Sino-Tibetan again (was: Re: Alexis on Wald ...)

[MARK ROBERT HALE]

----------------------------Original message----------------------------
 
Can I just ditto this -- as someone who simply took for granted
for many years that the estimates thrown around for Indo-European
and Austronesian (the two families I know best) actually had some
basis, I realized, upon investigation, that there is simply NO
empirical foundation to any of these 'time-depth' guesses. Why
experts continue to off-handedly cite these numbers (particularly
given their utter irrelevance to the linguistic issues involved)
is a total mystery to me.
 
I do believe that the normal 'loss of information' (coupled with
the difficulty of discerning borrowing and other influences in
the very distant past) means there probably is a time-depth beyond
which relatedness cannot be demonstrated. But to believe that we
have any idea what that time depth is is fantasy. And to believe that,
even if we -- through divine inspiration, e.g. (which seems to be the only
possible source for such knowledge at this point)  -- knew what
that depth was, we could reliably use it to *preclude* demonstrable
relatedness for any sets of languages (which would require knowing
what the time depth of the *putative* family was -- and since it
might not be a family at all, figuring out that number is going to
involve some pretty mysterious methods) seems likewise to be a folly.
 
That said, since "related" in my book means, when used of
languages, "having been shown to have greater than chance systematic
correspondences for which borrowing is an unlikely explanation",
"unrelated" means the opposite (i.e., 'not having been shown to have
greater than chance systematic correspondences for which borrowing is
an unlikely explanation'). I therefore see no problems with asserting
that language X and language Y are 'unrelated.' The term seems to
mean *exactly* the right thing, in fact. "Related" in some other
sense (e.g., having a common origin as a factual matter) is never
accessible to us as scientists and I would never be willing to assert
that, e.g., it is a matter of *fact* that the Indo-European languages
are related in that sense (since it is not knowable). It is a
hypothesis and the form of the hypothesis is as I stated it above:
the languages have been shown to have greater than chance systematic
correspondences which cannot be plausibly attributed to chance (broadly
construed) or borrowing. We can be very wrong about our standards
for "plausibility" & such like (which are the subject of methodological
debate), but there's no point in pretending that we're talking about
whether the languages have a common origin *in fact.*
 
The responsibility for amassing the evidence which would indicate that
two languages (or families) show such greater than chance systematic
non-loan cognancy is *always* on those who propose the idea, like the
evidence for ALL scientific hypotheses. That different scholars are
going to have differing standards of proof, different assessments of
the liklihood of borrowing or chance, etc. seems inevitable and
pretty much uninteresting -- this is true in every scientific enterprise
I've ever examined. That these matters turn into religious warfare
tells us a lot more about the pathetic state of the human cognitive
system than it does about any language families.
 
On Sun, 22 Feb 1998, Scott DeLancey wrote:
 
> > Hmmm... How do you arrive to an estimate of 6,000 years? You obviously do
> > not base you calculations on glottochronology (in order to avoid
> > unnecessary discussion I should say I don't believe in it either), but
> > then what is your *objective* method of calculating time depth for
> > established or reconstructed families? I am pretty much afraid that it is
> > based on a guess-work, isn't it?
>
> Bravo!  This cannot be said often or loudly enough.  It is downright
> scandalous, this propensity linguists have for casually attaching
> dates to language families.  We have no objective, operationalizable
> method for estimating time depths, and no excuse for claiming (or
> even imagining) that our eyeball estimates have any validity.