The Integrated Information Theory Needs Attention

<i>Human and Animal Minds: The Consciousness Questions Laid to Rest</i>

What does it feel like to be a bat? Is conscious experience of echolocation closer to that of vision or audition? Or do bats process echolocation nonconsciously, such that they do not feel anything about echolocation? This famous question of bats' experience, posed by a philosopher Thomas Nagel in 1974, clarifies the difficult nature of the mind–body problem. Why a particular sense, such as vision, has to feel like vision, but not like audition, is totally puzzling. This is especially so given that any conscious experience is supported by neuronal activity. Activity of a single neuron appears fairly uniform across modalities and even similar to those for non‐conscious processing. Without any explanation on why a particular sense has to feel the way it does, researchers cannot approach the question of the bats' experience. Is there any theory that gives us a hope for such explanation? Currently, probably none, except for one. Integrated information theory (IIT) has potential to offer a plausible explanation. IIT essentially claims that any system that is composed of causally interacting mechanisms can have conscious experience. And precisely how the system feels is determined by the way the mechanisms influence each other in a holistic way. In this article, I will give a brief explanation of the essence of IIT. Further, I will briefly provide a potential scientific pathway to approach bats' conscious experience and its philosophical implications. If IIT, or its improved or related versions, is validated enough, the theory will gain credibility. When it matures enough, predictions from the theory, including nature of bats' experience, will have to be accepted. I argue that a seemingly impossible question about bats' consciousness will drive empirical and theoretical consciousness research to make big breakthroughs, in a similar way as an impossible question about the age of the universe has driven modern cosmology.

Mental causation is vitally important to the integrated information theory (IIT), which says consciousness exists since it is causally efficacious. While it might not be directly apparent, metaphysical commitments have consequential entailments concerning the causal efficacy of consciousness. Commitments regarding the ontology of consciousness and the nature of causation determine which problem(s) a view of consciousness faces with respect to mental causation. Analysis of mental causation in contemporary philosophy of mind has brought several problems to the fore: the alleged lack of psychophysical laws, the causal exclusion problem, and the causal pairing problem. This article surveys the threat each problem poses to IIT based on the different metaphysical commitments IIT theorists might make. Distinctions are made between what I call reductive IIT, non-reductive IIT, and non-physicalist IIT, each of which make differing metaphysical commitments regarding the ontology of consciousness and nature of causation. Subsequently, each problem pertaining to mental causation is presented and its threat, or lack thereof, to each version of IIT is considered. While the lack of psychophysical laws appears unthreatening for all versions, reductive IIT and non-reductive IIT are seriously threatened by the exclusion problem, and it is difficult to see how they could overcome it while maintaining a commitment to the causal closure principle. Yet, non-physicalist IIT denies the principle but is therefore threatened by the pairing problem, to which I have elsewhere provided a response that is briefly outlined here. This problem also threatens non-reductive IIT, but unlike non-physicalist IIT it lacks an evident response. The ultimate aim of this survey is to provide a roadmap for IIT theorists through the maze of mental causation, by clarifying which commitments lead to which problems, and how they might or might not be overcome. Such a survey can aid IIT theorists as they further develop and hone the metaphysical commitments of IIT.

First-person experience cannot rescue causal structure theories from the unfolding argument

The purpose of this article is to provide an overall critical appraisal of Integrated Information Theory (IIT) of consciousness. We explore how it has evolved and what problems are involved in the theory. IIT is a hypothesis that explains the consciousness in terms of integrated information. It argues that the fundamental properties of experience can be properly analyzed and explained by physical systems’ informational properties. Throughout the last decade, there have been many advances in IIT’s theoretical structure and mathematical model. Also, like all hypotheses in the field of science of consciousness, IIT has given rise to several controversies and issues. In this context, IIT needs a critical survey. To this end, we first introduce fundamental concepts of IIT and related issues. After that, we discuss major transitions IIT has been through and point out related intra-model issues. Finally, in the last section, some theoretical, extra-model issues involved in IIT’s principles are presented. The article concludes by suggesting that, for the sake of future development, IIT should take metacognitive accessibility to experience more seriously.

The Equations for Consciousness: A Reply to “Tracking the Travels,” a Review of <i>Journey of the Mind</i>

The problem of explaining the relationship between subjective experience and physical reality remains difficult and unresolved. In most explanations, consciousness is epiphenomenal, without causal power. The most notable exception is Integrated Information Theory (IIT), which provides a causal explanation for consciousness. However, IIT relies on an identity between subjectivity and a particular type of physical structure, namely with an information structure that has intrinsic causal power greater than the sum of its parts. Any theory that relies on a psycho-phyiscal identity must eventually appeal to panpsychism, which undermines that theory's claim to be fundamental. IIT has recently pivoted towards a strong version of causal emergence, but macroscopic structures cannot be stronger causally than their microphysical parts without some new physical law or governing principle. The approach taken here is designed to uncover such a principle. The decisive argument is entirely deductive from initial premises that are phenomenologically certain. If correct, the arguments prove that conscious experience is sufficient to create additional degrees of causal freedom independently of the content of experience, and in a manner that is unpredictable and unobservable by any temporally sequential means. This provides a fundamental principle about consciousness, and a conceptual bridge between it and the physics describing what is experienced. The principle makes testable predictions about brain function, with notable differences from IIT, some of which are also empirically testable.

