arxiv: 2604.11482 · v1 · submitted 2026-04-13 · 🧬 q-bio.NC

Recognition: unknown

Integrated information theory: the good, the bad and the misunderstood

Adam B. Barrett , Borjan Milinkovic , Pedro A. M. Mediano , Fernando E. Rosas , Daniel Bor , Lionel Barnett , Anil K. Seth

Authors on Pith no claims yet

Pith reviewed 2026-05-10 15:05 UTC · model grok-4.3

classification 🧬 q-bio.NC

keywords integrated information theoryconsciousnessphipanpsychismphysical systemsproxiescontinuous fields

0 comments

The pith

Φ in Integrated Information Theory is not well-defined for real physical systems and has never been computed on them.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper summarizes the core ideas of Integrated Information Theory and addresses common misunderstandings about it. The authors argue that the central measure Φ does not directly indicate the level of consciousness and that the theory's implied panpsychism is coherent. They further contend that Φ is not well-defined for real physical systems, that only proxies have been computed rather than the measure itself, and that a reformulation using continuous fields is necessary to make the theory compatible with fundamental physics.

Core claim

IIT aims to derive a mathematical formula for consciousness from its fundamental properties and apply it to any physical system. However, the paper identifies that the quantity Φ is not synonymous with the amount of consciousness, that IIT supports a panpsychist ontology without contradiction, but crucially that Φ has not been defined or computed for any real physical system and only proxies exist. To integrate with physics, the theory must be recast in continuous fields instead of discrete elements.

What carries the argument

Φ, IIT's measure of integrated information based on the system's cause-effect repertoire.

If this is right

A high Φ value is not synonymous with more consciousness; a suite of quantities could provide a multi-dimensional characterization instead.
IIT implies a panpsychism in which space and time are tiled with substrates of proto-consciousness, but this is not problematic for the theory.
Only proxies for IIT measures have been computed on systems, not direct approximations of Φ.
For compatibility with fundamental physics, IIT requires reformulation in terms of continuous fields rather than discrete units.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

If the authors are correct, then existing empirical work claiming to measure Φ in brains or other systems is actually using indirect proxies that do not test the core theory.
This limitation suggests that IIT may need substantial mathematical redevelopment before it can make testable predictions about consciousness in physical systems.
Reformulating IIT with continuous fields could open connections to other physical theories of information integration in quantum or field contexts.

Load-bearing premise

The authors' interpretation of IIT's formal definitions and requirements for physical systems is correct and complete.

What would settle it

Performing an exact computation of Φ on a real physical system or deriving a continuous-field version of the theory that preserves its core predictions would challenge the claim that such a definition is impossible in the current form.

Figures

Figures reproduced from arXiv: 2604.11482 by Adam B. Barrett, Anil K. Seth, Borjan Milinkovic, Daniel Bor, Fernando E. Rosas, Lionel Barnett, Pedro A. M. Mediano.

**Figure 2.** Figure 2: Neural architectures associated with experiences of space and time. In this figure, [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

**Figure 3.** Figure 3: The panpsychism of IIT. IIT implies a tiling of space and time with substrates [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

read the original abstract

The integrated information theory of consciousness (IIT) is uniquely ambitious in proposing a mathematical formula, derived from apparently fundamental properties of conscious experience, to describe the quantity and quality of consciousness for any physical system that possesses it. IIT has generated considerable debate, which has engendered some misunderstandings and misrepresentations. Here we address and hope to remedy this. We begin by concisely summarising the essentials of IIT. Given IIT is supposed to apply universally, we do this with reference to an arbitrary patch of matter, as opposed to the usual system of discrete computational units. Then, after briefly summarising IIT's theoretical and empirical achievements, we focus on five points which we consider especially important for driving forward new theory and increasing understanding. First, a high value of the measure $\Phi$ is not synonymous with `more consciousness'. We describe how $\Phi$ might be replaced with a suite of quantities to obtain a multi-dimensional characterisation of states of consciousness. Second, we describe with nuance the distinct flavour of panpsychism implied by IIT -- whereby space (and time) are tiled with substrates of (proto-) consciousness -- and find this is not problematic for the theory. Third, $\Phi$ is not well-defined for real physical systems, and has not been computed on any real physical system. Fourth, so far only proxies for IIT measures have been computed, and not approximations. Fifth, for IIT to fit with current successful theories in fundamental physics, a reformulation in terms of continuous fields would be needed.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This review organizes existing IIT critiques into five clear points and flags that Φ has never been computed on real systems, but the central claim needs tighter derivation on why continuous physics breaks it.

