Re: Theta role?

Dirk:
> On Dec 14, 2004, at 7:41 PM, Tristan Mc Leay wrote:
> > On 15 Dec 2004, at 1.22 pm, Thomas R. Wier wrote:
> >
> >> And wrote:
> >>> Tom Wier:
> >>>> there's almost no agreement on how many theta
> >>>> roles exist, even among grammarians who practice within the same
> >>>> theoretical framework. IMO, three is the best bet, since to my
> >>>> knowledge there are no languages where an underived verb takes
> >>>> more than three arguments,
> >>>
> >>> I'll bet you plenty that in English at least one verb seems to
> >>> take four arguments...
> >>
> >> Perhaps I should have been clearer.  I was talking about verbs
> >> which take four *NP* arguments.  On some analyses, English verbs
> >> like 'buy' and 'sell' take four arguments, but one of those must,
> >> in every dialect I know of, be prepositional.  These constructions
> >> in Georgian and Abkhaz on the contrary involve four NP arguments.
> >> (I can't speak for Abkhaz, but in Georgian the constructions
> >> alternate with variants having a postposition for the fourth
> >> argument.)
> >
> > I think And was talking about a particular special case of anadewism,
> > eadewism (or aeadewism), '(an) English's already dunnit even worse'.
>
> No. And was playing with the verb 'bet', which according to some
> analyses takes four arguments.
Just so. It's not stricly a counterexample to Tom's revised claim,
because the fourth argument must be clausal rather than nominal.
But restricting the claim to exclude clauses seems rather arbitrary
to me, certainly if the claim is supposed to have bearing on the
number of theta roles a theory needs, since I can think of no
conceivable basis for supposing that clausal arguments don't bear
theta roles.

That said, the clausal argument alternates with "on + NP" ("I'll
bet you five quid on the outcome"), so one might argue that my
sentence is syntactically "I'll bet you plenty ON that in English
at least one verb seems to take four arguments", with a
(contentious but not ad hoc) rule deleting (the phonology of)
"on" when it has a clausal complement. I'd be fairly willing to
buy this analysis, and hence retract it as a counterexample to
Tom's generalization, with the added proviso that the systemic
relationship between prepositional complements and theta roles
is pretty mysterious.

--And.