November 07, 2003

Gall in the family

It's depressing that Greg Ross, the managing editor of the generally excellent American Scientist Online, has written such a badly-informed and credulous review of Peter Forster and Alfred Toth, 'Toward a phylogenetic chronology of ancient Gaulish, Celtic, and Indo-European'. PNAS (2003).

For a better appraisal, see Larry Trask's Linguist List review, Peter Forster's reply, and Trask's re-reply.

The American Scientist review starts out badly:

Ever since Darwin proposed an evolutionary tree to describe the descent of species, linguists have sought to apply the concept in their own field. ... Now historical linguists may stand to benefit by borrowing a second idea from evolutionary biology.

This gets the direction of intellectual influence exactly backwards. The well-known fact of the matter is that Darwin modeled his idea of "descent with modification" in biological evolution explicitly on what he took to be the obvious prior success of philologists in establishing "descent with modification" as the basis of the the history of languages.

"It may be worth while to illustrate this view of classification, by taking the case of languages. If we possessed a perfect pedigree of mankind, a genealogical arrangement of the races of man would afford the best classification of the various languages now spoken throughout the world; and if all extinct languages, and all intermediate and slowly changing dialects, had to be included, such an arrangement would, I think, be the only possible one. Yet it might be that some very ancient language had altered little, and had given rise to few new languages, whilst others (owing to the spreading and subsequent isolation and states of civilisation of the several races, descended from a common race) had altered much, and had given rise to many new languages and dialects. The various degrees of difference in the languages from the same stock, would have to be expressed by groups subordinate to groups; but the proper or even only possible arrangement would still be genealogical; and this would be strictly natural, as it would connect together all languages, extinct and modern, by the closest affinities, and would give the filiation and origin of each tongue."
("Origin of Species", 1st Edition, Chap. 13 Mutual Affinities of Organic Beings).

Indeed, this idea was already well understood by Thomas Jefferson almost a century earlier. See this chapter by Lyle Campbell for other, even earlier roots of the idea that languages form a "family tree" (scroll down to section 4, "The Rise of the Comparative Method").

I'm also fairly certain that the lexicostatisticians used algorithmic phylogenetic-tree-inducing techniques (Ross' "second idea") for language history before any such techniques were ever employed in biology. They certainly did so many decades before Forster and Toth came on the scene.

It's not fair to blame Ross for being ignorant of the past and present state of historical linguistics, but he could have asked someone with some linguistic credentials. If a couple of computational linguists wrote an article about applying language-modeling techiques to determining the structure of macromolecules, I'd expect Ross to consult with specialists in that area before deciding whether or not to take the authors at their word (in this case I believe that he'd discover that their word is good). When a couple of geneticists take a flying leap at Indoeuropean, I'd expect Ross to consult with a historical linguist or two rather than writing a puff piece based entirely on the article and an interview with its authors.

I'm not going to criticize Ross any further, or rehearse the problems with the Forster & Toth article in detail here -- but read Trask, read Ross, read Darwin, read Jefferson, and weep.

Our field needs to fire its public relations consultants and... What? We don't have any?

Posted by Mark Liberman at November 7, 2003 06:37 PM