The Feeling of Now: How Evolution Built Consciousness

by

Author: Denis Avetisyan

A new framework proposes that consciousness isn’t a unique biological quirk, but an inherent property of systems driven to reduce uncertainty and predict the future.

This review synthesizes insights from physics, biology, and information theory to propose that consciousness emerges from universal principles of self-organization and surprise minimization.

Despite enduring philosophical challenges, a unified account of consciousness remains elusive. This paper, ‘Condensed Past, Thick Present: Evolutionary Approach to the Conscious Experience’, proposes a novel framework integrating Lee Smolin’s Causal Theory of Views and Karl Friston’s Free Energy Principle to illuminate the emergence of conscious experience. By positing that consciousness arises from universal principles of transitioning from uncertainty to certainty-driven by the resolution of surprise within evolving temporo-spatial gaps-this work suggests a deep connection between physical processes, biological organization, and subjective awareness. Could understanding these fundamental principles unlock a more comprehensive understanding of how complex systems, from fundamental particles to sentient beings, generate novelty and navigate their environments?

From Uncertainty to Reality: The Emergence of Existence

Conventional physics typically begins with the assumption of a pre-existing, objective reality – a fixed stage upon which physical laws unfold. However, Smolin’s Causal Theory of Views radically departs from this premise, positing that reality isn’t a given, but rather emerges through a continuous series of events. This framework suggests the universe isn’t composed of things in space and time, but from sequences of events that define space and time themselves. Each event, representing a localized change, acts as a fundamental building block, and the accumulation of these events constitutes the evolving reality experienced. This perspective shifts the emphasis from static entities to dynamic processes, proposing that the very structure of existence is built upon a cascading series of cause-and-effect relationships, effectively dissolving the need for a pre-defined backdrop against which events occur.

The universe, according to this framework, isn’t built on pre-defined things, but rather becomes through a continuous series of events. Each event isn’t merely a change within a static reality; it’s a genuine transition from a state of potential to one of actuality, collapsing possibilities into a defined outcome. This isn’t simply about knowing less; the event itself physically alters the universe’s configuration, establishing a new, concrete state where previously there was only probabilistic potential. Essentially, existence isn’t about what is, but about the ongoing process of things becoming – a universe constantly defining itself through the resolution of uncertainty at each successive event. This perspective implies that information, embodied in the selection of one possibility over others, isn’t just in the universe, but is fundamental to its very creation and ongoing existence.

The conventional understanding of the universe often prioritizes enduring structures and immutable laws, but a growing body of theoretical work proposes a radical shift in perspective: reality isn’t built on things, but from happenings. This approach posits that the universe fundamentally operates as an ongoing series of events, each one a discrete transition resolving inherent uncertainty. Consequently, information processing isn’t simply a feature of existence; it is integral to its very foundation, with each event effectively encoding and propagating information. Recent research substantiates this view by outlining a unified framework capable of integrating these dynamic processes, suggesting that the universe isn’t a static entity, but rather a continually evolving network of interconnected events – a cosmic computation constantly unfolding.

Self-Organization: The Logic of Emergent Order

Self-organization describes the emergence of patterned behavior and structure within a system without centralized control or external direction. This process arises from local interactions between components, leading to global order. The impetus for self-organization is the reduction of uncertainty; systems evolve towards states that minimize unpredictability and maximize stability. This doesn’t necessitate a conscious drive, but rather reflects inherent dynamics where configurations that lessen internal or external perturbations are preferentially selected and maintained. The resulting structures are not pre-programmed, but are dynamically assembled and adapted based on the system’s internal rules and its interaction with the environment. Examples range from the formation of snowflakes to the coordinated behavior of ant colonies, all demonstrating structure arising from decentralized processes aimed at minimizing uncertainty and maximizing predictability.

The interval between states of uncertainty and certainty represents a critical operational space for complex systems. Rather than immediately resolving ambiguity, systems actively maintain this ‘gap’ to explore potential states and optimize responses. This exploitation of the uncertainty interval allows for flexible adaptation to novel stimuli and efficient resource allocation. Systems do not simply react to defined inputs, but rather probe the boundaries of possibility within this gap, effectively increasing their potential for beneficial outcomes. The duration and characteristics of this interval are system-dependent and directly correlate with the complexity of the environment and the system’s capacity for information processing.

