Re: [tied] Re: IE lexical accent

On Thu, 15 Jul 2004 enlil@... wrote:

> > There are many asyllabic endings in IE, [...]
>
> No kidding? And yes I derive them all from syllabic endings. There
> are etymological considerations that can't be ignored and they show
> that the endings were once syllabic. You ignore it so I guess we will
> have to agree to disagree. To me the etymologies of these suffixes are
> painfully clear but to each his/her/its/he-she's own.
>
> These particular endings may violate Suffix Resistance but they are
> regular under Syncope. Also it's transparent that the endings I've
> identified are the most common suffixes around in the language. They are
> naturally more prone to erosion just like in any other language.

Once syllabic - sure, but when was that? You could construe the same
argument on the basis of Sanskrit and claim that Proto-Indo-Iranian had
syllabic endings where Sanskrit shows -s, -m, -:/-i with the noun and -m,
-s, -t with the verb. You do not do that because you know it's wrong since
the same asyllabic endings are reflected by Iranian. You could then
construe the argument that any prestage of Proto-Indo-Iranian would have
to be posited with syllabic values for these morphemes, but again you
don't do this because the other branches also demand short forms without
underlying vowels and thus show that the argument is wrong. Now, as soon
as you are out of radar range where comparative linguistics cannot control
you, you shout, Here it is, there must be syllabic forms here! But nothing
tells you that the prestage you cut back to when rolling back the few
changes we call ablaut is different from PIE (and Sanskrit) in this
particular sense. It may be, but it may also not be. The simpler solution
is of course that it isn't, for that would mean one change less, but that
is more of esthetic value. What could really decide it would be evidence
from the closest relatives within Nostratic. It seems to me that the
flexives that do have credible counterparts outside IE have no more vowels
in Eurasiatic than they do in IE. That should mean that the vowelless
endings of the IE strong paradigm forms have not lost each lost a vowel by
the development of IE zero-grade. If they must of logical necessity be
derived from full syllables then they must have been reduced to single
consonants in an even older period than any to which we can assign
anything we can call ablaut in IE. And that reduction may then well belong
to a common prestage of a larger family.

> > So *-e is a *marker* of irrelevance? And the *-e means "don't bother to
> > look for the object"? That would be an antipassive;
>
> Wrong. Please listen properly. Think about how the language would shift
> from a contrast of intransitive/transitive to active/stative for a minute.

I'm listening and thinking, and I've got more than a minute. But it
becomes hard already at this point: IE is intransitive/transitive, and
some say it was earlier active/stative. Why am I to envisage a change from
what IE is to what it perhaps used to be? Still, I'll do it.

> There are actually two possibilities, similar but slightly different,
> when dealing with the origins of the perfect but I'm sure the perfect
> and stative are linked. The question is whether a seperate stative
> conjugation can be posited for the latest form of IE or not. Either way,
> the IndoTyrrhenian stative 1ps was *-ah (whereupon we must expect IE
> *-x as a result) and the perfect was *-ah-e (> *-xe). It is the perfect
> that is given the added *-e and this ending cannot be called an
> antipassive since we'd expect an _imperfect_ nuance to come from it then,
> not to mention that IndoTyrrhenian is an accusative language, not
> ergative, for which antipassives would be unlikely.

Perfect *and* stative? So the two are separate categories? I'm curious to
see what you may mean by stative if not the perfect. I am completely
convinced there were no such two separate categories in PIE, but I'll keep
an open mind (though I'm sure I have already been over whatever comes). I
stay unimpressed by any argument based on Etruscan. I have seen no serious
argumentation for its being a closer relative of IE than the branches
Greenberg assembled as Eurasiatic (leaving out Nivkh and Ainu however).
Experts of both IE and Etruscan I have spoken with declare themselves
unable to see any serious connection. Sadly, I cannot make them go public
on the matter. I also fail to grasp the terminology underlying the curious
statement that an antipassive interpretation of the perfect would make it
an imperfect (sic?). If this is a wrong reading of your wording (most of
my reactions to your statements have been just that you have said) you
should perhaps change it into something more comprehensible.

