Wanja Wiese

Vanilla PP for Philosophers: A Primer on Predictive Processing

The goal of this short chapter, aimed at philosophers, is to provide an overview and brief explanation of some central concepts involved in predictive processing (PP). Even those who consider themselves experts on the topic may find it helpful to see how the central terms are used in this collection. To keep things simple, we will first informally define a set of features important to predictive processing, supplemented by some short explanations and an alphabetic glossary. The features described here are not shared in all PP accounts. Some may not be necessary for an individual model; others may be contested. Indeed, not even all authors of this collection will accept all of them. To make this transparent, we have encouraged contributors to indicate briefly which of the features are necessary to support the arguments they provide, and which (if any) are incompatible with their account. For the sake of clarity, we provide the complete list here, very roughly ordered by how central we take them to be for “vanilla PP” (i.e., a formulation of predictive processing that will probably be accepted by most researchers working on this topic). More detailed explanations will be given below. Note that these features do not specify individually necessary and jointly sufficient conditions for the application of the concept of “predictive processing”. All we currently have is a semantic cluster, with perhaps some overlapping sets of jointly sufficient criteria. The framework is still developing, and it is difficult, maybe impossible, to provide theory-neutral explanations of all PP ideas without already introducing strong background assumptions. Nevertheless, at least features 1-7 can be regarded as necessary properties of what is called PP in this volume.

Perceptual Presence in the Kuhnian-Popperian Bayesian Brain

Anil Seth ’s target paper connects the framework of PP (predictive processing) and the FEP (free-energy principle) to cybernetic principles. Exploiting an analogy to theory of science, Seth draws a distinction between three types of active inference . The first type involves confirmatory hypothesis-testing. The other types involve seeking disconfirming and disambiguating evidence, respectively. Furthermore, Seth applies PP to various fascinating phenomena, including perceptual presence . In this commentary, I explore how far we can take the analogy between explanation in perception and explanation in science. In the first part, I draw a slightly broader analogy between PP and concepts in theory of science, by asking whether the Bayesian brain is Kuhnian or Popperian. While many aspects of PP are in line with Karl Popper ’s falsificationism, other aspects of PP conform to how Thomas Kuhn described scientific revolutions. Thus, there is both a sense in which the Bayesian brain is Kuhnian, and a sense in which it is Popperian. The upshot of these considerations is that falsification in PP can take many different forms. In particular, active inference can be used to falsify a model in more ways than identified by Seth . In the second part of this commentary, I focus on Seth ’s PPSMCT (predictive processing account of sensorimotor contingency theory) and its application to perceptual presence, which assigns a crucial role to counterfactual richness . In my discussion, I question the significance of counterfactual richness for perceptual presence. First, I highlight an ambiguity inherent in Seth ’s descriptions of the target phenomenon (perceptual presence vs. objecthood ). Then 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.

Jakob Hohwy: The Predictive Mind

The Predictive Mind by Jakob Hohwy is the first monograph to address the philosophical significance of what Hohwy calls the prediction error minimization framework (henceforth: PEM). The central claim of the book is that, on a conceptual level, perception, action, and cognition can be understood by reference to a single principle: prediction error minimization (p. 1; unless indicated otherwise, all page numbers refer to Hohwy 2013). The corresponding empirical hypothesis is that the brain implements a hierarchical generative model that generates predictions about sensory inputs and their hidden causes (it is called generative model because it models the causes of sensory data). When sensory signals arrive, only their divergence from the predictions has to be further processed. The general strategy of using predictions derived from generative models to compress and transmit information is also known as predictive coding. Perception is thus not construed as a purely bottom–up process, but rather as an (unconscious, subpersonal) probabilistic inference process, in which the brain derives from its model a hypothesis about what is going on in the world; the sensory input is then only used to test the accuracy of the hypothesis and to update the parameters of the model from which it is derived. Updates are performed in accord with principles of Bayesian learning. In the first part of the book (chapters one to four), Hohwy explains the core ideas of PEM in a way that is accessible to philosophers. He thereby builds a foundation for the discussion in the second and third part of the book, in which he applies PEM to a variety of cognitive phenomena (with a focus on perception). I will now describe the basic idea in a bit more detail, and then turn to some critical remarks, which will focus on the way Hohwy treats the possible conceptual unification of action and perception.

Ethics of Beliefs

This paper deals with the concept of positive learning (PL). The main goal is to provide a working definition of PL on which further refinements and extensions can be based. First, I formulate a list of desiderata for a definition of PL: I argue that a working definition of PL should (i) make the involved epistemic norms explicit, (ii) be flexible, and (iii) be empirically tractable. After that, I argue that a working definition of PL should focus on three basic epistemic norms (which I call Evidentialism, Degrees of Plausibility, and Non-Arbitrary Updates). Drawing on work on the ethics of belief and Bayesian inference, I highlight theoretical and empirical challenges that already follow from such basic assumptions. Finally, I formulate a working definition of PL based on the three epistemic norms and show that it fulfills the desiderata given above. Furthermore, I also provide a tentative agenda for future research that seeks to develop the notion of PL in various ways that are relevant to PLATO in general, and to a “practical philosophy of mind” in particular.

Towards a Mature Science of Consciousness.

In Being No One, Metzinger (2004[2003]) introduces an approach to the scientific study of consciousness that draws on theories and results from different disciplines, targeted at multiple levels of analysis. Descriptions and assumptions formulated at, for instance, the phenomenological, representationalist, and neurobiological levels of analysis provide different perspectives on the same phenomenon, which can ultimately yield necessary and sufficient conditions for applying the concept of phenomenal representation. In this way, the “method of interdisciplinary constraint satisfaction (MICS)” (as it has been called by Weisberg, 2005) promotes our understanding of consciousness. However, even more than a decade after the first publication of Being No One, we still lack a mature science of consciousness. This paper makes the following meta-theoretical contribution: It analyzes the hurdles an approach such as MICS has yet to overcome and discusses to what extent existing approaches solve the problems left open by MICS. Furthermore, it argues that a unifying theory of different features of consciousness is required to reach a mature science of consciousness.

Predictive Processing and the Phenomenology of Time Consciousness

This chapter explores to what extent some core ideas of predictive processing can be applied to the phenomenology of time consciousness. The focus is on the experienced continuity of consciously perceived, temporally extended phenomena (such as enduring processes and successions of events). The main claim is that the hierarchy of representations posited by hierarchical predictive processing models can contribute to a deepened understanding of the continuity of consciousness. Computationally, such models show that sequences of events can be represented as states of a hierarchy of dynamical systems. Phenomenologically, they suggest a more fine-grained analysis of the perceptual contents of the specious present, in terms of a hierarchy of nested temporal wholes. Visual perception of static scenes not only contains perceived objects and regions but also spatial gist; similarly, auditory perception of temporal sequences, such as melodies, involves not only perceiving individual notes but also slightly more abstract features (temporal gist), which have longer temporal durations (e.g., emotional character or rhythm). Further investigations into these elusive contents of conscious perception may be facilitated by findings regarding its neural underpinnings. Predictive processing models suggest that sensorimotor areas may influence these contents. [1] I am highly grateful to Martin Butz Martin Butz, Jakob Hohwy Jakob Hohwy, Marius Jung Marius Jung, Thomas Metzinger Thomas Metzinger, Mark Miller Mark Miller, Iuliia Pliushch Iuliia Pliushch, and Lisa Quadt Lisa Quadt for providing a number of very useful comments on drafts of this paper. Thanks to Robin Wilson Robin Wilson for excellent editorial help.