Re: [tied] The Rise of Feminines (aka Where's Waldo)

On Thu, 6 May 2004 enlil@... wrote:

> Sadly, nobody contested some irrelevancies with which Jens intended
> to oppose my idea of *-ix developping out of a fundamental need to
> distinguish the new feminine from an otherwise homophonous collective
> inanimate in *-x.
>
> However, perhaps this means that I've been elected to be Jens'
> arch-nemesis  >:)
>
>
> Jens:
> > But the only thing that makes a form thematic is the presence of
> > the thematic vowel. Thus, to create the thematic optative you
> > just add the athematic suffix //-yeH1-// to the thematic vowel,
> > this giving /-o-yH1-/. There are no flimsy considerations about
> > animacy to confuse this. If you want a collective from an
> > athematic stem you just add *-H2, as in prs.ptc. Gk. phéront-a,
> > Skt. bhárant-i. And if you want to make it from a thematic stem
> > you do the same, only from a thematic stem, as in *náw-a-H2.
>
> Now we're really confusing everything together into one big chaos
> soup, aren't we? (Or do we call it a "word salad"?)
>
> There exists a chasm between nominal and adjectival declension
> on the one hand and verbal morphology on the other.

I do not think there is any such basic difference, or that "thematic"
means different things with nouns and verbs.

> We cannot
> confuse the two by introducing insignificant examples using *bHer-
> for the simple fact that thematicism is not employed to make
> gender contrasts or any particularly important contrasts as it
> does in nouns and adjectives. So the fact that athematic *-ih-
> is used with thematic verbs doesn't shock me.
>
> I expect it.

That is not relevant.

> >> Apparently you missed stems ending in *-t-i- which form abstract
> >> stems, for example. It may also be at the end of *xawi- "bird"
> >> which is an animal most often found in flocks.
> >
> > That is simply not known. There is no basis for that statement.
>
> Our current understanding of IE morphology is the basis. We see *-i-
> whenever there are collective nouns. Very simple. I can hardly see
> how abstract *-ti- can be anything other than the substantive *-t-
> and a collective *-i-.
>
> An inescapable example is Hittite /dalugasti/ "length" where we see
> a derivative of "long" ending with the aforementioned suffix.

In my phonology the "abstractness" of formations in -i- is nothing but
a mirage: the -i- follows automatically from the lack of accent on the
thematic vowel. The form is made with the "contrastive accent", i.e. a
switch of the accent from the suffix to the initial which turns an
adjective into a substantive. IE adjectives form s-stem abstracts, which
is what the span dalugas- must reflect; from such nouns can be formed
adjectives in the form of to-participles, as Latin modestus, honestus or
augustus; substantivizations of that type have -ti-, as OCS dlUgostI,
Hitt. dalugasti-s. Thus the -i- is not an abstract-forming suffix, much
less a collective marker.

>
> In all, if our goal is to truly understand the morphology that we're
> all reconstructing here (and I for one am certainly interested in
> understanding it rather than using the ol' "we-simply-do-not-know"
> copout) then what other analysis of this word can there be? It can
> only sensibly be based on a verb *dlegH- "to extend" (sometimes with
> the the verbal *n-infix) which has been made into an o-grade aorist
> noun stem *dlogH-(e)s- "extension". Then a composite suffix *-ti- for
> abstract nouns, representing general ideas, was added. This abstract
> suffix is a synthesis of the substantive *-t- and the collective *-i-.
> (Afterall, what is more "general" and "abstract" than a collective?)
> This analysis properly produces the intended meaning: "length".

Well, this *is* word salad. There is no such thing as an "o-grade aorist
noun stem", especially not for anyone whose "goal is to truly understand
the morphology that we're all reconstructing here".

> So I'm afraid we do know because logic dictates that this is the only
> sense we can make of such constructs unless you can come up with
> better. That'll be your challenge for the next post.

Piece of cake.

> > I assume with a long tradition that the collective is the old form
> > of the feminine, only not "animated" or "individualized". If *all*
> > uses of *-H2 are "later", there is no basis for any of them.
>
> Say wha? The *-x was used at an early date even during MIE to
> form inanimate collectives (in the form *-hWa). It wasn't a new
> suffix; it's in fact VERY old and found in Tyrrhenian. That suffix
> was used right to the end. The animatized variant *-ex and its
> sister *-ix had developped in early Late IE, from what I can tell,
> and were always at the very least animate in gender. This is
> what /agricola/ tells us. There is nothing to counter because I'm
> simply restating what we already know about IE grammar! Your position
> contradicts all this.

