Re: [tied] Finnish hevonen "horse"

----- Original Message -----

From: Knut

To: cybalist@yahoogroups.com

Sent: Tuesday, October 30, 2001 11:07 PM

Subject: Re: [tied] Finnish hevonen "horse"

> The time depth we are talking about when comparing uralic and indo-european is so great that it is meaningless to try to find exact correspondances between such sub-phonemic elements as labialisation, aspiration, glottalisation, etc. Such elements can ossilate back and forth many times during this time span. What can be done, I think, is to see if elements in two words have common phonetic traits.

 

First, we are talking of _phonemic_ labialisation, etc. (*k and *kW are different phonemes). Secondly, if segments are allowed to "oscillate back and forth" according to your convenience, you are left with no criteria for distiguishing real and spurious cognates. There is a high price to pay for loose methodology. If "it is meaningless to try to find exact correspondences", this in fact means that the comparative method cannot be applied and no credible reconstruction is possible.

> According to my opinion "kVw" and "p" could very well be related, If you go back to a period before common IU and common UR. For example could the uralic form result from the development kVw>kw>pp>p.

 

> And with the H/S^: The change from s/S^ to h is so common that it could have happened also in periods before common IU and common UR. Then s could have been the original, and it could have changed into h in IU.

 

You can't propose a may-have-been development (let alone two developments, as above) based on a single pair of words. An isolated "match" is a contradiction in terms, since you need multiple instantiation to define a rule. You'd have to offer further convincing examples of PIE *h1 : PU *s^ (or *s) word-initially and PIE *k^w : PU *p between vowels. You say above that it's "meaningless" to look for such regular correspondences. If it were so, it would be sad news for anyone trying to reconstruct "Indo-Uralic", since the whole enterprise would have to be regarded as stillborn.

> And what actually is the difference between a labiovelar and the sequence kW? In either case, the sequence kw could very easily develope into a labiovelar.

*kW is a unitary phoneme and a single segment (labiovelar stop), whereas *k^w is a cluster of two phonemes realised as two segments (palato-velar stop plus labiovelar approximant). The two are consistently distinguished in all Satem languages (e.g. in Slavic *kW > *k but *k^w > *sv), and were evidently kept distinct in PIE, though they admittedly merged later in several branches, partly or completely.