more on aspect and objects

Thu May 2 14:24:00 UTC 1996

Hi, 

just a quick response to Dick Hudson's and Chris Pinon's messages on
the OBJtheta topic. 

Dick Hudson is, of course, entirely correct in pointing out that I
oversimplified the presentation of the link between syntax and aspect.
Given that it would haave been impossible to do justice to the entire
discussion that has been going on in the literature with regard to the
interaction between aspect and realization of NP arguments, I tried to
keep things as simple as possible.  However, I believe that there has
been quite a lot of ground gained since Dowty's seminal work, 
particularly with regard to "push" type verbs (push a cart, vs. push a
cart up the hill).   And as far as I can tell from the literature, the
general idea of what I am currently after can hold up.  I refer
interested parties to the following literature.

Carol Tenny.  1987. Grammaticalizing Aspect and Affectedness.  Phd Thesis.
              MIT.

Carol Tenny. 1994. Aspectual roles and the syntax-semantics interface.
             Dordrecht: Kluwer.

Manfred Krifka. 1992. Thematic Relations as links between nominal
reference and temporal constitution.  In I. Sag and A. Szabolcsi
(Eds.), Lexical Matters, 29-53.  Stanford, CA: CSLI.

(Carol Tenny also has an article in this volume, but I forget the
title). 

Verkuyl, H. 1972. On the Compositional Nature of the Aspects.
Dordrecht: Reidel. 

Verkuyl, H. 1993. A Theory of Aspectuality. Cambridge, Cambridge Univ.
Press. 

Gillian Ramchand. To appear.  Aspect and Predication: The Semantics of
Argument Structure. Oxford: Oxford University Press. 
(also available in an older version at a Stanford University
Dissertation)

Of these, I found Manfred Krifka's and Gillian Ramchand's work most
illuminating. 

More recently, Chris Pinon's dissertation is also out and relates to
some of the issues being discussed.  

Chris Pinon.  1995. An Ontology for Event Semantics. Phd Thesis,
Stanford University.  

However, as I have not as yet read it (I've owned it for a week now),
I can't extend the favor of being "plagued by doubts" with regard to
his world view, but can only respond to his message briefly.

The main points I should clarify with regard to his comments are.

1) I take Linking and Semantic Interpretation to be two very different
things.  In much of the literature, the two appear to be conflated,
with semanticists trying to do semantic interpretation off of thematic
roles (which doesn't really work), or syntacticians who are working on
linking trying to convince themselves they are actually making reference
to truly semantic notions when they bandy about thematic roles, and
worry about what being a patient vs. a theme really means.  

So, I don't believe in this statement:

> (A) The arguments of a verbal predicate can be characterized by certain
> semantic properties (what many call "thematic roles").

I believe that the arguments of a verbal predicate have semantic as
well as syntactic properties, and that a different set of properties
will be relevant to linking than for semantic interpretation.  Of
course, the difficult thing is that we really are at an interface
level here, and argument structure, i.e. the notion of thematic roles,
has been explicated to try to deal with this interface.  So, it is not
always clear what interacts with what, and to what degree the relevant
set of properties will overlap, but I think we should realize that the
set of properties relevant for linking and for semantic interpretation
are not identical -- one is looking more at the syntactic properties
of argument realization, the other more at the semantic side.  

2) Although I mentioned de Hoop's work on Strong vs. Weak Objects
admiringly, I should make clear that I don't, in my own work, follow
the details of her proposal.  In fact, in more recent literature, her
proposals have been criticized for being inadequate semantically.  
In particular, I don't follow her type-shifting approach.  
For some intelligent discussion of the interpretation of NPs from a
more creative perspective than the strict type-shifting Chris Pinon
outlines see Ramchand (as above), and van Geenhoven.

Veerle van Geenhoven. 1996. Semantic Incorporation and Indefinite
Descriptions. PhD thesis, Universitaet Tuebingen. 

