Consciousness (open question)

⑂ Comparison. The views on this question, side by side under the same headings. The table is not a score: the number of pieces of evidence says nothing about their strength or truth. Details and sources are on each view’s own page.

Views shownGlobal (neuronal) workspace theoryIntegrated information theory (IIT)Higher-order theories of consciousness

	Idea / theory / claim · 1988Global (neuronal) workspace theory	Idea / theory / claim · 2004Integrated information theory (IIT)	Idea / theory / claim · 2005Higher-order theories of consciousness
For everyone	According to this theory many specialised processes run in the brain at the same time and outside awareness. Information becomes conscious when it is broadcast to wide areas of the brain at once, that is, when it enters a shared "workspace". Baars proposed the theory in 1988 using a theatre metaphor. Dehaene and colleagues turned it into a model based on long-range neurons linking frontal and parietal regions in particular.	According to this theory, proposed by Giulio Tononi in 2004, consciousness is a system's capacity to generate integrated information over and above its parts. This capacity is measured by a quantity called Φ (phi). On the theory, consciousness is tied to tightly connected structures in posterior regions of the brain and is not unique to brains. Some researchers regard it as a principled theoretical basis for consciousness, while others criticise it on the grounds that its core claims cannot be tested.	According to higher-order theories, a mental state is conscious only when the mind represents that state to itself. Seeing a red apple, for example, becomes conscious only when accompanied by a higher-order representation registering "I am seeing red". The view was developed in philosophy by David Rosenthal; in neuroscience it is proposed that prefrontal regions play a role in this higher-order representation.
Claims	Information becomes conscious when it is broadcast to a wide brain network and becomes usable by many processes.	Consciousness is identical to the amount of integrated information (Φ) a system has that cannot be reduced to its parts.	A mental state is conscious only when accompanied by a suitable higher-order representation that one is in that state.
Supporting	Experiment In experiments with visible and masked (unconscious) words, only the seen stimuli produced a widespread, sudden rise in activity in fronto-parietal regions.	Experiment In the preregistered adversarial experiment, the presence of information about conscious content in posterior regions of the brain agreed with some of the theory's predictions.	—
Challenging	Experiment In a preregistered adversarial experiment, the frontal re-ignition at stimulus offset predicted by the theory was not observed, and some dimensions of conscious content were only weakly represented in frontal cortex.	Experiment In the same experiment, the sustained synchronisation in posterior regions predicted by the theory was not observed for the duration of conscious perception.	—
Positions	—	Researchers who signed a critical letter (2023, published 2025): IIT's core claims are untestable even in principle, so it should not be treated as a scientific theory. Proponents of the theory: IIT has testable predictions that have been tested in preregistered experiments; its counter-intuitive consequences deserve to be taken seriously as claims about the nature of consciousness.	First-order theories: Suitably processed representations in sensory areas are enough for conscious experience; higher-order representation enables reporting about experience, not experience itself.
Open questions	—	—	—
Sources	3 sources →	3 sources →	2 sources →