Aside from not knowing what the gene variants actually do, no one knows how precise the model Lahn used to date them is, Collins added.
Lahn's own calculations acknowledge that the microcephalin variant could have arisen anywhere from 14,000 to 60,000 years ago, and that the uncertainty about the ASPM variant ranged from 500 to 14,000 years ago.
That does not mean one population is smarter than another, Lahn and other scientists stressed, noting that numerous other genes are key to brain development.
"There's just no correlation," said Duke's Wray, calling education and other environmental factors more important for intelligence than DNA anyway. The information contained in the AP News report may not be published, broadcast, rewritten or redistributed without the prior written authority of The Associated Press.
Those criticisms are particularly important, Collins said, because Lahn's testing did find geographic differences in populations harboring the gene variants today.
They were less common in sub-Saharan African populations, for example.
"No matter how we [changed] the analysis or assumptions, we couldn't get a date of around 6,000 years," says Gray.
The resulting tree matches many existing ideas about language development.
"A different way to look at is it's almost impossible for evolution not to happen." Still, the findings also are controversial, because it's far from clear what effect the genetic changes had or if they arose when Lahn's "molecular clock" suggests — at roughly the same time period as some cultural achievements, including written language and the development of cities.
Lahn and colleagues examined two genes, named microcephalin and ASPM, that are connected to brain size.
Around this time, farming techniques began to spread out of Anatolia - now Turkey - across Europe and Asia, archaeological evidence shows.
The farmers themselves may have moved, or natives may have adopted words along with agricultural technology.