The Problem with Phi: A Critique of Integrated Information Theory
Open accessThe DOI matches its Crossref record; 59 of the 73 references it cites were found in a registry or an index; 21 of the 110 sources its claims cite were checked for support.
Abstract
In the last decade, Guilio Tononi has developed the Integrated Information Theory (IIT) of consciousness.IIT postulates that consciousness is equal to integrated information (F).The goal of this paper is to show that IIT fails in its stated goal of quantifying consciousness.The paper will challenge the theoretical and empirical arguments in support of IIT.The main theoretical argument for the relevance of integrated information to consciousness is the principle of information exclusion.Yet, no justification is given to support this principle.Tononi claims there is significant empirical support for IIT, but this is called into question by the creation of a trivial theory of consciousness with equal explanatory power.After examining the theoretical and empirical evidence for IIT, arguments from philosophy of mind and epistemology will be examined.Since IIT is not a form of computational functionalism, it is vulnerable to fading/ dancing qualia arguments.Finally, the limitations of the phenomenological approach to studying consciousness are examined, and it will be shown that IIT is a theory of protoconsciousness rather than a theory of consciousness.
Overview
CoChat Summary
TL;DR
Cerullo argues that Integrated Information Theory (IIT) fails in its stated goal of quantifying consciousness because its foundational principle of information exclusion rests only on Tononi's intuition, and because a deliberately trivial rival theory (Circular Coordinated Message Theory, CCMT) can reproduce IIT's key empirical predictions. The paper concludes that IIT is best understood as a theory of proto- or noncognitive consciousness (partial-panexperientialism) addressing the "Pretty Hard Problem" rather than consciousness as studied by most neuroscientists.
Key findings
- The principle of information exclusion — that consciousness is generated by ruling out alternatives — is the sole theoretical foundation of IIT, yet "no valid arguments outside of Tononi's intuition are advanced to support" it; at most, granting it would make integrated information necessary but not sufficient for consciousness.
- CCMT, built on an equally "self-evident" principle of information circulation (Omicron, O), makes the same predictions as IIT for split-brain syndrome and the cerebellum, which the author argues undercuts IIT's claimed explanatory power.
- Because IIT rejects computational functionalism (Tononi: "whether a system is conscious or not cannot be decided based on its input-output behavior only"), it is vulnerable to fading/dancing qualia arguments, and is "especially vulnerable" given its axiom of the unity of conscious experience.
- Aaronson and Tononi agreed that an n × n grid of XOR gates can generate √n Φ, so arbitrarily large Φ can arise in systems performing no complex information processing — suggesting IIT measures protoconsciousness, not cognitive consciousness.
Methods
- Conceptual/philosophical critique, not an empirical study: the paper analyzes IIT's theoretical foundations (information exclusion, the axioms of IIT 3.0), its claimed empirical support (split-brain, cerebellum, EEG correlates, effective connectivity), and its relationship to functionalism and panpsychism.
- Constructs a counter-theory, CCMT, as a demonstration that IIT's predictions are trivially reproducible, and draws on philosophy of mind arguments (fading/dancing qualia, Chalmers' hard vs. easy problem, Block's A/P consciousness distinction) plus the published Aaronson–Tononi exchange.
- No sample, dataset, or statistical analysis is reported.
Limitations
- The paper states no formal limitations section; it acknowledges that neither computational functionalism nor identity theory solves the hard problem of consciousness, and that the debate over panpsychism/panexperientialism is unresolved ("we do not need to resolve the debate... to understand some differences").
- Evident from the text: the central argument against information exclusion rests on the author's own contrary intuition rather than new empirical data, and CCMT is offered as a rhetorical counterexample rather than a developed theory.
- Evident from the text: the taxonomy of four consciousness types (Chalmers' primitive protoconsciousness, Rosenberg's noncognitive-consciousness, Block's access- and phenomenal-cognitive-consciousness) is the author's synthesis, and the claim that IIT is "irrelevant" to cognitive consciousness depends on that framing.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
9 Valid · 0 Not valid · 29 Not checked · 35 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“In either case, IIT appears to be identifying consciousness with a specific measure of information.”
