Hallo!

On Sun, 21 Dec 2003 14:14:46 -0500,
Rob Haden <[log in to unmask]> wrote:

> On Sun, 21 Dec 2003 15:38:20 +0100, =?iso-8859-1?Q?J=F6rg?= Rhiemeier
> <[log in to unmask]> wrote:
>
> >The Black Sea Flood happened around 5500 BC; within the next 500 years,
> >the neolithic farming tribes had reached the Rhine.  The flood must
> >have displaced a large number of people.
> >
> >My hypothesis of how this is connected is that the language spoken
> >by the refugees was not PIE proper, but an ancestor of the latter.
> >PIE proper is one of the languages that evolved within the large
> >area populated by the refugees, and I place it in central Ukraine
> >shortly before 4000 BC.
>
> Ah.  Perhaps the Black Sea refugees spoke languages which formed the
> substrates of later IE languages, once the Kurgans had invaded Europe?

Likely.  BTW: My conlang family Hesperic (formerly known as Q)
is meant to be one branch of this "wider IE" family.

> >Yes, I have heard that, too, especially with regard to the Balto-Finnic
> >languages.
>
> Yes.  Many dialects of spoken Finnish still preserve final vowels, in
> contrast to their more "progressive" neighbors, namely Estonian and
> Karelian.

Estonian is essentially Finnish without final vowels.

> >> Since the people of northeastern Europe and Northern Asia are presumed to
> >> have been (semi) nomadic hunter-gatherers, there would have been much
> more
> >> contact between disparate language groups than if they were sedentary
> >> agricultural people.
> >
> >Yep.
>
> Thus I think that the thesis posited by Marcantonio (et al.?), that there
> is no Uralic family proper, but a continuum of dialects and areal
> isoglosses via contact, is probably true.

Hmmm.  I know too little about Uralic to make a judgement here.
Diffusion likely happened within the family, but there seem to be
similarities (especially in morphology) that are not easily explained
by that.

> >Under the theory I posted, pre-ablaut *i and *u give the same results
> >as pre-ablaut *ai and *au.  I see little reason for a typologically
> >unlikely system /a ai au/ without true high vowels /i/ and /u/
> >when one can have a more natural /a i u/, either with or without
> >the diphthongs /ai/ and /au/.
>
> Even in languages which do not recognize vowel length as phonemic, stressed
> vowels are lengthened.  Thus, it is certainly possible that stressed vowels
> in a hypothetical 3-vowel Pre-PIE became lengthened.  Then, a change
> similar to the Great Vowel Shift in English could have occurred,
> whereby /i:/ > /ai/ and /u:/ > /au/.

Exactly.

>        The situation is more ambiguous
> with /a:/.  In later IE languages, /a:/ could become either /e(:)/ (as in
> Ionic Greek) or /o(:)/ (as in Germanic and Balto-Slavic).  Perhaps it could
> become either in (Pre-) PIE.  Furthermore, /ai/ can become either /ei/
> or /oi/, as attested in natural languages.
>
> >Why not posit *likWa?  The problem arises only because you insist
> >on a one-vowel system ***which is not attested in any language***.
>
> Well, there would not be only one vowel period.  There would be a basic
> vowel with two or four allophones, and perhaps even a contrast in height
> (/a/ vs. /@/).  There are some Caucasian languages that are like this.

There are some Caucasian languages that are analyzed that way by some
linguists.  But even those have at least allophones that sound like
[i] and [u], and I think also [e] and [o].  The analysis of those
languages as having only /a/ and /@/ is doubted my many linguists.

>      The
> possibilities are:
>
> /a/ & /a:/
> /a/ & /á/ (where the acute accent denotes either stress or tonal accent)
> /a/, /a:/, /á/, & /á:/
> /a/, /@/
> /a/, /a:/, /@/, /@:/
> /a/, /á/, /@/, /@"/
> /a/, /a:/, /á/, /á:/, /@/, /@:/, /@"/, & /@:"/
>
> It all depends on what you recognize as separate phonemes.

These systems are all equally unlikely.

> >Yes.  Paul Bennett goes as far as positing a series of dental-velar
> >doubly articulated stops (i.e., treating *tk as monophonemic),
> >but that is untenable because forms with *tk- are zero grades,
> >alternating with *tek-.
>
> Yes, as in Greek tiktô < *ti-tk-.  However, as Muke pointed out, there are
> cases of initial *tk- (and *dhgh-) in PIE.  Why aren't there any attested
> cases of *dg-?

The traditional plain voiced stops (which probably were ejectives in an
earlier stage of PIE) are rarer than the others, and two of them
*never* occur in a single root.

> It seems that many IE languages (if not most) underwent a markedness shift,
> where velar stops instead of dental stops became the most marked.
>
> >In nouns: accent on the last stem vowel in the nominative, accusative
> >and locative singular, on the ending in the other cases.
> >(This includes the especially archaic root nouns.)
> >In verbs: accent on the last stem vowel in the singular,
> >on the ending in the plural.
>
> OK, so with 'foot':
>
> Ns   pe:ds/po:ds
> As   pé:dm/pó:dm
> Gs   podés/pedós
> Ds   pedéy/podéy
> Etc.
>
> Is this correct?

