The Possible Role of Oscillators in General Intelligence

What allows cells to work together? Is it genetic instructions? Biological self-organization has been presupposed in, rather than explained, by our currently prevailing theories of life. The idea that organisms self-organize exclusively by means of a genetic code has been refuted. Instead, a complete explanation for life would require a formula by which organisms adjust their functional patterns of gene expression to the conditions of the present moment. Because organisms from many distinct genetic lineages can also spontaneously self-organize into living systems, this formula for downward causation must be universal to all life forms, which narrows the possibilities considerably. “Epistolution” is the idea that organismic self-organization and general intelligence are the same process, and it occurs by the merging and separating of critical oscillations in response to use and disuse. This testable conjecture means, if confirmed, that evolution can indeed be considered the result of intelligent behavior on the part of organisms which form biological Selves, if not on the part of the biosphere as a whole.

The idea that evolution is strictly blind is key to the coherence of modern evolutionary theory. By “blind,” I mean that the controlling logic of an organism, the influence that dictates its phenotype and behavior, is contained in the genetic material which has been assembled through random variation and natural selection through many generations, and not in response to the present moment. This form of causation is diachronic and reductive, in the sense that specific causes precede specific effects. This view requires that a very small molecule, usually DNA, carries the only causal influence on the next generation, and that only changes to this molecule are permanently assimilated as the cause of new adaptations. Because this posits that a tiny part has a controlling influence over the whole organism, this is called “upward causation.” In this view DNA is a blueprint for life. A person with this view might say, “My DNA interacts with the world through me.”

This theory of strict upward causation presupposes that living beings have the capacity for self-organization without explaining it. A more comprehensive explanation might go like this: An organism, through an unknown logic, exhibits purposive behavior. The results of this behavior are then naturally selected and affect all the inherited cellular structures, not only the DNA. This quasi-teleological alternative would mean that the controlling logic of an organism is instead contained in its responses to its immediate surroundings, an idea called “downward causation.” Much confusion has been caused by taking this metaphor for synchronous causation too literally [1]. Downward causation is simply the effect of many upward causes occurring at once, often interacting in complex ways. Larger scales incorporate more disparate causes, so any explanation of a process that extends across multiple scales will require this mode of thought to be complete. In this view, it is the present state of the universe that is the blueprint. A person with this view might say,“The world interacts with itself through me.”

This idea has previously been rejected because it appears to many to require a supernatural being, but I argue that this rejection is only a metaphysical prejudice. It is no more or less atheistic to presume that life is organized according to an arbitrary present than according to an arbitrary distant past. However, it does require some universal causal formula by which organisms could achieve downward self-organization. Since genetic material surely has much upward influence over phenotype and behavior, this downward view requires that changes acquired from the environment be both systematic and adaptive. If they were not, the heritable influence of these surroundings would spoil the adaptive value of genetic material.

Currently nearly all biological research assumes that strict upward causation on the basis of genetic material must ultimately be correct. This is myopic because Darwinian inheritance itself must presuppose downward causation to form a complete explanation even on its own merits. Darwin himself considered variation according to “use and disuse” essential to the theory he described in The Origin of Species. In addition, there are many well-known forms of life that falsify this upward explanation. A new approach to testing this universal downward causal formula will be demonstrated in this paper. I argue that in order to take downward causation into account we must explain self-organization through a process other than replication, and propose oscillation as a possible alternative.

Why strict replicator theory must be false

Advocates of strict genetic upward causation would acknowledge that there are many essential ingredients that must be present in order for an organism to be assembled and to function adaptively. These are elements such as space, nutrients, liquid water, amino acids, suitable gravity and temperature, sunlight, and so forth. These are necessary for life but not sufficient. The truly decisive cause, without which these elements would lie disparate and disorganized, according to this idea, is the information contained in DNA. In their view, DNA contains an informational code that causes these materials to rearrange themselves into a functional being. They explain self-organization by proposing that there is a script in DNA edited by many eons of natural selection that instructs the other elements to form into an organism when conditions are right.

