One of my favorite correspondents (thanks, Carol!) has sent me a NY Times story by Nicholas Wade about “a new way of linking languages, which [linguists] say has allowed them to reconstruct a network of the languages spoken in islands near New Guinea.”

The new method is designed for languages so old that little trace of their common vocabulary remains. It forges connections between languages through grammatical features, which change less quickly than words.
With the new tool, historians may be able to peer considerably further back in time than the 5,000 to 7,000 years or so that many linguists see as the limit beyond which no sure connections can be made between languages.
The authors of the new method say the relationships they can construct may be 10,000 years or older.
The researchers, who were led by Michael Dunn, of the Max Planck Institute for Psycholinguistics in Nijmegen, Holland, have published their work in the current issue of Science.

I know nothing about the Papuan languages, so I can’t evaluate the conclusions they come to, but it bothers me that the scientist quoted praising the approach and saying it will be “widely emulated” is a biologist, not a linguist. None of my fellow linguabloggers has discussed the story, so consider this a call for comment: anybody have an informed opinion on whether this is valuable or just another bit of linguistic cold fusion?


  1. Since they didn’t really describe the method, it’s hard to say, but isn’t using grammar something that’s always been doing when trying to link languages, anyway?
    I mean, look at all the possibility to link languages based on areal similarities, and so on. I guess I don’t see exactly what it is that’s “new” here.

  2. As I understand the article, it’s grammatical functions that are being considered cognate, regardless of form. (Though I know nothing about either in Papuan.)
    Exactly: the coincidence of geography and language does not necessarily imply historical linguistic unity. Compare Indo-European, Uralic, and Basque, all sharing the same small space of Europe. While some seek to unite the first two as Indo-Uralic, they do so still based on systematic formal and functional similarities.
    Seems to me that typology and diachrony are totally conflated by that study, as reported. Rather, diachrony is subsumed by typology?
    Now, it’s somehow in my emptied head that Papuan speakers — or is it Australian Aboriginal speakers — consciously change lexica and frequently. Is this what Dunn et al. are going for? Finding the patterns of replacement and extrapolating from there? Which I might find not so objectionable.
    Claire Bowern would know.
    (Looks like Michael Dunn is a synchronicist-syntactician, based on his bookcase.)

  3. David Costa says:

    With the caveat that I too know almost nothing about Papuan languages, I agree, this article seems a bit off. First, trying to link languages on the basis of grammatical features is hardly a new idea, and linguists have long known the risks involved with that; second, I ‘m rather troubled by the idea of a biologist telling the world about some ‘new’ linguistic technique that will soon be ‘widely emulated’.
    I’m also not sure what the point was of quoting Merritt Ruhlen, but all MSM articles on historical linguistics seem to do that…
    I find it a bit fishy that they published this in Science. Generally when a linguistic article gets published in a popular science journal like that, it’s often the case that the scholarship wasn’t rigorous enough to get it into a regular linguistic journal.

  4. I’m amazed they can write about this topic and not mention Johanna Nichols’s groundbreaking
    Linguistic Diversity in Space and Time
    By the way, one of her major findings was that, typologically, the typical New Guinea language looks very different from the typical Australian language, and even that the two regions are more similar to the Americas than they are to each other, contrary to the William Foley quote in the middle of this article.

  5. Well, I was trying to control my automatic suspicions about this kind of “breakthrough,” but it’s beginning to sound like I needn’t have. Waiting to hear from Claire, though…

  6. Wow, I’m famous ;). Sorry, today’s been hectic. Let me also start with the caveat that I don’t know much about Papuan languages either.
    Re Angelo’s comment, it’s been said that necronym taboos could contribute to rapid lexical change, but it hasn’t really been demonstrated. E.g. Alpher and Nash (1999) in the Australian Journal of Linguistics didn’t find huge lexical replacement in Guugu Yimidhirr between Cook’s wordlist in 1770 and recent fieldwork. There’s also some work by Alan Dench on correspondence mimicry, where speakers are aware of the correspondences between their language and neighbouring languages, and produce ‘expected’ forms on that basis.
    … I’ve now read the NYT article and the Science article. This is going to be a veyr long comment, so let me write it in a blog entry and I’ll link over here when it’s done.

  7. The same issue of Science has a related article: Pushing the Time Barrier in the Quest for Language Roots, by Russell Gray.
    Here’s the main piece: Structural Phylogenetics and the Reconstruction of Ancient Language History.
    Sorry, I don’t think these links will work for everyone.

  8. me again. Just to mention that Linguistic Diversity in Space and Time is the first reference in the Science article.

  9. right, I wrote up something here:
    or click on my name to be taken to
    (Angelo, how did you add a link in the comment? I used the html code and the preview stripped it)

  10. Here‘s the direct link. I don’t understand the code-stripping thing; I recall somebody else used to have a problem with it, but most of us seem to be able to use a href= with no problems.

  11. Thanks, Claire! I should have given it deeper thought, though I confess never having been taught advanced historical-linguistic theory.
    (Huh. I have no idea why I was able to post links.)

  12. I tried <a href=URL> Link title </a> and got highlighted text but no link. Or do you just put in href= and nothing else?

  13. Your problem is that you omitted the quotation marks; it has to be “URL” (I went into your earlier comment and added them, so the link works now). I don’t know why Claire’s didn’t work.

  14. Andrew Dunbar says:

    Now, it’s somehow in my emptied head that Papuan speakers — or is it Australian Aboriginal speakers — consciously change lexica and frequently
    I also know little about either Australian or Papuan languages, but I have done a little reading over the years.

    1. Many Australian languages do indeed replace words due to a taboo on the names of the recently deceased and similar sounding words. “Recently” may mean a much longer time that you’d expect.
    2. I have no idea if this also occurs in Papuan languages.
    3. Australian and Papuan languages are utterly unrelated and utterly dissimilar.
    4. Even within the Australian and (I believe but am not sure) Papuan languages, few family relationships have been proven. Australian includes the Pama-Nyungan family. The rest are labelled Non-Pama-Nyungan but are not a family.
    5. In the Torres Strait Islands between Australia and Papua New Guinea there are two indigenous languages. Kala Lagaw Ya spoken in the west is an Australian Pama-Nyungan language, Meriam Mir spoken in the east, is Australia’s only Papuan indigenous language.
  15. Re Andrew’s points 1 and 2, necronym tabooing lasts for different times in different areas. I believe it’s longest in the Central/Western Desert and a few communities in Arnhem Land, where it lasts years. However, it’s more common for the community-wide taboo to be much shorter (6 months, maybe?) and the taboo to last longer in front of immediate family. For example, when I was at Milingimbi in June 2004 someone called “Margaret” died, and one of my consultants, who shared that name, was known as Seraphina for the rest of my stay. When I went back this year she was being called both Seraphina and Margaret by different people, with Margaret being more common.
    PNG doesn’t have wide-spread necronym taboos, as far as I know.

  16. Heh, i just attended a talk by Gray, and he uses a lot of techniques common to biology. Unfortunately i dont know enough about the techniques or linguistics to give an informed comment, but the talk was pretty interesting.

Speak Your Mind