The traditional divide between inert matter and conscious mind is dissolving across disciplines. In fields as diverse as soft robotics, neurophenomenology, and bioengineering, researchers increasingly observe cognitive-like behaviors in non-neural systems. The discovery of sensing materials and intelligent matter demonstrates that the building blocks of life are not merely reactive but capable of functional evaluation, memory, and adaptive reorganization (Kaspar et al., 2021; Wang et al., 2020; Dixon et al., 2021; Hu et al., 2019).

This paper situates these findings within a continuum of cognition, arguing that the properties and behaviors of simple physical systems precede the basal cognition observed in living cells (Lyon et al., 2021; Jin et al., 2024) and, ultimately, the complex subconscious and conscious operations within the human body. In doing so, we offer a naturalistic, embodied, biologically inspired framework that may be used to inform the design of adaptive, embodied systems relevant to Artificial Intelligence (AI).

From its inception, AI was envisioned as an artifactual counterpart to human intelligence with the prospect of eventually surpassing it. The development of AI and research in human cognition have progressed concurrently and interactively. As Kari and Rozenberg (2008) explain, the learning process is two-way: while we design new technological systems, these systems, in turn, provide tools to test and refine our understanding of human mind and intelligence.

Materials like shape-memory alloys, piezoelectric polymers, and neuromorphic gels are capable of storing and processing information along with environmental interactions (Chortos and Bao, 2014; Prakash et al., 2023). These materials exhibit behaviors that, while non-conscious, are valence-sensitive as they respond differently to different stimuli in ways that resemble biological preference or avoidance.

Recent developments in neuromorphic engineering suggest that intelligent matter (Kaspar et al., 2021) can exhibit learning-like and adaptive responses, suggesting the idea that material substrates can support basic cognitive architectures. These materials provide a physical and conceptual basis for proto-cognition, i.e., early forms of information integration and action selection that precede life and scaffold its emergence.

Antecedent proto-cognitive processes in non-living matter do not imply cognition to such systems. Proto-cognition denotes physical and chemical precursor dynamics, such as adaptive material responses and constraint-driven reconfiguration, that enables the self-organization of living organisms but lack self-maintenance and intrinsic normativity. Cognition proper, as defined here, is restricted to living systems.

New research on cellular level shows memory as general property of cells. Cellular memory in non-neural cells involves corresponding molecular machinery previously found in neurons. Work of Kukushkin et al. (2024) experimentally demonstrates memory-like gene activation effects in non-neural cells, showing conserved molecular mechanisms underlying cellular learning. At the epigenetic level, cellular memory is supported by the propagation of DNA methylation and histone modifications through cell division, maintaining transcriptional programs essential for cell identity and differentiation (Espinosa-Martínez et al., 2024). This epigenetic cycle allows cells to “remember” environmental signals long after they cease, enabling stable phenotypic states and plastic responses. Further, microtubules and microtubule-associated proteins in cells contribute to intracellular signaling and structural plasticity relevant to memory consolidation and information storage (Flores and Liester, 2024). Disruption of these components can impair memory-related functions.

These findings characterize cellular memory as a multilayered phenomenon involving gene regulation, signaling networks, and epigenetic mechanisms, extending the concept of memory beyond the nervous system to all cell types. This has implications for understanding development, disease, and systemic embodied cognition.

Individual somatic cells in multicellular organisms demonstrate capacities for decision-making and learning at the biochemical and biophysical level. Immune cells, for example, evaluate threat, adapt responses based on past encounters, and communicate dynamically with other systems (Prinz and Priller, 2017). Stem cells integrate environmental cues to determine differentiation pathways, a process requiring multi-factorial assessment and response (Whited and Levin, 2019).

Michael Levin’s work on bioelectric signaling and morphogenetic fields reveals that tissues and organs participate in distributed information processing (computation), enabling body-wide decision-making during regeneration and development. These examples of somatic cognition are operating outside conscious awareness but are essential to adaptive behavior.

As multicellular coordination scales up, it gives rise to body-level intelligence that shapes affective states, intuitive actions, and physiological regulation. Neuroscientific models of interoception (Barrett and Simmons, 2015; Damasio, 1999) and enactivism (Varela et al., 1991) emphasize that conscious experience arises from an ongoing dialogue between the brain and the body.

