Three birds!

About two years ago, I posted the following to an internal company blog (which I’ve altered here by adding a fourth column to the table, to aid in further discussion):

Today, we celebrate Think Friday with something to ponder about language. Consider the words three and bird, in a number of Indo-European languages:

Language	Three	Bird	Cheese
English	three	bird	cheese
Spanish	tres	ave	queso
French	trois	oiseau	fromage
German	drei	Vogel	Käse
Russian	tre	ptitsa	cyr
Italian	tre	uccello	formaggio

Note that the word for “three” is essentially the same word in each language. Note that the word for “bird” is very different in each. Now, surely, way back some time, there was a proto-Indo-European gent on the hunt, who said to his companions the equivalent to, “Look! Three birds!” So... why did the word for “three” and the word for “bird” develop in two vastly different ways?

In the comments to that original post, there were a couple of suggestions:

1. Perhaps abstract concepts, like numbers, tend not to localize, but concrete things, like birds, adopt the local name for the thing.
2. Merchants and trade might be the reason. Perhaps “three” became common because merchants had to understand each other, but they didn’t sell generic “birds”, so no common name came about for that.

To look at both of those, I’ve added the “cheese” column. Cheese is a thing, not an abstract concept, and was surely traded by merchants as a generic item (“I should like to buy some cheese!”). But we see at least two unrelated words here: the English/Spanish/German word comes from the Latin word for cheese, “caseus”. The French/Italian word comes from the Latin “forma”, form, shape, referring to the way cheese is molded into a specific shape. I don’t know where the Russian word (and the apparently related Greek word, “tir”) comes from, but I bet (and hope) that The Ridger will tell us in the comments.

In any case, I’d thought the mercantile explanation was a sensible one, at least as a part of the answer. And Wednesday’s All Things Considered (a conversation with the wonderfully named Tecumseh Fitch) alerted me to two studies published in Nature, one of which directly addresses this question.

The first study looks at how words have changed over time, particularly looking at irregular past tenses and their tendencies to become regular — or not to. As the study decides, “The half-life of irregular verbs is proportional to the square root of their frequency.” In other words, the more often an irregular verb is used, the less likely it is to drift toward regularity. That makes intuitive sense: if you’re used to saying, “I get,” and “I got,” you’re not very likely to come out with “I getted,” unless you’re a small child who’s still got training wheels on his tongue.

The second study looks directly at the “three birds” question that I post, rendered in the paper’s abstract as “two tails”. The researchers find the same effect as was found in the first paper: frequent usage results in slower rates of evolution, and that this factor accounts for half of the effect on the evolution of words.

“The easiest way to think about it is that changes to the higher-frequency words are less likely to be accepted,” says Pagel. “If I say there are two guys coming over the hill to kill us, and in fact there are 20, we might end up dead.”

Well, OK, but mightn't you wind up just as dead if you say there are 20 rabbits coming, and in fact they are lions? But....

What this seems to say, then, is that what’s happening is not that local words are evolving into common ones, but, quite the opposite, that common words for things are, over long time periods, evolving into local words, and that this happens far more slowly for more frequently used words like “two” and “three”, and more quickly for less common words like “tail” and... “bird”?

Is “bird”, then, really a word that’s used sufficiently infrequently such that it evolved local versions quickly? Or is something else going on here? Professor Pagel and his team determined that use-frequency accounts for half of the effect on the rate of word evolution, so maybe the bird thing has been dominated by the other half.

“The analogy with darwinian evolution is crude, although not useless,” says Stephen Pinker, a language and cognition specialist at Harvard. “Some portion of the variance can be accounted for by a quantitative variable such as frequency, but much more remains unexplained. You can’t understand language change without looking at the cognitive psychology of the human brains that learn and transmit the words.”

Maybe, as someone has suggested to me, it’s just that local names for local bird species became the generic terms, giving different local names in different places, and, thus, for different languages.

Or maybe it’s something else entirely. Comments and speculation are, as always, welcome.