read the letter

The paper's main contribution is walking through five targeted clarifications on Integrated Information Theory using an arbitrary patch of matter as the example. It separates Φ from any simple notion of 'more consciousness,' notes the specific panpsychist flavor without calling it a problem, and stresses that only proxies have been calculated so far. The suggestion to replace Φ with a multi-dimensional suite is a modest but practical extension of prior comments in the literature. The writing stays close to IIT's published axioms and avoids overclaiming novelty, which keeps the arguments grounded. The distinction between proxies and approximations is especially useful for anyone trying to interpret existing empirical work. The soft spot sits in the third point. The authors state that Φ is not well-defined for real physical systems because it requires discrete units and perfect state knowledge, and therefore needs a continuous-field reformulation to match fundamental physics. This follows logically from the discrete partition and repertoire definitions they summarize, but the text does not supply an explicit step-by-step showing where the standard calculation becomes inconsistent or non-normalizable once states are replaced by continuous fields or densities. The claim therefore rests on the assumption that the current formulation cannot be extended without violating the axioms. That assumption may hold, yet a short formal sketch would make the incompatibility sharper. The paper is aimed at consciousness researchers and philosophers who already know the basics of IIT and want to avoid repeating common misapplications. It does not resolve open questions or add new mechanisms, but the clarifications are clear enough to help guide future empirical and theoretical work. I would send it to peer review; the logical structure is sound and the points are worth having on record even if one section could use more explicit derivation.

Referee Report

1 major / 2 minor

Summary. The paper summarizes Integrated Information Theory (IIT) by applying its axioms to an arbitrary patch of matter rather than discrete units, reviews its achievements, and advances five points: (1) high Φ does not equate to more consciousness and a multi-dimensional suite of measures is preferable; (2) IIT implies a nuanced panpsychism that tiles space-time with proto-conscious substrates and is not problematic; (3) Φ is not well-defined for real physical systems and has never been computed on any; (4) only proxies for IIT measures, not approximations, have been computed; (5) compatibility with fundamental physics requires reformulation in continuous fields.

Significance. If the distinctions between proxies and approximations, and the discrete requirements of the current Φ formulation, are accurate, the paper provides a useful clarification of IIT that can guide future work. It explicitly credits the theory's axiom-derived structure while identifying gaps in application to physical systems. The call for continuous-field reformulation and multi-dimensional characterization offers concrete directions for bridging IIT to physics and empirical neuroscience.

major comments (1)

[Point 3 and the subsequent discussion of continuous reformulation] Point 3: The central claim that Φ 'is not well-defined for real physical systems' and 'has not been computed on any real physical system' rests on the assertion that the current formulation presupposes discrete units and perfect state knowledge. No explicit derivation is supplied showing the precise step (e.g., in the cause-effect repertoire construction or the search over partitions) at which the calculation becomes undefined or inconsistent when states are replaced by continuous fields or probability densities over manifolds. This is load-bearing for both point 3 and the reformulation advocated in point 5.

minor comments (2)

[Abstract] The abstract refers to replacing Φ with 'a suite of quantities' for multi-dimensional characterization but provides no concrete examples or references to candidate quantities; adding one or two brief illustrations would improve clarity.
[Main text, after the IIT summary] The transition from the summary of IIT (applied to an arbitrary patch of matter) to the five points would benefit from explicit subsection headings or a short roadmap sentence to help readers track the logical flow.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful reading and constructive critique, which identifies a key area where our argument on the applicability of IIT can be strengthened. We address the major comment below and will revise the manuscript accordingly.

read point-by-point responses

Referee: Point 3: The central claim that Φ 'is not well-defined for real physical systems' and 'has not been computed on any real physical system' rests on the assertion that the current formulation presupposes discrete units and perfect state knowledge. No explicit derivation is supplied showing the precise step (e.g., in the cause-effect repertoire construction or the search over partitions) at which the calculation becomes undefined or inconsistent when states are replaced by continuous fields or probability densities over manifolds. This is load-bearing for both point 3 and the reformulation advocated in point 5.

