Dating the final IE unity

[X99Lynx at aol.com]

In a message dated 3/2/2000 4:29:38 AM, Rich Alderson quoted me:
> Is this differentiation is quantifiable?  Are you sure it works in your
> favor?....
> Following Renfrew, roughly 4000 years separates non-Anatolian PIE from
> Mycenaean (1200BC), Sanskrit (1000BC?) and Latin (500BC).  How
> 'differentiated' are those three languages?  On a scale of 1 to 10?

And then wrote:
>>Now, to me, and apparently to Mr. Stirling, this is a challenge to state how
>>easily these three languages are recognized as being related to one another.
>>Nothing in the query stands out as asking for "the time of separation" of
>>these languages,...

Well, even my quote given above includes a pretty good indication of what
kind of difference I was talking about <<roughly 4000 years separates
non-Anatolian PIE from Mycenaean (1200BC), Sanskrit (1000BC?) and Latin
(500BC)>>.

And of course Mr Stirling has continuously used selective examples of IEness
to somehow give dates of separation.  None of these examples have proved
anything about absolute dating and much less any absolute statements about
IE's final unity - which is after all the subject of this thread.

It's blindingly obvious that 'agnis/ignis' does not represent the typical
difference between Latin and Sanskrit.  And when you compare this "look how
similar" approach to what geologists, biologists and yes archaeologists
consider credible time measurement, it does not come off as particularly
viable.

Going back in the archives, one will see that my point has always been about
how clear the evidence was against Renfrew's dates.  I cannot take any of Mr.
Stirling's arguments about archaeological dating seriously.  They are just
oftern as not outright incorrect.  Go back to the beginning of this thread,
and you'll see that the first arguments against Renfrew were all based on
paleolinguistic dating of archaeological subjects.  Every single one of those
arguments is either weak or simply now wrong.  Everytime I dig into one I
find there's nothing there.  (With the possible exception of an argument that
a list member sent to me privately that may date the wheel in fact to some
stage of PIE.)

The other argument against Renfrew's dates is the supposed correlation
between differences in unreconstructed languages and the time those
differences represent.

I respect Rich Alderson's impressions of "time/differentiation" measures
(Hittite to Mycenaean/Sanskrit = 500years) as I do Miguel Carrasquer Vidal
([The difference between] "Hittite and either M.Greek or V.Sanskrit for me is
definitely 2000 years or more.").  But I really don't have to choose.

You see, my guess has always been that there is NOTHING in comparative
linguistics that in some scientific way clearly DISPROVES Renfrew's dates.
And although professional impressions have weight, they cannot be dispositive
on scientific issues.  Nothing so far has scientifically disproven Renfrew's
dates.

And despite what has been written on this last any number of times - advising
me to read textbooks and such - I haven't found a single instance of anything
resembling an objective, reproducible method of measuring time/difference in
languages.  If it is an art, that is fine.  But then the objection to
Renfrew's dates are artistic, not scientific.  (I have found exceptions to
what I've just written in the textbooks, but they all deal with
glottochronology - e.g., Robert Lee's formula for calculating 'time depth'
given by Lehmann (1992, p.176) at least attempts to quantify elements and
just as importantly offers reproducibility and a measure of possible error -
requirements these days in real science.)

In this context, the persistent use of examples of cognates to prove some
kind of time relationship is very troubling.  It's basic problem is
demonstrated - just as an example - in the post below:

In a message dated 2/25/2000 7:04:34 PM, rao.3 at osu.edu quoted:
>"Stanley Friesen" <sarima at friesen.net>
>> I know how similar cognate words are in Hittite and Sanskrit.
>>  After one abstracts out the differences in writing systems, they
>> are very little different at all.  Indeed, if the phonetic differences
>> were all there were, they would be more like dialect variants of
>> one language.... - much like the fact that in my dialect of English, "worm"
>> is pronounced almost identically to the reconstructed
>> PIE root it derives from [*wrm] - sans endings).

Here the author is saying that stripped of writing systems and if "the
phonetic differences were all there," THE COGNATES in Sanskit and Hittite
would sound not different than in two "dialect variants."  Of course in other
messages, the same author has also pointed out there is a 'dearth of
cognates' in Hittite, raising the question of how much of Hittite he is
comparing to Sanskrit.  Do the amount of cognates count against time?  What
about the non-cognates, do they count in easuring time?

But more importantly what do these cognates say about time differentation?
In the very same post, he tells us how much these cognates can change -

<<...much like the fact that in my dialect of English, "worm" is pronounced
almost identically to the reconstructed PIE root it derives from [*wrm]>>

5000-7000 years and the 'cognate' sounds the same?  What does that tell us
about the use of cognates to indicate time and change correlations?  How long
does it take agnis/ignis to stop sounding a lot like "ignite" in modern
English or "ognik" in modern Slavic?  This is like clocking something with a
watch that isn't running.

When I brought up the fact that its the differences that should be measured
and that agnis/ignis does not occur in Mycenaean or Hittite and that thay is
the key difference, I got the following reply:

On Fri, 25 Feb 2000 11:48:36 -0800, Rich Alderson wrote:
<<The answer to the question *as posed* is "Within the bounds of phonological
change in the individual languages, yes, Mycenaean and Hittite decline their
words for 'fire' similarly to Latin and Sanskrit. That is one of the defining
characteristics of the IE family, after all."

But that's not what you meant to ask, is it?>>

I'll take that as a measure of time/difference, if by thae above it was meant
that the 'phonological change in the individual languages' was going to be
equated to a reproducible, objective meausre of time and change in languages.
 Yes, that is what I meant to ask.

And the next question I'd have is how long does it take to come up with an
-xi conjugation - which I understand to be a real difference between Hittite
and the languages it is being compared to.

In the same post, I was quoted:
>> Well, it seems that Anatolian is in the picture when the evidence helps, but
>> x.not when it doesn't.

The defense offered was:

>But that's the way of *all* evidence in *every* discipline:  If there's
>nothing to be said by a particular witness, you don't bother to call her to
>the stand.

Unfortunately, that is not true when the witness has contrary evidence to
offer.  In many disciplines, failing to call her might then be considered
falsifying.  In 1978, I saw two economists censured and fined in federal
court for failing to mention contradicting evidence in their affidavits
before an appeals court.

There is much contradictory evidence here.  And as I said above, the position
I've taken doesn't require me to take sides.  I'm only contradicting the
statement that Renfrew's dates are 'impossible' given the linguistic
evidence.  At this point I can say that certainly is not true, by any
scientific standard.

Regards,
Steve Long