[tied] Re: Syncope

From: elmeras2000
Message: 31592
Date: 2004-03-28

--- In cybalist@yahoogroups.com, enlil@... wrote:
> Jens:
> > Your rules only work if you change the evidence accordingly.
> QAR doesn't "change the evidence". It doesn't change ANY
> evidence. You know well by now that it is based on the most
> widespread accent alternations and the way you contort
> your words to disqualify it is another form of deception.

No, I don't know that by now. This is a totally personal set of
ideas that are constantly undergoing change depending on where they
get caught. What is QAR? Quantitative ablaut rule? Quasi-something
accent rule? I've lost you. I am not playing any kind of deception
game to do that, it is far too easy for that.

> A mere *-a can regularize this accent yet you hypocritically
> paint it as some kind of "smokescreen" when in reality it
> would be impossible to use this same rule on, say, English
> to even come close to regularizing its accent. Since it is
> unlikely to regularize the accent as it does unless there
> isn't an element of truth to this Syncope that we even have
> proven exists in other positions, QAR is untouchably sound.

I need to see ten examples representing different word types of IE
which present a predictable accent by rule using your QAR thing and
do not do that otherwise. You must show your mileage.

> In IE, when accent alternations occur in a paradigm, QAR
> explains it. When not, there is nothing to explain because
> the paradigm is entirely regular and thus clearly secondary.
> The exceptions to QAR are no more damaging to it than the
> exceptions to Grimm's Law because the rule already does its
> intended job -- to explain as much as possible in one blow.

Accent change in paradigms is well enough explained by the widely
accepted rule that everything syllabic pulls the accent to the
following underlying vowel of the word. That gives you unchanged
accent in the strong forms of nouns and verbs, and on the following
segment in the weak forms of both.

Just to be sure I understand: Is it your theory that a single
underlying vowel in an added flexive is not enough to move the
accent, but that it takes two vowel to do that? In theory that could
conceivably be right, onlty then there is so much to do in the
analysis of the morphology that it will take a long time to have it
tested.

> Now, of the items with alternating accent that QAR cannot
> explain, additional corollaries must be created to explain
> them which leads to further discovery and further rules
> or tweakings.
>
> As one can see, you lack any material objection to my
> methodology. What you have an issue with is QAR itself,
> so you should stick to what objections you have with it
> based on the evidence in general. Yes, from time to time,
> I will say that a specific form that you force me to
> explain does not apply because the etymon is too recent.
> That's why roots like *kwon- are safer. No, we can't
> analyse them which is why they are called roots but words
> like *wertmn are immediately analysable and are likely
> to be more recent because they are clearly not warped by
> centuries of changes that obscure its etymology.

This is one point where your priorities are certainly not
acceptable. We cannot begin with the unknown. I do not know of a
rule, however ancient, that is contradicted by *wért-mn, unless I
take the unscholarly liberty to make one up on the basis of
intransparent material as you have selfavowedly done.

> One wouldn't expect all derived words in English to be
> of antiquity, like "antidisestablishmentarianism". We're
> more likely to expect a simple form like "bird" to be
> ancient. Due to your perverse reasoning, however, you'd
> expect me to not only explain "antidisestablishmentarianism"
> in terms of Old English grammar but if I logically can't
> do this task with a rule that I've devised then you say
> that my rule is invalid even when it explains the majority
> of simple, underived forms. That's plain crazy.

You would have to treat "bird" as a root and leave it unanalyzed.
The same with *wert-, but not with *wért-mn which does offer a small
field of investigation. And an analysis of English the begins with
your long "anti..-ism" is likely to produce much more and better
insight into the history of English and its derivations that any
amount of time spent looking at the indivisible word "bird".

>
> > No, if the root was athematic at the outset it would have
> > become *bhr-ént-i, not *bhér-o-nt-i.
>
> Ugh, I made an error. It should be *ber-ena because as you
> correctly point out, *a would disappear otherwise in the root,
> yielding **bHrent. After updating my theories, sometimes
> archaicisms creep into my analyses or when I switch
> from one stage to another where rules operate a little
> differently, it's easy to lose track.