Authors: We agree that an explicit step-by-step demonstration of the inconsistencies would make the claim more rigorous and directly support the call for reformulation. In the revised manuscript we will add a dedicated subsection (or appendix) that traces the breakdown: (i) the foundational requirement that a system consist of a finite set of discrete elements each possessing a finite repertoire of states; (ii) the construction of cause-effect repertoires, which are defined via conditional probability distributions over those discrete states and become ill-posed when states are replaced by continuous probability densities on manifolds; and (iii) the exhaustive search over bipartitions, which presupposes a countable collection of elements and cannot be performed without an arbitrary discretization that alters the original quantity. These steps render the current Φ undefined for continuous fields without a wholesale reformulation, consistent with point 5. We will also reiterate that all existing numerical work has been performed on discrete or discretized models rather than on any actual continuous physical system. revision: yes

Circularity Check

0 steps flagged

No circularity: critique summarizes IIT definitions without self-referential derivations or load-bearing self-citations

full rationale

The paper is a conceptual commentary that restates IIT's published axioms and measures (e.g., Φ computed via cause-effect repertoires on discrete units) and then offers five points of critique, including that Φ has never been computed on real physical systems and requires continuous-field reformulation. No mathematical derivation chain exists in the paper; claims rest on logical reading of external IIT literature rather than any equation or parameter introduced here that reduces to itself by construction. Self-citations, if present, are not load-bearing for the central assertions, which do not invoke uniqueness theorems or ansatzes from the authors' prior work. The argument is therefore self-contained against the theory's own stated definitions.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The paper introduces no new free parameters, axioms, or invented entities; it relies entirely on the standard mathematical structure and assumptions already present in the IIT literature it cites.

pith-pipeline@v0.9.0 · 5598 in / 1147 out tokens · 27244 ms · 2026-05-10T15:05:34.336581+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

47 extracted references · 6 canonical work pages

[1]

(2014) Why I am not an integrated information theorist (or, the unconscious expander)

Aaronson, S. (2014) Why I am not an integrated information theorist (or, the unconscious expander). Shtetl-Optimized. https://scottaaronson.blog/?p=1799

2014
[2]

(2023) Integrated information theory (IIT) 4.0: formulating the properties of phenomenal existence in physical terms

Albantakis, L., Barbosa, L., Findlay, G., Grasso, M., Haun, A.M., Marshall, W., et al. (2023) Integrated information theory (IIT) 4.0: formulating the properties of phenomenal existence in physical terms. PLoS Comput. Biol. 19(10): e1011465

2023
[3]

(2014) An integration of integrated information theory with fundamental physics

Barrett, A.B. (2014) An integration of integrated information theory with fundamental physics. Front. Psychol. 5, 63

2014
[4]

(2016) A comment on Tononi & Koch (2015) `Consciousness: here, there and everywhere?'

Barrett, A.B. (2016) A comment on Tononi & Koch (2015) `Consciousness: here, there and everywhere?'. Phil. Trans. R. Soc. B 20140198

2016
[5]

and Mediano, P.A.M

Barrett, A.B. and Mediano, P.A.M. (2019) The phi measure of integrated information is not well-defined for general physical systems. J. Conscious. Stud. 26(1--2), 11--20

2019
[6]

(2022) Does integrated information theory make testable predictions about the role of silent neurons in consciousness? Neurosci

Bartlett, G. (2022) Does integrated information theory make testable predictions about the role of silent neurons in consciousness? Neurosci. Conscious. 2022(1), niac015

2022
[7]

(2016) Are there levels of consciousness? Trends Cogn

Bayne, T., Hohwy, J., and Owen, A.M. (2016) Are there levels of consciousness? Trends Cogn. Sci. 20(6), 405--413

2016
[8]

(2018) On the axiomatic foundations of the integrated information theory of consciousness

Bayne, T. (2018) On the axiomatic foundations of the integrated information theory of consciousness. Neurosci. Conscious. 2018(1), niy007

2018
[9]

Casali, A.G. et al. (2013) A theoretically based index of consciousness independent of sensory processing and behavior. Sci. Transl. Med. 5, 198ra105

2013
[10]

Casarotto, S. et al. (2016) Stratification of unresponsive patients by an independently validated index of brain complexity. Ann. Neurol. 80, 718--729

2016
[11]

(2023) The fundamental tension in integrated information theory 4.0's realist idealism

