Testable or bust: theoretical lessons for predictive processing

Type Journal Article
Author Marcin Miłkowski
Author Piotr Litwin
URL https://link.springer.com/10.1007/s11229-022-03891-9
Volume 200
Issue 6
Pages 462
Publication Synthese
ISSN 1573-0964
Date 2022-11-02
Journal Abbr Synthese
DOI 10.1007/s11229-022-03891-9
Accessed 2022-11-04 12:05:41
Library Catalog DOI.org (Crossref)
Language en
Abstract Abstract

The predictive processing (PP) account of action, cognition, and perception is one of the most influential approaches to unifying research in cognitive science. However, its promises of grand unification will remain unfulfilled unless the account becomes theoretically robust. In this paper, we focus on empirical commitments of PP, since they are necessary both for its theoretical status to be established and for explanations of individual phenomena to be falsifiable. First, we argue that PP is a varied research tradition, which may employ various kinds of scientific representations (from theories to frameworks and toolboxes), differing in the scope of empirical commitments they entail. Two major perspectives on PP
cognitive theory may then be distinguished: generalized vs. hierarchical. The first one fails to provide empirical detail, and the latter constrains possible physical implementations. However, we show that even hierarchical PP is insufficiently restrictive to disallow incorrect models and may be adjusted to explain any neurocognitive phenomenon–including non-existent or impossible ones–through flexible adjustments. This renders PP a universal modeling tool with an unrestricted number of degrees of freedom. Therefore, in contrast with declarations of its proponents, it should not be understood as a unifying theoretical perspective, but as a computational framework, possibly informing further theory development in cognitive science.

Short Title Testable or bust

Source: Publications

Misidentification delusions as mentalization disorders

Type Journal Article
Author Adrianna Smurzyńska
URL https://doi.org/10.1007/s11097-022-09820-y
Publication Phenomenology and the Cognitive Sciences
ISSN 1572-8676
Date 2022-05-25
Journal Abbr Phenom Cogn Sci
DOI 10.1007/s11097-022-09820-y
Accessed 2022-05-26 07:58:37
Library Catalog Springer Link
Language en
Abstract The aim of this article is to analyze those theories that interpret misidentification delusions in terms of mentalization. The hypothesis under examination holds that a mentalization framework is useful for describing misidentification delusions when identification is thought to be partially based on mentalization. The article provides both a characterization and possible interpretations of such delusions, and possible relations between misidentification and mentalization are scrutinized. Whether the mentalization approach may explain or describe such kinds of mental disorders is considered, with the conclusion that while this approach is unsatisfactory, there is some room for future improvement.

Source: Publications

Piotr Kozak – Thinking with Images

The next meeting of the seminar Philosophy of Cognitive Science is planned for May, 12th, at 10:30 (AM Warsaw, CET). Our guest will be Piotr Kozak(IPS PAS). We will discuss a draft of the Introduction and Chapter 1 from his forthcoming book: Thinking with Images: Imagistic Cognition and Non-propositional Content. For the more interested we also attach Chapter 6.

From Introduction:
 The main research question of this book is: What is thinking with images? The question is analogical in its form to such questions as ‘what is thinking with language?’ It means that if we can ask whether we can think in or with language, then we can ask whether we can think in or with images.

The question follows from a commonsensical observation: when someone is asked how many windows are in his or her flat, he or she will probably form and inspect a mental image of the flat and count the windows. If an architect designs a house, then he or she designs the house with the help of drawings. If one tries to get from point to point B, one may use a map.

The examples above are instantiations of what can be called ‘thinking with images’. However, listing examples of imagistic thinking is a relatively easy task. The difficult task is to say exactly what thinking with images is.

Imagistic thinking is understood here in three ways: as a faculty, an act, and a mental state or an event. The faculty of imagistic thinking refers to the capacity of using images in thinking. The act of imagistic thinking is exercising this faculty. The mental state called ‘imagistic thought’ is a product of such an act.

