Re: [tied] IE *de:(y)- 'bind'.

--- In cybalist@yahoogroups.com, "Patrick Ryan" <proto-
language@...> wrote:

> Old Indian Sráyati, 'leans', suggests durative for the *-y to

me. Do you disagree?

Yes, I disagree. All forms of 'lean' have the /y/, be they durative
or punctual (or uncommitted). The stem s'ray-a- is a typical example
of the creation of a secondary present stem on the basis of the
subjunctive of the aorist, which is then in this case a root aorist,
explicable only from an original punctual meaning of the root *k^ley-
as something 'take up a leaning posture, take to slant'. The old
present stem had a nasal infix, as Gk. klí(:)no:, Lat. cli:no:, OHG
hline:n show.

>
> *k^el- is interesting, is it not? It looks like it has no

attested root present or root aorist, and instead is building its
inflection with -*to, which I know as a future/prospective formant.

From what material do you "know" that? What "inflection" are you
talking about? The dental extension seen in PGmc. *haltha-/*halda-
'slanting'? How do you know that is based on the inflection of a
verb? And if so, why is the *-to- not just the participle morpheme?
There is no -t- in Lith s^alìs 'side'. I can't make anything
interesting of a combination between *k^el- and *k^ley-. The fact
just seems to be that the IE roots are old words that have come to
have that form before they were expanded by the suffixal and
inflectional material of later stages. In principle it might be
possible to acquire some impression of the functional services
rendered by the root extensions (of such they are - the short forms
could also have lost the material not appearing in them), but if the
two variants means exactly the same, there is nothing to go on.
19 out of 130 76 out of 600

> Barring a final *-H or *-y, I assume that any *CVC root in PIE

is aorist (though there will probably be the apparent exception).

There are many exceptions. In fact, the statement is nowhere near
correct: *h1es- 'be', *h1ey- 'go', *bheH2- 'speak; shine', *ses-
'sleep', *sep- 'attend to', *yeH2- 'go', *wek^- 'wish', *tekW-
'run', *k^ey- 'lie' and no doubt others do form well-established
root presents. I count in the index of LIV 130 secured or suggested
root presents, 19 of which are from roots of the minimal structure
CVC. For root aorists, I count 600, 76 being from minimal roots.
Thus, most roots just are punctual, and so of course there are also
more punctual roots made from any given structure that one might be
interested in (for whatever reason). But actually the percentage of
CVC roots is lower among the aorists than among the presents, viz.
12.6 versus 14.6. Now that does not make me go crazy and claim the
opposite, for there is no reason to assume that this has any reason
associated with the root structures at all. Why would there be an
absolutely equal distribution at all? Most other things are not
equally big, why would these two groups be?

>
> I think the more productive approach is for me to put this forth

as a hypothesis to be tested by your selecting a *CVC root (not
ending in *-H or *-y that you believe is demonstrably _not_ aorist.

>
> What do you think?

I think it's a waste of time, but I just did that. You may respond
to it if you like.

> I have temporarily misplaced my Whitney. What is the aorist

recorded for *k^lei-, if you do not mind telling me?

3sg ás'ret, 3pl ás'riyan.

> If rules operated absolutely faultlessly, there would be no need

for professors to explain the exceptions.

>
> While I do claim that *k^ley- is inherently durative, I do not

rule out the possibility that it was lexicalized, and so could form
a root aorist.

> ***

Why do you keep on using the pseudo-technical word "lexicalized"
every time a form turns out to be an aorist against out predictions?
Lexicalizations are typically retentions of an older state of
affairs, so in such cases your statements about the original
situation should work even better.

>
> >
> > > When PIE started forming roots in *CVC, any *CVC that
> reflected
> > an earlier stem extension of a *CV root, would have been
> > lexicalized, and redefined as an aorist/injunctive.
> >
> > And what is *that* statement based on?
> >
> > ***
> > Patrick writes:
> >
> > The extraordinary lengths to which IE seems constrained to

form

> presents.
>
> I do not even understand the English of that sentence, What do

you

> mean?
>
> ***
> Patrick writes:
>
> There is no PIE tense that has so many different tense forming

devices as the present.