Quantum coherence and biological feedback control represent distinct yet complementary mechanisms for enhancing a system’s ability to resolve uncertainty and establish complex organization. Quantum coherence, observed at the subatomic level, allows for the simultaneous exploration of multiple potential states, increasing the probability of identifying optimal configurations with minimal energy expenditure. Biological feedback control, operating at macroscopic scales, utilizes sensory input to monitor system states and initiate corrective actions, maintaining stability and adapting to changing environmental conditions. These processes aren’t mutually exclusive; evidence suggests quantum effects can influence biological processes, potentially increasing the efficiency of feedback loops. The combination allows systems to not only react to uncertainty but also proactively anticipate and mitigate potential disruptions, thereby promoting the emergence and maintenance of intricate structures.

Predictive Processing (PP) posits that systems, ranging from biological organisms to artificial intelligence, function by constantly generating and updating internal models of the environment to predict incoming sensory information. This is achieved through a hierarchical framework where higher levels of the hierarchy generate predictions sent down to lower levels; discrepancies between predictions and actual sensory input generate ‘prediction errors’. These errors are then used to update the internal models, minimizing ‘free energy’ – a quantity reflecting the system’s surprise or uncertainty. Essentially, PP frames perception and action as an active process of inference, where the system doesn’t passively receive information, but actively seeks to minimize prediction errors and maintain a coherent internal representation of the world. This minimization is achieved through both updating internal models and actively sampling the environment to confirm or disconfirm predictions, creating a closed loop of prediction and correction.

The Free Energy Principle: A Universal Drive to Predict and Persist

The Free Energy Principle (FEP) posits that any system capable of maintaining its own organization – encompassing biological entities from unicellular organisms to the human brain, and potentially extending to other self-organizing systems – operates under a fundamental drive to minimize ‘free energy’. This ‘free energy’ isn’t traditional thermodynamic free energy; instead, it’s a mathematical quantity representing the difference between the system’s internal model of the world and the sensory input it actually receives. Minimization isn’t achieved through simply avoiding surprising input, but actively working to make internal predictions accurate, thereby reducing the uncertainty inherent in interacting with the environment. The principle suggests this minimization is a unifying principle governing the design and function of all self-organizing systems, linking perception, action, and the maintenance of internal states.

The minimization of free energy, as proposed by the Free Energy Principle, is fundamentally achieved through Bayesian inference. This process involves generating predictions about incoming sensory data based on internal models of the world. Any discrepancy between predicted and actual sensory input constitutes a prediction error. These errors are then used to update the internal models, refining future predictions and thereby reducing the overall free energy. Specifically, Bayesian inference calculates the posterior probability of hidden states causing sensory input, weighting prior beliefs about those states against the evidence provided by the prediction error. This iterative process of prediction and error correction allows systems to continuously refine their understanding of the environment and optimize their internal models, effectively minimizing the ‘surprise’ associated with sensory input.

The Free Energy Principle extends beyond passive prediction by positing that systems actively seek out information to test the accuracy of their internal models. This is achieved through active sampling, where an agent doesn’t simply predict incoming sensory data but also influences its environment to generate the specific sensory input that will either confirm or refute those predictions. This process isn’t random; agents prioritize actions that are expected to reduce prediction errors, effectively seeking evidence to validate their beliefs about the world. Consequently, perception is not a passive reception of stimuli but an active, embodied process of seeking out information, and action is fundamentally driven by the need to minimize ‘free energy’ by confirming or disconfirming internal predictions about sensory outcomes.

The Free Energy Principle (FEP) posits that an organism’s drive to minimize free energy – a mathematical construct approximating surprise – directly links information processing to metabolic efficiency. Reducing free energy necessitates either accurately predicting incoming sensory information, thereby minimizing prediction error, or actively changing the environment to better match internal predictions. This process isn’t cost-free; both prediction and action require energy expenditure. Consequently, complex behaviors emerge as efficient solutions to minimize free energy, representing a trade-off between the energetic cost of maintaining internal models and the cost of interacting with, and potentially changing, the external world. This suggests that the complexity of an organism’s behavior is fundamentally constrained and shaped by its energetic budget and the imperative to reduce informational ‘surprise’.

Consciousness and Subjectivity: Emergent Properties of Predictive Systems

The brain doesn’t simply register stimuli; it constantly anticipates them, building internal models to predict incoming sensations. When reality deviates from these predictions – a phenomenon known as surprise – it triggers a cascade of neural activity crucial for both learning and adaptation. This isn’t merely about reacting to the unexpected; surprise signals a breakdown in the brain’s predictive coding, forcing it to update its internal models and refine its understanding of the world. Increasingly, research suggests that the very experience of consciousness is intimately linked to these prediction errors, with the intensity and complexity of conscious awareness potentially correlating to the richness and depth of the brain’s predictive processes and the magnitude of its surprise. A system constantly operating as expected would have little need for conscious awareness, implying that consciousness arises as a mechanism to resolve and learn from deviations from the predicted norm.