§ The Problem with Phi: A Critique of Integrated Information Theory > IIT and Functionalism
What the sources say
- Partly supportedcites [1] G. Tononi, 2003, Measuring integrated information, BMC Neuroscience sourceExperimental — accuracy not yet measured
The Background defines Φ as a specific measure of information integration, and the Discussion links high Φ to the generation of consciousness. But the Discussion presents that link conditionally; it does not identify consciousness itself with Φ. (LLM, conf=0.82)
- Supportedcites [2] G. Tononi, 2004, An informational integration theory of consciousness, BMC Neuroscience source
In the Presentation of the hypothesis, the paper says consciousness corresponds to the capacity to integrate information and that its quantity is given by a complex’s Φ value. This supports the citing sentence’s hedged characterization; the paper separately describes informational relationships as determining the quality of consciousness. (LLM, conf=0.92)
- Not checked — no full text of the sourcecites [3] G. Tononi, 2008, Consciousness as integrated information: A provisional manifesto, Biol. Bull source
- Supportedcites [4] G. Tononi, 2008, The neural correlates of consciousness An Update, Ann N.Y. Acad. Sci source
In “Consciousness as Integrated Information,” the paper says that a system is conscious in proportion to its integrated information and introduces a specific measure to quantify it. This supports the citing sentence’s hedged characterization of IIT. (LLM, conf=0.92)
- Partly supportedcites [5] D. Balduzzi, 2009, Qualia: The geometry of integrated information, Computational Biology sourceExperimental — accuracy not yet measured
The Introduction says IIT identifies the quantity of consciousness with the amount of integrated information and describes Φ as a measure of it. But it says the quality of consciousness is specified by informational relationships, not by that single measure, so the claim is too broad if it refers to consciousness as a whole. (LLM, conf=0.84)
- Not checked — no full text of the sourcecites [6] C. Koch, 2011, A Test for Consciousness, Scientific American source
- Not checked — no full text of the sourcecites [7] G. Tononi, 2012, Integrated information theory of consciousness: an updated account, Arch Ital Biol doi:10.4449/aib.v149i5.1388
- Partly supportedcites [8] M. Oizumi, 2014, From the phenomenology to the mechanisms of consciousness: Integrated information theory 3.0, Computational Biology sourceExperimental — accuracy not yet measured
The Introduction and Abstract say IIT uses a value of integrated information to specify the quantity of an experience. They also say its quality is specified by a maximally irreducible conceptual structure (MICS). Thus, the paper does not identify consciousness in its entirety with a single measure of information. (LLM, conf=0.86)
hedgedstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Fading and dancing qualia thought experiments show that identity and non-computational functionalist theories allow the unity of consciousness to be interrupted at any arbitrary point [22].”
§ The Problem with Phi: A Critique of Integrated Information Theory > IIT and Functionalism
What the sources say
- Partly supportedcites [22] M. Cerullo, 2015, Uploading and Branching Identity, Minds and Machines sourceExperimental — accuracy not yet measured
In “Machine Consciousness” (page unavailable), the paper says these qualia thought experiments could interrupt the unity of conscious experience at any point under identity theory. It does not extend that conclusion to non-computational functionalist theories; instead, it says the experiments support functionalism and present no obstacles to it. (LLM, conf=0.89)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“The principal theoretical justification for IIT is the conjecture that, in perception, what is absent may be more significant to consciousness than what is present [2,3].”
§ The Problem with Phi: A Critique of Integrated Information Theory > The Theoretical Foundation of IIT
What the sources say
- Partly supportedcites [2] G. Tononi, 2004, An informational integration theory of consciousness, BMC Neuroscience sourceExperimental — accuracy not yet measured
In “Presentation of the hypothesis” (page unavailable), the seeing-light example says a conscious percept rules out an extraordinarily large number of alternatives, supporting the point about what is absent. But the paper presents both information and its integration as central to the theory; it does not identify exclusion of absent possibilities as IIT’s principal theoretical justification or say they are more significant than what is present. (LLM, conf=0.84)
- Not checked — no full text of the sourcecites [3] G. Tononi, 2008, Consciousness as integrated information: A provisional manifesto, Biol. Bull source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“However, this research is merely consistent with IIT rather than evidence in support of the theory.”