> ***

There are basically four: (1) Unmarked, i.e. root present, made
directly from roots whose semantics was durative already. (2)
Reduplicated, made from aoristic roots expressing actions which if
prolonged would involve repetition (*dhi-dhéH1-/*dhé-dH1- "be
putting several things"). (3) Nasal presents, which were originally
causatives, originally used in the middle voice to express a
roundabout action *k^l-n-u-tór 'is being occasioned to listen,
listens', OIr. ro-cluinethar). (4) ye/o-presents, apparently adding
a durative note to the bare meaning expressed by the root-aorist:
*gWm-yé-ti 'is coming' : aor. *e gWém-t 'came'. The inchoative
sk^e/o-presents which go with s-aorist apparently reflect the simple
addition of the *-s- and *-ye/o- the result of which seems to be
precisely *-sk^e/o-. The thematic present type *bhér-e-ti is in
origin only the subjunctive of the root aorist; the zero-grade
thematic of Skt. tudáti is a secodanry thematicization of an
athematic structure, either a root present or a root aorist (often
taken to be always from a root aorist, but that seems to be based on
a hasty impression of the simple fact that most roots are aoristic).
Other types are combinations or special changes that took place with
special root structures.

>
> <snip>
>
> > This is not only a pattern in IE but also in Sumerian, where
> durative/present notice require endings not needed for the
> aorist/past.
> >
> > Same for Egyptian: the oldest and simplest sDm=f form is a

past

> narrative.
>
> Not necessarily so for IE: An unmarked stem is a root aorist

only in

> the case of roots with inherently punctual semantics; with such
> roots the durative ("present stem") must be marked (by
> reduplication, nasal infixation, *-ye/o-, or suppletion). If the
> semantics of the root is durative in itself, the unmarked root

is

> used as a root present without any further marking; in that case

the

> punctual aspect ("aorist") must be marked (by *-s-,

reduplication,

> or suppletion).
>
> ***
> Patrick writes:
>
> I am sorry to have to doubt that. *pleH-, 'fill, forms a root

aorist ápra:t to be sure but the present tense is reduplicated *pel-
(píparti) not reduplicated *pleH-.

>
> I see nothing inherently aorist in 'pour'.
> ***

There is if it means to fill water into something which is empty
before you begin. The function of the aorist is to change the
situation, while the present aspect elaborates the situation already
given. You can't fill a pot without changing the situation - unless
you have already filled some of it, in which case the present stem
applies when you go on.

>
> >
> > And theoretically, also. I believe that nouns preceded verbs

as

> a category in all these languages. And in all these languages,
> singular is unmarked, which ties into a punctual (aorist) verb
> simple form.
> > ***
>
> I do not see the relevance of that statement. What do you mean?
>
> ***
> Patrick writes:
>
> I mean, at an early date, there were _no_ verbs per se/
> ***

I actually agree that the verbs we know are analyzable in terms of
original nouns, and their syntax then as that of nominal sentences.
That, however, does not exclude the earlier existence of real verbs
that have been lost when the nouns took over. A case of this is sen
in Modern Indic where the verb has been practically lost and
replaced by old nominal circumlocutions; that is no valid reason to
deny the existence of finite verbs in Sanskrit. As the nominal
phrases are growing into new verbs in Modern Indic, someone might
say that verbs are secondary in comparison with nouns; he would be
wrong.

> > For these pre-Nostratic roots, I think the only formant that
> could explain *H in most cases is *?a, stative.
>
> Why are you so occupied with explaining something which at most

does

> not demand an explanation at all? All we see is a root having a
> total of three consonants; there are many of those, and the fact
> that they exist does not pose a problem or call for an analysis.
> This is not affected in any essential way by the fact that one

or

> the other of the consonants of *deH1y- fails to show under

special

> conditions which are just regular points of neutralization.
>
> You need
> to demonstrate that some of the material of the root *deH1y- is

not

> used in forms where it *could* have been present.
>
> Only in the case
> of a serious demonstration of that kind can I seriously consider
> ascribing morphematic status to some of the material appearing

in

> the longer root form. I have not seen any such demonstration. I

have

> read and listened to hundreds of attempts at proving such

analyses,

> but it has been just too pitifully easy to dismantle them.
>
> ***
> Patrick writes:
>
> In my opinion, that is what dyáti does.
> ***

Again, I don't understand you: *What* does dyáti do which is
important here? Does dyáti show *absence* of some of the material of
*deH1y- under conditions where the absent material would have been
retained if it had been present in the first place? You may note
that the single attestation of the stem in the Rigveda, as I have
repeatedly pointed out, shows *presence* of all of its material, for
[adias] can only reflect *e-dH1i-e-s. A form without the laryngeal,
i.e. ***e-dy-e-s could not vocalize the /y/.

