Hallo!

On Sat, 15 Nov 2003 08:53:04 -0800,
Costentin Cornomorus <[log in to unmask]> wrote:

> --- Jörg Rhiemeier <[log in to unmask]>
> wrote:
>
> > I assume that by "IH", you mean "Indo-Hittite".
>
> Yes.
>
> > Hence, Talarian would occupy a position similar
> > to that of Tocharian.
>
> Yes. Sort of off to the side - it's not exactly
> Anatolian, it's not exactly IE either.

I see.

> > Quendic (to use the provisional name of the
> > family) is even farther
> > from the main IE stock than Anatolian, though
> > closer than Uralic
> > (if the latter is related to IE at all).
>
> Out of curiosity, have you worked out any of the
> conhistory of how all this came about?
>
> In the World, the premise relies on the Black Sea
> Flood to disperse a bunch of related peoples,
> known as the Punt, that would become IE and
> Semitic speakers (at least, there may be / may
> have been some others).

The common ancestor of IE and Elvish was spoken
to the north of the lake that existed in the Black Sea
basin before the Black Sea Flood, which is part of Elvish
mythology.  It is a case of mythology confirmed by
hard linguistic facts.

According to the mythology, the ancestors of the Elves
came from a lush, fertile plain at the shore of a big lake
in the east (you may recognize the Cuiviénen story here),
which was swallowed by the sea in a disaster caused by
a war between angelic powers.  The refugees moved west
and became the various peoples of Europe, the westernmost
of them being the Elves.

What regards Semitic, I doubt it has much to do
with the Black Sea Flood.  The languages most closely
related to Semitic are all spoken in Africa.

> > [PQ cases]
> >
> > Herein, the abbreviations "AS" and "OS" mean
> > "agentive stem"
> > and "objective stem", respectively.
>
> I like that very much.

Thank you!

> > [PQ gender/number vowels]
>
> So many vowels in Talarian end up as A or O that
> any system that may have existed was obliterated.
> As it stands, there are only animate and
> inanimate nouns.

In modern Nur-ellen, the genders are no longer marked
on the noun either, as all the final vowels have been lost.
Plural is marked by umlaut, while the dual has gone.
Anyway, the masc. and fem. genders were only used
for entities of the respective natural gender:

chvana `dog'
chvano `male dog'
chvane `female dog'

The Nur-ellen pronouns have retained their PQ forms
ma `I', tha `thou', so `he', se `she', sa `he/she/it',
etc., until today, simply because a loss of the vowel
would have left behind single consonants.
They inflected regularly in PQ, and in modern N-e,
they still have distinct objective forms in -m (e.g. mam `me').

> You can still see a/e ablaut in animate nouns,
> and there are some -as/-os variation in the
> nominative singular. But otherwise, the animate
> stem is in -a-.

There is a sort of ablaut in PQ, too, but I don't
know yet where exactly it occurs.  Certainly in derivations,
perhaps in present vs. aorist stems.  The idea is
that in the common ancestor of IE and Quendic,
there were three vowels *a, *i, *u which each had
a "strong" and a "weak" grade.  In IE, the strong grades
became *e, *ei, *eu and the weak grades, *0, *i, *u
(with the strong grades re-coloured by laryngeals).
According to Gamkrelidze & Ivanov, the o-grade originally
was an alternative form of the weak grade which occured
where zero would have yielded unacceptable consonant clusters,
and was later systematized as a separate grade.  Now, what
happened in PQ was that strong and weak grades differed by
the presence or absence of the vowel feature [+open].
Thus, strong grades were *a, *e, *o and weak grades were
*V, *i, *u wherein *V represents a "featureless" vowel that
assimilated to neighbouring vowels.  The laryngeal *h2 added
the feature [+front], thus changing *a into *e and *o into *ø,
while *h3 added [+round], thus changing *a into *o and
*e into *ø.  In positions where consonants were unproblematic,
the weak grade could vanish altogether.

In addition to that, there were i-umlaut (*a, *o, *u became
*e, *ø, *y before *i) and u-umlaut (*a, *e, *i became
*o, *ø, *y before *u).  Thus PQ had seven vowels:

*a, *e, *i, *o, *u, *ø, *y, each occuring short and long.

I am not sure yet about what happened to them in the
daughter languages.  In Nur-ellen, which has the same
inventory, they are preserved unchanged at least
in some positions.