My position is the same on *-H2 ("*-x"), except for the note of Etruscan
mysticism. My position differs radically with regard to the status of the
longer forms. Both *-e-H2 and *-iH2 are ultimately composed of the
thematic vowel and the collective marker *-H2. I assume that they were
originally accent variants: *-e-H2 accented, *-i-H2 unaccented, though
this is perhaps seen only indirectly: Vedic deva'- shows that devi':- was
thematic, and its inflection with gen. devya':s shows that the feminine
once had a mobile accent. In most formations, however, the suffix *-iH2
marks the feminine of athematic, not thematic, stems, which is very
surprising. I can offer only the explanation that the thematic vowel of
which the -i- is a reduced form is functionally a redundant adjectivizer
(I owe this idea to my student Benedicte Nielsen, currently in Cambridge).
I do not know why feminine adjectives have been redundantly marked as
adjectives and their masculine-neuter companions have not. In the "one"
part of the pronominal inflection that is based on *te-sm-e/o-, fem.
*te-s(m)-yeH2- (if the ad-hoc analysis with loss of -m- in the feminine is
correct), both genders are marked as adjectives; but that analysis has a
few unclear points of its own.

> By the way, a little nuance to this that confirms everything I'm
> saying about early Late IE. You see, the eLIE animate collective in
> *-&x-/*-ix- opposes the regular animate thematic *-&-/*-i- but this
> parallel is blurred after Schwa Diffusion, producing *-ex/*-ix and
> *-o-/*-i-. Something to think about.

I have. It's not true. The morpheme *-iH2 is not the collective
specifically belonging to stems in -i-.

>
> > None of this is anywhere near known: it is specifically not known
> > that the thematic neuter nom.-acc. sg. *-om was originally a
> > genitive plural; the genitive plural does not even *look* that
> > way.
>
> It's called adding two and two. We see thematic adjectives in the
> default case, the nominative, in *-o-s that are indistinguishable
> from genitives;

Not of the same words, except for Hittite.

> we see Anatolian confusing genitives and adjectives

No, but the point could be made that we see Hittite confusing genitive and
nominative of a-stems.

> (a possible IndoIE-Tyrrhenian areal feature, in fact); and we see
> stems ending in *-om conveying collective concepts like *yugom
> "yoke" and *kmtom "hundred".

The form in *-om is the singular of a thematic collective in *-e-H2, not
itself a collective. IE *pedo'm is a single 'footprint', not something
collective.

> We see an opposition between genitive
> singular *-os and genitive plural *-om;

No, there is no evidence for gen.pl. *-om with a short vowel in any
language that can show the difference.

> we also see a similar
> fortuitous opposition of nominative *-s and accusative *-m, two
> pairs of declensional suffixes just ripe for confusion and
> interchange.

Word salad. Is the point that the gen.sg. is originally a nominative, and
the gen.pl. is originally an accusative? What sense would that make? Or is
the intended message that the gen.sg. ending in *-s was animate and the
gen.pl. in *-m was inanimate? The latter point could make some sense,
since the plurals are lower on the scale of animacy than singulars. But if
that is meant, why not say so? And why not explain how it could assume
this particular appearance?

> That confusion produced animate adjectives in *-o-s
> and inanimates in *-o-m which became animate nouns in *-o-s and
> inanimate ones in *-o-m.

I see no confusion in that. It more looks like the intended system.

> Common sense. All the i's are dotted.

Far from it.

>
> > And what nonsense is meant by the argument anyway: neuters did
> > not distinguish sg. and pl., so they just took marked plural
> > endings and used them in their own lawless way?
>
> Exactly. The perceived "singular" and "plural" of collectives is
> semantically the same. The difference between "sky" and "skies"
> or "the water" (as in "the ocean") and "the waters" is trivial.
> This is a crosslinguistic fact.

The crosslinguistic fact is that this opposition does not have to be
marked, for it does not matter all that much. But IE was a language that
did mark the opposition between inanimate singular and collective.
Therefore, that it did not matter is just about the only thing that cannot
be assumed.

>
> So too is the same opposition found in IE. The normal use of *-x
> was simply to collectivize the meaning of a noun, not to pluralize
> per se anymore than English -hood in "neighbourhood". We know that
> inanimates had no special plural form. Speaking of which, a perfect
> example. We say "neighbour" (a natural animate) and with -hood, the
> new word conveys an abstract collection of "neighbours". It may even
> take on new secondary meanings, even more divorced from its derivative,
> such as "the general place in which a collection of neighbours live
> together".
>
> This is exactly like *-x.

Correct, but that was not the point that was made. The contention was that
collective markers add nothing to the function of inanimates. That does
not seem to be adequate for the language here concerned.

