Re: [tied] Re: IE lexical accent

On Tue, 20 Jul 2004 enlil@... wrote:

> > It is true that some nominal stems which show -o(:)- in their strong
> > forms have -e- in their weak cases. However, the collateral type with
> > strong-case -e:- which is parallelled by the verb shows, in my opinion,
> > that the underlying vowel is a long /-e:-/ for both alternants.
>
> Yes, "in your opinion" (aka "my idle assumption"). There is no *e: in
> either forms so you have to dream it up in order to get your results.

There certainly is an allomorph with /e:/ in the locative *pe:d-su 'at the
feet' reflected in OIr. i:s and Alb. peùr-posh 'down'. Also, the Avestan
loc.sg. daNm 'in the house' reflects a long vowel; thus there must have
been a long-e: form somewhere in the inflection, most probably in one a
locative case-form. Other weak paradigm forms with -e- alternate with
-e:-, as *ste':w-ti and *ye':kW-r. These forms are posited by practically
everybody; what are they doing wrong? If 'water' does not comply with this
picture, then 'water' is of a special type and so not suited as a basis
for the formulation of rules that are supposed to be repeated elsewhere in
the language.

> What I do is _accept_ that *o and *e are both there and that one of them
> is original. Picking the nominative form with *o as original is the only
> sensible choice. However, we can see that *o < *a (via Vowel Shift),
> otherwise many things aren't as phonetically motivated such as
> a-Epenthesis and the thematic vowel alternations.

Everybody else also accepts that *o and *e are there, but see they are not
alone. In 'foot' you have *po:d-, *pod-, *ped-, and *pe:d-. In 'man' you
have *Hne:r, *H2ner- and *H2nr-. For some reason you pick *pod- and
*H2ner- as underlying. I pick *pe:d- and *H2ner-. By so doing I get rules
I can use in the verb also.

> As for these examples not showing *i-vocalism, there's an obvious
> reason for it. My views on syllabification reflect the known morphology
> of these words such that the syllable boundary is placed at the same place
> as the union of stem and desinence. So *pedos = /ped.'os/ and *pedi
> /ped.'i/. Since monosyllabic nominal stems are always CVC, we can never
> expect *i reflected here because we will never have an open syllable.

That is completely circular. If there are morphological reason for the
assignment of the lable 'closed' to the first syllable of *pedos, then the
rule pertaining to its development is not phonological but morphological.
However, since there apparently is no change of thematic vowel to *i in
pretonic position it does not really matter that you do not find your
pseudo-rules working at your command in this particuar respect.

>
> > The lexical accent is unaffected by the shape of the flexive and only
> > concerns the stem, but the mobility of the inflexive accent is caused by
> > the attraction exerted by an underlyingly syllabic flexive.
>
> The previous post demonstrates that you're incorrect and you fail to
> explain the entire proterodynamic paradigm as well as purposely ignore
> sensible etymologies to these asyllabic desinences that anyone else
> can see are of demonstrative origin.

Nothing shows that. It's just a mantra you keep repeating.

> > This will be modified (further down) by the possible acceptance of
> > subphonemic mini-vowels
>
> This can't work for IE. Sorry. You may get away with it in EA or
> Uralic, if that indeed is what you mean to convey when you weave
> a silly long string of morphemes together without any vowels in
> between.

For IE it would work vacuously; the inserted vowels would be unaccented
and thus disappear again. What gives you the insight that this attempt to
accept one of your suggestions is also misguided?

> > Also *ud-n-ós would come out of **wed-n-ós with shift of the accent
> > from the root to the ending due to its vowel. And *wéd-n-s would come
> > out of **we:d-n-ós if that was the old form.
>
> This comes out of that, if this, but not that... yadayada. This should
> not be a difficult word to understand. It is clearly old having the
> heteroclitic alternation. It always had accent on the genitive because
> it conforms with QAR. There's no need to drum up pretend forms. The
> only form that is acceptable in the above besides *udnós is *wednós
> which shouldn't even be double-asterisked considering that that alternate
> form is reflected in Hittite. I don't understand your reasoning here, if
> we can even call it that.
>
> Where on earth is the reflex of **wedns? Why isn't _this_ double-
> asterisked?

One might say the same about (*)*wedno's. I never adopted consistent use
of double asterisks because I found I should then put three (or four?)
asterisks on further rconstructions that are based on comparison of two
double-asterisked forms. It just got out of hand, so I gave it up. I
sometimes use it for clarity though (and then tend to get it wrong). If
you accept *wedno's you would also accept *yekW-n-o's which has a
nom.-acc. *ye':kW-r. That ought to count for something, making it a fair
assumption that strong forms of such paradigm forms can have o- or
e:-vocalism. Now I have a rule that unites the two, so I am one up on the
rest. I can make *po':d-s the regular nom.sg. of *pe:d-, and I can make
*wo'd-r the collective of *we:d-r/n-; I can also make *ped-o's and
*wed-n-o's the genitive of *pe:d- and *we:d-n-. It takes introduction of
mobility, sure, but that is something we have anyway, as in Gk. nukto's,
Avest. n&mo: ( Iran. *dmah from *dm-o's replacing *dem-s. Forms like Hitt.
nekuz and *dems- which retain the phonetically regular structures are very
rare indeed and have only been preserved in very isolated cases. However
the real rules are found in such gems.

