14-08-03 13:21, Miguel Carrasquer wrote:

>>What about the shift r' > z^? Is the starting point there
>>attested?
>
> Twelfth century texts write <r>, thirteenth century texts already have <rf>
> (that's r + "long s"), later <rz> (= Czech <r^>, rhotic fricative). The
> change from /r^/ to /z^/ started around the 15th c., but didn't become
> normative until the 18th. century.

[This is really for Richard, not for Miguel, who speaks fluent Polish.]

In modern standard Polish, <rz> is pronounced as (flat apical
postalveolar) /z^/ in voice-preserving contexts and its voiceless
counterpart /s^/ before or after a voiceless obstruent, as well as
pre-pausally. The old pronunciation /r^/ still occurs (or has been
recorded within living memory) in some traditional dialects, not only in
Silesia, where it might be attributed to Czech influence, but also in a
number of enclaves scattered across western and northern Poland. Rhymes
and spelling mistakes prove that the merger of the sound of <rz> with
the postalveolar fricatives spelt <z.> and <sz> in accents close to the
normative pronunciation took place gradually in the course of the late
17th and early 18th c. The morphophonemic alternation of /r/ ~ /z^/ is
still entirely productive: underlying /r/ is transformed into /z^/ (or
/s^/) in palatalising derived contexts: <komputer, laser, radar>,
loc.sg. <komputerze, laserze, radarze>.

Piotr (loc./voc. <Piotrze>, formerly <Pietrze>)

P.S. Incidentally, the final /r/ of <Piotr> is a _completely devoiced_
trilled rhotic.