Re: [tied] Proto-Steppe Numerals

----- Original Message -----
From: "Glen Gordon" <glengordon01@...>
To: <cybalist@egroups.com>
Sent: Saturday, August 26, 2000 11:41 AM
Subject: Re: [tied] Proto-Steppe Numerals


> Glen: It's easy to "prove" things in a physical science
like Physics, Mathematics,
> Chemistry, etc. because they all deal with tangible things
and the answers
> to a given question are always absolute responses: "Yes",
"No", "10%",
> "50mm".
> In theoretical sciences like Psychology, Sociology or
Comparative
> Linguistics (and all the humanities), there is nothing
tangible that one can
> prove. Like I tell John repeatedly, you won't find a mummy
with a piece of
> prehistoric language clutched in his hands.

To begin with, mathematics is NOT a "physical science" but
an exact theory of logical deduction. It's ONLY in
mathematics that 100% certainty is possible in proving
anything, since mathematics studies the logic or abstract
relations and does not concern itself directly with the
messy "laws of nature".

The remaining sciences can be organised into a conventional
hierarchy depending on the complexity of the systems they
are interested in. The simpler configurations they study,
the more rigid abstract models can be employed and the more
quantification and computation is possible. However, not
even physics (not to mention chemistry, astronomy, cosmology
or geology) is as exact and doubt-free as pure mathematics.

The biological sciences deal with dauntingly complex sustems
and processes. They are not so easy to formalise and
consequently individual opinions may play a significant role
in evaluating answers to non-trivial questions (e.g. "Did
birds evolve from small maniraptors or from primitive
archosaurs?" "Were the Neanderthals a subspecies of Homo
sapiens?"). Of course the social and behavioural sciences
are the least objective of all. But in ALL sciences there
have been remarkable advances towards more objectivity,
greater reliability and deeper insight, mainly thanks to
their interlocking relationships. Scientific disciplines
form a spectrum, and the subdivisions we introduce for the
sake of convenience are arbitrary. Progress in physics, for
example, helps chemists, biologists and even archaeologists.
Your classification of science into "physical" and
"theoretical" (?) is artificial.

> So continuing to use the word "proof" as you do seems
nonsensical and
> arbitrary to me. "Proof" in such sciences is a matter of
opinion only, since
> it depends on the individual's tolerance to probability.
All one can do as a
> theoretical scientist is to travel down the most probable
paths to arrive at
> a conclusion, which by nature will forever be
probabilistic and uncertain to
> a degree... that includes our beloved IE hypothesis, of
course. The point is
> to make one's "hypothesis", which is a theoretical
science's version of
> "proof", the _most_ probable given the current
information.

> To give an example, if we were to ask ourselves "What is
the closest
> language related to Gilyak?", how might we go about this
answer? If we went
> about it your way, Piotr, by "proving" somehow that Gilyak
is related to
> another language, we'd certainly never get anywhere
because by insisting on
> "proof", we are insisting on an absolute answer that can't
exist because of
> the very nature of the question. Thus, as you do, we say
"Don't know" or "We
> can't say" which isn't an answer at all.

We do use the phrase "beyond reasonable doubt" rather than
"100%" when we assess the quality of a "proof" in historical
linguistics. Still, I'm against the postmodern tendency to
reduce science to "narratives" and personal opinions. As I
said, even in physics answers are uncertain to a degree,
though of course 99.93% certainty is better than 68.37%
certainty or a "certainty" that doesn't lend itself to
quantification at all. Comparative reconstruction typically
involves much speculation, but there are certain ways of
making it less speculative by sticking to formal discipline.
The work of ANY scientist, including linguists, is evaluated
by its objectivity and rigour. For example, in order to
assess the level of "background similarity" due to chance,
comparison with an outgroup may be necessary (that's why I
mentioned Quechua and Austronesian for the purpose of
illustration). Distributional arguments may help in
screening the data against "loanword contamination". It
means painstaking work, but who says linguistics is easy?
There is a vast literature devoted to these technical
issues. Comparative linguistics is not as helplessly
subjective as your statements imply.

"Don't know" is a modest answer -- not defeatist at all,
just cautious. You speculate that Gilyak (well, I do prefer
"Nivkh") is related to Korean and Altaic. But since you
don't provide convincing evidence to substantiate this claim
(I mean the sort of evidence that would convince others and
not only strengthen your own hunches), your opinion remains
highly subjective and you'll PROBABLY have to keep it for
yourself (but let your model evolve and we shall see).

> In this case, it would be more complex to say that Gilyak
isn't related to
> anything, since it most certainly HAS to be, unless we
have sufficient
> reason to believe that Gilyak is a language distinct from
all other human
> languages!

Advocatus Diaboli: -- I see, Glen, you tacitly ASSUME that
all languages are ultimately related in a non-trivial
phylogenetic sense (i.e., that they form a neat family tree
which is reconstructable at least in principle). However,
except in highly controversial classifications, there are
many isolated languages and tiny groups, so Nivkh would be
no exception. Is your tacit assumption warranted, then? What
if isolates and numerous small families were the NORM until
pretty recently (as they still are in some parts of the
world)? This possibility undermines the assertion that Nivkh
MUST have some extant genetic relatives.

Thanks for the examples of Altaic phonetic correspondences
(I'd STILL appreciate more lexical material). I'll need some
time to analyse them, but you can count on me in this
neverending debate.

Piotr