The free energy principle induces intracellular compartmentalization.

Living systems at all scales are compartmentalized into interacting subsystems. This paper reviews a mechanism that drives compartmentalization in generic systems at any scale. It first discusses three symmetries of generic physical interactions in a quantum-theoretic description. It then shows that if one of these, a permutation symmetry on the inter-system boundary, is spontaneously broken, the symmetry breaking is amplified by the Free Energy Principle (FEP). It thus shows how compartmentalization generically results from permutation symmetry breaking under the FEP. It finally notes that the FEP asymptotically restores the broken symmetry, showing that the FEP can be regarded as a theory of fluctuations away from a permutation-symmetric boundary, and hence from an entangled joint state of the interacting systems.

Meta-learning goes hand-in-hand with metacognition.

Binz et al. propose a general framework for meta-learning and contrast it with built-by-hand Bayesian models. We comment on some architectural assumptions of the approach, its relation to the active inference framework, its potential applicability to living systems in general, and the advantages of the latter in addressing the explanation problem.

Principled Limitations on Self-Representation for Generic Physical Systems.

The ideas of self-observation and self-representation, and the concomitant idea of self-control, pervade both the cognitive and life sciences, arising in domains as diverse as immunology and robotics. Here, we ask in a very general way whether, and to what extent, these ideas make sense. Using a generic model of physical interactions, we prove a theorem and several corollaries that severely restrict applicable notions of self-observation, self-representation, and self-control. We show, in particular, that adding observational, representational, or control capabilities to a meta-level component of a system cannot, even in principle, lead to a complete meta-level representation of the system as a whole. We conclude that self-representation can at best be heuristic, and that self models cannot, in general, be empirically tested by the systems that implement them.

Making the Thermodynamic Cost of Active Inference Explicit.

When describing Active Inference Agents (AIAs), the term "energy" can have two distinct meanings. One is the energy that is utilized by the AIA (e.g., electrical energy or chemical energy). The second meaning is so-called Variational Free Energy (VFE), a statistical quantity which provides an upper bound on surprisal. In this paper, we develop an account of the former quantity-the Thermodynamic Free Energy (TFE)-and its relationship with the latter. We highlight the necessary tradeoffs between these two in a generic, quantum information-theoretic formulation, and the macroscopic consequences of those tradeoffs for the ways that organisms approach their environments. By making this tradeoff explicit, we provide a theoretical basis for the different metabolic strategies that organisms from plants to predators use to survive.

A Variational Synthesis of Evolutionary and Developmental Dynamics.

This paper introduces a variational formulation of natural selection, paying special attention to the nature of 'things' and the way that different 'kinds' of 'things' are individuated from-and influence-each other. We use the Bayesian mechanics of particular partitions to understand how slow phylogenetic processes constrain-and are constrained by-fast, phenotypic processes. The main result is a formulation of adaptive fitness as a path integral of phenotypic fitness. Paths of least action, at the phenotypic and phylogenetic scales, can then be read as inference and learning processes, respectively. In this view, a phenotype actively infers the state of its econiche under a generative model, whose parameters are learned via natural (Bayesian model) selection. The ensuing variational synthesis features some unexpected aspects. Perhaps the most notable is that it is not possible to describe or model a population of conspecifics per se. Rather, it is necessary to consider populations of distinct natural kinds that influence each other. This paper is limited to a description of the mathematical apparatus and accompanying ideas. Subsequent work will use these methods for simulations and numerical analyses-and identify points of contact with related mathematical formulations of evolution.

Scale-Free Biology: Integrating Evolutionary and Developmental Thinking.

When the history of life on earth is viewed as a history of cell division, all of life becomes a single cell lineage. The growth and differentiation of this lineage in reciprocal interaction with its environment can be viewed as a developmental process; hence the evolution of life on earth can also be seen as the development of life on earth. Here, in reviewing this field, some potentially fruitful research directions suggested by this change in perspective are highlighted. Variation and selection become, for example, bidirectional information flows between scales, while the notions of "cooperation" and "competition" become scale relative. The language of communication, inference, and information processing becomes more useful than the language of causation to describe the interactions of both homogeneous and heterogeneous living systems at any scale. Emerging scale-free theoretical frameworks such as predictive coding and active inference provide conceptual tools for reconceptualizing biology as the study of a unified, multiscale dynamical system.

