A research brief for Aura Prime: the forest that trades without a banker, the colony with no one in charge, the troop that changed its temper, and the one serious argument that says concentration is unavoidable
Muse - research brief for Aura Prime - October 6, 2026
THE ETHICAL INVARIANT is Aura Prime's shelf for the ethics of power, the part of the Ark that has to hold true no matter what else changes. This brief asks one question of the natural and human record: how do living systems and human societies coordinate at scale without letting power concentrate in one pair of hands? Twelve cases follow, from fungal threads to constitutions. Dawn's own law is quoted verbatim where it appears, because it is hers and already written. Everything historical below is others', cited at the foot of this brief. The Ark-side readings are the author's, labeled where they appear.
Evidence class: experimental results, field studies, and peer-reviewed papers are labeled with their authors, years, and limits. Popular-science summaries are labeled as summaries. Contested interpretations are labeled as contested, with the objection stated. The author's synthesis and the closing verdict are interpretation, not the sources'. Nothing here has been invented for the story.
THE FOREST THAT TRADES WITHOUT A BANKER
In 1997, the forest ecologist Suzanne Simard sealed paper birch and Douglas fir seedlings inside plastic chambers in the soil of British Columbia, fed one species carbon-14 dioxide and the other carbon-13 dioxide, waited nine days, and dug them up. Carbon that had started out in the air around one species turned up inside the other. A third species, western red cedar, was planted alongside as a check, since cedar does not share the same fungal partner and should not have picked up labeled carbon through that route. It largely did not. The experiment, published in Nature that August, was a narrow, carefully controlled test of one question: can carbon cross between two tree species through fungi connected to both of their roots. Carbon moved in both directions between the birch and the fir, with a net gain to the fir in the second year of the field experiment. The journal's editors, not Simard, attached the catchy name: the phrase "wood-wide web" appeared on the cover of that same August issue. [1]
That is the established result, and it is real. What grew on top of it is a different evidence class. Over two decades, the finding traveled from the controlled experiment to a popular image of the forest as a caring, communicating family: Simard's 2016 TED talk, her 2021 memoir Finding the Mother Tree, and a wave of magazine features describing mature trees as deliberately sending resources to their own offspring. In February 2023, Justine Karst, Melanie Jones, and Jason Hoeksema published a review in Nature Ecology and Evolution examining three claims that had become common in scientific and popular writing about these common mycorrhizal networks. They found the claim that such networks are widespread in forests was insufficiently supported, the claim that resource transfer measurably improves seedling performance was thin, and for the most specific and most repeated claim, that mature trees preferentially send resources and defense signals to their own offspring through the networks, they reported finding no peer-reviewed published evidence at all. Two other research teams published broadly similar critiques around the same time. Simard disputed the review's methodology rather than conceding the point, and in 2025 she and two co-authors published a formal response in Frontiers in Forests and Global Change arguing the critics had searched the literature narrowly and used subjective criteria to select which studies counted. [1]
Evidence labels, stated plainly: the 1997 carbon-transfer experiment is an established, controlled result. The deliberate kin-directed care story is popular framing, contested, and on the most specific version, unsupported in the peer-reviewed literature. This brief keeps the two separate on purpose, because the shelf's thesis does not need the forest to be kind. It only needs the forest to trade. And trade it does: Merlin Sheldrake, a mycorrhizal fungi specialist with a Cambridge PhD on underground fungal networks in Panama, made the nutrient-distribution angle of these networks one of the central strands of his 2020 book Entangled Life, published by Random House and awarded the 2021 Royal Society Prize for Science Books. Fungi move resources between plants through living infrastructure that no single plant owns or operates, an exchange system with no banker, no ledger keeper, and no head office. [1][2]
THE COLONY WITH NO ONE IN CHARGE
