Re: THEORY: "Quirky" Case -- "Quirky" Subjects and "Quirky" Objects

Thank you very much, Markus.
I have just read a couple of articles by Halldor Sigurdsson of the
University of Lund and by Paul Kiparsky of Stanford University.
http://mercury.hau.ac.kr/kggc/Monthly/ReadingMaterial/Case.pdf
http://kybele.psych.cornell.edu/LU/Kiparsky-Cornell.pdf

I think that the reason my questions sounded confusing to some of the
responders on this thread was that I did not know any non-confusing
terminology.
Sigurdsson introduces terminology something like what you do;
but, although he (I am assuming gender because of the "-sson")
discusses the phenomenon I think you are calling "syntactical case",
he doesn't give it a name.
He distinguishes abstract or deep cases between "inherent" -- those
which are primarily semantic -- and "structural" -- those which are
primarily concerned with Argument Licensing (i.e. e.g. which nominal
is the "subject", which the "direct object", and which the "indirect
object".)
Since the core-arguments are pretty much all thematically-
unrestricted ("[-r]" in Lexical Mapping Theory terms), nearly
any "inherent case" could belong to the Subject or, if there is one,
the Object (or either Object if there are two).
Sigurdsson proposes to get around this problem by not
calling "Structural Cases" "cases", but instead calling
them "Argument licenses" or something like that.

In situations where a core-argument nominal has an
important "inherent" case, languages take one of five strategies
Sigurdsson mentions to assign "morphological case" (the same language
may take more than one strategy depending on the situation);
a. "Caseless" languages just skip the whole thing.
b. The "structural" abstract case can be expressed by its
corresponding morphological case, and the "inherent" abstract case
can be ignored.
c. The "inherent" abstract case can be expressed by its corresponding
morphological case, and the "structural" abstract case can be ignored
(one kind of quirky -- if this happens to the subject, it leaves the
nominative case available for assignment to the object, so we would
have a quirky object as well)
e. The "inherent" abstract case and the "structural" abstract case
can both be encoded into their respective morphological cases, and
the nominal can have both case-endings stacked

d. The "inherent" abstract case can be encoded into its morphological
case; the nominal's morphology reflects only its inherent case, not
its structural case; but it still carries its structural
case "invisibly", because it affects the syntax of the rest of the
clause (this is the other kind of quirky; and it is also one kind of
what you call "syntactic case", if I understand you correctly.)

Sigurdsson gives them the order b.c.d.e. because he hypothesizes that
any language which uses strategy c will sometimes also use b; any
language which uses d will sometimes use c or b; and any language
which uses e will sometimes use d or c or b.

I mentioned "d." last because it seems to be an example of what you
call "syntactic case" -- a kind of surface case which is marked not
by the morphology of the nominal itself, but by the syntax of the
clause or by the morphology of other words which must agree with its
("morphologically invisible") case.

Thanks for writing.

Tom H.C. in MI

--- In conlang@yahoogroups.com, Markus Miekk-oja <m13kk0@H...> wrote:
> >Very interesting.
> >What exactly is "morphological case",
> >what exactly is "syntactical case",
>
> I coined (afaict) the term syntactical case there in order to
enable the
> discussion of a certain phenomenon.
> Namely, the fact that non-nominative subjects in some languages
> syntactically behave like subjects, while in others they don't.
Therefore, I
> find it reasonable to say that non-nominatives that behave like
subjects (
> as far as pronominal binding, verb inflection, ellipsis in
coordinate
> sentences, etc...)
>
> For instance:
> Hann segist vera duglegur, en honum finnst verknefnið of þungt.
> he(NOM) claims to be diligent, but he(DAT) finds the homework too
hard
>
> Hann segist vera duglegur, en finnst verknefnið of þungt.
> He(nom) claims to be diligent but Ø(dat) finds the homework too
hard.
>
> Similarly, reflexive binding:
> Sigga<i> bardi mig með dukkunni sinni<i>/*hennar<i>
> Sigga(nom) hit me with doll(dat) self's<i>/*hers<i>.
>
> Eg bardi Siggu<i> með dukkunni *sinni<i>/hennar<i>
> I hit Sigga with dolls *selfs/hers
>
> As we see, the reflexive only refers back to the subject in
Icelandic.
> However, if the subject is of another case, it still refers back to
it:
> Henni<i> þykir broðir sinni<i>/*hennar<i> leðinlegur
> She(DAT) considers brother(nom) self's<i>/*hers<i> boring(nom).
>
> The situation can be reversed - German, for instance, seems not to
allow
> what Icelandic does, and I think one reasonable interpretation is
that the
> underlying syntax is different - German doesn't allow non-
nominatives in the
> syntactical position of subjects - (this can only be shown using
syntactic
> trees, I'm well aware that they can go first in a sentence in
German and
> thus look like they were in subject position).
>
> >and what languages don't have a one-to-one correspondence between
> >morphological cases and syntactical cases,
> >and wherein does the correspondence fail to be one-to-one?
>
> Quirky case subjects is exactly such a failure of correspondence,
esp. the
> Icelandic system is such. The German system differs in ways that
imho make
> the quirky subjects of German seem less subject-like than in
Icelandic.
> [sniipped stuff about reflexives]