>
> The *-e is a _transitive_ marker. It makes an intransitive a transitive.
> Since I can tell you still don't get it. I'll go step by step very
> slowly...

I think you mean it changes an essive into a possessive. That may have
something going for it.

>
> Stage 1: transitive / intransitive
>
>

This is the earliest form of IndoTyrrhenian. Here, the

> intransitive verbs are still marked by the set *[-a-h, *-a-t, *-a]
> while the transitive set is marked by *[-e-m, *-e-s, *-e].
> The accompanying "thematic vowel" may be understood as not
> technically belonging to the suffix, although the vowel is
> necessary nonetheless to avoid an invalid syllable shape. It
> must also be noted that transitive had *e-vocalism in the
> verb stem while intransitives had *a-vocalism.

You've lost here already. If this is in any way different from what we
find later on, should it not be given some motivation, if only as a
rhetorical pretext? You cannot expect a simple fiat to be accepted.

What does "the earliest form of IndoTyrrhenian" mean? If there was such a
thing as IndoTyrrhenian it would be defined by the stage in which it
split up and ceased to exist. All prestages of IndoTyrrhenian could be
called IndoTyrrhenian, so what in the world could the "oldest stage" of it
be? The first structured language of man? The oldest prestage of IE and
Etruscan that is not part of any other known family?

I do not accept that you have any basis for statements about what would
be "an invalid syllable shape" for a language which you can only reach in
this very indirect and imperfect way.

I see no basis for the assumed difference -e- : -a-, let alone for their
distribution as transitive vs. intransitive. I sense that your -a- is
later to emerge as the thematic vowel, but that is not a marker of
intransitivity anywhere that I have heard of.

> Stage 2: transitive => objective
> intransitive => subjective trans/subjective intrans
>
> So the first shift is a transition from a language that
> is concerned with the presence or absence of an object with a
> verb to a language that elaborates on whether the object is
> definite or indefinite/absent. This would be precisely when
> *-e was attached to the subjective. Think about it for a
> second. If the transitive should become the objective, there
> is no dilemma because an objective focuses on the object and
> the transitive always has an object. The objective verb implies
> transitivity. However, a subjective verb may or may not be
> transitive. So the subjective transitive was derived from the
> former intransitive. In the 1ps we obtain *-ah-e (with *-e
> being nothing more than the 3p oblique of *ei).

Why would the *-e be attached to the subjective which (in your dream
world) does not have a definite object to mark? It could not be used with
a definite object you say. Then *-e would be the marker of an *indefinite*
object. There seems to be only one, so it does not really say much about
the object, as its number or person. It apparently (as your story goes)
only says that there is an indefinite object. Where do you know all this
from? I see no basis for these statements at all.

>
> Still don't get it? Well, look at this example, you stubborn man,
> and stop thinking about the next jab to type:
>
> 1. */bara/ 'He carried' (intr)
>
> 2a. **/bara kewanata/ 'He carried (intr), a dog (partitive)'
> Sentence is disjointed between an intransitive
> verb and an unexpected indefinite object.
>
> 2b. */bar(a)-e kewanata/ 'He carried (tr) a dog'
> By adding the the oblique *e at the end, the verb
> phrase now announces the presence of an object that
> is naturally marked in the partitive when it follows.
> As a result we have an indefinite object with topical
> focus placed on the subject without disjointedness
> of these intertwined verb and noun phrases. And
> statives of course are intransitive and thus subjective
> so they are marked simply by *-ah in the 1ps.

I cannot accept a form "/bara/" or the statement that it would mean 'he
carried'. No basis has been offered for such an assumption.