> > Only animate nouns had an agentive stem, which
> > also means that
> > inanimate nouns have no agentive, genitive and
> > dative cases!
> > The objective stem always ended in a consonant;
> > the OS of an animate noun was AS + -m.
>
> Of course, a lot of Talarian inanimate nouns are
> sensically animate (like patar, father); I've
> been knocking around for a way to allow them a
> means of being used as agents.

In PQ, all semantically animate nouns were also grammatically
animate; there were a few grammatically animate nouns of
dubious semantic animacy, though.

>     An agent stem may
> well be the answer; it's a very elegant thing
> you've got going there!

Again, thank you!

> I'd already decided that sensically animate
> inanimate nouns (like patar) have a
> nominative/agent case in -m, which is also a
> common areal feature in other languages around.
> It would be no big deal to extend the -a- of the
> stem to -â-, thus giving patrâm for a
> nominative/agent of patar. Since inanimate nouns
> always have -m for both nominative and
> accusative, I suspect its accusative/patient
> would also be in -âm.
>
> This of course would yield interesting times when
> patar would still be nominative, as in stative
> verbs or any verb that doesn't require an agent;
> while patrâm would be usual for verbs requiring
> an agent.

This promises to become an interestingly convoluted system.

> Naturally, the accent will shift from PAtar to
> [log in to unmask]
>
> > OK, you might ask what "agentive" and
> > "objective" mean.
> > PQ was an active language (as are its
> > daughters), and the agentive
> > denoted the agent of the verb, and the
> > objective the patient.
>
> Talarian is also an active language.

I'd like to see some of its grammar.  Do you have
something to present on the list?

> > Now to verbs.  The tense/aspect/mood system is
> > still quite a mess, and I won't mention it
> > here.
>
> The messier the better! I'm afraid Talarian might
> be a little simplistic.

The PQ verb system will be inspired by the IE state
of affairs as well as the verb inflections in Tolkien's
languages and the screeve system of Georgian,
though it won't be as forbiddingly complex as the latter.
I know quite well which categories exist, but I am not
sure yet about the forms.  The tense/aspect/mood categories
are:

present
imperfect
future
conditional
subjunctive
aorist
aorist subjunctive
imperative

The aorist has similar functionas as in Ancient Greek,
including the narrative tense and the gnomic aorist.

As for now, I use the following markers, which are not fixed yet:

present            -a
imperfect          -Vn (derived verb)
                   -n  (root verb; infixed before final obstruent)
future             -u
conditional        -u suffixed to the imperfect
subjunctive        -i
aorist             V- (`augment'; the root vowel prefixed)
aorist subjunctive V- -i
imperative         -a but no agentive ending

In root verbs, the imperfect marker *n was suffixed to the root
if it ended in a sonorant, and infixed before the last consonant
if it ended in an obstruent.  Thus *meln- `loved', but *dant- `fell'.

There may be a difference between present and aorist stems
(if yes, probably by ablaut), but I am not sure yet.

> > [2 sets of conjugation endings]
>
> Interesting. Talarian has Active and Stative
> conjugations, with different sets of verb endings
> and pronouns for each.

What are the forms like?

> > The suffixes are thus:
> >
> > [PQ conjugation suffixes]
>
> Cool. Related to primary and secondary endings in
> some way?

They are cognate to the IE active (present/aorist) and
stative (perfect) endings, or the Hittite mi- and hi-
conjugations.  PQ *th corresponds regularly to PIE *t
(sort of Germanic/Armenian sound shift, or rather
conservativism if one accepts glottalic theory which I do),
PQ *h matches PIE *h2, while PIE *th2 is reflected
as *ch in PQ.

> > If the object (or subject of a stative verb) is
> > inanimate,
> > the singular endings are used regardless of the
> > number of the object.
>
> Any reason why?

This is AFAIK a rather common feature, which can also be
observed in several IE languages.

> How do laryngeals fit in either PQ or modern N-E?

The laryngeals have merged into a single phoneme *h,
which is preserved in some positions, but deleted with
compensatory lengthenig of the vowel in post-vocalic position.
The different laryngeals colour vowels differently,
as I have laid out above:

*h1a > *ha, *h2a > *he, *h3a > *ho

In modern Nur-ellen, most PQ *h's are intact.

What happens with the laryngeals in Talarian?

Greetings,

Jörg.