Questions: Integrated Information Theory of Consciousness

Question 1

The cerebellum has roughly four times as many neurons as the cortex and performs highly complex computations. Yet damage to the cerebellum rarely affects conscious experience. What does IIT predict about the cerebellum's phi (Φ)?

• A. High Φ — the cerebellum has more neurons and therefore integrates more information
• B. High Φ — complex computation is the best indicator of integration in IIT
• C. Low Φ — the cerebellum's modular, feedforward architecture means each sub-circuit processes independently
• D. Low Φ — the cerebellum is too small relative to the cortex to contribute to consciousness

Answer: (C)

IIT predicts that the cerebellum has low Φ because its architecture is highly modular and feedforward — each sub-circuit processes its inputs largely independently of the others. Integrated information measures how much the *whole system* generates beyond its *parts in isolation*. A modular system, no matter how computationally powerful, has low integration because you can recover its outputs by examining its parts separately. This is why computational complexity (options A and B) is not what IIT measures — a computer with massive parallel processing could have near-zero Φ.

Question 2

According to IIT, a standard digital computer with highly modular, parallel processing architecture is predicted to have very low consciousness despite processing vast amounts of information. What is the IIT explanation for this?

• A. Computers lack biological neurons and cannot support the biochemistry required for consciousness
• B. Computers process information too quickly for integration to occur across modules
• C. The computer's modules process information independently, so the whole system generates little information beyond what its parts generate separately
• D. Computers have no intrinsic causal power because they follow programmed instructions

Answer: (C)

IIT's measure Φ specifically captures information generated by the system *as a whole beyond its parts*. A computer with independent parallel modules can be decomposed into those modules with almost no loss of information about the system's state — meaning its Φ approaches zero. Biological consciousness (on IIT's account) arises from dense, recurrent, causally integrated networks like the thalamocortical system, not from the sheer volume of information processed. Biological substrate (option A) is not IIT's criterion; it is causal architecture that matters.

Question 3

According to IIT, the degree of consciousness of a system is proportional to how much total information the system processes per unit time.

• T. True
• F. False

Answer: False

False. IIT measures *integrated* information — how much information the whole system generates *beyond what its parts generate independently*. A system can process enormous total amounts of information and still have near-zero Φ if its architecture is modular (parts can be separated without losing information about the whole). The cerebellum and a modern computer are canonical examples: massive computation, low integration, low predicted consciousness. This is one of IIT's most counterintuitive — and controversial — implications.

Question 4

IIT implies that any physical system with sufficiently high integrated information (Φ) has some degree of conscious experience, even if it is not biological.

• T. True
• F. False

Answer: True

True. This panpsychist implication follows directly from IIT's axioms. If consciousness is identical to (not merely correlated with) integrated information, then any system with high enough Φ — regardless of its physical substrate — has conscious experience. This means thermostats, simple networks, and even some non-biological systems might have minimal consciousness if their causal structure is sufficiently integrated. This is not an accident or a bug in IIT; it follows necessarily from defining Φ as what consciousness *is*, not just what it correlates with.

Question 5

Why does IIT lead to panpsychist conclusions, and why do many philosophers find this implication troubling?

Model answer: IIT defines Φ such that it *constitutes* consciousness rather than merely correlating with it. Since Φ can be calculated for any physical system with a causal structure, any system with nonzero Φ — no matter how simple or non-biological — has some degree of consciousness by definition. The panpsychist conclusion (even rocks or simple circuits have minimal experience) follows necessarily. Philosophers find this troubling for two reasons: first, it seems deeply counterintuitive to attribute any experience to a thermostat or a simple logic gate; second, panpsychism raises the 'combination problem' — how do vast numbers of micro-conscious units combine into the unified, rich experience of a human subject?

The panpsychist implication reveals a deep tension in IIT: the same axioms that make it mathematically precise and testable also commit it to metaphysical claims that many find implausible. The theory's defenders argue that our intuitions about what can be conscious are unreliable; critics argue the implication is a reductio ad absurdum of the framework.