What we mean by "consciousness"
Phenomenal consciousness: the qualitative, subjective "what it is like" aspect of experience (pain, red, tasting coffee). Often called qualia.[1]
Access consciousness: information in the mind that is available for reasoning, reporting, decision-making, and control of behavior.[2]
Self-consciousness / reflective consciousness: the capacity to be the object of one’s thoughts (thinking “I am thinking this”).[3]
Minimal phenomenal experience / core awareness: simpler forms such as raw perception and wakeful awareness that do not require higher cognition.[4]
These distinctions help avoid conflating different questions: e.g., “How do we report an experience?” (access) vs. “Why does experience feel like anything?” (phenomenal).
Short history of thought about consciousness: ancient to modern, then 19th–20th century scientific developments, then late-20th / 21st century revolutions.
A. Ancient and medieval
Early thinkers (Indian, Greek, Chinese) treated consciousness within broader metaphysics and psychology: thinkers like Upanishadic authors, Buddhism, Aristotle, and early Chinese and Hebrew traditions discussed mind, perception, and self in ethical and cosmological contexts.[5][6]
Aristotle distinguished senses, imagination, and thought; some strands held that the soul/mind was distinct from the body, others treated them as integrated.[6]
B. Early modern (17th–18th centuries)
Descartes (dualism): famously separated res cogitans (thinking substance) from res extensa (extended matter). Conscious thought and subjective experience were taken as bedrock certainty (“I think, therefore I am”) but posed the mind–body interaction problem.[7]
Empiricists (Locke, Hume): emphasized experience as origin of ideas; Locke introduced the term “consciousness” in English to describe awareness of mental states; Hume was skeptical about a persistent, unchanging self.[8]
C. 19th century: psychology emerges
Structuralists (Wundt, Titchener) used introspection to study conscious experience’s elements; debates about reliability ensued.[9]
William James offered influential functionalist and phenomenological accounts: consciousness as a continuous “stream,” emphasising function and the adaptive roles of attention, memory, emotion.[10]
D. Early–mid 20th century: behaviourism, cognitive revolution
Radical behaviorism (Watson, later Skinner) downplayed or dismissed consciousness as unscientific because it is private and not publicly observable; the focus shifted to behavior and stimulus–response laws.[11]
The cognitive revolution (1950s–70s) reintroduced internal mental representations and processes; models of information processing and computation came to the fore.[12]
Philosophers like Gilbert Ryle criticized Cartesian dualism as a “ghost in the machine”; others explored functionalist accounts of mental states.[13]
E. Late 20th century: analytic philosophy and the "hard problem"
Thomas Nagel’s 1974 paper “What Is It Like to Be a Bat?” emphasized that subjective character resists purely objective description — the “what it’s like” point.[14]
Frank Jackson’s knowledge argument (Mary’s room) and other thought experiments emphasized that physical descriptions might leave out phenomenal qualities.[15]
David Chalmers (1995) popularized the distinction between the “easy problems” (explaining functions like discrimination, report, control) and the “hard problem” (why and how physical processes give rise to subjective experience).[16]
F. 21st century: interdisciplinary empirical and theoretical work
Neuroscience advanced with imaging, lesion studies, and manipulations (TMS, optogenetics), allowing correlations between brain states and conscious reports; new efforts try to identify neural correlates of consciousness (NCC).[17]
Theoretical frameworks (global workspace, integrated information, predictive processing, higher-order thought) aim to explain access and/or phenomenal consciousness.[18][19][20]
Debates about whether consciousness is reducible to neural processes, emergent, or fundamental to reality continue; panpsychism and physicalist emergentism have gained renewed attention.[21]
Contemporary theories — categories, core ideas, evidence, critiques I group current theories into families. Each family addresses different aspects (access vs. phenomenal), and some theories aim to do both.
