After ‘Consciousness’ Workshop
Date: January 27-28, 2026 (Tuesday – Wednesday)
Venue: Rm 10.13, Run Run Shaw Tower, Centennial Campus, HKU
Co-Organized by: HKU-AIH Lab & Berggruen Institute China
Speaker: Professor Keith Frankish, University of Sheffield
Abstract:
The concept of consciousness as a private inner show of phenomenal qualities is broken. It creates the Hard Problem without helping neuroscience or AI policy. We need to replace it. Taking my cue from Dennett, I propose a strange inversion. Consciousness isn’t a play on a mental stage; it’s the roar of the neural crowd. It’s the collective reactivity of neural subsystems–how stimuli modulate attention, perceptual set, memory, affect, cognition, motor control, and the rest. There’s no private phenomenal medium. There are just these reactions, some substantial enough to enable the overt responses we regard as indicative of consciousness.
I introduce two replacement concepts. First, reactivity patterns–substantial, multidimensional patterns of neural reactivity. These do the work phenomenal qualities were supposed to do, but they are measurable. Second, reactivity schemas–simplified internal models the brain uses to track its own reactivity. These explain why we fall for phenomenal realism. They create an introspective illusion, making us think we are acquainted with irreducibly subjective mental qualities when we’re really just monitoring our own complex neural reactions.
This changes how we think about AI. Instead of asking ‘Is this system conscious?’ we ask better questions: What sort of reactivity patterns does it generate? How substantial are these patterns? How do they compare to the ones generated by the human nervous system? Does the system model its own reactivity? These are questions engineers and policymakers can work with.
The view dissolves traditional puzzles of consciousness, including the Hard Problem and the problem of psychophysical harmony. Instead, it directs us to Dennett’s hard question: What does consciousness actually do? It lets us be generous in attributing consciousness to systems (going whole hog, as Cappelen puts it) without committing to spooky phenomenal properties. And it opens up research programmes in neuroscience, AI engineering, and the study of pathologies of consciousness, where reactivity or reactivity monitoring is impaired.
Turn around, stop looking for the inner show, and look at what the inner audience is doing. That’s where consciousness has been all along.
