Global Neuronal Workspace Theory (GNW) is the most empirically productive framework in consciousness science. It has generated more experiments, more replicated findings, and more clinical applications than any competitor. It is also a theory of information routing that explains nothing about consciousness itself.

That’s not a contradiction. GNW identifies the neural signatures that accompany conscious access — ignition thresholds, P3b components, fronto-parietal activation patterns. This is genuinely useful science. Clinicians use GNW-derived measures like the Perturbational Complexity Index (PCI) to assess disorders of consciousness. Experimentalists use GNW’s predictions to design rigorous studies.

But none of this tells us why any of it feels like something. And mistaking the correlation for the explanation has shaped the field’s experimental priorities for two decades.

What GNW actually claims

The theory, developed by Bernard Baars and later formalized by Stanislas Dehaene and Jean-Pierre Changeux, identifies consciousness with global broadcast. A mental state becomes conscious when it enters a “workspace” — a network of fronto-parietal neurons with long-range connections — and gets broadcast to all connected cortical areas.

Below a certain signal strength, information stays local and unconscious. Above the threshold, it “ignites” — triggering widespread activation that makes the information available for verbal report, flexible behavior, and memory encoding.

This is a clean, testable model. It predicts a sharp transition between unconscious and conscious processing (the ignition threshold). It predicts specific neural signatures (late fronto-parietal activation). It predicts that stronger, attended stimuli are more likely to become conscious than weak, unattended ones.

All of these predictions have been confirmed, more or less.

And none of them explain consciousness.

The radio analogy

A radio broadcasts information globally. When you turn up the volume past a threshold, the signal becomes audible throughout the room. The broadcast mechanism is real and measurable.

But the broadcast doesn’t create the music. It distributes it.

GNW tells us how the brain distributes information — how signals get routed from local processing to global availability. It does not tell us why this distribution produces subjective experience. “Information was broadcast to fronto-parietal neurons” is a description of what happened in the brain, not an explanation of what happened in the mind.

This matters because consciousness science is supposed to explain the hard problem: why does any physical process feel like something from the inside? GNW’s answer is, effectively: “it feels like something when the signal is broadcast globally.” But why? What is it about global broadcast that creates experience?

GNW has no answer. Not a wrong answer — no answer. The question is outside its scope. And that would be fine if the theory presented itself as a theory of cognitive access. But it presents itself as a theory of consciousness.

Four problems with GNW’s explanatory claims

1. The threshold is trivially true

GNW’s core empirical prediction — that stimuli above a certain threshold become conscious while those below it don’t — is trivially true of any information-processing system. In your email inbox, messages above a spam-filter threshold reach your attention; messages below it don’t. The filter mechanism is real, but it doesn’t explain reading comprehension.

“Strong signals are processed more thoroughly than weak signals” is a description of signal processing, not an explanation of consciousness. A hearing aid amplifies signals above a threshold. The amplification doesn’t explain hearing.

2. Consciousness is not unified

GNW claims that global broadcast creates a single, unified conscious state. But consciousness doesn’t work that way. Right now, you’re reading these words while simultaneously experiencing:

  • Peripheral vision you’re not directly attending to
  • Background sounds you’re half-aware of
  • Proprioceptive awareness of your body position
  • A vague emotional tone (boredom? curiosity? skepticism?)
  • Perhaps a background thought about something else entirely

These aren’t components of a single unified state. They’re semi-independent streams of varying intensity and clarity. Split attention is the norm, not the exception. Dissociative identity disorder, split-brain patients, and ordinary phenomenology all demonstrate that consciousness is layered, partial, and sometimes contradictory.

The “unity of consciousness” is an artifact of forcing experience through the bottleneck of verbal report. When you describe your experience in words, you linearize it into a single narrative. But the experience itself was never singular.

3. Flexible behavior doesn’t require a workspace

If GNW’s architecture is required for flexible behavior, then organisms without a fronto-parietal workspace should be automatons. They’re not.

Honeybees perform waggle dances encoding distance and direction to food sources. Jumping spiders plan multi-step detour routes to reach prey they can no longer see. Corvids — whose brains lack cortical columns entirely (they have a nuclear pallium instead) — use tools, plan for the future, and recognize themselves in mirrors.

