Re: [tied] Re: Syncope

On Sat, 20 Mar 2004 enlil@... wrote:

>
> Jens:
> > Let me help: The root was *pleH1-; ablaut reduction caused zero
> > grade *plH1-, and analogy created a new full-grade *pelH1-.
>
> So then, you're saying that *pleh-, *plh- and *pelh- all
> co-existed within IE itself just before the breakup of IE,
> right?

Yes.

> I can certainly see why disallowing *polh-no- would
> be to your advantage but its perpendicularity to your
> steadfast beliefs will not make it go away on its own.

It doesn't have to go away, for it is not present; it never was more than
a suggestion in the first place, and a wrong one it seems.

>
> > But stress *is* the obvious feature to have caused unaccented
> > short vowels to vanish.
>
> Thank you for agreeing. We both accept the alternative
> possibilities (eg: tonal accent) but we see the optimal
> solution (ie: stress). That's all I wanted. You are slowly
> being assilimated.

I am not yet a'sillymated to the point of accepting an incomplete
characterization of the IE accent which I sense is being anticipated here
for future reference. For the purpose at hand the stress part of the
accent is enough, sure, but I will not say anything that can later be
interpreted to mean that I said this is the full story. If I do that I
fear it might be misused as evidence against the assumption of an
intermediate stage with o-timbre between unweakened /e/ and zero. I am not
falling prey to that trick again.

> >> Yet you won't agree that reconstructing initial consonant
> >> clusters in a pre-Syncope stage is unnecessarily complex?

No, I told you already.

> >> I think you do agree, but you'd rather not admit it.

> > No, a priori it is only a possibility, and its degree of
> > complexity is not deterring,
>
> Yes, it *is* deterring for the simple fact that we can't
> distinguish a truely old pre-Syncope initial cluster, if
> there were any, from one genuinely caused by Syncope.

That is only right if you look the other way every time I bring in the
evidence of the infixal o-vocalism which indeed is diagnostic. It may be
right or wrong, but it does purport to decide precisely the question you
declare yourself unable to answer. It is a simple lie that I have not
produced evidence to elucidate this matter.

> Therefore, to be unbiased and scientific, we dare not make
> a distinction that we aren't qualified to make until we
> have some predictable method of doing this based on logic
> and facts.

But that is exactly what you are doing when you declare that, out of two
possibilities that were theoretically open before evidence was produced,
it is the one that opts for absence of unseen complication that is
correct. I could accept "... that is to be preferred", meaning some
division in the order of 60% : 40% or even 90% : 10%, but your insistence
that the priority is 100% : 0% is outright foolish. It is like not looking
when passing at a green light - you get killed the minute the unexpected
happens.

> Since we damn well know that some initial
> clusters were created out of syncope like *dyeu- for
> example, being transparently based on *dei-, we must
> unbiasedly presume that all initial consonant clusters
> were created in the same manner until evidence presents
> itself contradicting that hypothesis.

No. What we can't check (or don't bother to check, or won't accept any
checking of, as the case is with you), we can't know and shouldn't make
statements about based on ignorance.

> So far I've seen
> no such proof and everything that shows otherwise.

This is a lie. I have told you a number of times how we can indeed decide
this, and that the deciding happens to contradict your dreams. It is empty
rhetorics that you have seen everything showing otherwise; you have seen
nothing, I can even quote your own words to the effect that you have no
evidence one way or the other. You are generalizing this ignorance to the
entire field. You should not unfairly include me in this, for I can do
better.

> It is
> entirely bias to assume without reason that *stex _wasn't_
> produced by Syncope when other roots with initial
> clustering clearly were.

Not entirely, but I do not believe I have said that. If I am to express
myself on that point, the answer comes out in favour of a monosyllabic
root. The deciding evidence is the OP causative form sta:ya- 'make stand'
which can be regular IE *stoH2-eye-, but not **sotH2-eye-. If it is not
analogical the form shows that the first vowel of the root followed the
/t/. There are other reasons to believe that the first vowel was also the
last, i.e. only, vowel of the root. If the structure /sta:ya-/ *is*
analogical, there may be no evidence, and surely that is not license to
declare the opposite as you constantly do.