We reinterpret the subconscious not as a repressed psychological domain, but as a valence-sensitive network of dynamic, embodied processes. It includes habits, affective responses, visceral memories, and autonomic regulation, sub-personal systems that operate beneath awareness but scaffold consciousness. These processes can be understood as the integration of somatic and tissue-level cognition into whole-body intelligence.

Bridging from sensing materials properties to cellular and whole-organism cognition enables a continuum view of mind, where intelligence emerges through material integration and organizational complexity. This perspective supports a non-dualistic account of mind, grounded in bio-physical systems and scaling up through recursive embodiment and feedback.

Importantly, this approach understands cognition as a property of organized responsiveness, not necessarily symbolic reasoning. It models the mind as a material phenomenon rooted in living matter, capable of being emulated in adaptive, embodied artificial systems.

Current AGI paradigms tend to replicate human-level symbolic reasoning, often abstracted from embodiment. A continuum theory of biological minds suggests that AGI should instead emulate the scalable, material, and embodied properties of natural cognition. From neuromorphic substrates to decentralized learning, lessons from biology may help construct systems with genuine adaptive intelligence.

Table 1 presents a spectrum of cognition, intelligence, awareness and mind forms across living systems, from bacteria to great apes and humans. It shows how cognitive capacities and behaviors increase in complexity with evolutionary development increasing organisms capacity to control their environment.

By focusing on cognition as an emergent property of valence-sensitive organization—from molecules to organs to conscious experience, we find an alternative pathway for AGI development: one rooted in life’s own layered architecture of sense-making. Learning from nature, we focus on the self-organization and development of biological minds with increasing complexity of cognition, intelligence, sentience, awareness and consciousness.

Terms such as sentience, awareness, and experience are used in a graded and biologically grounded sense. At basal levels of life, these terms refer to functional, valence-sensitive regulation rather than to phenomenally conscious experience. Phenomenal consciousness, reflective awareness, and reportable subjective experience arise only under specific organizational conditions, typically involving nervous systems and integrative architectures. The use of experiential vocabulary at lower biological scales is not intended to attribute human-like subjective experience to single cells but to describe degrees of organized responsiveness relevant to the organism’s level of organization.

In what follows, the definitions of basic terms will be given and explained.

Definition: Information is a difference that makes a difference for the regulation of system–environment interactions.

Informational states are defined by their functional consequences for maintaining or modifying organized behavior, rather than by abstract symbols or semantic content. This allows information to be instantiated in biochemical, electrical, or mechanical configurations without presupposing representation or interpretation.

Definition: Computation is the physical transformation of information through the intrinsic dynamics of a system (information dynamics).

On this view, computation need not involve symbolic manipulation or centralized control. It is realized by the material and dynamical properties of physical systems themselves (natural computation). Unlike conventional digital computation, which operates over abstract symbols in engineered architectures, morphological computation in living systems is realized through continuous physical dynamics of embodied structures. This interpretation is supported by work which demonstrates how embodied dynamics can perform computational functions without representation or explicit control architectures (e.g., Dodig-Crnkovic, 2012).

Definition: Cognition is the process by which an organism acquires, transforms, stores, and uses information to regulate its behavior and interactions with the environment.

Cognition exists in all life forms, from bacteria to humans. It includes information processing, sensory input, and response coordination (Dodig-Crnkovic, 2025). It does not require a brain. Bacteria, fungi, and plants exhibit cognition through chemical and electrical signaling. For example, bacteria use quorum sensing to make group decisions (Ng and Bassler, 2009; Waters and Bassler, 2005). Fungi transmit electrical signals across their mycelial networks (Adamatzky, 2022; Adamatzky et al., 2023; Baluška and Levin, 2016). Recent research on plant electrical and chemical signaling provides evidence for biological bases for plant sentience (Segundo-Ortin and Calvo, 2021). Animal cognition is studied in biosemiotics, ethology, comparative psychology, behavioral ecology, etc. and is well documented in mammals, birds, reptiles, fish and invertebrates, including insects (Greggers and Menzel, 1993).

Definition: Sentience is the capacity of an organism to have valenced responses—meaningful experiences of preference for beneficial conditions over harmful ones, that in the first step is valenced response that distinguishes “good” from “bad.” Sentience reflects a preference toward beneficial states. The organism does not just react but internalize inputs of individual/subjective experience. Sentience ranges from basic (bacteria avoiding toxins and move toward nutrients) (Lyon, 2015); insects exhibit pain-like responses to injury, (Gibbons et al., 2022) to complex (Ginsburg and Jablonka, 2019). Sentience does not require language or self-awareness.

