A more basic version of agency? As if!

Abstract

At present there is just one definition [2] that attempts to explicitly naturalise the concept of (general, embodied) agency, and it is unapologetically autopoietic-enactivist. This fact constitutes a public challenge to other traditions in cognitive science. A ‘bare’ problem of defining agency remains, even after paring away hard phenomenological and normative problems by limiting the scope of the problem to describing ‘as-if’ agency (i.e. the external appearance of agency). Building on [2], I identify an extended list of criteria that a theory of agency (whether ‘true’ or ‘as-if’) should meet. I argue that autopoiesis is the wrong foundation even for ‘bare’ agency, let alone phenomenological and volitional agency; instead, I recommend starting with an ‘as-if’ definition that relates agency to some theory of embodied rationality, effectively providing a generalised version of Dennett’s intentional stance [9], and taking a step towards a rigorous formal definition.