BackgroundThe question of whether large language models (LLMs) possess consciousness has been increasingly debated. Integrated information theory (IIT) offers a quantitative framework for assessing consciousness through a measure of integrated information.MethodsThis study applied IIT principles to the architecture of transformer-based LLMs, focusing on causal integration, temporal persistence, and system irreducibility. Ablation experiments on Generative Pretrained Transformer 2 (GPT-2) were performed, selectively removing individual attention heads and measuring changes in perplexity as a behavioral proxy for integrated information to empirically approximate the measure of integrated information.ResultsThe ablation study of a single attention head produced minimal or negative changes in perplexity in four out of five representative sentences, indicating redundancy or noise. Only one sentence revealed a significant increase in perplexity change (ΔPPL +11.29), reflecting a localized but nonessential contribution. A comparison with biological systems demonstrated that LLMs meet the IIT criterion of differentiation, but fail to meet the criteria of integration, causal closure, and temporal persistence. These findings confirm that LLMs are architecturally decomposable, lack persistent internal states, and do not sustain global causal irreducibility. Philosophical considerations, including Searle’s Chinese Room argument, further support the idea that the linguistic fluency of LLMs arises from syntactic manipulation rather than semantic understanding.ConclusionCurrent LLMs do not satisfy the structural and informational requirements of consciousness under IIT. Although capable of simulating intelligent language, LLMs remain unconscious systems with a negligible amount of integrated information, underscoring the distinction between linguistic competence and conscious experience.

Integrated information theory (IIT) of consciousness proposes an identity between its causal structure and phenomenology. Through this assertion, IIT aims to explain consciousness by prioritizing first-person experience. However, despite its phenomenology-first stance, developments in IIT have overlooked temporality. As such, we argue that at present IIT’s phenomenological analysis is incomplete. In this critique, we show how IIT takes a non-identical illusionist stance towards the experiences of continuity, flow, and extent of our experiences. Moreover, in isolating temporal grains of experience to a single timescale, IIT misses out on capturing the multi-scale nested nature of temporal phenomena. Hence, we contend that IIT needs an axiom for time, through which its causal structure can be refined to account for temporal experiences. Here, we propose an axiom to address these concerns. We conclude by discussing how IIT may need to be revised if our concerns hold true.

Integrated Information Theory (IIT) is currently one of the most influential scientific theories of consciousness. Here, we focus specifically on a metaphysical aspect of the theory's most recent version (IIT 4.0), what we may call its idealistic ontology, and its tension with a kind of realism about the external world that IIT also endorses. IIT 4.0 openly rejects the mainstream view that consciousness is generated by the brain, positing instead that consciousness is ontologically primary while the physical domain is just "operational". However, this philosophical position is presently underdeveloped and is not rigorously formulated in IIT, potentially leading to many misinterpretations and undermining its overall explanatory power. In the present paper we aim to address this issue. We argue that IIT's idealistic ontology should be understood as a specific combination of phenomenal primitivism, reductionism regarding Φ-structures and complexes, and eliminativism about non-conscious physical entities. Having clarified this, we then focus on the problematic tension between IIT's idealistic ontology and its simultaneous endorsement of realism, according to which there is some kind of external reality independent of our minds. After refuting three potential solutions to this theoretical tension, we propose the most plausible alternative: understanding IIT's realism as an assertion of the existence of other experiences beyond one's own, what we call a non-solipsistic idealist realism. We end with concluding remarks and future research avenues.

Integrated information theory (IIT), developed by Giulio Tononi and colleagues, aims to provide a theory of consciousness with explanatory, predictive, and inferential power, starting from phenomenology itself. IIT contrasts with current approaches that start from the neural correlates of consciousness (NCC) with the hope of identifying generalized principles about the nature of consciousness. Instead, IIT first identifies the essential properties of every experience from consciousness itself. From these “axioms” of phenomenology, IIT then infers a set of requirements (“postulates”) for a physical system to be a substrate of consciousness, while the NCC serve as empirical data for evaluating IIT’s predictions. In particular, IIT predicts that the quality or content of an experience is structurally identical to the cause-effect structure of its physical substrate, and that the quantity or level of consciousness corresponds to the amount of intrinsic cause-effect power the substrate has onto itself. IIT, moreover, offers a theoretical formalism that, in principle, makes it possible to evaluate whether a physical system complies with the IIT postulates and thus forms a physical substrate of consciousness, to quantify the level of consciousness of such a system, and to provide a full account of its phenomenological structure in causal terms.

