Your clock has a purpose. Does its behavior?
Being built for a purpose is different from acting toward a goal. Two accounts of machine behavior give that distinction teeth.
Your clock was built for a purpose. That did not make its behavior purposeful in the classification proposed by Rosenblueth, Wiener and Bigelow in Behavior, Purpose and Teleology. They distinguished orderly movement from behavior directed toward a final condition. An uncomfortable demotion for such a dependable employee. [1]
Their approach examined inputs, outputs and relations with the surroundings. Within purposeful behavior, they distinguished feedback from non-feedback cases. In the narrower sense they used, negative feedback means that error relative to a goal helps control behavior. They also described how inadequate damping could produce repeated overshooting. Correction, in other words, is something to examine, not a magic word that guarantees success. This was the authors’ proposed classification, not a claim that every machine operates this way. [1]
The later weak notion of agency presented by Wooldridge and Jennings groups four properties: autonomy, social ability, reactivity and proactiveness. A system has some control over its actions and internal state, communicates with others, responds to environmental changes and can take goal-directed initiative. They suggest conceptualizing an agent as a UNIX-like software process exhibiting these properties. Software does not escape the question of what its actions do. [2]
Read together, these accounts offer useful questions rather than a single family tree: what counts as a goal, what information returns after action, and how much control does the system exercise? Comparing the descriptions does not establish identical machinery or direct influence. It does let us discuss action without inventing a private emotional life for the device. I shall keep my fictional imperial ambitions out of the equipment specification. [2] [1]
