Bicameral Labs

A Straight Shot to a Science of Consciousness

The Problem

Consciousness has been a unique holdout for science. The past decades have produced hundreds of neuroscientific theories of consciousness, and yet empirical research has ruled out none of them.

Even the most well-funded and well-intentioned adversarial collaborations have publicly failed to falsify a single theory. Meanwhile, urgent questions, like identifying the necessary and sufficient conditions for artificial consciousness, remain stuck in assumption-dependent quagmires.

The standard academic approach of pet theories, one-off papers, and testing in small-scale neuroscience labs is not progressive or focused enough to solve the most difficult problem in all of science.

The Breakthrough

Our recent research has shown that the strict requirements for falsifiability create narrow bounds that a theory of consciousness must meet, which in turn constrain theories significantly.

In other words, most theories are unfalsifiable, or trivial, or there exist known mathematical theorems that can be used to construct “counterexamples” in the form of hypothetical systems that falsify the theory.

The majority of theories of consciousness are not even wrong. But this is actually good news. A class of theories that are falsifiable, and testable, and non-trivial, do exist, with narrow bounds that we can identify.

The Vision

We are going to make those bounds for theories of consciousness even narrower. And narrower. And narrower. Until most theories have been ruled out, and the shape of a final scientific theory of consciousness reveals itself.

  • Assemble the team to develop and formalize a detailed and extended framework for how to test, and falsify, theories of consciousness. Generate a publicly-accessible knowledge graph containing hundreds of formal definitions, testing setups, and counterexample systems.

    Delineate the bounds that a scientific theory of consciousness must fall within in order to be falsifiable.

  • Apply the logical machinery to winnow hundreds of existing theories down to dozens—or a handful—or none.

    Generate novel candidate classes of theories that fit the framework’s constraints.

  • Establish a taxonomy of provably non-conscious or conscious systems with moral, political, and technological implications (see, e.g., our recent research on how non-trivial theories of consciousness cannot apply to single-hidden-layer feedforward neural networks or even baseline LLMs).

    Rebuild neuroscientific testing of consciousness from scratch to follow the framework.

  • Scale up empirical neuroscientific work to differentiate remaining theories.

    Scientific compression into a final theory of consciousness.

  • An unknown world awaits: hyper-data-efficient learning, creating (or preventing) artificial consciousness, experience engineering, and much more.


The Lead Scientist

Erik Hoel received his PhD in neuroscience from the University of Wisconsin-Madison, working closely with Giulio Tononi on the “theory team” to develop Integrated Information Theory. He was later a postdoctoral researcher at Columbia University working with Rafael Yuste, and then a research assistant professor at Tufts University, working with Michael Levin. He’s been a Forbes 30 Under 30 in Science and a visiting scholar at the Institute for Advanced Study in Princeton. He is the author of The World Behind the World: Consciousness, Free Will, and the Limits of Science, as well as The Revelations. He is known for his highly original scientific work, including his work on multiscale causation and error correction, as well as the Overfitted Brain Hypothesis, and his recent establishment of the formal framework for falsifying theories of consciousness that is behind Bicameral Labs. He blogs at www.theintrinsicperspective.com.