This widely accepted view is refuted by a simple observation. All phenotypes and behaviors depend on patterns of gene expression, and these patterns are affected by the environment. If they were not, then genetic clones (like many monozygotic twins) would be identical behavioral robots. Instead, twins develop differently. Using the word “environment” is a way of ignoring all the distinctive or peculiar interactions each organism has with its surroundings. Instead, we should think of this framework of interactions as an “umwelt,” a term invented by the biologist Jacob von Uexküll to describe the world in its relation to an individual. The word umwelt literally translates from German as the “world unto.” Imagine two twins who have applied together to study abroad, but one is assigned to France while the other ends up in Portugal. At the end of the year, the two twins’ speech abilities have diverged. The twins never spoke in a common gene-derived language; neither do they each speak in their own random gibberish. Rather, in addition to a shared language, one has become proficient in French while the other speaks Portuguese. Phenotypic plasticity, including learning and other acquired functions, depends at least partly on different gene expression patterns in neurons. In order for genetic material to be the decisive or determining cause of its own expression patterns, it would have to be able to reliably force different umwelts to express it exactly the same way, and it does not.

The notion that life is purely an emergent consequence of replicating molecules is also refuted by the fact that many organisms, ourselves included, have a stage during reproduction when genetic material is randomly divided and distributed, called “crossing over”[2]. This is why we do not necessarily inherit any particular set of genetic material from our parents, but rather a randomly spliced mixture. This means that the individual segment of DNA cannot possibly control the replication process because it is not an independent unit of Darwinian selection. The strictly upward causal explanation would forbid any step in the process which destroyed these replicators. Crossing-over, by dividing them randomly, often destroys them. Replicator theory could be saved from this refutation by positing that the whole cell is the replicator, but in this case it would describe reproduction, not replication [3].

There are several other refutations of the notion that there is no hereditary downward causation in biology. Organisms such as plants do not carry a separate germline, sealed off from external influences, but rather can reproduce themselves from somatic lineages of cells which then also reproduce any of the changes they have acquired during active life. The majority of all life is asexual, single-celled, mitosing life. In these cases there is only one cell present, which means the whole organism is often inherited. Organelles with their own genetic systems such as mitochondria and chloroplasts are hard to explain without supposing that two lineages have fused and now reproduce as one Darwinian unit [4]. There are also no naked replicators in nature; even viruses require a host cell to replicate, suggesting that life (and therefore replication) requires at least a cell membrane and probably much more. Some acquired traits can be inherited epigenetically for many generations [5]. Many organisms can and do edit their genomes [6]. In some cases, mutation rates have been found to be nonrandom with respect to function [7, 8]. All these facts directly contradict the strict upward-causation replicator theory of biology.

The DNA molecule contributes vital heritable templates for making proteins, without which life as we know it would cease. But attributing the determining cause of all phenotype and behavior to DNA is like saying the entire match of tennis is controlled by the ball boy. Yes, the match will come to a halt if the ball boy does not deliver the correct ball, but he cannot independently decide when and where the ball is needed, or how the game is played.

Why downward causation equals general intelligence

Without replicator theory, we are left with a distressingly incomplete explanation for life. If genetic material cannot cause any umwelt to express the information it contains in a single definite way, the basic explanation for genetic inheritance fails. If there is no such code, every miniscule change in the umwelt should be constantly disorganizing the causal effects of genetics, spoiling the heritability of traits and the functionality of cells.

Look at the situation from the point of view of a particular beneficial cistron (a segment of DNA that codes for a particular polypeptide). If this bit of DNA was the size of your body, the outer boundary of the cell would be as far away from it as New Orleans is from Mobile, Alabama. A causal path from this bit of DNA to the edge of the human organism and then to all that is in that human’s external surroundings is an unimaginably far journey. We can demonstrate the causal influence the presence of this cistron has on phenotype by testing; when we knock out certain cistrons, the resulting organism is sometimes, but not always, markedly different [9]. But the reverse must be true as well; changing the umwelt must cause changes in gene expression. If the path back to the pattern of gene expression involves an unpredictable causal influence from the umwelt, what guarantees that this cistron will be expressed to build its special protein in the right conditions, rather than at an inopportune time?