I wonder what you wisom will make of that tomorrow or the day after.
While in principle updating is noble and wise, it might be expected
to prompt a certain amount of modesty.

> > Again, your analyses are disqualifying the language you
> > are trying to attain insight into.
>
> No, rather you're purposely distorting QAR by picking
> apart any errors that I might make elsewhere or the
> fact that it isn't a TOE (Theory of Everything). I've
> already stated what the rule ACTUALLY is and it shouldn't
> be brought to task for not explaining everything. So your
> rhetoric isn't constructive.

You need to remind me what the rule is. It should be easy for you if
Occam helps you keep it simple.

> > Ah, do they now? I see that quite differently. They
> > introduce too many knowns for your darling theories to
> > survive.
>
> You must tame your bloodthirst for my head on a stick.

I have no such thing, I am addressing the matter under discussion,
but I observe a totally and very disturbing attitude on your part.

> What you just said here is ludicrous because you want
> me to make a rule that will explain all derivations,
> regardless of their proper chronology or whether those
> derivations even existed during the events prescribed.
> This is where the debate and your expectations become
> insane.

Well if you have that allpervading problem and other more
conservative theories don't, or only to a considerably lesser
extent, then your theories have lost the race and are not the ones
to be preferred. Still, they may contain grains of interesting
observations here and there, and that is what I am trying to tap
them for.

> > It's this way every time I point out blatantly contrasting
> > evidence to you; then the forms either did not exist, or
> > you cannot handle them for whatever reason. - I admit however
> > that I act the same way when you base yourself on the singular
> > of the perfect which to me just fails to show mobility where
> > I would have expected.
>
> Hmm.
>
> Then perhaps you can resolve this self-contradiction on your
> own time and not confront me with it in the future. As I
> always continue saying, rules aren't meant to explain
> everything. They're meant to explain as much as possible
> and yes, there ARE going to be exceptions. You constantly
> bring up the mere existence of these exceptions as some
> proof that my rules must be a farce. How senseless. I
> want my rules to be evaluated for what they are, not
> what they aren't.

No, it's a matter of proportions. I my assessment you have simply
started at the wrong end, making some broad cavalier assumptions on
points where they can't be checked and then proceeded to invent
accounts of whatever material has come to your attention afterwards
in such a way that it either fits or is excused. But now you are
confronted with a growing avalanche of well-known material that does
not fit in your brave new world but was already well enough
accounted for in theories you do not want to even consider.


> > Who are you to tell speakers of pre-PIE how they should
> > have formed their causatives?
>
> What does this emotive plea have to do with justifying
> the use of the O-fix rule here? You continue to speak of
> the "O" as a morpheme even though it hasn't been logically
> established as such. The "o" appears to be inserted into
> morphemes unexpectedly perhaps, but a morpheme? That's
> too much to accept considering that it doesn't even act
> like a normal IE morpheme, even if it were.

So call it a "co-affix", I guess that is the proper term of affixes
that are constantly combined with other affixes. They are off course
morphemes too. The causative is formed by simultaneous prefixation
of *O- and suffixation of *-éye-. It is an automatic matter that the
stem-final vowel is treated as the thematic vowel, as all such
vowels are. It is apparently also an automatic matter that the
presumed prefix in the vast majority of possible structures is
metathesized with the initial consonantism of the root and takes its
position immediately after it. The same coaffix is combined with the
thematic vowel in the derivation of the word-types tomós and tomé:,
IE *tomH1-ó-s, *tomH1-á-H2, and also in nominal derivatives from
suffixed primary formations such as kor-mó-s and pho:né: from IE
*kor-m-ó-s, *bhoH2-n-á-H2, derived from the already-derived veral
nouns */kér-mn./, */bháH2-mn./. It is not easy to say just what the
functional benefit of the o-part is, but that it is the same in
formal terms in all three is quite plain. Note that is also the same
if you just crudely call it "o-grade", for these are three sets of
cases whose "o-grade" works in a very special way not shared by
other "o-grade" formations of the language. And, in contrast to the
unenlightening and confusing term "o-grade" the analysis of the
three as a unit with highlighting of the unexpected behaviour of
their /o/ as if it comes from an asyllabic element gets at the core
of the matter and introduces precision where there was earlier
nothing but confusion. You would prefer a more exact definition of
the elements in functional terms, so go ahead, now you have
something to work on, which was not the case before. But it may be
like "bird", with the oldest parts of the language you can see only
so much.