A. Global Workspace Theories (GWT)
Core idea: Consciousness is the broadcasting of information to a global workspace — a set of cognitive systems (working memory, attention, language, planning) that access and use information once it is globally available. Nonconscious processes operate in parallel; consciousness integrates and forwards selected information for flexible control.[22]
Leading proponents: Bernard Baars (original GWT), Stanislas Dehaene and collaborators (neuronal global workspace, NGW).[22][23]
Evidence: Neuroimaging and EEG studies show late, sustained, wide-spread frontoparietal activations associated with reportable stimuli; P3 wave (a late EEG signal) correlates with conscious report in many paradigms.[24]
Strengths: Explains many “easy” problems: reportability, attention, working memory, decision-related functions. It makes testable neural predictions.
Weaknesses: Critics argue it may explain access consciousness but still leave phenomenal “what-it’s-like” unexplained; some experiments show dissociation between global frontoparietal activity and certain forms of vivid subjective experience (e.g., some sensory qualia might persist without frontoparietal activation), raising questions about sufficiency/necessity claims.[25]
B. Higher-Order Thought (HOT) and Higher-Order Representation theories
Core idea: A mental state is conscious if there is a higher-order representation (a thought or perception) that one is in that state. Consciousness arises when the system represents that it is having a first-order state.[26]
Proponents: David Rosenthal, others.
Evidence: Explains self-report and metacognition: conscious states involve an additional representational layer.
Strengths: Accounts for the introspective and reflective aspects of consciousness and explains the distinction between conscious and unconscious mental states.
Weaknesses: Critics ask whether higher-order states themselves need to be conscious, potentially causing regress, and whether HOTs capture raw phenomenology (qualia).[27]
C. Integrated Information Theory (IIT)
Core idea: Consciousness corresponds to integrated information — phi (Φ) — a quantity measuring how much a system is unified and informative above its parts. Systems with high Φ have richer phenomenal states; IIT tries to connect phenomenological axioms to a mathematical measure derived from system causal structure.[28]
Proponent: Giulio Tononi and collaborators.
Evidence: IIT generates testable predictions (e.g., loss of integration during anesthesia correlates with loss of consciousness; certain cortical areas contribute more). Empirical work finds reduced measures of integration in unconscious states.[29]
Strengths: Attempts to explain why experience is unified and quantifies consciousness; offers a principled, quantitative framework.
Weaknesses: Hard to compute Φ for realistic brains; critiques argue that IIT can ascribe consciousness to systems intuitively not conscious (panpsychist implications) and that Φ may not track subjective experience in all cases. Also, mapping mathematical structures to specific qualitative experiences is controversial.[30]
D. Predictive Processing and Bayesian Brain accounts
Core idea: The brain is a hierarchical prediction machine: it continually generates predictions about sensory inputs and updates beliefs by minimizing prediction errors. Conscious perception arises when prediction models are revised or when precision (confidence) of prediction errors is high, making information available for cognitive access.[31]
Proponents: Karl Friston and many computational neuroscientists; active inference frameworks extend this to action and motivation.
Evidence: Explains many perceptual illusions and suggests mechanisms for attention and expectation effects on perception; some modeling links predictive dynamics to conscious reportability.
Strengths: Integrates perception, attention, learning; gives mechanistic computational story. Offers principled explanation for expectation effects and top-down influence on perception.
Weaknesses: It's sometimes vague about phenomenal character; critics argue it needs additional specification to explain why certain prediction dynamics produce subjective feeling. Different implementations vary widely, making empirical falsification tricky.[32]
E. Recurrent Processing Theory (RPT)
Core idea: Recurrent or recurrently amplified neural processing in sensory areas (feedback loops) is essential for conscious perception; feedforward sweep can support nonconscious processing but recurrent activity enables the vivid, reportable experience.[33]
Proponent: Victor Lamme and others.
Evidence: Time-course EEG/MEG and masking experiments show early feedforward activation can occur without report, whereas recurrent activity correlates with awareness.[34]
Strengths: Makes clear temporal predictions and is more localist (sensory cortex) than global workspace approaches; supported by masking and timing data.
