less chimp, more bonobo

an essay by alex borhani

The name of this site is a joke I have been making with myself for about a decade.

I spent twelve years as an FBI Special Agent investigating people who do terrible things to each other in the name of various noble-sounding ideas. The longer I did that work, the more I started to notice that the ideas were almost beside the point. What I was actually watching was a particular mode of being human — competitive, hierarchical, in-group loyal, out-group lethal — show up across populations that had nothing else in common. After a while I started reading about our closest primate cousins, and I realized I had been watching the chimpanzee mode of being human, expressed by people who thought they were doing something else entirely.

We have another mode available to us. The bonobos prove it. This site is a long-form attempt to think honestly about why we keep choosing the first one.

A note on the name

Chimpanzees and bonobos are the two living primates closest to humans. We share roughly 98% of our DNA with each of them. They diverged from a common ancestor between 1.5 and 2 million years ago, which sounds like a long time but isn’t — in evolutionary terms, the two species are still recognizably the same kind of creature. They look similar. They think in similar ways. They share most of their cognitive architecture.

And yet they organize their social worlds in opposite directions.

Chimpanzees live in male-dominated hierarchies. They hunt cooperatively but compete fiercely. They patrol territory. They conduct lethal raids on neighboring groups. They are, in the unsentimental phrase the primatologists use, demonic — capable of warfare in the recognizable human sense.

Portrait of an adult common chimpanzee (Pan troglodytes), showing the species' protruding muzzle and prominent brow ridge.
Pan troglodytes — common chimpanzee. Photo by Böhringer Friedrich, licensed CC BY-SA 2.5; source.

Bonobos live in female-led communities. They share food readily. They resolve inter-group conflict through affiliation rather than warfare — sometimes, famously, through sexual behavior. It’s a joke I make with friends: if bonobos had been cast in the Planet of the Apes franchise, the movies would never have made it to theaters. Recent work has complicated the older picture — Mouginot et al. (2024) found that male bonobos engage in more frequent low-level aggression with other males than chimpanzees do, and a 2026 paper documented the first lethal outcome of an inter-community encounter — but the pattern is still distinct. Bonobo aggression is overwhelmingly male-male within a community, rarely involves coalitions, and almost never produces the coordinated lethal raids on neighbors that define the chimpanzee pattern. Bonobos are not the cartoon peaceniks of the early literature. They are an ape whose aggression takes a different form, in a different social structure, in a different resource environment — and the difference is exactly the point of the comparison.

Portrait of an adult bonobo (Pan paniscus) at Apenheul Primate Park, showing the species' darker face, fuller lips, and slighter build relative to chimpanzees.
Pan paniscus — bonobo. Photo by Natataek, licensed CC BY-SA 3.0; source.

Both species are correctly adapted to the environments that produced them. Neither is morally superior. But humans, unlike either of our cousins, are configurable. We carry both possibilities. Whichever one expresses depends, in large part, on the signals our biology is receiving about the world we live in.

This site argues that we have been receiving the wrong signals for a long time, and that the cost of that has been higher than we have admitted. Less chimp, more bonobo is, in the most literal sense possible, a proposal.

How I actually use this

Before the science, a practical note.

Over the years I’ve started, half-consciously, to place the people and groups I encounter on a spectrum running from chimp to bonobo. The longer I do it, the more useful it has become — not as a diagnosis of anyone, but as a way of locating behavior.

Hot-tempered, quick to anger, loud, territorial about possessions, prone to dominance display, escalates conflict rather than dissolving it — that is chimp-leaning behavior. Not bad. Not pathological. Just chimp-configured.

Patient, slow to react, empathetic, nurturing, easy to share with, defuses conflict before it has fully formed — that is bonobo-leaning behavior. Not saintly. Not weak. Just bonobo-configured.

It is a spectrum, not a binary, and almost nobody sits at one fixed point on it. The same person is chimp-mode in one context and bonobo-mode in another. Whole countries shift across generations. The signals reaching us push us up and down the spectrum constantly, and most of the time we do not notice it happening.

I use it on individuals. On the teams I work with. On companies, on cities, on states, on countries. It is not hard science and it is not meant to be. It is a heuristic that has helped me, and the way it has helped me most is by making it harder to read every difficult interaction as personal. The person leaning on the horn behind me is not, as a rule, a moral failure. They are a primate in chimp configuration, in a city that has probably been chimp-configuring them for a long time. That reframe does not excuse the behavior. It just locates it.

The rest of this essay is an attempt to explain why people and groups end up where they do on the spectrum, and what makes them shift.

1. The Gap

Opening observation. We live in objectively the most resource-abundant period in human history. More calories, more shelter, more medicine, more information, more connection are available to more people than at any prior moment in the human story. By every measurable input we have ever cared about, the species has won.

And yet the lived experience of most people in the most abundant societies is one of chronic scarcity, depletion, threat, and inadequacy. We are overworked, behind, and afraid. The gap between objective abundance and subjective experience is not a small or marginal phenomenon. It is the dominant emotional reality of modern life in developed nations.

Why this gap is the question worth asking. Most analyses of modern dissatisfaction reach for cultural, economic, technological, or political explanations. These analyses are not wrong, but they are downstream. The gap exists because we are biological organisms running on a stress-response system that did not evolve to measure objective conditions — it evolved to respond to signals about conditions. And the signals we have built around ourselves are, by an evolutionary measure, indistinguishable from the signals of severe scarcity.

The thesis stated plainly. This essay argues that the gap between abundance and felt scarcity is biological, that the mechanism is the interaction between cortisol and testosterone in response to chronic environmental signaling, that the configurability of this system is visible in our closest primate relatives, and that the implications for how we organize society are larger than any current public conversation acknowledges.

2. The Two Cousins

Setup. Humans share over 98% of our DNA with two living primate species: chimpanzees and bonobos. These two species diverged from a common ancestor between 1.5 and 2 million years ago. They look similar, they behave in similar ways across many dimensions, and they share most of their cognitive architecture. But they organize their social worlds in opposite directions.

Chimpanzee social structure. Male-dominated hierarchies, territorial aggression, lethal violence against neighboring communities, food sharing only under specific conditions, hunting as a status activity. The behavioral pattern is one of competitive engagement with limited resources.