§ The Problem with Phi: A Critique of Integrated Information Theory > The Explanatory Power of IIT
What the sources say
- Supportedcites [17] J. King, 2013, Information Sharing in the Brain Indexes Consciousness in Noncommunicative Patients, Current Biology source
The Results report an EEG study testing whether brain-scale information sharing distinguishes patients with and without clinical signs of consciousness. The Discussion reports that information sharing increased with consciousness state. The judgment that this is merely consistent with IIT, rather than evidence for it, is the citing authors’ own assessment, not a finding attributed to this paper. (LLM, conf=0.90)
- Not checked — no full text of the sourcecites [18] M Massimini, 2005, Breakdown of Cortical Effective Connectivity During Sleep, Science source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“Tononi’s response is not surprising given that IIT implies a form of panpsychism [1–8].”
§ The Problem with Phi: A Critique of Integrated Information Theory > The Epistemology of IIT
What the sources say
- Could not verifycites [1] G. Tononi, 2003, Measuring integrated information, BMC Neuroscience sourceExperimental — accuracy not yet measured
The Discussion section, “Information integration and basic neuroanatomy” (page unavailable), proposes high information integration as a requirement for consciousness and discusses the thalamocortical system. It does not say that consciousness is widespread or that the theory implies panpsychism. Searches for panpsychism and consciousness in non-neural systems found no passage establishing that implication. (LLM, conf=0.72)
- Supportedcites [2] G. Tononi, 2004, An informational integration theory of consciousness, BMC Neuroscience source
In “Implications of the hypothesis,” Tononi says experience is a fundamental quantity and that any physical system has subjective experience to the extent that it integrates information. The section also says consciousness is graded and should exist to varying degrees in most natural and artificial systems. This supports the characterization as a form of panpsychism, with information integration as the condition. (LLM, conf=0.92)
- Not checked — no full text of the sourcecites [3] G. Tononi, 2008, Consciousness as integrated information: A provisional manifesto, Biol. Bull source
- Supportedcites [4] G. Tononi, 2008, The neural correlates of consciousness An Update, Ann N.Y. Acad. Sci source
In “Some Implications” (page not supplied), Tononi says that any physical system with some capacity for integrated information would have some degree of experience, regardless of what it is made of. This supports describing IIT as implying a form of panpsychism, with the qualification that the system must have capacity for integrated information. (LLM, conf=0.92)
- Supportedcites [5] D. Balduzzi, 2009, Qualia: The geometry of integrated information, Computational Biology source
The Introduction says IIT ties the quantity of consciousness to integrated information generated by a physical system, rather than restricting consciousness to humans or biological organisms. The Model section says a system of elements with Φ>0 can form a complex; this supports the claim that IIT implies a form of panpsychism, though the paper does not use that term or say every physical object is conscious. (LLM, conf=0.85)
- Not checked — no full text of the sourcecites [6] C. Koch, 2011, A Test for Consciousness, Scientific American source
- Not checked — no full text of the sourcecites [7] G. Tononi, 2012, Integrated information theory of consciousness: an updated account, Arch Ital Biol doi:10.4449/aib.v149i5.1388
- Supportedcites [8] M. Oizumi, 2014, From the phenomenology to the mechanisms of consciousness: Integrated information theory 3.0, Computational Biology source
In “Simple systems can be conscious: A ‘minimally conscious’ photodiode,” the paper concludes that a two-element photodiode is conscious, albeit minimally, and says a thermistor could have the same minimal experience. This supports the broad characterization that IIT implies a form of panpsychism, though the passage does not use that term or say that every physical object is conscious. (LLM, conf=0.85)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Could not verify
44 more claims · 2 with their cited sources checked
Create a free account to see all 73 claims and to verify the 41 references still open.
- Probably not a substantial claim. 13
- It may state two separate claims as one. 14
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“both theories seem to implicate the corico-thalamic connections as necessary neural structures of consciousness (Tononi et al., 2016, p.452;Cerullo, 2015, p.5)”
From The Ergonomics of Learning: A Proposed Update to Bloom’s Taxonomy (2025) · cites this paper as Cerullo, 2015
Experimental — accuracy not yet measuredIn the citing paper
“Cerullo (2015) posits a more dynamic explanatory framework because it defines consciousness as information circulation in feedback loops.”
