From Automatism to Autonomy

Abstract

When we refer to something as automatic in ordinary language, we tend to speak of it as unconscious and working by itself —machinic, repetitive, needing no intervention or control from others to move along its natural course. If a process is automatic, we regularly assume that it happens independently of the human will. What is automated, in other words, will go on until non-human physical constraints prevent it from further labor, such as when the battery is dead in the robot or when the electricity goes out as the washing machine is running its usual course, or when one of its parts is worn out and needs repair. But if the machine “decides” that it is too tired or having a moody afternoon and wants to stop working mid-way through a task, we can’t help feeling very alarmed. Usually, we see automatism as precluding autonomy. Its automatic nature seems to suggest that it is, or ought to be, heteronomous in the sense that its course of action remains the same until it is told otherwise, e.g., when someone else turns the switch on or off. The contrast between the two statuses is prevalent in philosophical discourses as well, notably Descartes’ thought experiment that an automaton designed to look like an animal would be hard to distinguish from the real thing, but a machine that imitates humans would be far easier to detect, due to the latter’s language and general reasoning abilities, which reflect the fact that it is guided by immaterial mind.