consciousness · Part 3

Integrated Information Theory

Giulio Tononi's mathematical framework for consciousness — a theory that dares to measure experience.

A Bold Claim

What if consciousness could be measured? Not just detected, but quantified — given a number that tells you how conscious a system is?

This is the ambition of Integrated Information Theory (IIT), developed by neuroscientist Giulio Tononi. It is perhaps the most rigorous attempt to formalize what consciousness is.

The Core Idea: Phi (Φ)

IIT proposes that consciousness corresponds to integrated information — the amount of information generated by a system as a whole, above and beyond its parts.

The quantity Φ (phi) measures this integration. A system with high Φ is highly conscious. A system with Φ = 0 has no experience at all.

The Five Axioms

IIT starts from what we know most certainly — the properties of experience itself:

  1. Existence: Experience exists (cogito ergo sum)
  2. Composition: Experience is structured (it has parts)
  3. Information: Experience is specific (this experience, not that one)
  4. Integration: Experience is unified (not decomposable)
  5. Exclusion: Experience is definite (has particular borders)

Surprising Implications

IIT leads to conclusions that many find uncomfortable:

  • A sufficiently integrated network — even a non-biological one — would be conscious
  • A digital computer running a perfect brain simulation might have zero Φ (and thus zero consciousness)
  • Consciousness might be far more widespread in nature than we assume

Criticisms

IIT is not without critics. Computing Φ for any realistic system is currently intractable. Some philosophers argue it confuses correlation with constitution. Others question whether consciousness can be captured by any single number.

Yet IIT remains the most mathematically precise theory of consciousness we have — and its willingness to make testable predictions sets it apart from many competitors.