[tied] Re: Nominative Loss. A strengthened theory?

From: elmeras2000
Message: 31968
Date: 2004-04-16

--- In cybalist@yahoogroups.com, enlil@... wrote:

[...]
> Jens:
> >However, it does not add up:
> >
> > If *-yo is an endingless locative of the relative pronoun (Oh
> > dear, ending in *-o), and the s-forms are nominatives, the phrase
> > would be construed to mean "the eye at which there is a wolf".
> > That is nowhere near what is meant.
>
> It still is semantically the same as a possessive construct,
> although the nuance you give it attempts to obscure a relationship
> that still can be seen here. It shouldn't be translated as "there
> is a wolf" but it still gives us a possessive.

How else could an isolated dumping of a nominative of "wolf" be
understood? I can see an alternative that is worse: "an eye - hey, a
wolf - that's where it is". That could just conceivably be pieced
together as a sequence of disconnected utterances that could add up
to the actually intended meaning, and over time this could become
the syntactic norm of the language. It would demand some backing,
though, I'd say, and I can't see it has any.

> In fact, it may just as well be that the phrase was once
> mLIE */wlkW&-s ya hWa:kW-s/ which can then mean that *ya is
> in reality modifying the following noun. Hence "the wolf
> (*wlkW&s) at which (*ya) [is] the eye (*hWa:kWs)". Afterall,
> we do see the relative particle preceding the clause in IE.

"The wolf at which [is] the eye" would be a statement about the
wolf, not about its eye, as it is supposed to be. This is an error
of calculation, plain and simple. Your suggestion has failed at your
own test.

> Either interpretation will do in the end but we always need to
> have the *-yo portion in the locative to make any sense of it
> without leaping to conjecture that there is a case ending we
> can't see. You continue to struggle to get me to believe in
> something no one can see but you.

I always feel flattered when it is verbalized that I can see more
than others. For reasons beyond my comprehension nobody else has
seen the problem being cried out by the word-final vocalism of *so
and *-esyo/ *-osyo. Since the vowel concerned is identified as the
thematic vowel (certainly in the sense of "stem-final vowel"), the
known rule that this vowel is *-e when word-final ought to apply. I
think it does, and that this indicates that the vowel was not word-
final at the time it was settled as /o/. If that is not the reason
for the discrepancy with a known rule seen here, I would like to be
told what the real reason is. Something *has* to be assumed to get
this right. As long as you have made no suggestion as to what can
have caused this deviation from normal rules your job is not done,
or is seen to be incorrect. I think the latter.

> > There are lots of endingless locatives in IE, they just never
> > end in *-o. The only case in point I know for the thematic
> > vowel is the "augment" *(H1)é 'then' and perhaps the particle
> > *kWe 'and'.
>
> Erh, definitely not *kWe, but certainly *e, yes. You see, *e is
> the proper and ancient oblique form of *i- (just as *twe is the
> oblique of *tu:) but a locative *i developped out of analogy with
> the typical endingless locatives at the time. Then a new locative
> was built for all these demonstratives as the language changed
> once again.

You don't know any of this. I see no reason whatever to believe you
when you stand up on the podium and make such self-assured
declarations.

> As I said, the indicative in *-i and the locative *-i shows
> the endingless stem in plain view. Again, a locative interpretation
> of both particles makes the most sense since the indicative
> would once assert the existence of an action "at this/that time;
> now; then" and the locative... well, having it derive from a
> locative ending is completely natural.

There is a lot of mutually supportive evidence to show that /i/ is
an old enclitic variant of e/o. That means I accept the
identification of *-i as the unmarked stem, i.e. endingless
locative, of the pronoun *e/o-, meaning precisely now/then and
here/there. That makes the augment *e- and the present-ending
enlargement *-i old positional variants of one and the same
morpheme. I am aware that this is unsatisfactory with regard to the
functional difference which I cannot explain. I would welcome
sensible suggestions.

> > Word-final thematic vowels have the shape *-e. One way of
> > accounting for the irregular shape *-o would be to imagine [...]
>
> Stop! We don't imagine anything because you haven't shown that
> *-o in *-syo must derive from a "thematic vowel". You're solving
> problems that don't even exist again! It's your strange obsession.

Well, you also said this. Maybe you do not have the necessary
insight to understand the full implications of your own words, but
when you identified the part *-yo as a form of the relative pronoun,
you automatically identified the vowel it contains as a stem-final,
i.e. thematic, vowel. And you have done the same with *so. In this
point I agree with you, if only until you go back on your words or
disclaim their implications.