Bonobo social structure. Female-led affiliation, sexual behavior as conflict resolution, food sharing as default, no documented lethal aggression between neighboring groups until a single recent case (Sakamaki et al., 2026), and an aggression profile that — as Mouginot et al. (2024) documented in Current Biology — involves more frequent male-male within-group conflict than older literature suggested, but lacks the coordinated coalitionary violence that defines the chimpanzee pattern. The behavioral signature is one of cooperative engagement with shared resources, distinct from but not free of aggression.

The neuroendocrine evidence. The Wobber et al. (2010) PNAS finding is the load-bearing citation here. When age-matched male bonobos and chimpanzees are presented with the same dyadic food competition, they show different neuroendocrine responses. Bonobos show an anticipatory rise in cortisol — a stress-mediated avoidance response. Chimps show an anticipatory rise in testosterone — an aggression-mediated engagement response. Same challenge. Two completely different evolved strategies.

The evolutionary explanation, as currently debated. The Hare, Wobber, and Wrangham (2012) self-domestication hypothesis is the most-developed framework for the divergence. Bonobos evolved south of the Congo River in an environment with abundant terrestrial herb resources and without competition from gorillas. Chimps evolved north of the Congo in fruit-patchy, gorilla-competitive environments. The selection pressures were different, and over evolutionary time produced different neuroendocrine architectures. The hypothesis remains contested — Frans de Waal has been a long-standing critic, and the 2024 Mouginot finding of higher within-group male aggression in bonobos has prompted renewed debate — but the basic claim that different environmental pressures produced different social and endocrine profiles is well supported even where the specific mechanism remains under active discussion. Neither species is “better.” Both are correctly adapted to the environments that shaped them.

The point of the comparison. The two species prove that the same underlying primate biology — the HPA axis, the HPG axis, the basic emotional and social wiring — can produce dramatically different social configurations under different environmental conditions. Configurability is a property of great-ape biology, not a unique human invention.

Citations to develop: Wobber et al. (2010), Hare et al. (2012), de Waal (1997), Stimpson et al. (2016).

3. The Configurable Primate

Setup. Humans inherited the great-ape neuroendocrine architecture, but with a critical modification: we are far more plastic than either chimps or bonobos. Within a single human lifetime — and to some degree within developmental windows that close as we age — our stress response system reconfigures based on environmental input.

The HPA axis as the central regulator. Brief explanation of how the hypothalamic-pituitary-adrenal axis produces cortisol in response to perceived threat or scarcity, and how chronic activation produces dysregulation rather than simply elevated levels. The Dowd (2009) review showing that the SES-cortisol relationship is more complicated than intuition suggests.

The neural circuit. Cortisol does not act in a vacuum. The same stress response the HPA axis coordinates also reshapes which brain regions are in charge of behavior. Arnsten’s synthesis of the prefrontal cortex literature (2015, Nature Neuroscience) documents how chronic stress signaling progressively impairs PFC function — the deliberative, top-down region that would normally regulate the amygdala — and hands control to the amygdala and dorsal striatum, the reactive, habitual circuits. Prolonged cortisol elevation produces measurable structural changes in the hippocampus, including dendritic remodeling and, in sustained cases, volume reductions (Sapolsky, 1996; McEwen & Gianaros, 2010), weakening one of the brain’s primary mechanisms for contextualizing whether a signal actually represents threat. Liston, McEwen & Casey (2009, PNAS) demonstrated this experimentally in humans: chronic stress produces dendritic remodeling in the PFC that is reversible when stress is reduced. The shift is structural, not metaphorical — and crucially, in young adults at least, it is plastic.

Signaling below awareness. Most of this happens before conscious awareness arrives. The thalamic shortcut LeDoux (2003) mapped lets the amygdala detect threat in roughly twelve milliseconds, well before cortical processing produces a conscious interpretation of the same stimulus. Whalen et al. (1998, Journal of Neuroscience) showed that fearful faces presented below the threshold of conscious detection still activate the amygdala reliably. Critchley (2009) documented that interoceptive signals shape mood and decision-making without entering awareness at all. By the time the deliberative brain notices there is a question to ask, the cortisol cascade has already begun. This becomes the foundation of the limits-of-awareness argument developed in Section 7.5.

Developmental programming. Early-life environments shape the HPA axis through epigenetic mechanisms. Murgatroyd and Spengler (2011) on the epigenetic programming literature. The Royal Society 2019 theme issue on evolutionary medicine framing this as adaptive plasticity.

Differential susceptibility. Belsky and Pluess (2009). Some individuals are genetically more responsive to environmental signals than others. The same scarcity signal will produce a larger stress response in some people than others, for reasons that are partly genetic and partly developmental. This explains both individual differences in resilience and the population-level variation in how a single environment produces different outcomes.

Why this matters. The configurability is not infinite. Humans are not blank slates onto which any social structure can be imposed. But within the biologically available range, the environment we build determines a great deal of what we collectively become. The question is not whether biology is destiny. The question is which configuration our biology has been signaled into, and whether the signals are accurate to our actual conditions.

Citations to develop: Dowd et al. (2009), Murgatroyd & Spengler (2011), Belsky & Pluess (2009), Royal Society 2019 theme issue, Geronimus et al. (2006), Arnsten (2015), Liston, McEwen & Casey (2009), Sapolsky (1996), LeDoux (2003), Whalen et al. (1998), Critchley (2009).

4. The Dual-Hormone Hypothesis

Setup. Cortisol alone does not explain the social patterns we observe. The full mechanism involves the interaction between cortisol and testosterone — what the published literature calls the dual-hormone hypothesis.

The hypothesis stated. Mehta and Josephs (2010), Hormones and Behavior. Testosterone drives dominance-seeking and aggressive behavior, but only when cortisol is configured in a permissive way. The HPA axis (cortisol) and the HPG axis (testosterone) interact rather than operating independently. The combination of low cortisol and high testosterone predicts aggressive dominance-seeking. The combination of high cortisol with high testosterone often blocks or reverses that effect.

The empirical evidence. The dual-hormone literature is real and has been replicated in studies of incarcerated populations (Dabbs et al.), delinquent adolescents (Popma et al.), psychiatric inpatient aggression (Zerroug et al., 2025), and laboratory studies of social competition (multiple replications). The effect sizes are modest — the Dekkers et al. (2018) meta-analysis found marginal support overall — but the qualitative pattern is robust enough that the framework is taken seriously by researchers studying aggression.