> So we should
> throw the morpheme idea into your other bag of goodies
> like double-long vowels, "accent stealing" and such and
> chuck it into the large recycling bin where we put away
> other non-necessities.

The lengthening of already-long //e:// to produce the /o:/ we find
in nominatives is the only account I know that has really shown that
it works and allows morphological normality in the examples. The
initial accent rule, if that is what you mean, is demanded by the
fact that acrostatic paradigms are acrostative, yet alternate with
gradation in complete parallelism with the other type where é/zero
follows the accent. I have seen no other account for this that has a
chance to work. It was unexpectedly confirmed by the differentiated
accentuation of the o-infix formations made from derived stems.
Otherwise it is without any causation that light roots always accent
the thematic vowel and heavy roots always the root in such pairs as
kormó-s : tórmo-s. These are from *kér-mn and *tér&1-mn
respectively, from which are derived, before the operation of the
ablaut, *O-ker-men-é- and *O-terH1-men-é- respectively. With
metathesis they become *kOermené- *tOerH1mené-, from which ablaut
reduction makes *kOrmné- *tOrH1mné-. In these now the laryngeal is
suppressed (let me write it with lower-case h), and the cluster /-mn-
/ is reduced, so we get *kOrmé- *tOrhmé-, at which point the
vocalization of O to /o/ begins, starting in the heavy type which
needs it the most, giving *kOrmé- *torhmé-. At this point now the
initial accent rule comes into life, changing the set to *kOrmé-
*tórhme-, and finally the remaining asyllabic O's are also
syllabified, and the reduced h is lost, so we get *kormé-, *tórme-,
which with the rules applying to the thematic vowel, produce the PIE
nominative sg. *kormós, *tórmos, whence with few problems Greek
kormós, tórmos.


> >> [...] so don't you think it is wiser to apply more
> >> general rules FIRST before lumping the causative into
> >> this?
> >
> > That's what I did, so the answer is yes, only I did it,
> > you didn't.
>
> This is still the matter of the debate. It seems Syncope
> is our shared starting point but while I build my theories
> sequentially by concluding things built on previous
> conclusions, you appear to have a more scattered approach.

I don't think so, rather we disagree substantially as to which
conclusions can be allowed the status you call "previous". I find
you beginning at the wrong end going ever deeper into a mist of wild
guesses.

> QAR for example is built on Syncope, and Acrostatic
> Regularization is built on QAR.

This will need explicitation for all the relevant sets of the
morphology on which it is supposed to work. I suppose would know by
now if it had, but pray do show it again, foor it sounds important.

> I fail to see a similar
> linearity in your ideas. You have an O-fix rule
> dangling over here, a double-long vowel hypothesis
> over there, but these are just a bunch of theories that
> spontaneously appear with little to do with each other
> until you bring them together to maintain your convictions
> of a pre-IE replete with freakshow characteristics.

A language is not so simple that you can demand every rule to be
relevant for every single word. Rules apply to given environments,
which are not just offered by all words. This is clearly not fair.


> >> Well, it would affect accent in compounds if it were
> >> the second element of the word.
> >
> > Show us how. Accent in compounds is a wasps' nest, so we
> > can do with some enlightenment.
>
> Putting aside Acrostatic Regularization, does the second
> element of a compound not normally contain accent?

Not, far from it. Bahuvri:his, which in the opinion of some are
the "only" ancient compound type (a misguided view in my opinion,
but not without a lot going for it) constantly accent the first
member.


> > The root was demonstrably *H1es- with an initial laryngeal.
>
> Absolutely in IE but not necessarily so in pre-IE.

No, in pre-PIE it did not necessarily have the /s/ either. Is it
meant to be entertaining? It could also have had an element "/U/" of
unknown realization, which was lost in the meantime by a rule
of "/U/ deletion". hat's the point in making such arbitrary
guesses?

Jens