> > That does not necessarily make it flawed. All explanation in
> > internal reconstruction is circular, it is only a matter of the
size
> > of the circle.
>
> If so, then we must reject all the times you've accused me of being
> circular :) We can't have that, now can we?

We can if your circle has no extension. If your rules are only "ad
hoc", they are less good than if they are both "ad hoc" and "ad
illud", in the sense that their application can be repeated. That is
the principle we use. We are looking for the largest circle.

> Internal reconstruction
> should be based on what we observe and the more general the
> conclusions we can base our overall viewpoints on, the better.
> I don't see that as being "circular".

Doesn't matter what you call it, we are in basic agreement on this
point.

> Rather, if you insist on something that isn't there and provide
> other evidence that is equally not there to support it... that's
> really circular and has no place in internal reconstruction. Yet
> this is exactly what you're doing with **-syoz and **soz.

No, that is a direct lie. There *is* evidence, quite much actually,
to make us expect that word-final thematic vowels turn up as *-e,
not *-o. That is insisting on something which is there, namely this
principle applying precisely to this language as far was we have
advanced in our analysis of it at the present moment. That makes it
impossible to accept the forms *so and teh final part *-yo as
originally endingless forms, so they must have been something else.
There *is* evidence, quite strong actually, that nominatives of
thematic stems end in *-os (on the chronological layer of the
protolanguage, whatever its prestages may have had). There *is*
syntactical necessity (no less) to determine the form *so as a non-
neuter nominative, and something very close to that applies to the *-
yo part of the gen.sg. ending *-esyo, *-osyo. That makes the form
*so in *need* of a final *-s, and it makes it very probable that the
*-yo also once had a final *-s. Since that is at variance with what
we observe, we *need* a change from something like *sos to *so, and
of *-syos to *-syo, of a kind not contradicting rules we have. It is
a general fact that dissimilations occur, and that they are often
spontaneous and not bound by rules. That then offers the smooth
explanation we need: *so-s became *so, and *-syo-s became *-syo by
the same kind of dissimilation. If that is accepted, there are no
syntactic oddities in it all.

> > Hey, that was an objection raised against your analysis, not
against
> > mine. I am not sure it is valid. Even if the 'eye' of your master
> > sentence may also occasionally be the object of a larger phrase,
it
> > would be "the eye at which the wolf is" with the wolf constantly
in
> > the nominative. But that demands this underlying analysis which
we
> > have not seen you make.
>
> In order for you to propose *-syo having the case ending, you must
> show how "WHICH" is nominative, not "wolf" or "eye"! So by adding
> *-z/*-s to *yo, you're saying "the eye which the wolf is".

No, if the relative "particle" is enclitic it must hang on the first
word of the phrase, no matter which word it belongs to
syntactically. Celtic works that way with regard to the relative
pronoun, and for other enclitics all old IE languages show this rule
(Wackernagel's rule), and Serbo-Croation still works that way today.
In a language like that it is a matter of interpretation which of a
number of things is actually said. It's like when you don't know
where to translate the enclitic genitives in Vedic.

> Evidently senseless for the purpose of creating a genitive unless
> it's "the eye which is OF the wolf" making *wlkWos necessarily a
> genitive even though it's identical with the nominative. Yet, since
> there's nothing distinguishing the genitive and the nominative at
> this stage... how can this possibly be so without remembering
> that the nominative-looking word is a genitive? Are you suggesting
> that IE speakers were well versed in etymology?

In a way yes, for they could *hear* the difference between
nominative *-z and genitive *-s. But when that difference was
becoming neutralized they had to disambiguate which they apparently
did by drawing in the relative-pronoun circumlocution. If the
alternative is a unacceptable as you say, I thank you for it, for it
then amounts to proof that there once was a phonetic difference
between the two sibilants we observe in the nominative and the
genitive. That however is only if your logic is not flawed. I'm
afraid it is, for the oldest form of the thematic genitive was of
course the one in *-esyo which offers no clash with a nominative in
*-os.

> If not, they
> surely would forget what case it's in. All they'd no is that they'd
> have to come up with a structure that would convey without
ambiguity
> the genitive they wish to express. Hence *yo.

So they just took *yo, because it is not zero? Is that supposed to
explain why they chose exactly that?

> Your solution opens up a paradox. How do you feel, Pandora?

No, it respects the grammatical typology of the language as we know
it and avoids contradiction completely. It is simple too if that
counts for anything.

> > So now the first part of *wlkWos-yo is itself already a genitive?
>
> Well, if anything it could be either case for the very reason why
> *yo was added in the first place... to disambiguate the nominative
> and the genitive. However, in your analysis, you NEED to have
> *-s- be the genitive.

Yes.

