LINGUIST List 14.2047

Wed Jul 30 2003

Disc: Re 'Celtic Found to Have Ancient Roots'

Editor for this issue: Karen Milligan <karenlinguistlist.org>

Directory

1. Richard Ingham, Re: 14.2034, Disc: Re 'Celtic Found to Have Ancient Roots'

Message 1: Re: 14.2034, Disc: Re 'Celtic Found to Have Ancient Roots'

The ongoing LINGUIST LIST debate over the Forster & Toth article has
been fascinating, and has prompted me to look at the methods that the
authors say they used, which seem to be set out in the website:

http://www.mcdonald.cam.ac.uk/genetics/gaulish_tutorial.pdf

In particular, the passage at the top of p. 2 discusses how the
authors coded texts such as the beginning of the book of Genesis.

Unfortunately, I don't see an explanation of how to replicate this
procedure on other such texts. Even the criteria for initial coding
are not laid out clearly. For example, and to look no further than the
first word of the text, the linear counterpart of the word 'in' (as in
English 'in the beginning') is coded as B for German 'am' (as in 'am
Anfang'), but C for French 'au' (as in 'au commencement').

I don't see any principled account of why the coding here couldn't
just as well have been the other way round, or why both couldn't be
coded as B. The authors state that their coding considers only
lexemes, not grammatical morphology, so the morphological structure of
these items can't have been criterial. They say that linguistic coding
can be hard, which is the bit we did know, but they don't say much
about what they did to address the problem, which is the bit we'd all
surely like to know. It seems to me to be fair comment for a reviewer
to observe that there is a lack of specificity here, which is not only
evident in the article, but also in the supporting materials published
on the web.

I am certainly intrigued by the claims made by the researchers, but
I'm afraid I personally remain puzzled about how they obtained their
results.

 Richard Ingham

 Dr R.P.P. Ingham
 The University of Reading