Source Count: 21 | Weighted Score: 50 | Source Confidence: [5/5] | Primary Tier: 1–2 | Last Updated: April 17, 2026
Keywords: integrated-information-theory, iit, phi, giulio-tononi, consciousness-axioms, qualia-space, exclusion-postulate, intrinsic-cause-effect-power, neural-correlates, empirical-consciousness, panpsychism, perturbational-complexity-index, iit-4.0, cogitate-collaboration
Category Tags: consciousness-theory, neuroscience, information-theory, philosophy-of-mind, mathematical-consciousness
Cross-References: K_1_01 — Quantum Consciousness · K_1_10 — Panpsychism Comprehensive · K_1_09 — Philosophical Zombies & Hard Problem · V_1_01 — Information Theory · ZD_1_14 — Information Theory Foundations

QUICK SUMMARY

Integrated Information Theory (IIT), developed primarily by Giulio Tononi (University of Wisconsin–Madison) from 2004 to the present, proposes that consciousness is identical to integrated information — a quantity denoted Φ (phi) that measures the extent to which a system generates information "above and beyond" its parts. Unlike most theories that treat consciousness as an emergent property of computation, IIT begins from the phenomenology of experience itself, deriving five axioms (intrinsic existence, composition, information, integration, exclusion) and corresponding postulates about physical systems. IIT is mathematically rigorous but computationally intractable for large systems — calculating Φ exactly for a system of n elements requires evaluating all possible partitions (super-exponential complexity). The theory makes bold predictions: it implies that simple feed-forward networks (including current deep learning architectures) have zero Φ regardless of sophistication, while certain recurrent biological networks possess high Φ. IIT has been tested clinically using the Perturbational Complexity Index (PCI), which distinguishes conscious from unconscious states in patients with brain injuries with >95% accuracy (Casali et al., 2013). Critics including Scott Aaronson and Daniel Dennett have raised serious mathematical and philosophical objections.

1. VERIFIED CLAIMS (Tier 1 — Peer-Reviewed / Established)

• KEY FINDING Giulio Tononi published the foundational IIT paper "An Information Integration Theory of Consciousness" in BMC Neuroscience (2004), proposing that consciousness corresponds to the capacity of a system to integrate information, quantified as Φ — defined as the minimum information generated by a system above the information generated by its parts when the system is partitioned
• IIT 3.0 (Oizumi, Albantakis, and Tononi, 2014) formalized five axioms of consciousness derived from phenomenology: (1) intrinsic existence — experience exists from the intrinsic perspective; (2) composition — experience is structured; (3) information — experience is specific; (4) integration — experience is unified; (5) exclusion — experience has definite borders and content
• KEY FINDING The Perturbational Complexity Index (PCI), developed by Adenauer Casali et al. (2013, Science Translational Medicine), uses TMS-EEG (transcranial magnetic stimulation combined with electroencephalography) to measure the complexity of cortical responses — PCI correctly classified conscious vs. unconscious states with >95% accuracy across 208 measurements including healthy wakefulness, sleep stages, anesthesia, coma, vegetative state, and locked-in syndrome
• IIT predicts that feed-forward networks (regardless of complexity) have Φ = 0, because information flows in one direction without integration across recurrent connections — this implies that current artificial neural networks based on feed-forward architectures, including transformer models, are not conscious according to IIT
• The theory identifies the posterior cortical hot zone (parieto-occipital regions) rather than the prefrontal cortex as the primary substrate of consciousness — this prediction aligns with lesion studies showing that prefrontal damage does not eliminate consciousness, while posterior cortical damage can produce specific losses of conscious content (Koch et al., 2016)