Experiences are not neutral; they are perceived as “good” or “bad,” eliciting valenced responses. Sensory-based awareness implies that the organism processes sensory information in a way that affects behavior beyond pure reflexes. All living organisms possess a degree of sentience. Bacteria process information from the environment (cognition) and valuate it in terms of good/bad or attractive/repulsive. Table 2 presents types of sentience across life. All animals are sentient, but only some have self-awareness. Sentience is complex in organisms with nervous system, but simpler forms of sentience exist already in non-neural organisms.

Definition: Intelligence is the ability of an organism to learn, solve problems, and adapt behavior based on experience or environmental changes.

It is expressed on the individual level (octopus learning a task and using tools) (Godfrey-Smith, 2017) and collective level (bacteria in biofilms adapting to antibiotics) (Ben-Jacob, 2008; Ben-Jacob, 2009). Fungi adjust their growth patterns based on past and present nutrient availability (Fukasawa et al., 2024; Adamatzky et al., 2023). It does not require a brain—fungi and plants exhibit intelligence through adaptive behavior (Levin, 2023; Ginsburg and Jablonka, 2009). Bees learn and remember complex foraging routes (Greggers and Menzel, 1993).

Definition: Awareness → consciousness continuum is the ability of an organism to integrate sensory information, maintain a continuous state of responsiveness, and interact with the environment in a structured way.

This continuum ranges from basic environmental awareness to self-awareness and consciousness. It does not require thought or introspection—even bacteria and fungi are “aware” of their surroundings (Adamatzky et al., 2023). Simple organisms have sensory awareness, while complex organisms develop complex self-awareness and consciousness.

Bacteria detect chemical gradients and adjust behavior. Fungi sense nearby plants and redirect growth (Money, 2021). Dogs experience emotions and respond to social cues. Humans engage in self-reflection. In short, consciousness is not binary but a continuum, with different levels of complexity in different organisms (Damasio, 2010; Koch, 2012).

This graded perspective is related to what Ginsburg and Jablonka (2019) have described as the emergence of a “sensitive soul.” Within this view, the boundary between awareness and consciousness remains fluid and context-dependent. Some approaches place this boundary at the emergence of nervous systems, as they enable more integrated, flexible, and potentially self-referential forms of experience. In the present framework, awareness and consciousness are treated as points along a continuum of increasing organizational complexity.

Just like cognition, sentience, intelligence, and awareness → consciousness, the term “mind” is currently ill-defined and anthropocentric. If we want definitions that apply to all living organisms, and generalize to machines, we need generalized but precise explanations that allow for different levels of complexity across species.

Definition: Mind is the activity of an organism that processes information, integrates sensory input, regulates internal states, and generates adaptive responses.

The mind is the totality of cognitive, sentient, intelligent, and conscious functions working together.

This definition points out the dynamical and multi-tasking aspects of mind, which includes information processing (like cognition), regulation of behavior and adaptation (like intelligence) and integration of internal and external signals (like awareness).

The key features of a mind are information processing, signal integration, behavior regulation and adaptation (Maturana and Varela, 1980). Mind is not present only in a physical brain—it includes all information-processing mechanisms. It applies to both centralized (brains) and distributed/decentralized (fungal networks) systems. It exists in all living systems as a means they regulate themselves.

For example, fungal networks process information about nutrients and threats (Adamatzky et al., 2023; Money, 2021). Insect colonies function as “collective minds” that solve problems (Almér et al., 2015). Human brains engage in complex reasoning and creativity.

All biological organisms consist of cells, so “cellular minds” are the building blocks for all living minds, including human. Living organisms at different levels of complexity possess different forms of cognition and intelligence adapted to their environment. Awareness exists in many forms, from simple environmental sensing to deep introspection.

Brains are not required for cognition. Cognition is distributed information processing. Brains as well process distributed information. Consciousness appears on a continuum, with different levels of experience in different organisms.

Looking at processes of cognition, sentience, intelligence, awareness → consciousness and mind as a process integrating them all (Table 4) we see how those properties constitute a unified view of mind in living systems.

Instead of asking “Which organisms are conscious?” or “Which organisms are intelligent?,” the “diverse minds” approach suggests a more nuanced view. Different organisms exhibit different forms of cognition, adapted to their ecological needs. Intelligence emerges in multiple ways, through centralized brains or decentralized networks. Consciousness is a spectrum, without a strict boundary, with varying degrees of “subjective” (individual) experience. This understanding has a wide range of consequences.

