Rules in Programming Languages and Networks
Having found these numerous failed attempts to contrast networks and pro gramming languages in terms of rules, we are moved to question the motivation underlying these attempts. What reason is there to suppose that there is some grand difference between connectionist models and programming language mod els in their use of rules? Why have people thought it important for there to be greater differences between networks and Turing machines than there are be tween, say, Turing machines and LISP programs? Why are we not merely being deceived by the fact that units and connections do not look like computer pro grams?