Re: [tied] Enclosed Places

13-07-03 13:23, Richard Wordingham wrote:

> I only reported on the words Pokorny derives from PIE *gHerdH- and its
> ablaut variants. Piotr supports the alternative interpretation that there
> was no *gHerdH, only *g^HerdH, and that the forms in Satem languages that
> indicate *gH arise from borrowing, ultimately from Germanic.

I wouldn't die for this hypothesis, but it is certainly worth exploring.
The "garden" etymon is a very messy affair.

> How far back we can push the loan into Balto-Slavonic? Presumably it could
> go back to before the break-up of Satem.

It depends on what how we analyse the Germanic word. If *gardaz <
*g^HordHo-, the loan could be very old. If < *g^Hor-tó- (cf. Lat.
hortus, Celtic *gorto-), it would have to be post-Vernerian, and
borrowed either independently into Baltic and Slavic, or into Baltic via
Slavic. This would mean that Lat. urbs, Phryg. gordo- (for which the
meaning 'city' is more or less conjectural), not to mention Skt. gr.ha-,
may not be closely related to *gardaz and perhaps we should start
looking for alternative etymologies.

> In which case, could the Albanian
> form _garth_ 'hedge' be inherited from Common Satem? Pokorny does propose
> Indo-Iranian cognates, viz. Sanskrit gr.ha- 'house' and Avestan g&r&a- m.
> 'cave (as some sort of dwelling)'. Moreover, I don't recall anyone
> contradicting Sergei when he said, Albanian _garth_ is a native word (
> http://groups.yahoo.com/group/cybalist/message/14192 ).

Opinions vary. Cimochowski, who did not hesitate to treat <daltë> as a
loan from Slavic, wrote that <gardh> was "incontestablement hérité". But
he was merely echoing Jokl's opinion, and did not offer any arguments of
his own. What worries me is that I can't see how one could prove that
<gardh> is inherited (as would be the case if we had a Satem reflex in
Albanian). If we reconstruct the PIE word as *g^Hor-to- rather than
*g(^)HordHo-, we are _forced_ to assume a Germanic "diffusion centre"
from which the word spread across half of Europe.

> Is there any reason
> that this (or its Dacian cognate) need not be the source of the Romanian
> word?

See above.

> I did not quote the Slavic forms in /z/ (derived by inheritance from
> *g^HerdH); I quoted forms in /z^/. Alternations of /g/ and /z^/ occur in
> Slavic morphemes to this day.
>
> At what point did vowel gradation cease to be productive in Slavic? Or is
> it still productive today?

No longer productive in roots, though the alternations are still pretty
transparent. As for the consonants, the alternation *g/*z^ was still
alive at the time of early borrowing from Germanic. In some loans we see
velars palatalised before front vowels according to the pattern of the
"first palatalisation" (i.e. *g > *z^ rather than *dz), e.g. *z^elsti
(*z^eld-) 'pay a fine, compensate' < *geld-an- (Eng. yield, Goth. -gildan).

> (In English, it survives only in verb
> paradigms.) The question I want to ask is, 'Do Proto-Slavic *gardU and
> *z^rdU have the same morpheme?', but I'm not sure whether that is
> answerable.

To be precise, they are *gordU and *z^IrdI. A similar question can be
asked of Germanic: are Eng. <yard>1 'court' and <yard>2 'stick, rod'
related or not? If one follows Pokorny and lumps the the latter with
Goth. gazds and ON gaddr, then it must derive from *gazdjo: (with WGmc.
rhotacism) and has nothing to do with *gardaz. On the other hand, one
would like to relate <yard>2 to Slavic *z^IrdI (the meaning is
practically the same). I can't think of a satisfying solution at the
moment. I hope you can see what I mean by "a messy affair".

> After Gothic, the earliest and best documented Germanic language is Old
> English, and there is no trace of the 'hedge' meaning in recorded OE, but
> yet the 'hedge' meaning is what we find in German 'Zaun'. Thus the survival
> (or creation) of the meaning 'fence' should not be so surprising.

Note that although <town> means 'town' and nothing else in Modern
English, the dialectal verb <tine, tyne> (Kentish <teen>) < OE ty:nian <
*tu:n-jan- was found with the meaning 'enclose with a hedge or fence' as
late as the 19th century (see the OED entry). This is a very close
parallel to Slavic *gordU vs. *gorditi.

Piotr