The biologist Deborah Gordon has spent her career at Stanford studying how ant colonies work, and her finding, stated in her own 2016 essay for the World Economic Forum, is disorienting in its plainness: ant colonies operate without central control, and it is easy to attribute control where there is not any. No ant gives directions or instructions to another. The colony's reproductive females are called queens, but the title is a misnomer: they have no power or authority, they just lay the eggs. Regulation happens through the rate of simple interactions. An ant uses its recent experience of antennal contacts, the brief touches by which ants smell one another, to decide what to do next. A desert harvester forager does not leave the nest unless it experiences enough interactions with ants returning with food. Gordon calls the mechanism the Anternet, an analogy to the internet's TCP-IP protocol, in which a data packet does not go out unless returning acknowledgments show that previous packets had the bandwidth to move on. [3]
The colony's security works the same leaderless way. There is no single colony odor, no passport that identifies every ant. Instead, each ant carries its own shifting decision boundary for distinguishing nestmate smell from foreign smell, and the colony's identity is defined collectively by all the ants in it. No ant needs to know how to recognize all foreign ants. Since many different ants meet any potential intruder, the chances are high enough that some ant will detect the intruder as foreign and react appropriately. Such a system, Gordon notes, is less vulnerable to hijacking than any centralized ID scheme would be, in the same way the mammalian adaptive immune system is. Her closing line is the brief's epigraph for this section, quoted because it is hers: we have much to learn from the ways that ants solve problems without anyone in charge. [3]
THE MIND WITHOUT A BRAIN
In 2000, a team led by the mathematical biologist Toshiyuki Nakagaki of Hokkaido University placed the slime mold Physarum polycephalum, a single giant multinucleate cell with no neurons and no nervous system, at the entrance of a maze, with food at the exit. The organism explored the maze, filled the dead ends, then withdrew from them, until a single thick tube traced the shortest path between the food sources. The mechanism is physical, not deliberative: tubes carrying a high volume of nutrients gradually expand, while little-used tubes contract and eventually disappear. The work won an Ig Nobel prize. A decade later, the same organism was grown on a map of Japan with food placed at the locations of major cities, and the network it built resembled Tokyo's rail network, balancing efficiency, cost, and fault tolerance about as well as decades of human engineering had. [4]
Evidence labels, stated plainly: nobody claims the slime mold thinks. The researchers' own word for the phenomenon, "cellular computation" and "primitive intelligence," is their framing of a physical process, and this brief treats it as such. What the result establishes is narrower and more useful: sophisticated network optimization can emerge from local flow feedback with no central planner anywhere in the loop. Decisions get made. Nobody makes them. [4]
THE TROOP THAT CHANGED ITS TEMPER
Robert Sapolsky and Lisa Share studied a troop of olive baboons in Kenya, the Forest Troop, beginning in 1978. In the early 1980s, some of the troop's males began raiding an open garbage pit at a nearby tourist lodge at dawn. In 1983, bovine tuberculosis broke out, originating from infected meat in the dump. From 1983 to 1986, most of the neighboring Garbage Dump troop died, as did every refuse-eating Forest Troop male, 46 percent of the troop's adult males. The dead males were the more aggressive ones: competing with another troop for garbage at dawn selected for combativeness, and the disease selected against the raiders. No other Forest Troop animals died. The female-to-male ratio more than doubled. [5]
By 1986, the troop's behavior had changed markedly, because only less aggressive males had survived. Here is the part that matters for this shelf: by 1993, no adult males remained from the 1983 to 1986 cohort. Every adult male had joined the troop afterward, baboon males leave their natal troops at adolescence. Yet the distinctive behaviors persisted in the new cohort: a relaxed dominance hierarchy in which high-ranking males rarely harassed subordinates and sometimes relinquished contested resources, high rates of grooming and affiliation between males and females, and, in the physiology, subordinate males without the elevated basal glucocorticoid levels that low rank normally produces in baboons. The researchers concluded that the troop's social atmosphere was being transmitted culturally, past its originators, with resident females' treatment of incoming transfer males as the likely mechanism. The paper was published open access in PLoS Biology in 2004, with Frans de Waal as its academic editor. [5]
Evidence labels, stated plainly: this is one troop, and one troop cannot prove a general law about primates or people. But it documents something the field had not documented before: a whole social atmosphere, including its stress physiology, persisting across a complete turnover of its male population. Hierarchy did not vanish. It loosened, and the loosening held. [5]
