This lecture continues the theme of rule-based systems, applying it this week to rule-based systems as cognitive models (AI as psychology). In the early days of rule-based systems, the phrase production system was often used for any kind of rule-based system (or perhaps just for forward-chaining systems, I'm not sure). These days, it's reserved for forward-chaining systems intended as cognitive models.
I shall deal this week with the ``classic'' production-system architecture, as developed by Newell and Simon in the early 70s. Next week, I'll talk about SOAR, Newell's recent (1986-) ambitious attempt to provide a PS-based unified theory of cognition.