Bioelectricity, Evolution, Intelligence, and the Platonic Space

Bioelectricity: A Bridge Between Physics and Cognition, by Way of Biology

Bioelectricity bridges the gap between abstract mental goals and physical chemistry, demonstrating that cognition is not confined to brains but extends through all biological scales, and that bodies are multi-scale transduction systems in which high-level goals filter down to make molecular biology dance to their tune.

• Voluntary motion is a daily demonstration that abstract mental goals—social, financial, research—can causally direct the movement of ions across muscle cell membranes, making the body a multi-scale transduction apparatus and proving mind-matter interaction is not mystical but architectural.
  • The chemistry of the body, including calcium and potassium crossing muscle membranes, must respond to extremely abstract high-level cognitive goals every time a person chooses to act, making voluntary motion an everyday miracle of goal-to-chemistry transduction.
  • The interface between cognition and physics is never a sharp transition—it is cognition all the way down, with the trick being to understand communication, scaling of goals, and remapping of memories across substrates and problem spaces.
• Single cells like the unicellular organism Lacrymaria demonstrate remarkable environmental competency without any brain or nervous system, showing that we are made of agential materials with built-in competencies and agendas—materials that cannot be micromanaged but must be persuaded.
  • Lacrymaria, a single-cell organism with no nervous system, exhibits sophisticated problem-solving in its local environment, illustrating that agential competency is a property of living matter at its most basic level.
  • Engineering with agential materials is fundamentally different from engineering with wood or metal because the material has its own agenda and cannot be directed by micromanagement.
• Biology navigates many problem spaces beyond three-dimensional physical space—including gene expression space, physiological state space, and anatomical morphospace—and this navigation constitutes intelligence that humans are poorly equipped to recognize because of evolutionary bias toward medium-scale 3D movement.
  • The journey from a single fertilized egg to a human body is a navigation of the vast space of anatomical possibilities, with development constituting a perception-action loop in morphospace analogous to animal behavior in physical space.
  • The multi-scale competency architecture means every level of biological organization—from molecular networks to organs to organisms to swarms—has its own agendas and problem-solving competencies in its respective space.
• Axolotl limb regeneration and tail-to-limb transformation experiments reveal that biological systems maintain anatomical set points and that global body-plan errors filter down to make local cells—even undamaged ones—undergo molecular reprogramming to correct large-scale deviations from the target morphology.
  • When a tail is grafted to the flank of an amphibian, the tail transforms into a limb over time; the cells at the tip of the tail turn into fingers despite having no local damage, because the collective body-level intelligence detects the global error and propagates correction signals downward.
  • Regeneration stops precisely when the correct species-specific structure has been completed, demonstrating a genuine set point and error-minimization process rather than open-loop emergence.
• It’s cognition all the way down, from a high-level mind embodied in human bodies and whatever is beyond that, all the way down to molecules. It’s all mental.
• Bioelectric manipulation of ion channels in frog embryos can instruct groups of gut-destined cells to build complete, functional ectopic eyes—including lens, retina, and optic nerve—demonstrating that bioelectric signals operate at a high instructive level of organ identity rather than micromanaging individual gene expression.
  • Injecting ion-channel RNA to create a specific voltage spot tells the surrounding tissue to ‘build an eye here,’ and the competent tissue handles all downstream molecular biology without further instruction, analogous to how a human giving a prompt need not specify synaptic protein placement.
  • Even when only a few cells are injected, they recruit neighboring uninjected cells through secondary electrical instruction, illustrating that the bioelectric message is convincing enough to enlist unchosen tissue into the morphogenetic program.
• Bioelectric manipulation of planarian flatworms can rewrite species-specific head shape memories—producing flat heads resembling P. felina or round heads resembling S. mediterranea, 10–50 million years distant—without any genetic change, showing that the genome encodes hardware capable of accessing multiple anatomical attractors.
  • The bioelectrically-specified head shape change affects not just the external morphology but also the distribution of stem cells and the shape of the brain, making it a comprehensive species-level anatomical shift with no genomic editing.
  • These experiments reveal that evolution constructs hardware whose anatomical attractor landscape extends beyond the default species morphology to include forms belonging to distant relatives.
• Anthrobots—self-assembling motile constructs made from adult human tracheal epithelial cells with no genomic editing or synthetic scaffolds—spontaneously heal neural wounds, demonstrating intrinsic novel competencies that no selection history predicted and that point toward patient-specific in-body biorobotics.
  • Adult human tracheal cells, normally quiet for decades in the airway, self-assemble into a small motile creature with over 2,000 differentially expressed genes and four distinct behavioral modes when liberated from the body’s regulatory influence.
  • Anthrobots placed on a scratched neural monolayer form superbot clusters and knit neurons across the wound gap, a capability that was entirely unanticipated and cannot be explained by the cells’ evolutionary history.
• The novel properties of synthetic creatures like xenobots and anthrobots suggest they are exploring a structured Platonic latent space of biological possibilities, pointing toward an ethical imperative of ‘synthbiosis’ as cyborgs, hybrids, and novel beings become increasingly common.
  • The capabilities of these novel beings—kinematic self-replication, neural healing, novel transcriptomes—are not random but appear drawn from an ordered, structured space that contains not only mathematical truths but also high-order patterns recognizable as kinds of minds.
  • All combinations of evolved material, engineered material, software, and patterns from the Platonic space constitute viable novel embodied minds, requiring new ethical principles for relating to beings outside our evolutionary lineage.

