The main aim of this article is to give an assessment of prediction error minimization as a unifying theoretical framework for the study of social cognition. We show how this framework can be used to synthesize and systematically relate existing data from social cognition research, and explain how it introduces new constraints for further research. We discuss PEM in relation to other theoretical frameworks of social cognition, and identify the main challenges that this approach to social cognition will need to (...) address. (shrink)

Some naturalistic philosophers of mind subscribing to the predictive processing theory of mind have adopted a realist attitude towards the results of Bayesian cognitive science. In this paper, we argue that this realist attitude is unwarranted. The Bayesian research program in cognitive science does not possess special epistemic virtues over alternative approaches for explaining mental phenomena involving uncertainty. In particular, the Bayesian approach is not simpler, more unifying, or more rational than alternatives. It is also contentious that the Bayesian approach (...) is overall better supported by the empirical evidence. So, to develop philosophical theories of mind on the basis of a realist interpretation of results from Bayesian cognitive science is unwarranted. Naturalistic philosophers of mind should instead adopt an anti-realist attitude towards these results and remain agnostic as to whether Bayesian models are true. For continuing on with an exclusive focus and praise of Bayes within debates about the predictive processing theory will impede progress in philosophical understanding of scientific practice in computational cognitive science as well as of the architecture of the mind. (shrink)

It is an important feature of creatures like us that our various motivations compete for control over our behaviour, including mental behaviour such as imagining and attending. In large part, this competition is adjudicated by the stimulation of affect — the intrinsically pleasant or unpleasant aspects of experience. In this paper I argue that the motivation-affect system controls a sub-type of attention called 'alerting attention' to bring various goals and stimuli to consciousness and thereby prioritize those contents for action. This (...) view allows me to flesh out the global workspace theory of consciousness, as well as some of the phenomenal characteristics of conscious experience. (shrink)

According to Enactivism, cognition should be understood in terms of a dynamic interaction between an acting organism and its environment. Further, this view holds that organisms do not passively receive information from this environment, they rather selectively create this environment by engaging in interaction with the world. Radical Enactivism adds that basic cognition does so without entertaining representations and hence that representations are not an essential constituent of cognition. Some proponents think that getting rid of representations amounts to a revolutionary (...) alternative to standard views about cognition. To emphasize the impact, they claim that this ‘radicalization’ should be applied to all enactivist friendly views, including, another current and potentially revolutionary approach to cognition: predictive processing. In this paper, we will show that this is not the case. After introducing the problem, we will argue that ‘radicalizing’ predictive processing does not add any value to this approach. After this, we will analyze whether or not radical Enactivism can count as a revolution within cognitive science at all and conclude that it cannot. Finally, in section 5 we will claim that cognitive science is better off when embracing heterogeneity. (shrink)

Brains, it has recently been argued, are essentially prediction machines. They are bundles of cells that support perception and action by constantly attempting to match incoming sensory inputs with top-down expectations or predictions. This is achieved using a hierarchical generative model that aims to minimize prediction error within a bidirectional cascade of cortical processing. Such accounts offer a unifying model of perception and action, illuminate the functional role of attention, and may neatly capture the special contribution of cortical processing to (...) adaptive success. This target article critically examines this approach, concluding that it offers the best clue yet to the shape of a unified science of mind and action. Sections 1 and 2 lay out the key elements and implications of the approach. Section 3 explores a variety of pitfalls and challenges, spanning the evidential, the methodological, and the more properly conceptual. The paper ends (sections 4 and 5) by asking how such approaches might impact our more general vision of mind, experience, and agency. (shrink)

Does the material basis of conscious experience extend beyond the boundaries of the brain and central nervous system? In Clark 2009 I reviewed a number of ‘enactivist’ arguments for such a view and found none of them compelling. Ward (2012) rejects my analysis on the grounds that the enactivist deploys an essentially world-involving concept of experience that transforms the argumentative landscape in a way that makes the enactivist conclusion inescapable. I present an alternative (prediction-and-generative-model-based) account that neatly accommodates all the (...) positive evidence that Ward cites on behalf of this enactivist conception, and that (I argue) makes richer and more satisfying contact with the full sweep of human experience. (shrink)

