Re: [tied] Re: Caland [was -m (-n)?]

On Tue, 25 Nov 2003 23:22:16 +0000, elmeras2000 <jer@...> wrote:

>--- In cybalist@yahoogroups.com, Miguel Carrasquer <mcv@...> wrote:
>> On Tue, 25 Nov 2003 15:44:32 +0100 (MET), Jens Elmegaard Rasmussen
>> <jer@...> wrote:
>>
>> >And it is not only in adjective stems we observe reductions
>> >of *-Ce/o- to *-e/o- (and further *-i-), we see something of the
>kind at
>> >the end of compounds too (*newo-g^nH3ó-s for **-g^nH3-tó-s).
>>
>> Would you say that the variation *-ós ~ *-tós in (most) ordinals
>> illustrates the same phenomenon?
>
>No, I see no connection. My impression of the ordinals is that
>Szemerényi's analysis was correct: The morpheme is *-o- (thematic
>vowel), but metanalysis of *dek^mt-o-s led to a new morpheme *-to-
>which had at least spread to 'fifth' before the dissolution of the
>protolanguage. That appears to be a totally different story.

Allright. I feared you had abandoned the analysis through *dek^mt-os >
*dek^m-tos. Any other examples of -Cos > -os apart from neognós (if it is
that)?

>[...]
>
>> But what of the other half of the problem? The alternation r/m, for
>> instance, which (according to Collinge) was part of Caland's original
>> treatment (Avest. xrvi- vs. xru:ma-, xru:ra-, Av. tiGra- vs. Skt. tigma-).
>
>> If any *-Co- (or *-o-) will do as a source for *-i- in compounds (and
>> comparatives), then that's only a solution for half the problem. It does
>> not explain why *-mó- and *-ró- themselves alternate in these words (nor,
>> for instance, what the alternation of u-stem adjectives and ró- adjectives
>> is all about).
>
>There is no alternation of *-ro- and *-mo- except in formations
>where they have lost their meaning.

I just gave two examples of such an alternation. In what sense have they
lost their meaning?

>There is an alternation between
>*-ro and *-u- which both form adjectives. One could say they form
>positives. The two are distributed in what looks like the ruins of
>older complementary distribution: If the syllabic nucleus is a
>sonant, the suffix is *-u-; if the nucleus is a vocoid it takes *-ro-
>. There are some cases of doublets, but Brugmann's sizable list can
>really be arranged to show this.

I'm sorry, I can't see this distribution. The first five adjectives in
-ro- that come to mind (*h2r.gró- bright, *h2ugró- strong, *h1rudhró- red,
*kruh2ró- "grausam", k^ubhró- "shiny, pure") don't follow the rule, nor
does the second u-stem adjective that springs to mind (*swáh2dus "sweet")
[The first one, *bhr.g^hú- "high" does].

>I suspect that the origin of *-ro-
>and *-u- is ultimately the same and that, under the relevant
>phonotactic conditions, *-u- is the phonetic representative of both
>*-ro- and *-i-.

Allright, that's my point of departure too. So what do you think are the
phonotactic conditions?

>The suffixes *-mo- and *-no- alternate when from *-mn-o-, originally
>distributed by a dissimilatory principle seen by Johannes Schmidt: *-
>no- if a labial precedes on the word, otherwise *-mo-. In the
>passive participle, however, *-no- alternates with *-to-, a fact one
>would like to connect with the active participle *-ent- of which it
>is perhaps an adjectival derivative.

That's all either true (*mno > *mo/*no), likely (*-ent connected to
*to/*no) or worth investigating (*to and *no as variants).

However, I am not convinced that _all_ adjectives in *-mo- or *-no- derive
from *-mno-.

[...]

>Now, all of this is "reconstructive rather than descriptive" (I've
>forgotten who said that) even when applied to the protolanguage. It
>accounts for the connections between some very archaic derivatives
>which are old enough to have undergone some very old processes of
>phonetic changes. They have been overgrown by a host of later
>formations that did not abide by the same sound rules simply because
>the changes they reflect were over. That of course makes it a
>difficult and insecure undertaking to try and uncover the changes
>and the system they worked on. But, while we may have some hope of
>verifying the regular phonetic changes involved, if instead we put
>up a scenario in purely functional terms we do not have a prayer.

I don't think anybody _here_ is putting up a scenario in purely functional
terms. I don't even know what that means exactly.

My analysis, as I sketched it before, is one of phonetic change. We know
(at least in this group, if not outside it, it seems to be communis opinio)
that at some point in pre-PIE, final *-n developed into *-r. If the
adjectives in *-rós are thematizations of earlier *-r/*-n-stems (it is
perhaps significant, perhaps just a coincidence, that both in Latin and in
Anatolian stems in *-ros show forms in -er and -ar, respectively, which are
usually explained by assuming special soundlaws reducing *-ros > *-rs >
*-r), then that takes care of the connection between *-n/*-n-os and
*-r(-os) . What the connection is with the u-stems is more difficult to
see, except in Armenian, where -r, -u- and -n- happily coexist in the
paradigm of u-stem adjectives (barj-r, barj-u, barj-un-k`), as well as some
(neuter) nouns. My proposal is a set of soundlaws:

NAn. *bhérg^h-un > *bherg^hur ~> barjr (oblique stem barj- analogical)
[this is just the -n > -r soundlaw]

G. *bherg^h-ún-os > *bhr.g^héw-os > *bhr.g^hwós > barju
[stressed *ú labializes a following *n or *t: the result is *ún > *éw, *út

> *és, cf. *méh1nu:t-, *meh1núto:s > *méh1no:t-, *m(e)h1nésos].

pl. *bhérg^h-un-es(W) > barjunk` (perhaps through analogical
*bhérg^h-on-es, with normal n-stem plural -unk`, -an- < */n./ in the plural
oblique)

Now all that's required is a palatalization law to explain -i- < *n^.

=======================
Miguel Carrasquer Vidal
mcv@...