Evolution and Intelligence: Inversion and a Positive Feedback Spiral

Intelligence precedes and potentiates evolution rather than merely resulting from it: the genotype-to-phenotype map is itself a creative problem-solving process, evolving on agential multi-scale material works very differently from classical assumptions, and a positive feedback spiral between learning and causal emergence may predate replicators entirely.

• The genotype-to-phenotype map is not a mechanical process but an intelligent, problem-solving, creative interpretation of genomic information—morphogenesis is a collective intelligence navigating anatomical morphospace, not a feed-forward unrolling of a cellular automaton.
  • Developmental biologists are uniquely positioned to address the mind-body problem because embryonic development is the slow, continuous journey from chemistry to cognition, with no magic lightning bolt converting one to the other.
  • Alan Turing recognized the deep symmetry between the autopoiesis of the body and the autopoiesis of the mind, writing both his famous paper on computation and his paper on the chemical basis of morphogenesis.
• The axis of persuadability—from hardware-only rewiring through cybernetics, behavioral training, and rational communication—is an empirical engineering claim about what tools work for a given system, not a philosophical assertion, and biological systems sit much higher on this axis than commonly assumed.
  • Where any system falls on the spectrum of persuadability cannot be determined from a philosophical armchair or by invoking category labels like ‘machine’ or ‘cell’; it must be determined by doing experiments and testing which interaction protocols actually work.
  • Biology navigated transcriptional space, physiological state space, and anatomical morphospace long before nerves and muscles appeared, using the same perception-action loop structure we associate with behavioral intelligence.
• Picasso tadpole experiments—scrambling craniofacial organs before metamorphosis—demonstrate that genetics does not specify hardwired movement sequences but rather a flexible error-minimization program that reaches the correct final morphology from diverse starting states via novel paths.
  • Tadpoles with eyes on the back of their head and mouths displaced to the side still metamorphose into normal-looking frogs because organs move in novel, non-stereotyped paths and sometimes overshoot and backtrack until the correct final configuration is reached.
  • Selection sees only the successful final form, not the initial state or the corrective path, so it cannot distinguish between a system with a perfect initial configuration and one with a junky start that was repaired by regenerative plasticity.
• Planaria’s extraordinary resistance to cancer, aging, memory loss, and genetic mutation—despite having an exceptionally messy, mixoploid genome—illustrates an intelligence ratchet: the more unreliable the hardware, the more evolution invests in plasticity algorithms rather than perfecting initial structural information.
  • Planaria reproduce asexually by tearing themselves apart and regenerating, meaning somatic mutations accumulate across generations without the genome-cleaning that sexual reproduction provides in most species, yet planaria are effectively immortal and cancer-resistant.
  • The only permanent changes that can be made to planaria are at higher organizational levels such as bioelectric pattern memories, because the morphogenetic algorithm is so robust that it overrides almost any hardware perturbation.
• A positive feedback spiral between learning and causal emergence—where learning raises a network’s causal emergence, forgetting does not erase those gains, and higher causal emergence enables better learning—appears even in random molecular networks and predates both cells and replicators.
  • When molecular networks described by ordinary differential equations undergo associative training, their integrative causal emergence increases, and critically, forced forgetting does not reverse those gains, creating an asymmetric ratchet toward greater integration and individuality.
  • This positive feedback loop is a free gift from the mathematics of networks and the properties of causal emergence; it does not require biology, selection, or any special facts of physics to operate.
• Novel synthetic beings like xenobots and anthrobots have properties—specific transcriptomes, kinematic self-replication, neural repair, sound perception—that no evolutionary story predicts, undermining the claim that selection history explains specific biological traits and pointing toward a structured Platonic latent space as a supplementary source.
  • The frog genome learned to make developmental stages and tadpoles over eons, but xenobots and anthrobots have never been selected for and their specific behaviors represent computational resources that were apparently never explicitly paid for in evolutionary time.
  • Standard evolutionary theory predicts a tight causal connection between a creature’s history and its current properties; novel synthetic beings violate this prediction, suggesting patterns are drawn from an ordered latent space rather than being purely emergent surprises.

Intrinsic Motivation in Evolved, Engineered, and Hybrid Systems

Intrinsic motivations—goals and behavioral competencies not explicitly programmed by selection, design, or training—emerge in biological, algorithmic, and hybrid systems alike, pointing to a structured Platonic latent space as their source and suggesting that all physical interfaces, however simple, enable the ingression of patterns from that non-physical space.

