Re: [tied] *g'(h)- > d as aberrant outcome

Miguel Carrasquer wrote:

> On Sun, 09 May 2004 09:05:37 +0200, alex
> <alxmoeller@...> wrote:
>
>
>>> Why not? *(h1en-)joh3s- > pre-Alb. *(n-)gjes- regularly.
>>> From this are derived Alb. (n)gjesh- (s > sh) and Rom.
>>> (în-)ghes-.
>>
>> that will mean in the time of Roman empire the word should have

been

>> already *gjes at least. And a such form should have had the output
>> "g^e" in Rom ( with lost of final consonant too ).
>
> What final consonant?

"s" if the Albanian form was *gjes

>> "viézure", "mázãre" do not have the stress on the "suffix" as

well;

>> however, the words ares considered to be a compositum with
>> uffif -alle, -ulle ( in Alb. both words being suffixed with

"-ulle")

>
> Exactly. Neither word ends in -urã.

-ure = -urã

>
>>> In the third place, gh- should
>>> have palatalized in Romanian before a front vowel (or do you
>>> think that PIE *gh and Romanian <gh> mean the same thing?).
>>
>> No, I consider that the presence of "h" did not allowed the
>> palatalisation of the velars;
>
> Rubbish.

that "rubbish" does not explain why I should be wrong.

>
> No. /a/ gives /â/ before /n/, and /jâ/, /âj/ > /i/ (as e.g.
> in excambiare > *scâimba > schimba, anima > *âinima > inimã,
> etc.).

hmmm... again that?
exambiare > *scâimba with methatesis mbi > imb ? if not methathesis
then the second problem is the loosing of "i" in the group "-mbi";
anima > inima; the change here is not via "âinima" at all, at least
Rosetti so far I remember. The change of "â" to "i" is explained in
this case due the next "i" which made the "â" to open to "i", thus:
anima > ânima > inima, no *âinima in this bussiness.

NOTE: Rom. "âmbi", Alb "mbi" Latin "ambi"; theree is no change of
"âmbi" to "imbi". On this basis one has to see critical even the
explanation of anima > inima with the "â" > "i" due the next "i".
I will conclude there is as well as very little evidence that
/ja/,/âj/, /jâ/( the last one "jâ" is unknown in Rom.) show the
change to "i" form these examples. With indulgence one consider the
o.m. examples as evidence.

>> to sum up:
>>
>> against IE *ghend speaks Greko-Albanian "gl'ëndërë" which point to

a

>> previous *gl-
>
> Yes.
>
>> against Latin *glandula or Romanian *glandura speaks Alb. "ë"

(Latin

>> "u" or Rom "u" rendered as "ë"?)
>
> No. The normal development of /aN/ in Albanian is Tosk ëN,
> Gheg /ãN/ (dhëmb/dhãmb, llërë/llãnë, etc.), so that proves
> glan-. The unstressed final syllables give regularly
> Albanian -ërë (cf. for -u- > -ë- Alb. *medhur- > mjedhër,
> swek^uro- > vjehërr, smek^ru- > mjekër).

If we have regular "u" > "ë" in Alb., then the change is regular. What
I miss here are the Latin examples which show this change. I assume
"swek^uro" and "smek^ru" are not Latin examples:-)

>
>> against Latin "glandula" speaks Romanian "in" where one expected
>> "ian" (*ghiandurã instead of ghindurã)
>
> No. The normal development of glanC- in Rom. is (gl^ân- >
> ghjân >) ghin-.

No. just the first part: ClanC > ClânC is true if first "C" is not an
"g" or "k"; if first "C" is an "g" or "k" then the output is "KiânC"
or "Giank" with "clear" velars and the " do not reduce to "i". I
assume plangeo > plânge should show the rightness of this change^since
"â" remains "â" and does not become "i" kind of "*plinge".

unfortunately, there are not too many examples to proove your
argumentation; there is in Latin just "glans" which has this requested
form and no other words as *glanC, *glenC- *glinC- which will allow us
to trace back your derivation. The group "cl-" should act in the same
manner as "gl", thus let see if there is any "*clanC-", "*clenC-" or
"clinC-" in Latin which will help out:
-there is Latin clango, clanculo, clingo, clino
it seems to me none of them has any Romanian output thus , the
derivation is hard to trace back; that is: you can argue that "glanC"

> "ghinC" since there are not other examples to dissaprove it. But

this remains an "oneword"-rule.

Alex