BK x GT

 Tononi, IIT, and Kastrup: Does consciousness emerge by magic?

The passage raises one of the deepest problems in the philosophy of mind: How can physical processes that are supposedly non-conscious give rise to subjective experience?

The dispute between Giulio Tononi’s Integrated Information Theory (IIT) and Bernardo Kastrup’s analytic idealism is not merely about neurons or information. It concerns what consciousness fundamentally is, what a scientific explanation of consciousness must accomplish, and whether matter can explain experience at all.

Kastrup’s criticism is powerful because it targets the transition from unconscious physical processes to conscious experience. But to assess it fairly, we need to distinguish what IIT actually claims from the stronger interpretation Kastrup is challenging.

1. The central problem: Where does experience come from?

Imagine a brain containing billions of neurons.

• Individual neurons transmit electrical and chemical signals.
• Neural circuits process sensory information.
• Networks interact and integrate information.
• At some point, we find a conscious person experiencing the color red, feeling pain, or hearing music.

The scientific descriptions tell us what neurons do. But Kastrup asks a different question:

Why should any of this activity feel like something from the inside?

A neuron firing is a physical event described from the outside. Pain is an experience described from the inside. Even if we establish a precise correlation between a particular neural pattern and pain, we have not automatically explained why that pattern is accompanied by suffering.

This is related to the hard problem of consciousness, famously formulated by David Chalmers.

Three levels of explanation

Level 1: Neural mechanisms

How neurons fire, communicate, and form networks.

Level 2: Functional organization

How information is integrated, made available, or used by a system.

Level 3: Subjective experience

Why there is something it is like to see red, feel pain, or be aware.

The controversy is whether an account of the first two levels can adequately explain the third.

Kastrup argues that theories which begin with fundamentally non-conscious matter have a problem explaining how consciousness appears. IIT attempts to address this by making consciousness fundamental to a system’s intrinsic causal organization rather than treating it simply as an extra product of neural computation.

That distinction is crucial.

2. What does Tononi actually mean by Φ?

The mathematical heart of Integrated Information Theory

Tononi’s Integrated Information Theory is usually abbreviated IIT. Its central quantity, (phi), concerns how much a system’s cause-and-effect structure is irreducible to the structures of its parts considered separately.

In plain language, IIT asks:

How much causal power does a system possess as a unified whole that would be lost if it were divided into independent parts?

Consider two hypothetical systems.

System A: Mostly independent

Components perform separate operations. Splitting them apart may leave much of their individual causal activity intact.

System B: Strongly integrated

Components constrain one another’s possible states. Dividing the system destroys some of its unified cause-and-effect organization.

IIT proposes that the latter kind of irreducible causal organization is directly related to consciousness. In the theory, experience is not merely an output produced after information processing. Consciousness is identified with the system’s intrinsic, maximally irreducible causal structure.

That is a much more ambitious claim than saying that a computer becomes conscious whenever it has processed enough bits.

A critical distinction: information is not automatically consciousness

The word information can mislead us.

A computer can process enormous quantities of information without IIT necessarily assigning it a correspondingly large amount of consciousness. The theory concerns a specific form of causal integration, not simply data volume, complexity, connectivity, or computational speed.

In simplified terms:

• Information: Which possibilities does a system distinguish?
• Integration: How much of that information belongs to the system as an indivisible causal whole?
• Φ: How much integrated information is specified under a particular IIT formulation?
• Consciousness: The subjective experience IIT identifies with the system’s maximally irreducible causal structure.

The last step is the controversial one. The mathematical quantity does not, by itself, demonstrate that subjective experience exists. That identification is a fundamental commitment of the theory.

3. Kastrup’s first challenge: Where is the magical threshold?

Kastrup asks: At what exact point do unconscious neurons become conscious?

Suppose, for the sake of illustration, that a system’s integration gradually increases:

The alleged threshold problem

Low integration

Φ = 0.2

Is the system completely unconscious?

Higher integration

Φ = 0.8

Does some form of experience now exist?

The philosophical question

What physical or causal fact makes the difference between no experience whatsoever and experience?

The numerical values above are illustrative only, not measured thresholds or predictions of IIT.

There are actually three different questions hidden in Kastrup’s objection.

A. Is there a numerical threshold?

Does consciousness suddenly appear only when exceeds a particular universal value?

