Self, The World, And Cognitive Models
How can we reconcile ideas of consciousness? How can physics and philosophy reconcile in the mysteries of what 'Selfhood' is?
Contemporary debates in cognitive science continue to oscillate between computational metaphors and ecological approaches—each offering partial insights while missing crucial dimensions of lived experience. In this oscillation, we find ourselves trapped within an intellectual pendulum that swings from brain-as-computer to organism-as-detector, neither fully capturing the dynamic, relational nature of consciousness. For the sake of comparison between normative and quantative perspectives, let’s explore how Carlo Rovelli's relational interpretation of quantum mechanics and Alva Noë's enactivist theory of perception might help us transcend this dichotomy.
Rovelli's perspective in quantum physics challenges our fundamental assumptions about reality itself. His relational interpretation suggests that physical properties don't exist in isolation but emerge through interactions between systems. In other words, there are no absolute properties, only relations. This radical view dismantles the notion of an objective, observer-independent reality—a perspective with profound implications for how we conceptualise cognition.
Physics may set itself worlds apart from the domain of neuroscience and psychology, but research demands us to require more interdisciplinary freedom than ever. Computational approaches are vital in deepening our understanding of the universe around us, as complex algorithms from machine learning expose properties in data which the human eye (or even brain) neglects to pick out. Logic persists that non-statistical models may also afford us deeper insight from applying mechanics across a variety of sciences. So, if we apply quantum mechanics - or Rovelli's thinking - to cognitive science, we might ask: What if the mind isn't a computational device processing information about an external world, nor an ecological system detecting pre-existing affordances? What if, instead, cognition itself is fundamentally relational—neither located entirely "in the head" nor wholly determined by environmental structures?
This brings us to Alva Noë's enactive approach to perception. Noë challenges the prevailing assumption that perception involves creating internal representations of an external world. Instead, he argues that perception is a skilled activity—something we do, not something that happens to us. "Perception is not something that happens to us or in us," Noë writes. "It is something we do." Our perceptual experience arises through our active engagement with environments, mediated by our sensorimotor knowledge.
I have a confession: I've long been intrigued by these embodied approaches, yet remained skeptical about abandoning computational frameworks entirely. The mathematical elegance of Bayesian models seems to capture something essential about how we navigate uncertainty. However, Noë's perspective offers a compelling alternative that doesn't necessarily reject computation but situates it within a broader relational context.
Consider visual perception. The traditional computational view treats vision as a process of constructing detailed internal models from retinal input. The ecological approach emphasizes direct perception of environmental affordances. But Noë suggests something different: we perceive by actively exploring our environment through movement. We don't need detailed internal representations because the world itself serves as an "outside memory" that we access through skilled bodily activity.
Worryingly, both computational and ecological approaches inadvertently perpetuate a problematic subject-object dichotomy—either overemphasising internal processing or external determination. This dichotomy collapses under Rovelli's relational perspective, where properties emerge through interaction rather than existing independently. Similarly, Noë's enactive approach dissolves the rigid boundary between perceiver and perceived, emphasizing their co-constitution through dynamic engagement.
The standard metaphors—”brain as computer, perception as detection”—fail because they presuppose a world of discrete, pre-existing objects with fixed properties. But if Rovelli is right, reality itself is relational all the way down. There are no observer-independent facts, only interactions between systems. And if Noë is right, our perceptual experience isn't about representing or detecting a pre-existing world, but about skillfully engaging with a world that reveals itself through our active participation.
This perspective radically reframes how we should approach cognitive science. The question isn't whether brains compute probabilities or detect affordances; it's how organisms and environments co-constitute each other through ongoing, dynamic interaction. Cognition emerges not within neural tissue alone, nor is it determined solely by environmental structures, but exists in the relational space between.
Perhaps more problematically, both computational and ecological approaches struggle to account for the phenomenological dimension of experience—the lived, first-person perspective that remains frustratingly elusive to third-person methodologies. Noë's embodied approach helps bridge this gap by acknowledging that perception isn't something happening "in here" about something "out there," but is itself a mode of active engagement with the world.
Active inference, with its emphasis on organisms as hypothesis-testing agents, moves in the right direction. But even this framework often remains too computationally focused, missing the fundamental insight from Rovelli and Noë: reality itself, including our cognitive processes, is constituted through relations and interactions rather than existing as pre-given entities that interact secondarily.
The circularity in comparing brains to AI systems becomes even more apparent through this lens. We've designed AI systems based on computational metaphors that themselves emerged from particular ways of conceptualizing cognition. These metaphors were never faithful representations of brain function but conceptual frameworks with their own historical and cultural contingencies. Then, in a remarkable feat of circular reasoning, we point to these human-designed systems as evidence for computational theories of mind.
When we apply Noë's insight that perception is a skilled activity to the realm of artificial intelligence, we immediately see what's missing. Language models process vast amounts of text data but lack the embodied, exploratory engagement that characterizes human understanding. They don't perceive or act in any meaningful sense because they exist outside the relational matrix that constitutes conscious experience.
Rovelli's quantum relationalism offers a profound challenge to the metaphysical assumptions underlying both computational and ecological approaches. If properties emerge through interactions rather than existing independently, then the search for the "true nature" of cognition may be fundamentally misguided. There is no absolute fact of the matter about whether brains compute or detect—these are perspectives that become meaningful only in relation to specific experimental contexts and theoretical frameworks.
A more fruitful approach might embrace this relationality, recognizing that cognition isn't reducible to computational processes, ecological detection, or any other singular framework. Instead, it emerges through the dynamic interplay between embodied agents and their environments—a dance of mutual constitution that defies our traditional categories and metaphors.
In this view, the brain isn't a computer running cognitive algorithms, nor is it merely a detector of environmental affordances. It's a participant in relational systems that span traditional boundaries between self and world. Consciousness isn't something that happens within the skull but exists in the spaces between—in the ongoing, dynamic relations that constitute experience itself.
This relational perspective doesn't diminish the importance of neuroscience or computational modeling. Rather, it situates these approaches within a broader framework that acknowledges the irreducibly interactive nature of cognition. The brain matters tremendously, but not as an isolated computational device—as an organ of relation, enabling particular modes of engagement with environments and others.
The ultimate irony may be that our very attempt to understand cognition through technological metaphors—from clockwork to computers to neural networks—reveals something profound about consciousness itself: it is inherently relational, constituted through its engagements rather than existing as a thing-in-itself. Perhaps the most accurate model of consciousness isn't a model at all, but a recognition of its fundamentally participatory, enactive nature—a nature that reveals itself not through detached observation but through embodied involvement in the world we seek to understand.