Logical formalisms

The precursory ideas of McCarthy’s Advice Taker were developed by McCarthy and Hayes into the Situation Calculus: a logical formalism for reasoning about real-life situations. These ideas later motivated the development of the Event Calculus and a range of other action and fluent-based calculi. In such formalisms, events, world-states, actors, properties and physical entities are denoted by the objects, terms and functors of a symbolic logic. These formalisms therefore operate over an assumed domain of pre-categorized knowledge and actions are defined in terms of the changing properties of an environment.

For example, the following expression may be used to express that when someone drops a book in a room, that book can then be found in that room (Section 6.47 of Commonsense Reasoning):

HoldsAt(InRoom(a,r),t) =>
Initiates(LetGoOf(a,o), InRoom(o,r), t)

Action logics have found wide application in intelligent systems. For example:
  • General game playing, wherein systems must discover strategies and compete in novel games without human oversight
  • One-touch-mission systems, requiring the ability to modify plans on the fly when obstacles or problems are encountered
  • Intelligent characters in computer games that must communicate with human beings and independently discover plans to achieve their own ends