Re: [tied] Re: Laryngeal Loss (was Does Koenraad Elst Meet Hock´s

From: Miguel Carrasquer
Message: 17176
Date: 2002-12-14

On Fri, 13 Dec 2002 11:32:13 +0000, "Piotr Gasiorowski
<piotr.gasiorowski@...>" <piotr.gasiorowski@...> wrote:

>--- In, "tgpedersen <tgpedersen@...>"
><tgpedersen@...> wrote:
>> Yes, yes. Let me rephrase then:
>> Why did almost all of IE's and almost none of AfrAs' branches loose
>> their _original_ laryngeals at an early date, regardless of whether
>> they later developped new ones?
>I'll have to check the fate of the "original AA laryngeals" first
>(inasmuch as they are reconstructable). I'm by no means sure that they
>have survived so well.

The original Semitic laryngeals are not hard to reconstruct:

velar/uvular *x *G
pharyngeal/epiglottal *H *3
laryngeal(glottal) *h *?

In Phoenician, there was an early merger of *x and *H, and *G and *3,
which is why the Phoenician-derived alphabets only have 4 letters for
the laryngeals (our H, O, E, A). Akkadian only had a sign for /x/ (<
*x, *H), but may have had /h/ and /?/ as well (not fully reflected in
the spelling). Classical Arabic has retained the six Proto-Semitic
laryngeals, but in a language like Maltese (an Arabic dialect
relatively uninfluenced by Classical Arabic), the spelling only
reflects four (H < *H, *x, gH < *G, *3, h, q = /?/), and pronunciation
only two (written <gH> and <h> being usually silent).

New "laryngeal" sounds gave been created in several Semitic languages
(e.g. Hebrew k > k_ (/x/), g > g_ (/G/) or Aramaic *tL. (= Arab. /d./)
> q > 3. PAA *s^ is reflected in Akkadian as <s^>, but elsewhere as
/h/ (what one might call *h2).

What the original Afro-Asiatic laryngeals were is a harder question to
answer. As usual, the different reconstructions for PAA do not agree.

Ancient Egyptian had several "laryngeal" consonants: 3 'i (=j) ` h h.
x h_. According to Roessler's system, 3 reflects PAA *r, and 'i
usually PAA *y (plus a number of palatalized reflexes of other
consonants). Egyptian <`> reflects PAA *d, *d_ and *dz.

The PAA laryngeals would be reflected as:

*x = Eg. h_ or s^ (pal.)
*G = Eg. x or 'i (pal.)
*x. = Eg. h. or d_ (pal.)
*h = Eg. h
*? = Eg. 'i

Ehret, Orël & Stolbova do not agree, and, besides other differences,
think that Egyptian <`> reflects PAA *3 or *G.

In Chadic, it is doubtful whether PAA *x and *G are reflected (perhaps
they give k and g in some languages), and only *h and *? are secure.
Proto-Cushitic, as far as I can tell, appears to reflect all six
laryngeals (although mergers have usually taken place in the modern
languages), while Omotic has generally lost all trace of them.

Whether PAA really had six laryngeals is open to discussion. The
presence of an emphatic laryngeal *H (h., x.) strongly suggests that
the series *x *G *x. was originally not fricative, but a stop triad
similar to /*p *p. *b/, /*t *t. *d/, /*k *k. *g/. The logical
reconstruction would be a uvular series *q *q. *G" (as in part
suggested by Orël & Stolbova), which gave *X, *X., *R (> *x, *h. *G)
in Semitic-Egyptian-Berber (but stops in Chadic-Cushitic-Omotic?).
That leaves *h, *?, *3 [another triad] and perhaps *x2/*X2 (not from
*q but a separate velar and/or uvular fricative, like we similarly
also had dental and palatal fricatives *s, *s^ [perhaps also labial *f
and lateral *L]).

The PAA consonantal system then perhaps looked like this:

stop/affric. fric. son. semiv. -> Arabic
asp glot voi
labial *p *p. *b (*f) *m *w f b b -- m w
dental *t *t. *d *n t t. d n
*c *c. *dz *s s s. z s
palatal *c^ *c^. *dz^ *s^ *r *y t_ z. d_ h r y
lateral *tL *tL. (*dl) (*L) *l s^ d. -- s^ l
velar *k *k. *g *x k q j x
uvular *q *q. *G" (*X) x h. G x
*h *? *3 h ? 3

(the dental and palatal affricates *c *c. *dz; *c^, *c^., *dz^ are
perhaps ultimately derivable from palatalized dentals and velars)

If the origin of the consonantal triads *t *t. *d etc. lies in
original tonal/prosodic/suprasegmental distinctions, that would
explain the existence of an "aconsonantal" triad *h *? *3.

Miguel Carrasquer Vidal