Parsing I

• Finding linguistic structure
  • Morphology: structure within words (we will not cover this, sadly)
    atanifurahisha -> a-ta-ni-furah-ish-a `she will make me happy' (Swahili)
  • Syntax: structure within sentences
    • Grouping words into phrases
      (The pile (of leaves) (that Frog (had raked (for Toad)))) (blew everywhere).
    • Labeling lexical and phrasal categories
      leaves: noun, the pile of leaves that Frog had raked for Toad: noun phrase
    • Accommodating syntactic ambiguity
      • What if Frog is being chased by a big animal with many sharp teeth?
      • We will ride down this big hill on my sled.
      • We need more intelligent leaders.
      • Time flies like an arrow.
• Syntax vs. semantics
  • Syntactic anomaly: Book the table on is.
  • Semantic, but not syntactic, anomaly: Colorless green ideas sleep furiously. (but even this is interpretable)
• Syntactic parsing and semantic parsing: a matter of degree
• Frog ran through the woods so that Toad would not see him. -->

  [a superficial (purely syntactic) parse]
  (s (np (pn FROG))
     (vp (tense PAST)
         (iv RUN)
         (pp (p THROUGH)
             (np (det THE) (n WOODS)))
         (dep-clause
             (conj SO-THAT)
             (s (np (pn TOAD))
                (vp (aux WOULD)
                    (neg NOT)
                    (tv SEE)
                    (np (pron HIM)))))))
  
  [a deeper, more semantic parse]
  (RUN1 (event-type RUN)
        (agent FROG)
        (path (ref-point WOODS1)
              (dir THROUGH))
        (time (ref-point (^time TAKE1))
              (dir AFTER))
        (purpose (event-type PREVENT)
                 (agent FROG)
                 (theme (event-type SEE)
                        (experiencer TOAD)
                        (theme RUN1))))
  
  [an even deeper "parse"]
  {a movie in which Frog is running through the woods;
   we cut to an scene in his mind where Toad fails
   to notice him}