Re: [tied] n/r (was: PIE suffix *-ro - 'similar-with')

----- Original Message -----
From: "Sean Whalen" <stlatos@...>
To: <cybalist@yahoogroups.com>
Sent: Saturday, January 07, 2006 2:00 AM
Subject: [tied] n/r (was: PIE suffix *-ro - 'similar-with')


>
> --- Patrick Ryan <proto-language@...> wrote:
>
> > I can see no "individualizing" component in the
> > -*nt- suffix.
> >
> > What shocked me, after I had been regarding -*n as
> > an individualizing suffix
> > for many years, was the necessity of recognizing and
> > reconstructing a
> > second -*n as a collective - and believe me, folks,
> > I fought hard against
> > making this assignment.
>
> If you're talking about something like *wedn/r- vs.
> *itn/r- (water vs. a way) my explanation is a suffix
> -n- in the same category as dual and plural with
> "partial" meaning: "one of" for countables and "some
> of" for uncountables. Also suffix -t- for "complete"
> or "group" (*dekY_mt "all ten" *pe:Nt "all (five)"
> (with *pe:N(t)kWe being contaminated with kWe "and" in
> counting as *kWetwores).

<snip>

***
Patrick:

The short answer, Sean, is no - this is not what I mean.

But more completely:

I think everyone would agree to reconstruct an 'individualizing' meaning
for -*n.

But, although a 'collective' for -*n is rare, is there any real alternative
but to reconstruct such a meaning for *ne-, 'we'?

The question with which we must grapple here is which of these two meanings
makes more sense for the participial -*no and, for what I presume to be an
extension of it: -*nt.

If I have understood him, that is at the core of Piotr's suggestion; and I
agree with Jens: it is ingenious.

In what I have seen so far, I cannot justify a meaning of 'complete, all' to
be attached to -*t. What seems to be implicit there is 'habituality',
leading to an implication of 'futurity'.

The only process I have been able to identify that I seems to me to require
an interpretation of 'totality' is _complete_ reduplication. Partial
reduplication only seems to imply 'iteration' without the necessary
implication of 'totality'.

You will notice I am merrily mixing interpretations primarily attached to
'nouns' _and_ verbs because I believe that there was little if any
difference between these two categories in the _earliest_ PIE and that any
of the same morphemes were used for both with appropriate interpretive
modifications.

For 'dual', I have found only one morpheme that is unequivocal: -*y.

Of course, unfortunately, there is also a 'derivational' *-y and
'causative'/'necessitative' -*y but these are usually fairly easy to
distinguish.

To try to distinguish any of these by the following vowel is a thankless
task since all those mentioned above have the Ablaut 'vowel': *e/*o/*Ø.

The only two PIE morphemes, of the vowel quality for which in PIE we can be
sure is -*a: (*H2e, is you insist), which is the result of an earlier *ha,
'plural (uncountable)'; and -*a (*H2e), which is the result of an earlier
*ha from an even earlier different source: 'feminine'. There is a 'stative'
in -*?a which oftenest is seen in the stage of PIE we normally reconstruct
as -*6 (schwa), just a hint to let us know we are dealing with a former
'laryngeal'.

-*r has the 'intensive' nuance that almost everyone would agree on seeing;
in addition, it has a 'distributional' employment related to the idea of
'counting'. I have been regarding it as representing an 'indefinite' in 3rd
p. pl. verbal forms (passive) but I wonder if that is really correct. It
might be, instead, a 'definite' marker since it is tied to the idea of
'number'.

Previously, I would have said the -*r/-*n heteroclyte declension is a
conflation of 'indefinite'/'definite' but, just as of this writing, I am
beginning to wonder if I have been wrong. It may be that -*r represents the
'definite' and -*n, in this case, the 'indefinite' collective -*n which
would accommodate its employment in a participle more easily. I have
reopened the question in my thinking, primarily because of the -*l/*n
heteroclytes: *-n could be the 'indefinite' here since I can discover no
connection of -*l to 'indefinite' or to 'collective' but there is a definite
(pun!) connection to 'definite' in the form of its employment in
demonstratives.

> I reconstruct x etc. for H1/2/3 because they color
> V's as velar stops do differently early in various
> languages (phleg-/flag-); they dissapear before velars
> with the same place of articulation in some languages
> (*p(e)lxYúx kWid > *polwáxki > polláki in Greek); gh
> dissimilates to g before syllabic x in Greek
> (thugater-); x becomes syllabic before w but not y in
> Greek; combinations of round or palatalized velar and
> plain velar become all plain in Khowar and x has the
> same effect.

***
Patrick:

So do I (or plain /h/) but I do not believe in the 'coloring' unless, by it,
you mean 'retention of an earlier vowel-quality'.

Hittite shows rather unambiguously, however, that -*? must also be
reconstructed for an early stage of PIE.

I cannot really follow the rest of what you write here. If you wish to
pursue it, could you spell it out a little more expansively?

As for "thugater-", I believe this is the normal expectation from PIE
*dheug(h)-, in which the aspiration is preferentially lost: we see this root
again in MEng 'dug'.

***