>
> So, if you wish to claim what you claim, you must support
> it with something tangible because you can't insist on a
> distinction such as this that is otherwise unfounded and
> arbitrary.

Nice to be reminded that it isn't.

> How do you know that *stex- didn't have a lost
> vowel that broke up the cluster when others did? If you
> don't know, how can you assume it? You can't.

The story of the infixal o is all I've got. It is more than zero, which is
the position of the opposition. It still may be illusory though. What
"others did" what? I do not think any roots had vowels before the root
vowel we see in the full grade, but admittedly that is an extrapolation
which should not be insisted upon. Still, the movement of the infixal o
has yet to reveal a slot for it to move to positioned earlier in the root
than the full-grade vowel. Failing this I see no reason to assume multiple
vowels in roots, but the minute I see evidence for it I am ready to change
my mind accordingly. You may take this as a challenge to give me some
evidence.

> (Hold that thought. There's more on that below in relation
> to O-fix/a-Epenthesis.)

> > so there is no reason not to keep it open.
>
> The distinction that you make is unfounded and this
> added complexity is unmotivated. Complexity that
> is unmotivated is dismissable until proven to be
> required. You haven't proved it. Therefore, while a
> possibility I do admit, your view is dismissable.

It is unwise to dismiss possibilities. I dismiss theories when they have
ceased to be possible because we have come to know too much for them to
have any real chance anymore. That is plainly not the way you work.

> > Are you positing pre-ablaut *dheweghe-meneH1ené- for IE
> > *dhugh-m.H1nó-, Sanskrit duha:ná-?
>
> That's a bit ridiculous. This question assumes that
> *dHugHmhno- is a pre-Syncope stem for one thing. In
> opposition, this word is not sufficiently ancient in my
> mind because it contains a complex polysynthetic suffix
> *-mhno-, suggesting that this word is restricted to the
> final and post-Syncope stage of IE. The ending's
> etymology (down below) may also date it to relatively
> recent times.

The working of the ablaut on the input to the form proves that the
structure existed at the time the ablaut operated. So, if there were roots
ending in clusters (-wgh- in this case), and there were suffixes
beginning with consonants (verbal-noun forming -men- in this case), then
there must also have been triconsonantal clusters before the Schwundablaut
operated.

> Also, we shouldn't expect the mediopassive
> mood itself to have been necessarily very ancient.

Ancient enough: the middle-voice forms have been operated upon by the
ablaut, so they obviously existed at the time we are talking about.

> From
> what I've concluded so far, the mediopassive, while
> starting to form in MIE (although only as a phrasal
> pattern), still must not have been as extensive as it
> came to be.

The middle-voice forms were partly adapted to the endings of the active
voice already in pre-PIE times, but the basis was plainly a separate
middle-voice inflection already. I see no sign of deficiency in the
middle-voice inflection, and I do not see what can be taken to indicate a
smaller extension of the category for the time of the operation of the
ablaut. I find it completely unfounded to say that it must have had a more
limited extension at that time. This also answers the next paragraphs in
your posting.

>
> So there is another reason for me to be suspicious of the
> age of your given stem. Of course though, you'll without
> a doubt reject my analysis if I simply leave you with my
> MIE-based arguement. That's fair actually. So...
>
> If one visualizes or draws out the verbal system of
> Reconstructed IE on paper in grid form, we see a larger,
> outer layer of active-middle contrast that must surely be
> most recent in formation and it envelops the more internal
> contrast of durative-aorist-perfect which then must be more
> ancient (quite ancient actually). Gaps and inconsistencies
> in the more outlying contrast, specifically gaps in the
> mediopassive mood, also support my conclusion that the
> mediopassive and hence *-mhno- cannot be reconstructed
> very far back at all.

I am not aware of any gaps in the mediopassive inflection. It must have
been the Schwundablaut that caused the reduction of the postposition *H1en
'in' to *-H1n- when the collocation of verbal noun + postposition was
turned into an adjective by the addition of the thematic vowel. This is
the only analysis I have seen of the form that makes sense; it was made by
Fabrice Cavoto. The postposition *H1en does not appear to show zero-grade
elsewhere, so this unproductive allomorph of it must have been created
precisely here where it is phonetically regular. And it can only have been
created by vowel weakening resulting in zero-grade vocalism in unaccented
position, i.e. exactly the process we are talking about.

