21.302 quantifying evolutionary dynamics of language

               Humanist Discussion Group, Vol. 21, No. 302.
       Centre for Computing in the Humanities, King's College London
  www.kcl.ac.uk/schools/humanities/cch/research/publications/humanist.html
                        www.princeton.edu/humanist/
                     Submit to: humanist_at_princeton.edu

         Date: Wed, 17 Oct 2007 06:35:36 +0100
         From: "David L. Hoover" <david.hoover_at_nyu.edu>
         Subject: Re: 21.299 quantifying evolutionary dynamics of language

Having just read the article, I'd second Norman's concerns. Two
simple examples must suffice:

1. The authors use the dates 800, 1200, and 2000 as the dates at
which they check for the regular vs irregular status of a verb. The
earliest written records of Old English are usually dated to about
700, but clearly OE existed before that. The larger problem is with
the dates for Middle English and Modern English. It makes sense to
use 1200 as the beginning date for Middle English, but Modern English
is normally considered to begin around 1500. So their calculations
are based on a late beginning for Old, an early beginning for Middle,
and a 500 year late date for Modern English. Equations for rates of
change based on these figures are, to say the least, debatable, and
one would need further explanation of the choices.

2. The only date at which actually sampling for frequency is done is
for Modern English (from COBUILD, roughly 1980-present). The authors
merely looked for an entry in a reference book that indicated whether
a verb was regular or irregular in Old, Middle, and Modern English
(800, 1200, 2000). Thus the equations apparently assume that each
verb's frequency remained constant over the 1200 years they study. It
is obviously reasonable to suggest that infrequent verbs are less
resistant to regularization (they occur so seldom that it is easier
for children to use regular forms analogically without finding out
that the verbs are historically irregular). But one might have liked
a check of the relative frequencies of some verbs in Old English or
Middle English to see whether there is any close relationship between
the frequencies in OE and Modern English, for example.

The controversy may prove instructive, at any rate.
DLH

Humanist Discussion Group (by way of Willard McCarty
<willard.mccarty_at_kcl.ac.uk>) wrote:
> Humanist Discussion Group, Vol. 21, No. 299.
> Centre for Computing in the Humanities, King's College London
> www.kcl.ac.uk/schools/humanities/cch/research/publications/humanist.html
> www.princeton.edu/humanist/
> Submit to: humanist_at_princeton.edu
>
>
>
> Date: Sat, 13 Oct 2007 07:36:43 +0100
> From: "hinton_at_springnet1.com" <hinton_at_springnet1.com>
> >
>
> >Those here interested in the application of
> >statistical techniques to study language will be
> >interested in work done by Erez Lieberman and
> >colleagues at Harvard and MIT, an account of
> >which has recently been published in Nature 449
> >for 11 October, "Quantifying the evolutionary
> >dynamics of language", pp. 713-16,
>
>
>This has been getting a lot of negative criticism in the list HEL-L,
>the History of English discussion list.
>
>It would take a lot of space and time to repeat all the problems we
>have indicated with what we can make of the study from the little
>blurb, but a great deal of dissatisfaction has been expressed about
>the authors' lack of knowledge of the history of English (for
>instance, they could easily have found out Chaucer's dates),. about
>the fact that the relative rates of change in the past participles of
>common strong verbs versus uncommon ones has been known to historical
>linguists for at least a century, the dearth of valuable diachronic
>databases of English from which a careful study could be made, and
>even some of the authors' statistical remarks.
>
>In other words, the thing is getting a bad press among many
>historical linguists. (And many of us feel that Nature would do well
>not to accept papers in fields the editors and readers do not understand.)

-- 
       David L. Hoover, Professor of English & Webmaster
              NYU English Department, 212-998-8832
              http://english.fas.nyu.edu/page/home
Among all forms of mistake, prophecy is the most gratuitous.
                 George Eliot, Middlemarch