Discussion Details
| Title: | Tones and Genes: Scientific American |
| Submitter: | Bob Ladd |
| Description: | I refer to Roger Blench's comment on the paper that Dan Dediu and I (not
''Ladd et al.'' as Blench says) have just published in PNAS, on a possible link between population genetics and language typology. In reply to Blench's specific question (''Who referees articles for PNAS?''), I know the identity of one of our referees but I don't think it would be appropriate to reveal their name; I don't know the other two. The PNAS editor in charge of our submission was Henry Harpending. It is probably true that the genetic and statistical aspects of our paper got a closer look than the linguistic aspects. However, we can assure Blench that we are well aware of the difficulty of classifying typological traits into either-or boxes. We do not feel that such broad-brush classification is necessarily always inappropriate or misleading; for example, for similar global-overview purposes in the World Atlas of Linguistic Structures, Ian Maddieson classified tone into three boxes (no tone/simple tone/complex tone). We used the two-way classification for a very specific statistical reason, and in the paper we discuss the simplifying assumptions that this classification involves. We are also at pains to emphasize that we are not suggesting any sort of deterministic relationship between population genetics and linguistic typology. What we propose is an indirect link between individual genetic makeup and the slow process of language change: we hypothesize that genetic differences could lead to ''cognitive biases'' of one sort or another (which we freely concede are ill-defined and hypothetical, though plausible, as discussed at greater length in the paper), and that these biases could influence language acquisition or language processing in a way that might influence the direction of language change. Such an influence, if it exists, would coexist (and interact) with all the other factors that we know are involved in influencing language change - contact, conquest, and so on. Pace Blench, our hypothesis is in no way inconsistent with the fact (of which we are well aware) that ''highly tonal languages can be closely related to those with no tones.'' We have demonstrated a correlation, and have proposed a hypothetical mechanism to account for it. We hope shortly to do experimental work looking for the basis of the hypothesized ''cognitive bias''. We welcome critical comment on our work, ideally based on a reasonably careful reading of the actual paper rather than the Scientific American summary. More information is available on our web page, which aims to clarify the goals, methods and conclusions of the paper for those who have only seen secondary press reports: please go to http://www.ling.ed.ac.uk/~s0340638/tonegenes/tonegenessummary.html |
| Date Posted: | 04-Jun-2007 |
| Linguistic Field(s): |
General Linguistics
Historical Linguistics |
| LL Issue: | 18.1697 |
| Posted: | 04-Jun-2007 |