The contested points, addressed honestly. The effect sizes are smaller than the popular framing suggests. Some studies fail to replicate. The hypothesis is a useful framework, not a settled law. The essay should engage these limitations directly rather than overclaim.

Testosterone is a status amplifier, not an aggression hormone. The popular framing of testosterone as a “violence hormone” obscures a finding that has been well replicated across the past three decades and that sharpens the dual-hormone argument considerably. Mazur and Booth (1998, Behavioral and Brain Sciences) first articulated, and Eisenegger, Haushofer and Fehr (2011, Trends in Cognitive Sciences) synthesized, the status-seeking hypothesis: testosterone motivates individuals to acquire and defend status within their reference group, but the form that status-seeking takes depends on what the group rewards. Eisenegger et al. (2010, Nature) demonstrated this directly — women given testosterone in an Ultimatum Game made fairer offers, not more aggressive ones, because in that context status accrued to those seen as fair. In groups where aggression earns status, testosterone amplifies aggression. In groups where generosity earns status, testosterone amplifies generosity. Once you see testosterone this way, the dual-hormone pattern becomes clearer: a chronic-scarcity-signaling environment produces a group whose status norms tilt toward dominance display and out-group hostility, and the hormone amplifies those behaviors because they are what the group has been signaled to reward. The same molecule in a bonobo-mode environment, where status comes from affiliation and food-sharing, amplifies those instead. The hormone is doing exactly what evolution designed it to do — making the individual a more effective version of whatever the group already is.

The 2024 bonobo data as confirmation, not contradiction. Mouginot et al. (2024) found that male bonobos engage in more frequent low-level aggression with other males than chimpanzees do, and that more aggressive bonobo males have greater mating success. This result surprised researchers who had absorbed the popular “peaceful bonobo” framing, and it has been described in some quarters as a problem for the self-domestication hypothesis. Within the status-amplifier framework, it is exactly what one would predict. Male bonobos compete for mating opportunities in a social structure where coalitions are weaker and lethal coordination is absent, so individual male-male aggression remains the route to status — and the hormone amplifies it accordingly. What is missing in the bonobo case is not aggression itself but the coalitionary, lethal, out-group-directed pattern that emerges when males of a chronic-scarcity-signaled species form dominance hierarchies around contested resources. The species comparison is not chimps-violent / bonobos-peaceful. It is two species in which testosterone amplifies what status looks like in their respective social ecologies, and the contrast is in the form of aggression, not its presence.

Why this matters for the broader argument. Once you accept that cortisol and testosterone interact under chronic stress signaling, the chimp pattern stops being a metaphor and becomes a mechanism. The behavioral architecture of violent organized groups — hierarchies, dominance displays, in-group/out-group dehumanization, willingness to die for the group — maps onto what we would predict from a population of male primates whose dual-hormone systems have been chronically configured by unpredictable scarcity signaling.

The honest framing. This is not a claim that hormones determine behavior. It is a claim that hormones are a substantial input into the probability distribution of behaviors a group is likely to produce, and that this input has been systematically underestimated in conversations about social pathology.

Citations to develop: Mehta & Josephs (2010), Dekkers et al. (2018), Dabbs et al., Popma et al., Zerroug et al. (2025), Mazur & Booth (1998), Eisenegger et al. (2010), Eisenegger, Haushofer & Fehr (2011), Boksem et al. (2013), Diekhof et al. (2014), Carré & Olmstead (2015).

4.5 Individual Variability, the Gene-Environment Interaction, and the Sorting Dynamic

The question this section answers. If the signaling environment configures the population, why do two people raised in the same environment so often end up in different places? Why do siblings from the same household, neighbors on the same block, soldiers from the same unit produce such different behavioral outcomes? The configurability argument explains population-scale distributions; it does not, on its own, explain individual variance within those distributions. The honest answer is that individuals differ in how their biology processes environmental signals, and a substantial fraction of that difference is genetic.

The general finding. Twin and adoption studies, summarized in Rhee & Waldman’s (2002, Psychological Bulletin) meta-analysis of fifty-one studies covering nearly 100,000 individuals, place the heritability of antisocial behavior at approximately 0.40 to 0.50. Burt’s (2009, Clinical Psychology Review) review reaches the same range across more recent samples. The effect is consistent across study designs, sexes, and definitions of antisocial behavior, and it is larger for the most severe and persistent expressions. Aggression is not a culturally invented phenomenon imposed on a blank biological substrate. It has a measurable genetic component, and the component is substantial.

The best-replicated specific mechanism. Caspi et al.’s (2002, Science) study of the MAOA gene-by-environment interaction is, by some measures, the most-replicated finding in all of behavioral genetics. Male carriers of the low-activity variant of the MAOA gene who were maltreated in childhood were dramatically more likely to develop antisocial behavior in adulthood than either low-activity carriers who were not maltreated, or high-activity carriers who were. The interaction was confirmed by Kim-Cohen et al.’s (2006) meta-analysis and again by Byrd & Manuck’s (2014, Biological Psychiatry) meta-analysis of twenty-seven studies. Gallardo-Pujol et al. (2013, Genes, Brain and Behavior) reproduced the effect under randomized laboratory conditions. Newman et al. (2005, Biological Psychiatry) showed the same interaction in rhesus macaques carrying the orthologous variant, which means the mechanism is deep evolutionary, not recent-cultural.

What the interaction actually says. Carrying the low-activity MAOA variant alone does not predict antisocial behavior. Experiencing maltreatment alone does not strongly predict it either. The combination predicts it dramatically. The genotype calibrates how strongly the individual’s biology responds to environmental signaling about threat and resource availability. In a stable, well-resourced environment, the low-activity variant produces no observable behavioral signal. In an environment of chronic threat and instability, the same variant amplifies the response. Genes load the gun. Environment pulls the trigger. Without the environmental signal, the gun does not fire.

Heritability itself varies with environment. The most counterintuitive finding in this literature, and the one most relevant to the framework, is that the heritability of aggression is not a fixed number. Recent work from the European ACTION consortium (CORDIS-Aggression project) found heritability of childhood aggression ranging from approximately 0.20 in low-environmental-risk samples to approximately 0.60 in high-environmental-risk samples. The same genes express more strongly when the environment is more adverse. This is the configurability claim extended to the genetic level — not just that environment shapes behavior, but that environment determines how much genetic predispositions matter in the first place. Change the signal, and you change not only the distribution of outcomes but the slope of the line between genes and outcomes.

