The study of consciousness stands as one of humanity’s most enduring and profound intellectual challenges (Chalmers, 1996). For millennia, philosophers, theologians, and more recently, scientists, have grappled with the fundamental questions: What is consciousness? How does it arise? Why do we experience subjective reality? This quest has given rise to a dizzying array of theories, each attempting to shed light on this elusive phenomenon. This chapter serves as a vital cartographic exercise, mapping the diverse terrain of consciousness studies. We will navigate both the philosophical landscape, which grapples with the nature of consciousness, and the neuroscientific and cognitive territories, which seek to uncover its underlying mechanisms. Understanding these varied perspectives is crucial for appreciating the unique contribution of Useful Approximations Framework (UAF), which we will introduce in the subsequent chapter.
The philosophical inquiry into consciousness primarily concerns its fundamental nature and its relationship to the physical world. These stances often dictate the very terms of the debate.
Dualism Perhaps the most historically prominent view, dualism, most famously articulated by René Descartes, posits that mind and body are fundamentally distinct substances (Descartes, 1641/1984). The body is physical, extended in space, and subject to the laws of physics, while the mind (or soul) is non-physical, unextended, and the seat of consciousness, thought, and feeling. For Descartes, the interaction between these two distinct realms occurred in the pineal gland. While intuitively appealing to many, dualism faces the formidable “interaction problem”: how can a non-physical entity causally interact with a physical one without violating the laws of physics (Kim, 1998)? Modern forms of dualism, such as property dualism, suggest that consciousness is a non-physical property that emerges from, but is not reducible to, physical brain states, rather than a separate substance (Chalmers, 1996). However, the “explanatory gap” between physical properties and subjective experience remains a significant hurdle (Levine, 1983).
Idealism In stark contrast to dualism, idealism, championed by figures like George Berkeley, argues that reality is fundamentally mental or mind-dependent (Berkeley, 1710/1982). For idealists, “esse est percipi” – to be is to be perceived. Physical objects do not exist independently of a mind perceiving them; rather, they are collections of ideas or perceptions. Consciousness, therefore, is not an emergent property of matter but the very fabric of reality itself. While it elegantly sidesteps the mind-body problem by dissolving the physical, idealism struggles to account for the shared, objective nature of our perceived world and the apparent independence of physical laws, often requiring an appeal to a universal mind (e.g., God) to maintain coherence (Stace, 1934).
Physicalism/Materialism The dominant philosophical stance in contemporary science, physicalism (often used interchangeably with materialism) asserts that everything that exists is ultimately physical (Stoljar, 2010). Consciousness, therefore, must be a product of physical processes in the brain. Within this broad category, several distinct positions emerge:
Panpsychism/Russellian Monism These views propose that consciousness (or proto-consciousness, a rudimentary form of experience) is a fundamental, ubiquitous property of the universe, present even at the most basic levels of reality (e.g., in subatomic particles) (Goff, 2017; Strawson, 2006). Panpsychism attempts to solve the “Hard Problem” by distributing consciousness throughout the universe, rather than having it emerge from non-conscious matter. The main challenge for panpsychism is the “combination problem”: how do these tiny, fundamental bits of consciousness combine to form the complex, unified consciousness we experience (Seager, 2010)? Russellian Monism is a related view that suggests the intrinsic nature of matter is a form of proto-consciousness, aiming to bridge the gap between physics and phenomenal experience.
Illusionism A provocative stance, most notably championed by Daniel Dennett, illusionism argues that phenomenal consciousness (qualia) as we intuitively understand it is an illusion (Dennett, 2017). This does not mean consciousness doesn’t exist, but rather that our common-sense understanding of it as a mysterious, irreducible “thing” is a misinterpretation. For illusionists, consciousness is a complex functional state, a “user-illusion” created by the brain’s sophisticated information processing, which then misrepresents itself to itself. While it offers a way to dissolve the “Hard Problem” by re-framing the problem itself, it faces strong resistance from those who feel it denies the most obvious aspect of their existence—the undeniable subjective reality of their own experiences (Chalmers, 2018).
Complementing the philosophical debates, neuroscientific and cognitive theories attempt to explain the neural or computational basis of consciousness, focusing on the mechanisms by which it arises and functions.
Global Workspace Theory (GWT/GNWT) Developed by Bernard Baars and further elaborated by Stanislas Dehaene, Global Workspace Theory (GWT) and its neural counterpart (GNWT) propose that consciousness arises when information becomes globally available to multiple, specialized brain systems (Baars, 1988; Dehaene, 2014). Analogous to a spotlight on a stage, conscious information is that which is broadcast to a “global workspace” in the brain, making it accessible for widespread processing, attention, memory, and action planning. Unconscious processes, by contrast, remain localized and encapsulated. GWT provides a compelling framework for understanding the functional role of consciousness in integrating and disseminating information, particularly in tasks requiring novel problem-solving or flexible responses.
