--- In
cybalist@yahoogroups.com, "Richard Wordingham"
<richard.wordingham@...> wrote:
> However, _I_ don't know whether the 2s
> imperfect, is ever vr.ddhied for athematic stems ending in
> consonants.
My point has been all along that there is no lengthening before 2sg
*-s in any IE material we know. Most of it would be redundant
because of Osthoff's shortening in many branches, but Indo-Iranian
is clear: Skt. áhan, akar, ágan, á:nat., inj. bhet, ipf. ábrav:is,
ákr.n.os, arin.ak, Avest. mraos^, co:is^, var&s^-ca:.
This difference in morphophonemic behaviour was my main reason to
separate the nominative sibilant from that of the second person
singular. I then found them to differ also in their effect on a
preceding thematic vowel, as nom. *-os vs. 2sg *-es. The combination
of the two observations has brought me a lot of trouble.
Jens