Rate of Change

[Larry Trask]

--On Wednesday, June 6, 2001 1:06 am +0000 X99Lynx at aol.com wrote:

Folks, I'm buried in exam marking, and I don't have time now to respond to
all of Steve Long's points.  For now, I'll have to content myself with
replying to just one, on the comparative method.

> To say that I can't back up to that starting point, where everything is at
> best similarities and resemblances -- before any judgments about genetic
> relationships would have been made -- is to queer the game.  It makes the
> comparative method immune to critical analysis.

No; certainly not.  The comparative method is not immune to critical
analysis.  If it were, nobody would have any confidence in it.  Any
attempted application of the method must stand or fall on the quality of
the evidence offered to support it.  If that evidence is not good enough,
then the claimed conclusions will be rejected.

> I suspect part of the problem begins back when no analysis has yet been
> done and there can be nothing but resemblances between two languages.
> That's where the presumption starts that there WILL be only one genetic
> relationship and therefore there should be only one original systematic
> correspondence between the languages.  If you start by presuming that all
> "genetic" relationships can only relate to one parent, than all your
> results will tend towards that presumption, whether true or not.

Not so, I'm afraid.  The comparative method cannot possibly reconstruct an
ancestor that never existed -- not if it's competently applied, anyway.
Incompetent attempts are prominent in the fringe literature, of course, but
these are simply laughed away by professional linguists.

> (The Australian example you give -- Blake and "the clear patterns
> underlying the superficial absence of resemblances,..." could only happen
> in a historical and geographical context where relationships were already
> expected, for extrinsic reasons.

Again I disagree.  Of course, it made more sense for Blake to try to spot
patterns linking the troublesome Australian languages to other Australian
languages, rather than to, say, Eskimo-Aleut.  But he would not have found
those patterns if they had not already been there, waiting to be discovered.
He didn't create patterns that didn't already exist.  The comparative
method can't do that.

> I don't imagine you would have much
> patience with anyone establishing a close genetic relationship between
> Mexican and Phoenician despite "the superficial absence of resemblances"
> between the two languages.)

Depends on the evidence.  If "Mexican", whatever that is, and Phoenician
really do exhibit systematic correspondences, then those correspondences
are sitting there waiting to be discovered.  If somebody discovers them,
then that is the end of the matter.

I get the distinct impression that Steve believes that the comparative
method is a sham, a toy that does no more than to spit back the assumptions
built into it in the first place.  'Tain't so, Steve.

Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larryt at cogs.susx.ac.uk

Tel: (01273)-678693 (from UK); +44-1273-678693 (from abroad)
Fax: (01273)-671320 (from UK); +44-1273-671320 (from abroad)