From: willemvermeer
Message: 36407
Date: 2005-02-19
--- In cybalist@yahoogroups.com, Miguel Carrasquer <mcv@...> wrote:
WV had written:
> >It is the word "optionally" that is hard to reconcile with the MAS
> >conception here. The MAS view holds that at the earliest
> >reconstructible stage (which they assume was inherited straight
from
> >PIE), every morpheme had one of two inherent tones, either High or
> >Low. Contrastive stress didn't exist.
Then MC wrote:
> Where is this theory laid down? I must have missed it.
I'm afraid I can't quote chapter and verse right now, but if you
start from a system in which the prosodic pattern of a word form is
determined entirely by the inherent tones (H vs. L) of the
constituent morphemes, the place of the stress is redundant, hence
non-contrastive.
[On the possibility of the MAS system for PIE:]
> I don't think it makes any sense for PIE, which surely had
> stress (take zero grade, for example),
That is one point I would very much like to hear the opinion of the
cognoscenti about. I've written earlier that K really owes us an
update about how he views the connection between late Proto-Indo-
Uralic, the various phases of PIE and early Balto-Slavic. Some of it
is scattered in the literature, but I for one cannot manage to
combine it into a coherent whole.
> nor for the PBS
> soundlaws, such as Pedersen's law, which creates lateral
> mobility _of the stress_ in mobile C-stems and oxytone
> non-neuter vowel stems, or Hirt's law, which pulls the
> stress back if the syllable is _pretonic_ and contains a
> non-vocalized laryngeal [vocalized laryngeals _also_ have
> high tone, so clearly there can't be a Hirt's law without a
> concept of "stressed syllable"], or Winter's law, where a
> vowel is lengthened if _pretonic_. None of these laws make
> any sense if there was no such thing as a stressed syllable
> in PBS.
Not so fast. Arno Verweij and me found out that most of the rules
responsible for the rise of lateral mobility can be understood also
in terms of the analogical generalization of low or high tones under
fairly trivial conditions, e.g. generalization of L as the tone of
all accusative endings starting from a system in which some
accusative endings are H and others L. Both Hirt's and Ebeling's laws
can be reformulated in terms of neutralization of the contrast
between H and L in final syllables given certain phonetic conditions,
with the "architoneme" being realized more like L than like H.
I've not said it is pretty or preferable or even necessary, but it
looks at least feasible, so it may be better to keep an open mind as
long as so much is so uncertain.
Note also that in K's scheme the reflex of the laryngeals has nothing
to do with tone throughout the BSl period and that K sticks to
Winter's original formulation of what we now call Winter's law, which
does not depend on the place of the stress.
So what Arno and me found was that the BSl segment of K's theory
appears to be compatible in principle with the fairly long retention
of inherent tone the MAS people like to operate with. But it is just
an experiment all the same. It is not at all compelling, except that
it taught us that it would be a good idea to keep an open mind.
>
> The PIE tonal system, as far as it can be deduced from the
> Greek and Vedic evidence, was very much connected to the
> stress. Short stressed vowels had rising tone (udatta),
> stressed long vowels also had (long) rising tone, except
> when they were the result of contractions, in which case
> they had (super)long falling tone (circumflex). Post-tonic
> vowels had falling tone (svarita), all others were toneless
> (anudatta).
If you reverse the description, stress turns out to be an automatic
consequence of the tonal movement within a word form. If I'm
correctly informed, both the Indian and the Greek grammarians
described their systems in terms of tonal movements.
> Resonants (*m, *n, *l, *r, *w, *y) after a vowel (i.e.
> diphthongs) had falling tone in Balto-Slavic, which was
> possibly inherited from PIE. In that case, Greek has given
> up the feature for *m, *n, *r, *l (e.g. in <poimé:n> the -n
> behaves as a plain consonant, not affecting the tone, which
> was originally circumflex as in Lith. piemuo~).
This looks like an intrusion of foreign elements into the discussion.
> I can find no evidence in Baltic and Slavic that the
> stressed vowel had rising/high tone.
Well, to the MAS people the stress assignment rule ("konturnoe
pravilo")*is* the evidence. (Note that there is no direct link with
the tones earlier investigators have postulated for PIE and PSl. In
K's theory "acute" means 'glottalic' until very deep into the history
of the separate branches.)
> Dybo's law in Slavic
> and Saussure's law in Lithuanian rather suggest the
> contrary: a short stressed vowel had falling tone (Dybo's
> law might suggest that svarita vowels had rising tone, but
> this is not confirmed by Lithuanian).
Both Dybo's law and de Saussure's law are so recent within the
respective branches of BSl that the conditions they presuppose cannot
be projected back without further ado into the BSl period.
I'm not going to comment on your proposals. What I'm trying to do is
clarify the MAS scheme and K's theory, so that effective criticism
may some day become possible.
Willem