What this means for the framework, and what it explicitly does not mean. Individual genetic variation is real. Some individuals are more sensitive to scarcity signaling than others. This does not, however, license any of the following claims, all of which are sometimes attached to behavioral genetics in popular discourse: that some populations are genetically predisposed to violence (allele frequencies vary across populations but the G×E interaction means signaling environment determines expression, and population-level violence rates track environmental conditions rather than allele frequencies); that environmental intervention is futile for genetically predisposed individuals (the opposite is true — the G×E literature shows that removing the environmental signal prevents the outcome even in the most sensitive genotypes); or that the framework reduces to individual genetic destiny (population-scale distributions of behavior remain governed by population-scale signaling environments, and that is the leverage point this framework identifies).

The next move the literature makes: gene-environment correlation. The variation does not stay distributed evenly across the population. The same literature that establishes G×E also establishes a more sophisticated cousin called gene-environment correlation, or rGE (Plomin & Bergeman, 1991; Jaffee & Price, 2007; Knafo & Jaffee, 2013). rGE describes the documented tendency for individuals with certain genetic predispositions to end up disproportionately in environments that match those predispositions — not because the genes cause the environments, but because the genes shape the behavioral tendencies that influence environmental selection. Passive rGE: children inherit both genes and matching environments from their parents. Evocative rGE: genetically influenced behavior elicits matching responses from others. Active rGE, also called niche-picking: individuals seek out the social environments that fit the physiology they already have. The three forms compound, and the compounding is where the framework gets the rest of its reach.

The sorting dynamic in chimp-mode groups. A population exposed to chronic scarcity signaling does not produce chimp-mode behavior evenly distributed across its members. It produces a distribution in which the most physiologically responsive individuals — those whose genotypes most strongly amplify the scarcity signal — concentrate disproportionately in the groups that institutionalize chimp-mode behavior. The concentration is mutual. Predisposed individuals seek out groups that match their physiology because those groups feel right to them, providing the social structure their internal state is already running. The groups, in turn, select for and retain the most responsive members because those members are the most committed, the most willing to escalate, and the most stable in the group’s hierarchy. The behavioral-genetic literature has documented this dynamic most clearly in the gang-affiliation studies of Beaver, DeLisi, Vaughn and colleagues, where the same MAOA, DRD2, and DRD4 variants that predict individual antisocial behavior also predict gang affiliation and gang persistence beyond the effects of neighborhood environment alone. Dishion & Tipsord (2011, Annual Review of Psychology) describe the within-group amplification process under the heading of peer contagion and deviancy training: once predisposed individuals are co-located in a group context, the group amplifies what its members brought to it, producing behaviors more extreme than the signaling environment alone would predict.

What the sorting dynamic does and does not claim. The claim is not that chimp-mode groups are populated by genetically defective individuals. The claim is that within a signaling environment that activates chimp-mode behavior at all, the tail of the distribution of high-responders is more likely to occupy and remain in the groups that express the behavior most intensely. In a low-scarcity signaling environment, this tail is largely invisible at the behavioral level — the same individuals show no elevated antisocial behavior, because the predisposition has nothing to amplify. The MAOA literature is unusually clean on this point: low-activity carriers who were not maltreated as children show essentially the same behavioral profile as high-activity carriers. The genetic difference becomes a behavioral difference only in the presence of the environmental signal. Reduce the signal, and the sorting dynamic loses its input. The groups dissolve not because their members were rehabilitated individually, but because the conditions that selected for their concentration no longer exist.

The framework reaches farther with this layer in it. The earlier sections of the essay describe what happens to a population under chronic scarcity signaling. This section describes how the population sorts itself once the signaling is in place. The two together explain a pattern most discussions of social pathology cannot account for: why the groups that emerge under scarcity are not merely larger versions of the surrounding population, but qualitatively more intense versions of it. They are not random samples. They are the high-responder tail of the distribution, concentrated by mutual selection, amplified by within-group dynamics, and held in place by the environmental signaling that activated them in the first place. The upstream variable remains the signal. The downstream consequence is groups whose behavior cannot be addressed individual-by-individual, because the individuals are not the variable.

Why this strengthens the framework rather than complicates it. A common objection to environment-focused arguments about social behavior is some version of “but people are different, and some people will be violent in any environment.” The honest answer is: yes, individuals vary, and the variation has a real genetic component. But the same literature that establishes the genetic component also establishes that the genetic component is expressed through environmental signaling, and that individuals with the most reactive genotypes concentrate into groups that institutionalize the reaction. Without the signal, the predisposition stays latent and the sorting has no input. The framework’s claim is not that all individuals respond identically to the signaling environment. The claim is that the signaling environment shifts the distribution of outcomes, sorts the most responsive individuals into the groups that express the outcome most intensely, and produces the population-level patterns the rest of the essay describes. The individual-variability layer and the sorting layer make this claim stronger, not weaker, because together they identify the mechanism by which environmental change produces outcome change: not by overwriting individual biology, but by removing the input that activates it and dissolving the conditions that concentrated the responders.

Citations to develop: Rhee & Waldman (2002), Burt (2009), Caspi et al. (2002), Kim-Cohen et al. (2006), Byrd & Manuck (2014), Gallardo-Pujol et al. (2013), Newman et al. (2005), Ficks & Waldman (2014), CORDIS-Aggression / ACTION consortium publications, Plomin & Bergeman (1991), Jaffee & Price (2007), Knafo & Jaffee (2013), Burt, McGue & Iacono (2009), Beaver, DeLisi & Vaughn (multiple), Dishion & Tipsord (2011).

5. The Pattern in the Field

Setup. The most personally consequential observation of my federal career was that the structural architecture of violent extremist organizations did not depend on the ideology they had attached to themselves.

The observation, stated carefully. Religious extremism, secular ethnonationalism, gang structures, and organized criminal hierarchies share a behavioral skeleton: in-group cohesion, out-group dehumanization, ritual hierarchy, willingness to use violence on behalf of the group, and a narrative that justifies all of it. The ideological dressing varies. The skeleton does not.

The conclusion this forces. If the skeleton is the same across populations that share no ideology, no theology, no ethnicity, and no cultural history, then the variable producing the skeleton cannot be the ideology. It has to be upstream. The dual-hormone framework, combined with the chronic-scarcity-signaling environment that produces it, is a candidate for what is actually upstream.

