Giulio Tononi: Integrated Information Theory

Biography

Giulio Tononi was born in 1960 in Trento, Italy. He studied medicine at the University of Pisa and the Scuola Normale Superiore, where he earned his medical degree, and he completed his training in neurology and psychiatry. Tononi's early work focused on sleep: he was among the first to argue that sleep serves a fundamental function of synaptic homeostasis, allowing the brain to consolidate learning by renormalizing its synaptic connections. He spent time at the Neurosciences Institute in San Diego, where he collaborated with Gerald Edelman, and he joined the University of Wisconsin-Madison in 2004, where he is the David P. White Professor of Sleep Medicine and a professor of psychiatry. Since 2013 he has also been a Distinguished Professor at the Allen Institute for Brain Science, working with Christof Koch on testing the predictions of his theory.

Historical Background

Tononi developed integrated information theory (IIT) beginning in the 1990s, when the scientific study of consciousness was dominated by two broad approaches. On one side were functional and cognitive theories, which treated consciousness as a kind of information processing — exemplified by the global workspace theory of Bernard Baars and Stanislas Dehaene. On the other side were the philosophers and neuroscientists who insisted that consciousness poses a distinct problem: the hard problem of explaining why there is experience at all, and why experience has the particular structure it does.

Tononi's ambition was to build a theory that answers both questions. In a 2004 paper in BMC Neuroscience and in the 2008 paper "Consciousness as Integrated Information," he proposed that consciousness is not merely correlated with a kind of brain activity but is identical to a specific mathematical quantity: integrated information, denoted by the Greek letter phi. The theory was developed in detail in his 2012 book Phi: A Voyage from the Brain to the Soul and in a series of formal papers with his collaborators. It has since become the most prominent quantitative theory of consciousness, generating both enthusiastic support — most notably from Christof Koch — and intense scientific and philosophical criticism.

Core Ideas

The Axioms of Consciousness

IIT begins not with the brain but with phenomenology. Tononi argues that every conscious experience has five essential properties, which he calls the axioms: existence (experience happens); composition (experience is structured — it has parts); information (experience is specific — it is what it is, and not something else); integration (experience is unified — it cannot be decomposed into independent parts); and exclusion (experience has a definite content and a definite spatiotemporal scale). The theory then asks: what kind of physical system can account for a property with these features? The answer, Tononi argues, is a system that has a cause-effect power structure with corresponding properties.

Consciousness as Integrated Information

The central claim of IIT is that consciousness is identical to integrated information: a system is conscious to the extent that its current state generates information that is irreducible to the information generated by its parts. The degree of consciousness is measured by phi, a quantity defined in terms of the cause-effect repertoire of the system: how much the system's current state constrains its possible past states (causes) and future states (effects), over and above the constraints imposed by its parts considered independently. If phi is zero, the system is not conscious at all; the greater the phi, the more conscious it is. Each experience, moreover, corresponds to a specific structure — called the "quality space" — shaped by the system's cause-effect structure, which is why the experience of red differs from the experience of a musical note.

The Exclusion and the Distribution of Consciousness

IIT yields striking and controversial implications. Because phi is computed for a system as a whole — and because "the whole" must be the maximal set of elements with irreducible cause-effect power (the "exclusion" principle) — the theory implies that the brain's cortex, with its dense, structured connectivity, is the seat of human consciousness, and that consciousness is graded: a system with modest phi has modest consciousness. The theory also implies that consciousness is not confined to brains. Any system with the right kind of cause-effect structure — possibly including some artificial systems, and even simple physical systems with feedback loops — would be conscious to some degree, a conclusion that places IIT in the neighborhood of panpsychism and has generated enormous debate.

Sleep, Anesthesia, and the Measurement of Phi

Tononi's empirical program attempts to measure consciousness. He has argued that consciousness vanishes in deep sleep and anesthesia not because the brain's neurons are silent but because the brain's information integration collapses: different cortical areas lose their capacity to influence one another. Using transcranial magnetic stimulation (TMS) and EEG, Tononi and colleagues developed the "perturbational complexity index" (PCI), which estimates how much the brain responds to a direct perturbation as an integrated system. The PCI distinguishes conscious from unconscious states in coma patients, anesthesia, and sleep — one of the few consciousness measures with direct clinical applications.

Major Works

Phi: A Voyage from the Brain to the Soul (2012) is a fictionalized, Socratic presentation of IIT, tracing the theory's development through dialogues with the spirits of philosophers and scientists. The technical foundations are laid in "Consciousness as Integrated Information: A Provisional Manifesto" (2008) and the collaborative review "Integrated Information Theory: From Consciousness to Its Physical Substrate" (2016) in Nature Reviews Neuroscience. With his colleagues, Tononi has also published the formal mathematical development of IIT 3.0, the current version of the theory, in a series of papers including the 2014 PLoS Computational Biology article and the 2016 Physics of Life Reviews target article.

Philosophical Influence

IIT has transformed the scientific landscape for the study of consciousness. It is the only major theory that provides a formal, quantitative answer to the question of what consciousness is and where it resides, and it has inspired a large research program of empirical tests, theoretical refinement, and philosophical debate. Tononi's work has also reshaped the philosophical conversation. His insistence that consciousness must be characterized from its phenomenal properties first — a methodology he calls "the identity of phenomenology and cause-effect power" — resonates with the phenomenological tradition, while his conclusion that consciousness is widespread in nature aligns him with panpsychists like Galen Strawson and with the "consciousness is fundamental" position defended by David Chalmers and Christof Koch.

The theory has faced sharp criticism. Daniel Dennett has argued that IIT mistakes a mathematical construct for a real phenomenon; others have questioned whether phi is computable in practice and whether the theory's axioms are as self-evident as claimed. The "unfolding argument" from a group of researchers including Scott Aaronson challenges the theory's computational consequences. Nevertheless, IIT has passed a milestone few theories of consciousness reach: it makes quantitative predictions that researchers are actively testing.

Related Concepts

Tononi's theory is a central reference point for the philosophy of neuroscience and for the question of what is consciousness. Its details are treated in what is integrated information theory, and its empirical program is organized around the neural correlate of consciousness. Its implications for artificial systems bear on can AI be conscious and machine consciousness, and its panpsychist flavor connects to panpsychism. Whether IIT succeeds or fails, it has set the standard for what a serious theory of consciousness must attempt: to explain not just what brains do but why there is experience at all.