339 | Ned Block on Whether Consciousness Requires Biology

What It Covers

Philosopher Ned Block challenges computational functionalism in consciousness studies, arguing that biological mechanisms and subconscious processes may be necessary for phenomenal consciousness beyond mere computational input-output functions that current AI systems perform.

Key Questions Answered

• •Phenomenal vs Access Consciousness: Block distinguishes phenomenal consciousness (the subjective what-it-is-like quality of experience) from access consciousness (global information availability for cognition). This distinction matters because machines may achieve access consciousness through computation while lacking phenomenal experience entirely.
• •Computational Functionalism Limits: The view that specific computations alone create consciousness faces a critical challenge: simulating a process computationally does not preserve its physical properties. A computational simulation of gravity produces no gravitational force; similarly, computing consciousness-related functions may not generate actual conscious experience.
• •Electrochemical Processing Advantage: Human neurons translate between chemical and electrical signals across synapses, unlike purely electrical systems. Evolutionary evidence suggests purely electrical nervous systems in ctenophores became evolutionary dead ends, while electrochemical pathways enabled complex consciousness, indicating mechanism matters beyond computation.
• •Substrate Independence with Mechanism Dependence: Block rejects substrate dependence (requiring specific materials like carbon) but emphasizes mechanism dependence. Different ions could potentially replace calcium and potassium in neurons if they preserve the same electrochemical mechanisms, distinguishing his view from strict biological requirements.
• •First-Person Perspective Test: A more convincing consciousness test for AI would train systems exclusively on third-person data like encyclopedias, eliminating all first-person perspective examples. If such systems spontaneously develop first-person viewpoints, this provides stronger evidence for genuine consciousness than current Turing-style tests.