> > Rather than have the pronoun form *so produce a nominative you now let
> > it mark a genitive, and of all genders.
>
> It would be nice if before you reject, that you understand what you're
> rejecting. I've said clearly multiple times that a particle *sa marked
> the IndoTyrrhenian _animate nominative_. It didn't in any way mark the
> inanimate, which was bear. Obviously so since we don't find inanimates
> marked with *-s in IE itself, do we. Where do you dream this stuff up?
>
> The genitive is NOT marked and NEVER WAS marked with the same particle
> as the nominative. I had said that the IndoTyrrhenian genitive was a
> _suffix_ *-ase that ultimately derived from the Proto-Steppe ablative
> particle *si. This is to be distinguished from *sa, the Proto-Steppe
> general demonstrative that the ancestor of Uralic and EA have been
> shown to have used as a 3ps marker.
>
> Pay attention.

Sure, your suggestions should be evaluated for what they are, not what
they are not. Your stories about the nominative and the genitive have kept
me busy. I understand now that the nom. *-s is in your opinion (or "idle
assumption" if you will) a "clipped *so", the nom.sg.masc. _full form_ of
the pronoun *so-/*to-, and the nom.-acc. ntr. *-d of pronouns is a
"clipped *to-", the _stem_ of all the other forms of the paradigm. I have
pointed out the lack of parallelism in this but you explicitly and
emphatically don't care. I also understand now that the genitive marker
*-os was in your opinion earlier an ablative ending "*-ase". That makes
the nom.sg. and the gen.sg. completely different in origin, just as I have
always held them to be.

Now, you have also told us a story about the nom.sg. of thematic stems in
*-os, saying that this is in origin a misanalyzed genitive. I somehow got
this backwards, it *is* hard to remember since there are no facts to
tell one how the story was, it has to be remembered by heart.

> > In the old days a vowel was inserted you say, and later speakers could
> > handle the form without,
>
> ?? I can't be bothered to repeat myself and correct you when you don't
> even pay attention. You can reread the post you're blindly responding to
> and paraphrase properly next time. A vowel was inserted in IndoTyrrhenian
> to seperate a consonant-ending stem from a consonant-beginning suffix.

I'm afraid you will have to repeat yourself many times if you want anybody
to understand anything. I have gained a growing accepting audience around
me simply by patiently repeating the same explanations and answering the
same questions over and over again. Good theories are worth that. I for
one have now ceased to ascribe the opinion to you that the morphemes of
the nom.sg. and the gen.sg. are identical. I understand that this holds
only for the assumption of the genitive misanalysis which caused a
genitive in *-os to be resegmented as nominatives in *-o-s.

>
> > A desinence that fails to attract accent may be a desinence without
> > any vowel to do it with. That is simpler by any standard,
>
> Any standard except phonotactics, etymology, morphology, accentuation,
> etc. Syllable shape is the most basic part of a reconstruction so if we
> go willy-nilly on that, we really end up with a chaotic result. You
> lack a standard by simply stringing consonants in an absurd chain in
> every language you deal with and then say that the vowels are "sub-
> phonemic". Any sensible etymologies for these desinences which strongly
> appear to be mostly garden-variety demonstratives is dismissed by you
> as invalid. What rot.

You are taking standards to new levels. You appparently want to decide on
such fundamental things as phonotactic rules by looking the other way. It
does not matter in this connection whether I accept or reject the ultimate
etymological identity of the individual flexives. What matters is how many
vowels the protoforms had during the phases of the ablaut changes. You
insist there were vowels in the short endings, but you have to cut back
all the way to Nostratic or one of its early subsequent stages, and even
then you find only subphonemic vowels or even no vowels. Now, did the
poststage of this in which the IE ablaut changes took plac have no vowels,
subphonemic vowels or full vowels? You insist it had full vowels, and I
shake my head in disbelief. That's where we stand.

> > Elaborate please. But sure, most others don't know a first thing about
> > these matters. Where are the vowels you cling to, and what proves their
> > antiquity? For EA, I wrote the only book that exists on the subject.
> > What *are* you talking about? There are plenty of anaptyctic vowels of
> > a subphonemic status,[...]
>
> Alright, I guess my philosophy is that if someone is reconstructing
> a language, they do it by reflecting the phonology. If there are vowels
> there, I should see them in the reconstruction, even if they are
> "subphonemic". Also, while I suppose it could be possible that they
> are subphonemic in EA (although I'm skeptical), I really don't see that
> being the original state of affairs. In Uralic, we still don't have
> subphonemic vowels like you say. It can be understood as having a
> simple CV(C) pattern without problems. Your theory only complicates things
> unnecessarily.

The major fallacy is in the postulation of a definite phonotactic
straightjacket. If there is nothing but overriding phonotactic demands to
motivate the presence of a vowel, then it has no independent status, for
the vowel is then a mere consequence of the consonant skeleton and so
informationally zero.