Biological brains are increasingly cast as ‘prediction machines’: evolved organs whose core operating principle is to learn about the world by trying to predict their own patterns of sensory stimulation. This, some argue, should lead us to embrace a brain-bound ‘neurocentric’ vision of the mind. The mind, such views suggest, consists entirely in the skull-bound activity of the predictive brain. In this paper I reject the inference from predictive brains to skull-bound minds. Predictive brains, I hope to show, can be (...) apt participants in larger cognitive circuits. The path is thus cleared for a new synthesis in which predictive brains act as entry-points for ‘extended minds’, and embodiment and action contribute constitutively to knowing contact with the world. (shrink)

The target article sketched and explored a mechanism (action-oriented predictive processing) most plausibly associated with core forms of cortical processing. In assessing the attractions and pitfalls of the proposal we should keep that element distinct from larger, though interlocking, issues concerning the nature of adaptive organization in general.

About three decades ago, AI theory underwent a sharp turn as a consequence of criticism that pointed out the problem of externalism in the cognitivist position. Hubert Dreyfus, undoubtedly the main exponent of this criticism, opened the possibility of a Heideggerian reading using the frame problem to bring to light obscurities that otherwise would have been very difficult to detect. However, the question still remains of whether or not Heidegger’s philosophy can serve as the source of a positive contribution to (...) AI. In this paper, we argue that in the small measure in which such a task has been attempted, its orientation has been hampered by the omission of what, for Heidegger, was the central issue to be pondered: the question for being and the ontological difference. To propose a possible direction in which AI can be headed as a consequence of this novel perspective, we undertake a brief and schematic review of two published projects suggesting that they both manage to avoid the frame problem and also bear this Heideggerian outlook. (shrink)

We suggest that pain processing has a modular architecture. We begin by motivating the (widely assumed but seldom defended) conjecture that pain processing comprises inferential mechanisms. We then note that pain exhibits a characteristic form of judgement independence. On the assumption that pain processing is inferential, we argue that its judgement independence is indicative of modular (encapsulated) mechanisms. Indeed, we go further, suggesting that it renders the modularity of pain mechanisms a default hypothesis to be embraced pending convincing counterevidence. Finally, (...) we consider what a modular pain architecture might look like, and question alleged counterevidence to our proposal. (shrink)

Philosophical proponents of predictive processing cast the novelty of predictive models of perception in terms of differences in the functional role and information content of neural signals. However, they fail to provide constraints on how the crucial semantic mapping from signals to their informational contents is determined. Beyond a novel interpretative gloss on neural signals, they have little new to say about the causal structure of the system, or even what statistical information is carried by the signals. That means that (...) the predictive framework for perception can be relabeled in traditional, non-predictive terms, with no empirical consequences relevant to existing or future data. To the extent that neuroscientific research based on predictive processing is both innovative and productive, it will be due to the framework’s suggestive heuristic effects, or perhaps auxiliary empirical claims about implementation, rather than a difference in the information-processing structure that it describes. (shrink)

Philosophical discussion tends to converge on the view that narratives are at the center of the emotion of grief. In this article, I expand on this kind of view. On the one hand, I argue that key strands of phenomenological and neuroscientific studies suggest that grief consists in a complex emotional process of disconfirmation-and-updating of the narrative self-model. By heuristically drawing on an analogy between binocular rivalry and grief, I show that certain salient aspects of the phenomenology of grief, such (...) as self-alienation and ambiguity, are straightforwardly reflected by this account. On the other hand, I argue that this hierarchical approach has the resources to: (i) show that both narratable and unnarratable experiences of personal loss are both instances of grief; and (ii) differentiate between narratable and unnarratable experiences of grief, without compromising the view that self-constituting narratives are at the center of the phenomenon. (shrink)

In the past two decades, reinforcement learning has become a popular framework for understanding brain function. A key component of RL models, prediction error, has been associated with neural signals throughout the brain, including subcortical nuclei, primary sensory cortices, and prefrontal cortex. Depending on the location in which activity is observed, the functional interpretation of prediction error may change: Prediction errors may reflect a discrepancy in the anticipated and actual value of reward, a signal indicating the salience or novelty of (...) a stimulus, and many other interpretations. Anterior cingulate cortex has long been recognized as a region involved in processing behavioral error, and recent computational models of the region have expanded this interpretation to include a more general role for the region in predicting likely events, broadly construed, and signaling deviations between expected and observed events. Ongoing modeling work investigating the interaction between ACC and additional regions involved in cognitive control suggests an even broader role for cingulate in computing a hierarchically structured surprise signal critical for learning models of the environment. The result is a predictive coding model of the frontal lobes, suggesting that predictive coding may be a unifying computational principle across the neocortex. (shrink)