Theories of consciousness grounded in neuroscience must explain the phenomenal binding problem, e.g., how micro-units of information are combined to create the macro-scale conscious experience common to human phenomenology. An example is how single ‘pixels’ of a visual scene are experienced as a single holistic image in the ‘mind’s eye’, rather than as individual, separate, and massively parallel experiences, corresponding perhaps to individual neuron activations, neural ensembles, or foveal saccades, any of which could conceivably deliver identical functionality from an information processing point of view. There are multiple contested candidate solutions to the phenomenal binding problem. This paper explores how the metaphysical infrastructure of Integrated Information Theory (IIT) v4.0 can provide a distinctive solution. The solution—that particular entities aggregable from multiple units (‘complexes’) define existence—might work in a static picture, but introduces issues in a dynamic system. We ask what happens to our phenomenal self as the main complex moves around a biological neural network. Our account of conscious entities developing through time leads to an apparent dilemma for IIT theorists between non-local entity transitions and contiguous selves: the ‘dynamic entity evolution problem’. As well as specifying the dilemma, we describe three ways IIT might dissolve the dilemma before it gains traction. Clarifying IIT’s position on the phenomenal binding problem, potentially underpinned with novel empirical or theoretical research, helps researchers understand IIT and assess its plausibility. We see our paper as contributing to IIT’s current research emphasis on the shift from static to dynamic analysis.

Integrated Information Theory (IIT) stands out as one of the most promising theories for dealing with the hard problem of consciousness. Founded on five axioms derived from phenomenology, IIT seeks for the physical substrate of consciousness that complies with such axioms according to the criterion of maximally integrated information (Φ). Eventually, IIT identifies phenomenal consciousness with maximal Φ or, what is the same thing, with the strongest cause-effect power in the system. Among the scholars critical of this theory, some point to the so-called Intrinsicality Problem (IP), namely that consciousness cannot be an intrinsic property of the system because maximal Φ crucially depends on the possible existence of bigger values of Φ if the initial system is appropriately linked to or embedded in larger systems. Although proposals in the recent literature aim to solve the IP by going beyond reductionism and physicalism, none of them tackle the real issue, i.e., the insufficiency of IIT’s causal-metaphysical structure. This papers endeavors to provide a solution to the IP in IIT within a hylomorphist ontology that includes formal causation. Complementing IIT with formal causation provides the theory with a criterion of individuation that solves the IP and, by relaxing identification between maximal Φ and consciousness, it lends a more robust metaphysical structure. To wit, maximal Φ is a necessary but not sufficient condition for the existence of consciousness.

In this paper, we examine the causal framework within which integrated information theory (IIT) of consciousness makes it claims. We argue that, in its current formulation, IIT is threatened by the causal exclusion problem. Some proponents of IIT have attempted to thwart the causal exclusion problem by arguing that IIT has the resources to demonstrate genuine causal emergence at macro scales. In contrast, we argue that their proposed solution to the problem is damagingly circular as a result of inter-defining information and causation. As a solution, we propose that IIT should adopt the specific interventionist causal framework that we offer and show how IIT can harness this interventionist framework to avoid the causal exclusion problem. We demonstrate how our argument remains fully compatible with the methodology, empirical data, and conceptual aims of the theory.

Giulio Tononi's integrated information theory (IIT) proposes explaining consciousness by directly identifying it with integrated information. We examine the construct validity of IIT's measure of consciousness, phi (Φ), by analyzing its formal properties, its relation to key aspects of consciousness, and its co-variation with relevant empirical circumstances. Our analysis shows that IIT's identification of consciousness with the causal efficacy with which differentiated networks accomplish global information transfer (which is what Φ in fact measures) is mistaken. This misidentification has the consequence of requiring the attribution of consciousness to a range of natural systems and artifacts that include, but are not limited to, large-scale electrical power grids, gene-regulation networks, some electronic circuit boards, and social networks. Instead of treating this consequence of the theory as a disconfirmation, IIT embraces it. By regarding these systems as bearers of consciousness ex hypothesi, IIT is led toward the orbit of panpsychist ideation. This departure from science as we know it can be avoided by recognizing the functional misattribution at the heart of IIT's identity claim. We show, for example, what function is actually performed, at least in the human case, by the cortical combination of differentiation with integration that IIT identifies with consciousness. Finally, we examine what lessons may be drawn from IIT's failure to provide a credible account of consciousness for progress in the very active field of research concerned with exploring the phenomenon from formal and neural points of view.