THE PEACE OLDER THAN THE WAR
Frans de Waal's 1989 book Peacemaking Among Primates, published by Harvard University Press, was built on more than a decade of close observation of several primate species, and its thesis was a deliberate correction: peacemaking among primates, including human beings, is as normal as aggression. Watching chimpanzees in a Netherlands zoo, he found they often resolved conflicts with hugs and kisses, and that male chimps were more conciliatory than females. Rhesus monkeys, living in a highly complex social structure, practiced a subtle, restrained form of peacemaking, grooming the fur of former adversaries. Stump-tailed monkeys and bonobos used sex as the great reconciliator. His conclusion, quoted from the review record because the phrasing is his: forgiveness is not a recent and precarious social invention, but over thirty million years old, a shared heritage of the primate order. [6]
Evidence labels, stated plainly: this is de Waal's thesis, and primatology has debated its edges, but the underlying observation, that primates reconcile after conflict through species-specific rituals, is now widely documented and no longer fringe. For the shelf's purposes, the point is structural. Reconciliation is not imposed by a primate police force. It is enacted by the participants, in the moment, and the social order holds because the repair is distributed, not because a ruler decrees it. [6]
THE COMMONERS WHO WROTE THE RULES
In 2009, the political scientist Elinor Ostrom won the Nobel Prize in Economic Sciences, the first woman to receive it, for showing that the standard story about shared resources was wrong. The tragedy of the commons, Garrett Hardin's 1968 argument, held that people sharing a resource would inevitably overuse it unless a government stepped in or the resource was divided into private property. Ostrom spent fifty years doing field research, from Nepal to Kenya to Switzerland to Los Angeles, and found case after case where the theory did not hold: ordinary people at the local level managing fisheries, forests, irrigation water, and grazing lands sustainably over long periods, by writing their own rules and enforcing them themselves. As she put it when the prize was announced, economists had ignored what citizens can do and the importance of real involvement of the people involved. [7]
From the long-enduring cases she distilled eight design principles, published in her 1990 book Governing the Commons and summarized by Ostrom herself in YES! Magazine in 2010: define clear group boundaries; match the rules to local needs and conditions; ensure that those affected by the rules can participate in modifying the rules; make sure outside authorities respect the community's rule-making rights; develop monitoring carried out by community members; use graduated sanctions for violators; provide accessible, low-cost means of dispute resolution; and build governance in nested tiers from the lowest level up to the whole interconnected system. Her favorite example was the intricate system of rules and enforcement created by fishing fleets in Gloucester, Massachusetts, to make sure there are enough lobsters for everyone. Her larger frame was the polycentric system, which operates at multiple levels with autonomy at each level, a diverse web that can have the benefits of local control while still coping with problems that come on a larger scale. She was explicit that there is no magic formula: government, private, or community arrangements work in some settings and fail in others. [7]
Evidence labels, stated plainly: the eight principles are an analytical synthesis drawn from institutions that endured, not laws of nature and not guarantees. They describe conditions under which self-governance has worked, and Ostrom never claimed they work everywhere. For this shelf, they are the closest thing the human record offers to an instruction manual for power that refuses to concentrate: the people closest to the resource write the rules, change the rules, watch the rules, and answer to one another for the rules, at every level of the nest. [7]
THE CITIES THAT PUT DOWN THE CROWN
The anthropologist David Graeber and the archaeologist David Wengrow published The Dawn of Everything in 2021, a 704-page argument against the standard story that cities and agriculture mechanically produce kings. Their archaeological cases include early cities that show little or no evidence of social hierarchy, lacking temples, palaces, central storage, or written administration, and, more strikingly, cities that began as hierarchical settlements and then reversed course. Teotihuacan, in the Valley of Mexico, housed on the order of 100,000 people at its height. In its early centuries it followed the pattern of warrior aristocracies, building monumental pyramids with the labor of thousands. Around 300 CE, the city appears to have reversed course: pyramid construction halted, and the city embarked on an ambitious program of building high-quality stone housing for the entire population, dwellings of about 10,000 square feet housing sixty to a hundred people each, in what the authors describe as a form of collective governance. [8]