There is surprising evidence that introspection of our phenomenal states varies greatly between individuals and within the same individual over time. This puts pressure on the notion that introspection gives reliable access to our own phenomenology: introspective unreliability would explain the variability, while assuming that the underlying phenomenology is stable. I appeal to a body of neurocomputational, Bayesian theory and neuroimaging findings to provide an alternative explanation of the evidence: though some limited testing conditions can cause introspection to be unreliable, (...) mostly it is our phenomenology itself that is variable. With this account of phenomenal variability, the occurrence of the surprising evidence can be explained while generally retaining introspective reliability. (shrink)

Two challenges that face popular self-monitoring theories (SMTs) of auditory verbal hallucination (AVH) are that they cannot account for the auditory phenomenology of AVHs and that they cannot account for their variety. In this paper I show that both challenges can be met by adopting a predictive processing framework (PPF), and by viewing AVHs as arising from abnormalities in predictive processing. I show how, within the PPF, both the auditory phenomenology of AVHs, and three subtypes of AVH, can be accounted (...) for. (shrink)

Both mindreading and stereotyping are forms of social cognition that play a pervasive role in our everyday lives, yet too little attention has been paid to the question of how these two processes are related. This paper offers a theory of the influence of stereotyping on mental-state attribution that draws on hierarchical predictive coding accounts of action prediction. It is argued that the key to understanding the relation between stereotyping and mindreading lies in the fact that stereotypes centrally involve character-trait (...) attributions, which play a systematic role in the action–prediction hierarchy. On this view, when we apply a stereotype to an individual, we rapidly attribute to her a cluster of generic character traits on the basis of her perceived social group membership. These traits are then used to make inferences about that individual’s likely beliefs and desires, which in turn inform inferences about her behavior. (shrink)

In this paper, I examine an evolutionary approach to the action selection problem and illustrate how it helps raise an objection to the predictive processing account. Clark examines the predictive processing account as a theory of brain function that aims to unify perception, action, and cognition, but - despite this aim - fails to consider action selection overtly. He off ers an account of action control with the implication that minimizing prediction error is an imperative of living organisms because, according (...) to the predictive processing account, action is employed to fulfill expectations and reduce prediction error. One way in which this can be achieved is by seeking out the least stimulating environment and staying there. Bayesian, neuroscientific, and machine learning approaches into a single framework whose overarching principle is the minimization of surprise (or, equivalently, the maximization of expectation. But, most living organisms do not find, and stay in, surprise free environments. This paper explores this objection, also called the “dark room problem”, and examines Clark’s response to the problem. Finally, I recommend that if supplemented with an account of action selection, Clark’s account will avoid the dark room problem. (shrink)

This paper aims to offer an account of affective experiences within Predictive Processing, a novel framework that considers the brain to be a dynamical, hierarchical, Bayesian hypothesis-testing mechanism. We begin by outlining a set of common features of affective experiences that a PP-theory should aim to explain: feelings are conscious, they have valence, they motivate behaviour, and they are intentional states with particular and formal objects. We then review existing theories of affective experiences within Predictive Processing and delineate two families (...) of theories: Interoceptive Inference Theories and Error Dynamics Theories. We highlight the strengths and shortcomings of each family of theories and develop a synthesis: the Affective Inference Theory. Affective Inference Theory claims that valence corresponds to the expected rate of prediction error reduction. In turn, the particular object of a feeling is the object predicted to be the most likely cause of expected changes in prediction error rate, and the formal object of a feeling is a predictive model of the expected changes in prediction error rate caused by a given particular object. Finally, our theory shows how affective experiences bias action selection, directing the organism towards allostasis and towards optimal levels of uncertainty in order to minimise prediction error over time. (shrink)

