LINGUIST List 9.762

Thu May 21 1998

Disc: Time Depth

Editor for this issue: Julie Wilson <julielinguistlist.org>

Directory

1. manaster, Re: 9.757, Disc: Limits on Knowledge in Linguistics

Message 1: Re: 9.757, Disc: Limits on Knowledge in Linguistics

"Fidelholtz James L." <jfidelsiu.buap.mx> writes (inter alia):
> 
> 	I have published in a couple of articles the claim that
> vocabulary considerations limit us in principle to more or less 20,000
> years in time depth. The calculations are simple. Glottochronology
> supposes (on the basis of test cases like Vulgar Latin and others) a
> rate of change around 17 or 18% loss per millennium. A simple series
> of multiplications (.83 to the twentieth power) will give a residue
> well under 4% (2.4 to 1.8%) after 20 millennia. Others (I can't at
> the moment remember who) have calculated, using some fairly
> noncontroversial assumptions about the phonetic structure of languages
> (maybe it was Greenberg), that any two random unrelated languages are
> likely to score about 4% on the Swadesh test just by pure accident,
> with that then being the lower limit for showing a relationship other
> than chance. 

The point is that Swadesh himself did these calculations in 1960!!
Yet a decade later Bender did the same thing and I believe Bynon
has also tried to. But this is one my chief problems. Why
not have a whole string of articles reinventing transformational
grammar and ignoring Chomsky or reinventing generative phonology
and ignoring Halle? Surely there is a problem here. 

Of course, there still remains the good old
> comparative method, which, depending on what the residue is, might
> just be able to provide some evidence, and syntactic or morphological
> evidence, 

I am very happy to see this acknowledgment. Swadesh understood this
and my caouthors and I keep talking about this, but the urban myth
fails to mention this crucial distinction. In fact the urban
myth of authors liek Nichols and Hock insists that the limit
Jim refers to is precisely a limit on the comparative methods
in general. (I insist on the plural, because there is no one
method. The constant references to THE COMPARATIVE METHOD are
yet another urban myth).

Finally, thre are three points Jim ignores but which are crucial.
(These points were ignored by Swadesh and others too, of course!
although Greenberg has seen some of them).

One, if glottochronology is wrong (even only sometimes wrong)
that is if some languages evolve less fast than assumed, then
the calculations he cites lose all meaning. Yet this is a 
rather well-established fact I would have thought, as in the
case of Icelandic whose rate of vocabulary loss per millennium
is 1% or less on teh Swadesh list.

Two, in establishing language families, we do not look at
just two languages at a time nor do we ignore subgroups.
For example, IE is not arrived at by comparing Polish and
Sinhala, and the mathematical model Jim assumes (as did
Swadesh, Bender, and Bynon) ignores this. If we consider
more than 2 languages, the "limit" will be different.
Moreover, if we consider subgroups, this is what happens:
to show the relatedness of Polish and Sinhala, we actually
first relate Polish to the rest of Balto-Slavic
and of SInhala to the rest of Indo-Iranian
and so on, we then (assuming we had no acient texts in
Sanskrit or Old Church Slavic) reconstruct Proto-Slavic
and Proto-Indo-Iranina and then compare these to each
other. What now happens is of course that the loss of any given
vocabulary in any one language is compenstated in many
cases by its retention in another of the same branch.
For example, Sinhala is easy to relate to Hindi and
other Indo-Aryan languages and they in turn are easy to
relate to Iranian, and all of II is easy to relate to
Balto-Slavic because the loss of IE words for 'head',
'leg', and most or all kinship terms in Sinhala for example
does not matter if they are retained in Persian or some
other II language. Even without reconstruction per se
scholars like Strahlenberg and others since might be
able to recognize the IE relationship by reasoning about
such a chain of relationships.

Three, all teh calculations are based on a particular
100-word list. Changing or more to the point expanding
the list can make a difference. 

AMR