Theiling Online    Sitemap    Conlang Mailing List HQ   

THEORY: Bivalent Intransitive Verbs

From:Tom Chappell <tomhchappell@...>
Date:Tuesday, August 2, 2005, 19:48
Hello, everyone.

[WHY I'M ASKING]

I am thinking of putting in my first conlang a set of verbs that require two
nominal core arguments but take no direct object. I plan to call them bivalent
intransitive verbs -- bivalent because they require two core arguments,
intransitive because they take no direct object.

The alignment of the case system will be like this;
If S stands for the Subject of Monovalent clauses,
A stands for the Agent, and P for the Patient,
of Agent/Patient (mono)Transitive (bivalent) clauses,
and D stands for the Donor, R the Recipient, and T the "Theme" (moved object)
of Ditransitive clauses,
then
I'll have a case I'll call ABSOLUTIVE for S;
a case I'll call ERGATIVE for A and D;
a case I'll call ACCUSATIVE for P and T;
and a case I'll call DATIVE for R.

Allowable argument combinations will be;
ABS;
ERG and ACC;
ERG and ACC and DAT;
and ABS and DAT.

The thing is, I am going to want some sentences with ABSOLUTIVE subject and DATIVE
required second argument. The second argument, however, will not be a direct
object.
Prototypically, these verbs, when glossed in English, will have to emply a reflexive and a
preposition. For instance, a verb meaning "to concern oneself with" would have
the concerned person as the subject in ABS and the target of his/her concern as
the second argument in DAT.
I suppose it might be alright to call that second argument and Indirect Object, I'm not sure.

[ANADEW?]

When I "Google" on "Bivalent Intransitive", I get some interesting pages.  Among them are;
http://www.surrey.ac.uk/LIS/MB/WALS/Basque.htm
which says that Basque has Bivalent Intransitive verbs, among them a verb glossed "come";
http://www.ling.lu.se/disseminations/pdf/44/Kumakhov_Vamling.pdf
which says Kabardian has Bivalent Intransitive verbs, among them a verb glossed "wait for";
(It also says
Kabardian transitive verbs are polyvalent, and may correlate with up to five (5) NPs);
http://rspas.anu.edu.au/linguistics/WP/Bowden2.html
says Taba has "bivalent semi-transitive verbs" and "non-agentive bivalent verbs";
http://www.mpi.nl/world/anrep/98/anrep-98-9.html
says Olutec has some verbs, including 'sweat' and 'be angry', which are
surface-monovalent yet behave as if they were bivalent.

Does anyone on the list know anything about natlangs with Bivalent Intransitive verbs?

[HOW I CAME UP WITH THIS IDEA]

Dowty produced a list of five Proto-Agent properties and five Proto-Patient
properties for seeing how well any given nominal argument of any given
predicate fit into the Agent or Patient macrorole (have I got the terminology
right?) relative to that predicate and its other arguments.

Following Kiparsky, I am going to assign the argument with the highest Proto-A score the
feature [+HR], and assign all other arguments the feature [-HR]. I am going to
assign the argument with the highest Proto-P score the feature [+LR], and
assign all other arguments the feature [-LR].
Obviously, then, every clause that has any arguments at all, will have one and only one
argument with the [+HR] feature, and one and only one argument with the [+LR]
feature.
If an argument is [+HR, +LR], I will assign it the ABSOLUTIVE case.
Obviously any clause with only one argument will have its only argument in this ABS case.
If an argument is {+HR, -LR], I will assign it the ERGATIVE case;
and if a argument is [-HR, +LR], I will assign it the ACCUSATIVE case.
Clearly any clause with one or more arguments must have either an ABS argument or an
ERG argument, but not both;
and any clause with one or more arguments must have either an ABS argument or an
ACC argument, but not both.
Also, if any core argument has [-HR, -LR], I will assign it the DATIVE case.
Clearly if any clause has a third argument, that third argument must be DATIVE.

But, what if a clause does have exactly two arguments, and nevertheless the same
one of those two arguments gets both the [+HR] and the [+LR] feature? Then that
[+HR, +LR] argument will be ABSOLUTIVE, and the other, [-HR,-LR] argument will
be DATIVE.
In such a case, the ABSOLUTIVE argument will be the most active of the two
arguments, and will also be the most affected (effected?). So in a sense the
predicate is a reflexive event of the ABSOLUTIVE argument acting on itself,
causing itself to undergo something. But the DATIVE argument is still required
-- it cannot be skipped or omitted. It tells us something essential about the
focus or direction, or something like that, of the event.

That's why I thought of the prototype "to concern oneself with".
You can tell someone, "Don't concern yourself with the bill."
You can't tell someone, *"Don't concern yourself with" -- that's not a complete sentence.

-----

What does anyone think?

Tom H.C. in MI

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com

Reply

Patrick Littell <puchitao@...>