Computer generations of the past have always been

          differentiated by innovations in hardware.  We have gone from

          vacuum tubes to transisitors, then from integrated circuits to

          very large scale integration.[1] The "supercomputers" of today

          are considered "fourth-generation" even though their speed in

          number crunching is phenomenal.  Janet Raloff in her article

          "Approaching the Age of Reason" states that innovations of the

          fifth generation "in software, are expected to be as essential

          and integral to performance as the requisite innovation in

          hardware".[2]


            "Today computing is generally referred to as data processing"

          says Raloff.[3] Data is stored in computer memory as a series of

          1's and 0's: a series of electrical inpulses.  What

          differentiates fifth generation computers is that they "use

          knowledge instead of data as the basic units to be processed."[4]
          _________

          Fifth generation computers will not simply calculate, they will

          "infer, reason, and learn" Raloff says.  These computers will

          also be capable of an array of "user-friendly" tasks.  Raloff

          details:

               "...the ability to 'hear' and comprehend spoken,
               conversational language; the ability to translate
               between languages nearly as quickly as they are spoken
               or read; the ability to 'see' and comprehend the
               meaning of visual data such as maps, photographs, and
               hand-written script; the ability to search for and
               retrieve requested data from stores of knowledge filed
               locally or in some distant library---data that may only
               have been referred to casually; and finally the ability
               to solve problems even when the details about the
               problem are somewhat sketchy"[5]




                                        - 2 -










            These machines obviously possess a high level of intelligence,

          in fact, what is called Artificial Intelligence.  What are the

          main components of AI?  What are the problems involved with it?


            M.M. Waldrop recognizes the three main problems of AI as "the

          representation of knowledge,or memory, the control and use of

          knowledge, and the acqisition of knowledge, or learning".[6] I

          will focus on the question of representation of knowledge and the

          control of that knowledge in this paper.  How much must a

          computer know in order to exhibit intelligent behavior?  Where

          does it go from there?





          1 Repesentation of Knowledge
          ____________________________



          1.1 The Logic Approach


            "The oldest and simplest form of representation is symbolic

          logic" says Waldrop.  Symbolic logic is the use of a list of true

          statement