Typology and the phonetics of laryngeals

[Jens Elmegaard Rasmussen]

On Mon, 17 Apr 2000, Ante Aikio wrote:

[...]
> But
> there's a problem with *H1 as [h] from an Uralic point of view. There
> are quite a few loan words with *H1 > PU *S. A substitution [x] >
> retroflex [S] is phonetically sensible, given that the system in the
> receiving language doesn't have an unvoiced velar fricative or [h] -
> parallels are known. But [h] > [S] seems impossible in any circumstance.
> One would rather expect [h] > [k] or even [h] > zero.

> But then again, I guess it is not necessarily the case that the sound
> values of the laryngeals remained the same in all the daughter languages
> before they disappeared. The cases with IE *H1 > U *S seem to be
> Pre-Baltic and Pre-Germanic.

> I have also entertained the thought that there may be one loan word which
> points to a palatal(ized) value of *H1. PU *s´ijili 'hedgehog' might
> derive from PIE *H1eg´hi-l- (> German Igel). But this etymology must be
> considered uncertain since no parallel cases for PIE *H1 > PU *s´ have
> been found.

While one should of course be cautious about sweeping statements
straddling millennia, I have some worries with an Ich-Laut for *H1. First,
could a palatal spirant really avoid triggering a vowel (by svarabhakti or
otherwise) in the position after non-initial stops if both of the back
spirants *H2 and *H3 do produce one? Second, would its vocalization
product (by svarabhakti or direct vocalization) really by a centralized
vowel? Third, the reflexes of H1 and H2 do not follow anything remotely
reminiscent of the satem/centum isogloss. Fourth, if people are crying for
an /h/ because there are aspirates, isn't this our chance to given them
one?

   Even so, I cannot dismiss your suggestion of phonetic variation - that
could even have been there from the start, meaning that [h] may be just
one of the manifestations of /H1/, which would still leave room for [x^]
being another. I do not believe phonetic typology has reached a point
enabling it to exclude any such thing. And incidentally, the Greek
reflexes of CRH1C with /-Re:-/ are easier to udnerstand from [x^] than
from [h], since the latter would simply add voicelessness, but not
redirect the articulation to any other location than where the sonants are
themselves - and they all produce [a] when given an undisturbed course.

   But also FU should be kept open for interpretation in terms of
variation. I have always tended to see the substitution of sh-sounds for h
and the Finnish development of sh to h in the light of the Swedish
pronunciation of sj which varies a huge lot and certainly includes sounds
that would qualify excellently as substitutes for [h] if the language did
not have that already. Is it possible that the development of sh to h had
begun, on a limited scale, in Proto-FU already, so that a foreign [h]
could be perceived as a (substandard) manifestation of an sh-like
sound? That would be a story much like the Russian g for h.

   For what it's worth, I think the equation of siili and Igel is
brilliant.

Regards,
Jens