According to a report in the New York Times, frequently-used words evolve more slowly than rarely used ones:
Some words evolve rapidly, with a result that there are many different word forms, what linguists call cognates, for meanings across languages. "Bird," for example, takes many disparate forms across other Indo-European languages: oiseau in French, vogel in German and so on.
But other words, like the word for the number after one, have hardly evolved at all: two, deux (French) and dos (Spanish) are very similar, derived from the same ancestral sound.
Seems reasonable. In our travels across Europe, we found that "yes" and "no" were very similar in different languages -- until we got to Greece, where their word for yes was pronounced "neh," and the word for no was "ochi." But despite anomalies such as this, overall more frequently used words tend to be more similar across Indo-European languages. From the study abstract:
Here we use four large and divergent language corpora (English, Spanish, Russian and Greek) and a comparative database of 200 fundamental vocabulary meanings in 87 Indo-European languages to show that the frequency with which these words are used in modern language predicts their rate of replacement over thousands of years of Indo-European language evolution. Across all 200 meanings, frequently used words evolve at slower rates and infrequently used words evolve more rapidly.
Why might this be? One study author explains, in the Times article:
As to how frequency of word use would affect evolution, Dr. Pagel said a possibility is that if errors are made in speaking common words, they may tend to be corrected, precisely because they are so common and so important for communication.
It doesn't just apply to words. See work by Joan Bybee on fossilization and change of grammatical forms as well.
Since rarer words are less well known generally, then someone adopting an existing rare word as their own pet word, after having guessed wrongly at the original meaning, may use it freely with little chance of getting corrected. With their repetition of the word other people may begin adopting it with its new meaning.
This also works for pronunciation. At Jet Propulsion Laboratory, somebody years ago established that the Greek letter phi was said 'fee', and that became the accepted in-house pronunciation. Yet no member of Phi Beta Kappa would accept the practice.
Remember the dialect survey done by Bert Vaux at Harvard a few years ago? The words he used to distinguish between dialects in the U.S. tended to be for objects that are rarely discussed, such as (what I call) roly-polies, crawdaddies, dust bunnies, and rubbernecking.
Well, it seems the relationship between frequency and language change is not quite so intuitively obvious. on Sunday, Oct 14, Sally Thomason at Language Log posted a critique of a similar NYT's article titled 'Fitchifying Language Change' where she claims "regular sound change is indeed blind to frequency and all other nonphonetic contextual factors. So it is nonsense to say that frequent words resist change unless one qualifies the statement to exclude regular sound change."
Her entire post is well worth the read.
"At Jet Propulsion Laboratory, somebody years ago established that the Greek letter phi was said 'fee', and that became the accepted in-house pronunciation"
It also works in Russian-speaking world. At Moscow university (at least at Physics and CS departments) Greek letter omega turns to Omega (with O accented, instead of 'e'). May be it a language/culture-independent? (whats about China?)
Chris, I've read both the papers in Nature and the Language Log commentary, and I think that Thomason's response was a little harsh. Nothing in the papers (which focused on lexical-item replacement, and affix regularization) suggested that phonetic changes would follow the same frequency-based rules. Indeed, the critique was not really on the papers, as on Fitch's commentary on the papers! I think it's mostly a case of people being contrary because they can...