> >
> > However, having formed a stative, the root was redefined as

*CVC

> in PIE so that *deH- (and *daH-, etc.) were only felt as

statives

> selectively; and became lexicalized. Under those circumstance, a
> *CVH- root was regarded as full par with another root that might

be

> *CVr-, *CVn-, *CVl-, using other (originally) formants, or *CVbh-

.

>
> I do not understand these sentences, could you rephrase?
>
>
> ***
> Patrick writes:
>
> Certainly. The original impulse for *CVH (*CV?) was to form a

stative of a given *CV root.

I don't understand that either.

> At the PIE stage, this stative nuance was only selectively felt

so that *CVH was generally regarded as a punctual verb just as any
other *CVC.

So they marked the stative nuance, but were not really understood?
How come *you* understand it so much better? What source of
information gave you the insight that enabled you to make this funny
assessment?

> *CVbh- constitutes *CV + *bh, a compound; hence a punctual verb

root or root noun. *-bh forms no aspectual extension of a verbal
root - ever.

>
> *CVr-, *CVn-, *CVl- were also lexicalized when, in the majority

of cases, they derived from *CV stem modifications like *-H and *-y:
*-r, factitive, *-n and *-l, inanimate and animate ingressive. That
is not to say that these three could not also occur as genuine
compounds as in PIE *del-, 'split', pre-Nostratic *da-l-, 'hand-move
back and forth = cut'; Sumerian dal; Egyptian dn.

You're not getting through to me. You are obviously talking about
some analysis consisting in dissection of the IE roots that I have
never heard of, much less seen the need of myself. I hesitate to ask
you to be more specific, but I'm afraid there's no other way. So
what *do* you mean?

>
> > One has only to look at Slavic to realize that inflection

has

> been heaped on weakened inflection heaped on weakened inflection.
>
> What are you thinking of in Slavic that makes you say that, and

what

> relevance does it have here?
>
> ***
> Patrick writes:
>
> The sheer length of the forms.
> ***

I don't see inflection being heaped upon inflection to any
noteworthy degree. Nor do I see the relevance even it were true. I
do not think Slavic wordforms are generally longer than the
wordforms of Sanskrit, Greek, Gaulish or Runic Germanic.

>
> > I saw 'scarce' and assumed incorrectly that you were

referring

> to its distribution rather than its meaning. I also grant that

with

> this meaning, di:ná- is surely to be derived from *daHy-.
>
> I'm afraid that's not safe either. Actually Mayrhofer, following
> Kuiper, tentatively derives it from "DAY1" of Ved.
> dayate 'zerstören, zerfallen lassen', not with DA:4 'divide'

(ava-

> dyati) which I posit as *deH2y-. I'm sorry about the hasty

report of

> what is in fact confusing. The root is "DAY1" is said to be IE
> *deyH1- seen also in Gk. deilós 'timid, miserable'. The meanings

of

> di:ná- given are "spärlich, seicht, nicht tief (Wasser);

schwach,

> gering". The gloss 'scarce' is the first meaning given by Monier-
> Williams. If the root is *deyH1-, its zero-grade should of

course be

> *diH1- without any problems, and then the word has no connection
> with a long-diphthong root.
>
> ***
> Patrick writes:
>
> For our purposes, that really makes no difference to derive it

from *da:-, 'liquid, flow'. It just becomes a reason why this entry
in Pokorny should be revised to *daH-, *daHy-.

Of course it makes a difference! If the root is *deyH1- and the
participle is *diH1-nó-, then the word is not an example of what you
are talking about. Then there never was a sequence /Hy/ in this word.

> I guess you have anticipated my answer. If dayate

means 'zerfallen', I would attribute it to a *day-, connected with
*daH-; and both of these are connected to 'teilen' in this way.

That's what I wrongly read into Mayrhofer's words myself before I
looked more closely. It cannot stand.

> The absolute basal meaning of *da is 'side'. In *daH-, 'liquid',

we have the inanimate 'leak, liquidate';'produce something at the
side'; with 'divide', we have 'make something to be at the side'.

>
> It just struck me that all this suggests *da:nu- is 'glacial

melt' since other good words for 'river' exist.

And if yet another one is found, which one is then 'glacial melt'?
*da:nu- or the new word?

> So, my choice would be *day- / *dahy-.
> ***

So I've noticed.