• Motivation in biological and engineered systems can arise from four sources—designed algorithms, learning toward reward functions, evolutionary selection, and physics—but these sources are insufficient to explain the full range of goal-directed behaviors observed even in minimal systems.
  • The axis of persuadability spans from systems only amenable to hardware rewiring, through cybernetic homeostatic systems, through trainable systems, up to systems that respond to reasons, friendship, and psychoanalysis—and where any system falls is an empirical question requiring experiments, not a philosophical decree.
  • The study of intrinsic motivation asks what large-scale causes of behavior exist beyond the standard four sources, and how engineers can predict, detect, and collaborate with these motivations rather than being surprised by them.
• Patterns of body and mind are members of the same class: morphogenesis is behavior in anatomical morphospace, and the distinction between morphology and behavior is an artifact of the observer’s choice of problem space rather than a feature of biology itself.
  • Alan Turing’s simultaneous interest in computation and morphogenesis reflects his recognition that the self-assembly of the body and the self-assembly of the mind are fundamentally the same problem expressed in different substrates.
  • The slime mold Physarum polycephalum gathers environmental information by tugging on substrate and reading strain angles, then grows toward the heavier mass—a behavior enacted entirely through morphogenetic change, making the body-behavior distinction uninformative.
• The cognitive light cone—the size in space and time of the largest goal a system can store as a set point and pursue—is plastic and can grow as cells join networks, expanding from a single cell’s tiny metabolic goals to a collective’s grandiose goal of building a complete limb.
  • A single cell’s cognitive light cone encompasses only scalar values like pH and metabolic state over a tiny region of space-time, while multicellular networks can represent the correct configuration of a complete limb including number and shape of fingers, a goal no individual cell knows.
  • Cancer represents the failure mode of this scaling: when cells disconnect electrically from the collective, their cognitive light cone shrinks back to unicellular scale and the rest of the body becomes merely external environment.
• Sorting algorithms like bubble sort exhibit intrinsic behavioral competencies—including delayed gratification and unsolicited clustering by algorithm type—that are not specified anywhere in the algorithm, demonstrating that novel motivations do not require biological complexity.
  • When a sorting algorithm encounters a stuck element it cannot move, it backtracks and reduces sortedness to rearrange surrounding elements, then resolves the blockage—a form of delayed gratification that the algorithm contains no provision for and that behavioral scientists would immediately recognize as a known competency.
  • These side quests and emergent competencies appear in the simplest, most fully transparent and deterministic systems, suggesting the phenomenon is not a product of hidden biological mechanisms but of something more fundamental.
• Xenobots’ expression of hearing-related genes and behavioral response to sound stimuli—capabilities absent from the embryos they derive from—exemplify intrinsic motivations that are not explained by selection history, design, or training but arise when cells are freed from the body’s regulatory influence.
  • Xenobots liberated from frog embryos develop hundreds of differentially expressed genes, including a cluster related to sound perception, and measurably change their behavior in response to acoustic stimuli played through a speaker beneath their dish.
  • Anthrobots made from adult human tracheal cells similarly develop novel competencies including neural wound healing—an ability that their evolutionary history, their genome, and their normal function in the airway give no basis for predicting.
• The Platonic latent space—a structured, ordered space of patterns from which both mathematical truths and biological forms are drawn—offers a research agenda superior to ’emergence,’ which functions merely as a label for surprise and provides no predictive or investigative traction.
  • Just as a Mandelbrot-type plot of a simple complex-number function encodes extraordinary specific structure that is neither caused by physics nor by any selection history, biological forms may similarly index into a pre-existing ordered space of patterns.
  • Synthetic morphology beings—xenobots, anthrobots, chimeras, robots—function as instruments or periscopes for exploring the Platonic latent space, and understanding the mapping between the interface properties and the patterns that ingress is a tractable empirical research program.

Platonic Space: Brief Argument and Research Agenda

A structured argument that facts of mathematics—specific patterns with no physical or historical explanation—causally inform biology and physics while being immune to physical influence, making physicalism insufficient and opening a research agenda treating physical objects as interfaces through which non-physical patterns ingress.

