> I have looked with my eyes wide open, and I have seen it does not
> work. If I have been wrong in my assessment, somebody tell me
> exactly how.
The difference between an empty assertion and a valid viewpoint
involves a counterexample. I see a lot of off-the-cuff dismissals
but no actual counterexamples.
>> *kwan > *kwon-
>> ---------------
> I see no reason to assume *kawana.
The final vowel is motivated by previous penultimate accent
that regularizes the unpredictable accent in IE in general.
The alternation *kwon/*kunos shows the predictable change
in accent caused by the addition of the extra syllables of
the MIE genitive to the root (*kawana/*kawanasa). There
are so many other examples showing the basis for
penultimate accent and syncope that I've mentioned on this
forum that you should check the archives.
The vowel between *k and *w is needed because consonant
clusters are shown clearly to be the product of zerograde
(cf. *dei- > *dye:us). This gives us at least *kAwonA where
*A is some vowel. If all *a's derive from either *e or *o
as the evidence seems to indicate, we are left only with
*e and *o. But such a vowel system is simply nonexistent
and must be remedied.
If *o were merely *a once, this would immediately
work. Thus *o < *a and *kAwanA. Occam's razor would have
us reconstruct a _single_ vowel *& for all instances of
the unknown *A unless other vowels are shown to exist.
So far I see no such indication here for otherwise,
although I've recently concluded that one other vowel
(written *e) does occur in unstressed syllables distinct
from *a in MIE (as *pedos < *pet:asa and *-es [pl] <
*-es plainly show).
Hence *k&wan&. However since I write *a for *[&] and *e
for *[I], we've justified this very solution that you
lazily dismiss out of turn without explanation. I presume
it is because you can't even accept the Penultimate
Accent theory which is the basis of my entire understanding
of Pre-IE. As I say, it's necessary to explain the
unpredictable accent from *esti/*?senti to *peku/*pekeus.
In response to *osdos, *oxuyom and *polhu-:
> None of the o-forms can be shown to have existed in
> PIE.
Elaborate. Are you saying that none of these PIE forms
exist as roots?? Or are you saying that they cannot
be derived as I derive them? If so, why not?
> You are in essence saying that IE has no old initial
> clusters of any complexity, right?
Whole-heartedly, yes. Why wouldn't I? So many others have
come to this commonsense conclusion because zerograde
alternations prove that at least some consonant clusters
are caused by syncope. So if some are, is there reason to
not think that all are? No. The simplest solution wins
until proven otherwise: IE didn't have initial consonant
clusters in the past.
Since assuming some consonant clusters are inherited and
others are not is a more complicated position, the onus
is on those people to justify their viewpoint that
there is a distinction between two types of consonant
clusters. The mere "possibility" that there WERE
consonant clusters in MIE is irrelevent until proven to
be the case by concrete examples.
My conclusion is in fact perfect because Uralic didn't
have initial consonant clusters, the language group most
commonly thought to be closest to IE! Tyrrhenian is
also thought to be even closer to IE than Uralic and
any instances of initial consonant clusters are the
product of a roving quasi-initial accent that has
nothing to do with IE accentuation. My solution is in
line with the Bomhardian view of Nostratic.
> The position of the infixal /o/ is very plainly sensitive
> to the position of the root-vowel, for when it is
> metathesized it moves right into the position where the
> root-vowel is. That is, of course if it *is* metathesis,
> which I think it is,
"Think" is troubling here because it means "believe" and
belief has no place in science. If some modicum of logic
or even a plea to theoretical simplicity showed that this
was favourable, I could be sympathetic to the solution.
To consider what you say in entirety, that is, beyond
simply an observable pattern of early phonotactics, I
must be shown why it is a morpheme. There is flatly NO
evidence for a morpheme, since you fail to relate it to
anything existent in IE. So it looks as we are not
logically inclined to accept such an infix at all. The
basis for your analysis is demonstrated but not the
solution. Hence my exploration of a phonotactic solution
to your O-fix.
My solution uses the same underlying analysis as yours
but without the belief in a morpheme, nor a mysterious
consonant for what is otherwise clearly a vowel.
> this meaning that the old prefix moved to the position
> *before* the root vowel.
Another assumption is that this is a consonant and not
a vowel. Again, I fail to see the justification.
> Its ability to find the root vowel implies that there was
> only one root vowel at the time of the metathesis.
Assuming that we need metathesis. However, metathesis
like sporadic developments are used countless times by
Nostraticists to support some very wonky reconstructions
so I've become very unresponsive to these sorts of answers
because they can be abused if we don't justify their
usage. Here neither metathesis nor vowel-turned-consonant
switcheroos are transparent, undeniable solutions.
> If the o-form were only a phenotype of zero-grade it
> could not sensibly be expected to "remember" where the
> root vowel had disappeared from. Since it plainly does,
> the theory is wrong.
I'm sure you're misunderstanding my point and dismissing
my idea without understanding the full implications. No
memory is needed because all ablaut forms survive this
rule.
Other alternations of a stem existed in eLIE in other
word forms. For example, in post-Clipping eLIE a strong
case form like accusative *kwan-m opposed the genitive
with weak stem, *kun-as. These alternations are the
direct result of Syncope because they can be regularly
derived from a simpler alternation of *kawana/*kawana-sa
in MIE. However, a derived stem with weak syllabics
used as a strong form of a new paradigm was NOT allowed.
So we could derive a new word from *kwan- like *kwan-a-
"something that relates to dogs" but we cannot allow
a new stem like **kun-a- unless it is the weak case
forms of a strong *kwan-a- starting with a full syllable.
PIE *pleh-, which must derive from MIE *paleh-(em/es/a)
in the singular in order to properly yield *pleh-, also
will show forms like *plhas, again by simple Syncope
from *palah-asa. So we already have *pleh-/*plh- by this
time.
Since all *o's derive from *a due to typological reasons,
the O-fix also derives from *a by accepting the simplest
solution that it is a vowel until proven otherwise. In
a-Epenthesis, the *a is introduced into the first syllable
of the no-no "strong" zerograde form *plhas (with syllable
*l). STRONG forms are only affected. This makes
*palhas, not **plahas because *h is not syllabic, *l is.
The zero form *plh- continues to exist in paradigmatic
alternations as long as it is the considered a derived
"weak" alternate of a strong form in nominal or verbal
paradigms. For example, eLIE 1ps *pleh-&-m versus its
weak 1pp *plh-&-mes or ironically nominative *palha-s
and accusative *palha-m versus genitive *plh-as.
So there is no need for the language to "remember". This is
a deficient objection. All necessary ablaut forms continue
to exist unhindered after a-Epenthesis: *pleh-, *plah- (in
stative or perfect forms), *plh- (in weak forms) and
*palh- (from a-Epenthesis). These regularly change to
*pleh-/*ploh-/*plh-/*polh- after Vowel Shift.
How is this incorrect?
= gLeN