[tied] Re: Monovocalism: sequel

--- In cybalist@yahoogroups.com, enlil@... wrote:

> You continue to ignore important differences between Sanskrit and

IE.

> In Sanskrit, there is a merger of *e, *a and *o no doubt because
> a single vowel in Sanskrit, namely /a/, is attributable to all

three

> IE vowels. In IE, there is no such single vowel. Instead we must

ask

> "What is *e attributable to previously? And what is *o originally?"
> That in itself is sufficiently different from the situation in

Sanskrit

> to negate its application here.

It is quite obvious that *o, to the extent it respresents an IE
original vowel at all, is attributable to the same source as *e, and
that *o is a conditioned variant (perhaps even more than one) while
*e is the unconditioned form. That is totally unconnected with the
question of whether there were other original vowels than *e. If
there was also an original *a as a distinct unit at a stage
preceding the ablaut changes, I think we must confess that we do not
know whether that also alternated with *o by the same rules, or
perhaps did so by other rules, or jus didn't. I have suggested a few
examples of a-roots having o-infix, but that is not the root vowel
and so does not count.

> In IE, there is no "single vowel" evident. Instead we have to

invent one

> to suit our headstrong claims but the evidence shows that there is

a need

> for two vowels to explain the morphophonemics of the language.

If the two vowels are meant to be separate sources of PIE *e and *o,
this is not necessary. If *e and *a as separate root vowels is
meant, it is correct if the *a-roots are not secondary additions. If
they are, there does indeed seem to be only one vowel needed to
account for all that is old.

> All
> the complexity that we see in IE can't possibly come from a single
> vowel at any time within the last thousand or two thousand years

of IE's

> development. It's far too complex. The solution is fundamentally

absurd.

By what standards? The time frame is not known for PIE.
Proto-Germanic *a which even had to stay free of a number of other
vowels in the system, developed into quite many different vowels in
Old English already. Surely a single vowel could produce the
vocalisms we reconstruct for PIE, if *a may be kept out of the
account as a potential late-comer.

> THEMATIC VOWELS
> ---------------
> > It ought to be quoted as *kWér-tu-
>
> Really? Yet we have *kWr-to-. I thought you said this is simple
> thematic alternation. I think this qualifies as "contra-hoc".

It's simple enough to me. The accent on the root in the strong cases
of *kWér-tu-/*kWr-téw- is a concomitant feature of the accent; so is
the zero-grade of the weak cases and of the suffixally accented to-
participle. This is the way IE works when it *is* simple.

> Now, /u/ is a phoneme of the language, so it is not unexpected if
> > there is more than one thing that contains it. That seems to be
> > important for what follows:
>
> Yes, but *-eu-/*-u- is also a _morpheme_, as in *hWr-n-eu- "to

cause

> to move", *bHr-eu- "to boil", *pek-u- "herd", *gWehW-u- "cow" and
> *gWr-u- "heavy". It has some sort of transitive function like *-ex-

,

> probably absolutive somehow, marking the patient of transitive

actions

> and the agent of intransitive ones. Its relationship to

transitivity

> properly explains why we see both *-n-ex- and *-n-eu- for the

causitive.

I first wrote this:

In *bherH1w- 'boil, brew, ferment' the /w/ is a part of the root. In
*pek^-u- it is a suffix of a function which is not clear to me. In
*gWrH2-ú- 'heavy' it is an adjective-forming suffix. "Cow" is
generally posited as a root noun *gWo:w-, but the analysis with a
laryngeal by connection with Gk. bósko: is not unappealing and could
be right. The root analysis of Gk. órnu:mi, Ved. r.n.óti, Hitt.
arnuzzi, supposing they are identical which is far from generally
agreed upon, is very unclear. The w-part also appears in a number of
nominal derivatives, most notably Germ. *arwa-z 'quick' (ON o,rr),
Avest. arva- 'swift', Skt. árvan(t)- 'steed'; this example belongs
to the problematics of the o-infix formations which continue to
supply us with new information. The overall picture of this is
certainly not one of a *shared* morpheme *-eu-/*-u-.


