DATA HOMEWORK 4

This homework is on the information-structure annotations of the Switchboard corpus.

Skim this guidelines paper on the annotations:
Nissim, Malvina. 2003. Annotation Scheme for Information Status in Dialogue.
Go to the SWBD information-structure annotations in
```
/afs/ir.stanford.edu/data/linguistic-data/Treebank/LINK-swbd/backtrans
```
Not all the dialogues there are annotated for information structure; you can do
```
egrep -l '_old' * | more
```
to find out which ones.

Look at the file sw2041.backtranslated.mrg. Explain why "the aerobics class" is mediated in the following sentence:

( (S
 (SBAR-SBJ
  (WHADVP-N40040B (WRB how) )
  (S
   (NP-SBJ_MARKABLE_human_old_general (PRP I) )
   (VP (VBD met) 
    (NP_MARKABLE_human_old_ident_ANAPH2_ANTEC3 (PRP him) )
    (ADVP (-NONE- -N40040B) ))))
 (VP (VBD was) 
  (PP-PRD (IN through) 
   (, ,)
   (INTJ (UH uh) )
   (, ,)
   (NP_MARKABLE_nonconc_med_set
    (NP_MARKABLE_nonconc_med_set_ANTEC29 (DT the)  (NN aerobics)  (NN class) )
    (SBAR
     (WHNP-N40044D_MARKABLE_nonconc_old_relative_ANAPH29 (WDT that) )
     (S
      (NP-SBJ-N40043E_MARKABLE_human_old_ident_ANAPH3 (PRP he) )
      (VP (VBD used) 
       (S
        (NP-SBJ_MARKABLE (-NONE- -N40043E) )
        (VP (TO to) 
         (VP (VB teach) 
          (NP_MARKABLE (-NONE- -N40044D) ))))))))))
 (. .) (-DFL- E_S) ))
( (CODE (SYM SpeakerA7) (. .) ))

Write a pgraph on something interesting you notice anywhere in the data.