>
> It is perhaps most fair to restrict oneself to roots,
> rather than derivative stems when discussing and
> demonstrating the extent of Syncope.

No, that is silly. Roots have no accentual oppositions, they can only go
into the zero-grade if they are combined with something else that can take
the accent, that in fact is the whole point you were trying to make
yourself. Therefore, ablaut can only be observed and studied in inflected
forms and derivatives.

> The zeroed root
> *dHugH- itself can be derived from MIE *deuga which is in
> fact the 3ps of the verb.

No.

> The final vowel of this verb (which
> automatically lacked vocalic contrast in final position
> as opposed to initio-medial vowels where both *e and *a
> are required to make sense of IE) was retained due to Suffix
> Resistance, an avoidance of desyllabicization or annihiliation
> of one-syllable suffixes like 3ps *-a (> thematic *-e/o-) or
> participle endings *-ta (> *-to-) & *-na (> *-no-) for
> example.

No.

> Hence the survival of a contrast between athematic
> and thematic verbs from the MIE stage.

No.

> Looking at it another way, a grammatical analysis of MIE
> would bring us to the conclusion that the 3ps of a verb,
> whether it be ending in a vowel (*kWera "she creates",
> *palewa "it rains") or not (*wes "he remains", *ei "she
> goes") lacked any suffix for person.

No, they all have one.

> However the thematic
> vowel obviously came to be identified as a durative affix by
> the time of Syncope. So Suffix Resistance preserved what
> would have been an otherwise lost vowel.

No. The use of the conjunction "so" is an act of fraud.

> Surviving instances
> of MIE *a regularly become eLIE *& and later IE *e or *o
> depending on the voicing of the following consonant.

This is my rule applying to the thematic vowel and to that only. I am not
opposed to that part of the story.

> Since
> the resulting thematic eLIE *-&- came to be seen as a durative
> marker, it was dropped in non-durative formations or outside
> the conjugational paradigm when forming nouns, adjectives,
> etc.

There are many thematic stems that are not specifically durative. But if
it means something to you, the Skt. mid.ptc. duha:na- *is* of the present
aspect, i.e. durative. I am not pressing this point, for it is not
relevant at all.

> So the stem *dHugHmhno- by that token cannot be
> reconstructed to MIE as a whole.

If the form (along with many others) is embarrassing to certain
misconceptions of the history of PIE they should be diagnosed for what
they are and given up.

> If it were to exist in
> MIE based on the etymology of it, it would have to have been
> something crazy like **dauga-na-?an-ása (*-(a)na [locative]
> > *-om [gen pl], *?an (?) [agent] > *-hon- and *-(a)sa
> [genitive] > *-os [gen sg]) but we cannot connect it directly
> with *dHugHmhno- then by any sane rules and it would have
> become **dHughnhons- instead.

No, "then" we cannot and should not. But if we correct the basis we can,
and I see no reason why we shouldn't.

> I have no doubt that this
> word is too recent to be of relevance.

There is nothing in the form, let alone in your arguments, that tells me
so. The form indicates the oppposite.

>
> > What evidence is there for the root-final vowel in this?
> > What evidence is there for a root-final vowel anywhere?
>
> The presence or absence of thematic vowel in conjugation
> is the remnant of pre-Syncope *-a in the verb root. Thus
> *es-t < MIE *es while *bHer-e-t < MIE *bera. So the
> thematic vowel is the evidence.

If that were anywhere near true, thematic stems should not coexist with
athematic stems of the same morphemes. But they do, and the thematic
extension means something. With the verb it forms subjunctives, and with
nouns it forms adjectives.

>
> With a root like *leikW- "leave", the MIE equivalent
> necessitates a terminating vowel based on syllabics:
> *leikWa. Since there was a difference between MIE durative
> endings *-em/*-es/*-a and aorist *-am/*-as/(-a), Syncope
> caused a standardized loss of *a in the aorist (hence
> root-aorists). Just after Syncope the durative was
> *-&m/*-&s/*-&t in the singular versus aorist *-m/*-s/NIL,
> producing 3ps *leikW > *likW-t. (Btw, zero-grading of some
> aorist roots was a later, analogical process.)

