Internet Encyclopedia of Philosophy
Search
PRIMARY MENUSKIP TO CONTENT
Integrated Information Theory of Consciousness
Integrated Information Theory (IIT) offers an explanation for the nature and source of consciousness. Initially proposed by Giulio Tononi in 2004, it claims that consciousness is identical to a certain kind of information, the realization of which requires physical, not merely functional, integration, and which can be measured mathematically according to the phi metric.
The theory attempts a balance between two different sets of convictions. On the one hand, it strives to preserve the Cartesian intuitions that experience is immediate, direct, and unified. This, according to IIT’s proponents and its methodology, rules out accounts of consciousness such as functionalism that explain experience as a system operating in a certain way, as well as ruling out any eliminativist theories that deny the existence of consciousness. On the other hand, IIT takes neuroscientific descriptions of the brain as a starting point for understanding what must be true of a physical system in order for it to be conscious. (Most of IIT’s developers and main proponents are neuroscientists.) IIT’s methodology involves characterizing the fundamentally subjective nature of consciousness and positing the physical attributes necessary for a system to realize it.
In short, according to IIT, consciousness requires a grouping of elements within a system that have physical cause-effect power upon one another. This in turn implies that only reentrant architecture consisting of feedback loops, whether neural or computational, will realize consciousness. Such groupings make a difference to themselves, not just to outside observers. This constitutes integrated information. Of the various groupings within a system that possess such causal power, one will do so maximally. This local maximum of integrated information is identical to consciousness.
IIT claims that these predictions square with observations of the brain’s physical realization of consciousness, and that, where the brain does not instantiate the necessary attributes, it does not generate consciousness. Bolstered by these apparent predictive successes, IIT generalizes its claims beyond human consciousness to animal and artificial consciousness. Because IIT identifies the subjective experience of consciousness with objectively measurable dynamics of a system, the degree of consciousness of a system is measurable in principle; IIT proposes the phi metric to quantify consciousness.
Table of Contents
The Main Argument
Cartesian Commitments
Axioms
Postulates
The Identity of Consciousness
Some Predictions
Characterizing the Argument
The Phi Metric
The Main Idea
Some Issues of Application
Situating the Theory
Some Prehistory
IIT’s Additional Support
IIT as Sui Generis
Relation to Panpsychism
Relation to David Chalmers
Implications
The Spectrum of Consciousness
IIT and Physics
Artificial Consciousness
Constraints on Structure/Architecture
Relation to “Silent Neurons”
Objections
The Functionalist Alternative
Rejecting Cartesian Commitments
Case Study: Access vs. Phenomenal Consciousness
Challenging IIT’s Augmentation of Naturalistic Ontology
Aaronson’s Reductio ad Absurdum
Searle’s Objection
References and Further Reading
- The Main Argument
IIT takes certain features of consciousness to be unavoidably true. Rather than beginning with the neural correlates of consciousness (NCC) and attempting to explain what about these sustains consciousness, IIT begins with its characterization of experience itself, determines the physical properties necessary for realizing these characteristics, and only then puts forward a theoretical explanation of consciousness, as identical to a special case of information instantiated by those physical properties. “The theory provides a principled account of both the quantity and quality of an individual experience… and a calculus to evaluate whether a physical system is conscious” (Tononi and Koch, 2015).
a. Cartesian Commitments
IIT takes Descartes very seriously. Descartes located the bedrock of epistemology in the knowledge of our own existence given to us by our thought. “I think, therefore I am” reflects an unavoidable certainty: one cannot deny one’s own existence as a thinker even if one’s particular thoughts are in error. For IIT, the relevance of this insight lies in its application to consciousness. Whatever else one might claim about consciousness, one cannot deny its existence.
i. Axioms
IIT takes consciousness as primary. Before speculating on the origins or the necessary and sufficient conditions for consciousness, IIT gives a characterization of what consciousness means. The theory advances five axioms intended to capture just this. Each axiom articulates a dimension of experience that IIT regards as self-evident.
First, following from the fundamental Cartesian insight, is the axiom of existence. Consciousness is real and undeniable; moreover, a subject’s consciousness has this reality intrinsically; it exists from its own perspective.
Second, consciousness has composition. In other words, each experience has structure. Color and shape, for example, structure visual experience. Such structure allows for various distinctions.
