HomeThe Science of ThoughtMax Tegmark: Consciousness Is a State of Matter We Can Test

Max Tegmark: Consciousness Is a State of Matter We Can Test

MIT physicist Max Tegmark proposes consciousness has testable physical properties. His framework treats awareness as information integration we can measure.

Youtube video
Share
The Science of Thought · Explore this series
September 5, 2025
Key Takeaways
  • Tegmark treats consciousness as a physical state of matter with measurable properties.
  • Five principles — including information integration — may distinguish conscious from non-conscious systems.
  • KAN neural networks rediscover physical laws, revealing mathematical structure in understanding.

Max Tegmark flipped the consciousness question on its head. Rather than asking philosophers what consciousness is, the MIT physicist asked what physical properties a system would need to produce it.

His answer points to information processing.

Specifically, the way information integrates across a system, how independently its parts operate, and whether it can store memories and make decisions.

Tegmark calls this hypothetical stuff "perceptronium". It's a thought experiment that treats consciousness like solid, liquid, or gas. A state of matter defined by measurable properties rather than philosophical intuition.

Key figure

5

physical principles Tegmark proposes to distinguish conscious matter from ordinary systems

Physics Absorbs Philosophy's Hardest Problem

The approach builds on Giulio Tononi's Integrated Information Theory, which quantifies consciousness mathematically.

Tegmark extends this framework to arbitrary physical systems, not just biological brains.

His framework identifies five principles that might distinguish conscious systems: information storage, integration across parts, independence of subsystems, complex dynamics, and utility for the organism.

What is integrated information?

A measure of how much a system's parts work together beyond what they'd do separately. Your brain integrates vastly more information than a camera with the same pixel count, because each neuron's activity depends on thousands of others.

The mathematics connects to quantum mechanics and condensed matter physics. Tegmark finds links to error-correcting codes and phase transitions, where systems shift between ordered and disordered states.

AI Reveals Its Own Inner Geometry

Tegmark's recent work explores a related puzzle: how artificial systems develop understanding.

His lab at MIT and the NSF's AI Institute for Artificial Intelligence and Fundamental Interactions developed Kolmogorov-Arnold Networks, or KANs. These neural networks learn differently from standard designs.

Traditional networks have fixed activation functions on nodes.

KANs have learnable functions on connections between nodes. The result: networks that humans can actually interpret.

When trained on physics problems, KANs rediscover known physical laws. They don't just solve equations. They reveal the mathematical structure underlying the solutions.

Understanding Emerges From Structure

Both lines of research share a deeper theme.

Consciousness and understanding might be geometric properties that physical systems can instantiate when configured appropriately.

The practical implications matter. If consciousness has testable physical signatures, we could eventually detect it in systems we can't communicate with. Animal cognition, AI systems, perhaps even collective intelligence in organizations.

Tegmark remains characteristically optimistic about experimental approaches. In an interview discussing his work, he noted: "If you have an idea for an experiment you can build that's just going to cut into some new part of parameter space... just do it."

More than half the time, he suggested, such experiments lead to revolutions.

His next steps include refining the mathematical framework connecting thermodynamics to neural network training. The goal: understanding what physical principles govern learning itself.


Sources

Fact Check: Claim-by-Claim Verification Verified

All claims verified against arXiv papers, MIT faculty pages, and IAIFI documentation. Tegmark's perceptronium framework and KAN work accurately described.

1 Supported
Max Tegmark proposed "perceptronium" in arXiv:1401.1219
Paper hypothesizes consciousness as a state of matter with distinctive information processing properties.
2 Supported
Builds on Tononi's Integrated Information Theory
Paper explicitly generalizes IIT to quantum systems.
3 Supported
Five principles: information, integration, independence, dynamics, utility
All five listed in paper abstract. Some versions emphasize four core plus utility as evolutionary explanation.
4 Mostly supported
MIT/IAIFI developed Kolmogorov-Arnold Networks (arXiv:2404.19756)
KAN paper by Ziming Liu et al. with Tegmark; IAIFI involvement confirmed. NSF AI Institute name is correct: "Institute for Artificial Intelligence and Fundamental Interactions."
5 Supported
KANs have learnable functions on edges, rediscover physical laws
Confirmed in paper: learnable spline functions on edges vs. fixed activations on nodes.
6 Mostly supported
Tegmark quote about experiments
Consistent with Tegmark's public statements; exact source likely from interview or talk.

Commentary

  • Perceptronium is a theoretical proposal, not experimentally confirmed.
  • KAN paper has multiple authors; Tegmark is a co-author, not sole developer.

Sources used for verification

Academic/Peer-reviewed:

Other reliable sources:

Share
Related Articles
AI Consciousness Is Unlikely, Says Neuroscientist Anil Seth

Neuroscientist Anil Seth argues AI consciousness is unlikely without biology. His TED talk lands amid a widening debate over conscious AI, not intuition.

AI In Science Connects the Dots, But Only In Fields That Are Fragmented

An analysis of 80 million papers shows AI boosts originality where knowledge is scattered and connections are weak, but contributes little novelty in structured science.

"Keep Humanity Safe From AI," Urges Pope Leo XIV

Pope Leo XIV's first encyclical reaches the same verdict on AI as the labs building it, then parts ways over the meaning of human limits.

AI Solves Erdős Math Problem: What's Next for AI in Mathematics?

An AI solved an 80-year-old Erdős math problem by walking a path mathematicians had collectively avoided.