On Thu, 10 May 2001 19:27:04 +0100, "petegray"
<
petegray@...> wrote:
>But the real point might be that there never was an "add r" rule for the 2nd
>person in early Latin.
In Old Irish, -r is not added to the 2pl. In Hittite, -ri is rare in
2sg. forms. It looks as if -r was originally avoided in second person
mediopassive forms.
There is also an ending -m, which appears in Skt. middle second person
dual and plural forms (imperfect/aorist, optative, imperative), as
well as in the 3rd. person middle imerative. It does not occur in
first person forms.
My guess is that the *-r and *-m were originally avoided in the second
and first person respectively, precisely because they were,
respectively, second and first person markers.
We have a similar situation in modern Basque, where the first and
second person verbal dative agreement markers cannot occur in first
and second person subject verbal forms, e.g.:
<x> give(s) it to <y>,
ematen -- diat dinat diot -- dizut dizuet diet
,, didak -- -- diok diguk -- -- diek
,, didan -- -- dion digun -- -- dien
,, dit dik din dio digu dizu dizue die
,, -- diagu dinagu diogu -- dizugu dizuegu diegu
,, didazu -- -- diozu diguzu -- -- diezu
,, didazue -- -- diozue diguzue -- -- diezue
,, didate diate dinate diote digute dizute dizuete diete
, where x, y are I/me, you(m.), you(f.), he/him/she/her/it, we/us,
you(pol.), you(pl.), they/them, respectively.
It is not difficult to imagine how the PIE middle, with its semantics
of reflexive action, might have developed out of the debris of a
conjugation with (optional) incorporated indirect object, for which we
can tentatively set up the following grid:
1sg./pl. 2sg. 3sg.(/pl.?) 2pl.(?)
1sg. -- -h2a-r(i) -h2a(i) -h2a-dhw(i)
2sg. -th2a-m(i) -- -th2a(i) --
3sg. -(t)o-m(i) -(t)o-r(i) -(t)(e/oi) -(t)(e/o)-dhw(i)
1pl. -- -mwo-r(i) -mw(e/oi) -mw(e/o)-dhw(i) [*-me(s)dh-]
2pl. -dhwo-m(i) -- -dhw(e/oi) --
3pl. -nto-m(i) -nto-r(i) -nt(e/oi) -nt(e/o)-dhw(i)
Different languages chose different forms out of this grid to build up
their mediopassive conjugation(s). The optional *-i can thus be seen
as identical to the loc./dat. marker *-i (it was not originally a
present tense marker in the middle conjugation, as best shown by
Tocharian, which picked the forms in *-i for the 1 & 2 sg. middle
_past_ tense (*-(m)ai, *-tai)).
=======================
Miguel Carrasquer Vidal
mcv@...