Re: ergative? I don't know...

On Sun, 25 Oct 1998, Tim Smith wrote:

> David Durand wrote:
> >>> iii. active system
> >>> Agent       (A)
> >>> Patient     (P)
> >>> Subject     (S)
> >>
> >>Do these constitute three separate cases?  I don't know if T. fits this=
=2E
> >
> >These would be 3 separate cases. I was trying to list cases on the left,
> >and roles on the right, with case names chosen as traditional within tha=
t
> >kind of system. I guessed on the names of the cases for active systems
> >'cause I couldn't remember
>
[Tim wrote]:
=20
> With respect, this is not my understanding of what is generally meant by
> "active languages".  According to Barry J. Blake (_Case_) and R.L. Trask =
(_A
> Dictionary of Grammatical Terms in Linguistics_), active systems, like
> nominative-accusative and ergative-absolutive systems, have only _two_ ca=
ses
> (or rather, two types of marking, whether it's case marking on the nouns,
> agreement marking on the verbs, or whatever) for the three argument types=
 S,
> A and P.  All three types mark A differently from P.  Where they differ i=
s
> in how they mark S:
> - In nominative-accusative systems, S is marked the same as A.
> - In ergative-absolutive systems, S is marked the same as P.
> - In active systems, some S arguments are marked like A and others like P=
=2E
>=20
> Active languages differ in what criteria they use to determine whether th=
e S
> of a given verb takes A or P marking.  In some active, e.g. Guaran=ED, it=
's
> basically dynamic verbs vs. stative verbs; in others, e.g. Lakhota, it's
> basically volitional vs. non-volitional. =20
=09But then what is S?  Does the S category still depend on
=09transitivity as it does in ergative systems?  So that
=09S can function as an A in active languages if it's volitional,
=09but also only if it's intransitive?  And likewise,  it can
=09function as a P in active languages if it's non-volitional,
=09but then also only if it's intransitive?

=09I think I have finally identified my confusion about
=09active languages and the use of S.

=09I thought I had understood this, but I guess I'm still=20
=09operating under a misconception about S.  The term "S"
=09HAS TO MEAN INTRANSITIVE SUBJECT, doesn't it?  Now I
=09finally understand where I have gone wrong in labeling
=09my terms for T.  Under the system that I now have, it
=09is incorrect to call the non-volitional subject an
=09"S," because this only really properly refers to=20
=09intransitive subjects.  I then lead everyone astray when
=09I apply the term to non-volitional transitives, as in
=09=09Galleyli li zef oua--"the man (S) hears music."

=09What *I* mean by S is "participant," but a participant
=09is not necessarily an intransitive subject in Teonaht.

=09It looks as though the active languages that you describe
=09above, Tim, derive from or partake of ergative structures that
=09they have modified to include a volitional/non-volitional
=09or dynamic/stative component ON TOP of a distinction made
=09between a transitive and intransitive subject.

=09This explains why David Bell and others who have tried
=09to make me understand active and failed always appealed
=09to examples in ergative languages.

=09The question then remains:  if that is what S means, then
=09how is *S* functioning in nom/accusative languages?  It
=09would seem that S is non-existent in these, if NO distinction
=09is made between transitive and intransitive subjects.

=09WHAT IS "S"???  That's my big question.  It sounds to me
=09like it's an elusive thing... just a concept, a wildcard
=09that is either handed over to the nominative or to the
=09absolutive.  Does it have any real status outside of its
=09recognition as a "place holder" by linguists?

=09I need some new terminology.  Agent, Participant, and Patient
=09in my books will have to go by new abbreviations:  Ag, Pt, and
=09Ob perhaps.  Is there a system I can borrow that does this
=09or do I have to make it up?

