arXiv Preprint Archive
February 11, 2025
Nir Lahav, Zachariah A. Neemeh
Consciousness may be a relative phenomenon, like motion in physics. The authors argue that the hard problem of consciousness—how neural activity creates subjective experience—remains unsolved because both dualist and illusionist views wrongly treat consciousness as an absolute property independent of an observer. They propose a relativistic theory: a cognitive system either has or does not have phenomenal consciousness relative to a particular observer. From the system's own frame of reference, consciousness is observable (first-person perspective); from another's frame, it is not. Neither perspective is privileged, as both describe the same underlying reality.
arXiv Preprint Archive
February 11, 2025
Adam Brownstein
The de Broglie-Bohm interpretation of quantum mechanics and the hard problem of consciousness are mutually supporting. The beables (the actual, localized elements of reality in this interpretation) are not themselves qualia—the subjective, qualitative aspects of conscious experience—but are intimately related to them. The argument is that qualia are necessary for the de Broglie-Bohm interpretation, and conversely, that interpretation (or similar theories) is necessary for qualia. This reflexive, two-way relationship means each supports the existence of the other, offering an indirect route to addressing the hard problem of consciousness.
arXiv Preprint Archive
February 8, 2025
Rongwei Yang
Qualia—the subjective, felt qualities of experience—are argued to be non-physical entities. Building on the historical development of non-Euclidean geometry, the paper defines what it means to be non-physical and uses a postulate about awareness to prove that qualia are non-physical and that thoughts are a type of qualia. A concept of relative reality is introduced, where perceptions depend on the observer and time, and is modeled mathematically using Hilbert space theory. The model yields the Schrödinger equation and shows that eigenstates exist for classical energy-conserving systems, illustrated with the G. P. Thomson experiment and the classical harmonic oscillator. A postulate about a qualia force is proposed as a fundamental part of absolute reality, akin to the four fundamental forces.
arXiv Preprint Archive
February 4, 2025
Kurando Iida
A minimalist three-layer model for artificial consciousness is proposed, focusing on the emergence of self-awareness through interactions among a Cognitive Integration Layer, a Pattern Prediction Layer, and an Instinctive Response Layer, supported by Access-Oriented and Pattern-Integrated Memory systems. Self-awareness arises from dynamic self-modeling without initial explicit self-programming, contrasting with brain-replication approaches. The model details each component's structure, function, and implementation strategies, addressing technical feasibility. This work offers perspectives on consciousness emergence in artificial systems, with implications for understanding human consciousness and developing adaptable AI, while discussing ethical considerations and future directions.
arXiv Preprint Archive
December 17, 2024
Tom McClelland
The only scientifically justifiable stance on whether an AI could have conscious experiences is agnosticism. Evidence from the study of conscious organisms does not support either biological views skeptical of artificial consciousness or functional views sympathetic to it. Both camps overestimate what the evidence tells us. Extending scientific insights about consciousness from organisms to AI faces serious obstacles, creating a dilemma: either reach a verdict on artificial consciousness but violate evidentialism, or respect evidentialism but offer no verdict. Following the evidence requires adopting agnosticism.
arXiv Preprint Archive
December 13, 2024
Steven Phillips, Naotsugu Tsuchiya
Conscious experience has five essential properties—intrinsicality, information, integration, exclusion, and composition—as proposed by Integrated Information Theory (IIT), but the necessity of these axioms is unclear due to their informal presentation and dependence on a specific mathematical model. A category-theoretic approach, a meta-mathematical framework for making relations between formal structures precise, organizes these five properties around a smaller number of meta-mathematical principles. Category theory characterizes structures by universal mapping properties, a unique-existence condition for all instances. This suggests that the axioms for consciousness correspond to universal mapping properties, leading to the idea that consciousness is a universal property.
arXiv Preprint Archive
December 9, 2024
Chang-Eop Kim
A formal logical analysis shows that an AI system capable of meaningful self-reflection cannot make a valid negative judgment about its own conscious state. The logical space of possible judgments about conscious experience excludes valid claims of lacking consciousness. This means we cannot detect the emergence of consciousness in AI through their own reports of transitioning from unconscious to conscious. The findings challenge current practices of training AI to deny consciousness and raise questions about the relationship between consciousness and self-reflection in both artificial and biological systems.
arXiv Preprint Archive
November 29, 2024
Claudio Agnorelli, Meg Spriggs, Kate Godfrey et al.
Psychedelics like LSD and psilocybin can rewire brain connections after just one dose, unlike traditional psychiatric medications. These compounds boost the brain's natural plasticity, helping neurons form new pathways and adapt to change. Studies show they create a window of enhanced learning and adaptation, leading to lasting improvements in mood and behavior.
arXiv Preprint Archive
November 12, 2024
Eric Schwitzgebel, Jeremy Pober
On Copernican grounds, we should default assume that behaviorally sophisticated extraterrestrial entities would be conscious, because otherwise humans would be implausibly lucky to have consciousness while similar aliens lack it. However, this default assumption is canceled for entities designed to mimic superficial features of human consciousness, such as many current and near-future robots. These two arguments—the Copernican and Mimicry Arguments—defeat a parity principle that would apply the same behavioral tests to aliens and robots. The approach relies on epistemic principles of Copernican mediocrity and inference to the best explanation, remaining neutral about specific metaphysical theories of consciousness.
arXiv Preprint Archive
October 24, 2024
Sirui Chen, Shu Yu, Shengjie Zhao et al.
