Principled Comparative Method - a new tool

Jon Patrick jonpat at
Mon Aug 23 23:19:38 UTC 1999

    Date:       Sat, 21 Aug 1999 08:35:46 +0300 (EETDST)
    From:       Robert Whiting <whiting at>

    On Wed, 18 Aug 1999 X99Lynx at wrote:

I agree pretty well with everything Robert White said on appraising my method,
and I missed the email he is commenting on, so here I will just add a few


    >jonpat at wrote further:

    ><<Our data has to be the word set in the parent form
    (reconstructed words or real words) and then one word set for the
    each daughter language and the set of phonological transformation
    rules between each parent and daughter for each word in their
    chronological sequence.>>

    >I'm wondering if there isn't a possible flaw here in using <<the
    >parent form (reconstructed words...)>>.

    >Reconstructed words have already made assumptions about the
    >relationship between the parent and the daughter languages.  In
    >fact they are nothing but a presumed relationship between the
    >daughter languages.

    Steve is quite correct here.  A reconstruction is just that.  In
    the best case it is only the statistically most probable
    original relationship between the forms found in the daughter

In our case study we were very happy to use the Cantonese and Beijing
materials because the mother language is Middle Chinese which is NOT a
reconstructed language, although there is some small degree of uncertainty in
the source materials. Hence this data formed a check on whether or not our
method could produce sensible answers. The answer is 'yes', because it
produced a result that was consistent with the perceptions of the scholars in
the field, however, it also produced extra illuminations of the relationships
between the languages. A book is to be published in the next week or two
showing all the results. When I have the correct bibliographical info I will
send it to the list. The linguist who worked with us was John Newman(it was
his data).
In the case of reconstructed parent languages our method would be useful to
evaluate competing reconstructions. With only a single reconstruction from two
daughter languages the answers may provide an nice generalisation story about
the proximity of the daughters but the value of that is in hands of the

One important aspect of our method is the rigour it forces on the linguist to
explain EVERY piece of the data. Generally I find dealing with linguistic data
difficult because firstly, it is almost impossible to find a significantly
sized set of data in the one place and secondly, the data is invariable
incomplete in that the every rule of transformation for each word is present
with the data set. I actually think that linguists are very poor data managers
and on that dimension are very - how shall I say it - unrigorous.

I would have thought it is time for the historical linguistics community to
follow the example of the genome people and set up international databases of
historical languages, reconstructions, etc using agreed data organisations, so
that we can all get the data and agree on what data we are working on. I am
trying to initiate such a project for basque.

I f there other specific items from Bob's message you want me to comment on
please let me know
Prof. Jon Patrick			BH +61-2-9351 3524
Sybase Chair of Information Systems	FX +61-2-9351 3838
Basser Dept. of Computer Science
University of Sydney
Sydney, 2006	
Australia	    WEB:

More information about the Indo-european mailing list