IIT does not straightforwardly propose a universal numerical threshold at which an otherwise unconscious brain suddenly becomes conscious. It treats consciousness as graded and uses the concept of a maximally irreducible complex to identify the physical substrate associated with an experience. Different formulations of IIT also differ in their mathematical details.

B. Is there a system-selection rule?

Yes. IIT uses a maximum or exclusion principle: the relevant conscious complex is selected according to its maximal irreducible causal structure at the relevant scale.

This is not the same as saying that every neuron is unconscious until it receives one additional bit of information. The relevant entity is the system or complex, not a neuron crossing an arbitrary data-processing threshold.

C. Does the selection rule explain why experience exists?

This is where Kastrup’s criticism has its greatest force.

Even if IIT tells us exactly which system qualifies as conscious, it remains possible to ask why that system’s causal organization should constitute an experience at all.

The mathematical criterion may answer which physical system IIT associates with consciousness. Whether it explains why there is something it is like to be that system is a deeper question.

The important distinction: A threshold rule can be mathematically precise without explaining why crossing the threshold has subjective significance.

4. Kastrup’s second challenge: Correlation is not constitution

Suppose neuroscientists discover that a particular pattern of brain activity reliably accompanies the experience of seeing red.

They can measure:

• The neurons that respond to the stimulus.
• The timing and strength of their activity.
• The causal relationships between neural populations.
• The changes that occur when the stimulus is masked or awareness is lost.

These findings can be scientifically powerful. They may explain how the brain discriminates red from blue, uses color information, and reports what it sees.

But Kastrup asks a different question:

Why is the neural event accompanied by redness as experienced?

Consider the distinction:

The external description

A physical stimulus has a wavelength distribution. The eye and nervous system respond, and a measurable neural process follows.

The subjective description

Red is present in experience. There is a particular qualitative character to seeing it.

The scientific task is to explain the relationship between these descriptions.

An IIT proponent may answer that the two are not separate events connected by an unexplained bridge: the relevant experience is the intrinsic causal structure, described from its intrinsic rather than external perspective.

Kastrup’s response is that identifying two descriptions, or correlating them, is not automatically an explanation of their relationship. If physical processes are defined exclusively by their externally observable properties, it is not obvious how subjective experience follows from those properties.

That argument challenges the explanatory adequacy of physicalism. It does not, by itself, establish that IIT is false, because IIT explicitly tries to redefine the relationship between physical structure and experience.

The key question becomes: Does IIT genuinely explain the identity of neural causation and experience, or does it postulate that identity?

That is a more precise version of Kastrup’s objection.

5. Kastrup’s third challenge: Why should integration produce experience?

This is arguably the strongest part of the passage.

Suppose we know everything about a neural network:

• Every neuron’s state.
• Every connection between neurons.
• Every causal influence between neural components.
• Every possible state transition.
• The system’s calculated .

What additional fact would establish that the network feels anything?

IIT’s answer is radical: if the system satisfies the theory’s requirements, its intrinsic causal structure is the experience. It does not propose an additional mechanism that transforms integration into consciousness.

This produces two very different interpretations.

Interpretation A: Emergence

Physical processes produce consciousness when their organization becomes sufficiently sophisticated.

Kastrup’s objection: More sophisticated processing does not obviously explain why experience exists.

Interpretation B: Identity

Consciousness is not produced as a separate outcome; it is identical to the appropriate intrinsic causal organization.

IIT’s response: Demanding a further process that turns causation into experience misunderstands what consciousness fundamentally is.

Notice what has happened. The argument is no longer simply about whether consciousness appears at a threshold. It concerns what counts as an explanation in the first place.

If experience is identical to a certain causal structure, there is no separate moment when a non-conscious structure magically manufactures experience. But the identity claim itself requires justification.

This is a familiar issue in philosophy. Saying that water is H₂O can be explanatory because chemistry provides a network of independently testable relationships. Saying that experience is identical to integrated information must likewise earn its explanatory status. The identity cannot be established merely by defining both terms together.

6. The combination problem: Can billions of parts make one experiencer?

Kastrup’s challenge has a close relative: the combination problem.

Imagine that individual neurons, or more fundamental constituents of nature, possess some rudimentary form of experience.

How do these tiny experiences combine into one unified human perspective?

Why do you experience one visual scene rather than billions of separate neural experiences? Why is your pain experienced as a single, unified event rather than a collection of unrelated microscopic feelings?

IIT attempts to address this problem differently from traditional panpsychism. Rather than simply adding up the experiences of individual parts, it identifies a maximally irreducible complex whose integrated causal structure corresponds to a unified experience.