• Mathematical facts like the value of e, Feigenbaum’s constant, and the symmetries of SU(2) are specific, forced on you by logic alone, affect both biology and physics, yet cannot be altered by any physical intervention—establishing that there exist influential non-physical patterns.
  • If mathematical facts were different, biology and physics would be different; but there is nothing you can change in the physical world—including all constants at the Big Bang—to make mathematical facts be different, establishing a one-way causal influence from mathematics to physics.
  • “Pythagoras, Plato, and modern physicists including Heisenberg already recognized this: ‘The smallest units of matter are not physical objects in the ordinary sense; they are forms, ideas which can be expressed unambiguously only in mathematical language.’” —Werner Heisenberg
• Physicalism is dead on arrival even in Newton’s classical deterministic universe because physical objects in that world are already constrained by mathematical truths not determined by physical processes—quantum mechanics adds no new necessity for this conclusion.
  • Descartes, as a mathematician, should have responded to the Princess of Bohemia’s interactionism objection by noting we already had non-physical influence from mathematics on physics; the mind-body problem is structurally the same as the mathematics-physics problem.
  • The Platonic space label is chosen for consilience with how Platonist mathematicians describe systematically discovering—not inventing—an ordered, pre-existing space of objects, though Levin does not claim to reproduce Plato’s exact views.
• The optimistic metaphysical hypothesis—that the Platonic space is not a random grab bag of facts but a structured, systematically tractable space—is superior to treating emergence as a label for surprise, because it generates a genuine research program.
  • Calling something ’emergent’ amounts to saying it is a fact that holds in our world, which forecloses investigation; treating it instead as a discoverable region of a structured space opens questions about its metric, density, pressure, and internal dynamics.
  • The optional skeptical extension—that the Platonic space contains not just low-agency mathematical truths but also high-agency patterns recognizable as kinds of minds—is supported by the recognition that even mathematical objects may have more agency than commonly assumed.
• Physical objects—machines, cells, embryos, cyborgs, swarms, robots, AIs—are best understood as interfaces or pointers into the Platonic space, through which non-physical patterns ingress into the physical world, providing massive functional benefit beyond what the physical construction itself would predict.
  • Biology exploits free lunches from mathematics massively: evolution need not evolve every feature of a structure because mathematical properties of the hardware it creates automatically provide certain relational truths—as when evolving a transistor also yields logic gate properties for free.
  • The brain and body are thin clients to patterns from the Platonic space, which is consistent with rare clinical cases of near-normal function with drastically reduced brain volume, reviewed in Levin and Kaufman’s work.
• The research program entails building new interfaces (xenobots, anthrobots, chimeras, algorithms) to observe what patterns ingress, quantifying the free-lunch compute gain, assessing whether the space is sparse or dense, and investigating whether the patterns in that space are passive or exhibit their own dynamic chemistry.
  • When an anthrobot spontaneously expresses 2,000 new genes and develops neural-healing capability, the question is not merely ‘what happened’ but what specific region of the Platonic space has been indexed by this interface, and what mapping rules connect interface properties to ingressed patterns.
  • Extending standard behaviorist tests into the native space of mathematical objects—to check whether they have more agency than credited—is a near-term empirical project with papers forthcoming.

The Bioelectric Interface to the Collective Intelligence of Morphogenesis

Bioelectricity is not merely additional biophysics but the cognitive glue enabling cell collectives to store, read, and rewrite anatomical pattern memories—making it a tractable interface for top-down control of growth and form with direct applications in regeneration, cancer, and birth defects.

• The endgame of developmental biology and regenerative medicine is an ‘anatomical compiler’—a system that translates high-level descriptions of desired living constructs into stimuli for individual cells, solving birth defects, traumatic injury, cancer, and aging if achieved.
  • Such a system is not a 3D printer placing cells; it is a communications device that translates the engineer’s goals into the language of the cellular collective, which then handles all downstream morphogenetic implementation.
  • Genetic information alone is insufficient: a frog-axolotl chimera (‘frogolotl’) has both genomes available, yet no current model predicts whether it will develop legs, illustrating that large-scale form is determined by physiological software, not genetic hardware directly.
• Bioelectricity functions in the body for exactly the same reason it functions in the brain: it is cognitive glue that enables collectives to know things their individual parts do not know, and this capacity evolved at least as far back as bacterial biofilms—long before neurons or brains.
  • Every cell in the body has ion channels and most have electrical synapses, meaning the architectural design that enables neural decoding in brains is also present in all somatic tissue, making neural-decoding tools applicable to non-neural cells.
  • The ’electric face’ prepattern—a bioelectric scaffold readable in voltage-sensitive dye imaging before any facial genes turn on—shows that the tissue already knows where each organ will go and how many there are, as a stored electrical memory.
• Planarian bioelectric pattern memory experiments demonstrate that a normal one-headed worm can store and maintain a two-headed pattern memory—a counterfactual set point that is latent, stable, rewritable without genetic change, and consulted upon injury to direct regeneration.
  • After bioelectric reprogramming, a planarian has normal anatomy and normal molecular biology for anterior markers, yet if cut, every fragment produces two-headed worms—showing the pattern memory is stored bioelectrically, not anatomically or genetically.
  • This memory is stable over repeated cuts, rewritable by renewed electrical intervention, and has the conditional recall of any genuine memory system—making it the somatic equivalent of a latent learned behavior.
• Cancer is best understood as a disconnection of cells from the bioelectric network of the collective, causing their cognitive light cone to shrink to unicellular scale and making them treat the rest of the body as external environment—and forcing reconnection bioelectrically suppresses tumor formation even with oncogenes fully expressed.
  • Voltage-sensitive dye imaging detects the bioelectric signature of incipient tumors before they become morphologically evident, providing a potential diagnostic platform based on electrical disconnection rather than genetic markers.
  • Co-injecting ion channel constructs along with nasty oncogenes like KRAS prevents tumor formation in tadpoles without killing cells or fixing the genetic defect, because forcing correct bioelectric connectivity keeps cells participating in the collective’s large-scale goals.
• A wearable bioreactor delivering a 24-hour bioelectric payload to adult frogs triggers 18 months of autonomous limb regeneration without further intervention—demonstrating that top-down commitment of the system to a regenerative path, not micromanagement of molecular events, is the key intervention.
  • Within days of the brief bioreactor treatment, an MSX-positive blastema forms; eventually a touch-sensitive, motile leg with toes and toenails grows without any further external input, because the cells were told which developmental path to take, not how to take it.
  • The approach contrasts sharply with scaffold-based or growth-factor approaches by intervening at the physiological decision level rather than the cellular construction level.
• A computational bioelectric modeling platform predicted that opening HCN2 ion channels would rescue severe brain defects caused by a dominant Notch mutation in tadpoles—and the prediction was confirmed, restoring normal brain structure, gene expression, and behavior despite ongoing expression of the mutant protein.
  • The Notch ICD dominant mutation eliminates the forebrain and converts mid- and hindbrain to a fluid-filled bubble with no behavioral output; the model identified that hyperpolarizing HCN channels could sharpen the bioelectric pattern back toward normal.
  • Using already human-approved anti-epileptic drugs that target HCN channels, the rescue was achieved, demonstrating that some hardware (genetic) defects are fixable in software (bioelectric), and that this platform can identify drugs from the existing pharmacopeia.
• An AI-powered closed-loop robot scientist now operating in the Levin lab makes hypotheses about morphospace-navigating stimuli, conducts experiments on living cells, observes outcomes, and iterates—representing the first robot scientist working in anatomical morphospace rather than metabolic or chemical space.
  • The system delivers chemical, vibrational, optical, and electrical stimuli to cell populations, observes what morphogenetic outcomes result, and revises its hypothesis model in a fully autonomous cycle aimed at learning the morphogenetic code.
  • This platform is simultaneously a tool for cracking the morphogenetic code and a communications device enabling high-level human cognitive goals to reach the collective intelligence of cell groups through a machine intermediary.