>
> >
> > But as you point out, *-no would produce an identical form

from

> *deHy-.
> >
> > Also, Pokorny was no man's fool, would he not recognize as I
> instantly do, that the semantics most logically connect this

form

> with *daHy-?
> >
> > Could there be another di:ná- with a meaning appropriate to
> *deHy-? Perhaps not. That would be a smudge on Pokorny's sleeve,
> would it not?
>
> The question has been messed up now to the point that is has

lost

> its meaning.
>
> ***
> Patrick writes:
>
> Sadly, yes.
> ***
>
> > But respectfully, I think you have forgotten the purpose for
> which di:ná was introduced into the discussion.
> >
> > If di:- represents the zero-grade of *deHy-, and I grant

that it

> does, and that is why _I_ introduced it, why do we _not_ find it
> reflected in dyánti, where we expect zero-grade + -ánti? Would

not a

> zero-grade di:- + ánti produce **di(y)ánti?
> > ***
>
> It might, but doesn't have to, for the rule governing it makes
> allowances. In this case it actually does: the single Rigvedic
> attestation <adyas> is metrically [adi(y)as], just as you want

it to

> be. That can certainly not be construed to show that your
> expectations are *not* met.
>
> ***
> Patrick writes:
>
> Excuse me, but a rule that "makes allowances" is really not much

of a "rule".

> ***

Sievers' law - or lawlessness if you prefer - states that semivowels
following long sequences (V:C or VCC) can be vocalized, while they
cannot be vocalized after short sequences (VC). This poses
restrictions on the linguistic variation: dative plurals in -ebhyas
and -a:bhyas appear with the /y/ optionally vocalized or
consonantal, as the poet needed it for his metre; dative plurals in -
ibhyas, -ubhyas of i- and u-stems never vocalize the /y/. There is
no point in denying the existence of such rules - or of anything
that can be established empirically.

> >
> > > ***
> > > Patrick wrote:
> > >
> > > 'Fraid you'd say that.
> >
> > It's up to you to give a different impression if you want us

to

> > have that. You can't just say that you wished there were
> evidence to
> > show that the alleged zero-grade alternants of *deH1y- point

to

> *di-
> > and not to the *d&1y- you expect, you ought to point to such
> > evidence. And when informed that "-&y-" just never occurs
> > unambiguously in Indo-European, you ought to do better than
> insist
> > on some ambiguous reconstructions. The world has moved on

since

> > Pokorny who certainly wrote a masterful book, and one that

can

> still
> > be used today if only one knows how to adjust its

information.

> We do
> > not do any services to good scholarship if we refuse to face

the

> > very possibility that progress that has been achieved since

the

> > publication of a handbook half a century old. Since I am the
> author
> > of some of the analyses and rules that are claimed to be
> instances
> > of real progress in the present matter I of course ought to

show

> > some restraint in evaluating them. But I cannot accept that

they

> are
> > being brushed aside to make way for alternatives that are

based

> on
> > no serious evidence at all.
> >
> > ***
> > Patrick writes:
> >
> > On your first point, di:ná- points to a zero-grade form of
> *da/eHy- so I have demonstrated that.
>
> If you mean that you have demonstrated that di:ná- shows *-&y-

with

> vocalized laryngeal (schwa) followed by /y/ with which the schwa
> forms a diphthong, that just is not correct. Where laryngeals

appear

> in the environment CHyV, i.e. between a consonant and a
> consonantal /y/, the /y/ is vocalized and the laryngeal is kept
> consonantal.
>
> ***
> Patrick writes:
>
> I will pass on this. I really do not understand it.
> ***

That's too bad, for you were the one making a point based on it, and
now when it is spelled put to you, you do not understand it. Do you
want *me* to defend your mistakes for you? The question boils down
to this: Where do you find [&y] as you claim you have? Where is its
existence proved?

>
> > And, in Old Indian at least, it is *di:- before consonants.
>
> It may sometimes be, but mostly it is /di-/, reflecting IE *dHi-
> before voiced consonants, and *d&- before /t/. Forms with /i:/

may

> be analogical on the full-grade alternant which has /a:/: The

na:-

> verbs have active -ná:-ti, middle -ni:-té, reflecting IE *-né-H-

ti,

> mid. *-n-&-tór.
>
> >
> > Of course, in dyáti, we need a zero-grade before a vowel.

