Re: Diphthong Distributions

--- In cybalist@yahoogroups.com, "tgpedersen" <tgpedersen@...> wrote:

> > I believe the received wisdom is that PIE *dh, *d, *t > Hittite

t.

> > This doesn't look like that at all.

The received wisdom is indeed that, in initial position. Internally,
we have *d/dh > -t-, and *t > -tt- by Sturtevant's law. There are of
course also the special rules that *ti > zi, while *dhi > ti, and
apparently *di > si (Sius < *diews), which must have been separated
before the merger.

When we find a statistic preponderance of ti-/te- and da- (as in
tehhi, daitti, da:i), it is at least in part a matter of pure
graphics. The signs TA and DA are homophonous in Hittite, but DA and
TI are each one wedge simpler than TA and TI, and TE is very much
simpler than DÉ (DE is not used by the Hittites). Therefore,
the "default" way of writing /ta/, /te/ and /ti/ in Hittite is by
the signs DA, TE and TI (but see below).

> >
> > Now it's known that PIE *o > Hittite a and that PIE *dh,* d, *t
>
> > Hittite z before *e, *i (note the ske-verb *deh1-sk^- > zikk-).

The z- of zikkizi is not from *dhe- or *dhi-, but from dental + /s/,
the general preform being *dhH1-sk^é-ti. The first <-i-> may be
silent.

> > Therefore all those forms that have da- must come from *dhoh1-
> > (o-grade) and those in ti-, te- must come from *dHh1- (zero-

grade)

> > where the laryngeal protects the dental from affricatization.

That, then, is not necessary.

> > I think Hittite added an -i- to the verb stem (this is a classic
> > PIE long-vowel verb in CV:(i)-).

That is surely right. The -i- is in my opinion due to analogy with
long-diphthing roots like *speH1y- 'thrive' which in part had an
allomorphy *speH1y-V-/*speH1-C/*spH1i-: ispa:i, ispiyanzi. That led
to da:i, tiyanzi.

> > I can't understand why neither Oettinger nor Jasanoff has
> > something to say about this distribution of root initials
> > (Oettinger has a chapter on it which makes no sense to me).
> > If my analysis is true, we should give up the idea taken
> > from Sanskrit that present sigular has gun.a, pl. zero-grade
> > of the root and instead posit the root vowel like this (but this
> > is the hi-, not the mi-conjugation):
> >
> > 1sg zero
> > 2sg -o-
> > 3sg -o-
> > 1pl zero
> > 2pl -o-(?)
> > 3pl zero
> >
> > cf u-HI/au-HI//aus^-MI "see"
> > And what is more: the initial root-consonant alternated
> > (here d/t/z) with the ablaut. Now I have all the way believed
> > that this must be the case; as soon as ablaut alternation was
> > there, it must have started eating away at the preceding

consonant,

> > causing confusing alternation, which made people generalize one
> > consonant or the other, thus creating what to us now looks like
> > a consonant shift, as in kentum/satem, or, in this case,
> > decem/taihun.

This is all a gross over-interpretation of the facts. Tehhi and uhhi
reflect a simple sound change ai > e: before h. Kimball has a fine
chapter on diphthongs. She says they are basically retained before
resonants and s, 2sg daitti and autti being then analogical.

Still, it does look like a "consonant shift" of sorts when Oettinger
points out in his Stammbildung that in the oldest texts word-initial
t-/d- and k-/g- are etymologically surprisingly reliable. The later
default writing is with t- and k-. From my years in Erlangen I
remember hearing talk of a Hittite sound change gi- > ki- and di- >
ti-, this accounting i.a. for the tenacious ki-(i-)ir 'heart' and
tianzi 'they put'. The observation is Eichner's.

I wonder what one should make of the etymologically "correct"
spellings. The best I can think of is historical orthography going
back to a time before the merger was complete in initial position.

Jens