Phonetic Clues Hint Language Is Africa-Born

Interesting stuff...
http://www.nytimes.com/2011/04/15/science/15language.html?hp

Anyone know more about this?

Also discussed here: http://www.librarything.com/topic/114241

http://www.cslu.ogi.edu/~sproatr/newindex/atkinson.html

>1 timspalding:
Don't know but wish I did. Am passing on the link to some phonology colleagues.

Seems to me like a classic example of the devil being in the details, and a bitch to verify.

The theory sounds reasonable to me, but I bet some expert could point to some languages that don't fit.

A scholarly critique was posted on this thread: http://www.librarything.com/topic/114241.

Mark Liberman posted about it today over at Language Log: http://languagelog.ldc.upenn.edu/nll/?p=3090

8> I love it when a scholarly treatment quotes Dave Berry.

A Dave Barry reference? Says something about how "scholarly" it's likely to be. I don't need to follow the link. A: the topic alone warns me off. B: newspaper accounts are usually far from "scholarly". C: I've seen this mauled by actual scholars elsewhere.
Given all that, yes, since early ancesters of ours seem to have originated there, the origins of human speech likely did, too. This is NOT something actual evidence of is going to turn up short of the invention of a time machine. It's not like tracing DNA around the world.

10> You're missing out, then. He's doing a scholarly mauling of the mathematical bases for the notion, with an appropriate joke, drawn from Dave Barry, to close (not as part of the real discussion). I found it rather more interesting than the original proposal...but then, I'm a math and probability geek as much as a language one.

10> Mark Liberman is a well-known academic. He's Trustee Professor of Phonetics in the Department of Linguistics, and a professor in the Department of Computer and Information Sciences at the University of Pennsylvania.

I think I'd be prepared to go out on a limb and dare call him an "actual scholar".

What the others said. Language Log isn't a newspaper source, it's an academic blog.

I don't buy the methodology of this. I've commented here http://blogs.discovermagazine.com/gnxp/2011/04/the-african-ur-language/#more-108... (comment no.29). There isn't any absolute number of phonemes in a language - it's a matter of analysis, with many judgement calls. The number will be quite parsimonious in languages that have already been analysed for the purposes of creating an alphabet. I don't know anything about Polynesian languages but if there really are so few phonemes, something else (syllable structure?) must be doing a lot of work instead.

Just because there's judgment calls about the number of phonemes in a language, doesn't mean the concept of counting the number of phonemes in a language is a meaningless one. One might look at Hawaiian, and come up with 9 consonants or 10, instead of the standard 8, but never 22.

It's also significant that the counting wasn't done by the author.

16> But the observation that languages that are studied more (which tend to be the larger ones) are likely to have more identified phonemes is also interesting.

#17: But that's not true; as this study points out, African languages tend to have the most phonemes, not the Indo-European languages.