Re: [tied] Re: Everything except the kitchen sink

On Sun, 11 May 2003, Glen Gordon wrote:

>
> About *wlkWos
> ----------------------
> Jens:
> >Punctuation aside, we are talking about types here. A word-type that
> >has lost a vowel by syncope must have existed when the syncopation
> >rule operated. Semicolon.
>
> As I said, there is naturally a period of accent-based ablaut AFTER the
> syncope (and apocope) caused zerograding, naturally causing further
> zerograding AFTER that event.
>
> No, your statement is clearly false. So *wlkWos DID NOT NECESSARILY
> exist when the "syncopation rule" operated. The onus is still in your
> court to extend the date of this word further back in time by means
> of further proof that this "must" be so.

This boils down to the question what is evidence and what is not. Are we
allowed to assume that a wordform showing the effect of a given sound
change existed at the time the said change took place? My answer is yes -
unless there is a very good reason not to assume that. Such a reason could
be analogy whereby new coinings followed the pattern of preexisting items.
That's why I slightly changed the topic to speak of the *word type*
exemplified by the word wolf. If the word itself is not as old as the
process it appears to reflect, its model at least must be. That's all we
needed to talk about.

>
> >What does not escape me is the fact that the nominative is still
> >marked by /-s/ in Greek, Lithuanian and Latvian, to which its
> >Icelandic reflex -r could be added.
>
> Three out of a thousand languages. Wow!
> You're once again twisting probabilities to suit your unlikely theory.
>
> The well-known linguistic fact remains: A marked nominative is unstable.

You are forgetting the many languages that mark the nom.sg. in different
ways than by a retained *-s. Practically all the IE languages we know from
about the Birth of Christ have a distinctly marked nominative. They all
have it by inheritance from the protolanguage. Nobody would date PIE later
than 2000 BC, some would go as far as 4000 BC, but at the very least the
nom. marking has remained stable for some millennia. I do not know how
much further back in time our internal analysis takes us, but it looks to
me much like the difference between Old Irish and Ogam or between Old
Norse and the golden horn of Gallehus. That's just a few hundred years. It
could of course also be more, but we can't know. Nor can we know where the
limit goes for a nominative marker - definitely present - to stay alive.
If you know, please share your soyurce of information with us.

>
>
> Aorist
> --------
> >I guess that's one of the few things about s-aorist theories we now
> >know to be wrong. The suffixal parts of action-noun s-stems and the
> >sigmatic aorist are just too different in their morphophonemic
> >behaviour to be identical.
>
> Their differing morphophonemic behaviour is irrelevant and fully
> explainable.
>
> The *s-aorist was adapted solely from the strong case form of these
> noun stems and wouldn't have the *s/*t alternation seen in nouns. On
> the other hand, the lengthening by the s-aorist is caused by the syllabic
> reshaping and shortening of verb stems in CVCVC- to CVCC- (*bHer-&s-
> >*bHe:r-s-) and wouldn't have affected noun stems. Their different
> behaviours reflect the divergeant evolutionary paths of nouns and verbs
> in Late IE.

There's no way you can know that. Didn't Ockey forbid you to make empty
guesses?

> >They would also be unified under a theory ascribing this effect to
> >their common marker which consists in the sibilant phoneme that
> >lengthens.
>
> As I already said over and over, **z is unlikely to be used so extensively
> for the commonest morphemes when it doesn't exist elsewhere. There
> is no intuitive reason why **z should lengthen preceding vowels either.
> I can't stress enough how assumptive this theory is.

You don't know how frequent the *z was at the time it was a separate
phoneme. It's like trying to separate the different sources of English
/r/: Which ones are from *r, and which ones are Verner variant of *s, and
can enough cases of the latter be made out on synchronic basis for the
very phenomenon to be at all credible?

> Your idea loses far more than it gains.

Only if we make you do the counting.

>
> About *so/*to-
> ---------------------
> >>Here's your solution in a nutshell:
> >> 1. We _assume_ that *so- and *to- are the same word.
> >> Why?? What shows us that they are?? Pure assumption based on
> >> nothing.
> >
> >Hey, it's based on the correct observation that other stems are
> >inflected all through, and on the similarity between *so- and *to-
> >in sound and length. That is not zero. Change a feature or two by
> >whatever rule, and the problem is gone.
>
> Amateur linguistics. Any linguist knows "similarity" isn't evidence. Nuff
> said.

It still counts for something - dissimilarity can be much worse.

> >> 2. We then assume some more that *s & *t alternate initially in
> >> order to support the above groundless fantasy...
> >> Why?? There's no evidence of *s/*t alternation initially or even
> >> medially either! Where is Jens getting his ideas from? Thin air
> >> obviously.
> >
> >Medially there is, but I do not invoke that, for *so-/*to- is
> >special in other respects too.
>
> Medially isn't initially, so I rest my case. And even in "medial" positions,
> this alternation is caused by their being originally final, proven by
> the 2pp *-te which does not show any sibilantization, yet is medial
> and surely ancient.

