Evolution of language study savaged in Science: “suboptimal data, biased methodology, and unjustified assumptions.”
|February 16, 2012||Posted by News under language, News|
From “Out of Africa? Data Fail to Support Language Origin in Africa” (ScienceDaily, Feb. 15, 2012), we learn
In the beginning was the word — yes, but where exactly? Last year, Quentin Atkinson, a cultural anthropologist at Auckland University in New Zealand, proposed that the cradle of language could be localized in the southwest of Africa. The report, which appeared in Science, was seized upon by the media and caused something of a sensation.
Sure it did. The story had everything: The cave man, the exotic location, the opportunity to poach the Bible (“In the beginning was the word” gives the basic idea … ), tell stories, and draw pictures … Here’s a report that makes the agenda pretty clear as well:
“Our work shows that the claims some linguists have made for a really strong role of the innate structure of the human mind in shaping linguistic variation have been hugely oversold,” he says.
Now however, LMU linguist Michael Cysouw has published a commentary in Science which argues that this neat “Out-of-Africa” hypothesis for the origin of language is not adequately supported by the data presented. The search for the site of origin of language remains very much alive.
Atkinson postulated that the number of phonemes (basic sound unit in words, like consonants, and vowels) proportionately decreased in languages as they were spoken by people further from this southwest African “hot spot.” He compared it to the “founder effect” in genetics:
Biologists have observed an analogous effect, insofar as human genetic diversity is found to decrease with distance from Africa, where our species originated. This is attributed to the so-called founder effect. As people migrated from the continent and small groups continued to disperse, each inevitably came to represent an ever-shrinking fraction of the total genetic diversity present in the African population as a whole.
However, in Science, Cysouw threw cold water on the idea, noting that
if Atkinson’s method is employed to examine other aspects of language, such as the construction of subordinate clauses or the use of the passive mood, the results “do not point in the same direction.”
and could well point to eastern Africa or the Caucasus instead.
… linguists have long sought to throw light on the origin of language by analyzing patterns of language distribution. The problem is that such relationships can be reliably traced only as far back as about 10,000 years before the present.
The basic problem is that the analogy from language to biology is flawed. Biology runs on the many life or death rules of biochemistry; language doesn’t operate under such severe constraints. The capacity for creating language, once established, is constrained only by a few fundamental rules. Thus, change can be relatively easy and histories can be hard to postdict.
For that matter, the innate structure of the human mind has not been “hugely oversold”; language change can be challenging to explain apart from cultural change. Consider political correctness in our own day: How did the term “housewife” go from being a conventional statement of occupation on a census form to a term of abuse? That is not accounted for by the genealogy of the word but by a unique history, one that may be lost one day.
Here’s the article.
Abstract: We show that Atkinson’s (Reports, 15 April 2011, p. 346) intriguing proposal—that global linguistic diversity supports a single language origin in Africa—is an artifact of using suboptimal data, biased methodology, and unjustified assumptions. We criticize his approach using more suitable data, and we additionally provide new results suggesting a more complex scenario for the emergence of global linguistic diversity.
See also: Tom Bethell on Noam Chomsky’s dissent from Darwin