[tied] Re: h1,h2,h3 in Albanian

From: alexandru_mg3
Message: 42614
Date: 2005-12-22

--- In cybalist@yahoogroups.com, Piotr Gasiorowski <gpiotr@...>
wrote:
>
> alexandru_mg3 wrote:
The long form
> > was resulting from a lengthening e>e: after the laryngeal H was
> > lost: *bhe:r(H)g^-o (so you made 'a confusion' the source of e:
is
> > nor eh1r > erh1 as you suspected but e>e: caused by H>zero)
>
> This doesn't explain BSl. *-re:- (_not_ *-e:r-) or IIr. *-ra:-.
> Secondary full grades of the type CeRC instead of CReC are a dime a
> dozen. It was the "samprasarana" alternation CReC/CR.C that came to
be
> regarded as aberrant in post-PIE times as opposed to "regular"
> CeRC/CR.C, so if we see CeRC attested beside CReC, the latter
variant
> normally proves to be archaic and the former innovated.


I still maintain e>e: and not eh1>e: here, but I will check
further (based on your indication) if CReC is the archaic variant



> > There is no 'original' & in Albanian (even it was suspected, if
> > you know one please post the example) so in that 'ancient' time
h3 >
> > o or zero.
> > The Methathesis ur~ru,ul~lu,ir~ri are well reflected in
Albanian
> > so I don't see any issue to add ar~ra for rH clusters : *prh3-wo
>
> > *pra-wa > *par-a [based on rh3>ra>ar here]
> >
> > The Methathesis was trigerred by the non-stressed position of
> > rh3>ra
>
> It's all ad hoc. There's no evidence for a "quite recent"
metathesis of
> *ra > ar in Albanian.
[...]
> > Please note also the Fact that we don't have rr in Albanian
parë
> > shows us that the original form was *pra-wa and that the
Methathesis
> > is quite recent...
> We don't know if *rw > rr in Albanian. <arrë>, which is sometimes
> offered as an example, may be based on any kind of productive
> suffixation, e.g. *h2ar-nah2.


It's not ad-hoc : it's all related to the presence or r and not rr
that we have in : i parë (Rom. "i" primu' :))
The proof for *pra-wá is the single r here : first because rw > rr
is well attested despite your doubts (and appears not only in arrë)
(but 'we know' also lw > ll and usually both appears in parallel)...
And, and "here's the rough..." for you..."that"....in Addition
intervocalic -r- > rr (together with intervocalic -l- > ll (and (not
on the topic) with the Romanian rothacism -l- > r)).
So if you "will try to avoid" rw > rr (that is obvious in my
opinion) is not enough => you will need next to avoid also "V-r-V >
rr" ...that is 'too much' (or in the best variant for you, you will
discover a new rule)
So either you will explain the r here or your solution is not
ok...





> > Please note also the Fact that we don't have rr in Albanian >
> or *bHr.h1g^-ó- was only reshaped in a: in PAlb (like many other
o-
> > stems)
> > X-ró- adjectives&nouns means 'like X' so I don't see any -ró-
here
>
> You wouldn't see it in Albanian. My suggested dissimilation would
have
> happened already in PIE. Don't assume that whatever I write must be
of
> direct relevance to Albanian/Romanian problems. Sometimes I just
want to
> understand the PIE form itself.


I really assumed that you have talked about PIE not about PAlb
because otherwise I would have been said that "r is not lost in that
position" (...please don't consider me 'more idiot' that I am :) )




BTW, X-ró- doesn't mean 'like X'; the
> suffix forms deverbal adjectives, and is an alternative to (or even
a
> variant of) *-nó-. Cf. *h2r.g^-ó-/*h2r.g^-ró- for a similar case.
>
> Piotr
>

PIE suffix "-ro" is not a variant of "-no", Piotr, semantically is
very different ...and I still maintain that -ro means "like-X/similar-
with-X" even in a verbal context (doesn't matter the context in
fact) ...Please put some examples/contra-examples here to check
together if I'm right or wrong...


Marius