Genetic expression in cells is not random or systematic in an arbitrary way, instead it is highly targeted to enhance cellular function in a specific context. Cistrons are not expressed as a step in a particular algorithm as far as we can tell. In fact, the expression rates of the same protein in adjacent cells sometimes varies by three orders of magnitude [10]. Emergent complexity in biology is thought, according to prevailing views, to arise entirely from simpler rules governing small parts. But if this were true, the higher levels of organization would only show either chaos or blind and arbitrary patterns, not the reliable and immediate adaptation to external conditions that is so often found. Because it cannot explain the influence of context on genetic expression, the theory of life that forms our current paradigm is woefully incomplete. Without reliable downward causation, genetic expression would be unreliable, therefore genetic inheritance could not work, and therefore life could not exist.

I have argued that this unexplained downward causal path is what we call general intelligence [11]. All organismic behavior, including human behavior, requires a reliable pattern of gene expression. Thus it is safe to assume that in at least one case, our own, the downward causal path from umwelt to genetic material constitutes general intelligence. I don’t believe that general intelligence is separately constituted, and that it then controls gene expression. I suspect instead that our intelligence is contained at least in part in the patterns of gene expression in all of our cells, including neurons, and contained in all the consequences of those patterns of expression which constitute our behavior.

Why must there be another sort of downward causation for other species? Humans, including human brains, are merely swarms of cells, each of which can live independently in a petri dish for a time. Our bodies contain a trillion or more foreign cells, on which we depend for our physical and mental function [12]. Our behavior as an organism, then, is some form of scaled-up cellular behavior. Not only can many other organisms solve general problems, but even temporary cell assemblages such as slime molds can learn and remember solutions under certain conditions [13]. There is no need to posit more than one mechanism here, one for humans and another for non-humans. The general intelligence of simpler organisms may often be unrecognizable to us simply because their niches are different from our own, therefore their problem spaces are unfamiliar. Because we are multicellular mosaic holobionts, intelligence in our own case is a collective affair [14]. There is no reason to assume that the principle which unites these many units into an intelligent collective is any different for any other form of life.

One objection could be raised at this point. Isn’t general intelligence the ability to represent knowledge externally, in symbolic form? It is certainly true that humans have unmatched abilities to create external templates. But perhaps the ability to create and use the templates in the genetic library of nucleobases constitutes the same property. The enaction of active life entails a representation of knowledge in the language of the genome. Only humans use complex external symbols, but all life uses genes. If this ability to express genes to solve problems does not count as knowledge, then it means that the basketball player does not know how to shoot the ball unless he is able to completely describe this process verbally. A complete verbal description of exactly how this or any other skilled physical feat is done is, of course, impossible, yet few would suggest that these actions lack intelligence.

Why downward causation must follow a universal formula

Why not assume that for every segment of DNA there is a bespoke causal path downwards from umwelt to base pair? In this view there would be a unique chain of downward causes for genetic expression for every lineage of organisms, indeed for every base pair in every section of every genome. This is the implicit assumption of the current paradigm in biology today, since it supposes that an organized bioworld emerges due to genetics. For each segment of genetic material to be the causally complete executor of its own expression, it would have to precisely control its own unique downward causal path.

This widely held assumption is likely false because life that is very diverse at the level of the genome still organizes itself into multiple levels of function. Organized life exists as cells, tissues, organs, microbiomes, multicellular mosaic organisms, holobionts, superorganisms, swarms, and ecosystems. The fact that organisms as diverse as our own microbiota form obligate parts of our own bodies, akin to functional organs, suggests that there must instead be a universal formula that governs downward causation rather than an enormously high number of separate bespoke paths. Human microbiomes are acquired from the trillions of microbiota in each human’s umwelt, comprising a stable, functioning system as unique on the level of the operational taxonomic unit as a fingerprint [12]. Millions of diverse lineages of organisms all cooperate in the causal path that leads from a human-DNA cistron to the particular human’s umwelt and back to cistron.

