Re: Sanskrit Voiceless Aspirates - Etymology of 'likh'

From: Neeraj Mathur (neemathur_at_hotmail.com)
Date: 10/28/04 

Date: Thu, 28 Oct 2004 23:44:50 +0100

Hi,

Thanks very much for your help - this has been bothering me for a while!
I enjoy learning from you. I had a few further questions/comments,
interspersed below.

"Miguel Carrasquer" <mcv@wxs.nl> wrote in message
news:72qun0pohr3d8vieofjgq5id5aeungf44j@4ax.com...
> The voiceless aspirates are excluded because they are better
> explained as sequences of stop + laryngeal. The evidence of
> Indo-Iranian is sufficient to establish the presence of the
> laryngeal in roots where I-I has a voiceless aspirate, even
> when no such evidence exists from other branches.

I have perhaps been too much influenced by Szemerenyi, whose explanations
are marvellous, but conclusions are dubious. He says, I think, that in some
cases only can the I-I voiceless aspirate come from a stop + laryngeal
sequence. You're saying something much stronger. Surely, though, if the
laryngeal were in fact there, one could expect to find it vocalised in at
least some languages? You say below that "A voiceless stop + laryngeal
probably gave *ph, *th, *kh everywhere, but these sounds subsequently merged
back with *p, *t, *k in most of the IE area, except in Indo-Iranian." Is
this not the same as saying that there was a voiceless aspirate in
Indo-European itself (which may well have arisen from a Proto-Indo-European
stop sequence, just like the long vowels of IE). Surely, if you are claiming
that it counted as two phonemes in the IE period, then we should expect to
find vocalised laryngeals in at least some of the daughter languages.

> This is incorrect. Hittite and the other Anatolian
> languages had vocalized their laryngeals after a vowel or
> between consonnants, like all other IE languages. In other
> positions (e.g. initial), *h2 is always preserved. Only *h3
> (a rather rare sound in any case) is a matter of some
> controversy. In my opinion it was usually lost
> word-initially.

Again, I suppose I'm a bit too much Szemerenyi's child, and I have no
Hittite. He says that the initial analysis of Hittite seemed to show H1
being lost regularly in the environments where it would not be expected to
vocalise, but H2 and H3 being preserved regularly. He then goes on to say
that more detailed analysis, working with a larger Hittite corpus, seemed to
suggest that in fact, sometimes both H2 and H3 were not preserved when they
could be expected to give a consonant /h/, and that oddly enough in some
cases H1 was. He uses all of this to argue for a single laryngeal, which is
to be reconstructed only where Hittite has preserved it, and leaves aside
the Greek triple representation.

I suppose that, since his writing, the results he is thinking of could have
been reinterpreted, or perhaps etymologies could have been revised, to make
the Hittite evidence more regular.

> Not necessarily. A voiceless stop + laryngeal probably gave
> *ph, *th, *kh everywhere, but these sounds subsequently
> merged back with *p, *t, *k in most of the IE area, except
> in Indo-Iranian. We can conclude that from the fact that
> *kh gives *x, not *k, in Slavic and Armenian (at least
> word-initially), and that *th sometimes merges with *dh, not
> *t, in Greek (e.g. in some verbal endings).

I commented on this above. To summarise my thought, I think that: 1) if the
sequence was two phonemes in the IE period, then there should be some
situations where the laryngeal was between two stops, and thus was
vocalised, leaving only a non-aspirated stop in front of it; 2) if there are
no such cases, then we must argue that the sequence was a single phoneme in
IE - in other words, not a sequence at all, but a voiceless aspirate series.
The origin of such a series as stop + laryngeal would then have to be an
internal instruction within IE to go back to Proto-IE.

>>All of the above could be explained by an IE verbal root *likwH.
>
> The usual reconstruction is *reikH-.

Thanks; I forgot to consider the r/l confusion in Sanskrit.

>>Here, an
>>accent on the thematic vowel in the present could keep the laryngeal
>>pronounced consonantally;
>
> I'm not sure the accent has anything to do with it. A root
> stressed *réikH-e-ti would also have given Skt. *lé:khati.
> The essential thing is that the laryngeal was followed by a
> vowel.

I think I meant the accent because of the zero-grade of the root; I should
not have made it sound like that had anything to do with the development of
the sequence. Thanks for making that explicit.

> There are many roots that end in stop + laryngeal. *petH-
> "fall; fly", for instance.

On what evidence do you say that that root has a laryngeal at the end? The
Sanskrit root seems to be 'pat', not 'path'. Is there a Greek participle for
petomai? I can't seem to find it mentioned in LSJ, but I could be mistaken;
certainly, a verbal adjective of the sort *ptHtos would have a vocalised
laryngeal showing up somewhere. The Latin form peti:tus is not much help,
because of the long vowel; I would expect **petitus < *petatos to be the
regular form, if you allow analogical levelling to e-grade of the stem.

Do you know any other examples of a post-stop laryngeal in a verbal root,
one that I might have fewer immediate problems with?

> Pokorny gives likháti under *rei-, *reik(h)- "ritzen,
> reissen, schneiden".
<snip entry>

Thank you very much - this is useful info. You singled out the Slavic form
as evidence of a voiceless aspirate in that branch; could this be explained
as another Indo-Iranian loan? I'm afraid I don't know enough about the
developments of all the branches mentioned (I can just about deal with
Latin, Greek, Sanskrit and to some extent Germanic - I've a while to go
before I can be a useful Indo-Europeanist!) - is there any other evidence in
that list that would imply a laryngeal in the root?

Again, thanks for your time and your knowledge.

Neeraj Mathur

Relevant Pages

  • Re: Sanskrit Voiceless Aspirates - Etymology of likh
    ... Indo-Iranian is sufficient to establish the presence of the ... laryngeal in roots where I-I has a voiceless aspirate, ... >for settling on three laryngeals, despite the fact that this does not really ... >All of the above could be explained by an IE verbal root *likwH. ...
    (sci.lang)
  • Sanskrit Voiceless Aspirates - Etymology of likh
    ... I have a question about the voiceless aspirates in Sanskrit. ... for settling on three laryngeals, despite the fact that this does not really ... All of the above could be explained by an IE verbal root *likwH. ... Sanskrit voiceless aspirate would be expected to appear; ...
    (sci.lang)
  • Re: Sanskrit Voiceless Aspirates - Etymology of likh
    ... >> when no such evidence exists from other branches. ... of the laryngeal only through the voiceless aspirate in I-I. ... >find vocalised laryngeals in at least some of the daughter languages. ... He says that the initial analysis of Hittite seemed to show H1 ...
    (sci.lang)