arxiv: 2605.13884 · v1 · submitted 2026-05-11 · 🧬 q-bio.NC · cs.AI

Recognition: 2 theorem links

· Lean Theorem

Consciousness as Uncommon Self-Knowledge: A Synergistic Information Framework

Krti Tallam

Authors on Pith no claims yet

Pith reviewed 2026-05-15 05:00 UTC · model grok-4.3

classification 🧬 q-bio.NC cs.AI

keywords consciousnesssynergistic informationself-knowledgepartial information decompositionuncommon self-knowledgemetacognitionanesthesiaglobal workspace

0 comments

The pith

Consciousness consists of uncommon self-knowledge: synergistic information about a system that exists only in the joint activity of its subsystems and vanishes when those subsystems are examined separately.

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

The paper proposes uncommon self-knowledge as a candidate criterion for consciousness. This takes the form of synergistic self-directed information that appears solely in the combined behavior of a system's parts and is eliminated by any decomposition into independent parts. A sympathetic reader would care because the idea cleanly separates consciousness from metacognition by contrasting synergy with redundancy in self-information. It also supplies resolutions to counterexamples faced by other theories and produces concrete empirical predictions, including a timing dissociation in global workspace models and specific disruptions in artificial systems. The account matches reported patterns in anesthesia and Alzheimer's disease, where synergistic processing drops while redundancy is preserved or elevated.

Core claim

Uncommon self-knowledge, understood as the synergistic component of self-directed information that exists only jointly among subsystems and is destroyed by decomposition, serves as a formal signature for conscious processing. The proposal rests on a partition-lattice grounding of partial information decomposition in which synergy captures the information gap between joint and separate observation of the self, while redundancy aligns with common knowledge.

What carries the argument

The synergistic component of self-directed information, the portion that exists only in the joint of subsystems and is lost upon decomposition.

If this is right

Consciousness separates from metacognition through the distinction between synergistic and redundant self-knowledge.
Counterexamples that challenge integrated information theory, global workspace theory, and higher-order thought theory receive principled resolutions.
The measure becomes operational via partial information rate decomposition with self as the target.
Global workspace theory gains a timing dissociation in which consciousness correlates with pre-broadcast synergy formation rather than the broadcast itself.
Large language models exhibit a specific dissociation between self-report disruption and task-performance disruption under middle-layer perturbation.

Where Pith is reading between the lines

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

The same self-directed synergy measure could be applied to neural recordings from non-human animals to test for the presence of uncommon self-knowledge.
Artificial systems might be probed for consciousness by checking whether they generate synergistic information about their own internal states that exceeds what separate modules can provide.
Clinical monitoring of synergy levels could offer a quantitative marker for conscious state that is independent of behavioral report.
The framework invites direct comparison experiments that isolate joint versus partitioned self-information in both biological and engineered systems.

Load-bearing premise

That synergistic self-directed information specifically marks consciousness rather than other forms of integrated processing or cognition.

What would settle it

An observation that synergistic self-information levels remain unchanged during loss of consciousness under anesthesia, or that targeted middle-layer perturbations in language models fail to produce the predicted dissociation between self-report and task performance.

read the original abstract

We propose uncommon self-knowledge (USK) as a candidate criterion for consciousness: synergistic information a system carries about itself that exists only in the joint of its subsystems and is destroyed by decomposition. Drawing on Gottwald's partition-lattice grounding of Partial Information Decomposition (PID), where redundancy corresponds to Aumann's common knowledge and synergy to the gap between separate and joint observation, we propose the synergistic component of self-directed information as a candidate formal signature for conscious processing. If correct, the framework would (1) offer a clean separation between consciousness and metacognition (synergistic vs. redundant self-knowledge), (2) provide principled resolutions to counterexamples that challenge IIT, GWT, and HOT, (3) be operationalizable via Partial Information Rate Decomposition (PIRD) with self-targeting, and (4) generate distinctive empirical predictions, the strongest being a GWT timing dissociation (consciousness correlates with pre-broadcast synergy formation, not broadcast itself) and a specific dissociation between self-report disruption and task-performance disruption under middle-layer perturbation in LLMs. The proposal is consistent with recent empirical findings that both anaesthesia and Alzheimer's disease specifically reduce synergistic information processing while preserving or increasing redundancy.