None of the cells in a holobiont operates in a vacuum; they all influence one another to form a cognitive Self that defends its immune boundaries, heals wounds, completes embryological development, and carries out goal-oriented problem-solving. The Self is a biological concept as well as a psychological concept. The Self is the boundary of organization that an organism defends through its behavior. Holobionts sacrifice the well-being of individual cells in defense of the entire system, for example when immune responses destroy senescent cells. In cancer, somatic cells produce a foreign Self-within-a-Self, triggering an immune reaction from the body. The bulk of our immune cells are located in the gut, where they interact with the large population of microbiota that comprise our intestinal microbiome. Without some universal principle or formula of downward self-organization, how could this functional holobiont emerge from organisms with widely disparate genomes? If life depended on genetic instructions, at every turn the introduction of cells with separately evolved genomes into any living system would cause disarray, interrupting function in every cell.

Why downward causation is a prerequisite for life, and for intelligence

Downward causation is the sum of all upward causes. It is the result of the presence of all things. This idea neither requires a supreme being, nor contradicts scientific materialism. Rather it simply takes seriously the interaction of all forms of upward causation at higher scales. The idea that large-scale forces are decisive follows logically from imagining that all small things have consequences. If downward causation can be heritable, then it must be the primary form of causation responsible for self-organization.

Organismal development requires a genome, but the process is not micromanaged by the genome. For example, results from Michael Levin’s lab on bioelectric signaling in organisms like planaria which regenerate flexibly despite genomic diversity suggest that it is the cellular umwelt that is reprogramming the goals of all cells, not their genomes [14]. In these experiments, cell groups can be manipulated through bioelectricity (turning on and off gap junctions between cells) to express novel forms of regenerated morphology. This demonstrates that in at least one form of life, the factors controlling development and behavior are largely intercellular. There is no reason to believe, simply because the intercellular influences in other species have not been so easily deciphered, that other cells are controlled any differently.

Levin describes his experiments on planaria as “reprogramming the goals of cells [15].” It is only after these goals are met that reproduction can occur. Staying organized long enough to reproduce successfully is a prerequisite for natural selection. This problem has typically been explained in the reverse fashion by evolutionary theorists who have claimed that complex metabolism can only be the result of natural selection. But no one can deny that however the first life form was configured, it contained enough of a self-regulating process to successfully reproduce itself, and there could have been no natural selection until something reproduced. Once natural selection began, it’s easy to see the incentives for parts of the organism to evolve into hereditary particles and for those hereditary particles to develop further complexity. Whatever the formula for downward causation is, it must have been present in the first form of life.

The causal interactions of cell groups of any size in principle have no outer boundary. Every set of molecules we can imagine in a living system of any size is continually interacting with the set of molecules exterior to it. There is nowhere in nature for a causally independent agent to be located. The upward consequences of one cell’s interactions will be the downward causes of the interactions of neighboring cells. This is true not only of cells, but also true of organs, organisms, swarms and every other organizational level of life. Causal communication must in principle be ubiquitous not only between cells in the same organism, but potentially even between all cells in the entire biosphere. Denis Noble describes this by saying “there is no privileged level of causation in biology.” This is the central insight of systems biology. [16]

There is an additional reason why downward causation must be the primary, decisive logic of biology. Ashby’s law of requisite variety [17], developed during the cybernetics era, requires that when distinguishing between the controlling element of a system and the controlled element, the set of possible states of the controlled element must be fewer than the possible states of the controller. Otherwise, the controlled unit could drift into ungovernable motion not accounted for by the controller. In a system as large and complex as a human holobiont, a small molecule such as DNA could certainly have a decisive emergent effect, but it could not function as a controller for the larger system. The cell groups that influence cells must be larger and more complex than the single isolated cells. Likewise, the genome must be a transducer, not a controller, of its transcriptome. Cellular behavior must be controlled from the outside in, not from the inside out.