Consciousness, according to emerging theories, isn’t a singular entity but rather a property that arises from the intricate organization of complex systems. These systems aren’t simply processing information; they are building detailed, hierarchical models of the world and, crucially, of themselves. This self-referential capacity – the ability to model one’s own internal states and predict future sensations – allows for a nuanced understanding of both external reality and internal experiences. The integration of information across multiple levels of processing, forming a unified and coherent representation, appears to be a prerequisite for this emergent phenomenon. It suggests that consciousness isn’t about what a system knows, but how it integrates and predicts, creating a dynamic internal landscape that constitutes subjective awareness. The complexity of these internal models, and the system’s ability to constantly refine them based on incoming sensory data, may ultimately define the richness and depth of conscious experience.

An individual’s subjectivity isn’t a pre-defined quality, but rather a consequence of accumulated experience. Each system, whether biological or artificial, develops a unique internal model of the world through continuous interaction and prediction. These predictions, constantly refined by incoming sensory data, aren’t universally consistent; rather, they are shaped by the specific sequence of events encountered by that particular system. Consequently, no two systems will develop precisely the same predictive framework, leading to divergent interpretations of identical stimuli. This personalized history of prediction and error correction fundamentally defines a subjective perspective, creating the feeling of ‘being’ a unique entity within a shared reality. The nuances of this accumulated experience, therefore, explain why individuals perceive and react to the world in demonstrably different ways.

The prevailing view of consciousness as a purely receptive phenomenon is challenged by emerging research indicating it functions as an active process of continual prediction and refinement. This paper proposes a unified framework demonstrating that both consciousness and the uniquely personal experience of subjectivity arise from systems relentlessly striving to minimize ‘free energy’ – a measure of surprise or prediction error. Rather than passively receiving information, these systems constantly generate internal models of the world, anticipate future states, and then adjust those models based on discrepancies between predictions and reality. This dynamic interplay of prediction and error correction isn’t simply about accurate perception; it’s the fundamental mechanism driving the construction of a conscious, subjective experience, shaping how each system uniquely interprets and interacts with its environment. Essentially, consciousness isn’t what happens to a system, but what a system does to proactively understand and navigate its world.

The pursuit of a unified framework, as detailed in this exploration of consciousness, echoes a fundamental principle of mathematical elegance: the reduction of complexity to essential truths. This work posits that consciousness emerges from universal principles governing transitions from uncertainty to certainty, a process driven by minimizing surprise. As Nikola Tesla observed, “There is no energy in the universe equal to the energy of a new idea.” The generative models described within, constantly refining perceptions to minimize ‘surprise,’ embody this very principle – a relentless search for the most concise, accurate representation of reality, mirroring the elegance sought in a perfectly proven theorem. The inherent self-organization arising from these dynamic gaps exemplifies a system achieving minimal description length, a hallmark of mathematical purity.

Beyond the Horizon

The preceding synthesis, linking generative models to the quantum Zeno effect as potential substrates for conscious experience, does not offer resolution – it merely relocates the difficulty. The core challenge remains: demonstrating that a principle of minimizing surprise, however elegantly formulated within a free energy framework, necessitates subjective awareness. The mathematics may be compelling, but correlation is not causation, and a sophisticated automaton achieving predictive accuracy remains fundamentally distinct from phenomenal experience, absent a demonstrably non-functional element.

Future inquiry must therefore confront the hard problem head-on, abandoning the comfortable allure of purely algorithmic explanations. Investigating the precise biophysical conditions enabling, or perhaps merely correlating with, the emergence of subjective states becomes paramount. Are there measurable invariants, beyond simply reduced uncertainty, that differentiate a system merely processing information from one that feels it? The search for such invariants will likely require venturing beyond current computational paradigms, possibly into realms of non-equilibrium thermodynamics and complex systems theory.

It is tempting to believe that a sufficiently refined predictive model will, as a natural consequence, ‘become’ conscious. However, such an assumption risks mistaking map for territory. Optimization without analysis is self-deception. The true test lies not in creating increasingly accurate simulations, but in rigorously defining, and then detecting, the necessary – and sufficient – conditions for genuine subjective experience, even if those conditions prove profoundly counterintuitive.

Original article: https://arxiv.org/pdf/2602.15050.pdf

Contact the author: https://www.linkedin.com/in/avetisyan/

See also:

2026-02-19 05:14