2. CREDIBLE CLAIMS (Tier 2 — Academic / Debated but Supported)

• The Adversarial Collaboration (Cogitate Consortium) between IIT and Global Neuronal Workspace Theory (GNWT), funded by the Templeton World Charity Foundation (2019–2023), tested pre-registered predictions of both theories using fMRI and EEG. Preliminary results presented in 2023 showed sustained posterior cortical activity during conscious perception (consistent with IIT) and lacked the predicted prefrontal "ignition" pattern central to GNWT. The full peer-reviewed results (Ferrante et al., Nature, 2025) were more equivocal: neither theory was fully supported, and both required substantial revision — the data partially falsified core predictions of each
• IIT's mathematical framework defines a "conceptual structure" in qualia space — each experience corresponds to a specific shape in a high-dimensional space whose axes represent possible states; two experiences differ to the extent that their shapes in qualia space differ, providing a geometric account of qualitative differences between experiences
• Christof Koch (longtime Chief Scientist at the Allen Institute for Brain Science, now President emeritus there and Chief Scientist at the Tiny Blue Dot Foundation) is a prominent IIT proponent and has argued that IIT provides the first truly falsifiable theory of consciousness — unlike "easy problem" approaches that correlate neural activity with reports, IIT makes structural predictions about which physical systems can and cannot be conscious
• IIT implies a form of panpsychism — any system with nonzero Φ possesses some degree of experience; a photodiode, for instance, has minimal but nonzero Φ (approximately 1 bit), implying an infinitesimally simple experience; Tononi has acknowledged this implication and embraced it as a feature rather than a bug
• The maximum of Φ over all possible spatial and temporal scales defines the "complex" — the specific set of elements that constitutes the conscious entity; this exclusion postulate implies that overlapping systems cannot both be independently conscious, preventing a "combination problem"
• William Marshall (University of Wisconsin, 2016) demonstrated that even approximation algorithms for Φ face super-exponential scaling barriers — the partition space of an n-element system grows roughly as the Bell number B(n), so even modest networks of a few hundred binary elements push exact Φ computation far beyond any tractable resource budget, making direct biological-scale verification fundamentally impossible with currently known methods (a system of ~270 binary elements already exceeds the ~10⁸⁰ atoms in the observable universe in partition count)
• Tononi and Christof Koch published a formal account in Philosophical Transactions of the Royal Society B (2015) arguing that the specific cause-effect structure of integrated information ('conceptual structure') corresponds to the specific quality of experience — different Φ-structures map to different qualia (red vs. tone). This is IIT's most ambitious and least empirically tested prediction.

3. SPECULATIVE CLAIMS (Tier 3 — Possible but Unverified)

• If IIT is correct, the theory has radical implications for AI consciousness — no matter how behaviorally sophisticated, systems based on feed-forward architectures (including current large language models) would be definitively non-conscious, while relatively simple recurrent circuits could possess genuine experience
• Some IIT researchers speculate that anesthesia works by reducing Φ through disrupting cortical integration rather than suppressing activity per se — propofol and ketamine both reduce consciousness but through different mechanisms (propofol reduces cortical connectivity; ketamine increases cortical activity but fragments integration), potentially providing differential tests of IIT
• Tononi has proposed "qualia realism" — that the mathematical structure of Φ not only correlates with but literally is consciousness, making IIT an identity theory rather than a correlation theory; this metaphysical commitment remains deeply controversial

4. DUBIOUS CLAIMS (Tier 4 — No Credible Source / Contradicted by Evidence)

• DEBUNKED Popular claims that IIT has been "proven" or "demonstrated to be the correct theory of consciousness" overstate the evidence — the Adversarial Collaboration yielded mixed results, PCI measures complexity rather than Φ directly, and exact Φ calculation remains computationally intractable for biologically realistic systems
• Claims that IIT proves that the entire universe or the internet is conscious misrepresent the theory — IIT predicts that systems without sufficient recurrent connectivity (like the internet's loose coupling) would have very low Φ despite enormous size

