John Cowan writes:
 > Tim May scripsit:
 > > While further investigating adjective-final SOV langages* I found the
 > > underlinked lecture on the semantics of core cases.  I'm finding it
 > > The argument is that all core arguments to a verb take
 > > the roles of agent, location or theme seems to fit in well with my own
 > > thoughts on Rick Morneau's treatment of the same subject.
 > There are counterexamples like "The stone broke the window", where
 > the grammatical subject is an instrument.  RM discusses this sentence
 > type briefly but dismisses it as too rare in the world's languages
 > to worry about.  Since his purposes are MT and IAL rather than
 > linguistic theory, fair enough.

I'm going to have to agree with Christophe here: in the linguistic
semantics* fo the sentence, the role of the stone is agentive, not
instrumental.  In the broader view of the situation we imagine, it's
probably an instrument (although as Christophe point out we cannot
tell for sure) but that's not particularly relevent here.  If we can
deduce that the stone is an instrument in some other sentence
describing the same situation, that doesn't necessarily mean it's one

One may object on the ground that the stone cannot act deliberately,
but I'd say either that this is a different sense of the term "agent",
or a similar sense used metaphorically.  (After all, in DeLancey's
analysis (IIUC - this specific example isn't given) "window" is the
theme and "broken" a location contained within the verb, and if we can
accept "broken" as a location I'm sure we can accept "stone" as an

* Delancey (author of the page I linked to) explains his use of the
  term "semantic" thus:

|If our purpose is to explain linguistic structure and behavior, we
|are concerned only with those cognitive categories which are
|reflected in linguistic structure and behavior--which is what I mean
|when I say semantic.  If there is no linguistic test for a category
|in any language, then it is not a linguistic category.  So, no
|"classification of relationships between entities in the world" which
|is in fact "extra-linguistic", i.e. has no linguistic reflection, has
|any place in our investigations.