Then I slowly found out what you meant; then I wrote this:

You are very hard to follow when you change standard terminology 180
degrees. I understand from what follows that your word "transitive"
in fact means "passive". So, sure a cow is the one subjected to
grazing (Gk. bósko: is transitive in the active), if that is the
motive underlying the word, and a sheep is the one subjected to
shearing. I have never been aware of this, and it does not seem to
be something one can just read, but many adjectives in *-u- can
indeed be said to have a passive meaning: *tnH2-u- 'tense, thin'
from "stretched", *plH1-u- 'plentiful' from "filled", *H2mg^h-u-
'narrow' from "tightened, strangled" (Gk. ánkho:, Lat. ango:).
Their semantics do make them look like old passive participles. In
the pair, Vedic krát-u- 'skillfulness, strength' : Gk. krat-ú-
s 'strong', we have a substantival variant with post-accentual /u/,
the thing I have adduced to explain *kWér-tu-s as opposed to *kWr-tó-
s. I have no reply as yet to the natural question why the adjectival
form has *-ú- with the underived adjectives, but *-ó- with the to-
participle, nor to the equally burning question why some root
structures just refuse to form u-adjectives, and form them with *-ró-
instead (*H1rudh-ró- 'red', *bhudh-ró- 'alert', pik^-ró- 'sharp', *
H2ug-ró- 'strong', etc).

I am still confused by a wording like "transitive function like *-ex-
, probably absolutive somehow", even given your special relation to
normal terminology. Are you talking about a transitivizing-and-
detransitivizing morpheme? Does your *-ex- mark intransitive use of
a transitive derivative when added to the root of an intransitive
verb? Why would it do such an elaborate thing in one go? Where else
does IE show anything like that?

That "we see both *-n-ex- and *-n-eu- for the causative" is in my
opinion best explained by the analysis that the n-part, which is
what they have in common, originally marked the causative.

> Both are transitive suffixes attached to the durative-marking *n-

infix.

> To further substantiate the claim, we must observe *pek-u-, from
> transitive verb *pek- "to comb" and *-u- marking the patient of the
> action, hence "that which is combed" (suggesting the original

semantics

> of *peku- as specifically "herd of sheep"). In *gWehW-u- "cow",

the verb

> in question is intransitive *gWehW- "to graze" and so the noun is

the

> _agent_ of the action in this case.

The assessment of suffixal -u- as passive is not bad. It may even be
a real breakthrough, and if I come to find it has been I shall not
hesitate to acknowledge your share in this. It may turn out to be a
cousin of the thematic vowel, and of course I won't quote you for
that, but every little step is valuable and deserves to be duly
credited.

> While *-to-/*-ti- is a true alternation, *-tu- in *kWer-tu- is in
> reality the combination of *kWer- "make" + *-t- (substantive) and

this

> suffix *-u- (patient). The resultant meaning will still be almost
> exactly what we see for stems in *-to- and *-ti- and may be

converted

> to an adjective at will by accent alternation. Again, as an

adjective,

> the meaning will be almost if not completely identical with that of
> the *o/*i counterparts.

The form in *-tu- is semantically a perfect substantival counterpart
of the two others: a deed is what is done. This is one of the best-
known action nouns of IE.

> The supposed "thematic *u" is a mirage, void of deeper analysis.

If we give it your analysis it fits perfectly what you do not want
to connect it with. In essence you make the very same analysis, only
perhaps more detailed, and then dismiss it for one of the types. I
think its charm is that it explains all.

>> [Glen]:
> > [...] such as with the transitive to form *-nex-.
>
> Jens:
> > And is there a transitive morpheme in that?
>
> Yes, *-ex- as in *?y-ex- "to go" (*ei-), *duk-ex- "lead" (*deuk-),
> *mn-ex- "to be mindful of" (*men- "to think").

