[developers] portuguese hpsg

[Valia Kordoni]

if I may add, similar arguments for the treatment of so called symmetric 
predicates in English have been given in Stephen Wechsler's 1995 Stanford 
PhD thesis and subsequent book 'The Semantic Basis of Argument Structure' 
(http://cslipublications.stanford.edu/site/1881526682.html). The same I 
have observed in my 2001 Essex PhD thesis (Psych Verb Constructions in 
Modern Greek: a semantic analysis in the Hierarchical Lexicon; 
http://www.coli.uni-saarland.de/~kordoni/thesis.html) to be the case in 
Modern Greek, too, for related predicates.

Valia

On Tue, 14 Apr 2009, Ann Copestake wrote:

>
>
> crysmann at ifk.uni-bonn.de said:
>>  What I'd suggest is to use sense tagging to identify predicates that display
>> this behaviour so they can be detected downstream.
>
>> E.g. _marry_v_sym_rel or _be_v_sym_rel (as in "The winner is ...)
>
> Wholeheartedly seconded.  I actually think there's a good argument that this
> is the most principled thing to do, but I won't try and make that here ...
>
> Some/most of the symmetric predicates aren't fully symmetric, of course -
> standard example:
>
> Kim and Sandy collided.
> Kim collided with Sandy.
> The drunk collided with the lamppost.
> ? The lamppost collided with the drunk.
>
> Even in the case of highly symmetric predicates like `marry' - if it's
> possible to identify one participant as being more agentive than the other,
> that participant will tend to be subject.  So if we're talking about medieval
> societies where royal marriages were about alliances:
>
>  The king married the earl's twelve year old daughter.
>
> is more natural than the other way around.
>
> Ann
>
>
>
>
>
>
>