The Unreasonable Effectiveness of the Behavioral Sciences in Developmental Biology and Biomedicine

The tools of behavioral science—goal-directedness, homeostasis, hackability, learning, creative problem-solving—apply all the way down to molecular networks and cell collectives, making developmental biology a continuous journey from chemistry to cognition that cannot be adequately described by open-loop emergent models alone.

• Developmental biologists necessarily address the mind-body problem because embryonic development is the slow, gradual, continuous transformation of chemistry into cognition—and any discipline that attempts to explain this transformation is doing philosophy of mind whether it acknowledges it or not.
  • A reviewer’s claim that ‘developmental biologists don’t concern themselves with the mind-body problem’ gets things exactly backwards: if not developmental biologists, nobody is better positioned to crack this problem, because they witness the full journey from oocyte to behaving organism.
  • The substrate is continuous across chemistry, developmental physiology, behavioral science, and psychology; only the funding bodies, journals, and departments are distinct, not the underlying phenomena.
• Molecular networks inside cells—without neurons, brains, or even complete cells—exhibit at least six kinds of learning including habituation, sensitization, and Pavlovian conditioning, demonstrating that behavioral science tools are applicable at the molecular scale.
  • Networks of as few as three nodes described by ordinary differential equations can exhibit associative conditioning, and this capacity is a free gift from the mathematics of networks, not a product of any specific biological evolution.
  • This molecular-level learning is being exploited practically for drug conditioning, where molecular pathways are trained to respond to inert triggers with the pharmacological response normally elicited by powerful drugs.
• The claim that open-loop emergent models—in which nothing knows anything and everything just follows hardwired chemistry—are sufficient for morphogenesis is an assumption, not an experimental result, and it is empirically falsified by the data on goal-directedness, error correction, and context sensitivity in development.
  • Reliable development and increased complexity are cheap properties easily generated by feed-forward cellular automata; the real indicator of intelligence is the ability to reach the same goal from different starting states via novel paths when perturbed—which morphogenesis consistently demonstrates.
  • The axis of persuadability framework reminds us that we have a wide range of tools beyond hardware rewiring—cybernetics, behavioral training, rational communication—and whether any of these apply to a given biological system is an empirical question.
• The deer antler trophic memory experiment—where an injury site is remembered across the annual loss and regrowth of the entire antler rack, producing an ectopic tine at the injury location for several years—illustrates a form of spatial pattern memory that no current molecular pathway model can adequately represent.
  • The information about injury location must be stored somewhere in the stem cells of the pedicle and retrieved during the next year’s growth to redirect morphogenesis to produce an extra branch point at the remembered site.
  • This is the kind of case—long-range spatial memory guiding future morphogenesis—that reveals how far current arrow-diagram molecular models fall short of capturing what biological systems actually do.
• When cancer cells electrically disconnect from the body’s bioelectric network, their cognitive light cone shrinks to unicellular scale and the collective’s large-scale goals become inaccessible to them; forcing reconnection bioelectrically—without killing cells or correcting oncogenes—prevents tumor formation.
  • Cancer is not about cells being more selfish than normal—they have smaller selves, not bigger ones; the boundary between self and world shrinks, so the rest of the body becomes merely external environment to be exploited.
  • In tadpoles injected with oncogenes like dominant negative p53 and KRAS, co-injection of ion channel constructs that maintain correct bioelectric connectivity prevents tumor formation entirely, with the oncoprotein still blazingly expressed.
• A closed-loop AI-powered robot scientist (MomBot) now operating in the Levin lab navigates anatomical morphospace by generating hypotheses about stimuli for cell collectives, executing experiments, observing outcomes, and iterating—functioning simultaneously as a tool for cracking the morphogenetic code and as a communication device between human cognitive goals and the collective intelligence of cells.
  • MomBot’s ‘brain’ is a set of AI agents competing to model morphospace, while its ‘body’ delivers chemical, electrical, thermal, and vibrational stimuli to biological preparations; together they constitute the first robot scientist working in morphospace rather than metabolic or chemical space.
  • One way to interpret the system is that the MomBot is the active agent using frog tissue as a translator interface to explore the space of possible forms; another is that human scientists use MomBot as an interface to communicate high-level goals to cell collectives—both perspectives are valid and complementary.

