From the actor model to agent cognition

The purpose and structure of software agents are grounded in ideas that emerged from early computation models, particularly the actor model that was introduced by Carl Hewitt in the 1970s (Hewitt et al. 1973).

The actor model treats computation as a collection of independent, concurrently executing entities called actors. Each actor encapsulates its own state, interacts solely through asynchronous message passing, and can create new actors and delegate tasks.

This model provided the conceptual foundation for decentralized reasoning, reactivity, and isolation—all of which underpin the behavioral architecture of modern software agents.