Predictive processing theories are increasingly popular in philosophy of mind; such process theories often gain support from the Free Energy Principle —a normative principle for adaptive self-organized systems. Yet there is a current and much discussed debate about conflicting philosophical interpretations of FEP, e.g., representational versus non-representational. Here we argue that these different interpretations depend on implicit assumptions about what qualifies as representational. We deploy the Free Energy Principle instrumentally to distinguish four main notions of representation, which focus on organizational, (...) structural, content-related and functional aspects, respectively. The various ways that these different aspects matter in arriving at representational or non-representational interpretations of the Free Energy Principle are discussed. We also discuss how the Free Energy Principle may be seen as a unified view where terms that traditionally belong to different ontologies—e.g., notions of model and expectation versus notions of autopoiesis and synchronization—can be harmonized. However, rather than attempting to settle the representationalist versus non-representationalist debate and reveal something about what representations are simpliciter, this paper demonstrates how the Free Energy Principle may be used to reveal something about those partaking in the debate; namely, what our hidden assumptions about what representations are—assumptions that act as sometimes antithetical starting points in this persistent philosophical debate. (shrink)

In this paper I propose an account of self-knowledge based on a framework of predictive processing. Predictive processing understands the brain as a prediction-action machine that tries to minimize error in its predictions about the world. For this view to evolve into a complete account of human cognition we ought to provide an idea how it can account for self-knowledge – knowledge of one’s own mental states. I provide an attempt for such an account starting from remarks on introspection made (...) by Hohwy. I develop Hohwy’s picture into a general model for knowledge of one’s mental states, discussing how predictions about oneself can be used to capture self-knowledge. I further explore empirical predictions, and thereby argue that the model provides a good explanation for failure of self-knowledge in cases involving motor aftereffects, such as the broken escalator phenomenon. I conclude that the proposed account is incomplete, but provides a valuable first step to connect research on predictive processing with the epistemology of self-knowledge. (shrink)

Ordinary objects are experienced to endure over space and time, to not be collocated with each other, to be composed of proper parts, and to survive the loss of some of their parts. These qualities are on the one hand difficult to reconcile for theorists of perception and on the other hand pose a variety of problems when considered in isolation. Relying on the theoretical framework of predictive processing, this paper argues that we can use the category of a robust (...) predictive process to conceptualize qualities such as persistence and compositionality in a unified manner. Traditional problems concerning the structural properties of ordinary objects, such as the question of when two objects compose, can then be reformulated using this new category. (shrink)

According to the standard view, a belief is based on a reason and doxastically justified—i.e., permissibly held—only if a causal relation obtains between a reason and the belief. In this paper, I argue that a belief can be doxastically justified by a reason’s mere disposition to sustain it. Such a disposition, however, wouldn’t establish a causal connection unless it were manifested. My argument is that, in the cases I have in mind, the manifestation of this disposition would add no positive (...) epistemic feature to the belief: a belief that is justified after the manifestation of a reason’s causal powers must have already been justified before their manifestation. As a result, those who adhere to the standard causal view of the basing relation face a hard choice: they should either abandon the enormously popular view that doxastic justification has a basing requirement or modify their view of the basing relation. (shrink)

Predictive processing framework has found wide applications in cognitive science and philosophy. It is an attractive candidate for a unified account of the mind in which perception, action, and cognition fit together in a single model. However, PP cannot claim this role if it fails to accommodate an essential part of cognition—conceptual thought. Recently, Williams argued that PP struggles to address at least two of thought’s core properties—generality and rich compositionality. In this paper, I show that neither necessarily presents a (...) problem for PP. In particular, I argue that because we do not have access to cognitive processes but only to their conscious manifestations, compositionality may be a manifest property of thought, rather than a feature of the thinking process, and result from the interplay of thinking and language. Pace Williams, both of these capacities, constituting parts of a complex and multifarious cognitive system, may be fully based on the architectural principles of PP. Under the assumption that language presents a subsystem separate from conceptual thought, I sketch out one possible way for PP to accommodate both generality and rich compositionality. (shrink)

