pre-IE k > H

[Carol F. Justus]

Annelies Kammenhuber never really endorsed the laryngeal theory because of
words where one scribe might use a sign with a 'k' equivalent where others
usually used the Hittite 'h' that confirmed Saussure's hypothesis. I don't
have a list of such words, but someone may.

Hittite scribes, of course, were not all native speakers of Hittite, nor
were Hittite texts written in a monolingual context. (I think someone
mentioned six written languages at Hattusa; there were seven, but you might
argue for losing Sumerian since it was only written, not spoken. But
Palaic, both dialects of Luwian, Hattic, Hurrian, Akkadian, and Hittite
certainly were, even if Hattic died out in the early years.) Without
abandoning the laryngeal theory, such scribal variants as those between k
and h may have their own tale to tell.

Some years ago, Yakov Malkiel, exhaustively studying the details of the
Romance languages, took a position often reduced to the slogan "every word
has its own history" because he found so many explanations for exceptions
to regular sound correspondences. While both the laryngeal theory and the
Neogrammarian principle of sound correspondence remain landmarks in the
science of language, competing forces have not been systematically factored
in for IE in much depth. In this context the Hittite k ~  h alternation,
for example, has not been systematically accounted for. This has left
Greenberg some room for working with the pre-IE system of sounds. It would
be interesting to do for IE what Malkiel did for Romance linguistics,
namely to amass some data on the anomalies, especially if they turned out
to show sub-systematicity beyond what the 19th century did so well. We are,
after all, in the 21st century! Pat's question is not a bad one.

Carol

>Dear Indo-Europeanists:
>
>In Greenberg's new book, _Indo-European and Its Closest Relatives_, the
>proposal is made on pp. 59-60 that pre=IE or early IE k became a
>"laryngeal".

>Greenberg cites Latin costa, 'rib', and Greek ostou^s, 'bone', among others,
>as examples of "a number of roots which require proto-forms with k alongside
>of H".

>I was, of course, aware of *kost- and *ost(h)-, for which I would prefer
>another explanation, but I am unaware of numerous examples suggesting this
>relationship.

>Anyone know of a few more?

>Pat