The careful framing on religion. Religion is real. Religious belief is meaningful and motivating to billions of people, most of whom are not violent. The variable that distinguishes a peaceful religious community from a violent extremist one is rarely the religion itself. The same traditions produce dramatically different outcomes in different resource-signaling environments. What changes is not the belief. What changes is the biological state the community is operating in, and the primate response pattern that activates.

The implications for counterterrorism, briefly. Most counterterrorism intervention targets the ideological surface — the recruitment narratives, the radicalization pipelines, the messaging. If the framework here is correct, these interventions operate too far downstream to be efficient. The upstream intervention is the resource-signaling environment that makes communities susceptible to chimp-mode configuration in the first place. This is not an argument for abandoning ideological intervention. It is an argument for adding an upstream variable that has been systematically ignored.

The anchor. Twelve years in counterterrorism, leadership of the cyberterrorism unit, observation of the structural parallel across populations that should have nothing in common. Stated as observation, not theory.

5.5 Homophily, Group Stability, and the Stickiness of Configuration

The question this section answers. Sections 4 and 4.5 explain how the signaling environment configures individual physiology and how the most responsive individuals sort into chimp-mode groups. Section 5 describes the structural pattern those groups exhibit across ostensibly unrelated populations. What remains unexplained is why the groups, once formed, are so durable — why they persist through changes in membership, why they reproduce themselves across generations, and why simply reducing the environmental signal often fails to disperse them on any reasonable timescale. The missing layer is the social-psychology phenomenon of homophily, and the bidirectional stickiness of physiological configuration once it has been established.

Homophily, briefly. McPherson, Smith-Lovin & Cook’s (2001, Annual Review of Sociology) canonical synthesis documents that human social ties form preferentially between individuals who resemble each other across many dimensions — demographic, attitudinal, behavioral, and increasingly, what we would now describe as physiological-state dimensions. The mechanism is not primarily conscious choice. It is unconscious assessment: humans process subtle cues about each other’s arousal state, threat sensitivity, and stress-response posture through fast, pre-conscious neural pathways (the amygdala, insula, and orbitofrontal cortex are well-documented in this role — LeDoux 2003; Whalen et al. 1998; Critchley 2009). Within seconds of entering a room, a person has implicitly assessed who else in that room is operating in a compatible state, and social ties form accordingly. The matching happens below conscious awareness, and the resulting clusters are remarkably stable.

State, not type. The framing matters here, and the framework requires precision. Homophily does not sort people by genotype. It sorts them by current state. Two high-responder individuals in a low-scarcity environment cluster with low-responders because all of them are in roughly the same low-arousal state — the genetic predisposition has nothing to amplify, so the behavioral signal is absent and homophily reads them as compatible with everyone else. The same two individuals in a high-scarcity environment cluster preferentially with other high-responders because their states have diverged from the low-responders, whose physiology has not been activated as strongly. The matching is real but it is not a property of the individual. It is a property of the configuration the individual is currently running, which is itself a product of the signaling environment plus the genetic sensitivity to that signal.

Why groups stay together once formed. Once a chimp-mode group has formed through the sorting-and-homophily mechanism, it reproduces its own state through interaction. Members’ physiological states reinforce each other through the social signaling that occurs within the group — shared threat-vigilance, coordinated arousal during in-group rituals, the calming effect of being among others in a similar state. Dishion & Tipsord (2011, Annual Review of Psychology) documented this phenomenon under the heading of peer contagion: within-group interaction amplifies whatever state the group brought to it, producing behaviors more extreme than the individual members would have produced alone. The group becomes a homeostatic system for its own configuration. Members who drift out of state are pulled back in. Members who cannot be pulled back are expelled. This is why simply removing the original environmental signal does not, on any short timescale, disperse the group — the group has become its own signaling environment.

The bidirectional stickiness. The configurability claim of the broader framework is true at the population scale and over long timescales. At the individual scale and over short timescales, configuration is sticky in both directions. A person who has spent years in chimp-mode arousal does not instantly become comfortable in a low-signal environment when the signal is removed. The physiology has built itself around the signal, and its absence produces its own dysregulation — what the trauma literature calls hyperarousal mismatch, what veterans describe as the difficulty of civilian life after combat, and what clinicians observe in people leaving high-control religious groups or extremist organizations. The same stickiness operates in reverse: a low-responder who is placed in a chronically high-arousal environment for long enough often shifts toward higher baseline arousal themselves, and may then have difficulty returning to the lower-arousal baseline even after the environment changes. The configuration follows the environment, but it does so with a substantial lag. Lupien et al. (2009, Nature Reviews Neuroscience) documented the persistence of early-life HPA-axis programming and the difficulty of overriding established setpoints through later environmental change alone.

What this means for the framework. Two things, one strengthening and one tempering. The strengthening: the framework now reaches deeper into explaining real-world group persistence. Chimp-mode groups are not held in place by the environmental signal alone — they are held in place by the environmental signal that produced them and the homophily-stabilized internal social dynamics that reproduce their state once formed. This explains why such groups can outlast the conditions that created them, why they reproduce across generations, and why interventions targeting the environmental signal alone often fail to produce visible group-level change on policy-relevant timescales. The tempering: the framework should not promise quick population-level change from environmental intervention alone. The signaling environment is the upstream lever, but the downstream groups have their own inertia, and the timeline of dispersal is not the same as the timeline of signal reduction.

The implication for intervention design. Environmental signal reduction is necessary but not sufficient at the group level. Where chimp-mode groups have formed and stabilized themselves, intervention has to address both the upstream signal and the within-group homeostasis — not by removing the individuals (the sorting dynamic will reproduce the group from the next cohort if the environmental signal remains) and not by changing minds within the group (cognitive intervention operates downstream of the physiological state that holds the group together), but by some combination of signal reduction at the population scale and physiological resetting at the individual scale, sustained over a timescale that respects the stickiness of established configuration. This is more demanding than “just change the environment.” It is also more honest about what change at this scale actually requires.

Citations to develop: McPherson, Smith-Lovin & Cook (2001), Dishion & Tipsord (2011), Lupien et al. (2009), LeDoux (2003), Whalen et al. (1998), Critchley (2009), Centola & Macy (2007) on complex contagion, Christakis & Fowler (2007, 2008) on social network dynamics.