Cea, I., Negro, N., and Signorelli, C.M. (2023) The fundamental tension in integrated information theory 4.0's realist idealism. Entropy 25, 1453

2023
[12]

(1995) Facing up to the problem of consciousness

Chalmers, D.J. (1995) Facing up to the problem of consciousness. J. Conscious. Stud. 2, 200--219

1995
[13]

(2025) Consciousness science: where are we, where are we going, and what if we get there? Front

Cleeremans, A., Mudrik, L., and Seth, A.K. (2025) Consciousness science: where are we, where are we going, and what if we get there? Front. Sci. 3. doi:10.3389/fsci.2025.1546279

work page doi:10.3389/fsci.2025.1546279 2025
[14]

Cogitate Consortium., Ferrante, O., Gorska-Klimowska, U. et al. (2025) Adversarial testing of global neuronal workspace and integrated information theories of consciousness. Nature. https://doi.org/10.1038/s41586-025-08888-1

work page doi:10.1038/s41586-025-08888-1 2025
[15]

(2024) Why does time feel the way it does? Towards a principled account of temporal experience

Comolatti, R., Grasso, M., and Tononi, G. (2024) Why does time feel the way it does? Towards a principled account of temporal experience. iScience 28, 113434

2024
[16]

and Changeux, J.P

Farisco, M. and Changeux, J.P. (2023) About the compatibility between the perturbational complexity index and the global neuronal workspace theory of consciousness. Neurosci. Conscious. 2023(1), niad016

2023
[17]

Ferrarelli, F. et al. (2010) Breakdown in cortical effective connectivity during midazolam-induced loss of consciousness. Proc. Natl. Acad. Sci. 107, 2681--2686

2010
[18]

(2024) Dissociating artificial intelligence from artificial consciousness

Findlay, G., Marshall, W., Albantakis, L., David, I., Mayner, W.G.P., Koch, C., and Tononi, G. (2024) Dissociating artificial intelligence from artificial consciousness. arXiv:2412.04571

work page arXiv 2024
[19]

(1998) Non-Markovian character of ionic current fluctuations in membrane channels

Fuli \'n ski, A., Grzywna, Z., Mellor, I., Siwy, Z., and Usherwood, P.N.R. (1998) Non-Markovian character of ionic current fluctuations in membrane channels. Phys. Rev. E 58(1), 919

1998
[20]

and Seth, A.K

Gomez-Marin, A. and Seth, A.K. (2025) A science of consciousness beyond pseudo-science and pseudo-consciousness. Nat. Neurosci. 28, 703--706

2025
[21]

and Tononi, G

Haun, A. and Tononi, G. (2019) Why does space feel the way it does? Towards a principled account of spatial experience. Entropy 21(12), 1160

2019
[22]

IIT-Concerned., Klincewicz, M., Cheng, T. et al. (2025) What makes a theory of consciousness unscientific? Nat. Neurosci. 28, 689--693

2025
[23]

and Tull, S

Kleiner, J. and Tull, S. (2021) The mathematical structure of integrated information theory. Front. Appl. Math. Stat. 6, 2020

2021
[24]

(2024) A landscape of consciousness: toward a taxonomy of explanations and implications

Kuhn, R.L. (2024) A landscape of consciousness: toward a taxonomy of explanations and implications. Prog. Biophys. Mol. Biol. 190, 28--169

2024
[25]

(1983) Materialism and qualia: the explanatory gap

Levine, J. (1983) Materialism and qualia: the explanatory gap. Pac. Philos. Q. 64, 354--361

1983
[26]

(2024) Intrinsic units: identifying a system's causal grain

Marshall, W., Findlay, G., Albantakis, L., and Tononi, G. (2024) Intrinsic units: identifying a system's causal grain. bioRxiv 2024.04.12.589163

2024
[27]

Massimini, M. et al. (2005) Breakdown of cortical effective connectivity during sleep. Science 309, 2228--2232

2005
[28]

Massimini, M. et al. (2009) A perturbational approach for evaluating the brain's capacity for consciousness. Prog. Brain Res. 177, 201--214

2009
[29]

(2019) Measuring integrated information: comparison of candidate measures in theory and simulation

Mediano, P.A.M., Seth, A.K., and Barrett, A.B. (2019) Measuring integrated information: comparison of candidate measures in theory and simulation. Entropy 21, 17