From The Ergonomics of Learning: A Proposed Update to Bloom’s Taxonomy (2025) · cites this paper as Cerullo (2015)
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ The Problem with Phi: A Critique of Integrated Information Theory > The Epistemology of IIT
“Dehaene and Naccache’s [30] workspace framework of consciousness”
§ The Problem with Phi: A Critique of Integrated Information Theory > References
“29. Dehaene S, & Naccache L. (2001). Towards a cognitive neuroscience of consciousness: basic evidence and a workspace framework In Dahaene (Ed.), The Cognitive Neuroscience of Consciousness (pp. 1– 37). Cambridge, MA: MIT Press 11164022 30. Gazzaniga M., Ivry R., & Mangun G. (2002). Cognitive Neuroscience: The Biology of the Mind 2nd edition New York: W. W. Norton & Company.”
Citations
Cites
James, W. (1913) · Henry Holt and Company eBooks · doi:10.1037/11059-000 · Cited by 25,226
The conscious mind: in search of a fundamental theory
(1996) · Choice Reviews Online · doi:10.5860/choice.34-2091 · Cited by 4,147
PINKER, S. (1999) · Annals of the New York Academy of Sciences · doi:10.1111/j.1749-6632.1999.tb08538.x · Cited by 3,738
(1988) · Cited by 3,656
On a confusion about a function of consciousness
Block, N. (1995) · Behavioral and Brain Sciences · doi:10.1017/s0140525x00038188 · Cited by 3,364
Towards a cognitive neuroscience of consciousness: basic evidence and a workspace framework
Dehaene, S. (2001) · Cognition · doi:10.1016/s0010-0277(00)00123-2 · Cited by 2,498 · cited 3×
“In mammalian brains, consciousness is associated with awareness, memory, and executive function [23–37].”
BackgroundMethodFacing up to the problem of consciousness
Chalmers, D. J. (1996) · Toward a Science of Consciousness · doi:10.7551/mitpress/6860.003.0003 · Cited by 1,966
An information integration theory of consciousness
Tononi, G. (2004) · BMC Neuroscience · doi:10.1186/1471-2202-5-42 · Cited by 1,958 · cited 3×
“The principal theoretical justification for IIT is the conjecture that, in perception, what is absent may be more significant to consciousness than what is present [2,3].”
BackgroundBreakdown of cortical effective connectivity during sleep
Massimini, M., Ferrarelli, F., Huber, R., Esser, S. K., Singh, H., & Tononi, G. (2005) · Science · doi:10.1126/science.1117256 · Cited by 1,534
Consciousness as integrated information: a provisional manifesto
Tononi, G. (2008) · The Biological Bulletin · doi:10.2307/25470707 · Cited by 1,529 · cited 10×
“In other words, what makes the experience different is that when the person looks at a uniformly dark room, in addition to registering that the room is dark, they are able to rule out every other state they are capable of perceiving [2,3].”
BackgroundInfluential
Research works printed but not linked: 3
- G. Tononi, 2003, Measuring integrated information, BMC Neuroscience, 4
- D. Griffen, 1997, Panexperientialist physicalism and the mind-body problem, Journal of Consciousness Studies, 4, 248
- G. Strawson, 2006, Realistic monism–why physicalism entails panpsychism, Journal of Consciousness Studies, 13, 3
Cited by
Items 101–117 of 117
Anesthesia for non-traditional consciousness
Deverett, B. (2023) · Frontiers in Human Neuroscience · doi:10.3389/fnhum.2023.1146242
“Some critics view IIT as insufficiently specific, citing simple systems that meet its criteria without convincingly demonstrating consciousness (Cerullo, 2015).”
BackgroundWhat We Are for Us, What We Are for Others: Consciousness and Identity
López-Segura, P., & Alonso, T. O. (2023) · Logic, argumentation & reasoning · doi:10.1007/978-3-031-46742-4_12
Informational Models of the Phenomenon of Consciousness and the Mechanistic Project in Neuroscience
Baetu, T. M. (2024) · Erkenntnis · doi:10.1007/s10670-024-00792-w · Cited by 1
“3 An alternative solution would be to provide independent empirical evidence that Φ satisfies the desideratum of content validity; but, as critics (Bayne, 2018; Cerullo, 2015; Merker et al., 2021) point out, no evidence for content validity is provided either.”