6. The Fabricated Signal

Setup. Modern industrial economies do something no prior human society did at scale: they generate continuous fabricated scarcity signals in environments of objective abundance.

Examples. Financial anxiety in conditions of historically high real income. Status competition mediated through algorithmic social comparison. Time poverty in a world that has automated away most physical labor. Housing scarcity in markets that have functionally adequate supply but unaffordable pricing. Information overload that produces a chronic low-grade threat-monitoring state. The list is long.

Why these signals matter. The HPA-HPG response system cannot distinguish fabricated scarcity from real scarcity. The cortisol response to financial anxiety is biochemically identical to the cortisol response to genuine food scarcity. We are running on chemistry that does not know the difference between a hostile environment and a notification.

The cumulative effect. Allostatic load. McEwen’s framework. Geronimus on weathering. The cumulative biological cost of running a stress-response system in a chronically activated state, regardless of whether the activation is warranted. The cost includes cardiovascular disease, immune dysfunction, cognitive impairment, depression, and the social configurations that follow from chronic dual-hormone dysregulation at population scale.

The unsettling implication. A society can be objectively wealthy and physiologically poor at the same time. By every external measure, it is winning. By every internal measure, it is at war.

Citations to develop: McEwen on allostatic load, Geronimus et al. (2006), the broader chronic-stress-and-health literature.

7. What This Does Not Claim

Setup. A careful argument requires acknowledging its limits. This section preempts the strongest objections and clarifies what the framework is and is not asserting.

What it does not claim. That hormones determine behavior. That testosterone causes aggression directly — the hormone amplifies whatever the group has been signaled to reward, not aggression as such. That all violence is reducible to neuroendocrine state. That the historical patterns of gendered leadership are biologically optimal or normative. That capitalism or any specific economic system is responsible for the fabricated-scarcity problem. That this framework explains everything about human social organization. That the chimp configuration is universally bad or the bonobo configuration is universally good — both are correctly adapted to specific environments.

What it does claim. That chronic, unpredictable scarcity signaling produces measurable physiological dysregulation, including dual-hormone configurations associated with aggressive dominance-seeking. That this dysregulation shapes social behavior at both individual and population scale. That the configurability visible in our closest primate relatives demonstrates the range of social patterns biologically available within great-ape architecture. That humans are uniquely plastic within this range. That the environment we build is the dominant input into which configuration our biology expresses. That we have been systematically underestimating this variable in conversations about social pathology, and that taking it seriously would change what we consider promising interventions.

7.5 The Limits of Awareness

Setup. If most stress signaling occurs below conscious awareness, and if chronic signaling progressively shifts decision-making out of the prefrontal cortex and into more reactive circuits, then awareness is necessary but not sufficient. This section makes the claim explicit.

The mechanism mismatch. Cognitive insight operates in the deliberative brain. The dysregulation operates in a different system entirely — older, faster, and structurally remodeled by years of chronic input. Knowing intellectually that one’s stress response is disproportionate to objective conditions does not, on its own, unsubscribe the body from the cascade. The system that would need to reconfigure does not speak the language of conscious intent.

The empirical picture. Lupien, McEwen, Gunnar & Heim (2009, Nature Reviews Neuroscience) document the persistence of early-life HPA programming: setpoints established in development are sticky and difficult to override through later cognitive intervention. Cole’s social genomics work on the Conserved Transcriptional Response to Adversity (CTRA) shows that the inflammatory gene-expression pattern responsive to subjective social threat tracks perceived conditions rather than actual ones, and can persist after objective circumstances improve. The broader trauma and chronic-stress treatment literature documents a consistent clinical observation: cognitive understanding does not, on its own, reset the physiological imprint. The therapies with the most consistent effects on chronic-stress physiology — trauma-focused cognitive-behavioral therapy, somatic and exposure-based modalities, sustained contemplative practice — all involve repeated behavioral or somatic engagement, not pure insight.

The honest caveats. Cognitive reappraisal does reduce acute sympathetic activation (Gross, 1998). Mindset interventions can modify the stress response within sustained framings (Crum, Salovey & Achor, 2013, JPSP). Cognitive-behavioral therapy has robust effects on symptom report. None of this contradicts the central claim. It clarifies it: awareness is useful, especially as a doorway. It is not, by itself, an intervention sufficient to reconfigure a system that operates faster than consciousness and recovers more slowly than thought.

Why this matters for what follows. Most current public conversation about chronic stress prescribes individual cognitive remedies — meditation apps, gratitude journals, mindset work. These prescriptions are not wrong, but they are downstream of where the dysregulation actually lives. The intervention with the largest leverage is environmental: changing the signal itself rather than the response to it. That is the argument the remainder of the essay develops.

Citations to develop: Lupien et al. (2009), Cole (2013, 2019), Gross (1998), Crum, Salovey & Achor (2013), Killingsworth & Gilbert (2010), Mani et al. (2013, with replication caveats).

8. What Would Change

Setup. If the framework is even partially correct, what follows?

At the individual level. Awareness is the doorway, not the remedy — for the reasons developed in Section 7.5. What individuals can usefully do is recognize that the felt experience of scarcity is a signal-driven cascade rather than identity, sustain whatever behavioral practices reliably soften the autonomic baseline, and locate themselves accurately in the larger system that is producing the signal in the first place. Most of the leverage is not here.

At the community level. Auditing the signal environment of institutions, neighborhoods, and workplaces. Identifying which features are producing dual-hormone responses that are out of proportion to actual conditions. Specific applied cases to be sketched in subsequent drafts.

At the civic level. A different conversation about urban design, public safety, education, and economic policy that takes the biological substrate seriously. Not as the only variable, but as a variable.

The honest closing. No single intervention changes the population-scale signal environment. This is a multigenerational project. The first step is making the variable visible. The second step is asking what each of us, in our specific position, could do to dampen the chronic-scarcity signaling for the people we have responsibility for. The bonobos at the Jacksonville Zoo are a small invitation to remember what a primate society organized around abundance and cooperation actually looks like. They are not a prescription. They are a proof of possibility.

8.5 Back to the spectrum

The framework above gives you a mechanism. The simpler reading I opened with — chimp behavior, bonobo behavior, a spectrum, signals that shift people along it — is the daily-use version of that mechanism. They are the same thing said at two different scales.