> > Nonsense again, these things are remains of things once of greater
> > importance ("that" opposed to "one", being generalized). What is being
> > suggested here is deliberate *introduction* of an irregularity without
> > any purpose.
>
> The purpose is to keep animacy and inanimacy seperate. To keep the
> gender contrast intact.

That really *is* wrong. You wrote: "In that way, there is no paradox
because *newax would then have only been used as a _feminine_ adjective
with the neuter always being *newom, singular AND plural." But if the
neuter plural had not been formed at the time the feminine was, your
entire position is virtually given up. You have been fighting to disprove
the antiquity of the feminine vis-a-vis the collective, and now - as an
argument - the opposite is being insisted upon. And that really is an
impossible position, if the feminine is supposed to have been absent from
Anatolian, for the neuter plural in *-a-H2 is present in Anatolian.

> > That's neither here nor there. If the added *-H2 conveyed the
> > idea of the feminine (even "feminine athematic" [...]
>
> No, _*-ex_ was the feminine while *-x was a collective suffix used
> to form inanimate stems.

That does not fit any facts I know. On what basis can one make such a
statement?

> > the -i- part is still left unaccounted-for.
>
> It's a collective. You're just too stubborn to listen and absorb this
> because it would mean that you're wrong and that the feminine is
> not reconstructable for the oldest layer of IE as you fantasize.

Maybe I am, I do think I should be. For the non-reconstructability of the
feminine ought to be based on something better and more solid than the
desire to prove me wrong. As I inspect the arguments it is now found to be
based on only that. This only has the modality of "could be", in my view
forcedly so, and it is apparently chosen on this list just to avoid siding
with me. Where is the "must be" argument that settles it? How did you make
up your mind? Is it all just based on a childish desire to distance
yourself from me? I have produced some potential "must be" arguments, but
they point the other way. Why are they all no good, while every whim of
nonsense is preferable provided only it is used against me?

>
> The nouns ending in *-oi- are probably implicated in this analysis of
> *-i- as well as the nouns ending in *-ti- such as *stex-t-i-. Again,
> as we see, a substantive *stex-t- being given a collective *-i-.
> Painfully obvious and it validates the meaning of the stem to boot.

There is hardly such a derivation. I can find a stem *steH2-t- only (in
its zero-grade alternant) in Lat. superstes, -stitis, allegedly from
*-stat-s 'standing', and a *staH2-ti-/*st&2-ti- 'station, standing place,
stand' is not "a collection of standing persons" by any obvious analysis
that will prove the -i- a collective marker. Also, Lat. -stes may itself
be a ti-stem reflecting *-st&2-ti-s with syncope as in compos from
*-potis. Quite generally, forms such as Vedic vajra-bhr.-t- 'carrying a
thunder-bolt' cannot reasonably be taken to contain a non-collective
counterpart of bhr-ti- 'act of carrying, sustenance' in the second part.
As so many times before, this is no basis for what it is being misused to
support.

> > Then what is the -i-? An "inanimate stem", we're told; but why
> > would that be added to mark the feminine?
>
> No, specifically a _collective_ which may be of either gender, not
> specifically inanimate. The point is that it collectivizes just
> like *-x does. The union of *-ix merely produces yet another
> "collectivizer" but it is specifically used for animate nouns.

What is the point in calling a morpheme that turns up as one half of a
feminine marker and never in collectives, not a feminine, but a
collective?? This cannot reflect a desire to be objective.

> >> As per above: *-ix didn't exist prior to its use as a feminine.
>
> Sorry. I worded wrong. Rather *-ix must have existed during *-ex as
> another collective animate.

"During" meaning "simultaneously with"? That does not answer the question
why one was completely ousted from collectives, being found only in
feminines, while the other one is found in both collectives and feminines.
How can that sensibly be taken to reflect forms that were specifically not
feminine, but collective?

> This is the summary of what I'm thinking
> in chronological order as of this moment:

> ITyr  => - *-hWa is used as a collectivizing suffix. It becomes
>             the Tyrrhenian inanimate plural *-qo (Etruscan -cHva
>             as in /un-cHva/ and /pulum-cHva/). Yes, Tyrrhenian
>             preserved animate-inanimate gender and might survive
>             right down to Etruscan.
> MIE   => - *-hWa is still used as a collectivizing suffix
> eLIE  => - Syncope affects *-hWa (in line with animate nominative
>             *-sa), producing *-x.

I do not accept the relevance of Etruscan to "explain" Indo-European. So I
start here:

>           - Thematicization: *-x => animate *-&x

The segment "thematic vowel + /H2/" is not animate, it's thematic, even
thematic collective.