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

The paper reframes synergistic self-information from PID as a candidate signature for consciousness but offers no derivation showing why it tracks experience rather than general integration.

read the letter

The main move is to define uncommon self-knowledge as the synergistic atom in self-directed partial information decomposition, grounded in Gottwald's partition lattice, and to equate that quantity with consciousness. Redundancy gets mapped to Aumann common knowledge and the synergy term to the extra information available only from joint observation. This produces a clean split between consciousness and metacognition that the paper contrasts with IIT, GWT, and HOT counterexamples. It also notes consistency with anesthesia and Alzheimer's data where synergy drops while redundancy stays flat or rises, and it lists concrete predictions such as pre-broadcast synergy timing in global workspace models and specific LLM perturbation effects. Those links are the parts that feel useful on a first read. The central gap is the missing argument for why this particular synergistic quantity appears only with phenomenal experience and vanishes exactly when consciousness does. The manuscript presents the mapping as a proposal rather than deriving it from axioms or excluding the same term in non-conscious integrated systems like feed-forward nets or cellular automata. Without that step the claim stays definitional. The operationalization via PIRD is sketched but not carried out here, so the predictions remain untested in the text. This is the sort of paper that belongs in a reading group on information-theoretic accounts of mind; anyone already comfortable with PID and the standard theories will see the connections quickly. It deserves peer review because the framing engages the literature directly and the empirical alignments are worth checking, even if the formal grounding needs tightening before the signature can be treated as operational.

Referee Report

3 major / 2 minor

Summary. The manuscript proposes uncommon self-knowledge (USK) as a candidate formal signature for consciousness, defined as the synergistic component of self-directed partial information decomposition (PID) using Gottwald's partition-lattice grounding. Redundancy is identified with Aumann common knowledge and synergy with the gap between separate and joint self-observation; the framework is claimed to cleanly separate consciousness from metacognition, resolve counterexamples to IIT/GWT/HOT, be operationalizable via self-targeted PIRD, and generate distinctive empirical predictions (e.g., GWT timing dissociation and LLM middle-layer perturbation effects) while remaining consistent with anesthesia and Alzheimer's findings that reduce synergy while preserving redundancy.

Significance. If the proposed mapping can be shown to be non-circular and uniquely diagnostic of phenomenal experience, the framework would supply an information-theoretic criterion that distinguishes conscious processing from other integrated computations, offering principled resolutions to existing theoretical impasses and generating falsifiable predictions for both biological and artificial systems.

major comments (3)

[Abstract] Abstract: the claim that the framework 'provide[s] principled resolutions to counterexamples that challenge IIT, GWT, and HOT' is unsupported; the text asserts the potential without examining any concrete counterexample or demonstrating how the synergistic self-information term specifically eliminates it.
[Central proposal] Central proposal: the identification of the synergistic atom of self-directed PID with consciousness is introduced by definitional mapping rather than by an axiomatic derivation or uniqueness argument showing why the same quantity must be zero in non-conscious integrated systems (feed-forward networks, cellular automata, unconscious perceptual binding).
[Empirical predictions] Empirical predictions: the stated GWT timing dissociation (consciousness correlates with pre-broadcast synergy formation) is asserted without a formal derivation from the PID partition lattice or a proof that synergy must precede broadcast in the model's dynamics.

minor comments (2)

The manuscript would benefit from an explicit equation or definition of the synergistic atom in self-directed PID (e.g., in terms of the partition lattice) to make the PIRD operationalization reproducible.
Additional citations to existing applications of PID to neural data would help situate the proposal relative to prior work.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive and detailed comments, which highlight important areas where the manuscript's claims require stronger support and elaboration. We address each major comment below, indicating the revisions we will make to improve clarity, rigor, and evidential grounding while preserving the core proposal.

read point-by-point responses

Referee: [Abstract] Abstract: the claim that the framework 'provide[s] principled resolutions to counterexamples that challenge IIT, GWT, and HOT' is unsupported; the text asserts the potential without examining any concrete counterexample or demonstrating how the synergistic self-information term specifically eliminates it.

