Okay. I've now read in full the article by Professor Pagel et al. and discussed the BBC article with him by e-mail. While I still can't help but laugh at the William the Conqueror quote, I hereby assign most of the blame for the kerfuffle to a combination of crossed wires and bad reporting. The article, yes, had something to do with the English language, but not an awful lot. It had more to do with the descent of certain words from proto-Indo-European into twenty-seven (yes, that's right, 27) different Indo-European languages.
See, here's what I think the problem is. If I were to say to a reporter that the word "heart" is 10,000 years old, that means different things to different people. Having studied the history of the English language (and of others), I can tell you that "heart" is pretty much the same word as "cardio" because of a series of sound changes that turned ancestral words with a "k" sound into more modern words (of Germanic descent) with an "h" sound. We can say with a pretty good degree of certainty that the proto-Indo-European word for "heart" was probably some kind of inflected form of a word that may have sounded an awful lot like "kart" or "card."
But if you tell someone who doesn't study this stuff that the word "heart" is 10,000 years old, chances are they'll think you could take a time machine back and use the word in its present form to talk to "cavemen."*
So when Professor Pagel says that the word "I" is a very old word indeed, it means that some cognate of the word "I" (maybe "ego" or "ich" -- like "cardio" for "heart") was probably spoken 10,000 years ago. It might have sounded less like the modern word "I" than like "itch" or "eggo," but it would be the same word.
This is what the reasearch is getting at.
The most interesting finding in the article, so far as I can tell, is that of the counterintuitive stability of the most repeated words. One might think rather that, given the frequency with which they're spoken (upwards of 35,000 times per million words, as opposed to most words, which are used fewer than 100 times per million words), these words would change most quickly. But instead it seems that their repetition, simplicity and lack of inflection protect them quite well as the language evolves.
As for what the article says about English, it does reiterate the findings of an earlier study that says that the most frequently used modern English words are more likely to be Old English in origin (as opposed to Anglo-Norman French, or Norse, or Latin), but as for the article itself, it tends to stick to the abstract and focus on the descent of cognate words from truly ancient languages to our own.
Perhaps Proffessor Pagel saw the glazed-over look of confusion in the reporter's eye, and tried to dumb things down for the sake of a simple explanation, like the way simple.wikipedia.org tells you that atoms are "very, very small" -- but I can't help but think that no matter what actually happened, one thing is certain: the BBC reporter didn't understand what was going on, and now, neither will most of his or her readers.
Oh and if you ever do get a time machine and go back to talk to our ancestors, remember two things: One, if you're going back to talk to William the Conqueror (nee "the Bastard") take an Old French dictionary, not an Old English one, and two, if you're going back 10,000 years, just remember that you're about as likely to be understood saying "I, Tarzan" as you would be saying "Leggo my Eggo."
*No, not primitive nomadic people pretty much identical to ourselves. Cavemen. Like in the Geico ads.