.pf-button.pf-button-excerpt { display: none; }
Source: Cognitive Science in Search of Unity

Without more theory, psychology will be a headless rider

Type Journal Article
Author Witold M. Hensel
Author Marcin Miłkowski
Author Przemysław Nowakowski
URL https://www.cambridge.org/core/product/identifier/S0140525X21000212/type/journal_article
Volume 45
Pages e20
Publication Behavioral and Brain Sciences
ISSN 0140-525X, 1469-1825
Date 2022
Journal Abbr Behav Brain Sci
DOI 10.1017/S0140525X21000212
Accessed 2022-02-10 08:48:42
Library Catalog DOI.org (Crossref)
Language en
Abstract Abstract
We argue that Yarkoni's proposed solutions to the generalizability crisis are half-measures because he does not recognize that the crisis arises from investigators' underappreciation of the roles of theory in experimental research. Rather than embracing qualitative analysis, the research community should make an effort to develop better theories and work toward consistently incorporating theoretical results into experimental practice.

Source: Publications

Turing’s conceptual engineering

The next meeting of the seminar is planned for February, 17th, at 10:30 (AM Warsaw, CET). Our guest will be Marcin Miłkowski (IPS PAS). We will discuss a draft paper: Turing’s conceptual engineering.

Abstract: Alan Turing’s influence on subsequent research in artificial intelligence is undeniable. His proposed test for intelligence remains influential. In this paper, Turing’s conception of how to understand intelligence is analyzed as an instance of conceptual engineering that rejects the role of the previous linguistic usage but appeals to intuition pumps instead. Even though many conceive of his proposal as a prime case of operationalism or behaviorism, it is more plausibly viewed as a stepping stone toward a future theoretical construal of intelligence. In addition, his own conceptual network is analyzed through the lens of distributional semantics over the corpus of his written work. As it turns out, Turing’s conceptual engineering of the notion of “intelligence” is indeed quite similar to providing a precising definition with the aim of revising the usage of the concept.

Get a Google Meet link by emailing pnowakowski@ifispan.edu.pl

Turing’s conceptual engineering

.pf-button.pf-button-excerpt { display: none; }
Source: Cognitive Science in Search of Unity

Incorporating free energy models into mechanisms: the case of predictive processing under the free energy principle

The next meeting of the seminar is planned for January, 20th, at 11:00 (AM Warsaw, CET). Our guest will be Michał Piekarski (UKSW Warsaw). We will discuss a draft paper: Incorporating free energy models into mechanisms: the case of predictive processing under the free energy principle.

Abstract: There is a view emerging in the philosophy of science that research practices in science can be characterized in terms of discovering and describing mechanisms. Mechanistic explanations are based on the identifying the underlying mechanisms that generate a target phenomenon and strategies understood as decomposition of these mechanisms. Recently, there has been a discussion among mechanists about the necessity to include constraints and free energy flows into the explanations, as constitutive components of mechanistic explanations. This is directly related to the existence of control mechanisms that are non-autonomous and entail the existence of heterarchical networks. I refer to this as the ‘constrained mechanisms approach’. This paper examines the extent to which this approach can be applied to the predictive processing framework, which is now an influential process theory, offering a computational description of perceptual and cognitive mechanisms in terms of hierarchical generative models approximating Bayesian inference. In other words, I examine whether the constrained mechanisms approach can be applied to the framework in which control mechanisms play an important explanatory role. I will argue that predictive processing models based on the free energy principle are amenable to this approach. In practice, this means that free energy principle offers a normative explanatory framework for predictive processing, and that in turn, this framework offers a biologically plausible account of the manner in which the principle is implemented in terms of hierarchical generative models and heterarchical active mechanisms. These analyzes are of great importance for those approaches that undermine the explanatory status of the free energy principle.


Zgłoszenia należy wysłać na mail: pnowakowski@ifispan.waw.pl

.pf-button.pf-button-excerpt { display: none; }
Source: Cognitive Science in Search of Unity

Buscicchi – In the Mood for Hedonic Phenomenalism

The next meeting of the seminar „Philosophy of Cognitive Science” is planned for December, 2nd, at 10:00 (AM Warsaw, CET). Our guest will be Lorenzo Buscicchi (University of Waikato). We will discuss a draft paper: In the Mood for Hedonic Phenomenalism.