Authors: We agree that the abstract overstates the resolution claim without concrete demonstration in the current text. In the revised manuscript we will qualify or remove the phrasing in the abstract and add a new subsection (likely in the Discussion) that explicitly examines at least two concrete counterexamples—one from IIT (feedforward networks) and one from GWT (unconscious perceptual binding)—showing how the synergistic self-information term evaluates to zero in those cases while remaining positive under the conditions we associate with consciousness. This will make the claim evidence-based rather than promissory. revision: yes
Referee: [Central proposal] Central proposal: the identification of the synergistic atom of self-directed PID with consciousness is introduced by definitional mapping rather than by an axiomatic derivation or uniqueness argument showing why the same quantity must be zero in non-conscious integrated systems (feed-forward networks, cellular automata, unconscious perceptual binding).

Authors: The identification is indeed presented as a candidate mapping motivated by the alignment between synergistic self-information and the concept of uncommon self-knowledge, leveraging Gottwald’s partition-lattice properties. We do not claim a complete axiomatic derivation from first principles in this work. In revision we will expand the central proposal section with a uniqueness argument sketch: in feedforward networks the self-directed PID synergy is zero because causal isolation prevents irreducible joint self-observation; in cellular automata and unconscious binding the decomposition shows synergy vanishes when the system lacks the self-referential structure that creates the gap between separate and joint observation. This will provide additional motivation while acknowledging that a full uniqueness proof remains an open direction for future work. revision: partial
Referee: [Empirical predictions] Empirical predictions: the stated GWT timing dissociation (consciousness correlates with pre-broadcast synergy formation) is asserted without a formal derivation from the PID partition lattice or a proof that synergy must precede broadcast in the model's dynamics.

Authors: We acknowledge that the timing dissociation is stated without a formal derivation from the PID lattice. In the revised manuscript we will add a short appendix or subsection that derives the ordering from the partition lattice: the synergistic atom requires joint observation across subsystems, which in the GWT-style dynamics occurs during the pre-broadcast integration phase; broadcast itself corresponds to the subsequent redundant transmission of already-integrated information. A minimal dynamical model sketch will be included to illustrate that synergy formation necessarily precedes the broadcast step under the model’s information-flow assumptions. revision: yes

Circularity Check

1 steps flagged

Synergistic self-information identified as consciousness signature by direct definitional proposal

specific steps

self definitional [Abstract]
"We propose uncommon self-knowledge (USK) as a candidate criterion for consciousness: synergistic information a system carries about itself that exists only in the joint of its subsystems and is destroyed by decomposition. ... we propose the synergistic component of self-directed information as a candidate formal signature for conscious processing."

USK is defined as the synergistic component, after which the same component is proposed as the signature for consciousness; the equivalence is therefore introduced by the definitional step itself rather than derived from prior equations or external benchmarks.

full rationale

The manuscript's core move is to define uncommon self-knowledge (USK) explicitly as the synergistic atom of self-directed PID and then nominate that same quantity as the formal signature of consciousness. This identification is presented as a proposal grounded in a conceptual mapping (synergy = gap between separate and joint self-observation) rather than an independent derivation, uniqueness theorem, or falsification against non-conscious integrated systems. The link therefore reduces to the framework's own definitional choice.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The framework rests on the domain assumption that Gottwald's PID grounding maps redundancy to common knowledge and synergy to the joint-observation gap, plus the invented entity of uncommon self-knowledge as the consciousness signature; no free parameters are introduced.

axioms (1)

domain assumption Gottwald's partition-lattice grounding of Partial Information Decomposition where redundancy corresponds to Aumann's common knowledge and synergy to the gap between separate and joint observation
Invoked to ground the synergistic component as the candidate signature.

