Representation Crisis: Commentary on “Investigating the concept of representation in the neural and psychological sciences”

Inês Hipólito

Macquarie University, Philosophy Department

In cognitive science, the notion of mental representation, stands as a central beacon, not due to unanimous agreement, but rather because it serves as the conceptual nexus through which diverse theoretical trajectories can be mapped out in their understanding of cognitive behaviour.  Cognitive psychology, in its mid-20th century rendezvous with computer science, moved away from philosophical moorings (O’Donohue and Kitchener, 1996; Slote, 2020). It became mainstream in psychology sciences by adopting, to a large extent, the computational toolkit of its era of information processing, inputs and outputs, and algorithms, coined then “cognitivism” (Currie, 1999). Much of the historical trajectory of cognitive psychology and the discipline itself owes its intellectual debt to the analogy between human behavior and human-engineered computational artifacts. This analogy has been further fortified by the discipline of philosophy of mind (Colombo and Sprevak, 2019). Be it the Turing Machine, which inspired the Modularity of the Mind framework (Fodor, 1983; Carruthers, 2005; Zeimbekis and Raftopoulos, 2015), or Parallel Distributed Processing which gave rise to Connectionism and Neural Networks (McClelland, 1988), or Bayesian Statistics underpinning predictive coding (Pezzulo, Parr and Friston, 2022) within the philosophical framework of predictive processing (Clark, 2015; Hohwy, 2020)—all represent distinct approaches endeavoring to elucidate the emergence of a mental representation from the “computational” mind, hence not necessarily incompatible with each other (Hipólito and Kirchhoff, 2023).

While computational simulation models of the technological kind described above are invaluable for scientific epistemic gain, researchers assuming them as more than a looking glass to formulate representations of the natural world, ardently contend that the natural world is a world of computation. Within this realism of computation (of any variety), the mind metamorphosed into information processing modules or Bayesian entities. This reasoning is licensed by the assumption that computation exists “in the wild” beyond human practices. What unites all these theories is the ambition to furnish an account of how the “computational” mind begets a mental representation (Smortchkova, Dołrega and Schlicht, 2020).  Yet, as Wittgenstein eloquently admonishes, this endeavor is akin to expecting an eye to scrutinize itself. That the means by which we look at the world – computational tools – is the world itself. In our human endeavors, we forge instruments to compile edifices of knowledge, some more robust in their certitude than others. This undertaking involves the generation of theories through representational cogitation, ranging in degrees of reliability, as well as the cultivation of tools such as computational simulation models to gauge the plausability of our ruminations (Guest and Love, 2019; van Rooij, 2022; Guest and Martin, 2021; Guest and Martin, 2023).

Such computational simulations find application across diverse domains grappling with complex systems, from astrophysics to neuroscience, in the scientific practices to proffer scientific representations of the natural world (Juarrero, 2023).  The contemporary theories nestled under the aegis of Embodied Cognitive Science (ECogSci) (Newen, Bruin and Gallagher, 2018) — Embodied, Embedded, Enactive, Ecological — cast aside the notion of computation as an autonomous entity existing apart from human enculturated practices. The collective stance of theories within ECogSci rebuffs the premise that cognitive behavior can be exhaustively expounded through mental representation as a consequence of sole neural processes. ECogSci also rejects that neural processes are computational in kind.  Instead, ECogSci aligns more closely with mathematical dynamical and complex systems theories (Favela, 2020). Cognitive life, beyond neural processes, concerns the full agent enculturation. Throughout one’s lifespan, individuals develop mental representations, which are essentially semantic constructs. This process is heavily influenced by the socio-cultural context in which it occurs (Hutto, 2005; Hutto et al., 2020; Miyahara, 2021; Rolla and Figueiredo, 2021; Hipólito and van Es, 2022). These representations play a significant role in shaping mental health (Maiese, 2023; Nielsen, 2023).

Favela and Machery’s (2023) (F&M) examination of the concept of “representation” in cognitive science presents a significant stride towards rectifying a long-standing gap in the field’s understanding. One commendable aspect of F&M’s study is its inclusive approach, involving a diverse cohort of participants hailing from psychology, neuroscience, and philosophy. This interdisciplinary perspective affords a comprehensive panorama of how the notion of representation is perceived across these distinct but interrelated fields. This robust experimental design effectively prompts participants to articulate their comprehension of “representation” through practical scenarios. The data collected bears the hallmarks of reliability and validity, bolstering the credibility of the findings. It hence presents a compelling invitation to explore alternative frameworks and methodologies in cognitive science signaling a shift away from cognitivism.  Taken seriously, it could precipitate a reevaluation of prevailing cognitivist theoretical models, paving the way for more nuanced and accurate conceptualizations.

