Possessor Raising/Double Subject constructions

George Aaron Broadwell g.broadwell at ALBANY.EDU
Fri Nov 15 20:50:25 UTC 2002


Dear LFG list readers,

I have some questions about the sort of constructions that are usually
labelled possessor raising or double nominative constructions. I have in
mind cases like the following :

1.)	Choctaw

John-at im-ofi'-at  im-illi-ih.
John-nom his-dog-nom dat-die-tns

'John, his dog died (on him.)'

Both 'John' and 'dog' are explicitly marked nominative.  There is a
dative/benefactive applicative prefix on the verb.


2. ) Amharic 

Lïj-u 	   doro-w 	  tä-yzä-ll-ät.
son-his  chicken-def    pass-catch-applic-3ms

(approximately) 'His son, a chicken was caught for.'  

Amharic requires that all definite objects be marked accusative, so in
examples of this sort, both "his son" and "the chicken" appear to be in the
 nominative.

Both the nominative NPs in these sentences pass a number of subject tests
in the language (case, control of agreement, control of switch-reference
marking (in Choctaw)).  We can also show that the first nominative NP does
not form a constituent with the second NP.  The first NP
(benefactive/possessor) must precede the second NP (theme/patient).

My question:

What is the best LFG approach to sentences of this type? 

My thinking has been running along the following lines --

In normal cases, it seems pretty clear that there is a principle like
Functional Uniqueness that applies to GFs. So normally we don't expect two
OBJ functions or two PRED functions in an f-str.

Bresnan (2001:69-70) notes that functions like ADJUNCT, FOCUS, and TOPIC
may take sets as their values.  FOCUS and TOPIC belong to the set of DFs.
But SUBJECT also belongs to that set of overlay functions.  Is it coherent
to think of a set-valued SUBJECT function as well?  In the Choctaw and
Amharic cases, it seems that one of the apparent SUBJECTs bears an argument
relation to the predicate, while the other does not.


Any comments or pointers to discussion would be much appreciated.

Thanks,
Aaron
-------------------------------------------------------------------------
George Aaron Broadwell,  g.broadwell at albany.edu
Dept. of Anthropology, Program in Linguistics and Cognitive Science 
UAlbany, Albany NY 12222
(518)-442-4711   Web page: http://www.albany.edu/anthro/fac/broadwell.htm
--------------------------------------------------------------------------
"Every absurdity has a champion to defend it." -- Anonymous



More information about the LFG mailing list