The 3sg active of the root aorist has full grade. This is outrageous. The
thematic vowel is not a durative-aspect marker at all.

>
> > And what evidence is there for a vowel between /n/ and
> > /H1/ here?
>
> The suffix -mhno- is a composite suffix. Since there are
> many such composite suffixes recognized in IE, my
> statement is well-[gr/f]ounded.
>
> Here, we should recognize the ending as a string of *-m-,
> *-hn- and an adjectivizing accented thematic.

"Adjectivizing accented thematic", very good!

> The first,
> *-(o)m- is the genitive plural normally used to derive
> collective nouns (*yug-om = "that/those which is/are
> yoked").

Nonsense.

> It is optionally reinforced by *-hon-.

Nonsense. The structure *mluH-mnH1n-o-, Vedic. bruva:na- 'being spoken',
is the obvious product of *mlewH-men- 'act of speaking' + postposition
*H1en 'in' + adjectivizing morpheme *-e/o- (Cavoto). Without the
postposition the form would be either the accusative or the locative of
the verbal noun; with the postposition the form would mean something like
'en parlant', and with the thematic vowel it will mean 'who/what is in
speech' referring to the object of a speech act.

> Both
> *-om and *-hon- were zero-graded according to normal
> Late IE quantitative ablaut rules because of the
> accent-stealing thematic vowel.

Correct concerning the phonetic principle, but the underlying forms are
unacceptable.

> The combined effect is
> a participle ending *-mhno- but while the suffix itself
> may conceivably derive from a hypothetical MIE form (even
> though I really don't think it is), there's no guarantee
> when the suffix was applied to the verb here or whether
> it follows ancient or more recent rules. Chances are,
> the rules the word demonstrates are relatively recent.

It follows the most archaic rules we know in deleting one of the nasals of
the underlying form *-men-H1en-o- > *-mnH1no- > *-mH1no-.

> So one could never derive any sensible MIE form out of
> *dHugH-mhno- as a whole based on both its syllabics and
> internally reconstructed morphology but we can predict
> the MIE equivalent of *dHeugH- itself without problems.

You give no evidence to make us believe that complicated roots did not
form participles as early as lighter roots. It takes a *reason* to make me
believe that. You are in effect disregarding half the language and making
your rules in a way that accomodates only the part of the language you
like.

>
> >> I in no way stated that IE itself cannot tolerate such
> >> mediofinal clusters. I merely stated that the stage previous
> >> (the Mid IE stage) only tolerated a CV(C) syllable structure.
>
> > I understood that and objected to it, referring to pre-ablaut
> > pre-PIE.
>
> Since you misunderstand lots of things in your comments
> above about my theory, you can hardly object to something
> you don't understand yet, can you?

This is unfair again. You cannot object to being understood, and that's
what you were this time.

> Part of it is my fault
> because I really need to put up a website on my current
> ideas so that I can refer them to it. There's a lot to it.

A lot of what?

>
> > If the roots are really *steH2- and *ped- they would both
> > be monosyllabic. That could be a principle which you would
> > miss if you insist on a choice based only on a dream of
> > simplicity. When it is observed that the o-infix formations
> > move the prefix into the position where the monosyllabic
> > theory places the root vowel it is confirmed that that
> > vowel was indeed the first vowel in the root, and that
> > some roots consequently had initial clusters already in
> > pre-ablaut times.
>
> An irrelevant arguement. The stem *stex- could never be at
> risk of a-Epenthesis and the o-grade *ste-stox- is merely
> ablaut, a more ancient and seperate process that has
> nothing to do with this phonotactic restructuring or
> "affixing".

An irrelevant objection, for, unlike you, I am taking pains not to make
such mistakes. Bringing in the o-vocalism of the perfect is a gross
mistake. The perfect does not have the o-vocalism of the infix formations,
These things should be kept separate.

> We must seperate instances of *o-grade which
> were used for the perfect-stative in conjugation and
> those that clearly have nothing to do with that.