The two Sanskrit roots i- and ya:- 'go' are not connected. The
former is *H1ey- with an initial laryngeal, the latter has none in
that position and is just *yeH2- (if not *yaH- with /a/, in which
case the laryngeal cannot be determined). The reduplicated and
augmented forms of ya:- exclude the acceptance of an initial
laryngeal for this root. Nor by the way is ya:- any more transitive
than i- is, both are typically intransitive verbs. Latin du:co: and
the form seen in compounds like e:-duc-a:-re do not show any
differences of semantics, so why would the -a:- of the longer form
be a marker of transitivity, or of verbal voice at all? The root
form *mneH2- may indeed be transitive (but then not specifically
absolutive), but it seems to stand without support and not as an
observable principle.

> > Both *-ne-w-/*-n-u- and *-ne-H-/*-n-H- have the old
> > factitive/causative morpheme *-n(e)- [...]
>
> No, the *n-infix is said to be a present marker by most.

I didn't make this up. Elizarenkova analyzed Ved. inóti 'sends' as a
causative of éti 'goes', apparently not knowing that the infixed
nasal is quite generally used to form causatives in Hittite. In
Hittite and the rest alike, it is used to form factitives from
adjectives, this giving with u-adjectives the type in *-né-w-. The
nasal infix is used in Germanic and Balto-Slavic to form ingressives
from adjectives, which is what the *middle voice* of the factitive
originally meant. Thus, the nasal infix is not just a present or
durative marker. But for some reason it is not used in the aorist
which is quite regularly of the unmarked kind (root aorist).

> > Adjectives are frequently u-stems, why I just do not really know,
>
> Because you're being stubborn. As I said *-u- is a most possibly
> an absolutive marker. So the *u-stem adjectives are derived from
> nominal stems. We could theoretically do it with *bHébHr-u- "that
> which is brown; beaver" (*bHer- "to be brown, dark") and make
> *bHebHr-ú- "brown; beaver-like" out of it.

Apart from its unpleasant ad hominem note, I find the idea very
appealing. Perhaps it would even be *bhé-bhr-u- 'beaver' : *bhi-bhr-
ú- 'brown'. Adams/Mallory gives that, only with *bhe- throughout,
but the facts they quote do not show it. Even so, the second hyphen
is not very well motivated here; there is little of a serious nature
to show that the -u- does not belong to the root.

> > I only know *-ne-H- from presents made from roots in H, as Ved.
> > krin.á:ti, OIr. crenaid 'buys' from *kWri-ná-H2-ti.
>
> Yes, and this *-ex- is a transitive marker.

No, the *-H2 (sic, not *-eH2) belongs to the root which is *kWreyH2-
in its full grade. All derivatives from it have that segment. Your
analysis makes the illusion that you know a shorter form with
intransitive function. Where did you find that?

> > The w/u of *k^l.-né-w-ti 'hears' is a part of the root *k^lew- as
> > analysed by Saussure.
>
> Yes, but Saussure didn't catch everything. He's long dead now. Let
> go of his ghost and let his spirit free :) The interesting fact is
> that we see *kleu-/*kl-n-eu- as we do. It's tempting giving these
> facts to allow ourselves to segment the verb into *kel- and *-eu-

This is outrageous. The parallel analysis of the three Sanskrit
nasal-infix types is one of the mainstays in the morphophonemic
analysis on which laryngeal theory rests. It is something we cannot
trade for a whim that does not lea to any insights.

...

> Well, that just makes sense both morphologically and semantically.

The

> verb *kal- is "to call out (to someone)", so if *-u- has absolutive
> function *kal-u- would yield "that which is called" (ie: "that

which

> hears the cry") and by making a verb out of it, we get *kl-eu- "to
> hear". We subsequently get an unexpected derivative *kl-n-eu- that
> only makes sense if the stem is analysed in this way, as a

denominal.