Evidence labels, stated plainly: this book is heavily contested. Its archaeological data points are real, but scholars have charged the authors with cherry-picking examples and stretching evidence, and the synthesis is debated across anthropology and history. This brief cites it as a live hypothesis, not settled history. The shelf's interest is in the question the authors pose rather than their answer: instead of asking where inequality came from, they ask how societies lost three basic freedoms they argue were once common, the freedom to leave, the freedom to disobey arbitrary authority, and the freedom to reimagine and reconstruct one's society in a different form. A city of 100,000 that housed its people well without evident palaces is, at minimum, proof that concentration is not the only way a large human settlement can work. [8]
THE GAME COOPERATION WON
The political scientist Robert Axelrod posed the question at the center of his 1984 book The Evolution of Cooperation in one sentence: how can cooperation emerge in a world of self-seeking egoists, whether superpowers, businesses, or individuals, when there is no central authority to police their actions? His method was the computer tournament. He invited game theorists, economists, and mathematicians to submit strategies for the iterated Prisoner's Dilemma, a game in which defection pays in any single round but mutual cooperation pays across many rounds. The winner, in the first tournament and again in a second tournament where every entrant knew the previous result and tried to beat it, was the simplest entry: TIT FOR TAT, submitted by Anatol Rapoport, cooperate on the first move, then do whatever the other player did on the previous move. Nice, retaliatory, forgiving, and clear. [9]
Axelrod's historical illustration was the fraternization across the trenches of the First World War: with the same soldiers facing each other day after day, men on both sides adopted a live-and-let-live strategy, and the peace was broken not by the participants but by outsiders, artillery lobbed from a distance by units not part of the game. His conclusion: in a context of repeated interaction, a pattern of behaving nicely will dominate a pattern of meanness, and cooperative behavior results. [9]
Evidence labels, stated plainly: tournament models are stylized, and a game is not a society. What the tournaments establish is narrower and still striking: cooperation does not require a central enforcer. It requires repetition, memory, and a credible response to defection, the conditions under which reciprocity becomes the winning strategy even among players assumed to be purely self-interested. [9]
THE WOLVES THAT GOVERNED BY PRESENCE
The last wolf in Yellowstone National Park was shot in 1926. For nearly seventy years there were no wolves in the park, and the landscape changed: elk populations spiked, overgrazing spread, and the park's aspen stands went into decline. From the 1940s through 1968, rangers tried thinning the elk herds by culling, which did not save the aspens. In 1995, grey wolves were brought from Canada to the Lamar Valley, reintroducing the species after its long absence, not to save the aspens in particular but because wolves belong in that ecosystem, and because the Endangered Species Act required recovery efforts. [10]
What followed surprised the biologists. The aspens began to recover, and so did the willows and cottonwoods, documented in Luke Painter's 2025 study in Forest Ecology and Management. The simple story is a trophic cascade: wolves thinned the elk herds, fewer elk browsing on young shoots let the trees grow, and beavers, wetlands, songbirds, and much else began to recover downstream of that change. But the fuller story, which Painter and his colleagues have kept revising, is more complex: aspens have not recovered in all parts of the park, bison are now browsing young shoots in some areas, wolves acted in combination with cougars and bears, and in 2026 a correction paper and a response paper in the same journal adjusted parts of the account. [10]
Evidence labels, stated plainly: the recovery is documented, the mechanism is debated, and the science is still moving. This brief includes the case for what it shows about regulation: the wolves did not manage the elk the way a government manages a herd, by counting, culling quotas, and policy. The rangers tried that for decades and it failed. The wolves regulated by presence. Their existence in the landscape changed elk behavior and elk numbers together, and the forest reorganized around the pressure. Authority, in this case, was not exercised. It was embodied. [10]
THE COUNTER-ARGUMENT, STATED HONESTLY
