From: elmeras2000
Message: 32718
Date: 2004-05-18
--- In cybalist@yahoogroups.com, enlil@... wrote:
.
>
> So what's the obsession with making up a new phoneme? Why isn't
> a category of allophone good enough for you people? Nobody's
> supporting their opinion with sensible facts of any kind.
>
>
> > No! If the vowels were always different, this minimal pair
provides
> > no evidence of allophony! It simple shows that thematic /o/
and /e/
> > are different!
>
> No, it does show at the very least allophony because the majority
> pattern is that thematic *-e- becomes *o before a _voiced segment_.
> We never see a thematic vowel *o before *dH or *g but always before
> a voiced phoneme. However, *s is the only consonant that breaks the
> rule.
>
> The thematic vowel, btw, has to have been once one and the same
vowel
> as in *to-s and *te-syo so there's no logical escape from accepting
> that *e/*o < *& (or some vowel of your choice). Surely, *to- is not
> a different pronoun from *te- so why the alternation? Because *tos
> was once pronounced [toz] and tesyo had unvoiced *s. In that way,
> *s is now in line with the rest of the pattern.
>
> So Jens is justified in using the pattern to surmise a "z".
However,
> he _assumes_ that it is a seperate sound from *s rather than simply
> an allophone and he does this by decree of his own. Therefore, he
> would reconstruct *toz and *tesyo. In that sense, he's gone too far
> because he hasn't shown that they ARE seperate phonemes.
I must say I am being widely understood quite correctly here which
is rewarding and rare in itself.
In the real debate we may perhaps be at a stalemate if we go on
passing the buck and sweeping the problem under the next carpet. I
find it hard to regard word-final voicing a preferred solution, but
who am I to tell?
We know that at one time in the prehistory of PIE the marker of the
2sg and the marker of the nominative acted differently on their
phonetic environments. The nominative caused lengthening, the 2sg
did not; the nominative selected preceding thematic vowel *-o- (or
its prestage) just like voiced segments, while the 2sg ending took *-
e- just like voiceless segments.
One possibility is now the one I have suggested by positing nom. *-z
as opposed to 2sg *-s. That would cost a phoneme, but that's all. Is
that to be avoided at all costs? I can't see why it should be, but
that is not really the point; the point is whether it reflects the
truth. So the alternatives should be considered with an open mind,
If the nominative marker was just /-s/, realized with subphonemic
voicing [-z], the nominative is all right, but then the 2sg ending
must be given a different form. We would like to posit *-t for that
anyway at some point, and it should of course be considered if it
could still be *-t at this stage. That would of course mean that
forms in *-d cannot be posited with /-t/ also, nor forms showing
final *-t. I have no problem accepting *-d for the pronominal neuter
a simply /-d/, but there is a clash between 2sg *-s and 3sg *-t
which cannot both be /-t/. So, do we conjure 3sg *-t into something
else too, or do we posit *-s : *-z? I do see indications that the
3sg marker was once a cluster /nt/ or rather a unit phoneme /Nt/,
but I have not ventured to bring that in before now (some might say
I have phonemes enough). To my great surprise the distribution of
the variants of this element (if such it is) seems to be sensitive
to structural differences which only came about by the working of
the ablaut (of the "Syncope" brand), so perhaps it does not run as
deep as I first thought. Then we may have /-s/ : /-t/ : /-Nt/ later
becoming *-z, *-t, *-Nt which could survive as *-s, *-s, *-t. The
lengthening and o-selecting processes would then have to be dated to
the stage with *-z, *-t, *-Nt.
Note that for this stage 2sg *-t is posited only to avoid a phonemic
opposition /s/ : /z/ because we don't like that. And since the 2sg
is *-t, the 3sg must still have an older form than its end product *-
t, so that's why I posit that as *-Nt. I do this simply to test all
possibilities I can make sense of.
Now, if in the stage with *-z, *-t, *-Nt, thematic vowel *-o- (or
its prestage) was selected by the voiced *-z (leading to *-oz > *-
os), and not by the voiceless *-t (leading to *-et > *-es), the
question about the 3sg arises. How could *-Nt not act like a voiced
segment? I consider this point fatal: If *-z, *-t, *-Nt have been
correctly posited, the change of the last two to *-s, *-t must have
occurred before the selection of the variant of the thematic vowel,
and then *-e-z (> *-oz > *-os), *-e-s, *-e-t follows with logical
necessity.
Other ways of avoiding this have been suggested, mainly by the
assumption of added independent pronouns. That, however, is not the
most likely of assumptions:
The nominative lengthening operated in a chronological stage when
unaccented short vowels were in the process of weakening, for when
lengthenend they survive as *-o:- in opposition to accented *-é:-.
The cause of the difference in accentuation in paradigms, i.e. the
rightward accent shift before desinences having an underlying vowel,
had already occurred in an earlier period and must be taken as a
given at this later stage. The three forms we are talking about all
belong to the "strong" forms of paradigms and so cannot be posited
with preserved vowels at this stage without causing additional
problems. They may of course do just that, and then the merry-go-
round does not stop here either.
Then, the way I see it is this: Provided the PIE endings nom. *-s
(with variants) : 2sg *-s : 3sg *-t are vowelless consonants at the
stage we are talking about, as seems to be the case, they can hardly
have been anything other than *-z : *-s : *-t. This demands a
phonemic constrast between /z/ and /s/ in the stage in which the
variation of the thematic vowel arose.
I do not think English is an adequate parallel here, for English [-
z] only occurs after originally voiced segments. All three variants
[-s], [-z] and [-iz] come from -es which took on voice in cases
where the environment supported that. To that environment belonged
the preceding vowel, while in the case of pre-PIE *-z there is quite
strong evidence that there was no such vowel present at the relevant
time. And yet the nominative sibilant caused lengthening, even after
voiceless consonants (*nép-o:t-s, *wó:kW-s). If a special phonetic
feature is needed for this, it must have been located in the
sibilant. This all falls into place if there was an independent
phoneme /z/ at the time.
The argument that /s/ is such an extremely frequent phoneme in IE
and so should be preferred if possible, is not logical. If there is
a disproportionate frequency of /s/, that may be because /s/ had
more than one source. Unfortunately, we have no yardstick by which
to measure how frequent sibilant phonemes ought to be in this
language.
I think we need, or at least could benefit from, an independent /z/
in the account of the ablaut of the sigmatic aorist too, and I also
think some individual word stems just happen to show a lengthening
effect which, among known factors, can only be ascribed to the
sibilant.
Jens