Dialog 001: Utterance 1 and 2

Modeling Dialog

The dialog is given to us as 
  an ordered set of natural language sentenses
  We imagine there is some form of NLP to change
  these into something more structured.
  We do not
    explain how this happens.  It is done by some other system.
  We do
    define some possible outputs such a system might produce,
  We call these utterances.

A conversation is a sequence of urrerances
From each utterrance we determine one or more speech acts  
  A speech act is
    the hearers understanding (capta) of the intended meaning
    of the speakers utterance (data)
  different listeners may derive different speech acts 
    after hearing the same utterance
Then a listener updates its world model
  as a result of hearing the utterance

In our dialog are four [agent]s:
  M (a medic) P1 (patient of M)
  C (a computer (advisor to M on medical problems)
  O (an observer who overhears a dialog between M and C)

and lots of 'things': eg (= I1 the-injury) (of P1)

====================================================================
In the beginning (by magic?)

C M are in a conversation C1
which is being overheard by O
O knows about them both
they know about each other
they trust each other

This is a description of the head of O
This assumes
  that the name used by O for himself is 'O'
  Common Knowledge
    are facts of O that he will attribute to all agents
    (IMPLIES (FACT-OF |common-knowledge| fact)
      (CONSEQUENCE-OF (declare A (language-of fact)) fact) )


  --In O------------------------
  (AGENT? O)
  (CONVERSATION? C1)
  (OVERHEARING C1 O)
  (PARTICIPANT C1 M)
  (PARTICIPANT C1 C) 
  (CAN-HEAR C1 (list-mak O M C))

     --In |common knowledge|-------
        (MEDIC? M)
        (AGENT? C)
        (MUTUAL-TRUST M C)
        (CAN-HELP C [medical-problem])
        (WITHIN-HEARING-DISTANCE M C)
     --In M------------------------ 
        --In C------------------------ 

     --In C------------------------ 
        --In M------------------------ 

====================================================================

The dialog
----------
[U1] We have an injured person.
[U2] Where is the injury?

====================================================================
[U1] We have an injured person.

(hears O (say M (declarative-sentence 
  (have we an-injured-person)
)))        

    ------------------------------------------------- 
    ; utterance (as recieved from the NLP system)
    ; the sentence as understood by the NLP is like
    ; "We have a person who is injured"
    ; which is more ackward to say
    ; this all depends how NLP treats adjectives

    (1st C1 U1)
    (isA U1 utterance) 
    (said-by U1 M)
    (said-to U1 C) 
    (content-of U1 DS1-1)
      (isA DS1-1 declarative-sentence)
      (statement DS1-1 REL1-1)
        (isA REL1-1 relation)
        (nlp-verb REL1-1 have)
        (nlp-subject REL1-1 we)
          (isA we pronoun)
          (pronomial-referent we C)
          (pronomial-referent we M)
        (nlp-direct-object REL1-1 P1)
          (isA P1 person)
          (problem-of P1 injured)
    (timestamp U1 T1)

    ------------------------------------------------- 
    ; the resulting speech acts (as objects)

    ; M asks C for help with this medical problem
    (isA SA1-1 request-for-help)
    (from SA1-1 U1)
    (subject SA1-1 medical-problem)

    ; M informs C that M has an injured-person  
    (isA SA1-2 inform) 
    (from SA1-2 U1)
    (content SA2-1 REL1-2)
      (isA REL1-2 relation)
      (relsym REL1-2 injured)
      (arg1 REL1-2 P1)
        (isA P1 patient) 

    ------------------------------------------------- 
    ; Common Knowledge
    
    (isA M medic)
    (isA C agent)  
    (isA O agent)
    (mutual-trust M C)   
    (can-help C medical-problem)
    (within-hearing-distance M C)
    -new-------------------------
    (isA PR1-1 medical-problem)
    (hypothesis PR1-1 REL1-2)
    (solution PR1-1 REL1-3)
      (isA REL1-3 relation)
      (relsym REL13 stabalized)
      (arg1 REL1-3 P1)

    ------------------------------------------------- 

  --In O--------------------------- 
  (isA C1 conversation)
  (overhearing O C1)
  (participant C1 M)
  (participant C1 C)
  (all-can-hear C1 O M C)
  -new-----------------------------
  REL1-2
  (problem M PR1-1)

  --Expectations of O--------------
  (belief M REL1-3))
  (belief C REL1-3))

     --In M------------------------ 
     --Goals of M------------------
     REL1-3
        --In C------------------------ 
        REL1-3

     --In C------------------------ 
     --Goals of c------------------
     REL1-3
        --In M------------------------ 
        --Goals of M------------------
        REL1-3


====================================================================
[U2] Where is the injury?

(hears O (say C (question-where-is
  (where-is the-injury)
)))        

    ------------------------------------------------- 
    ; utterance

    (2nd C1 U2)
    (isA U2 utterance) 
    (said-by U2 C)
    (said-to U2 M) 
    (content-of U2 Q2-1)
      (isA Q2-1 question-where-is)
      (what Q2-1 the-injury) 
    (timestamp U2 T2)

    ------------------------------------------------- 
    ; the resulting speech acts

    (isA SA2-1 where-on-body) 
    (from SA2-1 U2)
    (content SA2-1 I1)
      (isA I2-1 injury)
      (of I2-1 P1)