The general principles of perceptuo-motor processing and memory give rise to theNow-or-Never bottleneckconstraint imposed on the organization of the language processing system. In particular, the Now-or-Never bottleneck demands an appropriate structure of linguistic input and rapid incorporation of both linguistic and multisensory contextual information in a progressive, integrative manner. I argue that the emerging predictive processing framework is well suited for the task of providing a comprehensive account of language processing under the Now-or-Never constraint. Moreover, this framework presents a stronger (...) alternative to theChunk-and-Passaccount proposed by Christiansen and Chater (2016), as it better accommodates the available evidence concerning the role of context (in both the narrow and wider senses) in language comprehension at various levels of linguistic representation. Furthermore, the predictive processing approach allows for treating language as a special case of domain-general processing strategies, suggesting deep parallels with other cognitive processes such as vision. (shrink)

In this paper we argue that predictive processing (PP) theory cannot account for the phenomenon of affect-biased attention prioritized attention to stimuli that are affectively salient because of their associations with reward or punishment. Specifically, the PP hypothesis that selective attention can be analyzed in terms of the optimization of precision expectations cannot accommodate affect-biased attention; affectively salient stimuli can capture our attention even when precision expectations are low. We review the prospects of three recent attempts to accommodate affect with (...) tools internal to PP theory: Miller and Clark’s (2018) embodied inference; Seth’s (2013) interoceptive inference; and Joffily and Coricelli’s (2013) rate of change of free energy. In each case we argue that the account does not resolve the challenge from affect-biased attention. For this reason, we conclude that prediction error minimization is not sufficient to explain all mental phenomena, contrary to the claim that the PP framework provides a unified theory of all mental phenomena or the brain ‘s cognitive functioning. Nevertheless, we suggest that empirical investigation of the interaction between affective salience and precision expectations should prove helpful in understanding the limits of PP theory, and may provide new directions for the application of a Bayesian perspective to perception. (shrink)

We present a multiscale integrationist interpretation of the boundaries of cognitive systems, using the Markov blanket formalism of the variational free energy principle. This interpretation is intended as a corrective for the philosophical debate over internalist and externalist interpretations of cognitive boundaries; we stake out a compromise position. We first survey key principles of new radical views of cognition. We then describe an internalist interpretation premised on the Markov blanket formalism. Having reviewed these accounts, we develop our positive multiscale account. (...) We argue that the statistical seclusion of internal from external states of the system—entailed by the existence of a Markov boundary—can coexist happily with the multiscale integration of the system through its dynamics. Our approach does not privilege any given boundary, nor does it argue that all boundaries are equally prescient. We argue that the relevant boundaries of cognition depend on the level being characterised and the explanatory interests that guide investigation. We approach the issue of how and where to draw the boundaries of cognitive systems through a multiscale ontology of cognitive systems, which offers a multidisciplinary research heuristic for cognitive science. (shrink)

Sympathetic magic features strongly in virtually all religious traditions and in folk customs generally. Scholars agree that It is based on the association of ideas perceived as external, mind-independent causal realities, as connections mediating causal influence. Moreover, religious folk believe that this mediation involves forms of supernatural agency. From a psychological perspective, the key question revolves around the principles by which the cognitive system deems some of its content to reference the external world and other content to constitute internal mental (...) forms of activity like thoughts, feelings and attitudes. The paper proposes that the critical factor has to do with the balance between two distinctly different kinds of cognitive content: representations of things (mentation arising in the form of something other than itself), as distinct from registrations referencing the intrinsic phenomenal properties of the mental state itself. The balance between these two determines whether content is perceived as external worldly event or a form of internal mental content. (shrink)

Philosophers and psychologists assume that mental imagery is a cognitive state, that it represents things as being a certain way. However, I argue that imagery is a conative state: it represents things as to be made a certain way. I challenge the traditional assumption by targeting an increasingly popular cognitive account that identifies mental imagery, such as inner speech, with predictions of sensory input. This predictive account faces both empirical and theoretical problems. The account not only fails to capture the (...) salience effects exhibited by mental imagery, it also cannot explain why subjects should not spend their lives minimizing prediction error by engaging in fantasy. These shortcomings highlight what I call the conative profile of imagery, which includes its salience, motivational, and performance effects. I argue that an alternative, conative view of imagery, according to which mental images are goal states, is best able to account for its conative profile. (shrink)

