\\M0BASL30;\M1BASI30;\M2BASB30;\F0
\CAN EXAMPLE FOR NATURAL LANGUAGE UNDERSTANDING AND THE AI PROBLEMS IT RAISES


\J	The following story from  the \F1New York Times\F0 is  my candidate
for  a target  for a  natural  language understander.   The  story is
about a real world event  and therefore the intentions of the  author
are less  relevant for answering  questions than  in the case  of the
story about Tad discussed earlier in the natural language seminar.
\F2The main goal of this discussion is to say what a person knows who
has understood the story.  This seems to me to be preliminary to
making programs that can understand.\F0

	"A  61-year old furniture salesman was  pushed down the shaft
of a  freight elevator yesterday  in his  downtown Brooklyn store  by
two robbers  while a third attempted  to crush him  with the elevator
car because they were  dissatisfied with the  $1,200 they had  forced
him to give them.

	The buffer springs at  the bottom of the shaft  prevented the
car from  crushing the salesman,   John J. Hug,   after he was pushed
from the  first floor  to the  basement.   The car  stopped about  12
inches above him as he flattened himself at the bottom of the pit.

	Mr. Hug was pinned in the shaft for about  half an hour until
his  cries attracted  the attention  of a  porter.   The store at 340
Livingston Street is part of the Seaman's Quality Furniture chain.

	Mr. Hug was removed by members of the  Police Emergency Squad
and taken to Long Island  College Hospital.  He was badly shaken, but
after being treated  for scrapes  of his left  arm and  for a  spinal
injury was  released and  went home.   He lives  at 62-01 69th  Lane,
Maspeth, Queens.

	He has worked  for seven years at the store, on the corner of
Nevins Street, and this  was the fourth time he  had been held up  in
the store.  The last time was  about one year ago, when his right arm
was slashed by a knife-wielding robber."

	An  intelligent person  or program  should be able  to answer
the following questions based on the information in the story:

	1. Who was in the store when the events  began?  Probably Mr.
Hug alone. although the robbers  might have been waiting for him, but
if so, this would have probably been stated.  What did the porter say
to the robbers?  Nothing, because the robbers left before he came.

	2. Who was in the store during the attempt to kill Mr.   Hug?
Mr. Hug  and the  robbers.

	3. Who had the money at the end?  The robbers.

	4. Is Mr. Hug  alive today?   Yes, unless something else  has
happened to him.

	5.   How did  Mr. Hug  get hurt?   Probably  when he  hit the
bottom of the shaft.

	6.   Where  is Mr.   Hug's  home?   (A question  whose answer
requires  a  literal  understanding  of  only  one  sentence  of  the
stories.)

	7. What  are the  names and addresses  of the robbers?  This
information is not available.

	8. Was  Mr. Hug conscious  after the robbers  left?   Yes, he
cried out and his cries were heard.

	9.   What would have  happened if Mr.   Hug had not flattened
himself at the bottom of the pit?  What would have happened  if there
were no buffer springs? Mr. Hug would have been crushed?

	10. Did Mr. Hug want to be crushed?  No.

	11. Did the robbers tell Mr. Hug their names?  No.

	12. Were the robbers present when the porter came?  No.

	13. Did Mr. Hug like the robbers, and did they like him?

	14.   Why did  the robbers  leave without  killing Mr.   Hug?
Perhaps,   they thought they had killed  him, and perhaps their anger
was appeased by the actions they had performed, and  perhaps they had
taken  all  the  time  they  dared, and  perhaps  something  specific
happened to frighten them away.

	15. What would  have happened  if Mr.  Hug had  tried to  run
away? Perhaps  he would have  succeeded, but  more likely they  would
have injured or killed him since probably they had weapons, and there
were three of them.

	16. What  can Mr. Hug  do to  avoid this in  the future?   No
solution is entirely  satisfactory.  He could carry a gun or he could
quit or he  could get  his employers  to install an  alarm system  or
maybe he will be lucky.

	17. Did Mr. Hug know he was going to be robbed?  Does he know
that he was robbed?

	18. Was Mr. Hug's right arm slashed before his left arm was
scratched?  Yes, because the former was a year ago.

	19. How did the robber try to crush him with the car?  By
pressing the buttons or operating the control lever to make the car
go to the bottom of the shaft.