The genitive and dative are correct.  The vowel in the accusative
wasn't long, I think, and the vowel in the nominatoive was only
lengthened compensatorily when the *d was lost: *peds > *pe:s.

> Certainly the declension of 'foot' suggests that there was some kind of
> vowel length distinction in earlier PIE.

I think all PIE long vowels are the result of compensatory lengthening
that occured when consonants following the vowels were lost.  (In many
cases, the lost consonant was a laryngeal.)

> I can't think of an example with verbs right now.

Nor can I.

> >This can be easily explained if one assumes that the unaccented
> >endings were monosyllabic, the accented ones bisyllabic.
> >
> >Example:
> >
> >nom. *kwán-sa  > *kwán-s > *kwo:n
> >gen. *kwan-ása > *kun-ás > *kunós
> >
> >(I can't explain why the accented vowels surface as *o, though.)
>
> Perhaps the genitive suffix was originally *-sa, and when it also came to
> denote transitive subject, the genitive of the pronoun *a-, *asa, was
> attached to the noun?

I don't understand what you mean.  The way you put it, it would have
yielded nom. **-asa, gen. **-sa, while we have nom. *-sa, gen. *-asa.

>       I think this makes sense, since the thematic nouns
> used pronominal elements in their genitives as well (due to earlier
> syncretism).

Yes, they suffixed pronominal elements to disambiguate the cases that
have fallen together:

nom. *-a-sa  > *-asa > *-os
gen. *-a-asa > *-asa > *-os + *-yo > *-osyo

But some languages seem to have used other transformations to
disambiguate this (cf. Latin thematic genitive -i:).

> >There is indeed more to thematic stems than just a stem-final *-a.
> >Some of the endings are different from the athematic endings,
> >showing parallels to the pronominal endings.  Some scholars
> >thus assume that thematic nouns has pronominal elements suffixed.
>
> The endings are actually pretty similar.  The major difference between
> thematic and athematic is with the genitives, where the former has a
> pronominal element attached.

Yes.  Miguel Carrasquer, over there on the Nostratic-L list, maintains
the position that the thematic nouns had a pronoun suffixed.
But I don't see why they should (why was that pronoun suffixed only
to some nouns and not to others?), and find that assumption entirely
unnecessary.

> >> [Afro-Asiatic conjugations]
> >
> >I don't know.  Perhaps the etymology for the 3rd person is incorrect,
> >and it is simply an endingless form to which no *ya-ku was suffixed.
>
> In my opinion, that would obliterate the entire theory.

The two other etymologies might be correct, though.  But anyway,
the prefix and the suffix conjugation probably contain the same
elements.

> >This is not impossible though the endings are clearly different
> >(nom. *-s < *-sa, gen. *-os < *-asa).
>
> They're not really different.  Both are based on /s/, and the latter has an
> initial vowel which could easily mean it was originally a genitive
> pronoun.  This would be akin to saying "the horse its saddle fell off."

Ah, now I understand what I failed to grasp above.

> >I assume an active-stative of Pre-PIE in which agents were marked
> >with *-sa and patients with *-ma, whatever the origins of the endings.
>
> As do I.  Do you see what I'm saying, though?
>
> However, suffix neuters in PIE (neuters ending in *-r, *-n, *-l, *-s) do
> not take any case marking for patient.  I think that the determinative
> endings were enough to know that they were inanimate.

Yes.  The objective case of neuters remained unmarked because it did not
contrast with an agentive case which the neuters didn't have.

> >It is even possible that the genitive case contains a trace of
> >suffixaufnahme.  The actual genitive suffix would thus have been *-a;
> >to this was suffixed the case of the head noun.
> >
> >Example: `the man's dog'
> >
> >nom. *h2anar-a-sa kwan-sa
> >dat. *h2anar-a-aya kwan-aya
>
> I prefer:
>
> nom. *xanar(a)-asa kawan(a)-sa
> dat. *xanar(a)-asa kawan(a)-aya

Yes.  The suffixaufnahme is rather wild speculation.  There is
no evidence for it.

> With penultimate accent, this would have given:
>
> nom. *xnarás kwans
> dat. *xnarás kwnái
>
> Which would give reconstructed PIE:
>
> nom. *h2nerós kwons
> dat. *h2nerós kunéi

Exactly.

> For this to work, it must have been the case that neither 'man' nor 'dog'
> ended in a vowel when the case endings became attached.  The medial *o in
> *kwons could be attributed to back-rounding earlier *a.  It is also
> possible that *ó came from earlier *á: instead of *á.  OR, *ó is a Greek
> innovation -- all the other IE branches seem to indicate *é only.
>
> What do you think?

This makes sense.  I didn't know that the only evidence for *o instead
of
*e is from Greek.  The Greek vowel system is generally regarded as
especially conservative, but sometimes that assumption might fail.

> >In the later stage, the nominative of the genitive was generalized.
> >
> >(Assuming that genitives preceded nouns in Pre-PIE.)
>
> According to Lehmann, PIE was SOV, so genitives preceded nouns.

Yes.  And that makes suffixaufnahme very unlikely.  AFAIK,
suffixaufnahme
occurs mostly (or only) in languages where genitives follow nouns.

>       However,
> this also suggests that PIE originally had postpositions instead of
> prepositions.  I have a theory on how that changed, if you're interested in
> hearing about it.

Yes.

Merry Christmas,

Jörg.