{{CSCI 587 Test 2} \hfill {April 15, 1997}}

NLP in Prolog: Given the grammar below:

   S    -> NP VP
   NP   -> pronoun | art noun | NP PP
   VP   -> verb NP | VP PP
   PP   -> prep NP

Give two parse trees for ``I saw the man with a telescope.''

Construct an RTN that recgonizes the sentences generated by this grammar.
Write prolog code for a parser based on this grammar that will generate a parse tree as it parses. Assume the lexical productions, i.e., ``noun $\rightarrow$ man'' etc. are already handled.
```
   S    -> NP VP
   NP   -> pronoun | art noun | NP PP
   VP   -> verb NP | VP PP
   PP   -> prep NP
   
```

Assuming the grammar below:

   S    -> NP VP | NP aux VP
   NP   -> art noun | art noun noun
   VP   -> verb NP 
   art  -> the
   noun -> water | can
   verb -> water | hold
   aux  -> can

Give the chart for the parse of the sentence ``The water can can hold water.'' after the second ``can'' has been processed.

Features:
1. What does the feature SUBCAT \_np\_np mean for a verb?
2. Give a SUBCAT feature for the verb receive.
3. What does it mean for VAR to be a ``head'' feature for S?
Semantics:
1. Simplify:((($\lambda$ x (x*y +5*x + 3)2)
2. Simplify:(($\lambda$ x (x A))($\lambda$ y (Q y)))
3. Represent: ``likes Mary'' as in ``John likes Mary'' using a lambda expression.
Thematic Roles: What are the thematic roles in the sentence ``John gave a valentine to Mary at the party'' ?

Using the rule

        (VP   VAR ?v   SEM ($\lambda$  a3 (?semv ?v a3)))
        $\rightarrow$ \hfill 

         
        (V    SUBCAT \_np   VAR ?v   SEM ?semv)
        (NP   SEM ?semnp)

Show the SEM feature generated for VP assuming:

        (V    SUBCAT \_np   VAR e3   SEM KISS1   VFORM PAST)

        (NP   SEM (NAME m1 ``Mary''))