Weaknesses: Critics argue that recurrent local activity alone may not explain reportability and higher cognitive integration; some phenomena show wide frontoparietal involvement even when recurrent sensory processing is present.
F. Higher-order and First-order hybrids; Multiple Drafts Model
Daniel Dennett’s "Multiple Drafts" rejects a central theater for consciousness. Consciousness consists of many parallel drafts (interpretations) with no privileged inner observer; report depends on narrative and cognitive access.[35]
Strengths: Demystifies the need for a Cartesian theater; emphasizes distributed processing and language-mediated reporting.
Weaknesses: Critics say it downplays the felt aspect of experience and has trouble explaining immediacy/phenomenal quality.
G. Panpsychism and Constitutive Panpsychism
Core idea: Consciousness (or proto-consciousness) is a fundamental property of matter; complex consciousness arises from combinations of micro-level proto-experiences.[36]
Proponents: Galen Strawson, Philip Goff, integrated by some interpretations of IIT.
Strengths: Offers a way to dissolve the hard problem by positing consciousness as fundamental, avoiding emergence from purely non-experiential matter.
Weaknesses: Hard to test empirically; raises combination problems (how micro-experiences combine into unified macroscopic consciousness), and can seem counterintuitive.
H. Biological and Enactivist approaches
Core idea: Consciousness is an embodied, enactive process grounded in sensorimotor loops, organism–environment coupling, and biological autonomy rather than merely computation. Consciousness arises from an organism’s active engagement and normative life processes.[37]
Proponents: Francisco Varela, Evan Thompson, and some philosophers of mind and cognitive science.
Strengths: Connects consciousness to real biological function and behavior; highlights the role of action, emotion, and embodiment.
Weaknesses: Can be underspecified about neural mechanisms; critics want clearer mapping to neural data and how subjective experience arises.
I. Quantum and exotic theories (less mainstream)
Some propose quantum effects in microtubules or quantum collapse play roles in consciousness (e.g., Penrose–Hameroff Orch OR). These are highly controversial and lack robust empirical support.[38]
Most neuroscientists regard classical neural computation as sufficient unless specific testable quantum mechanisms are demonstrated.
Neural correlates of consciousness (NCC)
Definition: Minimal neural mechanisms jointly sufficient for a specific conscious content.[39]
Empirical strategy: Compare brain states when stimulus is consciously reported vs. not, controlling for stimulus and behavior. Methods include fMRI, EEG/MEG, single-unit recordings, lesion studies, and perturbations (TMS, optogenetics).
Key findings:
Early sensory responses (
100 ms) often dissociate from awareness; later widespread activity (200–300 ms and beyond) in cortex, especially frontoparietal regions, correlates with reportable awareness in many paradigms.[24][34]Local recurrent processing in sensory cortices is necessary for some conscious contents; global ignition or broadcast across networks correlates with report and flexible access.[22][33]
States like deep sleep, anesthesia, and vegetative states show reductions in measures of integration and complexity; some metrics (perturbational complexity index) predict levels of consciousness.[40]
Caution: Correlates are not automatically explanations. Determining sufficiency/necessity requires causal interventions (stimulation, lesions) and careful experimental design. Also, differences exist between content-specific NCCs (what makes a particular perception conscious) vs. full NCCs (what supports being conscious at all).
The hard problem and philosophical responses
The hard problem: Why should physical processes give rise to subjective experience at all? Why is there something it is like to be a system with those processes?[16]
Responses:
Physicalist reductions: phenomenal properties will eventually be explained in physical terms (many philosophers/neuroscientists optimistic).
Eliminativism: deny phenomenal properties or recast them as mischaracterized cognitive states (controversial).
Dualism: accept irreducible mental substance/properties (traditional Cartesian versions are less popular in science).
Panpsychism: see above.
