Re: IS PIE * DERU EXCLUSIVELY INDO-EUROPEAN ?

To assert a *d-prefix for PIE, the minimum requirement would be to
show the existence of other *C-prefixes. It is hardly possible that
PIE had this category of modification for just one morpheme.

That was the argument against s-mobile being a prefix. So perhaps both
*s and *d were prefixes. A d-prefix might also explain the word
initial pt found in Greek where other languages point to p (first with
devoicing of d- and then metathesis).
===========
Tsalam? t?ob

This is the most interesting discussion we 've had for six months.

In the case of Greek pt- or kt-,
I consider a -t- infix.
Example : Root *pu-
English spew
LAtin spu-tare
Greek p-t-u-ô
ARnaud
================
Also in Greek there are
z-initial words where one would expect an initial y-. In Greek dy > z.
There are several examples of words where we expect an initial nasal,
but find initial d instead. For example, PIE had *(e)noh1men yet
Hittite has a reflex atiman.
===========
Both -n- and -t- are infixes.
in this root H_H-
ARnaud
============

PIE has *nebHos yet we have Luwian
tapas^a and Lithuanian debesis. Then there's the famous Slavic devyni
meaning nine though we'd expect it to have an initial nasal. All these
could be explained with a *d-prefix that causes the loss of an initial
nasal.
PIE also has some *k/0 alternations (perhaps a *k prefix), *w/0
alternations (a *w prefix or infix), and *y/0 alternations (a *y
prefix or infix). There are also some variations involving initial *l
suggesting an *l prefix. For example, PIE *(d)ak^ru but Latin lacrima
and PIE *(e)noh1men Hittie lamen.
===============
Great !

Arnaud