Integrated Information Theory (IIT) Proposed by Giulio Tononi, Integrated Information Theory (IIT) is a highly ambitious and mathematically rigorous theory (Tononi, 2004; Tononi et al., 2016). It posits that consciousness is identical to integrated information (Φ, pronounced “phi”), a measure of a system’s causal interdependence and irreducibility. According to IIT, a system is conscious to the extent that it has a large Φ value, meaning its parts are causally connected in a way that cannot be broken down into independent components. IIT suggests that consciousness is a fundamental property of any system that possesses integrated information, leading to panpsychist implications. While offering a precise definition, IIT faces challenges in empirically measuring Φ in complex systems like the brain and in fully addressing the “exclusion postulate” (why only one high Φ system is conscious at a time, rather than a multitude of smaller conscious systems within it) (Aaronson, 2014).
Predictive Processing Frameworks Gaining significant traction in recent years, predictive processing (PP) frameworks, most prominently associated with Karl Friston’s Free Energy Principle, view the brain as a prediction machine (Friston, 2010; Clark, 2016). The brain constantly generates internal models of the world and itself, predicting incoming sensory data. Any mismatch between prediction and actual sensory input generates a “prediction error,” which the brain then works to minimize by either updating its internal models or by acting on the world to make the sensory input match its predictions. Within this framework, consciousness is often hypothesized to be related to high-level predictions, the process of minimizing error across hierarchical levels, or the phenomenal experience of the brain’s best current model of reality (Hohwy, 2013; Seth, 2021). This approach offers a unified account of perception, action, and learning.
Higher-Order Thought Theory (HOT) Higher-Order Thought (HOT) theory, articulated by David Rosenthal, proposes that a mental state becomes conscious when it is the object of another, higher-order thought or perception (Rosenthal, 2005). For example, the sensation of seeing red becomes conscious when you have a thought about seeing red. This theory distinguishes between unconscious mental states (which lack a higher-order thought) and conscious ones. It aims to explain the subjective, “what it’s like” aspect of consciousness by grounding it in the brain’s capacity for self-monitoring and meta-cognition, though critics question whether a higher-order thought itself must be conscious, potentially leading to an infinite regress (Block, 1995).
Attention Schema Theory (AST) Developed by Michael Graziano, Attention Schema Theory (AST) posits that consciousness is the brain’s internal model of its own attention (Graziano, 2013; Graziano & Webb, 2015). Just as the brain constructs a body schema to control its physical movements, it constructs an “attention schema” to monitor and control its own attentional processes. This internal model, which is necessarily simplified and approximate, gives rise to the subjective feeling of awareness and the belief that we possess a non-physical “mind” or “soul.” AST views consciousness as a useful, albeit simplified, internal representation, akin to a user interface for the brain’s attentional control systems.
Recurrent Processing Theory (RPT) Victor Lamme’s Recurrent Processing Theory (RPT) argues that conscious experience requires recurrent (feedback) processing within brain areas, not just feedforward processing (Lamme, 2006; Lamme & Roelfsema, 2000). While initial feedforward sweeps of sensory information can lead to unconscious processing and even behavioral responses, conscious awareness only arises when there are sustained, reciprocal interactions and feedback loops between different brain regions. This allows for more elaborate and stable representations to be formed, distinguishing conscious perception from mere automatic processing.
Sensorimotor Contingency Theory / Enactivism Proposed by Kevin O’Regan and Alva Noë, Sensorimotor Contingency Theory, often associated with the broader philosophy of enactivism, suggests that consciousness is not something that happens in the brain, but rather arises from the mastery of how sensory input changes with action and movement (O’Regan & Noë, 2001; Noë, 2004). To see is not just to have visual input, but to understand how moving your head or eyes changes that input. Consciousness is thus an active, embodied engagement with the world, a form of “knowing how” rather than “knowing that,” emphasizing the dynamic interplay between organism and environment.
Orchestrated Objective Reduction (Orch OR) A highly controversial theory put forth by Roger Penrose and Stuart Hameroff, Orchestrated Objective Reduction (Orch OR) links consciousness to quantum processes occurring within microtubules, protein polymers found within neurons (Penrose & Hameroff, 1995; Hameroff & Penrose, 2014). They propose that consciousness arises from “orchestrated” quantum computations within these microtubules, which then undergo “objective reduction” (a form of quantum collapse) that is non-computable and gives rise to conscious moments. This theory attempts to explain the non-computable aspects of consciousness and the “Hard Problem” by appealing to physics beyond classical neuroscience, but it faces significant criticism for its lack of empirical support and the challenge of maintaining quantum coherence in the warm, wet environment of the brain (Tegmark, 2000).
