From: Arnaud Fournet
Message: 60432
Date: 2008-09-27
----- Original Message -----
From: "Andrew Jarrette" <anjarrette@...>
>The *arH- root (*H2erH3, I believe), as in Latin <aro:> "I till,
>plow, cultivate", Greek <aróo:> "I plow, till", MIr <airim> "I plow",
>OHG <eren>, OE <erian>, ON <erja> etc. "to plow", Lith <ariù> "I
>plow", OCS <orjõ> "I plow", together with Arm <arawr> "plow", Toch
><a:re> "plow", Lat <ara:trum> "plow", MIr <arathar> "plow", Greek
><árotron> "plow", ON <arðr> "plow", OCS <ralo> (Pol <radl/o>) "plow",
>etc. (these all from Pokorny, under *ar(&)-). Absent from Indo-
>Iranian, as are many European-IE agricultural terms (explained by
>some as due to the Indo-Iranians' origin from those Indo-Europeans
>who were pastoralists on the eastern steppes east of the Caspian or
>thereabouts).
>AJ
============
This root is basically the same as Semitic z_r_& 'cultivate'.
As you can see, we can equate :
- H2 with z
- r with r
- H3 with &ayin.
NB : I'm not saying that H2 is _always_ z, nor that H3 is &ayin.
We can also see that Kartvelian 'to grow' is /zil/ with z.
This word had *z as initial.
The main issue is what happens with *z in PIE ?
The phoneme is not supposed to exist, because the correspondence for it is
not obvious.
Personally, I have come to think that :
- in western PIE (Celtic, Latin) *z is equivalent to H2
- in eastern PIE *z is equivalent to *y
- in Greek the situation is very complex ; H2, y or Z.
For example :
Arabic zimam 'bridle' : Latin amentum, : Skrt yama
Hebrew zeqeq 'purify' : Skrt yajna 'sacred'.
Etc.
It's not rare that Latin has a strange -s- in that case :
aro = sar-io < *z_r_&
sacred < *z_q-
I think that a branch of Indo-European substrate to Latin kept *z as *s.
This feature causes doublets like ab-ies = sap-in
But this is another matter.
Then, what is my point ?
If *H2erH3 from *z_r_& __were__ inherited, in the same fashion as the word
'sacred' is inherited,
we should expect an alternation between H2 and y in the western and eastern
parts of the Indo-European family,
What do we have ?
One family Indo-Iranian has no trace of this root,
The other languages all have H2, the form expected for the western half.
There is no alternation at all.
My conclusion :
Eastern languages where we should expect *y but where we find the reflect of
H2 have borrowed this word from the western IE languages.
Germanic probably borrowed this word from NWB and Celtic when it entered
Europe.
This cannot be a cognate.
As a matter of fact, Armenian is irregular : arawr reflects an irregular
retention of H between *arHtor > ar-a-wr. Hence loanword.
Next, Tocharian are does not mean 'to plough' but 'a plough'.
The verb radical is used as a noun : absurd hence loanword.
Conclusion about PIE :
When this word penetrated the indo-european family,
this family was already spread about,
For that matter, PIE dispersal must be pre-neolithic.
PIE split _before_ this word and agriculture came into being.
I hope my reasoning was about clear !?
Arnaud
========
> I am not well-informed about the Stone Age and its divisions, I
> didn't know that agriculture has been limited to the Neolithic, since
> ca. 8000 B.C. (nor do I know the reason for this -- and I am
> personally a little skeptical of the reliability of archaeology).
> Although I have supported you in your idea that PIE is much older
> than conventionally supposed (although my reasons are different, I
> think) nevertheless 8000 B.C. sounds reasonable as a possible genesis-
> time for PIE to me at least (most will say later than that, right?).
> AJ
==============
Yes, I suppose the conservative siders will vote for -4 000 BC.
But I believe none of their arguments withstand a thorough analysis.
The problem is - 8000 is too late.
PIE must have split before that time.
Otherwise a word like H2erH3 should have the alternation H2/y.
Arnaud