NLP; recognizing syntactical parts

I followed with interest the thread "Just starting out.. newbie help :)" 
and it made me believe you won't be extremely annoyed by another newbie 
asking questions :)

My problem:  I need to index texts against a list of relevant words. 
Doing this by hand for a number of small sample made me think that 
ranking the relevance of a text by the frequency of words that are 
subjects or verbs gives somewhat better results than simply counting the 
frequency of all words parts that match the list.

My question: is there any module that would help with recognizing 
syntactical parts in a text written in English?

Searching with Google got me a lot of results that were course outlines 
or assignments. On CPAN  I found only the Lingua::* modules, but if 
there is one that fits what I need, then probably I failed to see it 
among the over 600 hits ...

thank you,

Emil Perhinschi