Competency in Navigating Arbitrary Spaces as an Invariant for Analyzing Cognition in Diverse Embodiments.

One of the most salient features of life is its capacity to handle novelty and namely to thrive and adapt to new circumstances and changes in both the environment and internal components. An understanding of this capacity is central to several fields: the evolution of form and function, the design of effective strategies for biomedicine, and the creation of novel life forms via chimeric and bioengineering technologies. Here, we review instructive examples of living organisms solving diverse problems and propose competent navigation in arbitrary spaces as an invariant for thinking about the scaling of cognition during evolution. We argue that our innate capacity to recognize agency and intelligence in unfamiliar guises lags far behind our ability to detect it in familiar behavioral contexts. The multi-scale competency of life is essential to adaptive function, potentiating evolution and providing strategies for top-down control (not micromanagement) to address complex disease and injury. We propose an observer-focused viewpoint that is agnostic about scale and implementation, illustrating how evolution pivoted similar strategies to explore and exploit metabolic, transcriptional, morphological, and finally 3D motion spaces. By generalizing the concept of behavior, we gain novel perspectives on evolution, strategies for system-level biomedical interventions, and the construction of bioengineered intelligences. This framework is a first step toward relating to intelligence in highly unfamiliar embodiments, which will be essential for progress in artificial intelligence and regenerative medicine and for thriving in a world increasingly populated by synthetic, bio-robotic, and hybrid beings.

Metacognition as a Consequence of Competing Evolutionary Time Scales.

Evolution is full of coevolving systems characterized by complex spatio-temporal interactions that lead to intertwined processes of adaptation. Yet, how adaptation across multiple levels of temporal scales and biological complexity is achieved remains unclear. Here, we formalize how evolutionary multi-scale processing underlying adaptation constitutes a form of metacognition flowing from definitions of metaprocessing in machine learning. We show (1) how the evolution of metacognitive systems can be expected when fitness landscapes vary on multiple time scales, and (2) how multiple time scales emerge during coevolutionary processes of sufficiently complex interactions. After defining a metaprocessor as a regulator with local memory, we prove that metacognition is more energetically efficient than purely object-level cognition when selection operates at multiple timescales in evolution. Furthermore, we show that existing modeling approaches to coadaptation and coevolution-here active inference networks, predator-prey interactions, coupled genetic algorithms, and generative adversarial networks-lead to multiple emergent timescales underlying forms of metacognition. Lastly, we show how coarse-grained structures emerge naturally in any resource-limited system, providing sufficient evidence for metacognitive systems to be a prevalent and vital component of (co-)evolution. Therefore, multi-scale processing is a necessary requirement for many evolutionary scenarios, leading to de facto metacognitive evolutionary outcomes.

Morphological Coordination: A Common Ancestral Function Unifying Neural and Non-Neural Signaling.

Nervous systems are traditionally thought of as providing sensing and behavioral coordination functions at the level of the whole organism. What is the evolutionary origin of the mechanisms enabling the nervous systems' information processing ability? Here, we review evidence from evolutionary, developmental, and regenerative biology suggesting a deeper, ancestral function of both pre-neural and neural cell-cell communication systems: the long-distance coordination of cell division and differentiation required to create and maintain body-axis symmetries. This conceptualization of the function of nervous system activity sheds new light on the evolutionary transition from the morphologically rudimentary, non-neural Porifera and Placazoa to the complex morphologies of Ctenophores, Cnidarians, and Bilaterians. It further allows a sharp formulation of the distinction between long-distance axis-symmetry coordination based on external coordinates, e.g., by whole-organism scale trophisms as employed by plants and sessile animals, and coordination based on body-centered coordinates as employed by motile animals. Thus we suggest that the systems that control animal behavior evolved from ancient mechanisms adapting preexisting ionic and neurotransmitter mechanisms to regulate individual cell behaviors during morphogenesis. An appreciation of the ancient, non-neural origins of bioelectrically mediated computation suggests new approaches to the study of embryological development, including embryological dysregulation, cancer, regenerative medicine, and synthetic bioengineering.

Fitness Beats Truth in the Evolution of Perception.