F&M’s study illuminates a pivotal aspect: the palpable uncertainty among researchers about which manifestations of brain activity entail representations. While those entrenched in theoretical musings might find themselves delving into intricate contemplations of behavior viewed through a computational lens, their counterparts in applied neuroscience exercise caution when delineating the potential forms of mental representation intertwined with neural activity. This hesitancy reverberates as an empirical affirmation of the collective tenets upheld by ECogSci theories: mental representation is an agent’s enculturated skill.

What are the causes of the representation crisis?

What sustains the persistent belief in the field regarding the search for ’representations’ within the brain? One plausible explanation lies in the historical trajectory of cognitive psychology. The symbiotic relationship it forged with computer science in the 1950s, coupled with its subsequent detachment from philosophy came at that cost. This historical momentum, coupled with the inherent scientific inclination towards seeking reductionist explanations, may contribute to the field’s continued adherence to the representation-centric narrative. Additionally, the allure of precision and tangibility that comes with viewing cognition through the lens of conceivably “tractable” mental representations may have further entrenched this perspective, despite the lack of empirical substantiation and neuroscientists’s cautious attitude (F&M) (Hipólito, 2022). The combination of historical precedent, epistemic preferences, and the allure of reductionism from mental representation to ‘neural representation’ is instrumentalized within the academic ecosystem, exerting substantial influence on various aspects of scientific research:

Researcher Incentives: Researchers may prioritize mental representation due to its historical significance, anticipating recognition and influence.

Funding Influences: Funding bodies may favor studies on mental representation, aligning with established theories and historical prominence.

Journal Priorities: Journals may lean towards publishing studies challenging established theories, potentially prioritizing mental representation.

Peer Review Process: Studies invoking mental representation as explanatory assumptions may face less scrutiny, as reviewers may be more familiar with this framework.

F&M’s study underscores the necessity for a meticulous reevaluation of cognitivist models, potentially culminating in a substantive revision or even a reconsideration of the underlying theoretical framework. Neglecting the significance of F&M’s research constitutes a regression, as it – against scientific evidence for mental representation – perpetuates historical trends, aligns with epistemic biases, and succumbs to reductionism, all while inhibiting critical inquiry towards a Representation Crisis.

The Representation crisis – a new contender in the ring of challenges – joins the ranks of cognitive psychology’s recent tribulations, including the WEIRD problem, which highlights the over-reliance on Western, Educated, Industrialized, Rich, and Democratic populations for research, generalising textbook psychology on the basis of 12% WEIRD population (Henrich, 2020; Fuentes, 2022). In addition to this, the Replication Crisis shows that once awe-inspiring cognitive psychology textbook studies are now struggling to perform the same magic twice (Wiggins and Christopherson, 2019). Finally, the widely recognized measurement challenges also highlight the questionable measurement practices (Flake and Fried, 2020). As Meier (2023) aptly points out, “Historical references provide evidence that psychology’s measurement problems are not random occurrences, but periodic problems that appear, fade away without resolution, and are later rediscovered” (p. 1). Cognitivism has found itself stumbling under the weight of its own assumptions – including the notion of representation.

The convergence of these crises prompts contemplation on how many crises must converge to precipitate a paradigm shift in cognitive science, from cognitivism towards one that is inclusive of cultural value and diversity. This revelation, that minds might not be akin to rigid, preprogrammed computers, holds the promise of an exhilarating leap forward in comprehending the kaleidoscope of cognitive diversity across cultures. With the shackles of one-size-fits-all models and broad generalizations cast aside, we stand on the precipice of not only conquering the replication crisis but also earnestly addressing the clarion call for neurodiversity practices in research and in mental health (Chapman, 2020). The future of cognitive science is one of vibrant potential, where every mind is recognized and celebrated for its unique enculturated tapestry of embodied cognition. Mental representation is but one thread in the rich fabric of cognitive experience.

References

Carruthers, P. (2005). The case for massively modular models of mind. Contemporary debates in cognitive science, ed. R. Stainton, 205-25.

Chapman, R. (2020). Defining neurodiversity for research and practice. Neurodiversity studies: A new critical paradigm, 218-220.

Clark, A. (2015). Radical predictive processing. The Southern Journal of Philosophy, 53, 3-27.

