The article excerpt you posted discusses two separate but related things. First:
NICHOLAS WADE Once upon a time, there were very few human languages and perhaps only one, and if so, all of the 6,000 or so languages spoken round the world today must be descended from it.
If that family tree of human language could be reconstructed and its branching points dated, a wonderful new window would be opened onto the human past. Yet according to historical linguists, the chances of drawing up such a tree are virtually nil and those who suppose otherwise are chasing a tiresome delusion. Languages change so fast, say linguists, that their genealogies can be traced back only a few thousand years at best before the signal dissolves completely into noise. The linguists? problem has recently attracted biologists who have developed sophisticated mathematical tools for drawing up family trees of genes and species. Because the same problems crop up in both gene trees and language trees, they are confident that their tools will work with languages, too
Oh I am so glad to read this! The press has tended to report the most extravagant claims, like the flurry of attention during 1990-1992 about the Proto-World hypothesis and similar theories reducing the global extent of language into just a few macrofamilies, and reconstructing actual forms for such flimsy entities as Proto-Eurasitic (I'm not sure if I got the name right, it was Greenberg's answer to Proto-Nostratic), Proto-Amerind, and even Proto-World . There is a real time-depth problem in such deep reconstructions, when the role of chance plays a greater and greater factor, when there are so many parameters one could play fast and loose with. I'm willing to allow a little bit of credence to Proto-Nostratic, but the anything over 9,000 BC is very dubious to me. I think of it like carbon dating, when the method becomes very unreliable when you reach a certain threshold.
Plus, there is a simple problem with the evidence. Languages die out. The languages in existence today are the descendents of only a small fraction of languages that existed 10,000 years ago. Whole language families that once existed have died out. Look at how many of the languages attested only 5,000-3,000 years ago like Sumerian, Hurrian, Etruscan, Aquitanian (e.g. Old Basque), Elamite are unrelated to any known living language. So we are without critical evidence of what the linguistic situation was like 10,000 years ago -- much less 100,000 years ago, or even 200,000 years ago.
Indo-European family
This is different because of the Indo-European family has a shallower time depth and is universally accepted. I did not read the recent article that dated Proto-Indo-European much earlier, and I am a little surprised by the dating since Proto-Afro-Asiatic is probably best dated to 8,000-9,000 BC. (Proto-Nostratic, while is quite dubious to me, would fall even earlier, like 14,000 BC). There are other datums other than the "wheel" example. The domestication of dogs and horses are another: Proto-Indo-European *wlkwo "wolf" is the source of *kwon "dog" (whence English hound and Latin canis, and Sanskrit svan), and the latter is the probable source of *ekwo "horse" (whence Latin equus and Sanskrit asva). This would point to a later date for Proto-Indo-European, as the horse was domesticated around 5,000-4,000 BC. Domestication is suggested because the word for "horse" is likely derived from the name of another domesticated animal. However this can be disputed by treating the -kwo- element as a morpheme simply meaning "animal". It is also important to recognize that even in reconstructed PIE, there are already dialect divisions evident -- such as the famous satem-centum division as well as other more finer distinctions. And then there is the loss of the pharyngeals which occurred in most languages except Hittite. So Proto-Indo-European should be treated as a language like English that existed for several thousand years, but went through several stages in its own history.