Abstract: The present paper addresses the debate between hedonic phenomenalism and hedonic intentionalism, the two main theories of pleasure. More specifically, this paper advances an objection to hedonic intentionalism, the cheerfulness objection. Following it, the mood of cheerfulness lacks intentionality. If it is not the case that every pleasure is intentional, it follows that we cannot group pleasures together through a shared attitude. The globalist strategy—the most prominent move advanced to save the intentionality of moods—is rejected. This strategy seems phenomenologically implausible: it simply does not seem to respect our experience of cheerfulness. Thus, cheerfulness ends up being a damaging counter-example to hedonic intentionalism. Given that, I advance that hedonic phenomenalism is still the most plausible account of pleasure.

To receive the Zoom link, please contact Dr. Przemysław Nowakowski (prrono@wp.pl)

.pf-button.pf-button-excerpt { display: none; }
Source: Cognitive Science in Search of Unity

Special Issues of Theory & Psychology and Synthese

Our research project “Cognitive science in search of unity” led to the publication of two special issues of peer-reviewed journals. We are happy to announce that recently we fulfilled this goal. The special issues are ready. First of them, entitled “Mechanisms in psychology: The road towards unity?” is out in Theory & Psychology, while the second, called “Explanations in cognitive science: Unification vs pluralism”, was just published in Synthese.

The focus of the special issue of Theory & Psychology, edited by Marcin Miłkowski, Mateusz Hohol and Przemysław Nowakowski (2019), is on explanatory mechanisms in psychology, especially on problems of particular prominence for psychological science such as theoretical integration and unification. Proponents of the framework of mechanistic explanation claim, in short, that satisfactory explanations in psychology and related fields are causal. They stress the importance of explaining phenomena by describing mechanisms that are responsible for them, in particular by elucidating how the organization of component parts and operations in mechanisms gives rise to phenomena in certain conditions. The purpose of this special issue, broadly construed, was to solicit original papers from defenders and opponents of mechanistic explanation and theorists of psychology and neuroscience who address problems of special prominence for the psychological community.

The special issue of Theory and Psychology opens with an introductory paper where we (i.e., Marcin Miłkowski, Mateusz Hohol and Przemysław Nowakowski) introduce the concept of mechanism and explore what psychology gains from mechanistic explanation. In the following paper, Eric Hochstein (University of Victoria) provocatively claims that an experimenter interested in cognitive mechanisms should be a good metaphysician. His contribution is entitled “How metaphysical commitments shape the study of psychological mechanisms”. In the next paper, entitled “Phenomenology and mechanisms of consciousness: Considering the theoretical integration of phenomenology with a mechanistic framework” Marek Pokropski (University of Warsaw) claims that phenomenological analysis can supply descriptions of phenomena that are explained mechanistically analogically to traditional functional analysis. William Bechtel (University of California San Diego) is an author of the following paper called “Resituating cognitive mechanisms within heterarchical networks controlling physiology and behavior”. The author proposes that a primary function of cognition is not building highly adequate representations of the surrounding world on the basis of sensory input, but providing the organism with behavioral control. In the next paper, entitled “Model-based cognitive neuroscience: Multifield mechanistic integration in practice” Mark Povich (Washington University, St. Louis) points to a framework that integrates two levels of cognition, namely, the computational/algorithmic (traditionally accounted by cognitive psychology) and the implementational one (primarily investigated by neuroscience). Next, Paweł Gładziejewski (Nicolaus Copernicus University) explores the issue of explanatory unification of cognition under the free energy principle (FEP). In a paper “Mechanistic unity of the predictive mind”, the author claims that the FEP delivers only (contrary to some proponents of this perspective) a functional sketch or schema, which may be implemented by many distinct neural mechanisms. In the subsequent contribution, Sabrina Golonka and Andrew D. Wilson (Leeds Beckett University) deal with “Ecological mechanisms in cognitive science”. Their main message is that a neo-Gibsonian framework of action and perception can be reconciled with mechanistic analyses. The two papers that conclude the issue are more skeptical about the prospects of mechanistic explanation of the mental. In the first, Matteo Colombo (Tilburg University) and Andreas Heinz (Charité – Universitätsmedizin Berlin) explore conceptual problems of psychiatry. In their paper “Explanatory integration, computational phenotypes, and dimensional psychiatry: The case of alcohol use disorder”, the authors propose elucidating mental maladies by taking into account clinically relevant properties of a computational phenotype, such as the tension between model-based and model-free control. Finally, Lawrence Shapiro (University of Wisconsin–Madison)  is sketching “A tale of two explanatory styles in cognitive psychology”. He claims that although the new mechanism delivers a vital strategy of explanation in psychology, it does not mean that traditional functional analysis is redundant and thus should be completely rejected in the field. We believe that the above summarized papers provide a crucial update to the theory of mechanistic organization and unification, a number of new applications and extensions, and critical views of mechanistic explanation. 

