THEORY: Alignment of ditransitive with monotransitive case roles.
|From:||Tom Chappell <tomhchappell@...>|
|Date:||Tuesday, May 17, 2005, 23:36|
Seven questions are at the end; skip to them if you don't want to read the re-cap of theory first.
I recently read about a typology of aligning the case roles of ditransitive core
arguments with those of monotransitive core arguments, that parallels the
ACCUSATIVE/ERGATIVE/ACTIVE dimension (which is about aligning the case roles of
the monotransitive core arguments with that of the intransitive core argument).
Most people on this list have heard that the two core arguments of a monotransitive
clause can line up with the sole core argument of an intransitive clause in,
basically, two different ways.
Namely, if the Agent of a (simple active declarative) monotransitive clause shares its
case with that of the Subject of an intransitive clause, but the Patient is in
a different case, the case of the Patient is called "ACCUSATIVE";
but if the Patient of the monotransitive clause is in the same case as the
intransitive core argument, while the Agent is in a different case, the case of
the Agent is called "ERGATIVE".
(This leaves aside three main kinds of complications: first, since some languages
both mark the case on the NP and have agreeement or cross-reference marked on
the Verb, it is possible to have an ACCUSATIVE agreement system with an
ERGATIVE case system; second, some "split-ergative" languages are ACCUSATIVE in
some tenses, aspects, moods, or illocutionary forces, but ERGATIVE in others;
third, some languages have a "split-intransitive" system, in which some
intransitive subjects are treated like monotransitive Agents, but others are
treated like monotransitive Patients.)
Independently of its position on the ACC/ERG/Active position, a language can re-use the
cases of the monotransitive clause to fill two of the three core roles of a
ditransitive clause, in more than one way.
For the sake of reducing wordiness, let me introduce some abbreviations.
The sole core argument of an intransitive clause will be called S
The Agent core argument of a monotransitive clause will be called A
The Patient core argument of a monotransitive clause will be called P
The Recipient/Beneficiary/Maleficiary core argument of a ditransitive clause will be called R
The Theme (object located or moved) of the ditransitive clause will be called T
The Donor participant/actant of a ditransitive clause I will call D.
Except for the D for Donor I got these abbreviations from other people; I don't know
whether or not any of them are standard.
In the discussions I have read, everybody assumes all languages make the equation D=A.
Languages with D=A, T=P, and a third case for R, are called Indirect languages.
Languages with D=A, R=P, and a third case for T, are called Secondary languages.
Languages with D=A in which some monotransitive clauses treat P as T, and some
monotransitive clauses treat P as R, are called Split-P languages.
In an Accusative Indirect language, the D=A=S case is NOMINATIVE, the T=P case is
ACCUSATIVE, and the R case is DATIVE. The T is called the Direct Object and the
R is called the Indirect Object. There is a sort of "applicative voice" called
"Dative Movement", in which a ditransitive clause is transformed into a
monotransitive clause with the Indirect Object promoted to ACCUSATIVE, and the
erstwhile Direct Object demoted to an oblique adjunct requiring an adposition.
In an Accusative Secondary language such as Modern English, the D=A=S case is
NOMINATIVE, and I don't know what to call the R=P case or the T case. The R of
a ditransitive clause is called the Primary Object, and the T of a ditransitive
clause is called the Secondary Object. The transformation called "Anti-Dative
Movement" changes a ditransitive to a monotransitive, promoting the T to P
while demoting the R to an oblique requiring an adposition. Example:
Ditransitive "I gave John an apple" is transformed by anti-dative movement to
monotransitive "I gave an apple [to John]". "I" is the Donor which becomes the
Agent; "John" is the Primary Object (R) which gets demoted to the oblique "to
John" requiring the adposition "to"; "an apple" is the Secondary Object (P)
which gets promoted to P.
There is a parallel between, on the one hand, the Accusative/Ergative distinction
with its Passivization for Accusative languages and Anti-Passivization for
Ergative languages, and, on the other hand, the Indirect/Secondary distinction
with its Dative Movement for Indirect languages and Anti-Dative Movement for
In Ergative languages, the S=P case is called Absolutive (because it
prototypically is marked by a zero case ending). The Anti-Passive voice
transforms a monotransitive clause with A in Ergative and P in Absolutive, into
an intransitive clause with the erstwhile A promoted to Absolutive, and the
erstwhile P demoted to an oblique, which may be implicit, or if explicit must
be accompanied by an adposition.
In Accusative languages, the S=A case is called Nominative. The Passive voice
transforms a monotransitive clause with A in Nominative and P in Accusative,
into an intransitive clause with the erstwhile P promoted to Nominative, and
the erstwhile A demoted to an obliqu, which may be implicit, or if explicit
must be accompanied by an adposition. In English, "John thanked Mary" gets
transformed to "Mary was thanked [by John]": "John" is the A, "Mary" is the P,
"by" is the adposition.
I hope you can see the parallel.
Now, for my speculative, possibly fictional, language types.
Why should D=A be an absolute universal? It seems to me D=A and R=A should be
equally likely. If that is so, there are really several other alignment types:
(1) D=A=S, R=P, T (Accusative Secondary)
(2) D=A, R=P=S, T (Ergative Secondary)
(3) D=A=S, R, T=P (Accusative Indirect)
(4) D=A, R, T=P=S (Ergative Indirect)
(5) D, R=A=S, T=P (Accusative -- unnamed new subtype)
(6) D, R=A, T=P=S (Ergative -- unnamed new subtype)
(7) and (8): Accusative Split-A and Ergative Split-A types of languages, in which
all monotransitive clauses treat P as T, but some treat A as D and others treat
A as R.
(1) Aren't there some "Standard Average European" or otherwise familiar languages
with split-P or Secondary type?
(2) If not, don't some of them have fairly obvious relics of Split-P type hanging
around? (examples would be Accusative/Indirect languages having certain
monotransitive clauses whose only object must sometimes be in the DATIVE case,
without any ACCUSATIVE argument).
(3) Don't some S.A.E. or familiar languages seem to have vestiges of a Split-A
type as well? These might be, say Accusative Indirect languages having certain
monotransitive clauses whose subject must sometimes be in the DATIVE case,
while its object is still in the ACCUSATIVE, and there is no NOMINATIVE
(4) Are their attested natural languages that are in each of the following pairs of types:
(a) Accusative and Secondary
(b) Accusative and Indirect
(c) Accusative and Split-P
(d) Ergative and Secondary
(e) Ergative and Indirect
(f) Ergative and Split-P
(g) Split-S and Secondary
(h) Split-S and Indirect
(i) Split-S and Split-P
(5) How possible would a language that was both Split-A and Split-P at the same
time be? Such a language would have three kinds of monotransitive clauses:
those in which A was treated like D and P was treated like R; those in which A
was treated like D and P was treated like T; and those in which A was treated
like R and P was treated like T.
(6) Are there in fact any attested NatLangs with R=A or Split-A type?
(7) Does anyone besides me find these ideas interesting?
Thanks to anyone who answers.
I sort of expect some of the answers to be "Definitively, no one knows yet."
Tom H.C. in MI
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around