Re: Five questions about H. Rix et al.'s _LIV_

--- In cybalist@yahoogroups.com, "Francesco Brighenti" <frabrig@...> wrote:

> 4) that of the remaining 34% of the roots listed and discussed in LIV, nearly half have been
> reconstructed using laryngeal segments, and that only 18% of the total can, therefore, "form evidence, in principle, for genuine linguistic correlations, because their reconstruction meets the three-witnesses criterion and do not make recourse to laryngeal segments"?

It's not for nothing that cynics have described etymology as a subject where vowels count for nothing. However, 'cat' and 'dog' differ in more than the vowel, which is something that the author deliberately ignores. He also overlook ablaut systems as persistent elements in Indo-European daughter languages - what is compared is not so much matching ablaut in cognate words as similar ablaut systems.

For example, when he mutters about the lack of attestation of the perfect o-grade/zero-grade alternation between the singular and plural, he totally overlooks its pervasiveness in Germanic strong verbs, which is there irrespective of whether the verbs in question go back to Indo-European.

On the present v. perfect contrast of /e/ and /o/, he appears to have overlooked Brugmann's law in Indic, with the 3s perfect forms such as caka:ra.

The number of laryngeals does seem disturbingly high in IE verb roots. On the other hand, it perhaps should not be so disturbing - the occurrence of laryngeal-type sounds in Ancient Egyptian is also very high.

> 5) that the number of IE linguistic laws used to reconstruct the verbal roots in LIV is, thus, higher than the number of "genuine" verbal roots as defined at point 4)?

For detailed reconstruction, one has recourse to a great many laws. Try counting the number of laws needed to derive French from Latin! However, not all these laws depend on evidence from words of Indo-European origin.

Richard.