Yes, the perfect should be kept *out* of it! The infixal o appears in a
totally different set of derivatives. You are only making it worse in the
following.

> For
> example, there is little that's intrinsically stative or
> perfect about *ohWuyom, especially considering that it's
> not a normal looking *o-grade verb if it were so. It's also
> impossible to ignore the probable connection with the
> noun *hWawi-. This surely cannot be confused with the
> perfect *o-grade.

Well, it's not me doing it.

> Similarly, *osdo-, while based on *sed-
> is clearly not based on what would be its normal o-grade,
> *sod-. So you are lumping o-grades together that don't fit.

I am notorious for doing just the opposite. People all over the place ask
me why I want them to distinguish between different *kinds* of "o-grade".
My answer is that the linguistic material simply demands that. This is
apparently one point in which I have got through to you; only, it is as if
you have read the headline but not the real message.

> Mixing *o-grades caused by syllabics versus those caused by
> simple vowel alternations in conjugational paradigms is
> haphazard.

Syllable structure does not create IE /o/ of any kind. And there are no
"simple vowel alternations in conjugational paradigms" of relevance for
this. There is a change of underlying /e/ to /o/ in the strong forms of a
number of reduplicated verbal formations, the apparent rationale of which
is dissimilation. But thank you for spelling this out to the extent it is
acceptable. It is a craze in IE Studies at the moment to ascribe a
laryngeal-deleting effect to *every* IE /o/, which is just a mistake that
may stay with us for a long time if we do not fight it well enough. It has
been here too long already.

> The O-fix, in fact, if anything is a better
> example of a theory that is overly simplistic which
> demands further complexity to account for these facts.

You are not saying what facts you mean, or I have missed it in the fray. I
am used to be criticized for making this thing *too* complicated; since it
now is the opposite that is wrong, I may perhaps be doing it about right.

> In eLIE, just after Syncope, *stex- is to be trivially
> reconstructed as *steh-, so there is clearly then no
> motivation to apply a-Epenthesis to restructure the
> unproblematic syllabics.

Of course there is no motivation to apply "a-Epenthesis", for there is no
such phenomenon in IE. That is only a dream of yours, a terribly wrong one
at that.

> The a-Epenthesis (or the
> unmodified O-fix) does not bar us from formulating an
> MIE stem for *stex- that conforms to an ordered CV(C)
> pattern. We in fact have two logical options: *asteha
> or *sateha, both conforming to CV(C). Take your pick.

I am dismissing both if the evidence I see is conclusive; if it is not, I
am accepting neither.

>
> Another thing. Since your theory as is apparently lacks
> a credible motivation for the infixing in the first place
> despite the sober rationality of the analysis, we must
> remedy the glaring flaw.

I see what's coming; forget it, it just is not phonetically conditioned.

> We then see that syllabic theory ellucidates on the
> cause of this process because it can be predicted by the
> syllabics of the word at that point in time when and
> where to apply the "O-fix".

It cannot, the infixal o-form is opposed ot zero-grade of the same element
(minus the infix, that is).

> There is a resistance against
> clusters of more than two consonants, putting aside the
> pseudofix *s- of course. The a-Epenthesis process (merely
> a modified O-fix) shows that an eLIE syllable then could
> only be shaped as (C)CV(C)(C), not quite as Japanese-like
> as MIE but not as complex as we see in later words like
> *bHe:rst (CVCCC).

The complexity of a putative "*bHe:rst" is at the end, but the idiotic
"a-Epenthesis" is supposed to alleviate complexity in initial position.
There is plainly no rhyme or reason in this.

>
> Since the necessarily modified O-fix rule shows a
> syllabically simpler IE in the past, your rule when
> improved in fact seems to betray your arguement.

This is silly, you are reasoning as if the history of PIE lasted only two
days, with one old and one later stage. If you do not mean that, you are
not reasoning at all.

>
>
> > Let me be clearer still: You are confusing the theoretical
> > maximum with the applicable maximum.
>
> I'm aware of the two.

Then you should act accordingly.

> On the other hand, you're confusing
> theory with reality.

