On Sun, 1 Jun 2003, Glen Gordon wrote:
>
> Me, being characteristically sarcastic:
> >Apparently, Jens thinks that Uralic speakers could read in the
> >5th millenium BCE.
>
> Jens:
> >Still, they must have counted the days, so we can use the verb in its old
> >meaning 'count'.
>
> It's good to know that you haven't lost your mind completely.
> Alright, "to count" would be acceptable as long as there is
> evidence to support that definition. I don't quite see how
> "to count" becomes "to read". What is the in-between?
It means 'count ' even today. The common ground may be 'gather
information'. Russian c^itat' 'read' : c^islo 'number', etc.
>
>
> >It did stop, if only after I had presented the full basis, for fear
> >that it would otherwise appear to be without foundation.
>
> And where _is_ this foundation you speak of? You know full well
> that it's easy to make up some drunk idea that *gHWn-me <
> **gHWen-d-me-g but where on earth is the proof? How do you
> support **-g?? Where did that **-d- come from?
>
> I'm naturally not interested in what evidence in Eskimo-Aleut
> supports some god-awful pre-IE construct! If we are talking
> about IE, let's talk about _IE_.
Of course the theory was inspired from a squinting across the fence to
outside connections, but I did not invent the idea of such connections,
and in fact there are observations that go a long way towards supporting
them in a very stong way.
It demands an explanation that the dual and plural endings of the verb
ending in *-e are "weak" paradigm forms with accent movement, if the
nominal animate nom.-acc. dual in *-e are "strong" forms without accent
movement. The basic rule is that underlyingly vocalized endings move the
accent while endings consisting only of consonantal material do not. Thus,
if I derived the nominal *-e from a secondary propvowel inserted before
the ending *-H3 in a much the same way as Finnish -e is derived from *-ek,
then I come under obligation to explain why there *is* accent movement
where the underlying structure, based on the function of the output, would
be something ending in *-g-m-g og *-d-m-d or the like (including its
apparent replacement *-d-m-g). Now, what is so dramatically different
between *-C-g, *-C-m on one hand and *-C-g-m-g, *-C-d-m-g on the other?
Well, that's an easy one: structural complexity, the sheer number of
consonants. It will be relatively easy to go on having a structure like
*-C-g for some time, but it will be asking too much of the speakers to
have them use *-C-g-m-g etc. without the help of intervening vowels. I
therefore assume that the propvowels are not equally old: The vowel
inserted in *-C-g > *-Ceg > *-Ce was only inserted *after* the working of
the mobility-causing rule shifting the accent onto (or in the direction
of) desinences with vowels, while the much bigger sequences *-C-g-m-g etc.
had vowels already at the time of the operation of the mobility rule and
so shifted the accent.
That accounts for the difference in the effect excerted by the filling
vowels of the endings. Now, why do I not make do with onlt *two*
consonants in the long endings, i.e. why do I not just posit, say, *-m-g,
*-t-g, *-m-d, *-t-d (granting for a minute that I am justified in my lover
fore these elements)? I can't do that because the inserted vowel - which
here is *very* old - comes right after the person marker. That makes the
1du end in *-meg, the 1pl in *-med. Now that threw out the baby with the
bathing water, for it is the contact between /m/ and a preceding dual
number marker that has caused /m/ to change to the /w/ we find in the 1du
forms. That change is seen in the personal pronouns 1du *nH3we 'us two'
vs. 1pl *nsme 'us' where the w/m is the accusative marker (it also appears
in the other persons), and we now also find 1pl *-me opposed to 1du *-we
with the verb. I have therefore suggested that it was a dual number marker
*-g- that changed a folwoing /m/ into /w/ (I'm writing in ASCII here, the
-g- is meant to be a fricative, and so is the -d- in all of these forms).
In th personal pronoun the "-g-" can be relatively safely identified with
the known phoneme /H3/, and the change m > w /H3__ can be seen also in the
old verbal noun appearing in the Gk. infinitive dowenai 'to give' as
opposed to themenai 'to put'; the antiquity of dowenai is borne out by
Ved. da:vane 'to give' and Lith. dovana 'gift'.
So, even if we do not see the "-g-" directly in the 1du forms, we do see
its shadow in the change of the personal marker from /m/ to /w/. Now, why
does it no surface to show you that it was really there? That is, why do
the elements "-g-" and "-d-" of *-C-g-me-g and *-C-d-me-g not surface if
they do in *mweg-me, *mwed-me, later *negme, *nedme which yield PIE
*nH3we, *nsme in the end with the number markers intact? Well, because
-C-g-m-, -C-d-m- are more complex structures than -gm-, -dm-: In -C-g-m-,
-C-d-m- the number markers -g-/-d- were interconsonantal already before
the ablaut brough about zero-grade to make it even worse. Therefore it
runs counter to no known observation if we assume that these elements were
lost when interconsonantal but retained when adjacent to vowels. The rule
may have to be refined, but this is as much of it as we can see for now.