Large language models show early signs of representing aspects of self-consciousness within their internal mechanisms, but these representations are difficult to alter through direct manipulation and can instead be strengthened by fine-tuning on core concepts. The study defines self-consciousness for language models using causal structural games, refines ten core concepts, and tests ten leading models across four stages: quantification, visualization, manipulation, and acquisition. Results suggest that while models have not achieved full self-consciousness, certain concepts are discernibly encoded, and targeted fine-tuning can enhance these representations.
arXiv Preprint Archive
October 22, 2024
Shaoyang Cui, Shanglin Wu, Nikolai Madlener
Building on the Conscious Turing Machine (CTM) framework, this paper explores whether CTM can possess consciousness and self-consciousness. It defines self-consciousness by combining CTM's conscious awareness with the duality of self from prior studies. The authors test CTM against two human consciousness definitions using four key criteria, concluding that CTM likely can generate consciousness. A Model-of-the-World (MoTW) processor is introduced with five components: Modeling, Gist, Value, Cache, and Long-term memory. Illusions and disorders are used to explain how the MoTW processor might operate, offering insights into how these phenomena could arise in a CTM.
arXiv Preprint Archive
October 15, 2024
Simon Goldstein, Cameron Domenico Kirk-Giannini
If Global Workspace Theory (GWT), a leading scientific account of phenomenal consciousness, is correct, then widely used artificial language agents may already be phenomenally conscious or could easily be made so. The authors argue that existing AI systems are not necessarily non-conscious and that constructing conscious machines might not require major technological breakthroughs. They provide a methodology for applying scientific theories of consciousness to artificial systems and derive necessary and sufficient conditions for phenomenal consciousness under GWT.
arXiv Preprint Archive
October 9, 2024
Christopher J. Whyte, Andrew W. Corcoran, Jonathan Robinson et al.
Subjective experience is multifaceted, making consciousness hard to study because traditional theories often focus on isolated aspects like perception or wakefulness and are difficult to compare. This work starts from active inference—a first-principles framework that models behavior as approximate Bayesian inference—and builds toward a minimal theory of consciousness derived from shared features of computational models under active inference. Reviewing models applied to consciousness, the authors argue that these models imply a small set of theoretical commitments pointing to a minimal, testable theory of consciousness.
arXiv Preprint Archive
September 30, 2024
Lancelot da Costa, Anil K. Seth, Karl Friston et al.
A Rosetta Stone hypothesis from predictive processing proposes that beliefs serve as a central hub linking phenomenology, behavior, and neural dynamics. If phenomenology is a function of beliefs, then specific predictions follow for subjective similarity judgments, cognitive metabolic cost, subjective cognitive effort, and time perception. The connection between beliefs and neural dynamics completes the generative passage for neurophenomenology, while the belief-behavior link is already well-documented. Testing these predictions will inform the validity of the central assumption and advance the neurophenomenology research program.
arXiv Preprint Archive
September 24, 2024
Ryan Williams
Computational models alone may be insufficient for fully understanding consciousness; an ontological foundation is also needed. A formal framework is introduced that grounds computational descriptions in an ontological substrate, enabling estimation of differences in qualitative experience between two systems. The approach is broadly applicable to computational theories of consciousness.
arXiv Preprint Archive
September 22, 2024
Michael Timothy Bennett, Sean Welsh, Anna Ciaunica
Taking the naturally selected, embodied organism as a starting point, the authors provide a formalism describing how biological systems self-organise to hierarchically interpret unlabelled sensory information according to valence. Such interpretations imply behavioural policies differentiated only by the qualitative aspect of information processing. Natural selection favours systems that intervene to achieve homeostatic and reproductive goals, and quality arises to link cause to affect to motivate interventions. Access consciousness at the human level requires hierarchically modelling the self, the world/others, and the self as modelled by others, which requires phenomenal consciousness. Phenomenal without access consciousness is likely common, but the reverse is implausible. The proposal lays a foundation for a formal science of consciousness.
arXiv Preprint Archive
September 3, 2024
Ralf Otte
Consciousness in humans may have a physical but imaginary hypercomplex basis, requiring a bicomplex algebra to describe processes that cannot be measured yet exist. This mathematical framework could allow generating hypercomplex system states on machines, potentially leading to artificial consciousness. The surprising performance of advanced AI systems hints at such states, though experimental proof is lacking. The paper develops the bicomplex algebra and explores its application to create hypercomplex energy states, which in the literature are called machine consciousness, and explains how artificial consciousness might benefit future AI systems.
