CONCEPT
Processing vs. Enacting
Evan Thompson’s sharpest diagnostic distinction—between a system that manipulates representations without any stake in the outcome and an organism that brings forth a world of significance through its embodied activity—the fault line along which every honest claim about AI must be placed.
The distinction between processing and enacting is the load-bearing wall of
Evan Thompson’s entire intellectual edifice, and it is the distinction that the AI moment has made simultaneously most important and most difficult to maintain. A system processes when it manipulates symbols according to rules, producing outputs that are evaluated externally for their utility—the system has no stake in the quality of what it produces, no experience of getting it right or wrong, no inside from which the output means anything. A system enacts when it brings forth a world of significance through its own activity: it makes sense of its environment in terms of its own needs, acts on the basis of that sense-making, and is constitutively changed by the quality of the engagement. The bacterium navigating a sugar gradient is enacting a world in which sugar is food;
large language models generating text about nutrition are processing representations of a world