There is an alternation between initial/medial /(-)t-/ and final /-s/, but
apparently not all t's nor all s's undergo it. Not all English r's
alternate with s.

>
> The sibilantization of *t occurs very early, that much is self-evident,
> (in IndoTyrrhenian) and it occurs before weak case suffixes were even
> attached to noun stems.

These are relative concepts. I indeed believe the changing t/s
morphophoneme has -s- before weak-case endings. That is one of the
arguments for assuming an old word boundary where the thematic vowel
stands.

>
>
> >>My solution is as follows:
> >> 1. We assume that *so was simply an undeclined animate deictic
> >> added to an already declined paradigm using *to-.
> >> That's exactly what we see. The stem *so is only used in
> >> _animate_ functions and is never declined with case endings.
> >>
> >This is not a system the language shows us elsewhere, therefore it
> >is as inane, baseless and groundless as your impression of anything
> >you believe you have to fight.
>
> Aha! Firm proof that your views are offcenter.
>
> Anybody who has any amateur inkling of IE whatsoever will know that
> the paradigm of English "to be" is a merger of *bHeu- (be, been), *es-
> (am, is, are) and *wes- (was, were). Now, how can one possibly
> answer your misguided question above if we exchange *so/*to for
> "be/am/was" and "IE" for "English"?

Easily. One can see that be/am/was are more of a challenge than so/to
(come on!). On the other hand, if be/am/was *could* be paralleled by other
cases in the same language the testimony it gave would be *very* strong
indeed. I am sure many have tried to find at least some repetition of the
allomorphy elsewhere, but have been unsuccessful. By comparison, so/to is
not at all that hard, especially if "one-time events" are permitted.

>
> Point is, we can't seriously respond to an inane question as this.
> As with IE, English doesn't "show us elsewhere" such a system. It's
> a one-time event; It happens. So therefore, by your logic, English is
> inane, I suppose.

I would like to know what kind of *internal* analysis you would propose
for be/am/was. You appear to be saying that it is insoluble, and in this
respect offers a parallel to so/to. But how can so/to then suddenly be
soluble?

> >
> >Now, the nominative masculine to which *sah2 would be formed
> >by pure analogy would be expected to be *sos.
>
> No it wouldn't. The adoption of *so into the *to-paradigm occured
> at the same period of time as the adoption of the marked
> nominative. So *so fails to be explicitly marked by *-s because
> it occurs during the adoption of this marker.

You don't know any of this of course. And even if you would happen to be
right in your guessing, the form *so would indeed be expected to be *sos
if it is confronted with the fem. *se-h2.

>
> Another reason is that it would be both phonetically and
> grammatically redundant. It would be like saying "this-this" or "thith"
> instead of just "this".
>
> And did I not mention that *so was UNdeclined and that the feminine
> *sax is not from the earliest animate-inanimate stage of IE? The
> feminine has no bearing here and so *so is quite clearly undeclined.
>
>
> >What we expect is *to-s. I therefore assume that [...] the old regular
> >form *to-s was assimilated to *so-s.
>
> I will turn the tables on your double-standard thinking: "This is not a
> system the language shows us elsewhere, therefore it is as inane,
> baseless and groundless as your impression of anything you believe you
> have to fight."

Why would spontaneous assimilations have to be excluded from PIE? All
other languages have some, and this is pretty much what they look like.

>
> We do not see this assimilation elsewhere nor is it an inevitable
> conclusion.

I am very sure that if a series of sound changes of the same (modest)
magnitude as *tos > *sos > *so could make regularity out of be/am/was,
then it would be generally accepted common knowledge which would hardly
be doubted by a soul.

>
> Yes we do expect *to-s, but it was replaced by _another_ _seperate_
> stem, the undeclined *so -- A more straightforward self-evident theory
> because we don't need to further explain why *so has no *-s, or why it
> has *o (It simply does), or why **tos assimilated in this one special and
> highly speculative case.

The spontaneous changes that will have to be assumed for *so, *sah2 are
not so far-fetched if it is considered that these two forms of the
paradigm - and only these - were proclitic, while the rest of the paradigm
is accented. That's the state of affairs in Greek, and also the Toch.
fem.sg.nom. shows a reduced form which is commonly explained from
proclisis (an analysis I know from Werner Winter). Proneness to phonetic
interference à la distance is of course greater in reduced speech.

You can't know that *so just happened, giving you a free ride in your
darling idea of identifying the nominative marker with the stem of a
pronoun that itself escapes your morphological control. If the slight
one-time deviation from our expectations I have set up to make expected
*to-s, via *sos, into the known form *so is essentially right, then your
entire concept of the morphology of IE is erroneous.

>
> Conclusion
> ---------------
> Jens continues to respond with unfocused rebuttals, full of double
> standards, half-truths, probability-twisting and a penchant for the
> more absurd over the mundane. Will it end? Stay tuned next post...

Why stop when there is still so much to be done?

Jens