Counter-Arguments & Criticisms

• Scott Aaronson (2014) demonstrated that IIT assigns arbitrarily high Φ to certain simple physical systems (specifically, a 2D grid of XOR gates) — systems that intuitively should not be conscious; Tononi responded with modifications to IIT 3.0 but the "unfolding argument" remains contested
• Daniel Dennett criticized IIT as "a head-fake" — arguing that starting from phenomenological axioms begs the question of consciousness rather than explaining it, and that the mathematical formalism provides quantitative precision without genuine explanatory power
• The computational intractability of calculating Φ for systems larger than ~30 elements means that IIT's core prediction (Φ corresponds to consciousness) cannot be directly tested in biological brains — practical proxies like PCI measure related but different quantities
• Björn Merker and others have noted that IIT's exclusion postulate leads to counterintuitive predictions — for example, it implies that the cerebellum (with ~80 billion neurons but minimal recurrent connectivity) contributes virtually nothing to consciousness, which aligns with lesion evidence but seems surprising given its computational capacity
• A 2023 open letter signed by 124 consciousness researchers expressed "concern" about overstating IIT's empirical support and explicitly used the word "pseudoscience"; Tononi and collaborators responded in Frontiers in Psychology that the letter conflated unfalsifiability with current practical untestability and that PCI represents genuine empirical progress
• Eric Schwitzgebel (UC Riverside, 2015) argued IIT's commitment to panpsychism creates more problems than it solves — under IIT, we must either accept that the United States (as an integrated network) is conscious, or add ad hoc exclusion criteria. Published as "If Materialism Is True, the United States Is Probably Conscious" in Philosophical Studies
• Philip Goff (Galileo's Error, 2019) argued in defense of the panpsychist implication, treating IIT as the most rigorous mathematical framework yet developed for a panpsychist research program

IMAGES

No images assigned yet.