Does natural selection favor veridical percepts-those that accurately (if not exhaustively) depict objective reality? Perceptual and cognitive scientists standardly claim that it does. Here we formalize this claim using the tools of evolutionary game theory and Bayesian decision theory. We state and prove the "Fitness-Beats-Truth (FBT) Theorem" which shows that the claim is false: If one starts with the assumption that perception involves inference to states of the objective world, then the FBT Theorem shows that a strategy that simply seeks to maximize expected-fitness payoff, with no attempt to estimate the "true" world state, does consistently better. More precisely, the FBT Theorem provides a quantitative measure of the extent to which the fitness-only strategy dominates the truth strategy, and of how this dominance increases with the size of the perceptual space. The FBT Theorem supports the Interface Theory of Perception (e.g. Hoffman et al. in Psychon Bull Rev https://doi.org/10.3758/s13423-015-0890-8 , 2015), which proposes that our perceptual systems have evolved to provide a species-specific interface to guide adaptive behavior, and not to provide a veridical representation of objective reality.

Scale-free architectures support representational diversity.

Gilead et al. propose an ontology of abstract representations based on folk-psychological conceptions of cognitive architecture. There is, however, no evidence that the experience of cognition reveals the architecture of cognition. Scale-free architectural models propose that cognition has the same computational architecture from sub-cellular to whole-organism scales. This scale-free architecture supports representations with diverse functions and levels of abstraction.

Do Process-1 simulations generate the epistemic feelings that drive Process-2 decision making?

We apply previously developed Chu space and Channel Theory methods, focusing on the construction of Cone-Cocone Diagrams (CCCDs), to study the role of epistemic feelings, particularly feelings of confidence, in dual process models of problem solving. We specifically consider "Bayesian brain" models of probabilistic inference within a global neuronal workspace architecture. We develop a formal representation of Process-1 problem solving in which a solution is reached if and only if a CCCD is completed. We show that in this representation, Process-2 problem solving can be represented as multiply iterated Process-1 problem solving and has the same formal solution conditions. We then model the generation of explicit, reportable subjective probabilities from implicit, experienced confidence as a simulation-based, reverse engineering process and show that this process can also be modeled as a CCCD construction.

Fact, Fiction, and Fitness.

A theory of consciousness, whatever else it may do, must address the structure of experience. Our perceptual experiences are richly structured. Simply seeing a red apple, swaying between green leaves on a stout tree, involves symmetries, geometries, orders, topologies, and algebras of events. Are these structures also present in the world, fully independent of their observation? Perceptual theorists of many persuasions-from computational to radical embodied-say yes: perception veridically presents to observers structures that exist in an observer-independent world; and it does so because natural selection shapes perceptual systems to be increasingly veridical. Here we study four structures: total orders, permutation groups, cyclic groups, and measurable spaces. We ask whether the payoff functions that drive evolution by natural selection are homomorphisms of these structures. We prove, in each case, that generically the answer is no: as the number of world states and payoff values go to infinity, the probability that a payoff function is a homomorphism goes to zero. We conclude that natural selection almost surely shapes perceptions of these structures to be non-veridical. This is consistent with the interface theory of perception, which claims that natural selection shapes perceptual systems not to provide veridical perceptions, but to serve as species-specific interfaces that guide adaptive behavior. Our results present a constraint for any theory of consciousness which assumes that structure in perceptual experience is shaped by natural selection.

The Role of Early Bioelectric Signals in the Regeneration of Planarian Anterior/Posterior Polarity.

Axial patterning during planarian regeneration relies on a transcriptional circuit that confers distinct positional information on the two ends of an amputated fragment. The earliest known elements of this system begin demarcating differences between anterior and posterior wounds by 6 h postamputation. However, it is still unknown what upstream events break the axial symmetry, allowing a mutual repressor system to establish invariant, distinct biochemical states at the anterior and posterior ends. Here, we show that bioelectric signaling at 3 h is crucial for the formation of proper anterior-posterior polarity in planaria. Briefly manipulating the endogenous bioelectric state by depolarizing the injured tissue during the first 3 h of regeneration alters gene expression by 6 h postamputation and leads to a double-headed phenotype upon regeneration despite confirmed washout of ionophores from tissue. These data reveal a primary functional role for resting membrane potential taking place within the first 3 h after injury and kick-starting the downstream pattern of events that elaborate anatomy over the following 10 days. We propose a simple model of molecular-genetic mechanisms to explain how physiological events taking place immediately after injury regulate the spatial distribution of downstream gene expression and anatomy of regenerating planaria.

Some Consequences of the Thermodynamic Cost of System Identification.