> A nominative is senseless, yet it would be
> nominative by default unless disambiguated in the first place.

It would grow to become ambiguous, yes, and so the disambiguation is
well motivated.

> Adding *-yo by itself doesn't do this because it yields "The wolf
> which is the eye" yet it can't be the genitive until it's made
> different from the nominative. No sense!

Not if this particular kind of univerbation, after the analogical
change of *-esyo to *-osyo in substantives, was reserved for
genitival constructions, which is of course the only role we find it
in. This could even motivate the choice of an irregular allegro
pronunciation. If <*wlkWos yos HwkWo:s> (vel sim., we're talking
structure now) could also mean "the wolf which is an eye", or "the
eye that is a wolf" instead of "the eye which is the wolf's". But if
we accept the existence of the old form *WlkWe-s up to this point
there is no basis for all this (but then there is a basis for some
other things which your theory cannot survive either).

> In my interpretation, *wlkWos may be either case. It has no
> bearing in the end with the analysis, really. Whether it's "the
> eye (nom) at which (loc) the wolf is (nom)" or "the eye (nom)
> at which (loc) the wolf is of (gen)", it makes little difference.
> The meaning still conveys possession. Granted it may be wonky.

What is the intended meaning of "the eye at which the wolf is of"??
That does not look like anything that can make sense here. Also "the
eye at which the wolf is", while not logically different from the
intention, looks very odd indeed. Importantly it is the exact
opposite of the parallels used to back it, viz. language having a
local construction of possession. Russian says <u menja kniga> "at
me is a book" for 'I have a book', and not "I am at a book". Can you
point to a language that does this, for that is what you are
assuming for your pre-PIE construct, swearing on a bible of typology
at that?

> So, if we take the above suggestion that *yo may in fact
> mark the FOLLOWING noun, the possessed, then we either get
> "the wolf (nom) at which (loc) is the eye (nom)" or "of
> the wolf (gen) at which (loc) is the eye (nom)". It may
> even mean "of the wolf TO which is the eye" since the
> dative and locative were both endingless at the time.

But "the eye" is the head of the construction which varies according
to its role in the larger syntactic frame. You just killed that.
This is not a language anymore.

> Regardless of any single way we slice it, it all indicates
> possession. The analysis of *wlkWos as a genitive or nominative
> then is immaterial to my view while necessarily a genitive in
> yours, to add to the necessity that a case ending must exist
> where we never find it. Your view is clearly not optimal.

As opposed to the alternatives which are all impossible, it may have
had too little competition.

> So *ya was certainly the original locative of *yas (> *yos).
>
>
> > These are arrogant decrees issued on no basis at all.
>
> If "decreeing" that *so is undeclined for case when it truly is,
> then you'll never get it. It's your own arrogance you should
> worry about.

Leave that to me, for now I'm more worried about contradictions of
rules that have been settled already. Of course they are also open
to question, but I don't see these questions being addressed. It
remains that, as long as a thematic stem surfaces with *-e in
endingsless forms, *so and *yo cannot be that. IIf they can, the
basis of the thematic vowel allomorphy should be addressed in an
open and proper way. You are simply changing the subject every time
it come up. That is hardly a sign of arrogance on my part.

> Your greatest fear is the fear of being wrong and
> I see the anger building up as you cannot come to terms with it.
> I guess I'd do the same if I saw being wrong as a personal failure
> rather than a character-building strength.

This is alien to the topic. It is germane however, by being also
something you don't know a thing about.

> Since _I've_ admitted many times of being wrong and I never heard
> you once doing the same when you have been, your continuing
> accusation about decrees in light of **z is really coo-coo-for-
> cocoa-puffs.

I have been very wrong to take you so seriously as I have in the
many fruitless debates we have had. You actually have been issuing
decrees. Say something well-motivated, and I'll agree.

> > Rather if you go to the extreme of appointing *yo an endingless
> > locative of irregular shape used in an illogical syntactic
> > construction, why not do the same with *so which is structured
> > just the same?
>
> Obvious: *so has absolutely no case endings what so ever. We see
> that *yo- does. Ergo, mLIE *sa was not the locative of anything.
> Rather, *ta (with the optional addition of postparticles like
> *bHi or *dHi) was its corresponding locative.

How can you know it was not the locative that survived as the sole
remain of a defective stem? I'm not advocating this, however, but
that is for other reasons. I see *so as a reflex of what it is
functionally, the nominative masc. of *to-. I know this is only a
suggestion of minor value on which I would not venture to build
much. However, I cannot just watch you constructing almost the
entire prehistory of IE morphology on the mere hope that the
suggestion, easy as it is, is necessarily wrong.

Jens