However, critics can still ask whether the theory explains the unity of experience or formally characterizes the system it takes to instantiate that unity.

Kastrup takes a different route. In analytic idealism:

• Consciousness is fundamental, rather than something produced by non-conscious matter.
• Individual minds are relatively bounded or dissociated regions of a more fundamental consciousness.
• The brain is the observable, external appearance of processes within consciousness, rather than an object that creates consciousness from nothing.

His approach attempts to reverse the explanatory direction.

IIT asks how a particular physical system’s organization relates to experience.

Kastrup asks how individual, bounded experiences arise within a reality that is fundamentally experiential.

This does not mean analytic idealism has solved every problem. In particular, explaining how individual minds become dissociated from a universal consciousness raises its own questions. But it avoids the specific starting assumption that experience must emerge from something entirely non-experiential.

7. Does IIT really imply that unconscious neurons suddenly light up?

Not in the simple form suggested by Kastrup’s rhetoric.

There are three important corrections.

First, a neuron is not necessarily the relevant unit of consciousness. IIT looks for the maximally irreducible complex, which may involve many interacting elements.

Second, IIT does not simply equate more information with more consciousness. The relevant mathematical quantity concerns irreducibility of causal structure, not the quantity of information processed in an everyday sense.

Third, a system-level change does not necessarily imply a mysterious physical event. A system can change its causal organization as its components and connections change. The disputed issue is whether the theory has justified identifying the resulting organization with subjective experience.

Kastrup’s word magic is therefore a philosophical accusation, not a demonstrated mathematical defect.

Yet it points to a legitimate question: if the theory identifies a physical structure as conscious, what independent grounds justify that identification?

The distinction matters because a theory can be mathematically rigorous while making a metaphysical assumption that is not itself established by its equations.

8. A major challenge to IIT: Can the mathematics over-attribute consciousness?

This concern is not unique to Kastrup.

In a prominent critique, philosopher of science and consciousness theorist William James-style? No—the relevant critic is William Seager? More directly, a 2015 paper by William A. E. (Bill) not necessary to name here? The established critique is Aaronson’s challenge to IIT’s predictions for simple logical networks, alongside later philosophical critiques.

Scott Aaronson, a computer scientist, argued that certain simple artificial networks could obtain surprisingly high values of integrated information despite lacking the sort of sophisticated organization ordinarily associated with consciousness. A peer-reviewed critique also challenged whether validly measures consciousness.

The concern is straightforward:

If a system qualifies as conscious according to a mathematical criterion, but the criterion also attributes consciousness to systems that appear intuitively unpromising, the theory needs to explain why those counterintuitive results should be accepted.

IIT’s defenders can respond that ordinary intuitions are not a reliable test of which systems have experience. That is a reasonable caution: intuition alone cannot settle consciousness.

But the opposite caution is equally important: a mathematical measure does not become a validated measure of consciousness merely because a theory defines it that way.

The crucial issue is whether the theory’s predictions can be independently tested and whether they distinguish consciousness from other physical properties.

9. What does the latest experimental evidence say?

A significant development is the 2025 Nature study, Adversarial testing of global neuronal workspace and integrated information theories of consciousness. It involved 256 participants and compared predictions from IIT and Global Neuronal Workspace Theory using multiple methods of brain measurement.

The findings were mixed.

• Some observations aligned with predictions made by IIT.
• Other results challenged important IIT predictions, particularly the prediction involving sustained synchronization within posterior cortical regions.
• Global Neuronal Workspace Theory also faced challenges to key predictions.

This is important because IIT is not merely an abstract philosophical proposal. It makes claims about the neural organization associated with consciousness that can be investigated experimentally.

However, the study did not establish that consciousness is fundamentally physical, that IIT is entirely false, or that analytic idealism is true. It tested specific predictions of two theories, not every possible metaphysical interpretation of consciousness.

Read the original study: Nature (2025), Adversarial testing of global neuronal workspace and integrated information theories of consciousness⁠.

