Munda in Early NW India

Douglas G Kilday acnasvers at hotmail.com
Wed Jun 13 09:06:35 UTC 2001


Vidhyanath Rao (21 Apr 2001) wrote:

>I am not sure if "Douglas G Kilday" <acnasvers at hotmail.com> is for or
>against *gh etc being affricates. But I will use his post to to hang my
>questions on.

>> [...] A language
>> with an inventory of stops comparable to Sanskrit ought to have 3 or 4
>> distinct fricatives in addition to /h/.

>Does that mean that Sanskrit itself `should' have had 3 or 4 distinct
>fricatives?  [Nor am I sure what distinct means: Sanskrit had voiceless
>fricatives produced at various places, but these where all allophones of
>/s/.
>There is the sound commonly transcribed h, but this is voiced and results
>from (Brugmannian *gh', and less often from *dh. Then there are three
>sibilants.]

What I meant, and should have specified, was _phonemically_ distinct. I am
labeling voiced /h'/ and the sibilants /s/, /s./, /s'/ as "fricatives".
Macdonell calls them "spirants". In my usage, "fricatives" are continuants
produced with audible friction, and "sibilants" are fricatives in which the
primary articulation is formed by the extensible part of the tongue.

I don't have the advantage of formal instruction in Sanskrit. Macdonell
describes Visarga /h./ as a "hard spirant" (i.e. unvoiced fricative) and, if
I am reading correctly, [x] and [f] are "obsolete" allophones of Visarga,
which sounds [h] when unassimilated. Hence it appears that Sanskrit has 5
phonemically distinct fricatives.

>>[...] OTOH if the traditional (Brugmannian) voiced aspirates were
>> "really" voiced fricatives, the transition to voiced aspirates in Indic
>> could be viewed as systematic substrate-induced fortition, if the
>> Munda-substrate hypothesis is in the ballpark.

>How do we explain Germanic? Or for that matter, Iranian where *dh etc
>became d etc (at least intervocalic)? And if substratum explanation is
>in the ballpark, why did the aspirated series go to a voiced fricative?
>[in Vedic, this already had happened with intervocalic *dh in verbal
>endings such as mahe (< PI-Ir *madhai), as well as all intervocalic PI-Ir
>*jh and this speads to all (intervocalic) voiced aspirates in MIA.]

It's not hard to find examples of voiced fricatives becoming voiced stops.
This is sufficiently common that no substratal invocation is necessary: I
certainly wouldn't posit "Pre-Chicagoan" substrate to explain "Da Bears" or
"Da Bulls". In the hypothesis under discussion, Brugmannian *dh etc. are
presumed to have _originally_ been voiced fricatives, which became voiced
aspirates in Indic only. The phonetic systems in Iranian and MIA are thus
viewed as more conservative than the Sanskrit system, and there is no need
to explain "the aspirated series going to a voiced fricative". The trouble
lies in explaining the reverse, namely the voiced aspirated series in Indic,
and this is where substrate may be in the ballpark.

Of course, if PIE is presumed _not_ to have had aspirates, there is the
difficulty of explaining why both Greek and Sanskrit avoid successive
aspirated stops. Some form of Grassmann's Law must have operated in Late
South PIE. Since I don't understand the phonetic basis for Grassmann's Law
in the first place (which looks like the result of a cyclic suprasegmental),
I am neither "for" nor "against" *gh etc. being the traditional voiced
aspirates, at least in Late PIE.

My main point was that traditional PIE is short on fricatives and long on
"laryngeals". Historically, unvoiced fricatives lead a precarious existence,
like frogs living in a toilet-bowl. The next phonetic shift may flush one of
them down the larynx to become /h/, which is subject to further reduction
into a glide or even outright loss. I think the H-series in PIE represents
undetermined unvoiced fricatives, probably varying greatly among dialects
and more rapidly in time than most other PIE sounds. Their characterization
as "laryngeal" properly belongs to their final stage as fricatives, and I
doubt that any synchronic speech had more than two literally laryngeal
fricatives (fortis and lenis, perhaps). Otherwise we must conclude that PIE
was spoken by a tribe of ventriloquists (which could, of course, help
explain the difficulty in locating their homeland).

DGK



More information about the Indo-european mailing list