Colombo, M., & Sprevak, M. (2019). Introduction to Handbook. In Introduction to Handbook.  In M. Sprevak & M. Colombo (Eds.) The Routledge Handbook of the Computational Mind.

Currie, G. (1999). Cognitivism. A companion to film theory, 105-122.

Favela, L. H. (2020). Dynamical systems theory in cognitive science and neuroscience. Philosophy Compass, 15(8), e12695.

Flake, J. K., & Fried, E. I. (2020). Measurement schmeasurement: Questionable measurement practices and how to avoid them. Advances in Methods and Practices in Psychological Science, 3(4), 456-465.

Fodor, J. A. (1983). The modularity of mind. MIT press.

Fuentes, A. (2022). WEIRD Indeed, but there is more to the story: anthropological reflections on Henrich’s “The Weirdest people in the world”. Religion, Brain & Behavior, 12(3), 284-290.

Guest, O., & Love, B. C. (2019). Levels of representation in a deep learning model of categorization. BioRxiv, 626374.

Guest, O., & Martin, A. E. (2021). How computational modeling can force theory building in psychological science. Perspectives on Psychological Science, 16 (4), 789-802.

Guest, O., & Martin, A. E. (2023). On logical inference over brains, behaviour, and artificial neural networks. Computational Brain & Behavior, 1-15.

Henrich, J. (2020). The WEIRDest people in the world: How the West became psychologically peculiar and particularly prosperous. Penguin UK.

Hipólito, I. (2022). Cognition without neural representation: dynamics of a complex system. Frontiers in Psychology, 12, 643276.

Hipólito, I., & Kirchhoff, M. (2023). Breaking boundaries: The Bayesian Brain Hypothesis for perception and prediction. Consciousness and Cognition, 111, 103510.

Hipólito, I., & van Es, T. (2022). Enactive-dynamic social cognition and active inference. Frontiers in psychology, 13, 855074.

Hohwy, J. (2020). New directions in predictive processing. Mind & Language, 35(2), 209-223.

Hutto, D. D. (2005). Knowing what? Radical versus conservative enactivism. Phenomenology and the Cognitive Sciences, 4(4), 389-405.

Hutto, D., Gallagher, S., Ilundáin-Agurruza, J., & Hipólito, I. (2020). Culture in Mind. An Enactivist Account. Not Cognitive Penetration but Cultural Permeation. Culture, mind, and brain: Emerging concepts, models, applications, 163-187.

Juarrero, A. (2023). Context changes everything: how constraints create coherence. MIT Press.

Maiese, M. (2023). Are All Mental Disorders Affective Disorders?. Passion: Journal of the European Philosophical Society for the Study of Emotions, 1(1), 31-49.

McClelland, J. L. (1988). Parallel distributed processing: Implications for cognition and development. Depts. of Computer Science and Psychology, Carnegie Mellon University.

Meier, S. T. (2023). Persistence of measurement problems in psychological research. Frontiers in Psychology, 14, 1132185.

Miyahara, K. (2021). Enactive pain and its sociocultural embeddedness. Phenomenology and the Cognitive Sciences, 20(5), 871-886.

Newen, A., De Bruin, L., & Gallagher, S. (Eds.). (2018). The Oxford handbook of 4E cognition. Oxford University Press.

Nielsen, K. (2023). Embodied, Embedded, and Enactive Psychopathology: Reimagining Mental Disorder. Springer Nature.

O’Donohue, W., & Kitchener, R. F. (Eds.). (1996). The philosophy of psychology. Sage.

Pezzulo, G., Parr, T., & Friston, K. (2022). The evolution of brain architectures for predictive coding and active inference. Philosophical Transactions of the Royal Society B, 377(1844), 20200531.

Rolla, G., & Figueiredo, N. (2021). Bringing forth a world, literally. Phenomenology and the Cognitive Sciences, 1-23.

Slote, M. (2020). Between psychology and philosophy: east-west themes and beyond (p. 215). Springer Nature.

Smortchkova, J., Dołrega, K., & Schlicht, T. (Eds.). (2020). What are mental representations?. Oxford University Press.

van Rooij, I. (2022). Psychological models and their distractors. Nature Reviews Psychology, 1(3), 127-128.

Wiggins, B. J., & Christopherson, C. D. (2019). The replication crisis in psychology: An overview for theoretical and philosophical psychology. Journal of Theoretical and Philosophical Psychology, 39(4), 202.

Zeimbekis, J., & Raftopoulos, A. (Eds.). (2015). The cognitive penetrability of perception: New philosophical perspectives. OUP Oxford.