10. IIT versus Kastrup: The deepest differences

Question	Tononi’s IIT	Kastrup’s analytic idealism
What is fundamental?	Experience is identified with intrinsic causal organization; the theory formalizes its physical substrate.	Consciousness is fundamental; matter is its extrinsic appearance.
What is the brain?	The relevant physical substrate of experience is associated with a maximally irreducible causal complex.	The brain is the observable appearance of processes within consciousness.
Why is experience unified?	The theory links unity to irreducibility and the structure of the relevant complex.	Individual experience is a bounded or dissociated perspective within a wider consciousness.
What is Φ?	A mathematical quantity associated with integrated causal structure.	A physical-system measure that, by itself, does not explain why experience exists.
What is the main explanatory challenge?	Justifying the identification of causal structure with subjective experience.	Explaining individual minds, dissociation, and the relation between experience and the regularities described by physics.

The positions are not simply opposites in every respect. IIT also gives experience a fundamental role in its theoretical starting point, and its defenders reject the idea that consciousness is merely an accidental by-product of computation. The disagreement concerns the precise relationship between experience, physical structure, and the nature of reality.

11. The hidden assumption in Kastrup’s argument

There is a subtle but important issue in the quoted passage.

Kastrup describes neurons as unconscious and then asks how additional information suddenly makes them conscious.

But how do we know that the relevant physical constituents are fundamentally unconscious?

If we begin by assuming that all matter is intrinsically non-experiential, the emergence of experience looks mysterious from the outset.

If we begin by assuming that experience is fundamental, the question changes: we must explain the organization of experience into individual minds and the relation between those experiences and the physical world.

Neither starting point is neutral.

IIT attempts to bridge the gap through a proposed identity between experience and intrinsic causal structure. Analytic idealism attempts to dissolve the gap by rejecting the premise that reality is fundamentally non-conscious matter.

The decisive question is whether either approach can provide sufficient explanatory and empirical justification for its starting assumptions.

12. The strongest version of the criticism

Kastrup’s point can be reconstructed as a three-step argument:

Premise 1: Physical organization can be described mathematically.

We can specify neural states, causal interactions, and measures of integration.

Premise 2: Mathematical specification does not automatically explain subjectivity.

Knowing the structure of a system does not, without further argument, establish why it has a first-person perspective.

Conclusion: IIT needs to justify its identity claim.

It must explain why its proposed causal structure constitutes experience, rather than merely stipulating a mathematical criterion for consciousness.

The argument is strongest when directed at the explanatory gap. It is weaker if interpreted as showing that IIT literally proposes an arbitrary numerical threshold at which ordinary neurons abruptly acquire consciousness.

The latter misrepresents the theory; the former identifies a serious philosophical issue.

13. Ten deep-dive questions worth pursuing

These questions can guide a rigorous discussion of IIT and Kastrup without assuming either position is correct.

1. The hard problem: Can a third-person description of causal structure ever entail a first-person fact about experience?

2. Identity versus emergence: If consciousness is identical to a causal structure, what evidence could establish that identity?

3. The threshold question: Does consciousness have a genuine all-or-nothing onset, or can a system’s capacity for experience change gradually?

4. The combination problem: Does IIT explain the unity of consciousness, or does it select a system whose unity is already built into the theory?

5. The exclusion principle: Why should one maximally irreducible complex take precedence over other overlapping systems?

6. Artificial consciousness: If two systems behave identically but have different intrinsic causal structures, should they have different experiences according to IIT?

7. The empirical challenge: Which experimental result would genuinely count against IIT rather than merely requiring a revised interpretation?

8. Analytic idealism: How could the hypothesis of a universal consciousness be tested against alternative metaphysical explanations?

9. Neuroscience and metaphysics: Do correlations between brain damage and changes in experience establish that brains produce consciousness, or only that brain processes are systematically related to experience?

10. Explanatory progress: Which theory explains more while assuming less: a theory that identifies consciousness with physical causal organization, or one that treats consciousness as fundamental?

Final assessment

Kastrup’s passage is rhetorically forceful, but its most important insight is not that IIT contains a magical numerical threshold.

It is that a precise mathematical description of a physical system and an explanation of subjective experience are not automatically the same thing.

IIT offers a substantive proposal: consciousness is identical to a particular kind of maximally irreducible causal structure. Its success depends on whether this identification can be defended conceptually and tested empirically.

Kastrup offers a different proposal: consciousness is fundamental, and brains are the external appearance of processes within consciousness. Its success depends on whether it can explain the structure of individual minds and the relationship between consciousness and the physical world.

Neither view wins simply by pointing to the other side’s unanswered questions.

The deepest issue remains this: Are we trying to explain how unconscious matter produces consciousness, or have we misunderstood the nature of matter by assuming that consciousness must be produced in the first place?

That is the philosophical fault line separating the two approaches.