NAL User's Guide

User's Guide of NAL

Input/Output Language

In this demonstration, NAL implements the following formal language, Narsese. The input and output of the system are Narsese judgments.

           <judgment> ::= [<statement> [frequency-value, confidence-value]]
          <statement> ::= <relation>(<term>, <term>)
                        | <compound-statement>
                        | <term>
               <term> ::= <word>
                        | <variable>
                        | <compound-term>
                        | <statement>
           <variable> ::= <independent-variable>
                        | <dependent-variable>(<independent-variable>^*)
           <relation> ::= inheritance
                        | similarity
                        | implication
                        | equivalence
                        | instance
                        | property
                        | inst_prop                                // instance-property
 <compound-statement> ::= negation(<statement>)
                        | conjunction([<statement>, <statement>⁺]) 
                        | disjunction([<statement>, <statement>⁺]) 
      <compound-term> ::= ext_set([<term>⁺])                       // extensional set
                        | int_set([<term>⁺])                       // intensional set
                        | ext_intersection([<term>, <term>⁺])      // extensional intersection
                        | int_intersection([<term>, <term>⁺])      // intensional intersection
                        | ext_difference(<term>, <term>)           // extensional difference
                        | int_difference(<term>, <term>)           // intensional difference
                        | product([<term>, <term>⁺])
                        | ext_image(<term>, <term>)                // extensional image
                        | int_image(<term>, <term>)                // intensional image

The frequency-value is a real number in [0, 1]; the confidence-value a real number in (0, 1).

User Interface

The program can be invoked in the following ways:

revision(J1, J2, J).
Judgment J is the result of a revision between judgments J1 and J2. The three judgments all have the same statement in them.
choice(J1, J2, J).
Judgment J is the result of a choice between judgments J1 and J2.
inference(J1, J2, J).
Judgment J is the conclusion derived from judgments J1 and J2 as premises.
inference(J1, J).
Judgment J is the conclusion derived from judgment J1 as single premise.