Micro-conditions in biology are highly stochastic, while macro-conditions are highly stable. At the level of individual genetic sequences in each cell, expression rates are highly unpredictable. At the level of the ecosystem, however, changes happen very slowly and large sets of conditions hardly change at all. This is why biology cannot be based on the consequences of particular small parts, because if this were so the stochasticity or blind emergent patterning would only become compounded as you rose to higher levels, with no sensitivity to context. Instead it must be based on accommodation to macro-conditions. Once again, imagine yourself as a beneficial cistron, adapted to reliably benefit your host. Now looking outward into a downward-caused bioworld you would see only regular patterns of synchronized events above you to the external, stable conditions. Everything that was chaotic at a tiny scale would resolve itself into orderly larger patterns because of its intimate reliance on the stability above. The whole system would correct the activity of the tiny parts rather than being driven awry by their stochasticity. A cause that you exerted upward would be returned reliably back to you (or perhaps to a cell near you) at just the right moment for your special protein to be expressed just when it was needed for the larger functionality of the whole living system. This orderly arrangement across scales is in fact what we do observe.

What characteristics must the downward formula have?

The universal formula for downward causation must be composed from elements that are present in every living cell. It must also be present in some scaled-up form in all multicellular Selves. What universal characteristics does intelligent life have that could also be common to all the single cells in nature?

All organisms have some critical oscillators, or chemical reactions that cycle over and over on schedules that interact with one another. Perhaps it is not the algorithm, or set of steps, that carries the downward adaptation to context into the genome at all, but rather the timing of those steps. A biological oscillation can serve as a form of anticipation, and in what does general intelligence consist if not in accurate anticipations? Not everything in biology oscillates, of course, but this does not prevent oscillations from being crucial, and possibly overlooked, causal forces. Some philosophical confusion about downward causation can be removed by supposing that the changes of function in cycles of sleeping, waking, and dreaming serve to separate many upward and downward effects in time. Even the lifespan itself of an organism can be seen as an extended oscillation. Perhaps there is a causal formula that merges or separates these oscillations based on external conditions, and thus forces life forms with widely divergent evolutionary histories to work together in spontaneous functional groups. There can really be only one candidate for this formula when you look at the problem in this way: “If used, reinforce; else if disused, mutate blindly” [11]. It is worth noticing that genetic evolution follows this “epistolution” formula, which supports my contention that genetic causation is only part of a more universal process.

Physiological conditioning also follows this formula. The body forms more resilient structures and processes by regular pressure or stress. Even broken bones often heal into a structure that is stronger than the original. This general physiological phenomenon extends to the formation of habits and addictions in the behavioral domain. Practice makes perfect. Organisms that engage repeatedly in routines become established in those routines, and can perform them with greater proficiency. The converse of this process is atrophy. Unused structures degrade and become waste or vestigial remnants. Infrequently performed behaviors are forgotten or become more difficult to execute. While not all aspects of physiology respond to use and disuse, perhaps all critical oscillations do.

All highly intelligent organisms are dependent on sleep for their cognitive performance. Sleep is now understood not as a passive stage of life but as a highly active stage of whole-body physiological function that repairs cellular structures [18]. This process of daily repair is necessary, presumably, because damage is also necessary. But why must intelligent beings damage themselves so extensively during active life that they must become prone, unconscious, and unresponsive for, in our case, one third of their lives? Why not evolve to repair on-the-go? Surely it would be of great adaptive benefit to shorten the time spent in such a vulnerable condition in order to compete for resources and mates? Perhaps sleep allows the organism to distinguish between use and disuse.

Organisms deprived of energy live longer. Results from yeast to flies to worms to humans suggest that the chemical reaction that comprises an organism’s lifespan can be extended by reducing caloric availability. This reaction seems to be, like physiological conditioning, a beneficial response to stress [19]. If an organism’s lifespan itself is an oscillation, it seems to be universally slowed by starvation. Many theorists have explained lifespan through antagonistic pleiotropy [20], but a phenomenon can be compatible with genetic effects without being explained by it. All phenotypes are supported by genetics, but the question is why do they occur? The oak and the octopus have comparable numbers of protein-coding sequences which are presumably all pleiotropic [21, 22], and they reproduce at roughly similar rates, but an oak becomes less likely to die each year (biologically younger) until it has lived hundreds of years, while some octopuses live only one year before senescing. Unlike aged oaks, young octopuses are masters of plasticity and learning. By supporting senescence, a given genome is ensuring its persistence in the future by disfavoring immediate replication in the current generation in exchange for the downward-caused acquisition of function in the next generation.