Tim again:
>  Furthermore, all active languages
> seem to have some arbitrary exceptions to whatever criteria they use.  To
> further complicate matters, a few active languages have what's called "fl=
uid
> intransitivity", in which some intransitive verbs can take _either_ A or =
P
> marking for their S argument, depending on the degree of volitionality in
> the specific situation.
>=20
> The kind of system you're talking about, with three distinct types of
> marking for the three argument types, is generally referred to (by Blake,
> Comrie, and others) as a "tripartite" system.  Such systems are fairly ra=
re,
> and are apparently always limited to some subset of NPs in a given langua=
ge;
> there don't seem to be any natlangs that have such a system for all class=
es
> of NPs.  (Although one of my conlangs, Meitzanathein, does have an
> across-the-board tripartite system.)
=09=09Teonaht is definitely tripartite in its use of
=09=09articles to define these argument types:

=09=09=09Le zef, li zef, and il zef
=09=09=09The man, (agent) the man (participant)
=09=09=09and the man (object or patient)

=09=09But it is nom/acc. in its use of PRONOUNS--and I
=09=09want to keep it that way.  Nothing does more violence
=09=09to a beloved and long established conlang than to
=09=09dicker with its pronouns.  I have, however, introduced
=09=09a variation on the emphasized pronouns, as some of you
=09=09may have noticed when I set up the incomplete pronoun
=09=09chapter on my web:  Yry firrimby  "I grateful" is
=09=09now the participant marking.  But the variant "yryi"
=09=09indicates the agentive.  Yryi kerreca, "I disapproving."
=09=09You can be grateful because of circumstances outside
=09=09your agency... but you actively disapprove of something.=09=09
=20
> However, I agree with what I gather is your main point: that the system
> Sally has devised for Teonaht doesn't exist in any known natlang (which i=
s
> why it's hard to figure out what to call it, though to me "split nominati=
ve"
> sounds reasonable), but that it nevertheless doesn't sound particularly
> "unnatural". =20
Firrimby!  Firrimby!  Firrimby!  <G>

In fact, it sounds natural enough to make me wonder if there
> might have been such languages but there just don't happen to be any
> currently extant.  (I've often wondered which of the apparent gaps in
> language typology are "real" -- that is, rooted in actual language
> universals -- and which are accidental, caused only by the vagaries of wh=
at
> languages happen to have gone extinct at particular times.)
And to what these languages have been exposed to, and to the degree that
they have been caught and captured by grammarians.  To address this
question, Tim.... Teonaht is an entirely "artificial" language.  But many
natural languages bend to artificial embellishments, some of them reaching
down as far as basic grammar.  We do know that English underwent a great
deal of facelifting in the eighteenth century, so that some basic rules
about double negatives and many other aspects of "correct grammar" were
agreed upon and doctored.  Look at the many gyrations that the structure
of the ordinary sentence underwent in the history of written Welsh.  In
the thirteenth century, the "abnormal sentence," never spoken by the
populace, was the norm in writing.  It was a fad and it disappeared.  And
I have heard that modern Turkish was drastically doctored up by its
grammarians.  This seems to me to be typical of languages and literacy.=20
One thing I know for sure about Teonaht is that it is literate... and
bends compliantly to literate embellishment (except in the case of its
pronouns). In fact it is hardly spoken at all. ;-)  Were it a natural
language, I could easily see it adopting a volitional distinction in its
nominative articles, and even in its verbs in a written tradition--if
there were some language it admired and wanted to copy.  English did this
with its imposition of periodic structure on its sentences (copying the
Latin).  Written traditions DO affect spoken traditions, as we can see by
the pronounciation of "fault," originally spelled faute, the "l" added
later on as an "etymological respelling" (they thought it was related to
L.  fallita), pronounced /fOt/ and rhymed with "ought," and now pronounced
with the "l."  Same story with "calm."  I know people who pronounce it as
it was wont: /cam/, and others, who've capitulated to the respelling and
pronounce it as I do: /caUlm/. Couldn't other languages redo their basic
grammar along the same lines?  No one knows exactly why the London dialect
threw out perfectly good _hem_ and _hir_ and adopted the Northern
variations: _them_ and _their_.

Sally
++++++++++++++++++++++++++++++++++++++++++++++++++
Sally Caves
scaves@frontiernet.net
http://www.frontiernet.net/~scaves/teonaht.html

Rin euab ouarjo vopy vytssema tohda uo zef:
ar al aippara brottwav; ad kemban aril yllefo
brotwav fenom; vybbrysan brotwav an; he ad=20
edirmerem brotwav kronom.

"A cat and a man are not all that different.
Both are on my bed; both lay their head on their
arm; both have mustaches; both purr when they
sleep."
++++++++++++++++++++++++++++++++++++++++++++++++++