>
>
> Stage 3: objective => imperfect active / imperfect stative
> subjective => perfect active / perfect stative
>
> An extra dimension of active/stative developped cleaving the
> original subjective/objective contrast into a box of four cells.
> Subjectivity also shifted in semantics such that a focus on
> the definiteness of an object translated into the definiteness of
> the verb, or rather the verb's single or abrupt event. Object
> indefiniteness consequently was easily handled by case marking
> alone anyway (through the means of either animate accusative *-m,
> although unmarked for inanimates, or the partitive in *-ata).
>
> More recognizable names for the four categories mentioned above
> are durative (imperfect active), aorist (imperfect stative),
> perfect (perfect active) and stative (perfect stative). Both
> the durative and aorist had their slightly different *m-sets
> while the perfect and statives had their slightly different
> *x-sets.
>
> I suppose that had the stative continued on into IE, we could
> easily get this four-way system to collapse into the mi- and
> hi-classes in Anatolian while also collapsing into a 3-way
> durative-aorist-perfect system in the rest of IE by simply
> merging the perfect and stative together.

You apparently operate with the following categories:

durative = imperfect active
aorist = imperfect stative
perfect = perfect active
stative = perfect stative

This does not look right. Anatolian has its hi-conjugation, and we cannot
find the counterpart of the perfect. The Rest has its perfect, and we
cannot find the counterpart of the hi-conjugation in it. What happened to
the quest for simplicity here? It is not even difficult to achieve. In my
view the hi-conjugation is a special turn taken by the perfect. If I were
to give myself the leeway with which you have operated in the above I
could imagine a thousand ways in which this could be correct. As long as
there is a way (and we need only one) it is not advisable to separate the
two.

The aorist could hardly have been assigned a less adequate functional
description than the one you have given it. The aorist is used about
punctual events. It is purely a matter of semantics whether a given root
forms a root present (which it does if it means something durative) or a
root aorist (which it does if it means something momentaneous). In both
cases the subjunctive structure can be transferred to the use as a present
indicative. There is no underlying structural difference in stem formation
between a root present and a root aorist.

Your system has no place for the middle voice of the present (aka
durative, aka imperfect) and aorist. Still, these categories are just as
fundamental as the rest.

One also wonders on what basis the perfect has been called active in
opposition to a definition of the ancestor of the hi-conjugation which is
put down as stative. I could better understand the opposite although that
would not be very adequate either.

>
> > What in-betweens moved them to desinential position in Indo-European?
>
> In a nutshell: Changes in word order. Seperate evolutions of Nostratic's
> ergative and largely analytic natures in the diverging Nostratic
> language groups must be recognized and finally detailed in more depth
> than has been done up to now.

If this is to make any sense, that indeed needs to be done.

>
> In AfroAsiatic, the absolutive forms of the pronouns were prefixed
> to form the stative conjugation. This is easy to understand considering
> that verbs with absolutive subjects are intransitive and intransitives
> can evolve into statives because statives are, well, intransitive anyway.
> Kartvelian used the absolutive and ergative contrast to mark the
> transitivity of the verb naturally, again through the use of prefixing.
> This would suggest a primarly SVO or SOV word order in Nostratic.

I am sure this is part of the truth. However, the account has not
established the connection if the elements come out in a different order
than the one it postulates. If such crude errors are accepted as
inconsequential what are we to do with the details we try to get right?

> The other languages are a seperate group. Sumerian first of all
> split away. By the time Elamite had split away from Proto-Eurasiatic
> as it headed towards Central Asia, both ergative and absolute pronouns
> were _suffixed_ to the verb stem to mark transitivity. It was probably
> at this point that ergativity shifted towards accusativity as well.
> Dravidian diverged next and opted to generalize the nominative form of
> pronouns (formerly absolutive) while discarding the oblique ones
> (formerly ergative). The Proto-Steppe language that arrived to Central
> Asia by about 9,000 BCE however chose to discard the absolutive set of
> pronouns while generalizing the oblique forms to the nominative, hence
> Proto-Steppe *mu and *tu versus Dravidian *ya:n (< *u-n) and *ni:n (<
> *nu-n). Proto-Steppe also kept the intransitive-transitive conjugation
> system intact (transitive *-im/*-it/*-i vs intransitive *-uh/*-un/*-u).
> This latter system is the basis on which I further explain the origins
> of IE's system.
>
> Those are the in-betweens.

I would like to see a version stripped of the fictional parts. It would
probably be very short.