What I notice, now that the science has been laid out, is that the spectrum sharpens. It is no longer just a heuristic. The hot-tempered colleague, the territorial neighbor, the country that pivots into ethnonationalism — these are not separate phenomena loosely metaphorized as chimp behavior. They are the chimp configuration expressing, in different settings, through the same underlying dual-hormone-and-signal architecture. What looked like a soft analogy turns out to be a hard mechanism.

That sharpens the question of what to do.

The honest answer, for any of us, is small. We notice the signals reaching us. We notice the signals we send. We reduce, where we can, the chimp-configuring inputs we are responsible for — in our households, in the institutions we run, in the conversations we have with strangers. We do not transform the population-scale signal environment by ourselves. We just move our small piece of it a little, and stay alert to where on the spectrum the people we care about are sitting today.

Less chimp, more bonobo is not a prescription anyone hands out to anyone else. It is a direction. The work is to move yourself, and your part of the world, a little further along it.

9. A Note on the Author’s Position

I am not a scientist. I am a former federal counterterrorism investigator who became convinced the field I had trained in was looking at the wrong variable, and who spent a decade reading the science to understand what I had been observing. What this essay offers is a synthesis of existing literature filtered through field experience, not original research. The component findings are peer-reviewed; the synthesis is mine. In honest terms, what is presented here is a framework — an integrative model that draws together established hypotheses (the dual-hormone hypothesis, the self-domestication hypothesis, allostatic load, the CTRA, the social neuroendocrinology of status) into a coherent account of why objective abundance can produce felt scarcity at population scale. It is not yet a theory in the strict scientific sense, because the synthesis itself has not been independently tested. It generates testable predictions, and I offer it for the conversation it might start, not as settled science. The framework is open to revision. The citations are provided so the reader can do the same work I did, and reach their own conclusions.

9.5 The Limits of This Framework, and What Would Change My Mind

The kind of science this is. Before listing the limits, the kind of claim being made has to be named. This framework operates under what philosophers of science call historical-science epistemology — the same epistemic standard that governs plate tectonics, paleontology, evolutionary biology, cosmology, and large parts of climate science. In these fields, the events being explained are not available for controlled experimental manipulation. You cannot run a randomized trial on the divergence of bonobos and chimpanzees two million years ago. You cannot replicate the Pleistocene environment that shaped human stress physiology. The data is fragmentary, the causal chains are long, and the gold standard of contemporary lab science — falsification through experimental manipulation — is structurally unavailable. This is a real epistemic constraint, and it applies to any framework that reasons from deep evolutionary or environmental history to present-day behavior, including this one.

What that does and does not mean. What it does not mean is that historical sciences are unscientific. Plate tectonics was not proved by experiment; it was confirmed by the convergence of seismic data, paleomagnetic records, fossil distribution patterns, ocean-floor age mapping, and eventually direct GPS measurement of continental drift. Each line of evidence alone was suggestive. Together they were overwhelming. The standard for historical science is convergence of independent evidence across multiple methodologies, not falsification through experimental manipulation. We should not avoid synthesis just because a single direct cause-and-effect chain is unavailable. The synthesis itself, if it integrates enough independent lines of evidence, is the contribution. What it does mean is that frameworks of this kind are inherently provisional, that their strength scales with the number of independent lines they integrate, and that they remain vulnerable to revision as any one of those lines develops.

The specific limits of the framework as it stands. Honesty requires naming each load-bearing claim and the extent to which the supporting literature can carry it.

The dual-hormone hypothesis. The most recent meta-analysis (Dekkers et al., 2019) found a statistically significant but very small interaction effect (r ≈ −0.06), with evidence of publication bias, analytic flexibility, and underpowered individual studies. Replication attempts have produced mixed results, and at least some studies report reversed-direction findings. The hypothesis is not refuted, but it is not "well-established" in the sense the essay sometimes implies. The honest position is that dual-hormone is one mechanism among several plausible mechanisms by which stress physiology might shape status-relevant behavior, that the effect is real but small at the population level, and that the framework would not collapse if dual-hormone were replaced by an alternative mechanism — the configurability claim is independent of the specific molecular pathway.

Candidate-gene-by-environment interaction. The MAOA × maltreatment finding (Caspi et al., 2002) lives within a research paradigm that has been substantially discredited since 2019. Border, Johnson et al. (2019, American Journal of Psychiatry) examined eighteen historical candidate genes for depression across multiple large samples and found essentially no support for the original associations or for candidate-gene-by-environment interactions. The candidate-gene approach has been largely retired in favor of genome-wide association studies. The MAOA case is one of the better-supported exceptions within that retired paradigm, with two confirmatory meta-analyses (Kim-Cohen et al., 2006; Byrd & Manuck, 2014) and rhesus macaque replication (Newman et al., 2005), but it remains an exception in a literature whose other findings mostly have not held up. The honest position is that the MAOA × maltreatment interaction is plausibly real but should not be presented as bulletproof, and the framework should lean more heavily on the more robust twin-and-adoption heritability estimates (Rhee & Waldman, 2002) than on any single candidate-gene finding.

The primate comparison. The chimp-bonobo contrast is rhetorical and pedagogical scaffolding for the framework, not its load-bearing science. The dual-hormone mechanism, the allostatic-load literature, the homophily literature, and the gene-environment interaction findings are independently established in humans and do not require the primate comparison to function. The 2024 Mouginot finding has complicated the contrast, the self-domestication hypothesis remains contested, and the genetic-distance-to-behavioral-similarity inference is a leap the framework cannot fully justify. The primate comparison is useful for communicating the configurability claim. It is not what the framework rests on.

"Fabricated scarcity signaling." This concept is doing significant conceptual work in the framework without being rigorously operationalized. What counts as a fabricated scarcity signal? Income inequality, status anxiety, social media use, advertising saturation, housing-debt structures, job precarity, news-cycle exposure — each of these has its own literature and its own measurement challenges, and the framework treats them as roughly interchangeable inputs to a single biological response system. A more rigorous version of the framework would need a dose-response curve: if signal X is reduced by Y percent, what's the predicted reduction in chimp-mode behavior at population scale? The framework as written cannot answer that question, and the policy implications in Section 8 are correspondingly soft until the operationalization improves.

