stanford nlp - nlp - How do I identify if a part of sentence is an answer to Who, What, How or Why? -

i'm trying assign every word in sentence role: actor, predicate, who, what, where, when, how or why.

for example, "the robot brought me orange juice" should tagged "the/actor robot/actor brought/predicate me/who an/what orange/what juice/what ."

i'm able detect actor, predicate, , when using semantic role labelling. i'm having problem who, what, how , why.

is there tool in nlp identifies if phrase in sentence answer w-question? how tell if noun person or thing?

i know there name entity recognition, identifies proper nouns. tells obama person, not teacher person.

i think looker part-of-speech tagger. have added standford-nlp-tag give standford-pos-tagger link:

http://nlp.stanford.edu/software/tagger.shtml

for more details how tagging done take @ documentation. on basis on tags can answer questions who/whom/what etc.. in case: x brought y z -> article "an" determines x active/nominative, y passive/dative , z passive/accusative, while x brought y z example change things (y acc., z location)

by way: "thing" can identified comparison semantic network wordnet if model not trained context.