Third is the axiom of information: the way an experience is distinguishes it from other possible experiences. An experience specifies; it is specific to certain things, distinct from others.
Fourth, consciousness has the characteristic of integration. The elements of an experience are interdependent. For example, the particular colors and shapes that structure a visual conscious state are experienced together. As we read these words, we experience the font-shape and letter-color inseparably. We do not have isolated experiences of each and then add them together. This integration means that consciousness is irreducible to separate elements. Consciousness is unified.
Fifth, consciousness has the property of exclusion. Every experience has borders. Precisely because consciousness specifies certain things, it excludes others. Consciousness also flows at a particular speed.
ii. Postulates
In isolation, these axioms may seem trivial or overlapping. IIT labels them axioms precisely because it takes them to be obviously true. IIT does not present them in isolation. Rather, they motivate postulates. Sometimes the IIT literature refers to phenomenological axioms and ontological postulates. Each axiom leads to a corresponding postulate identifying a physical property. Any conscious system must possess these properties.
First, the existence of consciousness implies a system of mechanisms with a particular cause-effect power. IIT regards existence as inextricable from causality: for something to exist, it must be able to make a difference to other things, and vice versa. (What would it even mean for a thing to exist in the absence of any causal power whatsoever?) Because consciousness exists from its own perspective, the implied system of mechanisms must do more than simply have causal power; it must have cause-effect power upon itself.
Second, the compositional nature of consciousness implies that its system’s mechanistic elements must have the capacity to combine, and that those combinations have cause-effect power.
Third, because consciousness is informative, it must specify, or distinguish one experience from others. IIT calls the cause-effect powers of any given mechanism within a system its cause-effect repertoire. The cause-effect repertoires of all the system’s mechanistic elements taken together, it calls its cause-effect structure. This structure, at any given point, is in a particular state. In complex structures, the number of possible states is very high. For a structure to instantiate a particular state is for it to specify that state. The specified state is the particular way that the system is making a difference to itself.
Fourth, consciousness’s integration into a unified whole implies that the system must be irreducible. In other words, its parts must be interdependent. This in turn implies that every mechanistic element must have the capacity to act as a cause on the rest of the system and to be affected by the rest of the system. If a system can be divided into two parts without affecting its cause-effect structure, it fails to satisfy the requirement of this postulate.
Fifth, the exclusivity of the borders of consciousness implies that the state of a conscious system must be definite. In physical terms, the various simultaneous subgroupings of mechanisms in a system have varying cause-effect structures. Of these, only one will have a maximally irreducible cause-effect structure. This is called the maximally irreducible conceptual structure, or MICS. Others will have smaller cause-effect structures, at least when reduced to non-redundant elements. Precisely this is the conscious state.
b. The Identity of Consciousness
IIT accepts the Cartesian conviction that consciousness has immediate, self-evident properties, and outlines the implications of these phenomenological axioms for conscious physical systems. This characterization does not exhaustively describe the theoretical ambition of IIT. The ontological postulates concerning physical systems do not merely articulate necessities, or even sufficiencies, for realizing consciousness. The claim is much stronger than this. IIT identifies consciousness with a system’s having the physical features that the postulates describe. Each conscious state is a maximally irreducible conceptual structure, which just is and can only be a system of irreducibly interdependent physical parts whose causal interaction constitutes the integration of information.
An example may help to clarify the nature of IIT’s explanation of consciousness. Our experience of a cue ball integrates its white color and spherical shape, such that these elements are inseparably fused. The fusion of these elements constitutes the structure of the experience: the experience is composed of them. The nature of the experience informs us about whiteness and spherical shape in a way that distinguishes it from other possible experiences, such as of a blue cube of chalk. This is just a description of the phenomenology of a simple experience (perhaps necessarily awkward, because it articulates the self-evident). Our brain generates the experience through neurons physically communicating with one another in systems linked by cause-effect power. IIT interprets this physical communication as the integration of information, according to the various constraints laid out in the postulates. The neurobiology and phenomenology converge.
Theories of consciousness need to account for what is sometimes termed the “binding problem.” This concerns the unity of conscious experience. Even a simple experience like viewing a cue ball unites different elements such as color, shape, and size. Any theory of consciousness will need to make sense of how this happens. IIT’s account of the integration of information may be understood as a response to this problem.