Even simple life forms have a “mind” if we define it as the system integrating level of cognition, intelligence, and awareness. Evolution leads to higher minds developed with greater intelligence and more complex consciousness (Jablonka et al., 2014).

Instead of being a physical object, the mind is an emergent process—it arises when cognition, sentience, intelligence, and consciousness interact in a dynamic, self-regulating way. This means that minds can exist in decentralized systems such as bacteria colonies and fungi (Lyon, 2015; Adamatzky, 2022). Minds can exist without self-awareness as in plants and insects. The complexity of a mind depends on the depth of cognition, intelligence, and consciousness.

As biological minds exist on a spectrum, we can categorize organisms by the depth of their mind-related capacities. Instead of asking “Which organisms have minds?,” a better question is: “What kind of mind does an organism have?” The mind is not an object but a process— an active, embodied, predictive interface between organism and world (Clark, 2016). It exists in different degrees across all life forms.

Empirically motivated proposals for unified frameworks of cognition across living systems have emerged from observations of cognitive-like behavior spanning biological scales (Dennett, 2017; Levin, 2022, 2023). However, these proposals have not articulated a systematic explanatory account of the generative mechanisms through which diverse cognitive capacities arise and relate to one another within a single coherent framework.

A related attempt at unifying mind and matter is Nakajima’s cognizers system (CS) model, which extends the notion of cognition hierarchically across physical, chemical, and semiotic levels by defining cognition as a relational state change (“related state change”) (Nakajima, 2024). Within that framework, semiotic cognition is emphasized as the level most closely aligned with conventional cognitive and mental phenomena, while physical and chemical “cognitions” describe context-dependent state transitions in nonliving systems. The present work (ICON) shares the aim of theoretical unification, but differs in locating cognition primarily at the level of living organization and treating behavior, learning, and minimal experience as present from the onset of life and progressively complexified with increasing biological organization. ICON locates cognition at the level of biological organization, grounding meaning and normativity directly in the viability conditions of living systems. Physical and chemical substrates support cognitive functions such as memory, recall, information processing, and communication. Table 6 presents the comparison between ICON and CS models.

Identifying cognition with living organization establishes a two-fold empirical connection: an explanatory role, in which diverse biological phenomena are unified under a common organizational account, and a predictive role, in which the framework constrains expectations about how adaptive behavior should change under experimental conditions.

Explanatorily, the framework accounts for empirical findings that are difficult to reconcile with current views restricting cognition to neural or representational architectures. In particular, work on bioelectric regulation of morphogenesis demonstrates that multicellular tissues can store, recall, and correct target anatomical patterns through distributed physiological networks, independently of nervous systems (Levin, 2014; Whited and Levin, 2019). Such phenomena exhibit error correction, goal-directed regulation, and context sensitivity, and are therefore more naturally understood as manifestations of multiscale organizational cognition than as the by-products of local molecular dynamics alone.

Predictively, the framework yields empirically testable expectations. If cognition depends on self-maintaining, multiscale organization rather than on localized information processing alone, then systematic disruption of organizational coupling, such as perturbations of gene regulatory networks, bioelectric signaling, or intercellular connectivity, should impair adaptive regulation even when local signaling mechanisms remain intact. This logic is consistent with experimental results showing loss of morphogenetic competence and stable anatomical outcomes following targeted disruption of bioelectric patterning and physiological coordination (Levin, 2014; Whited and Levin, 2019), and with predictive-processing interpretations of organisms as problem-solving systems rather than passive stimulus–response machines (Pezzulo and Levin, 2016).

Established paradigms in developmental and regenerative biology provide concrete test cases for these predictions. The remarkable regenerative abilities of planarian flatworms, such as whole-body regeneration from small tissue fragments, cannot be explained solely by local gene expression or molecular gradients. Experimental work shows that these competences depend on long-range bioelectric signaling and forms of anatomical memory that enable error-correcting pattern regulation (Levin, 2014; Whited and Levin, 2019) which exact mechanisms remain underexplored. Within the proposed framework, such phenomena are predicted to arise from cognitive capacities of cell collectives, whereby cells store, evaluate, and update information about target morphology and adjust their behavior accordingly, which reframes theoretical understanding in a cognitive–organizational terms (Levin, 2022).

In this way, the proposed framework links explanatory unification with experimentally grounded predictions, situating cognition as a property of living organization that is already under active empirical investigation.