>           - To avoid gender homophony in athematic stems,
>             the animate collectives use *-i-x (a synthesis of
>             two collectivizing endings).

There seems to be no evidence for -i- as a collective marker.

> mLIE  => - Schwa Diffusion: *-&x remains because *x is voiceless.

In my language: The thematic vowel is /e/ in *-e-H2. This is correct.

> IE    => - gender system remains one of animate vs inanimate,
>             with *-ex/*-ix used for animate collectives and
>             *-x for inanimate collectives.

Inadequate. The distribution is, thematic (collective and feminine)
*-e-H2, athematic collective *-H2, athematic feminine *-iH2.

> after => - feminine derived from animate collectives and therefore
>             most stems in *-ex/*-ix are now specifically feminine.

Same error repeated.

>
> > Frightfully complicated, really. For *-o:n and *-onts are two
> > stages in the development of the same form. And both stages
> > spawned feminines, which turn up as *-n-iH and *-nt-iH respectively.
> > Thus if feminines were created before *-o:n was replaced by *-ont-s,
> > any language showing *-ont-s must also have shared in the feminines
> > of the *-o:n type. Surely it is now simpler to accept the feminine.
>
> No. These feminines can develop semi-independently regardless.
>
> I'm afraid in your obsessively linear way of thinking, it would
> certainly appear frightfully complicated, yes. However, I see
> your flaw. You haven't learned the principle of wave-particle
> duality in linguistics and only think in terms of language
> branching.

Well, that stops it all. If the quest for strict laws is itself
disqualified, as it is by this "duality", there is no basis for any
serious discussion. Then anything goes, just shout duality anytime you
are wrong.

> Wave-particle duality is the understanding that particles can equally
> be thought of as waves coming from a source.

And so on.

>
> > No. Comparative evidence is not just zero. None of the
> > i-adjectives has any apparent justification from within
> > Anatolian.
>
> Why not? Anatolian derives from IE. IE had adjectives in *-i-,
> so why can't Anatolian, even if a later development of *meg-i-
> is somehow un-IndoEuropean?

The form *meg^-iH2 is an inherited lexicalized feminine form of a specific
adjective. If that turns up in an individual branch of Indo-European that
branch has inherited it and has consequently known the feminine gender,

>
> >> >> 7. Anatolian's Motions-i, used often in nonneuter strong cases,
> >> >>    must show *-ix, for what else could it be
> [...]
> > No, that's a total word salad. Caland is about a stem-replacing
> > -i- which is neutral in this respect.
>
> Your "feminine" *-ix is my "animate" *-ix. So why do you need the
> feminine here anyways? I don't get it. An animate *-ix would do
> just fine. If a thematic variant with *-i- doesn't work, so be it.
> But there's no reason why *-ix can't be used as a general _animate_
> at the IndoAnatolian stage.

This throws out the baby with the bathwater. If *-iH2 was just a stem
forming adjectives and is retained as such in an occasional Anatolian form
or two, how could it come to be restricted to feminine in non-Anatolian?
And if the collective function, allegedely of both of its parts, was lost
sight of in both branches, why was it then also a collective marker, viz.
*-H2, that was used to mark the feminine of thematic stems? Now the core
that really made sense in the connection between the collective and the
feminine has been thrown away. That cannot be taken seriously.

> > Nor does quom have the same stem as tum. Or quam as tam. So this
> > would be a third pair of this kind.
>
> That's all nice, but you're relying on ambiguities. Does Hittite /-n/
> represent *-m or *-n? What does the second /a/ in /mahhan/ actually
> represent as a vowel?

Look who's talking! Who just identified acc. *-om with the gen.pl. *-o::m?
Who made *-yo an endingless locative in conflict with its actual use with
and phonetic rules applying to its alleged material? And how unambiguous
is Hitt. -as anyway?

> For all we know /man/ and /mahhan/ could
> derive from something specifically postIE such as an accusative in
> early Anatolian *ma-m with a locative *ma-(h)an based on postposed
> *en "in" and we all know that similar postpositioning occurs with
> other locative particles *bHi and *dHi, among possible others. This
> is a competing hypothesis that you can't get rid of. Therefore
> /mahhan/ does not serve to adequately oppose what I'm saying. Find
> another example.

It must reflect something, and a reconstruction *mo-m, *ma-H2-m would be
completely in line with quom/quam and tum/tam. The masc.acc. certainly has
a companion in Skt. yam, Celtib. iom and the Old Irish nasalized relative
sentences. Also English then matches tum. However, I do not really know a
pronoun *me/o- very well (surely *mo- 'my' is off), although
Nostraticists are generally less reserved on that point.

Jens