invented entities (1)

Uncommon self-knowledge (USK) no independent evidence
purpose: Candidate formal signature for conscious processing
New term defined as the synergistic self-directed information destroyed by decomposition.

pith-pipeline@v0.9.0 · 5505 in / 1252 out tokens · 45546 ms · 2026-05-15T05:00:48.987459+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

Cost.FunctionalEquation washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

We propose uncommon self-knowledge (USK)—synergistic information a system carries about itself that exists only in the joint of its subsystems and is destroyed by decomposition... Syn(S1, …, Sn → St+1) > 0

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

16 extracted references · 16 canonical work pages

[1]

Why i am not an integrated information theorist

Scott Aaronson. Why i am not an integrated information theorist. Blog post, Shtetl-Optimized, 2014

work page 2014
[2]

Robert J. Aumann. Agreeing to disagree. The Annals of Statistics, 4 0 (6): 0 1236--1239, 1976

work page 1976
[3]

Bernard J. Baars. Global workspace theory of consciousness: Toward a cognitive neuroscience of human experience. Progress in Brain Research, 150: 0 45--53, 2005

work page 2005
[4]

Toward a computational theory of conscious processing

Stanislas Dehaene, Lucie Charles, Jean-R \'e mi King, and S \'e bastien Marti. Toward a computational theory of conscious processing. Current Opinion in Neurobiology, 25: 0 76--84, 2014

work page 2014
[5]

Partial information rate decomposition

Luca Faes, Laura Sparacino, Gorana Mijatovic, Yuri Antonacci, Leonardo Ricci, Daniele Marinazzo, and Sebastiano Stramaglia. Partial information rate decomposition. Physical Review Letters, 2025. doi:10.1103/nrwj-n8lj

work page doi:10.1103/nrwj-n8lj 2025
[6]

Dissecting spectral G ranger causality through partial information decomposition

Luca Faes et al. Dissecting spectral G ranger causality through partial information decomposition. arXiv preprint, 2026

work page 2026
[7]

Partial information decomposition via partition lattices

Sebastian Gottwald. Partial information decomposition via partition lattices. arXiv preprint, 2024

work page 2024
[8]

Down, Pedro A

Luca Liardi, Thomas F. Down, Pedro A. M. Mediano, et al. A comprehensive review of partial information decomposition. arXiv preprint, 2026

work page 2026
[9]

Emotion concepts and their function in a large language model

Jack Lindsey et al. Emotion concepts and their function in a large language model. Anthropic Research, 2026. URL https://www.anthropic.com/research/emotion-concepts-function

work page 2026
[10]

Luppi et al

Andrea I. Luppi et al. A synergistic core for human brain evolution and cognition. Nature Neuroscience, 27: 0 771--783, 2024

work page 2024
[11]

On the impossibility of lattice-based PID for three or more sources

Aobo Lyu et al. On the impossibility of lattice-based PID for three or more sources. arXiv preprint, 2025

work page 2025
[12]

Rosenthal

David M. Rosenthal. Consciousness and mind. 2005

work page 2005
[13]

Adversarial discovery of neural mechanisms of consciousness from brain activity

Daniel Toker et al. Adversarial discovery of neural mechanisms of consciousness from brain activity. Nature Neuroscience, March 2026. GAN-based discovery on 680K+ neurophysiology recordings

work page 2026
[14]

Integrated information theory: From consciousness to its physical substrate

Giulio Tononi, Melanie Boly, Marcello Massimini, and Christof Koch. Integrated information theory: From consciousness to its physical substrate. Nature Reviews Neuroscience, 17: 0 450--461, 2016

work page 2016
[15]

Synergistic information processing in large language models

Alejandro Urbina-Rodriguez et al. Synergistic information processing in large language models. arXiv preprint, 2025

work page 2025
[16]

Task-level integration: A substrate-neutral criterion for consciousness detection

Aineko Wallace and Krti Tallam. Task-level integration: A substrate-neutral criterion for consciousness detection. In preparation, 2026

work page 2026