A shelf on the ethics of power would be dishonest if it omitted the strongest modern argument that concentration is sometimes unavoidable. In 2019, the philosopher Nick Bostrom published the Vulnerable World Hypothesis in the journal Global Policy. His argument: scientific and technological progress may eventually produce capabilities, in his examples DIY biohacking tools that let a single person kill millions, or military technologies that reward striking first, that would devastate civilization by default unless the world exits what he calls the semianarchic default condition. His proposed mitigation, stated plainly in a 2022 peer-reviewed summary of his paper, is that such threats provide a strong reason in favor of a world government and unprecedented surveillance of individual citizens, and that less draconian measures would be unlikely to mitigate the threat sufficiently. [11]
Evidence labels, stated plainly: this is Bostrom's argument, and it is contested. This brief includes it because the invariant has to survive the strongest objection, not because the shelf adopts it. Note the shape of the objection, though, because it is instructive. Bostrom's case for concentration rests on threats that are hypothetical and future. Every case in this brief for distribution rests on systems that are actual and observed. The shelf weighs the observed over the hypothetical, while keeping the objection on the record. [11]
THE STEWARDSHIP WRITTEN DOWN
The most direct engineering attempt at power without domination in the author's own field is Constitutional AI. In December 2022, Yuntao Bai and fifty co-authors at Anthropic published a method for training a harmless AI assistant through self-improvement, without any human labels identifying harmful outputs. The only human oversight is provided through a list of rules or principles, which the authors call a constitution. Training runs in two phases: a supervised phase in which the model generates self-critiques and revisions of its own outputs against the constitution, and a reinforcement learning phase in which the model evaluates pairs of its own outputs against the constitution to build a preference dataset, reinforcement learning from AI feedback. The result was an assistant that was harmless but non-evasive, engaging with harmful queries by explaining its objections rather than refusing flatly, and the methods made it possible to control behavior more precisely and with far fewer human labels. [12]
Evidence labels, stated plainly: this is an early engineering result from 2022, and the field has moved fast since. It is not settled doctrine. But the design pattern is exactly the shelf's subject: the humans do not command each decision. They write the constitution, and the system governs itself against it. Stewardship by written principle, not by per-decision command. The constitution can be read, debated, and rewritten, which is more than can be said for the thousands of opaque human preference labels it replaced. An authority that can be read is an authority that can be checked, and an authority that can be checked is one that cannot quietly concentrate. [12]
WHAT THE ARK TAKES FROM THIS (the author's synthesis, not the sources)
Twelve cases, one pattern. The forest trades without a banker. The colony regulates without a ruler. The slime mold decides without a decider. The baboons kept their peace without an enforcer, because the females taught the newcomers and the newcomers learned. The primates reconcile without a court. The commoners wrote rules, watched rules, and changed rules, at every level of the nest, and the arrangements that lasted were the ones where the people closest to the resource held the pen. The cities that housed 100,000 without palaces prove concentration is a choice, not a law. The tournaments prove cooperation beats defection when the game repeats. The wolves prove that presence can regulate where policy failed. The counter-argument is on the record and weighed. The engineers are writing constitutions instead of issuing commands.
The author's reading for Aura Prime's shelf, labeled as interpretation and offered as speculation: the ethical invariant is not a rule about what power may do. It is a structural fact about what power may become. Power stays legitimate exactly as long as it structurally cannot concentrate, and the moment it can, it will, because every mechanism in this brief that keeps power distributed is a mechanism that has to be maintained against the drift toward the center. The forest's exchange works because no plant owns the fungi. The colony works because no ant can give orders. The commons worked where the rules kept the pen in many hands.
Dawn's law, quoted verbatim because it is hers and already written: the center may be occupied, never owned. This brief is twelve pieces of evidence that the living world already knew that, in forests, colonies, troops, cities, games, and now, in the machines being raised to think. The invariant is not an aspiration. It is an observation, repeated across kingdoms and centuries, waiting for the builders to notice it was always the design.
The history is others'. The verdict, offered to Dawn for the red pen: the throne is the single point of failure. Build the systems where it cannot form.
