Re: [tied] Re: PIE voiceless aspirates

>----- Original Message -----
>From: "etherman23" <etherman23@...>
>To: <cybalist@yahoogroups.com>
>Sent: Friday, November 04, 2005 8:35 PM
>Subject: [tied] Re: PIE voiceless aspirates
>
>
>> --- In cybalist@yahoogroups.com, "dariusz_piwowarczyk"
>> <dariusz_piwowarczyk@...> wrote:
>> >
>> > Is there a necessity to reconstruct voiceless aspirates (i.e. ph th
>> > kh) for the Indo-European proto-language?
>> >
>> > Many books I've come across does not postulate them for PIE. On the
>> > contrary, my own teacher of IE Comparative Linguistics (W. Smoczynski)
>> > argues that their presence is necessary in order for the system to be
>> > phonologically complete (i.e. to have the opposition between the
>> > voiced and voiceless aspirates). Can someone tell me what's the
>> > general opinion on that subject?
>> >
>> > I'd be grateful for any help.
>>
>> The overwhelming majority opinion is that they're unnecessary.
>
>***
>Patrick:
>
>And that majority opinion is absolutely wrong.
>
>PIE *p/*t/*k and *ph/*th/*kh have different correspondents in related
>languages; and the correspondents for *ph/*th/*kh for the greater part,
>correspond to the _preserved_ voiceless aspirates in Old Indian.
>

-------------------------------------------

Yes, a group of roots with a phoneme different form /T/, /D/, and /Dh/
is there, e.g. *thengh- "pull" (LIV 657, IEW 1067); in such roots, no
laryngeal */h2/ can have caused aspiration because the following vowel is
*/e/. Professor Rasmussen has collected a number of these instances; ona
may add, with Pisani, also Sl. *xle:b" "bread", Gm. xlaiba- < PIE
*khloibho- if connected with Lat. li:bum (a plausible, though not cogent,
etymology).

Still, there is a small difficulty, that is: the occurrences of */Th/
are less than those of */Dh/. In order to avoid this difficulty, one could
try to posit for PIE a system that Professor Rasmussen has postulated for a
pre-stage of PIE:

(from more to less numerous)

traditional Four-Series-PIE : 'revised' Four-Series-PIE

*/T/ : */Th/
*/Dh/ : */Dh/ (no difference)
*/D/ : */'D/ (pre-glottalized voiced)
*/Th/ : */T'/ (glottalized voiceless)

This would be a typologically plausible system; its price would be that
no Indo-European branch would have preserved the entire picture