Acceptance of explanatory gaps: some adopt a pragmatic stance that science progresses even if the hard problem remains philosophically thorny.
Empirical tests and progress
Experimental paradigms that advanced the field:
Binocular rivalry and multistable perception: same stimulus leads to alternations in conscious content, useful to dissociate sensory input from awareness.[41]
Masking paradigms (backward masking): reveal timing and processing necessary for awareness.[34]
No-report paradigms: study brain signals linked to experience that do not depend on explicit report (addresses confound that report processes cause some neural signals).[42]
Perturbational approaches (TMS, optogenetics): causally manipulate circuits to test necessity and sufficiency of networks for awareness.[43]
Clinical and translational work:
Consciousness assessment in disorders of consciousness uses behavioral scales, imaging, and perturbational complexity indexes to detect covert awareness and prognosis.[40]
Anesthesia studies probe how different drug classes disrupt consciousness and associated network measures.
Where the field is now — consensus and open questions
Consensus points:
Consciousness is tightly linked to brain activity; certain neural signatures track conscious states.
There is a useful distinction between access and phenomenal aspects; many theories cleanly handle one better than the other.
Research is increasingly empirical and mechanistic, combining theory, measurement, and intervention.
Open and contested issues:
The hard problem: subjective experience’s ultimate metaphysical status is unresolved.
The precise neural mechanisms — local recurrent processing vs. global broadcasting vs. integrated information — remain debated; different paradigms favor different mechanisms.
Measuring consciousness quantitatively across species and systems is challenging; IIT and alternatives propose metrics, but no single metric is universally accepted.
The criteria for attributing consciousness to nonhuman animals, fetuses, and AI are unsettled and have ethical implications.
Practical implications
Medicine: improving diagnosis/treatment of disorders of consciousness and anesthesia monitoring.
AI and ethics: debates about whether advanced AI could be conscious and how to detect it.
Neuroscience and psychiatry: understanding consciousness informs models of perception, attention, and mental disorders.
Suggested reading and representative sources (footnotes)
Nagel T. “What is it like to be a bat?” The Philosophical Review. 1974;83(4):435–50.
Block N. “On a confusion about a function of consciousness.” Behavioral and Brain Sciences. 1995;18(2):227–47.
Metzinger T. “The Ego Tunnel.” (2009). Explores self-models and self-consciousness.
Damasio AR. “The Feeling of What Happens: Body and Emotion in the Making of Consciousness.” (1999).
Varela FJ, Thompson E, Rosch E. “The Embodied Mind: Cognitive Science and Human Experience.” (1991).
Robinson TJ. “Aristotle and the Philosophy of Mind.” (1998). (For historical context on ancient thought.)
Descartes R. “Meditations on First Philosophy.” 1641.
Locke J. “An Essay Concerning Human Understanding.” 1690. Hume D. “A Treatise of Human Nature.” 1739–40.
Wundt W. “Principles of Physiological Psychology.” 1874.
James W. “The Principles of Psychology.” (1890).
Watson J. “Psychology as the Behaviorist Views It.” Psychological Review. 1913.
Newell A, Simon HA. “Computer science as empirical inquiry: Symbols and search.” Communications of the ACM. 1976. (Context on cognitive modeling)
Ryle G. “The Concept of Mind.” 1949.
Nagel 1974 (above).
Jackson F. “Epiphenomenal Qualia.” 1982; and his “What Mary Didn’t Know” later papers.
Chalmers DJ. “The Conscious Mind: In Search of a Fundamental Theory.” (1996).
Koch C, et al. “Neural correlates of consciousness.” Nat Rev Neurosci. 2016;17(5):307–321.
Baars BJ. “A Cognitive Theory of Consciousness.” 1988. Dehaene S, Changeux JP. “Experimental and theoretical approaches to conscious processing.” Neuron. 2011;70(2):200–27.