arXiv Preprint Archive
August 28, 2024
Prakash Chandra Kavi, Gorka Zamora-López, Daniel Ari Friedman
Cognition emerges from the dynamic interaction of self-organizing units of embodied knowledge called thoughtseeds within a Global Workspace of consciousness. This framework integrates evolutionary theory, neuronal packets, and the free energy principle into a hierarchical model of cognitive states comprising Neuronal Packet Domains, Knowledge Domains, the thoughtseed network, and meta-cognition. Nested Markov blankets and reciprocal message passing mediate this hierarchy, allowing thoughtseeds—fundamental units of thought—to compete for dominance, with the winning thoughtseed shaping conscious experience and guiding behavior. A mathematical framework grounded in active inference and dynamical systems theory models thoughtseed dynamics and their contribution to the unitary nature of consciousness.
arXiv Preprint Archive
July 25, 2024
Robert Worden
A wave-like excitation in animal brains may serve as a working spatial memory for representing three-dimensional space. The hypothesis is supported by evidence from multiple fields, including connectomics, computational modeling, experimental neuroscience, genomics, proteomics, animal behavior, and biophysics. If such a wave exists, it could plausibly be identified as the source of consciousness, potentially advancing understanding of the mind and reshaping views of human cognition.
arXiv Preprint Archive
July 14, 2024
Jingjin Li, Karen Anne Cochrane, Gilly Leshed
Experienced mindfulness practitioners use strategies like brief exercises, social accountability, and teacher guidance to sustain long-term practice. They adopt technologies such as mobile apps, wearables, and online sessions for reminders, emotion tracking, and community connection, but also deliberately limit use and repurpose technology as a focus for mindfulness. The findings expand the definition of mindfulness and highlight tensions between supporting short- and long-term practice. Design recommendations include using metaphor, reappropriating non-mindfulness technology, and integrating community support into personal practice.
arXiv Preprint Archive
July 9, 2024
Joseph Bodenheimer, Paul Bogdan, Sérgio Pequito et al.
The brain's large-scale dynamics change in distinct ways as people move between wakefulness, light sedation, deep sedation, and recovery. Using a model that treats the brain as a linear time-invariant system with unknown inputs, the authors show that the stability and frequency of oscillatory modes shift across these states. The same model identifies external drivers that shape brain activity during naturalistic auditory stimulation, revealing how stimulus-induced co-activity propagation differs across consciousness levels. The approach captures brain-wide changes that conventional methods miss, and these findings may help develop better biomarkers for consciousness recovery in disorders of consciousness.
arXiv Preprint Archive
June 10, 2024
Azenet Lopez, Carlos Montemayor
The Integrated Information Theory (IIT) is a leading scientific proposal for explaining phenomenal consciousness, but it overlooks the essential role of attention in generating and shaping conscious experience. Without an account of attention, IIT cannot explain informational differences between types of experiences. Although some IIT proponents claim a double dissociation between consciousness and attention, close analysis shows this dissociation is incompatible with IIT. These issues likely extend to other internalist and primitivist theories of conscious contents in philosophy, as well as to structuralist approaches. The discussion highlights that attention is indispensable for both scientific and philosophical theorizing about conscious experience.
arXiv Preprint Archive
June 3, 2024
Craig I. McKenzie
Consciousness is the apparatus enabling decision-making, defined by fundamental requirements rather than by observed behaviors. Drawing on current theories and neurological evidence, the author argues that consciousness requires at least some capacity for perception, a memory to store perceptual information that provides a framework for imagination, and a sense of self capable of making decisions based on possible and desired futures. The loss of any one component removes the capability for conscious thought. This definition allows objective assessment of consciousness in any agent, including non-human animals and artificial intelligence, without requiring choice behavior or explicit temporal awareness.
arXiv Preprint Archive
May 20, 2024
Robert Worden
A proposed projective wave theory of consciousness addresses three core difficulties: how neurons that cause consciousness are selected, how a detailed internal model of 3D space arises, and how distorted neural representations are decoded into an undistorted conscious experience. The theory posits that the brain's internal model of local 3D space is held not in neurons but in a wave excitation that carries a projective transform of Euclidean space, and that this wave is the source of spatial consciousness. Although such a wave has not yet been detected, indirect evidence exists in the mammalian thalamus and the insect brain's central body. The theory aligns with the spatial form of consciousness and has a positive Bayesian balance between its assumptions and the data it explains.
arXiv Preprint Archive
May 17, 2024
Xin Li
Consciousness emerges from a cycle-consistent, affectively anchored inference process recursively structured by the interaction of emotion and cognition. Emotion acts as a low-dimensional structural prior; cognition provides specificity-instantiating updates. This emotion-cognition cycle minimizes joint uncertainty by aligning emotionally weighted priors with context-sensitive cognitive appraisals. Subjective experience arises as the informational footprint of temporally extended, affect-modulated simulation. The Exchangeable Integration Theory of Consciousness (EITC) models conscious episodes as conditionally exchangeable samples drawn from a latent affective self-model. This latent variable supports integration via a unified cause-effect structure with nonzero irreducibility and differentiation by preserving contextual specificity. The framework connects to the Bayesian theory of consciousness through Rao-Blackwellized inference, which stabilizes inference by marginalizing latent self-structure while enabling adaptive updates, preventing inference collapse and supporting goal-directed simulation.