Neural dynamic concepts for intentional systems

AbstractHow may intentionality, the capacity of mental states to be about the world, emerge from neural processes? We propose a set of theoretical concepts that enable a simulated agent to have intentional states as it perceives, acts, memorizes, plans, and builds beliefs about a simulated environment. The concepts are framed within Dynamic Field Theory, a mathematical language for neural processes models at the level of networks of neural populations. Inspired by Searle's analysis of the two directions of fit of intentional states, we recognize that process models of intentional states must detect the match of the world to the mind (for ``action'' intentions) or the match of the mind to the world (for ``perceptual'' intentions). Neural representations of Searle's condition of satisfaction implement these detection decisions through dynamic instabilities that are instrumental in enabling autonomous switches among intentional states.

Return to previous page