We cannot analyse words this way. Analyses must be paralleled to be
acceptable. This one isn't.

> > I do not think you can peel out a common semantic core for the

u's

> > of *pek^-u- and "cow" - certainly not one you can use with the

last

> > element of *kWér-t-u-.
>
> I just did, my analysis supported by various examples, and I even
> supplied an external equivalent in neighbouring Tyrrhenian.

I don't give a hoot for Etruscan, but the passive meaning of *-u-
fits quite nice here. This was most unexpected and very rewarding.

> Concerning a u-reduplication of *ku-kleu-:
> > Right about what? I wasn't asserting anything;
>
> You were and _are_ asserting that thematic vowels alternate not

only

> between *o and *i but *u as well.

Yes.

> The u-reduplication you mentioned
> is logically immaterial to this assertion. However, thank you for
> letting me know of this. It was something that I hadn't really

pondered

> despite being long ago aware of /s'us'rava/.
>

/s'us'rá:va/. Nice to hear.

> > If you insist on calling the vowel of reduplications with -i- or
> > -e- "thematic", then -u- also is, but only in that uncommon

sense.

>
> No, only in a literal, superficial sense. I use the term in an
> etymological sense because this is what we discussing, etymology.

Not understood. The etymology of thematic connects it with theme
which means stem; it's a stem-final and often stem-forming vowel.

> While
> I see that *i and *e/*o even in reduplications derives from a
> non-morphemic 'thematic vowel' (aka schwa), the instances of *u and
> even *u-reduplication cannot be attributed to this stage of Pre-IE
> because they don't derive directly, if at all, from the eLIE

thematic

> vowel.

No, it's not known as schwa either; that term is also engaged for
other purposes already. Reduplications and suffix-final vowels
should be kept apart, at the very least in the terminology. If you
find they have something in common, you can still say so without
rebaptizing any of them.

> The *u-reduplication is simply *i-reduplication with a twist.

Yes, precisely!

>
>
> > Indeed I account for that with the infix theory.
>
> Alright, but there is a danger in adding vowels at will simply to
> repair against what we do not find and before you attack my QAR for
> the same reason, at least here there is an accent pattern that
> predicts a final vowel and that by this addition we regularize the
> accent. The basis for the addition of vowel in QAR is the accent.
> In contrast, by repairing a lack **bHr-o- with infixing, what is

the

> basis? It isn't even an impossible form when we have things like
> *pro. So the o-infixing is purely ad hoc. Empty assertion again.

That is precisely what forces us to accept there is something in
*bhor-ó- which is not in *pro. This is empiry, the opposite of what
you make of it.

>> > Why "syllabic"? Did I say that?
>
> You stated that *-to- was thematic *-t- and *-no- was thematic *-n-

,

> but I fail to see what *-n- would be otherwise in this context!
> If you take away the vowel in *-no-, we get a syllabic *n. ???
> I said "*-r instead" because we don't find a morpheme *-n anywhere.
> We find *-r instead due to Rhotacization in Mid IE of final *-n. So
> if anything, one might relate *-no- to *-r. Even so, I don't think
> they are connected. So what IS your position here, if any? What

were

> you trying to tell me in the first place?

That formations with the suffix *-t- are in complementary
distribution with n-stems, which is a fact. You have made your own
salad out of it.

> > The addition of the thematic morpheme accounts very well for the

model

> > it creates.
>
> I and many others can't entertain that notion for the _fact_ that

there

> is no singular semantic attributable to such a thing. It is a

vowel,

> that's all it is, that's all we see, even if we look deeper into

pre-IE.

> You insist on it; I still don't get a lucid answer as to why.

I don't run a mission here. If you can't sense the function of the
thematic vowel where I believe I can, so be it. I'd say it appears
in modifying derivatives: ordinals are modifying, so are other
adjectives, and so is the subjunctive.

Jens