Octopuses, with their distributed nervous system and no central workspace analogous to mammalian cortex, solve novel problems, learn by observation, and appear to experience something resembling curiosity.

If global broadcast were the mechanism of consciousness (rather than a feature of mammalian neural architecture that happens to correlate with it), these organisms shouldn’t be capable of flexible, context-sensitive behavior. They are.

4. Consciousness can override attention

GNW places attention as the gateway to consciousness: information becomes conscious when attention selects it for global broadcast.

But involuntary conscious experience refutes this. PTSD intrusions are vivid, fully conscious experiences that override deliberate attentional control. You can’t gate them. Chronic pain cannot be removed from consciousness by redirecting attention, no matter how hard you try. Dreams occur without attentional deployment.

The attentional gateway model gets the causal direction wrong. Conscious attention — the experience of “deciding to pay attention” — follows the actual attentional selection, which is an implicit, automatic process. Your brain’s implicit filtering system decides what to process before your conscious mind “decides” to attend to it. The conscious experience of choosing to pay attention is itself a product of the system, not a cause.

The substrate specificity problem

GNW requires specifically mammalian fronto-parietal infrastructure. The “global workspace” is defined in terms of cortical architecture — long-range connections between prefrontal and parietal cortex, with specialized neurons capable of sustaining ignition.

But if consciousness is about the functional role of global broadcast, then any architecture that achieves similar information integration should produce consciousness. Small-world network properties ensure that any cortical region is reachable within 3–6 synaptic hops regardless of long-range connectivity. The “global workspace” may be a substrate-specific optimization — faster routing — rather than a consciousness mechanism.

This is where the theory faces a dilemma. Either GNW is about a specific neural architecture (in which case it can’t be a general theory of consciousness), or it’s about the functional property of global information availability (in which case most nervous systems already have it and the theory loses its predictive specificity).

What actual explanation looks like

An actual theory of consciousness would tell us why broadcasting produces experience — not just that it correlates with it.

The Four-Model Theory offers one such account. In FMT, global broadcast accelerates information integration within what the theory calls the “explicit” simulation layer. But the simulation — not the broadcast — is where experience resides. Consciousness is what it’s like to be a system running a real-time self-simulation at the right dynamical regime.

Under this view, GNW identified a real mechanism (information broadcasting) that supports consciousness by feeding the simulation. But it mistook the information pipeline for the experience itself. It’s like discovering that roads are necessary for commerce and concluding that roads are commerce.

Other theories attempt genuine explanation too. IIT tries to explain consciousness through intrinsic causal power. Predictive processing locates it in prediction error minimization. These attempts may be wrong, but at least they’re attempting to answer the right question.

GNW isn’t even trying. It identified the plumbing and declared the plumbing is the water.

The cost of confusing correlation with explanation

This matters beyond academic philosophy. Two decades of consciousness experiments have been shaped by GNW’s framing. The field has gotten extraordinarily good at measuring when things become conscious — ignition timing, P3b latency, fronto-parietal activation patterns.

The field has made almost no progress on why any neural process produces experience.

This is partly because GNW’s correlational success created the illusion that the explanatory work was done. If your most successful theory says “consciousness = global broadcast” and your experiments keep confirming the neural signatures of global broadcast, it feels like progress. And it is progress — in neuroscience. But it’s not progress in consciousness science.

The COGITATE adversarial collaboration — where GNW and IIT were put to a direct pre-registered test and neither won cleanly — may be a healthy correction. But the field’s experimental infrastructure is now optimized for access-consciousness research, not for Hard Problem research.

Conclusion

GNW is the best theory of cognitive access and information routing in the brain. It is not a theory of consciousness. These are different claims, and calling it a theory of consciousness has been productive for cognitive neuroscience while diverting consciousness science from its central question.

The field needs theories that explain why — not just when and how — neural processing produces experience. GNW’s empirical tools remain invaluable. Its explanatory claims should be retired.

For an alternative framework that attempts to explain why processing feels like something, see The Four-Model Theory of Consciousness — An Introduction. The full theory is described in the research paper and in The Simulation You Call “I”.