> Whether or not we distinguish OBJ from OBJtheta as GFs, that has
> essentially nothing to do with the Strong/Weak distinction, as
> understood above.    
   (above = as in Helen de Hoop's original proposal) 

So, yes, as related to Helen de Hoop's original proposal, it is indeed
hard to draw the correspondences.  But not once one takes the view of
Strong and Weak that van Geenhoven or Ramchand articulate.  In
particular, for van Geenhoven, the syntactic distinction does not
correspond one-to-one to the semantic interpretational possibilities.
These tend to be more complex.  Which is exactly what I was after:  a
two-way distinction between OBJ and OBJtheta at a syntactic level,
that is then fed into a semantic interpretation component, where the
interpretive possibilities may be more complex (and also depend on
what kind of aspectual argument the OBJ was, i.e., what kind of
annotation it bore -- hence the idea of "piping" the information
through). 

But, I won't even try to do justice to the complexities involved in this
topic, but want to note that I take a slightly different view of this
statement. 

> In sum: a verbal predicate fully determines the thematic properties of its
> arguments, does not determine the semantic types of the NPs that realize
> those arguments in the syntax, and only partially determines the GFs of the
> NPs that realize those arguments in the syntax (the latter by LMT, if you
> wish).

1) It depends on what you define the thematic properties to be, to
decide whether you see a verb as fully determining them.  If you
decide to make a primitive, thematic, distinction between patients and
themes, for example, then yes, you do.  But, if you follow the
tendency evident in much of the literature on argument structure, of
separating out a "lexical semantic" and an "aspectual" contribution
(e.g., Grimshaw 1990, Ritter and Rosen 1993, Li 1995), then this is
not the case.  In this case, you can take something like Theme to be
underspecified, and then let the aspectual contribution decide whether
it is affected and gets interpreted more like a patient, or not (for a
very clear statement of this proposal, see Ramchand).  

2) Semantic Types of NPs are, of course, not determined by linking. 
That was part of the point of my posting. The two should be kept
separate.

3) The verb's lexical semantics only partially determine the GFs of
the NPs, yes.  They provide some of the information needed, but not
all of it.  Which is why I proposed to expand linking theory to
include the possibility of factoring in aspectual information.

>  (3) Nadyaa-ne ghore-ko bheje
>      Nadya-Erg horses-Acc sold
>      `Nadya sold the horses'   (telic)
>
>  (4) Nadyaa-ne ghore      bheje
>      Nadya-Erg horses-Nom sold
>      `Nadya sold (the) horses' (telic)
>      `Nadya did horse-sellinug' (atelic)

> What I don't understand is why the Strong/Weak distinction helps us to
> explain the contrast in (3)-(4).  In fact, it simply doesn't.  In fact, it
> even obscures the problem.  In Krifka's theory, for example, one could
> account for telicity of (3) in terms of the fact that the *nominal
> predicate* representing 'the horses (ghore-ko)' has quantized reference,
> and for the atelicity (on the relevant reading) of (4) in terms of the fact
> that the nominal predicate representing 'horses (ghore)' has cumulative
> reference.  (Of course, there would be much more to the story, but the
> Strong/Weak distinction would play no role in it.)

Well, what I haven't told you (on this list), is that the "-ko" marked
objects and the unmarked objects in Urdu (and Turkish) in fact exhibit
different distributional patterns.  The Strong/Weak distinction, which
is a syntactic distinction, can help express this.  The semantics then
must build on this.   So, conversely, what is missing from Krifka's
theory, while it delivers the right semantics, is what the syntax of
the clause looks like (he doesn't address syntactic issues in his
paper, which is a bit frustrating for a syntactician).  

Finally, just to put in a bit of a plug for a project I'm currently
involved in: my co-editor Willi Geuder (Tuebingen) and I are hoping to
bring out a book with CSLI by this fall that will address many of the
issues brought up here.  It's called "Argument Projection: Lexical and
Syntactic Constraints", and will hopefully address exactly some of the
interface problems that keep coming up.

Miriam