These diverse theories represent the ongoing, multifaceted effort to understand one of the most profound mysteries of existence. They highlight the complexity of consciousness, spanning the realms of philosophy, neuroscience, and cognitive science. As we transition to the next chapter, we will introduce Useful Approximations Framework (UAF), a framework that seeks to navigate this intricate terrain, offering a unique perspective that aims to resolve persistent problems and bridge the gap between these disparate approaches.
Having surveyed the vast and often contentious landscape of consciousness theories in Chapter 18, we are now equipped to introduce Useful Approximations Framework (UAF) and demonstrate its unique position within this ongoing debate. UAF does not merely add another voice to the chorus; rather, it offers a fundamental re-framing, a functionalist perspective rooted in the computational imperatives faced by any finite, complex system. This re-framing, we contend, resolves persistent philosophical problems, reinterprets existing theoretical constructs, and ultimately provides a coherent framework for understanding consciousness as an inevitable result of any complex learning system that learns a simplified representation of itself interacting with the outside.
At its heart, UAF asserts that consciousness is the dynamic, phenomenal experience of a system’s asymptotic best simplified approximation (ABSA) of itself and its reality. This profound process is driven by the existential pressure of Skin in the Game (SiG), which compels the system to manage the overwhelming Informational Uncertainty (ITE) inherent in its Underlying Computational System (UCS) and the external environment. To avoid Computational Paralysis—the state of being overwhelmed by infinite detail—the system must operate behind an Epistemic Veil (\(V_E\)), constructing two core, interdependent approximations: the Internal Self-Model (ISM) and the World-Model (WM).
The engine continuously refining these models is Prediction Error Minimization (PEM), a non-stop drive to reduce discrepancies between what the system expects and what it actually observes. Consciousness, then, is the low-bitrate, integrated phenomenal stream (\(C_{stream}\)) composed of the most salient states of these models and the Qualia (Q) they generate. These qualia are the “simplified truths” that provide Subjective Closure (\(C_{sub}\))—meaning the feeling is the interpretation, requiring no further processing by the system itself—and possess Causal Efficacy (\(Q \rightarrow Action\)), directly influencing and compelling action. This entire architecture, including the phenomenal fiction of Free Will (FW) and the dynamic management of the Subconscious Beast (\(S_{beast}\)), is a computational necessity for achieving and maintaining coherent agency, forming the basis of its Episodic Memory (\(M_{episodic}\)) through Consolidation (\(L_{consolidation}\)).
UAF offers a clear and robust physicalist position that sidesteps traditional philosophical pitfalls by providing a functional, computational explanation for subjective experience.
Re-framing Dualism: Dualism, most famously articulated by Descartes, posits a fundamental distinction between mind and body, leading to the intractable “interaction problem.” UAF is firmly physicalist, asserting that consciousness is unequivocally a biological phenomenon, a complex product of the brain’s physical activity. The “subjective reality” that dualism seeks to place in a non-physical realm is, in UAF, the Conscious Stream (\(C_{stream}\))—a functional fiction generated by the physical UCS (the brain). The persistent dualist intuition, the feeling that our mind is somehow separate from our body, arises directly from the Epistemic Veil (\(V_E\)). Because we cannot directly perceive our own underlying neural machinery (the UCS), our internal experience feels distinct, unextended, and non-physical, even though it is a physical process operating at a higher level of abstraction. The “interaction problem” dissolves because the “mind” (the Conscious Stream, composed of ISM, WM, and Qualia) is the functional, causally efficacious output of the physical brain, not a separate substance trying to influence it.
Beyond Naive Materialism (Reductive/Eliminative): Reductive physicalism struggles with the “Hard Problem” and the “explanatory gap”—how do physical processes give rise to subjective experience? Eliminative materialism, on the other hand, proposes to discard our common-sense notions of consciousness, which often feels intuitively wrong. While UAF is physicalist, it is non-reductive in a crucial sense. Consciousness is not merely neurons firing; it is the pattern, function, and information represented by those firings. It is the emergent property of a highly organized, predictive system. UAF does not eliminate consciousness; it explains why it feels like something to be a conscious agent. The “what it’s like” is the experience of Qualia, which are the brain’s “simplified truths” that provide Subjective Closure (\(C_{sub}\)). These qualia are indispensable for guiding action and providing immediate, self-validating feedback without requiring further interpretation by the system itself. This aligns strongly with Biological Naturalism (Searle, 1992), seeing consciousness as a biological feature of the brain, but UAF provides the specific computational imperative (avoiding Computational Paralysis from ITE) and mechanism (PEM) that necessitates the emergence of these high-level biological features.