You've got that right. I am in fact trying to uncover the real past, not
just have a dream come true.

> With theory, we attempt to approximate
> reality but we can do no more than approximate it as best
> we can. It would be like making a polyhedron with as many
> sides as is possible in order to approximate a sphere.
>
> Your complaint is by analogy like complaining that a cube
> isn't a sphere. Well, of course a cube isn't a sphere, but
> a theory has to start at the beginning before it can
> properly evolve and adapt to new facts. There is simply
> no other place to start but at the bottom and the only
> reason why it's lonely at the top is because no one's
> got there yet :)

I am not complaining about approximations per se. I do criticize the
formulation of crude approximations on points where we already know
better. The field is in possession of a lot of pertinent information which
you do not use, and apparently do not want to even notice.

>
> > You should not assume greater simplicity than the facts
> > allow,
>
> While I appreciate and even agree with your methodological
> stance in principle, you're now speaking in general terms,
> void of specifics as to what exactly you object to in my
> theory.

Forgot your message? You insist that all initial clusters have lost vowels
between the consonants. If you assume they have and still see the o-infix
moving past the first slot to find its position only after the second
consonant of specific roots, then you are disregarding important
information, even to the point that you formulate bombastic statements in
glaring contradiction of the facts.

> If the previous text was proof against something,
> I failed to see the strength in it. If anything, you seem
> to be wanting to overcomplicate things for no reason.

Oh, now you joined the crowd. Some will welcome you there. I care little
which way you choose to get it wrong.

> > but you do, even before you take a proper look at the
> > facts. And some of the time you have bad luck.
>
> I think it's fair to say that we *all*
> have "bad luck". It's the natural part of learning and
> discovery. Even artificially intelligent programs make
> errors and their mistakes are the only things that
> help them learn. It's only logical, really. I will have
> no shame in being burned by the fire that elders told
> me not to touch if it causes me to learn from it.

I don't see you learning even when given the best of opportunities by
being told flat out. Or are yoy saying, in essence, that your opposition
is only for the sake of the argument, just a sort of parlour game, and
that you do not really mean what you say?

>
> > [Reconstructed Indo-European] incidentally can also have
> > more than two consonants in medial position, even before
> > the operation of the ablaut, so that point of your account
> > is also incorrect.
>
> To say that because I think that rules in IE itself don't
> apply to an earlier stage makes my theory wrong is
> completely absurd. Yes I can see that *wertmn has three
> consonants in a row. So what?

You implied that words couldn't have that.

> I can also see that it is a
> derived stem that may or may not be reconstructable to MIE.

A structure *wert-men must be reconstructable for the immediate pre-ablaut
stage of which we are talking here. The inflectional accent never moves to
a position *between* the final consonant of a root and a form of the
suffix *-men-, so there was no hidden vowel in that position. Then this
word had three underlying consonants in a row even before the operation of
the Schwundablaut. And if this particular lexeme was not formed yet (which
you can't know), surely others of the same shape were. It is not a
terribly important point to me, you just chose it to be yet another place
where you insist on messing things up.

> Apparently you ignore this purposely. Again, you're using
> derived stems to prove me wrong in an illogical fashion.
> The root *wert- by itself is indeed pressable into my
> CV(C) grid. A root like that would suggest MIE *werta.

No, the "*-a" was not there.

> Being that we are, at least for the purposes of proving
> or disproving each others views, both ignorant of the
> exact age of derived stems like *dHugHmhno- or *wertmn,
> they cannot be allowed in this courtroom. Agreed?

No, you are speaking only for yourself.

> We should stick to roots since they are the heart and
> basis of later IE.
>
>
> > My biased imagination actually lost. I used to dream some
> > of the dreams you still believe, [...]
>
> Well, I'm glad that you found God but Confuscious say
> "A man not aware of his dreams is asleep." I think it
> was Confuscious. No wait, maybe it was in a fortune
> cookie I ate from Jim's Chinese Take-Out. Well, whatever.
> You gotta take wisdom where you can find it nowdays :\

You show very little readiness to take any wisdom that is handed to you. I
have taken pains to consider all of your arguments and pseudoarguments on
their own merits. I have found little of any value.

Jens