2019
[30]

(2022) The strength of weak integrated information theory

Mediano, P.A.M., Rosas, F.E., Bor, D., Seth, A.K., and Barrett, A.B. (2022) The strength of weak integrated information theory. Trends Cogn. Sci. 26(8)

2022
[31]

(2019) Is consciousness intrinsic? A problem for the integrated information theory

M rch, H.H. (2019) Is consciousness intrinsic? A problem for the integrated information theory. J. Conscious. Stud. 26(1--2), 133--162

2019
[32]

Mudrik, L. et al. (2025) Unpacking the complexities of consciousness: theories and reflections. Neurosci. Biobehav. Rev. 170, 106053

2025
[33]

Olcese, U., Muckli, L.F., Boly, M. et al. (2024) Accelerating research on consciousness: an adversarial collaboration to test contrasting predictions of the integrated information theory and predictive processing accounts of consciousness -- version 2. Open Science Foundation. https://doi.org/10.17605/OSF.IO/4RN85

work page doi:10.17605/osf.io/4rn85 2024
[34]

Re, V.L. et al. (2021) Role of transcranial magnetic stimulation (TMS) combined with electroencephalography (EEG) in disorders of consciousness (DOC). J. Neurol. Sci. 429, 118507

2021
[35]

(2021) Consciousness and complexity: a consilience of evidence

Sarasso, S., Casali, A.G., Casarotto, S., Rosanova, M., Sinigaglia, C., and Massimini, M. (2021) Consciousness and complexity: a consilience of evidence. Neurosci. Conscious. 2021(2), niab023

2021
[36]

(2000) Measuring information transfer

Schreiber, T. (2000) Measuring information transfer. Phys. Rev. Lett. 85(2), 461--464

2000
[37]

(2009) Explanatory correlates of consciousness: theoretical and computational challenges

Seth, A.K. (2009) Explanatory correlates of consciousness: theoretical and computational challenges. Cogn. Comput. 1, 50--63

2009
[38]

(2015) Granger causality analysis in neuroscience and neuroimaging

Seth, A.K., Barrett, A.B., and Barnett, L. (2015) Granger causality analysis in neuroscience and neuroimaging. J. Neurosci. 35(8), 3293--3297

2015
[39]

and Bayne, T

Seth, A.K. and Bayne, T. (2022) Theories of consciousness. Nat. Rev. Neurosci. 23, 439--452

2022
[40]

(2019) Evaluating approximations and heuristic measures of integrated information

Sevenius Nilsen, A., Juel, B.E., and Marshall, W. (2019) Evaluating approximations and heuristic measures of integrated information. Entropy 21(5), 525

2019
[41]

(2022) In search of lost time: integrated information theory needs constraints from temporal phenomenology

Singhal, I., Mudumba, R., and Srinivasan, N. (2022) In search of lost time: integrated information theory needs constraints from temporal phenomenology. Philos. Mind Sci. 3. doi:10.33735/phimisci.2022.9438

work page doi:10.33735/phimisci.2022.9438 2022
[42]

and Edelman, G.M

Tononi, G. and Edelman, G.M. (1998) Consciousness and complexity. Science 282(5395), 1846--1851

1998
[43]

(2004) An information integration theory of consciousness

Tononi, G. (2004) An information integration theory of consciousness. BMC Neurosci. 5, 42

2004
[44]

Tononi, G., Boly, M., Massimini, M. et al. (2016) Integrated information theory: from consciousness to its physical substrate. Nat. Rev. Neurosci. 17, 450--461

2016
[45]

(2023) Only what exists can cause: An intrinsic view of free will.https://doi.org/ 10.48550/arXiv.2206.02069 75 APREPRINT- 14THMAY, 2026

Tononi, G., Albantakis, L., Boly, M., Cirelli, C., and Koch, C. (2022) Only what exists can cause: an intrinsic view of free will. arXiv:2206.02069

work page arXiv 2022
[46]

Tononi, G., Boly, M., Grasso, M. et al. (2022b) IIT, half masked and half disfigured. Behav. Brain Sci. 45, e60
[47]

Tononi, G., Albantakis, L., Barbosa, L. et al. (2025) Consciousness or pseudo-consciousness? A clash of two paradigms. Nat. Neurosci. 28, 694--702

2025