🚀 go-pugleaf

In article <1993Jan22.191906.3648@leland.Stanford.EDU>, alderson@cisco.com (Rich Alderson) writes:
> In article <1993Jan21.094836.5120@memstvx1.memst.edu>, connolly@memstvx1 writes:

[stuff omitted]
>> There remains, I think the Greek
>>genitive plural.  *Any* dissyllabic form we like could be posited to produce
>>this.  Perhaps we should consider *-oHn (with syllabic n, since the laryngeal
>>was a consonant)?  Here too the circumflex would be a peculiarly Greek
>>development.
> 
> But you don't address either the disyllabic reading of the Sanskrit genitive
> plural, or the Germanic data, skimpy as it is.

One error: I should have posited *-oHm for the genitive plural.  This would
have been dissyllabic in PIE, with syllabic -m.  If the laryngeal was
preserved in Vedic times -- perhaps only between syllabics, as here --
the dissyllabic readings are entirely predictable.  Now I can't say
that this is right, since I simply concocted the form without much thought
or any checking, but it seems to fill the bill.

As for Germanic: I just dragged out my Prokosch to double-check the geni-
tive plurals in the older dialects, but I don't see anything there to
indicate dissyllabic pronunciation.  Except for Gothic -e:, which is just
plain weird, the endings are identical to the nominative masculine singular
of n-stems, which seems to go back to PIE -o:n (monosyllabic, with lengthened
grade of the vowel).  My proposed -oHm would probably produce the attested
genitive plural, but I don't think the Germanic forms evidence dissyllabicity.
Am I missing something?  Ah yes: The Old Norse masculine and neuter n-stems
have nominative -i or -e, while genitive plurals generally have -a -- but
the *feminine* n-stems have NSg. -a.

>>2. Quite certainly the primary PIE stress was not *always* one of pitch.  You
>>mention "non-reduction of vowels in the earliest-attested languages" to
>>demonstrate the lack of a strong "dynamic" accent.  This argument carries
>>considerable weight.  But equally certainly, the quantitative ablaut
>>alternations (e:0, ei:i, eu:u, eH:H, eHi:Hi, en:n, etc.) argue *for* a strong
>>dynamic accent at a still earlier period, since these alternations are well
>>reflected in Greek and Sanskrit.
> 
> The "proto-language uniformity" issue raises its ugly head.
> 
> Hmm.  Sanskrit and Greek are part of a well-known Central group sharing a large
> number of morphological innovations and several phonological ones as well--in
> fact, Greek is the most "satem" of the "centum" languages at its earliest
> attestation, with all those secondary assibilations.
> 
> Maybe the pitch accent both show was a Central innovation, and ought not to be reconstructed for the whole family?

Sounds plausible, and I had considered that the pitch accent may have been
confined to certain dialect areas; but I don't know how one could prove
such a thing.

> In any case, whether pitch is a Centralism or not, it's clear that the stage at
> which quantitative ablaut arose is (probably much) earlier, since there are
> reflexes of it in every IE language, and it is only productive morphologically
> --that is, it defines morphological distinctions, but is not predictable phono-
> logically.

Right.  But then, the (probable, if limited) pitch accent wasn't predictable
either.
 
> So we're left with the possibility of a pitch or a *weak* stress (like modern
> French) at the time of dialectical breakup, given that phenomena consonant with
> a *strong* stress do not occur.

Agreed.

>>> ...in the [Greek] optative, I would claim
>>>that it is a reflex of a fairly late retention of the original laryngeal
>>>(compare the athematic alternation -ye:/-i:).
>>...
>>I also happily concur with the late laryngeal retention.  It should be obvious
>>that many laryngeals were *not* lost in PIE times, much less during the
>>transition from "Indo-Hittite" to PIE, as Sturtevant and others have claimed.
>>Unfortunately, most Indo-Europeanists still try to sweep laryngeals under the
>>rug, either denying them or pretending they don't matter.  They do.
> 
> Sorry.  I'm still sensitive to the reaction this suggestion got when I made it
> as a student.  I'm not used to people agreeing with me.  :-)

I don't know what it is about laryngeals that make so many linguists ner-
vous; but since the obvious reflexes in the attested dialects are so diverse
and plainly language-specific, they absolutely *must* have survived through
Indo-European into the earlier stages of a great many IE dialects.  Naturally,
this does not mean that the full phonemic inventory of laryngeals survived
unchanged in all positions everywhere.  Mergers and splits seem likely,
loss in certain positions, etc. -- the usual sort of things that happen with
any phoneme.  In Germanic, sequences such as VHj and VHw show varying
developments, depending on the laryngeal and the Ablaut grade of the vowel.
PIE oHj (any laryngeal) and /eAj/ [aAj], with a-coloring laryngeal, yield
Germanic ajj, a process known as "Verschaerfung" ("sharpening") -- but
eEj (non-coloring laryngeal) yields e:j.  This suggests that E after e
dropped, causing compensatory lengthening, but not after PIE o (Gmc. a),
where, since A yields the same reflex, E and A may have merged while
still laryngeals (or after, for that matter).

Too bad they gave you a hard time about this in grad school.  I had it easier:
I was in Germanic philology, not Indo-European, and my advisor pretty much
believed in laryngeals himself.

> I don't have a firm enough grasp on the current literature on tonogenesis to
> say for certain but isn't it (at least usually) the case that when a pitch
> accent arises from a stress accent, this is in company with loss of syllables
> usw.?  (I mind the Norwegian and Swedish contours, for example.)

I don't know, but you can add the "Rheinische Schaerfung", a tonal distinction
in North Rhine dialects of German triggered by loss of unstressed syllables,
to your list.

--Leo Connolly