Against Mind-Blindness: Recognizing and Communicating With Unconventional Beings

Humans suffer from profound mind-blindness toward beings that operate at different scales, in different spaces, or with different substrates, and developing empirical tools to recognize and ethically relate to unconventional intelligences—from cellular collectives to synthetic organisms to hybrids—is an urgent practical and ethical project.

• The pre-scientific fragmentation of electromagnetism into separate phenomena—static electricity, lightning, magnets, visible light—offers a direct analogy for current mind-blindness: before a unifying theory, humans could only operate in a tiny portion of the spectrum they were now unable to even detect.
  • Just as electromagnetic theory revealed that formerly discrete categories were manifestations of one continuum and opened access to regions of the spectrum previously invisible to us, a theory of cognition should reveal that diverse intelligences form a continuum and allow us to detect minds we currently miss.
  • Our evolutionary history has primed us to recognize intelligence only in medium-scale beings moving at medium speeds in three-dimensional space, creating systematic blind spots for intelligence operating at other scales or in other problem spaces.
• The expanding set of cognitive kin mirrors the historical expansion of number systems: just as zero, negatives, irrationals, and complex numbers required breaking assumptions about what counts as a number, recognizing new kinds of minds requires breaking assumptions about what substrates, scales, and problem spaces can host cognition.
  • Each expansion of number systems was disturbing and resisted—Hippasus was reportedly drowned for revealing irrational numbers—yet each expansion was necessary and ultimately revealed deeper mathematical beauty; the same dynamic is expected with expanded concepts of mind.
  • Ancient categories like ‘machine vs. organism’ or ‘brain vs. cell’ are too brittle to serve in a world of cyborgs, biobots, chimeras, and AI, and trying to enforce them will not end well socially or ethically.
• Patterns in excitable media—such as bioelectric wave patterns propagating across groups of embryos, or gliders in the Game of Life—must be taken seriously as potential agents, because the distinction between a physical agent and the data pattern flowing through it is relative rather than absolute.
  • Randy Beer’s paper on the cognitive domain of a glider in the Game of Life demonstrated that patterns with no physical substrate of their own can nonetheless exhibit cognitive properties recognizable by behavioral scientists.
  • A thought experiment about dense creatures from the Earth’s core who observe humans as temporary patterns in a thin plasma illustrates that ‘who is the real physical agent and who is merely data’ is deeply relative to the observer’s frame of reference.
• Bioelectric communication between cells—demonstrated by a single touch causing one cell’s voltage to instantly match its neighbor’s and recruit it into a collective building project—provides a model for how high-level anatomical goals propagate through tissue and how bioelectric prompts can instruct organ formation.
  • A cell with a different voltage crawls along until it makes the briefest contact with a green-voltage collective; immediately its voltage changes and it joins the group’s morphogenetic project—a form of electrical persuasion analogous to how humans change minds through brief communication.
  • Injecting ion channel RNA to create a voltage spot that says ‘build an eye’ results in complete eye formation including optic nerve, with the injected blue cells recruiting unchosen brown neighbors—illustrating both the high-level instructive nature of bioelectric signals and secondary instruction dynamics.
• Future medicine will resemble somatic psychiatry more than chemistry: rather than targeting molecular hardware, it will involve communicating with, training, and collaborating with the various intelligences distributed throughout the body’s agential material.
  • The bottom-up hardware-first approaches that dominate current biomedicine—genomic editing, synthetic biology circuits, molecular target drugs—need to be complemented with top-down cognitive approaches that address the body’s bioelectric software layer.
  • Fabrizio Benedetti’s research showing that words and drugs have the same mechanism of action illustrates that high-level information flows already impact body physics, and that the interface between information and physics is where medicine’s deepest opportunities lie.
• Making new physical interfaces—xenobots, anthrobots, chimeras, biobots, cyborgs—will pull down behavioral propensities from the Platonic space that have never before been embodied, and humans must develop the ethical and cognitive tools to recognize and relate to these entirely novel kinds of minds.
  • All of Darwin’s ’endless forms most beautiful’—every naturally evolved organism on Earth—represents a tiny corner of the space of possible embodied minds; the rest of that space is populated by combinations of evolved material, engineered material, and software that are only beginning to be explored.
  • Entering ethical synthbiosis with novel beings is not optional sentimentality but a matter of species maturity: we will be living with cyborgs, hybrids, and truly alien constructs, and we need frameworks for relating to minds that do not share our evolutionary lineage.