First,

> we might ask why we want a zero-grade here. In dyánti we expect

zero-

> grade; in the third person singular, we normally expect full-

grade,

> is that not correct?
>
> That depends on what present type it is. This is a class VI

present,

> the type tudáti, tudánti. That is generally explained as a
> thematicization of an old root aorist.
>
> ***
> Patrick writes:
>
> Are you claiming that is what dyáti is?
> ***

Yes.

>
> > I would be very interested to know why that is not the case

in

> the third person singular if you have a good idea.
> >
> > We have the form Old Indian gáyati, 'sings', from *geHy-.

This

> seems to me to be the natural response of *CVHy- to a thematic

third

> person singular. Why is it different?
>
> 'Sing' forms the present stem gá:ya- from *géHy-e-. The Sanskrit
> root vowel is long.
>
> > Now I have _repeatedly_ asked you why we see an apparent

zero-

> grade in the third person singular form and you seem to be

avoiding

> an answer. Why is that?
>
> Because it is a class VI present stem.
>
> ***
> Patrick writes:
>
> I was aware of tudáti but also what Beekes writes: "There are

also present tense forms with zero grade in the root, but it is not
certain whether these go back to PIE."

>
> So, it seems, we are back where we started.
>
> I am claiming dyáti, Class VI now, which calls for a zero-grade

of the root, should be *diyáti if derived from *deHy- (*de6y-) based
on the zero-grade di:ná, from whatever it might come, and you
introduce the (in my opinion) ad hoc rule that *-6y resolves always
to *-y.

No. I observe that wherever one may expect schwa + /y/ to appear,
what we really find is the reflex of [Hi]. That is also what we have
in [diyáti] which is *dH1i-é-ti with a subphonemic glide between the
[i] and the following vowel. The [i] itself is (optionally) syllabic
because of the preceding cluster (and only optionally so because
that is the way Sievers' rule works).

But if *your* stand is that dyáti/diáti cannot be of PIE age, you
should not use it for reconstructions that go even further back in
time.

>
> I have confess to being Austrian German which inclines (*klei-)

me to dislike anything ad hoc viscerally. Tut mir herzlich leid.

> ***

I have not noticed any dislike for ad hoc solutions in your own
suggestions. Are you being objective at all?

>
> ***
> > Progress is in the eye of the beholder, is it not? I think

that

> progress would constitute looking beyond PIE to Nostratic and
> earlier. You do not. Am I seeing the future or only my own
> reflection in the mirror? Only time will tell.
>
> I do indeed see potential progress brought about by looking

beyond

> into Nostratic, but one cannot "explain" the obscure by

obscuring it

> even further. If your insight into Nostratic has enabled you to

see

> rules and connections of an enlightening kind, you should be

able to

> show us that. One does that by pointing to actual observations

and

> explaining their relevance. I have not seen you make any attempt

at

> anything of this nature. Until some sensible-looking evidence is
> produced I certainly refuse to give priority to what must appear

to

> be just empty assertions over the results of enduring and

careful

> analysis applied to the linguistic material we *can* handle.
>
> > I absolutely deny that my alternatives are based on "no

serious

> evidence at all".
>
> Then it's time to produce some. You may talk here.
>
>
> ***
> Patrick writes:
>
> "You may talk here." Jens, you do have a sense of humor!
>
> Can you give me any indication of what might constitute 'proof'

for any point I have made or attempted to make here?

>
> Seriously.

I have plenty of proof against it, so frankly no. Proof positive
would be some word which could only be based on a zero-grade
sequence and then has it surfacing as Latin -ae- or Greek -ei-/-ai-/-
oi- before a consonant, or as Latin -a- or Greek -e-/-a-/-o- before
a vowel (with loss of *-y- between the two vowels). Or the same from
any other IE branch. I know of no such examples. From *da:y-
'teilen' Pokorny gives *d&i-lo- as the immediate preform of OCS
de^lU 'Teil' and Gothic dails. Others posit *daHi-lo-, but that
ought to have developed into *dayH-lo- which would have given an
acute intonation in Slavic which is not found (SbCr. dì`jeo shows
circumflex). Note that Germanic d- is not expected with IE *d-, and
note also Lith. dalìs 'part' without the i-part. There may be
loanword relations involved in this, and the l-part may have formed
an athematic stem, in that case however of a type not immediately
transparent. This is not transparent enough to allow a serious
reconstruction for PIE, much less for states older than that.

Jens