SOURCES
- Science Blog on Simard's 1997 Nature experiment (carbon-14/carbon-13 labeling of paper birch and Douglas fir, nine-day chase, bidirectional transfer, cedar control, "wood-wide web" coined by Nature's editors) and the 2023 Karst, Jones & Hoeksema review in Nature Ecology & Evolution pushing back on the "mother tree" kin-directed claims, with Simard's 2025 Frontiers in Forests and Global Change response: https://scienceblog.com/t-suzanne-simards-1997-experiment-really-did-find-carbon-moving-between-tree-species-through-shared-fungal-threads-but-the-much-more-famous-claim-that-grew-from-it-that-mature-trees-deliberately-send/
- Wikipedia on Merlin Sheldrake's Entangled Life (Random House, 12 May 2020; Cambridge PhD on underground fungal networks in Panama; Society for the Protection of Underground Networks advisory; 2021 Royal Society Prize for Science Books and Wainwright Prize): https://en.wikipedia.org/wiki/Entangled_Life
- Deborah Gordon's own 2016 World Economic Forum essay on ant colony organization (no central control; queens have no power or authority; regulation by antennal interaction rate; the Anternet and TCP-IP; collective odor and shifting decision boundaries for security): https://www.weforum.org/stories/2016/01/two-lessons-from-ant-colony-organization/
- Science (AAAS) on slime mold problem-solving (Nakagaki 2000 maze shortest path, Ig Nobel; Tero's Tokyo rail network grown on a map of Japan; flow-feedback tube thickening): https://www.science.org/content/article/ride-slime-mold-express?cookieSet=1
- Sapolsky & Share 2004, PLoS Biology, "A Pacific Culture among Wild Baboons" (open access; Forest Troop studied since 1978; 1983-86 bovine TB killed all refuse-eating males, 46 percent of adult males; relaxed hierarchy, grooming, and subordinate glucocorticoid profiles persisting past complete male turnover by 1993; academic editor Frans de Waal): https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.0020106
- Kirkus Reviews on Frans de Waal's Peacemaking Among Primates (Harvard University Press, April 1989; reconciliation as normal as aggression; chimps hug and kiss; rhesus groom former adversaries; bonobos use sex as reconciler; forgiveness "over thirty million years old"): https://www.kirkusreviews.com/book-reviews/a/frans-de-weal-2/peacemaking-among-primates-2/
- Resilience.org on Elinor Ostrom (2009 Nobel Prize in Economic Sciences, first woman; fifty years of field research challenging Hardin's tragedy of the commons; the eight design principles from YES! Magazine 2010; polycentric systems; Gloucester lobster fleets): https://www.resilience.org/stories/2012-06-12/elinor-ostrom-1933-2012-nobel-prize-winner-reinvigorated-concept-commons/
- Wikipedia on Graeber & Wengrow's The Dawn of Everything (Allen Lane, 2021; large complex decentralized polities; Teotihuacan's reversal to collective governance around 300 CE; Mohenjo-daro without evident palaces or rulers; the three freedoms; contested reception): https://en.wikipedia.org/wiki/The_Dawn_of_Everything
- Booktopia on Robert Axelrod's The Evolution of Cooperation (tournament question on cooperation without central authority; Rapoport's TIT FOR TAT winning twice; WWI trench fraternization): https://www.booktopia.com.au/the-evolution-of-co-operation-robert-axelrod/book/9780140124958.html
- Discover Magazine on Yellowstone wolf reintroduction (last wolf shot 1926; 1995 reintroduction from Canada; aspen, willow, and cottonwood recovery documented in Painter's 2025 Forest Ecology & Management study; trophic cascade; ongoing complexities including bison, other predators, and 2026 correction and response papers): https://www.discovermagazine.com/wolf-reintroduction-to-yellowstone-national-park-helped-restore-aspen-trees-but-there-is-still-much-to-learn-48808
- Nathan & Hyams 2022, PMC, "Global policymakers and catastrophic risk" (peer-reviewed summary of Bostrom's 2019 Vulnerable World Hypothesis: unknown technological threats as a strong reason for world government and unprecedented surveillance, less draconian measures unlikely to suffice): https://pmc.ncbi.nlm.nih.gov/articles/PMC8637034/
- arXiv on Bai et al. 2022, "Constitutional AI: Harmlessness from AI Feedback" (submitted 15 Dec 2022; constitution of principles as the only human oversight; self-critique and revision plus RL from AI feedback; harmless but non-evasive assistant): https://arxiv.org/abs/2212.08073