Something about Lachmann

In whichever formulation, Lachmann's law (stem vowel is lengthened in
the ppp of verbs ending in voiced stop) contains a hidden flaw: in
order to effect said change, the old voiced stem, the voiced auslaut
of which according to common knowledge was devoiced in the ppp already
in PIE, has to come alive again, only to expire again after doing
that. Also, the unvoicing of that final consonant is progressive,
while the voiced aspirates in Sanskrit progressively influence the
following eg. *-t-. That's a mess.

Suppose PIE had the rule that all stop + stop contacts influenced
progressively? So that
*-gh + d- -> *-gdh-
*-g + d- -> *-gd-
*-k + d- -> *-kt-

and later the *-kt- variant spread by analogy, in the various
languages, driven by a wish to have a uniform identifiable morpheme in
the ppp.

(Actually, above I really meant
*-gh + d- -> *-z^dh-
*-g + d- -> *-z^d-
*-k + d- -> *-s^t-

with later restoration to stop in most languages; since there is no
phoneme **z^h, that would explain the loss of aspiration in
Bartholomae's law)

Latin is full of pseudo-ppp's in -idus; they are perhaps the regular,
non-standardized outcome?


Torsten