Re: [tied] Retroflex Series in PIE

--- Piotr Gasiorowski <gpiotr@...> wrote:

> On 2006-08-08 08:10, Sean Whalen wrote:
>
> > Since some languages distinguish between final
> *-er
> > and *-r but Hittite doesn't
>
> Which languages show an athematic neuter with final
> *-er for 'feather'?

Since many languages often replace them with
thematic forms of lose neuter/m/f distinction there's
controversy in anything I could say. I'd give
Armenian t`ir as the most likely. Are you saying
Hittite doesn't merge final er/ar?

> > it shows nothing about the
> > PIE form. Other aspects of "feather, fly, etc"
> need
> > to be explained.
>
> What aspects, namely?

As I already wrote, the alternation between t/r in
Gk/Skt. Also why *petros > Skt pat(t)ra- but OIr ette
(if no R. > t. after t., etc); why *pternos > Gk
pte:nos (irregular loss of r unless R. or some other
consonant substituted).

> >>> and *gWertí- "act of consumption".
> >> What particular cognate set points to such a
> >> reconstruction?
> >
> > It's not my reconstruction;
>
> Well, whose is it? I don't recognise it as a valid
> reconstruction at all.

In the first paragraph I described some traditional
reconstructions and then gave some things wrong with
them. I gave my own explanations below. These are
older ones because I didn't want to get into newer
ones which I think are even worse and need more
explanation of why they're wrong.

> > I had *t. > *R. in this
> > word below.
> >
> > borá: 'food' Gk; vora:re 'to devour' Lat
>
> These are unproblematically derived from PIE
> *gWerh3- (present stem
> *gWr.h3-é/ó-, cf. Ved. giráti).

Active in r. but passive in t. > R. It doesn't make
sense to derive, for example, both "die" and "kill"
from *mer- when the forms show differing changes that
can be explained by r./t., etc.

> > gerti 'to drink' Lith (not *girti), etc.
>
> This is a secondary athematic present replacing, in
> Baltic, an older one
> of the tudáti-type. The Lithuanian tone is acute,
> betraying a lost
> laryngeal (as if from *gWérh3-ti for inherited
> *gWr.h3-é-ti, the latter
> preserved e.g. in Slavic *z^Ir-oN 'I devour').

The infinitive in a few languages is from a former
noun; since "drinking" and "being drunk" describe the
same action the more common word was probably used as
the inf. to the act. even though originally pass.
(probably when sonorants voice so no r/R distinction).


The tone probably comes from V > V: before voiceless
R, etc., right before R/r merge (just as V > V: before
D not DH right before they merge).

> > gWet.tós 'swallowed, noun'
> > gWet.t.ós
> > gWet.ós
> > gWeR.ós
>
> The actual deverbal adjective in *-tó- can be seen
> in Gk. bro:tós and
> Lith. gi`rtas, both regularly derived from
> *gWr.h3-tó-

I don't deny *gWr.h3- 'swallow down' vs *gWeR.- 'be
swallowed'.

> <borá> is the same formation as <tomé:> <
> *tomh1-áh2. Both the o-grade
> and the final accent are expected in the whole type
> and don't require
> any special explanation.

Why would "devouring" and "being devoured" have the
same form? Why no o>u in Gk? How can you know what
accent to expect when Gk and Skt data have different
accentation (most of) which is unexplained? Also some
data (often difficult-to-discern traces) from other
languages with more complex changes make it impossible
for IE languages to be derived from one with a maximum
of one tone per word.



__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com