(no subject)

----- Original Message -----

From: Miguel Carrasquer Vidal

To: cybalist@egroups.com

Sent: Sunday, November 26, 2000 5:09 AM

Subject: Re: [tied] How many laryngeals?

> In any case, the lengthening rule did not apply to *-Vs endings (nom.
*-is, *-us, *-os), only to *-Cs (where I'll define C to exclude
*-(V)y- and *-(V)w-).

 

Why? We have *dje:us for /djew-s/, if you need a clear example.

>[And speaking of *ei ~ *i, *eu ~ *u Ablaut, has anyone ever made the
suggestion that this may in fact reflect an older stage **i: ~ **i,
**u: ~ *u?  The idea doesn't seem so strange to me: /i:/ > /ai/ and
/u:/ > /au/ are perfectly commonplace developments [e.g. English], and
so is shortening of unstressed long vowels [e.g. English again].
So what we know as *ei ~ *i, *eu ~ *u Ablaut may have developed as:

*i: > *i (unstressed)
*i: > *ai > *ei (stressed)
*u: > *u (unstressed)
*u: > *au > *eu (stressed)

Just a thought.]

 

Actually, this "Great Vowel Ship" scenario has been considered (e.g. by Schmitt-Brandt, if I remember correctly), but of course there's a long way to go from a suggestion to a fully developed and cogently argued theory. There are lotsa messy details to be handled.

>I was going to say something about *k^e:r(d) (Hitt. <kir>) in my
previous message, but I left it out in the end.  I see it as
confirming evidence that nom.sg. *-s was voiced (voiced *-d lengthens
the vowel, too).  Same goes for nom. *pern > *pe:r(r) [with **-n > *-r
and lengthening caused by voiced-final *-Cr] > Hitt. <pir>.

 

But what about *wokW-z > *wo:kWs? Do you suggest *wogWz as the intermediate stage at which the lengthening was possible? Why is there no lengthening in vocatives like *ph2ter or *k^won?

> As to *wedo:r and *ph2uo:r, I don't think they are the normal
nom./acc. forms (which are *wodr and *pah2wr).  Rasmussen thinks they
are old collectives (*weder + *-h2, *peh2uer + *-h2), and that *-h2
has the same lengthening effect as nom.sg. *-s, but I'm not sure I
agree (although Hitt. <wida:r> is indeed the coll. (n.pl.) form).

 

They may not be "normal" (since their function is marked) but are lengthened Nom./Acc. forms nevertheless. They often serve as collectives or what I call "augmentatives" (water : a lot of water). The "earth" word may be another example if originally inanimate (as in Hittite) -- then *dHgHo:m is like *wedo:r. Cf. also *dek^m (< *dek^nt-?) : *-dkont (in names of decads; here the stem-final cluster blocks the lengthening). I don't believe these forms originally ended in a laryngeal. I suggest instead that ablaut (due to stress shift) and suffixation (with the stress pattern unchanged) were alternative ways of forming inanimate collectives. To connect *wo(:)dr with *wedo:r, Rasmussen needs to assume both ablaut and a suffix; the latter assumption seems gratuitous to me.

>> What I'm gonna say now is unorthodox, but -- shh... -- I doubt if *-eh2 > *-a: is true word-finally. Compensatory lengthening preserves syllable weight, and final consonants are typically "weightless" (non-moraic) under normal conditions. I'd rather expect *-e:h2 > *-a:h2 > *-a: and *-eh2 > *-ah2 > *-a (Nom. and Voc., respectively, exactly like *ph2te:r, Voc. *ph2ter, cf. z^ena vs. z^eno in Slavic). Greek also reflects unlengthened Nom./Acc. coll. n. *-ah2 > *-a (cf. eugene:s m./f. vs. eugenes n.).

> Interesting.  But what about Latin short -a in the nom.sg.?

 

Latin is the odd man out here. Whatever happened in Latin (Nom./Voc.sg. levelling? generalisation of iambic shortening? a combination of both?) was a local innovation.

>> It's extremely doubtful if Latin -ik-s can be regarded as a direct reflex of *-ih2-s. Martinet was a great enthusiast of "laryngeal hardening" in derivations like *seneh2-s > senex (: *seneh2-to-s > sena:tus), but these analyses have been shown to be flawed in too many respects to be taken seriously.

> I've only seen this from Martinet's side (and that only in a rather
superficial fashion: his "Des steppes aux océans"), but the only flaw
I can independently discover is Skt. sanaj- "old".  As to feminine
-ix, what's wrong with it?  BTW, have you seen Rasmussen's article
about "laryngeal hardening" in Slavic [bic^I, bric^I, kljuc^I, zlUc^I,
vrac^I]?

The Latin evidence is discussed at length in "The Evidence for Laryngeals" (ed. by Werner Winter). Martinet is found guilty on many counts of arbitrary morphological cutting and of forcing laryngeals into innocent stems. Let me add from myself that there are inherent chronological paradoxes in analyses that allow *-eh2s to change into *-eks without colouring the vowel. The colouring must be common IE rather than branch-specific -- it's simply not trivial enough to be assumed as an automatic process that took place in each branch separately. On the other hand, the postulated cases of "hardening" are branch-specific (and more or less sporadic, or Latin would have no words like pa:sco: < *peh2-ske-). Why, then, do we get [-eks] rather than *[-aks] in senex? And where do forms like seneo:, sene:sco:, senectus, senio:r fit in? Martinet simply selects the items that seem to support his analyses and ignores or dismisses the inconvenient rest.

 

If -i:k-s < *ih2-s (why the length?), it seems strange that the -k- that originated in the Nom.sg. was generalised so uncompromisingly across the paradigm. I'd expect some archaic k-less forms to survive (e.g. as vocatives). Latin certainly restructured original i:-stems in various ways (cf. neptis 'granddaughter' -- why not *neptix?). Suffixes involving *-k are extremely productive in most branches. In Slavic they are used even today with any noun that might be hard to decline. A Polish child may experience trouble with the synchronically irregular dative of kot 'cat', which is kotu (rather than expected *kotowi), so children get round the difficulty by using the regularly declined diminutive kot-ek, Dat. kot-k-owi. The -(e)k- is just a morphological "buffer" helping regular inflections to attach to any root. It is now mainly a diminutive, but was once used primarily as a "regulariser" (*k^(e)rd- 'heart' -> *sIrd-Ik-o-, *ab(o)l- 'apple' -> *abl-Uk-o-, *wlkW-i: 'she-wolf' -> *vIlc^-ik-a).

 

A general remark: some historical linguists seem to be anxious to make laryngeals so tangible that no critic could possibly deny their presence -- isn't it some atavistic angst inherited from the times when the laryngeal theory was less securely established?

 

Piotr