Re: Evidence for Nostratic? (was Re: Proto-Uralic?)

En réponse à Rob Haden :


>I understand all this.  What I meant was that I would like to see
>reconstructed (or reconstruct myself) a form of PIE that could have
>actually existed.  Nothing perfect, mind you, but something better than the
>chimera that is 'accepted reconstructed PIE.'
But *that*'s what we have. The only thing is that we *have* to accept an
error margin in time. It's impossible to do otherwise, unless you travel
back in time with a tape recorder. PIE is the most precisely reconstructed
Protolang we have, and it's just *impossible* to do better, for the simple
reason that it's all we can do with the evidence we have.

>   More specifically, I'd like
>to use reconstruction to answer the following questions conclusively (if
>possible):
>
>1. What was/were the origin(s) for PIE grammatical Ablaut?
>2. What was the origin for the sigmatic nominative?
>
>(etc.)
But that's impossible to do, since we don't have any evidence to start with
(internal reconstruction helps a little, but it has its limits too. As
always, you cannot reconstruct what disappeared without a trace, so you
just have to accept that some things we will never be able to explain. That
doesn't make PIE a chimera. It almost certainly was as we reconstruct it.
Only more so, since it probably had features which we just cannot reconstruct).

>I've been attempting, off and on, to reconstruct the stage of PIE before
>all of the daughter languages (including Hittite et al.) separated.
That's already done.

>Glottochronology is pretty important in reconstructing proto-languages.  We
>have to try to pinpoint, as accurately as we can, when changes occurred.
Already done. PIE as reconstructed right now is, given all the evidence we
have, the most accurate reconstruction we can make. Believe me, people have
been working on it for more than a century. What we have is not just an
unsatisfactory consensus. It's the best we can do without entering into
unreliable fantasy.

>Of course, I know that the techniques are far from perfect, but I think
>that we can come up with something better than what is currently
>reconstructed as 'PIE' (and someday we will).
We can't, because that's all we can do with the evidence we have. The
evidence is limited, and *no* technique can reconstruct what has
disappeared without a trace. Once again, without time travel, we're pretty
much stuck with what we have. But what we have is pretty much as hard as
any fact. If it doesn't please you, too bad, but you cannot just dismiss it
just because it doesn't please you. And it certainly doesn't make it a
chimera. Everything we know about PIE is pretty certain. What we don't
know, we cannot even guess with the evidence we have, so it's useless to
even try, as it would be pure speculation, with no value. Unless we
discover another branch of Indo-European even closer to PIE than Hittite,
and written in a script we can read, we're pretty much stuck with what we have.


>My point was not to deny the existence of laryngeals entirely.  I certainly
>think that 'laryngeals' existed in PIE; what frustrates me is that
>seemingly no attempt is made to figure out exactly what they are.
Because we just have no way to make those attempts. We know they existed,
but no evidence points out what their actual value was, except that they
were indeed guttural. That's all we *can* get with the evidence we have.
And internal reconstruction cannot help here because it has already been
pushed to its limits. Once again, we are at the problem that we cannot
reconstruct what has disappeared without a trace.

Also, you have to realise that even the reconstructed sounds that have been
given a "meaningful" label are actually also uncertain. We can make
educated guesses about their phonetic value based on what they've become
and what we know about the usual direction of sound changes and language
universals (which, you must always remember, are easily broken even in
current languages, and thus can only give us simple guidelines but no
certainty), but that's all they are, educated guesses. Why do you think
their are two competing theories (the classic and the glottalic ones) about
the actual phonetic values of the PIE stops? Once again, short of time
travel, we will never be able to be sure what the exact value of the
reconstructed sounds are. By the very nature of the research, we're stuck
at the phonemic level, and can make at best educated guesses (something
which is often enough) about the phonetic level.

>   Instead,
>they seem to be just accepted as these 'mystery phonemes' that have certain
>alleged effects.
Because there is just *no* way, from the evidence we have, to learn more
than what we know about them, and any attempt to guess their actual
phonetic value is what is it: guess work. And guess work without evidence
to back it up is certainly not scientific work.

>   What I'd like to see is some genuine, honest debate as to
>what the actual 'laryngeals' were.
Already done. The current laryngeal theory is the result of such debate.
You seem to be annoyed that we couldn't get further that, but you forget
that it's *all* we can get with what we have.

>   I've presented my own ideas on this
>topic, and I'd like to see what others think.  This goes for other subjects
>on the topic of PIE too.
Of course, it's not a reason not to discuss it on the list, especially
since such speculations are great for conlang ideas :)) . But you must
never forget that what you're doing is pure speculation, and that unless we
discover new documents from the past, the current accepted form of PIE is
the *best* we can get, and there's *no* way to enhance it.

>Perhaps we should move this discussion to a Yahoo group or something, as it
>has relatively little to do with conlanging (although you could say that
>PIE is something of an a posteriori conlang).  I would be happy to create a
>group for that purpose.
I think personally that such discussions are very welcome on Conlang. After
all, it's also a place to gather and exchange ideas, and a very nice source
of conlanging ideas. So a discussion on PIE is never unwelcome, as long as
everyone knows where they stand :) .

>Oh, and by the way, I've noticed something rather interesting about Latin
>and Greek.  In root nouns, where Latin has -e- for PIE stressed vowels,
>Greek has -o-.  For example, Latin pes, pedis 'foot' vs. Greek pous,
>podos 'foot'; also Latin dens, dentis 'tooth' vs. Athenian odous,
>odontos 'tooth'.
It was one of the first evidence that brought linguists to the
"monovocalic" theory of PIE, that it that it had a single vowel (i and u
being syllabic versions of the [j] and [w] sonants) which could take null,
"e" and "o" grades. It was the only way to explain the ablaut processes
within IE languages, but also between cognates in different IE languages,
just like the examples you gave. It's a strange theory indeed, but it's not
as if there was nothing to back it up. There are other (non-IE, but maybe
related - and certainly closely related if you accept the Nostratic
hypothesis -) languages with such a vocalic system (Kartvelian languages,
of which Georgian is an example - but not a good one here because it
doesn't have the vocalic system I describe ;))) . Sister languages of
Georgian do have it, though, and reconstruction can account for Georgian's
"added" vowels -), and they are still alive today (however, in those
languages the different grades are usually /a/ and /@/. But once again,
remember that *e and *o are just label for reconstructed phonemes. We have
only some idea of the actual phonetic contents, based on their outcomes in
different languages. Basically, the frontness and unroundedness of *e and
the backness and roundedness of *o we are quite sure of, but their actual
height is more of a guess).

Christophe Grandsire.
http://rainbow.conlang.free.fr
You need a straight mind to invent a twisted conlang.