Re: [tied] Thracian place-names

George Hinge wrote:

> It is not necessary to date the metathesis to Old Iranian. There may
> have been a tendency towards metathesis in many different dialects of
> the vast Scythian-Sarmatian dialect continuum. The sources of the
> Scythian dialect of the 7th-3rd cent. BC are too few to exclude the
> presence of the metathesis here. However, in the Iranian personal
> names of the Greek inscriptions from the Northern Black Sea Coast
> dating to the Roman age, there are numerous examples of metathesis,
> e.g.:
> - Phurtas (CIRB 101, 1278, 1282, 1283), Purthaios (IOSPE 1(2).43, 83,
> 99, 130, 176), Purthakes (IOSPE 1(2).86, 101, 102) = Ossetic fyrt,
> furt "son", Avestic puþra-
> - Sorchakos (CIRB 1282) = Ossetic syrx, surx "red", Avestic suxra-
> - Pharnoxarthos (CIRB 1245, 1282, 1286) = Ossetic (æ)xsar "power,
> strength", Avestic xšaþra-
> Perhaps this isogloss was primarily "East Scythian" (around Don, =
> Sarmatian), whereas "West Scythian" (around Dnepr) retained stop +
> liquid. We have examples of metathesis in Olbian names, it is true,
> but they are due to eastern influence then.

That _is_ possible. I'd just like to point out that this highly
characteristic metathesis (also in clusters such as *-k^w- > *-sp- >
Oss. -fs-) is entirely regular in Ossetic but isn't even a common East
Iranian process. That's why I wouldn't be surprised to see it in a close
relative of Ossetic in the "Middle Iranian" period (Alanic
[Proto-Ossetic] furt 'son' can in fact be regarded as attested), but I
have doubts about projecting it too far back.

>>*-afri- > "Sarmatian" *-aifr- --> Slavic *-e^prU ...

> What is the evidence of Iranian "umlaut" (i-metathesis) in Scythian
> and Ossetic?

Vowel modifications triggered by the presence of a high vowel in the
next syllable are common in Iranian (already in Avestan, and including
Ossetic). I'm not sure if there are examples of diphthongisation not
blocked by an intervening cluster. This particular diphthongisation must
be dated rather late if it is to account for Slavic *-e^str-, *e^pr- <
*-aistr(i)-, *-aifr(i)- < *-astri-, *-afri-, while not showing up in the
Greek renderings of the names (with -astri-, -apri-). In other words, it
would have been "para-Ossetic" rather than "Scythian".

Piotr