One of the relevant aspects of Seth’s discussion is the way in which it highlights interesting links to theoretical precursors of PP. In doing so, he broadens the historical context in which the framework is usually situated. However, these considerations are not just relevant for the history of science, they also constitute a theoretical underpinning of several ways in which Seth has recently developed PP accounts of various phenomena. Due to limited space, I can only address some of these here. In particular, I will focus on his three interpretations of active inference, and on his PP account of perceptual presence. In so doing, I will also try to take the analogy between explanation in perception and explanation in science a little further than it has previously been taken.
In section 2, I will briefly summarize Seth’s view on the connection between cybernetics and the free-energy principle. One of the results of his considerations is that a distinction can be drawn between three types of active inference. The first type involves confirmatory hypothesis-testing. The other types involve seeking disconfirming and disambiguating evidence, respectively. Seth does not say much about what it takes to disconfirm or falsify a hypothesis or model. Furthermore, he seems to suggest that not all types of active inference he distinguishes are currently part of PP (at least in the version described by Karl Friston’s FEP): “[t]hese points represent significant developments of the basic infrastructure of PP” (Seth 2014, p. 3).[1] In section 3, I will provide clarification of the notion of falsification by referring to the works of Karl Popper, Imre Lakatos, and Thomas Kuhn. I will also provide examples to show that different types of falsification are part and parcel of PP, not extensions of the basic infrastructure. In section 4, I point out an ambiguity in Seth’s account of perceptual presence (perceptual presence vs. objecthood). After this, I suggest that counterfactual richness may not be the crucial underlying feature (of either perceptual presence or objecthood). Giving a series of examples, I argue that the degree of represented causal integration is an equally good candidate for accounting for perceptual presence (or objecthood), although more work needs to be done.