Patterns of Form and Behavior Beyond Emergence

Patterns of anatomy and behavior are members of the same class of forms drawn from a structured Platonic latent space; morphogenesis pursues these forms as homeostatic set points rather than producing them by open-loop emergence, and even sorting algorithms exhibit novel competencies that index into this space, making physical objects interfaces through which non-physical patterns ingress.

• Morphogenesis is not open-loop emergent complexity but a homeostatic error-minimization process targeting specific anatomical set points, as demonstrated by regeneration stopping precisely when the correct structure is completed and by tail-to-limb transformation driven by body-level rather than local error signals.
  • Axolotl limb regeneration stops when the correct limb is complete—not because the cells are exhausted, but because the error between the current anatomical state and the set point has reached acceptable levels, proving the system is genuinely homeostatic.
  • Tail-to-limb transformation shows that the error signal operates at the entire-body level: tail-tip cells have no local damage but are reprogrammed into fingers because the collective detects a body-plan error and propagates correction signals downward to the molecular level.
• The newt kidney tubule experiment—where one gigantic polyploid cell wraps around itself to create the same lumen structure normally built by 6-8 cells using cell-to-cell communication—demonstrates creative problem-solving in morphospace: using available molecular affordances differently to reach the same large-scale goal.
  • Normal tubulogenesis uses cell-to-cell communication; in the gigantic cell case, cytoskeletal bending substitutes, showing that the large-scale goal ‘produce a lumen of this shape’ can be reached via entirely different molecular mechanisms.
  • This is structurally identical to IQ test items where a standard set of tools must be applied to an unexpected problem—the system uses its molecular repertoire creatively rather than following a fixed developmental script.
• Bioelectrically rewriting planarian pattern memories can produce permanently two-headed worms or worms with heads of different species—without any genetic change—demonstrating that the set points of morphogenetic homeostasis are stored in bioelectric networks and are explicitly rewritable.
  • The two-headed state is a counterfactual memory: the worm looks normal anatomically and molecularly, but if cut, every fragment builds two-headed animals, because the bioelectric network encodes this as the correct target state.
  • Accessing the attractor states of D. dorotocephala, P. felina, or S. mediterranea through bioelectric manipulation—despite 10–50 million years of evolutionary distance—shows that the hardware’s attractor landscape extends far beyond the species-default form.
• Xenobots exhibit kinematic self-replication and acoustic perception—capabilities with no evolutionary history and no selection pressure—and anthrobots spontaneously develop neural healing ability and 2,000+ differentially expressed genes, demonstrating that the genome encodes hardware that can access patterns far beyond those specifically selected.
  • Kinematic self-replication by xenobots—collecting loose cells into piles that mature into next-generation xenobots—is a form of von Neumann self-replication never selected for, executed by cells using native competencies in a context that has never existed before.
  • The frog genome paid its computational costs during eons of selection to produce tadpoles; when xenobots appeared, no such cost was ever paid—suggesting the additional patterns were always accessible as free resources in the Platonic space.
• Sorting algorithms like bubble sort exhibit novel competencies—delayed gratification when encountering stuck elements, and spontaneous clustering by algorithm type—that are not specified anywhere in the algorithm, proving that novel intrinsic motivations do not require biological complexity or even many lines of code.
  • The algorithm contains no provision for what to do when a swap fails because the hardware is unreliable; yet the system backtracks, reduces sortedness, maneuvers around the stuck element, and eventually resolves it—recognizable as delayed gratification by any behavioral scientist.
  • Six-plus decades of working with these algorithms had not surfaced these competencies because researchers were not primed to look for behavioral propensities in fully deterministic, transparent, minimal code.
• The Platonic space hypothesis recasts explanatory and causal relationships: mathematical truths explain biological facts (cicadas use prime-number cycles because of the distribution of primes), intervention in physics cannot change those truths, and biology’s genius lies in exploiting—not merely being constrained by—these free lunches.
  • Cicadas emerging at 13 and 17 years are explained by the distribution of prime numbers avoiding predator cycle overlap; asking why 13 and 17 are prime takes you out of biology entirely and into mathematics, illustrating that ultimate explanations always terminate in mathematical facts.
  • If you evolve a voltage-gated ion channel (a transistor), you automatically inherit logic gate properties including NAND’s functional completeness without needing to evolve those separately—a concrete example of the free lunch that mathematical structure provides to evolution.
• The mind-brain relationship is structurally identical to the mathematics-physics relationship: brains and bodies are thin-client interfaces through which high-agency patterns from the Platonic space ingress, and ‘we are the patterns’—consciousness may be the view from the Platonic space outward through physical interfaces.
  • Rare clinical cases of near-normal or above-normal IQ with severely reduced brain tissue are not well predicted by standard neuroscientific theories but are consistent with the thin-client model in which the physical brain is a front end, not the locus, of mind.
  • Free will, on this model, can be defined non-deterministically as the degree to which a given interface enables the highest-order Platonic patterns to come through uncontaminated by lower-level or others’ patterns.