Thagard, P. R. (2019) · Oxford University Press eBooks · doi:10.1093/oso/9780190678739.002.0007
Ritter, F. E. (2019) · Oxford University Press eBooks · doi:10.1093/oso/9780190678739.002.0011
Thagard, P. R. (2019) · Oxford University Press eBooks · doi:10.1093/oso/9780190678739.002.0005
Thagard, P. R. (2019) · Oxford University Press eBooks · doi:10.1093/oso/9780190678739.002.0004
Thagard, P. R. (2019) · Oxford University Press eBooks · doi:10.1093/oso/9780190678739.002.0008
Wagemann, J. (2024) · Adaptive Behavior · doi:10.1177/10597123241263607
“…mere signal feedback or “ reentry ” (Edelman & Gally, 2013), initially separate networks should coordinate so that speci fi c oscillation patterns correlatable with phenomenal experiences result and how such a principle should be realized neuronally (Cerullo, 2015; Mindt, 2017; Wagemann, 2011).”
Panpsychism and dualism in the science of consciousnessPreprint
Yurchenko, S. B. (2024) · doi:10.31219/osf.io/ksc8x
Dadashkarimi, M. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5615010
Shibah, S. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5761142
From Randomness to Recursion: A Geometric Theory of Free Will in an Indeterministic WorldPreprint
Rafiei, S. K. S. (2025) · Research Square · doi:10.21203/rs.3.rs-6770593/v1
From Neural Complexity to Relational Personhood: Rethinking Consciousness in Brain Organoids
Picozzi, M., Pegoraro, R., & Carrara, A. (2026) · AJOB Neuroscience · doi:10.1080/21507740.2026.2649173
“Moreover, significant criticisms have challenged IIT’s empirical testability and conceptual coherence (Cerullo 2015; Fleming et al. 2023).”
Rafiei, S. K. S., Anar, M. A., Shahinzadeh, A., Asgari, S., Hosseinpour, M., Rafiei, M., Sadr, S. H. N., & Kashkoli, M. M. (2026) · Theory in Biosciences · doi:10.1007/s12064-026-00478-7 · Cited by 1
“The two frameworks may therefore converge in some regimes while diverging sharply in others (Tononi et al. 2016 ; Albantakis et al. 2023 ; Tononi 2015 ; Oizumi et al. 2014 ; Mediano et al. 2022 ; Cerullo 2015 ).”
Grigoryan, G. A., & Balaban, P. M. (2026) · Biosystems · doi:10.1016/j.biosystems.2026.105967
The Ergonomics of Learning: A Proposed Update to Bloom’s Taxonomy
Curtis, A., & Johansson, A. (2025) · Intersection: A Journal at the Intersection of Assessment and Learning · doi:10.61669/001c.142595
Metadata
- Authors
- Cerullo, M. A.
- Corresponding author
- Cerullo, Michael A.
- Venue
- PLOS Computational Biology
- Publisher
- Public Library of Science (PLoS)
- Published
- 2015-09-17
- Volume
- 11
- Issue
- 9
- Pages
- e1004286
- ISSN
- 1553-734X (ISSN-L)
- 1553-7358
- PMC
- PMC4574706
- OpenAlex
- W2193208712
- Cited by
- 114
- References
- 73
- Citations by year
- 2026: 4 · 2025: 9 · 2024: 10 · 2023: 10 · 2022: 7 · 2021: 15 · 2020: 8 · 2019: 31 · 2018: 12 · 2017: 5 · 2016: 2
- FWCI
- 9.55
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- PLoS Computational Biology (publishedVersion, cc-by)
- Europe PMC (PubMed Central) (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- PhilPapers (PhilPapers Foundation) (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Social Sciences › Arts and Humanities › Philosophy › Philosophy and Theoretical Science
- Life Sciences › Neuroscience › Cognitive Neuroscience › Embodied and Extended Cognition
- Social Sciences › Arts and Humanities › History and Philosophy of Science › Philosophy of Biology and Scientific Explanation
- Keywords
- integrated information theory, consciousness, integrated information, qualia, computational functionalism, phenomenology, proto-consciousness, philosophy of mind, epistemology of consciousness
- MeSH terms
- Algorithms, Consciousness, Humans, Information Theory, Models, Neurological, Biomedical Research, Split-Brain Procedure
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.