Levin states, “evolution does not produce machines built for a special purpose, but problem-solving machines [23].” Adjusting oscillations in response to use and disuse could be the universal feature that explains this difference. Perhaps this logic incorporates even the apparently chaotic exponential growth of gut bacteria into a reliable system of predictable stimuli. A strain of microbe that is increasing in the gut provides a form of reinforcement from use, while the space of decreasing strains or strains that are diversifying provides a form of mutation or atrophy from disuse. This formula could be the bridge between multicellular life and microbiome dynamics that allows their cooperation.

Experiments on rhythmic protein expression suggest that there are many independent clocks in every organ of the human body [24]. The nervous system can also be seen as a network of oscillators. Neurons fire and recharge over and over in order to harmonize behaviors and physiological processes with external conditions. A central nervous system is considered the seat of “agency,” but octopuses exhibit the intelligence of dogs or greater without highly centralized decision-making [25, 26]. Instead of executing commands, perhaps a nervous system is bringing the rhythmic changes of the external world inward and reflecting its patterns in expressing the genome. This might explain why neurons appear to use only approximate, not precise computations, to produce intelligent behavior. If cognition is seen as the hallucinatory rhythm of critical oscillators moving in time to the expected world, it might explain the placebo effect, and optical illusions.

What is dreaming?

Self-organization means change toward an intended goal-state, an organized state; it means a pseudo-teleology. This requires a reference frame. Life can be said to be ordered, but with reference to what? The proposal above shows how waking use and disuse could form the frame for organization with reference to the external world, but a further step is also necessary. Self-organization also requires a Self. Knowledge requires a subject. Repair from active use and disuse could explain downward causation, but it cannot explain the establishment of a coherent boundary within which genetic material could remain functional. If all the causal influence comes from the outside, why wouldn’t the system quickly become incompatible with the information that is stored in the genome? It must also be organized with reference to itself.

Complex intelligent lineages of life not only share the cognitive necessity of sleep, but they share dreaming. Even cephalopods share the necessity of a REM-like stage during their daily bout of sleep [27]. I argue that this shows a convergent evolution on this solution for cognitive performance. It suggests there is something about general intelligence that requires the organic process that leads to this hallucinatory sensory experience. Perhaps REM is a repair process for use and disuse that occurs while cut off from the external world, thus cementing the cognitive individual. Perhaps REM is a form of internal downward causation, a synthesis process for the biological Self.

Imagine an organized group of cells attempting to achieve a goal. During a sports game the team members act not only individually but with collective intentions. There is always a stage where the team meets in a private “huddle” with itself to communicate, bond, and solidify itself as a unit of action. Without these huddles, the members’ individual purposes would drift as the game progressed, and their private ambitions might come into conflict. Why should a group of cells require a private conference of this sort? It is required if they must cooperate to form an independent unit of Darwinian selection. A holobiont may be formed by trillions of lineages of cells in coordinated function, but the ones that reproduce from a common genome must share harmonized goals. I could write this conjecture as all cells that reproduce together dream together.

Evolutionary consequences

If intelligence can be said to consist in the repair and atrophy of all cellular structures in response to use and disuse, then this also applies to the genome itself. The genome is repaired from damage just like any other organelle, cell, or tissue. This repair is part of the normal cyclical processes that cells enact, and it applies not just to somatic cells but to sex cells as well. This means that whatever genetic variation is inherited is the result of an intelligent action by the cell. The source of all variation, both heritable and nonheritable, is intelligence, therefore the source of the variation that natural selection acts upon is intelligence. This is a radical departure from our paradigmatic notions of biology, but it is an inescapable consequence of the realization that all the plasticity of a life form is a potential source of evolutionary novelty, not only the genetic plasticity, and even genetic plasticity is a result of intelligent responses to the umwelt. It is a consequence of understanding that the behavior of non-genetic material is what guides the expression, use, and repair of genetic material, and therefore its heritable changes. It is also a consequence of understanding that much more than genetic change is heritable. Not all of the changes in cells are heritable, of course, or even remembered for long. Forgetting is a crucial part of learning. But all of the actions of a cell are intelligent, therefore evolution results from intelligence.