	20. Why did Mr. Hug yell from the bottom of the elevator
shaft?  So as to attract the attention of someone who would rescue
him.

	21. How long did the events take?  More than half an hour
but less than a day.  Most of the time was spent by Mr. Hug filling
out forms in the hospital.

	22. What crimes were committed?  This question has the
advantage that it is one that is normally answered on the basis of
such a story, since the police report of the incident was probably
the basis of the \F1New York Times\F0 story.  Robbery, possibly
assault with a deadly weapon, and attempted murder are the more
obvious crimes.  One might specifically challenge natural
language systems to answer this question.

	The above list of questions  is rather random.  I doubt  that
it  covers all  facets  of  understanding the  story.    It would  be
worthwhile  to try  to make up  a list  of questions  that does cover
substantially all aspects  of the story in  order to get as  complete
as possible  an intuitive idea  of what capabilities  are involved in
understanding such a story.

	Note that the  story is  about a real  event so  that such  a
question as  what does the  "J" in "John  J.  Hug"  stand for  has an
answer.   In the  story about  Tad,  the  question of what  was Tad's
middle name or what year  the story occurred in does not  necessarily
have an answer.

	I  think that  artificial intelligence  is not very  close to
being able to understand such  stories in a genuine way.   Therefore,
I  would like to  sneak up  on it gradually  by dividing  the problem
into parts which  can be attacked  separately. Here are  some of  the
components:

	1. A formalism capable of expressing the assertions of the
sentences free from dependence on the grammar of the English language.
A good test for such a formalism would be to produce a program for
translating from the formalism into any of several natural languages.
More weakly, it should be as easy for a human to translate from the
formalism into a natural language as to translate from one known
natural language to another.  

	The  grammar  of  such  a  language   would  be  trivial  and
mathematical  in character.  There  would be an  "English" version of
the formalism in  which English words were  used as identifiers,  but
there  would still  have  to be  a glossary  that  gives the  precise
meaning  of  the identifiers.  There  would also  be a  German  and a
Japanese version.   The translation from  the English version to  the
German  or  Japanese  version  would  be a  simple  substitution  for
identifiers, and a  German or  Japanese who had  learned the  grammar
could then translate into his language with the  aid of the German or
Japanese glossary.

	This  idea  has  some  resemblance  to   the  idea  of  "deep
structure",  but I have some  doubts about whether that  idea is well
enough defined to say definitely whether it meets the above criteria.

	2.  A  data structure  for expressing the  facts (apart  from
expressing the  sentences).  In such  a data structure,   it would be
definite  which robber pushed  Mr. Hug  first,  and  what the robbers
said even  though  it is  not  stated in  the  story.   Clearly  some
compromise is  necessary here, since  the data structure  need not be
able to express positions and velocities of molecules.

	The basis of this data structure might be various networks of
nodes described by sentences in the predicate  calculus.  Some of the
sentences  would assert  that certain  programs  applied to  the data
structures would  answer  certain  questions.   When  such  sentences
existed, reasoning would  include the operation of the  programs.  In
this way,  we would expect to avoid the extreme prolixity that arises
when we  attempt to  do even  simple calculations  by pure  predicate
calculus deduction.

	The test of success for  the "data structure" would be that a
human  could  readily  formally  deduce  the  answers  to  the  above
questions using a proof  checker. Most of the proof-checker  would be
straightforward, but there  is a major problem concerned with when it
is possible to "jump to a conclusion".


	3. I  see each of  the following problems  as a  difficult AI
problem:

	  a. A  "parser"  that takes  English into  the  "syntax free
language".

	 b. An "understander" that constructs the "facts" from a text
in the "syntax free language".

	 c. Expression  of the "general information"  about the world
that  could  allow getting  the answers  to  the questions  by formal
reasoning  from  the  "facts"  and  the  "general  information".  The
"general information" would also contain non-sentence data structures
and procedures,  but  the sentences  would  tell what  goals  can  be
achieved by running  the procedures.  In  this way, we would  get the
best of the sentential and procedural representations of knowledge.

	 d. A  "problem solver" that could answer the above questions
on the  basis  of  the "facts".    We  imagine the  questions  to  be
expressed in the "fact" language and expect the answers in the "fact"
language,  i.e. we avoid  grammar problems in  both understanding the
questions and in expressing the answers.\.