We cannot posit the preforms without these elements, for they are demanded
by the function and also by the phonological change of /m/ to /w/, and
they tally neatly with the demand for extra complexity that emanates from
the accentual mobility. I admit that I would not have made the analysis
without the support of the Finnish forms which plainly show the underlying
structure that makes it all add up.
>
>
> Concerning the adjective to describe Jens' new "theory":
> >May I suggest 'good' or 'constructive?
>
> We can only call it "good" if it is founded on logic -- It isn't.
> We can only call it "constructive" if, after being substantiated,
> it leads us to a new understanding of pre-IE -- It doesn't.
>
> May I suggest "crappy"?
> This theory is pure fantasy and unbecoming for a linguist.
>
>
> >The dual is nowhere on the advance, but plainly has the status
> >of a fragile archaism wherever we see it.
>
> True somewhat, although you taint it with biased interpretations
> without substantiation. What we can immediately say unbiasedly
> is that the dual was not fully formed by the time IE dissolved.
> On the surface, we can say nothing about an "advance" or an
> "erosion", or whether it is truely an "archaism" or not.
The erosion goes on to its complete obliteration in the full light of
history during the attestation of a number of Slavic languages, of Greek
and of Iranian and Indic languages; also Irish and Germanic languages give
facets to that picture. That is not a tainted presentation of the facts.
Outside IE I know the disappearance of the dual can be followed in
Greenlandic.
>
> We must have more to go on than just this sole arguement if
> we are to discourage other competing possibilities. Some other
> competing possibilities are:
>
> 1. The dual is recent and its non-transparency is the
> result of a sound change or the like.
>
> 2. The dual is recent and its origins have not been
> properly and fully considered.
>
> Personally, I side with the consensus that the dual is a late
> innovation that didn't have enough time to fully form in IE.
> Since we can be sure based on a lot of evidence that the
> Anatolian group branched out earlier than the others, its
> lack of dual (among other things), indicates that the dual is
> indeed a recent innovation.
>
> Further, it's an overstatement that the origin of the dual
> isn't "transparent". It all stems from the numeral *dwo:u
> "two". The ending *-u is analogical and based on the
> opposition between prefixed *dwo- and the full form
> *dwo:u, which makes it look like there is an ending where
> there isn't. So etymologically speaking, there is no dual
> ending on *dwo:u, nor is there one on *okto:u "eight"
> but it came to be perceived this way by IE speakers. As a
> result, natural pairs began to be marked in the new dual,
> such as "eyes", "ears", etc.
Yes, congratulations, it's reading a gross overstatement into my words if
you read them to mean that dual are *never* the least bit transparent. I
can analyze quite a lot on a good day, and some dual forms are no
excptions; still many are. Let me hear what system you see in the dual
forms of the middle voice and the perfect. Moreover, the dual seems to
have its most stubborn stronghold in the personal pronouns; that has
nothing to do with paired body parts.
> From this, the dual system developed further, being optionally
> markable in the verbal system. (I say "optionally" because
> we could just as well use the plural pronominal endings on
> dual objects and that's what we sometimes find attested.)
>
> What you see as "fragile archaism", I see as a "fragile innovation".
>
>
> >What do you mean? 'Two' in Uralic is something like *kakta. How is
> >the powerful argument construed which you apparently derives from
> >that?
>
> It is the same powerful arguement that shows that Norse /vit/ has a
> similar etymology.
This is nonsense: Nobody called /vit/ an IE dual (that form was *we), but
the fate of vit in Icelandic (and in part Norwegian) illustrates that an
*old dual* can become a *later plural*. Now, that's the *opposite* of what
you need to illustrate.
>
> >I do in fact reason that way, meaning that possibilities should be
> >kept open until we know they are not true.
>
> That's why your reasoning is disturbing and misguided. I don't object
> to imagination and is needed to find new ideas. However, we don't
> base our theories on imagination.
>
> In fact, it should be logic first and foremost that determines the
> shape of our theories. By your admission, you side with imagination,
> making your theories chaotic.
>
>
> >That, however, always has a simple reason, viz. that those
> >suggested are demonstrably flawed or simply did not add up in the
> >first place.
>
> And your baseless pre-IE morphology replete with **-g and **-d-
> "adds up"?? Come on now. To substantiate these markers I can
> only accept evidence _within_ IE itself, but there is none.
Oh yes. I'we told oyu before, now I've just told you again.
>
> The onus isn't up to me to show that your ideas, or any other
> cockamamey ideas, aren't probable!
>
> It's up to _you_ to prove that your theory _is_ probable.
I think the onus is on all those genuinely interested to go and search for
the truth in an unbiased way. But if you see it this light of petty
oneupmanship, that's your privilege, and to that I'll say: Oh yes, that's
why I've done just that.
Jens