According to IIT, the physical state of any conscious system must converge with phenomenology; otherwise the kind of information generated could not realize the axiomatic properties of consciousness. We can understand this by contrasting two kinds of information. First, there is Shannon information: When a digital camera takes a picture of a cue ball, the photodiodes operate in causal isolation from one another. This process does generate information; specifically, it generates observer-relative information. That is, the camera generates the information of an image of a cue ball for anyone looking at that photograph. The information that is the image of the cue ball is therefore relative to the observer; such information is called Shannon information. Because the elements of the system are causally isolated, the system does not make a difference to itself. Accordingly, although the camera gives information to an observer, it does not generate that information for itself. By contrast, consider what IIT refers to as intrinsic information: Unlike the digital camera’s photodiodes, the brain’s neurons do communicate with one another through physical cause and effect; the brain does not simply generate observer-relative information, it integrates intrinsic information. This information from its own perspective just is the conscious state of the brain. The physical nature of the digital camera does not conform to IIT’s postulates and therefore does not have consciousness; the physical nature of the brain, at least in certain states, does conform to IIT’s postulates, and therefore does have consciousness.
To identify consciousness with such physical integration of information constitutes an ontological claim. The physical postulates do not describe one way or even the best way to realize the phenomenology of consciousness; the phenomenology of consciousness is one and the same as a system having the properties described by the postulates. It is even too weak to say that such systems give rise to or generate consciousness. Consciousness is fundamental to these systems in the same way as mass or charge is basic to certain particles.
i. Some Predictions
IIT’s conception of consciousness as mechanisms systematically integrating information through cause and effect lends itself to quantification. The more complex the MICS, the higher the level of consciousness: the corresponding metric is phi. Sometimes the IIT literature uses the term “prediction” to refer to implications of the theory whose falsifiability is a matter of controversy. This section will focus on more straightforward cases of prediction, where the evidence is consistent with IIT’s claims. These cases provide corroborative evidence that enhance the plausibility of IIT.
Deep sleep states are less experientially rich than waking ones. IIT predicts, therefore, that such sleep states will have lower phi values than waking states. For this to be true, analysis of the brain during these contrasting states would have to show a disparity in the systematic complexity of non-redundant mechanisms. On IIT, this disparity of MICS complexity directly implies a disparity in the amount of conscious integrated information, because the MICS is identical to the conscious state. The neuroscientific findings bear out this prediction.
IIT cites similar evidence from the study of patients with brain damage. For example, we already know that among vegetative patients, there are some whose brain scans indicate that they can hear and process language: When researchers prompt such patients to think about playing tennis, the appropriate areas of the brain become activated. Other vegetative patients do not respond this way. Naturally, this suggests that the former have a richer degree of consciousness than the latter. When analyzed according to IIT’s theory, the former have a higher phi metric than the latter; once again, IIT has made a prediction that receives empirical confirmation. IIT also claims that findings in the analysis of patients under anesthesia corroborate its claims.
In all these cases, one of two things happens. First, as consciousness fades, cortical activity may become less global. This reversion to local cortical activity constitutes a loss of integration: The system no longer is communicating across itself in as complex a way as it had. Second, as consciousness fades, cortical activity may remain global, but become stereotypical, consisting in numerous redundant cause-effect mechanisms, such that the informational achievement of the system is reduced: a loss of information. As information either becomes less integrated or becomes reduced, consciousness fades, which IIT takes as empirical support of its theory of consciousness as integrated information.
c. Characterizing the Argument
IIT combines Cartesian commitments with claims about engineering that it interprets, in part by citing corroborative neuroscientific evidence, as identifying the nature of consciousness. This borrows from recognizable traditions in the field of consciousness studies, but the structure of the argument is novel. While IIT’s proponents strive for clarity in the exposition of their work by breaking it down into the simpler elements of axioms, propositions, and identity claims, the nature of the relations between these parts remains largely implicit in the IIT literature. To evaluate the explanatory success or failure of IIT, it should prove helpful to attempt an explication of these logical relations. This requires characterizing the relationship of the axioms with the postulates, and of the identity claims with the axioms, postulates, and supporting neuroscientific evidence.
The axioms, of course, count
© Copyright Internet Encyclopedia of Philosophy and its Authors | ISSN 2161-0002
…partial copy