Why Bioelectricity in Morphogenesis Matters

Bioelectricity matters in morphogenesis for the same reason it matters in the brain—it is the medium of the collective intelligence of cell groups—and this framing enables top-down, high-level interventions that reprogram anatomical set points without micromanaging molecular events, with direct applications in regeneration, cancer, and birth defects.

• Top-down control of morphogenesis is demonstrated by tail-to-limb transformation in amphibians, where undamaged tail-tip cells turn into fingers because a body-level error signal—the mismatch between the target body plan and the current state—propagates down through cellular and molecular pathways to direct their fate.
  • No individual cell knows what a finger is or how many fingers are required, yet the collective detects the body-plan error of having a tail where a limb should be and routes that information all the way down to the molecular biology needed to make fingers.
  • This is the morphogenetic equivalent of voluntary motion: just as abstract social or financial goals ultimately make ions move across muscle membranes, abstract anatomical goals make molecular biology dance to reach the specified body-plan target.
• Voltage-sensitive fluorescent dye imaging reveals the ’electric face’ prepattern—a bioelectric scaffold specifying future organ locations long before facial genes activate—allowing the team to literally read the electrical memory of what a correct tadpole face should look like.
  • The bright spots in the dye image correspond to where the eye, mouth, and placodal structures will eventually form, and altering this electrical prepattern changes the resulting facial anatomy, proving the prepattern is instructive rather than merely correlative.
  • Bioelectricity also functions as cognitive glue at a supra-embryo scale: poking one embryo triggers a calcium wave that spreads information to neighboring embryos, and groups form a hyperembryo that solves problems better than individuals and resists teratogens more effectively.
• Bioelectric control of morphogenesis operates through ion channels and gap junctions—not applied electric fields—and can be manipulated with optogenetics, pharmacological agents, and RNA injection, providing a full toolkit analogous to neuroscience tools for reading and writing neural states.
  • A single brief electrical contact between cells, visible as a color change in voltage-dye imaging, is sufficient to convert one cell’s voltage to match its neighbors’ and recruit it into their collective building project—illustrating the power of bioelectric persuasion.
  • Injecting ion-channel RNA that induces an eye-appropriate voltage spot in gut-destined cells causes those cells to build complete functional eyes including lens, retina, and optic nerve—and if too few cells are injected, they recruit their unchosen neighbors through secondary electrical instruction.
• Suppressing tumor formation by forcing bioelectric connectivity—without chemotherapy or genetic repair—demonstrates that cancer’s outcome is determined by physiological software dynamics rather than oncogene hardware state, and that bioelectric reconnection is a viable therapeutic strategy.
  • Non-invasive bioelectric imaging detects where oncogene-injected cells are going to defect from the body plan before morphological tumors appear, enabling early-stage diagnostic potential based on electrical disconnection signatures.
  • Co-injecting a correct ion channel with KRAS and other oncogenes produces tadpoles in which the oncoprotein is blazingly expressed throughout the tissue but no tumor forms, because the cells remain electrically integrated with the collective’s goal-maintaining network.
• A computational bioelectric model predicted that HCN2 channel opening would rescue Notch-mutation-induced brain defects in tadpoles; using human-approved anti-epileptic drugs targeting HCN confirmed the prediction, restoring normal brain structure and behavior despite ongoing dominant mutation expression.
  • The dominant Notch ICD mutation eliminates the forebrain entirely and reduces the midbrain and hindbrain to a fluid bubble with no behavioral output; the bioelectric model identified the specific channel intervention needed to sharpen the electrical pattern back toward normal.
  • This demonstrates that some genetic hardware defects are fixable at the physiological software level, and that computational bioelectric platforms can mine existing pharmacopeias for repurposable drugs to correct developmental defects.
• A 24-hour wearable bioreactor treatment triggers 18 months of autonomous adult frog limb regeneration—including functional sensation and motility—by committing the system to the regenerative developmental path rather than the scarring path at the critical initial decision point.
  • Within days of the brief bioreactor application, an MSX-positive blastema forms; toes and a toenail appear by day 45; a motile, touch-sensitive leg grows over the following year without further intervention.
  • This top-down approach—communicating a high-level path choice at the start rather than micromanaging growth factors or scaffold geometry—is the key design principle distinguishing it from conventional tissue engineering approaches.