Drift vs. common innovations; satem

[ECOLING at aol.com]

Edward Sapir had the very useful concept of "drift".
The concept simply acknowledges that there can be causes
for shifts, and that changes may happen by stages,
in degrees, a language can move a part of the way towards a shift.
This can be the case as a matter of fact,
without our necessarily being able to know all the details.
It is a useful way of thinking.

(I am of course *not* suggesting that we posit things
for which we have no evidence, simply that we do not rule out
hypotheses consistent with a small amount of data available to us,
posited developments for which we have actual known parallels.)

I believe the concept of "drift" helps to synthesize the
two views, that of family tree and that of dialect area,
which we find so often debating each other, each having
a part of the truth in view.

(I strictly distinguish "dialect area" in my thinking
from the concept of "Sprachbund" or areal effects,
which is appropriate when the languages involved are unrelated,
or when the effects are demonstrably after a period of separation
and return to contact so the results are different for reasons
of intervening changes.  The mere overlap of waves of innovation
across a persisting dialect network are somewhere in between?)

Consider the obvious possibility that there were several
*stages* in developments *towards* a satem-shift
(NOTE THAT WORD *towards*).

(In what follows I am neglecting Tocharian and
the obvious other small details.  Do not misunderstand
the simple terminology used merely to explain the concept.)

If some of the exact details can be improved,
someone else please do so, and then critique the
best version you can create.

***

Common innovation: stage 1 of satem drift

the entire area of the eastern IndoEuropean shared a portion
of the drift, including even Baltic and Greek.
Did not go far enough to be manifested in Greek (at all?).

***

Common innovation: stage 2 of satem drift

Armenian, Balto-Slavic, Indo-Iranian shared this stage of the drift

***

Common innovation: stage 3 of satem drift

Balto-Slavic and Indo-Iranian shared this stage of the drift

***

Common innovation: stage 4 of satem drift

Slavic and Indo-Iranian shared this stage of the drift

***

Common innovation: stage 5 of satem drift

Indo-Iranian shared this stage of the drift

***

Even possibly stage 6 of satem drift

Did Iranian go farther than Indic?

***********************************************

I always try to think in concrete terms, not to resolve real
questions via strict definitions (where the definitions answer
the question and therefore make the question less an empirical one).

So the following does make more precise wherein certain languages differ,
but does not satisfy me fully.

>When the issue of satemization is defined very strictly, it consists - not
>of the assibilations and spirantizations - but of the merger of velars and
>labiovelars (with the old palatals showing distinct reflexes), as opposed
>to kentum lgs., where palatals and velars merged and labiovelars show
>autonomous reflexes. If this definition is fololwed strictly, Armenian and
>Albanian are out (no merger at all); Baltic and Slavic are mostly in (but
>not entirely, because of the sizable number not-at-all satemized elements).
>Which leaves us with Indo-Iranian as the only real and true-blue satem
>branch (and if we look at the remnants of labiovelars there, even that is
>not entirely sure).