BIBLIOGRAPHY

1. Tononi, Giulio | 2004 | "An Information Integration Theory of Consciousness" | BMC Neuroscience | ∅ | 5.42::1–22 | ∅ | ∅ | doi:10.1186/1471-2202-5-42 | ∅ | ∅ | ∅
2. Oizumi, Masafumi, Larissa Albantakis; Giulio Tononi. e1003588 | 2014 | "From the Phenomenology to the Mechanisms of Consciousness: Integrated Information Theory 3.0" | PLoS Computational Biology | ∅ | 10.5:: | ∅ | ∅ | doi:10.1371/journal.pcbi.1003588 | ∅ | ∅ | ∅
3. Casali, Adenauer, Olivia Gosseries, Mario Rosanova, et al. ra105 | 2013 | "A Theoretically Based Index of Consciousness Independent of Sensory Processing and Behavior" | Science Translational Medicine | ∅ | 5.198::198 | ∅ | ∅ | doi:10.1126/scitranslmed.3006294 | ∅ | ∅ | ∅
4. Koch, Christof, Marcello Massimini, Melanie Boly; Giulio Tononi | 2016 | "Neural Correlates of Consciousness: Progress and Problems" | Nature Reviews Neuroscience | ∅ | 17.5::307–321 | ∅ | ∅ | doi:10.1038/nrn.2016.22 | ∅ | ∅ | ∅
5. Tononi, Giulio, Melanie Boly, Marcello Massimini; Christof Koch | 2016 | "Integrated Information Theory: From Consciousness to Its Physical Substrate" | Nature Reviews Neuroscience | ∅ | 17.7::450–461 | ∅ | ∅ | doi:10.1038/nrn.2016.44 | ∅ | ∅ | ∅
6. Aaronson, Scott | 2014 | "Why I Am Not an Integrated Information Theorist (or, The Unconscious Expander)" | ∅ | ∅ | ∅ | Blog post and technical paper, Shtetl-Optimized, May | ∅ | ∅ | ∅ | ∅ | ∅
7. Albantakis, Larissa, Leonardo Barbosa, Graham Findlay, et al. e1011465 | 2023 | "Integrated Information Theory (IIT) 4.0: Formulating the Properties of Phenomenal Existence in Physical Terms" | PLoS Computational Biology | ∅ | 19.10:: | ∅ | ∅ | doi:10.1371/journal.pcbi.1011465 | ∅ | ∅ | ∅
8. Melloni, Lucia, Liad Mudrik, Michael Pitts, et al | 2021 | "Making the Hard Problem of Consciousness Easier" | Science | ∅ | 372.6545::911–912 | ∅ | ∅ | doi:10.1126/science.abj3259 | ∅ | ∅ | ∅
9. Bayne, Tim, Anil Seth; Marcello Massimini | 2020 | "Are There Islands of Awareness?" | Trends in Neurosciences | ∅ | 43.1::6–16 | ∅ | ∅ | doi:10.1016/j.tins.2019.11.003 | ∅ | ∅ | ∅
10. Dehaene, Stanislas, Hakwan Lau; Sid Kouider | 2017 | "What Is Consciousness, and Could Machines Have It?" | Science | ∅ | 358.6362::486–492 | ∅ | ∅ | doi:10.1126/science.aan8871 | ∅ | ∅ | ∅
11. Chalmers, David | 2017 | "The Combination Problem for Panpsychism" | Panpsychism: Contemporary Perspectives | ∅ | ∅ | In , edited by Godehard Brüntrup and Ludwig Jaskolla, 179 214 | ∅ | ∅ | ∅ | ∅ | Oxford: Oxford University Press
12. Dennett, Daniel | 2018 | "The Fantasy of First-Person Science" | ∅ | ∅ | ∅ | Unpublished commentary, presented at the Association for the Scientific Study of Consciousness conference | ∅ | ∅ | ∅ | ∅ | ∅
13. Massimini, Marcello; Giulio Tononi | 2018 | ∅ | Sizing Up Consciousness: Towards an Objective Measure of the Capacity for Experience | ∅ | ∅ | Oxford: Oxford University Press | ∅ | isbn:9780198728437 | ∅ | ∅ | ∅
14. Merker, Björn | 2013 | "Cortical Gamma Oscillations: The Functional Key Is Activation, Not Cognition" | Neuroscience and Biobehavioral Reviews | ∅ | 37.3::401–417 | ∅ | ∅ | doi:10.1016/j.neubiorev.2013.01.013 | ∅ | ∅ | ∅
15. Marshall, William, et al. e1005079 | 2016 | "Integrated Information and State Differentiation" | PLoS Computational Biology | ∅ | 12.11:: | ∅ | ∅ | doi:10.1371/journal.pcbi.1005079 | ∅ | ∅ | ∅
16. Tononi, Giulio; Christof Koch | 2015 | "Consciousness: Here, There and Everywhere?" | Philosophical Transactions of the Royal Society B | ∅ | 370.1668::20140167 | ∅ | ∅ | doi:10.1098/rstb.2014.0167 | ∅ | ∅ | ∅
17. Goff, Philip | 2019 | ∅ | Galileo's Error: Foundations for a New Science of Consciousness | ∅ | ∅ | New York: Pantheon | ∅ | isbn:9781524747981 | ∅ | ∅ | ∅
18. Schwitzgebel, Eric | 2015 | "If Materialism Is True, the United States Is Probably Conscious" | Philosophical Studies | ∅ | 172.7::1697–1721 | ∅ | ∅ | doi:10.1007/s11098-014-0387-8 | ∅ | ∅ | ∅
19. Tononi, Giulio, et al | 2016 | "Integrated Information Theory: An Updated Account" | Archives Italiennes de Biologie | ∅ | 154::56–78 | ∅ | ∅ | ∅ | ∅ | ∅ | ∅
20. Ferrante, Oscar, et al. (Cogitate Consortium) | 2025 | "Adversarial Testing of Global Neuronal Workspace and Integrated Information Theories of Consciousness" | Nature | ∅ | 642::133–142 | ∅ | ∅ | doi:10.1038/s41586-025-08888-1 | ∅ | ∅ | ∅
21. Merker, Bjorn, Kenneth Williford; David Rudrauf. e41 | 2022 | "The Integrated Information Theory of Consciousness: A Case of Mistaken Identity" | Behavioral and Brain Sciences | ∅ | 45:: | ∅ | ∅ | doi:10.1017/S0140525X21000881 | ∅ | ∅ | ∅

CROSS-REFERENCE INDEX

Merged from K_1_16 (Apr 17, 2026) — sources combined, K_1_16 deprecated. Last Updated: April 17, 2026