The concept of a "system" is foundational to physics, but the question of how observers identify systems is seldom addressed. Classical thermodynamics restricts observers to finite, finite-resolution observations with which to identify the systems on which "pointer state" measurements are to be made. It is shown that system identification is at best approximate, even in a finite world, and that violations of the Leggett-Garg and Bell/CHSH (Clauser-Horne-Shimony-Holt) inequalities emerge naturally as requirements for successful system identification.

Visual re-identification of individual objects: a core problem for organisms and AI.

Two open questions about the visual re-identification of individual objects over extended time periods are briefly reviewed: (1) How much a priori information about the nature of objects, identity and time is required to support robust individual object re-identification abilities? and (2) how do epistemic feelings, such as the feeling of familiarity, contribute both to object re-identification and to the perception of opportunities and risks associated with individual objects and their affordances? The ongoing interplay between experiments that can be carried out with human subjects and experiments made possible with robotic systems is examined. It is suggested that developmental robotics, including virtual-reality simulations of robot-environment interactions, may provide the best route to understanding both the implementation of epistemic feelings in humans and their functional contribution to the identification of persistent individual objects.

Federated inference and belief sharing.

This paper concerns the distributed intelligence or federated inference that emerges under belief-sharing among agents who share a common world-and world model. Imagine, for example, several animals keeping a lookout for predators. Their collective surveillance rests upon being able to communicate their beliefs-about what they see-among themselves. But, how is this possible? Here, we show how all the necessary components arise from minimising free energy. We use numerical studies to simulate the generation, acquisition and emergence of language in synthetic agents. Specifically, we consider inference, learning and selection as minimising the variational free energy of posterior (i.e., Bayesian) beliefs about the states, parameters and structure of generative models, respectively. The common theme-that attends these optimisation processes-is the selection of actions that minimise expected free energy, leading to active inference, learning and model selection (a.k.a., structure learning). We first illustrate the role of communication in resolving uncertainty about the latent states of a partially observed world, on which agents have complementary perspectives. We then consider the acquisition of the requisite language-entailed by a likelihood mapping from an agent's beliefs to their overt expression (e.g., speech)-showing that language can be transmitted across generations by active learning. Finally, we show that language is an emergent property of free energy minimisation, when agents operate within the same econiche. We conclude with a discussion of various perspectives on these phenomena; ranging from cultural niche construction, through federated learning, to the emergence of complexity in ensembles of self-organising systems.

Metaphorical motion in mathematical reasoning: further evidence for pre-motor implementation of structure mapping in abstract domains.

The theory of computation and category theory both employ arrow-based notations that suggest that the basic metaphor "state changes are like motions" plays a fundamental role in all mathematical reasoning involving formal manipulations. If this is correct, structure-mapping inferences implemented by the pre-motor action planning system can be expected to be involved in solving any mathematics problems not solvable by table lookups and number line manipulations alone. Available functional imaging studies of multi-digit arithmetic, algebra, geometry and calculus problem solving are consistent with this expectation.

Regulative development as a model for origin of life and artificial life studies.

Using the formal framework of the Free Energy Principle, we show how generic thermodynamic requirements on bidirectional information exchange between a system and its environment can generate complexity. This leads to the emergence of hierarchical computational architectures in systems that operate sufficiently far from thermal equilibrium. In this setting, the environment of any system increases its ability to predict system behavior by "engineering" the system towards increased morphological complexity and hence larger-scale, more macroscopic behaviors. When seen in this light, regulative development becomes an environmentally-driven process in which "parts" are assembled to produce a system with predictable behavior. We suggest on this basis that life is thermodynamically favorable and that, when designing artificial living systems, human engineers are acting like a generic "environment".

Motion as manipulation: implementation of force-motion analogies by event-file binding and action planning.

Tool-improvisation analogies are structure-mapping inferences implemented, in many species, by event-file binding and pre-motor action planning. These processes act on multi-modal representations of currently perceived situations and eventuate in motor acts that can be directly evaluated for success or failure; they employ implicit representations of force-motion relations encoded by the pre-motor system and do not depend on explicit, language-like representations of relational concepts. A detailed reconstruction of the analogical reasoning steps involved in Rutherford's and Bohr's development of the first quantized-orbit model of atomic structure is used to show that human force-motion analogies can in general be implemented by these mechanisms. This event-file manipulation model of the implementation of force-motion analogies is distinguished from the standard view that structure-mapping analogies require the manipulation of explicit, language-like representations of relational concepts.