Seems that what they're calling "Numbers" are the same as our quantity descriptors. Small number, medium number, and large number. Seems reasonable, I'm no anthropologist, but I think that numbers really start when you have a lot of trade going on, when you have to KNOW that 5 ears of corn is worth 1 basket of peas.
The previous study had the same basic flaw: they asked the Piraha to count objects that they never normally had to deal with (it was batteries, I think).
What westerners often forget is that many cultures have different numbering systems for different types of things.
If they asked instead, "how many children do you have", or "how many people are there in that hut", they would most likely discover (shock! horror!) that the Piraha count people exactly as you or I. (If we know the individuals we can count up to 10 or so, if we don't, we count up to five or six, then switch to "many").
These experiments look designed to prove something bogus, namely that counting is not an innate skill.
My blog
If I might come in with a computing/neural perspective...
I think that baboons counting 1/2/many is an indicator of the difficulties with bioneural networks: As fundamentally analog systems, they can't subdivide values finely and retain accuracy for any length of time. Thus, they can store 0/2, 1/2 and 2/2 over time, but for more than that they just set an "overflow bit:" there's a lot of 'em.
You can observe the same thing in humans. Look at your mouse cursor, right now - is it on the left or right half of the screen? Obvious. Which third? Easy enough. Which fourth? A little harder. You couldn't really tell me which tenth it's on without measuring. It gets really difficult because your brain's analog systems have difficulty accurately dividing something up that finely.
From that perspective, I think that counting (which implies an increasingly accurate absolute reference for "one" as the max rises) was something born of necessity, because brains are bad at absolute comparisons. They're really good at comparing short-term differentials (there's an edge here, this texture is different, there are more hunters now than immediately before), but they drift almost without bound over time - thus the baboon's arithmetic fudges that "many - many = zero." It's great for adaptability, but bad for being able to hold more than a few single-digit numbers in your head.
If you're trying to show that Amazonians aren't inferior to us, I agree. If you're trying to show that they're superior, I disagree.
Each of us knows what we need to know. Getting "food and water and safety" is the primary task of every individual in a society like that, and you betcha they know a lot about it. We live in a very very specialized society, where a person can spend his whole career getting letters and numbers to appear on a screen correctly and never know where his food comes from.
Trying to get a programmer to live as an Amazonian is more hazardous than trying to get an Amazonian to live as a programmer, precisely because most of the Amazonian's "job" is "try to stay alive." And it is very hard - I'm sure their life expectancies are shorter than ours. If syntax errors made computers explode into shrapnel, it would be more even.