>
>
> > The inflectional accent actually reveals that. In addition, it should
> > not be just assumed without proper reason that all IE consonants were
> > once followed by vowels. We have good reason to believe that an optative
> > form like *dwis-iH1-ént had some such shape as *dweys-yeH1-ént before
> > the ablaut produced zero-grade of two of its elements; it is quite
> > another matter when some analysts run amok and posit some such thing as
> > *deweyese-yeH1e-énete.
>
> No. If you're talking about MIE, the form would be expected to be
> **t:waisa-yah-éna-ta. Of course I don't feel that the optative is
> anything but a recent conjugation anyways. The optative looks suspiciously
> like an extension of a verb stem marked with *-ye-.

As so often it is a major embarrassment to your explanations that they
have to disregard the material we do have.

> > There is no ablaut difference between present and aorist in IE.
>
> Precisely. So why should there be a difference in thematic vowel?

There isn't. The thematic form of *kWe'r-t is *kWe'r-e-t, like that of
*gWhe'n-t is *gWhe'n-e-t. The generally accepted reason we do not find the
structure *kWe'r-e-t as an aorist injunctive is that the aorist
subjunctive became a thematic present, not aorist.

> That's
> simply because the thematic vowel for the durative was *-e- and survived
> Syncope while the aorist had the thematic vowel *-a- which didn't.

This is just a funny dream of yours. You have not given any motivation for
it.

> The
> root aorist may work fine without the extra *-a-... but it still doesn't
> explain the differences with thematic vowel. The thematic vowel is just
> not a morpheme by any standards no matter how much we squint so this is
> the simplest solution possible and Syncope takes care of the rest.

So you believe there was originally an athematic present and a thematiic
present which were formally distinct but functioned the same? And also
both an athematic aorist and a thematic aorist of identical function,
which however coalesced phonetically?

>
>
> On EA, Jens:
> > Well, Eskimo and Aleut did in fact use fillers with some consonants,
> > and after clusters. But that was a nonphonemic vowel,
>
> Alright, thanks for the clarification. When you string consonants together
> without vowels I naturally assume you aren't reconstructing vowels which
> looks completely absurd and unpronounceable.
>
>
> > *aluR > PE *aluq 'footsole' (WGr. /aluq/, Chapl. aluq)
> > *aluR-m > PE *allum (WGr. allup, Chapl. alum)
> > *aluR-d > PE *allut (WGr. allut, Chapl. alut).
>
> Hmm, but I think I see a pattern. Why wouldn't it be...
>
> *aluR *nat¤R *cavig
> *alRu-m *natR¤-m *cavig¤-m
> *alRu-d *natR¤-d *cavig¤-d

The metathesis interpretation of the geminating type belongs to the
prehistory of the field by now. It is easily falsified by the fact that
the cluster it operates with exist in their own right and develop
differently. That hiwever cannot be seen ont he basis of Greenlandic or
the Eastern Canadian dialects alone, but Western material is clear. The
second and third columns are correct (except that -Ï- is not a phoneme).
Note that I as speaking of underlying forms, not reconstructions of an
immediate predialectal Proto-Eskimo stage (which certainly had *-q, *-k,
and *-t already).

>
> > I also find anaptyxis obvious for a Finnish example like:
>
> But what does Finnish have to do with Uralic and ultimately with
> older stages when there is so much time gap?

If Finnish has had syncope, how is the rule to be formulated? I do not
think such an event has been even suggested for any of the high-powered
early loanwords the family supplies.

> On the vowel of the Proto-Steppe plural that I reconstruct as *-it:
> > In my estimate it is phonemically zero. Even in later periods it
> > has not merged with any of the independent vowels.
>
> It wouldn't. The *i disappears before vowel-ending stems because there
> can only be one vowel per syllable.

That causes a philosophical dilemma. You just can't know that the vowel
has been lost where you don't see it and not inserted where you do. What
you can see is that it is nonphonemic, for the form cannot exist without
it. It does not contrast with zero.

> > There is the complication that the IE form is a nominative plural, not
> > just a plural.
>
> Yes! It was! Well, more specifically, the plural was marked in the
> nominative and accusative, the strong cases but not the weak ones. This
> has been my position for a while now. It is shown by the fact that the
> plural is wonky in IE weak cases while very much standard and more
> expected in form in the nominative and accusative. Afterall, if the
> genitive had *-os in the singular, we should expect plural **-oses right?

Wrong, of course. The genitive plural should not have a nominative plural
morpheme hanging on it.

> No, for the clear reason that the plural was only marked in the nominative
> and accusative in earlier stages... and only in the _animate_ gender,
> mind you.

I do not accept dictation.

>
> This makes sense given that the subject and the object, the foci of
> these cases, are the main characters in the sentence. Their plurality
> then will tend to be of more topic interest then the number of an
> indirect object.

That does make sense, and it may be the ultimate cause of the
idiosyncratic formation of the oblique cases of the plural. But if we do
not understand those of the singular too well either it may be of little
help. Also I consider it a wrong principle to assume that forms are
particularly young and secondary if they defy analysis; the opposite would
be more logical.

Jens