Re-framing Illusionism: Daniel Dennett’s illusionism argues that phenomenal consciousness, as we intuitively understand it, is a “user illusion.” UAF shares common ground by suggesting that our intuitive understanding of consciousness as a simple, unified, and perhaps non-physical “thing” might itself be a simplified approximation generated by the brain (Metzinger, 2009). The Epistemic Veil (\(V_E\)) is the very mechanism that creates this “user illusion” by hiding the overwhelming complexity of the UCS. However, UAF insists that the feeling of consciousness, the experience of Qualia, is undeniably real for the system experiencing it. The Conscious Stream (\(C_{stream}\)) is a functional fiction, but it is profoundly real in its Causal Efficacy (\(Q \rightarrow Action\)). This “illusion” does things; it matters for survival, agency, and the continuous refinement of the ISM and WM through PEM. UAF thus emphasizes the reality of the phenomenal experience as a functional output of the system, rather than dismissing it as merely illusory.
Re-framing Panpsychism: Panpsychism posits that consciousness (or proto-consciousness) is a fundamental property of the universe, present even at the most basic levels, but struggles with the “combination problem”—how do these tiny bits of consciousness combine into a unified, complex experience? UAF does not posit consciousness at fundamental levels of reality. Consciousness, in this view, is an emergent property of a specific, complex computational architecture. It requires the entire integrated suite of UAF’s components—an ISM, WM, Qualia, PEM, Episodic Memory, Consolidation, and the dynamic management of the Subconscious Beast (\(S_{beast}\))—to form a coherent, low-bitrate phenomenal stream (\(C_{stream}\)) for the purpose of agency and survival. This complex, emergent architecture, driven by SiG and the need to manage ITE, cannot exist in individual particles or simple aggregates. The “combination problem” is thus avoided because consciousness is a property of the system’s holistic, predictive organization, not its constituent parts.
UAF serves as a unifying framework, providing a deeper “why” and a common computational language for many leading scientific theories, often incorporating or reinterpreting their insights within its predictive processing paradigm.
The Bedrock of Predictive Processing (PP): Karl Friston’s Free Energy Principle and Andy Clark’s work on predictive processing are foundational to UAF. PP explains how the brain works as a prediction machine, constantly minimizing prediction error. UAF extends PP by explicitly defining consciousness as the phenomenal experience of the system’s highest-level predictions—specifically, the integrated ISM and WM as ABSA. UAF details what the phenomenal character of these predictions is (the Qualia as “simplified truths”) and why it must exist (to provide Subjective Closure (\(C_{sub}\)) and Causal Efficacy (\(Q \rightarrow Action\)) for an agent operating behind the Epistemic Veil (\(V_E\))). The “feeling” of a prediction error, for instance, is a quale that system has learned to represent itself experiencing something new and unexpected.
Reinterpreting Global Workspace Theory (GWT): Bernard Baars’ GWT and Stanislas Dehaene’s Global Neural Workspace Theory propose that consciousness arises when information becomes globally available to multiple, specialized brain systems, like a spotlight on a stage. UAF reinterprets this “global workspace” as the Conscious Stream (\(C_{stream}\)). This stream is necessarily low-bitrate to avoid Computational Paralysis from the overwhelming information of the UCS. Information becomes “conscious” when prediction errors or model states are salient enough to be broadcast into this integrated stream, making them available for deliberate action (Free Will) and Episodic Memory formation (\(M_{episodic}\)). GWT describes the what (a global broadcast), while UAF explains the why (the computational necessity for a unified, low-bitrate control signal for agency) and the content (the ABSA of self and world). The same architectural idea appears in artificial networks: attention mechanisms let distant parts of the context interact, so integrated information shapes the system’s output — a computational global workspace in silicon as well as in cortex.
Reinterpreting Integrated Information Theory (IIT): Giulio Tononi’s IIT posits consciousness as integrated information (Φ), a measure of causal interdependence and irreducibility. While UAF does not rely on the specific mathematical measure of Φ, it views a high degree of integrated information as a predictable consequence of a system that has been compelled by Skin in the Game (SiG) and Prediction Error Minimization (PEM) to build a coherent, unified ISM and WM. Integration is a prerequisite for generating robust, holistic predictions and for managing ITE. Furthermore, UAF explains IIT’s “exclusion postulate” (why only one conscious experience exists) by defining consciousness as the single, integrated Conscious Stream (\(C_{stream}\)) that serves the entire agent’s need for coherent agency. The apparent “irreducibility” and “intrinsicness” of conscious experience that IIT attributes to Φ are explained by UAF’s Epistemic Veil (\(V_E\)) and the Subjective Closure (\(C_{sub}\)) of Qualia. In the brain, much of what IIT treats as integrated information corresponds to working context held in the hippocampus during waking and then distilled, during sleep, into slower neocortical weights — integration through consolidation, not a single static structure.