This paper questions two prima facie plausible claims concerning switching in the presence of ambiguous figures. The first is the claim that reversing is an instantaneous process. The second is the claim that the ability to reverse demonstrates the interpretive, inferential and constructive nature of visual processing. Empirical studies show that optical and cerebral events related to switching protract in time in a way that clashes with its perceived instantaneity. The studies further suggest an alternative theory of reversing: according to (...) such alternative, seeing the same thing in multiple ways is a matter of uncovering what is already present to the senses through visual search. (shrink)

This article attempts to clarify the commitments of a predictive coding approach to perception. After summarizing predictive coding theory, the article addresses two questions. Is a predictive coding perceptual system also a Bayesian system? Is it a Kantian system? The article shows that the answer to these questions is negative.

Standard models of visual perception hold that vision is an inferential or interpretative process. Such models are said to be superior to competing, non-inferential views in explanatory power. In particular, they are said to be capable of explaining a number of otherwise mysterious, visual phenomena such as multi-stable perception. Multi-stable perception paradigmatically occurs in the presence of ambiguous figures, single images that can give rise to two or more distinct percepts. Different interpretations are said to produce the different percepts. In (...) this paper, I argue that a non-inferential account of visual perception is just as capable of explaining multi-stable perception. I propose an embedded understanding of vision, and show how the embedded account can, after properly qualifying them, use the explanatory resources of the inferential view to explain just what such a view explains. (shrink)

Ambiguous figures pose a problem for representationalists, particularly for representationalists who believe that the content of perceptual experience is non-conceptual (MacPherson in Nous 40(1):82–117, 2006). This is because, in viewing ambiguous figures, subjects have perceptual experiences that differ in phenomenal properties without differing in non-conceptual content. In this paper, I argue that ambiguous figures pose no problem for non-conceptual representationalists. I argue that aspect shifts do not presuppose or require the possession of sophisticated conceptual resources and that, although viewing ambiguous (...) figures often causes a change in phenomenal properties, this change is accompanied by a change in non-conceptual content. I illustrate the case by considering specific examples. (shrink)

There is a view on consciousness that has strong intuitive appeal and empirical support: the intermediate-level theory of consciousness, proposed mainly by Ray Jackendoff and by Jesse Prinz. This theory identifies a specific “intermediate” level of representation as the basis of human phenomenal consciousness, which sits between high-level non-perspectival thought processes and low-level disjointed feature-detection processes in the perceptual and cognitive processing hierarchy. In this article, we show that the claim that consciousness arises at an intermediate-level is true of some (...) cognitive systems, but only in virtue of specific constraints on their active interactions with the environment. We provide ecological reasons for why certain processing levels in a cognitive hierarchy are privileged with respect to consciousness. We do this from the perspective of a prediction-error minimization model of perception and cognition, relying especially on the notion of active inference: the privileged level for consciousness depends on the specific dispositions of an organism concerned with inferring its policies for action. Such a level is indeed intermediate for humans, but this depends on the spatiotemporal resolution of the typical actions that a human organism can normally perform. Thus, intermediateness is not an essential feature of consciousness. In organisms with different action dispositions the privileged level or levels may differ as well. (shrink)

In this article, we show how the phenomenon of self-deception when adequately analyzed, can be incorporated into a predictive processing framework. We describe four strategies by which a subject may become self-deceived to account for typical cases of self-deception. We then argue that the four strategies can be modeled within this framework, under the assumption that a satisfying account of motivation is possible within predictive processing. Finally, we outline how we can ground this assumption by discussing how such a systematic (...) model of motivation may be developed using the distinction between a world-model and a self-model. (shrink)

If beliefs and desires affect perception—at least in certain specified ways—then cognitive penetration occurs. Whether it occurs is a matter of controversy. Recently, some proponents of the predictive coding account of perception have claimed that the account entails that cognitive penetrations occurs. I argue that the relationship between the predictive coding account and cognitive penetration is dependent on both the specific form of the predictive coding account and the specific form of cognitive penetration. In so doing, I spell out different (...) forms of each and the relationship that holds between them. Thus, mere acceptance of the predictive coding approach to perception does not determine whether one should think that cognitive penetration exists. Moreover, given that there are such different conceptions of both predictive coding and cognitive penetration, researchers should cease talking of either without making clear which form they refer to, if they aspire to make true generalisations. (shrink)