Balto-Slavic needs laryngeals too (the different development of VRHC and VRC sequences). Some linguists would dispute the assertion that the laryngeals were lost in all branches (Hamp has long argued that some Albanian /h/'s reflect laryngeals [not that I'm convinced], and there are also some unexplained /h/-onsets in Armenian). The parallel dropping of the laryngeals in the various branches can be treated as a manifestation of the natural life cycle of dorsal, pharyngeal or glottal fricatives. They are dropped more easily than other consonants, being already close to the end of the usual lenition trajectories (/x/ > /h/ > zero). Latin (secondary) aitches were lost in Proto-Romance; French ("tertiary") h aspiré in Germanic loanwords was quickly lost; Greek lost its spiritus asper a long time ago; Middle English preconsonantal [x] and [ç] are now mainly reflected as vowel length, and some accents ave lost the phonemic haspirate haltogether. In many Iranian languages /h/ < *s has disappeared, etc., etc., etc. All these changes are certainly independent and have affected /h/ (or /x/) of various origin. The evidence suggests that aitch-dropping is in (typically ?) just a matter of time. Old aitches go and new ones arise from new lenitions, then these too are lost.
In several branches of IE consonantal laryngeals must have existed in some form ca. 2000 BC or later, so Hittite is not so exceptional; it was merely recorded sufficiently early. This means that the lenition-cum-loss was a long process, which was still incomplete in some lineages many hundred years (or even a few millennia) after the disintegration of PIE.
Piotr
----- Original Message -----
From: "P&G" <petegray@...>
To: <cybalist@yahoogroups.com>
Sent: Wednesday, December 11, 2002 9:12 PM
Subject: Re: [tied] Does Koenraad Elst Meet Hock´s Challenge?
> >Why did all IE branches except Hittite drop the laryngeals, leaving only
> traces in the
> >vowels?
>
> That's a fascinating question. But there are more traces than just in the
> vowels - sufficient for us to know that each dialect dropped the laryngeals
> independently.
> (a) Skt needs laryngeals early on to explain various things (eg aspiration,
> scansion, present tense formations, vowel length);
> (b) Latin needs laryngeals early on to explain the apparently irregular
> perfects in -ui (all of which without exception [I think!] are on stems with
> laryngeals - but the -u- perfects are not even proto-Italic!
> (c) Greek needs them likewise.
> So why did they all decide that laryngeals were just for Christmas?
>
> Peter