HOW MUCH OF PHYSICS IS JUST COMPUTATION

We explore an emergent interpretation of the action integral and the l agrangian in physics, and discuss its connection with the concept of ' amount of computation'. We give an abstract definition of action, and argue (1) that it provides a general, model-independent characterizati on of 'amount of computation'; and that (2) the action of physics is a special case of this general action. Much as entropy quantifies the l ack of information one has about the state of a system, action quantif ies the lack of information about the system's law-or, equivalently, i ts behavior. In this approach, action is to dynamics what entropy is t o statics. (C) 1998 Academic Press Limited.