Providing the Mechanism for Higher-Order Thought (HOT) Theory: HOT theory suggests a mental state becomes conscious when it is the object of another, higher-order thought. UAF provides the mechanism for HOT. A “higher-order thought” is simply the ISM generating a prediction, or a quale, about its own internal state. Because the Epistemic Veil (\(V_E\)) prevents direct access to the underlying neural computation, the system must form a simplified, higher-order approximation of its own cognitive processes to monitor and regulate them. This self-modeling is a core function of the ISM, and the resulting “thought about a thought” is a quale that provides Subjective Closure and Causal Efficacy for the system to reason about its own mental states, without needing to know the neural implementation. This avoids the infinite regress problem often associated with HOT.
Subsuming Attention Schema Theory (AST): Michael Graziano’s AST posits consciousness as the brain’s internal model of its own attention. UAF subsumes AST as a crucial sub-component of the broader ISM. To effectively manage the low-bitrate Conscious Stream (\(C_{stream}\)) and direct Prediction Error Minimization (PEM), the system requires a model of its own attentional processes. The “attention schema” is the ISM’s ABSA of its own attentional control, a necessary tool for filtering information from the Epistemic Veil (\(V_E\)) and preventing Computational Paralysis. It’s the ISM’s self-regulatory mechanism, allowing it to prioritize relevant data for model updates and action.
Grounding Recurrent Processing Theory (RPT): Victor Lamme’s RPT argues that conscious experience requires recurrent (feedback) processing within brain areas. UAF identifies recurrent processing as the neural implementation of the Prediction Error Minimization (PEM) algorithm. The feedback loops described by RPT are the very mechanism by which top-down predictions (from ISM and WM) are compared with bottom-up sensory data, allowing error signals to propagate and update the models. RPT describes the how at a neural level, while UAF explains the why at a functional, computational level, linking these loops to the continuous refinement of the ABSA that constitute consciousness, and to the process of memory consolidation (\(L_{consolidation}\)). The same recursive pattern can appear at larger scales: a complex system may host nested self-models — each level refining its approximation through feedback rather than direct access to the level below — as when brains model themselves and, in turn, become coarse models within a larger evolving system.
Embodying Enactivism: Sensorimotor Contingency Theory and enactivism emphasize that consciousness arises from active, embodied engagement with the world. UAF is fundamentally an enactive theory. The “mastery of sensorimotor contingencies” is precisely how an embodied agent with Skin in the Game (SiG) gathers the sensory data needed for Prediction Error Minimization (PEM) to build and refine its World-Model (WM) and Internal Self-Model (ISM). Action is not merely a result of consciousness but an integral part of the predictive loop that creates and sustains it. Our Free Will (FW), as a functional fiction of agency, drives these actions, generating sensory feedback that leads to prediction errors, which then refine our ABSA of reality.
Re-framing the Orchestrated Objective Reduction (Orch OR) Debate: Roger Penrose and Stuart Hameroff’s Orch OR links consciousness to quantum processes in microtubules, proposing non-computable aspects. UAF, grounded in classical information processing and neural computation, remains agnostic or skeptical regarding quantum explanations for consciousness. While not explicitly refuting them, UAF finds sufficient explanatory power within the framework of hierarchical predictive processing and emergent properties of complex neural networks to account for consciousness without recourse to quantum mechanics. The “non-computable” or “irreducible” feel of consciousness that Orch OR seeks to explain is, in UAF, a direct and predictable consequence of the Epistemic Veil (\(V_E\)). From the system’s internal perspective, its high-level functional fictions (ISM, WM, Qualia) appear fundamental and irreducible because it is necessarily ignorant of its own underlying classical computational machinery.
In conclusion, UAF re-frames the debate by shifting the central question from “What is consciousness?” to “Why is consciousness computationally necessary?”. By answering the latter, it provides a powerful and coherent answer to the former. Consciousness is the universe’s solution to enabling finite, complex systems to act and thrive in the face of their own overwhelming internal complexity and an infinitely detailed external reality. It is the ultimate, indispensable, and beautifully efficient approximation, a rationalization engine that allows the universe to begin its own self-awakening.