This is not an argument for supernatural design or monotheism, or any sort of refutation of materialism. If this argument has religious implications they are something like what is called pantheism, the idea that the Universe is God and God is only visible through scientific empirical means, by conjecturing good explanations and testing them. Most scientists would have to agree with this creed I think, simply because all people have to accept some fundamental ideas, no matter how skeptical and empirical they consider that they are. As Popper would have reminded us, there is can be no empiricism at all without fundamental ideas of where to look and what to measure. This definition of intelligence does not involve any more foresight, on the part of the cell, than the tiny horizon of interaction that form the cells miniscule umwelt. And the narrow, momentary foresight that a cell can muster is hardly a grand cosmic plan.

The more formidable foresight of more complex organisms is of course part of the process that shapes the material which evolution works on as well, but in these cases the intentions that govern organic change are no more extensive than the intention an organism has to survive and to reproduce and to pursue whatever other whimsical goals it chooses in its momentary cognitive performance, the “struggle for life” in Darwin’s words. An organism only knows what it knows; it only knows what its umwelt can teach it. This theory does not propose that the biosphere as a whole has a will or an intention that pushes organisms toward an eventual telos as global living system, except in the sense that all organisms are responding to the present moment in a way that respects the known laws of physics. This physical forcing is, in a metaphorical way, a sort of guiding larger “higher power,” but only in the ordinary way that scientific materialism allows. Whether the biosphere as a whole can be said to have an umwelt and act as an intelligent system can be left for others to work out; that concept is a bridge too far even for me. As far as I can see, the boundary of intelligent behavior is the boundary of the cognitive and individual biological Self, which is also usually or perhaps always the boundary of the Darwinian unit of selection.

What does it mean, then, that evolution is intelligent? It means that the living bioworld is integrated with itself. It means that we, as organisms, are sensitive to the changes that happen around us, and that our responses are then felt throughout the living system. All organisms are likewise set up this way in regard to their neighborhood and to their neighbors. It means that in an ultimate sense we are responding to all life, and all life is responding to us. And we are more than responding; we are accommodating. We are competing with other organisms, but a crucial part of that competition is avoiding, as much as possible, mutual annhilation. We are engaged in a multidimensional search for better ways of doing things: mostly better for ourselves, but also by realizing and reacting intelligently to other beings, we are opening up new possibilities that further elaborate the opportunities for life. This does have implication for morality, in fact it explains how morality can emerge from biological living systems. That will be the subject of a further essay. For now it suffices to remind the reader that this version of systems biology intelligence is empirically testable, and to appeal for help from anyone who in reading this essay becomes interested in helping to produce evidence that it does or does not exist. My own test is described in this deck of instructions for creating a matrix of oscillators in a virtual umwelt, and evaluating the intelligence of their behavior. This is my first attempt at formulating a test; better progress will be made later on this front by sharper minds than my own.

Acknowledgements:

Barrie McLune, Stephen Newton Hailey, Mike Skiba, Michael Levin, Sam Goto, Ella Hoeppner, Russell Bonduriansky

References

1. Kim, J., Making Sense of Downward Causation, in Downward Causation, C.E. Peter Andersen, Niels Finnemann, Peder Christiansen, Editor. 2000, Arhaus University Press: Oxford.

2. Godfrey-Smith, P., Darwinian populations and natural selection. 2009, Oxford ; New York: Oxford University Press. viii, 207 p.