The special issue of Synthese (2021), edited by Marcin Miłkowski and Mateusz Hohol, is a contribution to the debate between the defenders of explanatory unification and explanatory pluralism, that has been ongoing from the beginning of cognitive science and is one of the central themes of its philosophy. The debate is focused on the following questions: Does cognitive science need a grand unifying theory? Should explanatory pluralism be embraced instead? Or maybe local integrative efforts are needed? What are the advantages of explanatory unification as compared to the benefits of explanatory pluralism? 

The special issue of Synthese opens with an introductory paper, where we (i.e., Marcin Miłkowski and Mateusz Hohol) discuss the background of the above questions, distinguishing integrative theorizing from building unified theories. We also show that unification in contemporary cognitive science goes beyond reductive unity, and may involve various forms of joint efforts and division of explanatory labor. The following two papers explore computational and mechanistic modes of explanation in cognitive (neuro)science. In a paper entitled “The methodological role of mechanistic-computational models in cognitive science”, Jens Harbecke (Witten/Herdecke University, Witten, Germany) claims that a satisfactory explanatory model of a cognitive architecture should integrate phenomena at all the levels (in the mechanistic sense), accounting simultaneously for computational processes involved in the relevant component parts. Then, Lotem Elber-Dorozko and Oron Shagrir (The Hebrew University of Jerusalem) draw from the reinforcement learning phenomenon to explore the problem of “Integrating computation into the mechanistic hierarchy in the cognitive and neural sciences”. The next paper is entitled “Representational unification in cognitive science: Is embodied cognition a unifying perspective?”. Its authors, Marcin Miłkowski and Przemysław Nowakowski (IPS PAS), claim that even if embodied cognition fails as a proposal of the grand unification of cognitive science, it shows that unification constitutes a notable virtue of research traditions in the Laudan’s sense. Next three papers explore the dynamical approach to cognitive processes. In the first of them, called “Appraisal of certain methodologies in cognitive science based on Lakatos’s methodology of scientific research programmes”, Haydar Oğuz Erdin (Bogazici University) reveals shortcomings in Chemero’s non-representationalism. To counterbalance, the next paper, “Resonance and radical embodiment” by Vicente Raja (University of Antwerpen), is enthusiastic about the dynamical explanation. Raja claims that the unification power of the resonance-based framework is evidenced by bridging theories of behavioral dynamics and neural reuse. Next paper, called “Make up your mind: octopus cognition and hybrid explanations” by Sidney Carls-Diamante (Konrad Lorenz Institute for Evolution and Cognition Research), shows that some cognitive phenomena require the use of distinct explanatory accounts, wherein some parts/operations within a mechanism are described dynamically, while others are described in a representational way. As editors, we are proud that this paper received the Werner Callebaut Prize from the International Society for the History, Philosophy and Social Studies of Biology. The penultimate paper investigates innateness, that is one of the unifying general principles in a number of cognitive explanations. More precisely, J. Brendan Ritchie (KU Leuven) undertakes a defense of the frequently challenged notion of nativism, which identifies inborn cognitive skills with “not learned” ones. The paper is entitled “What’s wrong with the minimal conception of innateness in cognitive science?”. The final contribution, “Linguistics and the explanatory economy” by Gabe Dupre (University of California Los Angeles), focuses on generative linguistics. The author shows that in cognitive science, some theories could be considered complementary, and that they should not be overgeneralized. 

We hope that these papers contribute to a deeper understanding of the importance of unification and pluralism in cognitive science. The issue of whether unification is to be preferred over integration is far from settled. Nonetheless, as papers in this special issue also attest, unification does not boil down to reductive explanation. At the same time, the way unification or coordination is approached depends, obviously, on the general framework that one adopts in understanding satisfactory explanations.

.pf-button.pf-button-excerpt { display: none; }
Source: Cognitive Science in Search of Unity

Azərbaycanın ən yaxşı onlayn kazinosu sizə ən böyük ilk depozit bonusunu mostbet azerbaycan təklif edir.