Tononi G. “An information integration theory of consciousness.” BMC Neuroscience. 2004;5:42. Tononi G, Grossberg S. Follow-ups and IIT developments.
Friston K. “The free-energy principle: a unified brain theory?” Nat Rev Neurosci. 2010;11:127–138. (Predictive processing)
Goff P. “Consciousness and Fundamental Reality.” 2017 (panpsychist defense).
Baars BJ. (above)
Dehaene S. “Consciousness and the Brain: Deciphering How the Brain Codes Our Thoughts.” (2014).
Sergent C, Dehaene S. “Is consciousness a gradual phenomenon?” Trends Cogn Sci. 2004. P3 literature: Polich J. “Updating P300.” 2007.
Koch C, Tsuchiya N. “Attention and consciousness: two distinct brain processes.” Trends Cogn Sci. 2007.
Rosenthal DM. “A Theory of Consciousness.” 2005.
See critiques collected in commentaries on HOT theories (e.g., Gennaro, 2012).
Tononi 2004 (above); Oizumi M, Albantakis L, Tononi G. “From the phenomenology to the mechanisms of consciousness.” PNAS. 2014.
Mashour GA, et al. “Conscious processing and the global neuronal workspace hypothesis.” Neuron. 2020. (Discusses empirical tests and predictions.)
Critics of IIT: Bayne T., Massimini M., and others in various commentaries; see also Koch, Massimini & Tononi debates.
Clark A. “Surfing Uncertainty: Prediction, Action, and the Embodied Mind.” 2016. Friston K papers.
Hohwy J. “The Predictive Mind.” (2013) for detailed defense; see also critiques by Michel & Morales on tying to phenomenology.
Lamme VAF. “Why visual attention and awareness are different.” Trends Cogn Sci. 2003. Lamme’s recurrent processing work.
For masking and timing: Dehaene, Sergent, Lamme literature; Koivisto & Revonsuo reviews.
Dennett DC. “Consciousness Explained.” 1991.
See Goff P. (above) and Strawson G. “Realistic monism: why physicalism entails panpsychism.” (2006)
Varela, Thompson & Rosch (1991); Thompson E. “Mind in Life” (2007).
Penrose R, Hameroff S. “Orchestrated Objective Reduction” (Orch OR) — see reviews and strong critiques in neuroscience literature.
Crick F, Koch C. “What is the function of the claustrum?” Phil Trans R Soc. 2005; also early NCC framing by Crick & Koch.
Casali AG, et al. “A theoretically based index of consciousness independent of sensory processing and behavior.” Science Translational Medicine. 2013. (Perturbational Complexity Index)
Blake R, Logothetis NK. “Visual competition.” Nat Rev Neurosci. 2002. (Binocular rivalry).
Tsuchiya N, Wilke M, Frassle S, Lamme VAF. “No-report paradigms.” Trends Cogn Sci. 2015.
Studies using TMS and optogenetics in consciousness research — see reviews by Dehaene, Koch, and Mashour.
Quick comparison table (conceptual)
GWT / NGW: explains reportability and cognitive integration; strong neuro evidence; may not fully solve qualia.
IIT: attempts quantitative mapping to experience; ambitious but technically and philosophically contested.
RPT: local recurrent processing explains perceptual awareness; well-supported temporally.
Predictive processing: a powerful computational framework; needs specifics to address phenomenal character.
HOT: explains introspection and metacognition; struggles with qualia’s felt quality.
Panpsychism: sidesteps hard problem by changing basic ontology; faces combination and testability issues.
Final thoughts
Consciousness research is pluralistic: multiple frameworks currently coexist because they emphasize different phenomena and use different methods. This pluralism is fruitful: empirical tests increasingly discriminate among predictions (timing, loci, necessity).
Progress is most likely to come from combined strategies: careful experimental paradigms that dissociate access/report from subjective feeling, causal perturbation, theoretical rigor (mathematical measures where possible), and philosophical clarity about what counts as an explanation.