3. Godfrey-Smith, P., The Replicator in Retrospect. Biology and Philosophy, 2000. 15: p. 403–423.

4. Sagan, L., On the Origin of Mitosing Cells. Journal of Theoretical Biology, 1967. 14: p. 255–74.

5. Heard, E. and R.A. Martienssen, Transgenerational epigenetic inheritance: myths and mechanisms. Cell, 2014. 157(1): p. 95–109.

6. Shapiro, J.A., Evolution : a view from the 21st century. 2011, Upper Saddle River, N.J.: FT Press Science. xi, 253 p.

7. Martincorena, I., A.S. Seshasayee, and N.M. Luscombe, Evidence of non-random mutation rates suggests an evolutionary risk management strategy. Nature, 2012. 485(7396): p. 95–8.

8. Monroe, J.G., et al., Mutation bias reflects natural selection in Arabidopsis thaliana. Nature, 2022. 602(7895): p. 101–105.

9. Hillenmeyer, M.E., et al., The chemical genomic portrait of yeast: uncovering a phenotype for all genes. Science, 2008. 320(5874): p. 362–5.

10. Noble, D., Lecture to Cancer and Evolution Symposium, “Cellular Darwinism: Regulatory Networks, Stochasticity, and Selection in Cancer Development”. 2020.

11. Munford, C.S., Epistolution: How a Systems View of Biology May Explain General Intelligence. Artificial General Intelligence. AGI 2021. Lecture Notes in Computer Science, 2022. 13154.

12. Douglas, A.E., Fundamentals of microbiome science : how microbes shape animal biology. 2018, Princeton: Princeton University Press. viii, 236 pages.

13. Reid, C.R., et al., Slime mold uses an externalized spatial “memory” to navigate in complex environments. Proc Natl Acad Sci U S A, 2012. 109(43): p. 17490–4.

14. Levin, M., Technological Approach to Mind Everywhere: An Experimentally-Grounded Framework for Understanding Diverse Bodies and Minds. Front Syst Neurosci, 2022. 16: p. 768201.

15. Levin, M., Mike Levin — Live ALife 2020 Keynote — Robot Cancer. 2020, YouTube.

16. Noble, D., Dance to the tune of life : biological relativity. 2017, Cambridge ; New York: Cambridge University Press. xviii, 283 pages, 12 unnumbered pages of plates.

17. Ashby, W.R., Requisite variety and its implications for the control of complex systems. Cybernetica, 1958. 1:2: p. 83–99.

18. Anafi, R.C., M.S. Kayser, and D.M. Raizen, Exploring phylogeny to find the function of sleep. Nat Rev Neurosci, 2019. 20(2): p. 109–116.

19. Mattson, M.P. and R. de Cabo, Effects of Intermittent Fasting on Health, Aging, and Disease. Reply. N Engl J Med, 2020. 382(18): p. 1773–1774.

20. Austad, S.N. and J.M. Hoffman, Is antagonistic pleiotropy ubiquitous in aging biology? Evol Med Public Health, 2018. 2018(1): p. 287–294.

21. Gray, M.W., Lynn Margulis and the endosymbiont hypothesis: 50 years later. Mol Biol Cell, 2017. 28(10): p. 1285–1287.

22. Albertin, C.B., et al., The octopus genome and the evolution of cephalopod neural and morphological novelties. Nature, 2015. 524(7564): p. 220–4.

23. Levin, M., ICLR 2022 Workshop | From Cells to Societies: A conversation with Dr. Michael Levin. 2022, YouTube.

24. Zhang, R., et al., A circadian gene expression atlas in mammals: implications for biology and medicine. Proc Natl Acad Sci U S A, 2014. 111(45): p. 16219–24.

25. Shigeno, S., et al., Cephalopod Brains: An Overview of Current Knowledge to Facilitate Comparison With Vertebrates. Front Physiol, 2018. 9: p. 952.

26. Richter, J.N., B. Hochner, and M.J. Kuba, Pull or Push? Octopuses Solve a Puzzle Problem. PLoS One, 2016. 11(3): p. e0152048.

27. Frank, M.G., et al., A preliminary analysis of sleep-like states in the cuttlefish Sepia officinalis. PLoS One, 2012. 7(6): p. e38125.