Odp: Grassman's Law.

From: Piotr Gasiorowski
Message: 194
Date: 1999-11-06

junk
 
----- Original Message -----
From: markodegard@...
To: cybalist@eGroups.com
Sent: Saturday, November 06, 1999 4:45 PM
Subject: [cybalist] Grassman's Law.

Grassman's Law.

From: Comparative Indo-European Linguistics: An Introduction, by Robert S.P. Beekes (John Benjamins, 1995), p.128.
 

Grassman's Law says that an aspirated sound loses its aspiration when followed by another aspirated sound; two aspirated sounds in one form, then, could not be found together. In Sanskrit, for example, dh - dh became d - dh.

Strictly speaking, "Grassmann's Law" is the general name for the TYPE of dissimilation that involves the loss of aspiration in a consonant followed closely by another aspirated sound. Aspiration requires some extra articulatory effort, and various languages are known to prohibit its multiple occurrence in a single root. Grassmann's Law (GL) operated in Greek and Sanskrit, and presumably also in Tocharian.
 
Here is a concrete example: the root *dheigh-/*dhoigh-/*dhigh- 'shape, form, knead' had an initial aspirate, as evidenced by its Germanic and Latin reflexes (Gothic ga-digan, English dough, Latin figura). In Sanskrit we find dehi: 'wall, dam' with the following history:
dh..gh > dh..jh (satem palatalisation) > d..jh (GL) > d..h
In Greek the corresponding form is teikhos 'wall':
dh..gh > th..kh (Greek devoicing of aspirates) > t..kh (GL)
In Greek, GL also deletes an aspirate reflecting PIE *s:
segh- > hekh > ekh
This, as well as its interaction with some other Greek changes (the palatalisation of stops before *j), proves that GL operated at a rather late date there. Its comparison with the change in Sanskrit makes it clear that the process operated independently (with different results, though with the same phonological motivation) in the two languages.
 
The evidence for Tocharian is more subtle. Tocharian lost all voice and aspiration contrasts, so that, for instance, *k, *g and *gh are all represented as Tocharian k. But *t, *d and *dh did not fall together completely. *t merged with *dh as Tocharian t, but *d changed into ts. This minimal difference alows us to detect traces of GL in Tocharian:
dh..gh > d..gh (GL) > ts..k (NOT t..k, Tocharian devoicing)
Examples:
Tocharian tsik- 'form' < *dheigh-
Tocharian tsäk- 'burn', Sanskrit dahati, but Latin foveo 'warm (up)' with f < *dh (the root is *dheghw-/dhoghw-)
 
As you can see, the Tocharian version of GL is of the Sanskrit, not the Greek type. This doesn't mean that Tocharian is closely related to Sanskrit; the two languages are too dissimilar in other respects. It's just a case of a parallel tendency producing similar results.