"centum"/"satem" "exceptions" [was Re: Northwest IE attributes]

[roslyn frank]

At 08:36 PM 2/3/00 EST, you wrote:
>>frank at uiowa.edu writes:

><< 1) what then is the standard reconstruction proposed for Skt. pu:r-, Greek
>> <polis> and Baltic <pilis>? And the prototype meaning assigned to it?

>-- *pelh(x), "fort, fortified place"..e.

Could you elaborate a bit more in order to explain why the phonologicial
shape of the Gk. and Baltic items take precedence over that of the Skt.
pu:r-?

Are there a general set of rules that show the regular correspondence of
Sk. words in <pu:r> to Greek <poli(s)> and Baltic <pilis>, i.e., that there
are other examples of the same transformations?

>There's a possible cognate in Armenian -- 'k'alak', meaning "city".

>>2) is this set considered a good candidate for admission to the (P)IE
>>lexicon? Stated differently, does attestation in Sanskrit, Greek and Baltic
>>languages suffice for a data set to be considered part of the (P)IE lexicon?

>-- a dialect word of the south and east of the PIE world, at least.

>There's also *uriien, 'fort', which gives Mycenaean 'rijo', promontory, and
>Tocharian 'ri', 'town'.

Do these examples imply that there was a loss of the initial plosive in the
last example and that one could posit an earlier *<uri> or perhaps *<iri>
for Tocharian? What are the data sources for the reconstruction *uriien?
Do you mean to suggest by citing this example that there was a loss of the
initial plosive in the case of Tocharian. If so, how is that explained? Is
the correspondence regular between Sanskrit and Tocharian with respect to
the loss of the initial plosive, i.e., does it occur with other words.

Then with respect to the prototype meaning, is the choice of "fort,
fortified place" based on the fact that such a location/structure would
antedate an urban site such as  "town" or "city"; or is there some other
basis for this choice?

And back to the candidacy of this item for admission into the (P)IE
lexicon, does the presence of the various and sundry lexemes that make up
Buck's entries under "town" (19.15) and "fortress" (20.35) indicate that it
is assumed that there is no recoverable/identifiable (P)IE etynom for the
concept "fort, fortified place"? In the case of <burg>, et. al. I once read
that they believed that the referential object to which it once applied as
an Iron Age "hill-fort", although I don't recall the exact citation.

Thanks in advance for the info.

Roz Frank