The prior literature. Several of the framework's component moves have been made by named figures whose work this essay borrows from without always making the relationship explicit. Robert Sapolsky has spent his career arguing for the centrality of social-hierarchy stress in producing behavioral pathology; his synthesis in Behave overlaps substantially with what this essay attempts. Mullainathan & Shafir's Scarcity covers the cognitive-scarcity case. Evolutionary mismatch theorists (Lieberman, Eaton et al., others) cover the modern-environment-versus-evolved-physiology case. The framework here arrived at its synthesis through field experience rather than literature review, which is part of why these positions developed somewhat independently, but the synthesis sits inside a larger conversation, and a reader familiar with that conversation should know where the synthesis is original and where it converges with arguments already in print. The originality is in the cross-disciplinary assembly applied to counterterrorism-domain observation, not in the individual components.

What would change my mind. A framework that explains everything explains nothing, and a defensible version of this work has to specify what evidence would, in principle, refute it. Under historical-science epistemology, the falsification standard is not "show me a controlled experiment" but "show me an empirical pattern that the framework predicts should not exist." The patterns that would meaningfully shift my position on this framework include:

Population-level signal reduction producing no measurable change. If a population's fabricated-scarcity-signal environment were reduced substantially — through housing-cost intervention, income stabilization, reduction of social-comparison exposure, or some combination — and observable measures of chimp-mode social configuration (rates of organized violence, dominance-hierarchical group formation, in-group/out-group hostility) did not change over a generational timescale, the framework's central claim would be in serious trouble. The stickiness layer (Section 5.5) gives the framework some time to produce visible change, but it does not give it infinite time. If signal reduction reliably produces no downstream change, the framework is wrong about what the upstream variable is.

Cross-cultural data showing no environment-to-configuration relationship. If careful cross-cultural comparison showed that societies with very different scarcity-signal environments produced statistically indistinguishable rates of organized violence, dominance hierarchy, and in-group/out-group dynamics, the framework would have to be substantially revised. The current evidence (Wilkinson & Pickett on inequality, the CTRA literature, the cross-cultural epidemiology of violence) goes the other direction, but this is the kind of finding that would matter.

Specific neuroendocrine refutation. If the dual-hormone hypothesis, the allostatic-load model, and the CTRA framework were all to fail replication in large pre-registered studies, the molecular mechanism the framework claims would be in question. The framework could potentially survive by replacing the specific mechanism with a different one, but its current expression would have to be substantially rewritten.

Twin-study or natural-experiment evidence against configurability. If twin studies showed that monozygotic twins raised in radically different scarcity-signal environments produced indistinguishable behavioral configurations, the configurability claim itself would be in serious trouble. The current twin-study literature (heritability of aggression ranging from ~0.20 in low-risk environments to ~0.60 in high-risk environments) goes the other direction, but the configurability claim depends on this differential, and it has to be defensible.

The honest closing. This framework is offered as a synthesis under historical-science epistemology, with explicitly stated limits, explicitly stated falsification conditions, and explicit acknowledgment of the prior literature it draws from and overlaps with. It is meant to start a conversation, not to end one. The strongest version of the work is the version that knows what could be wrong with it. The reader who finishes this section knowing the limits is better equipped to evaluate the framework than the reader who only sees its strengths — and that asymmetry, I think, is the right one to leave the reader with.

Citations to develop: Dekkers et al. (2019) on dual-hormone meta-analysis, Border et al. (2019) on candidate-gene replication failure, Duncan & Keller (2011) on cG×E methodology, Sapolsky (Behave, 2017), Mullainathan & Shafir (Scarcity, 2013), Lieberman (The Story of the Human Body, 2013), Wilkinson & Pickett (The Spirit Level, 2009; The Inner Level, 2018).

Citations

Load-bearing references for the argument above. The full bibliography will be developed across drafting sessions.

Primary

  • Wobber, V., Hare, B., Maboto, J., Lipson, S., Wrangham, R., & Ellison, P. T. (2010). Differential changes in steroid hormones before competition in bonobos and chimpanzees. PNAS, 107(28), 12457–12462. https://doi.org/10.1073/pnas.1007411107
  • Hare, B., Wobber, V., & Wrangham, R. (2012). The self-domestication hypothesis: Evolution of bonobo psychology is due to selection against aggression. Animal Behaviour, 83(3), 573–585. https://doi.org/10.1016/j.anbehav.2011.12.007
  • Mehta, P. H., & Josephs, R. A. (2010). Testosterone and cortisol jointly regulate dominance: Evidence for a dual-hormone hypothesis. Hormones and Behavior, 58(5), 898–906. https://doi.org/10.1016/j.yhbeh.2010.08.020
  • Dekkers, T. J., Agelink van Rentergem, J. A., Meijer, B., Popma, A., Wagemaker, E., & Huizenga, H. M. (2018). A meta-analytical evaluation of the dual-hormone hypothesis: Does cortisol moderate the relationship between testosterone and status, dominance, risk taking, aggression, and psychopathy? Neuroscience & Biobehavioral Reviews, 96, 250–271. https://doi.org/10.1016/j.neubiorev.2018.12.004
  • Sapolsky, R. M. (2005). The influence of social hierarchy on primate health. Science, 308(5722), 648–652. https://doi.org/10.1126/science.1106477
  • Dowd, J. B., Simanek, A. M., & Aiello, A. E. (2009). Socio-economic status, cortisol and allostatic load: a review of the literature. International Journal of Epidemiology, 38(5), 1297–1309. https://doi.org/10.1093/ije/dyp277

Supporting

  • Murgatroyd, C., & Spengler, D. (2011). Epigenetic programming of the HPA axis: Early life decides. Stress, 14(6), 581–589. https://doi.org/10.3109/10253890.2011.602146
  • Belsky, J., & Pluess, M. (2009). Beyond diathesis stress: Differential susceptibility to environmental influences. Psychological Bulletin, 135(6), 885–908. https://doi.org/10.1037/a0017376
  • Geronimus, A. T., Hicken, M., Keene, D., & Bound, J. (2006). “Weathering” and age patterns of allostatic load scores among Blacks and Whites in the United States. American Journal of Public Health, 96(5), 826–833. https://doi.org/10.2105/AJPH.2004.060749
  • McEwen, B. S. (1998). Stress, adaptation, and disease: Allostasis and allostatic load. Annals of the New York Academy of Sciences, 840(1), 33–44. https://doi.org/10.1111/j.1749-6632.1998.tb09546.x
  • Stimpson, C. D., et al. (2016). Differential serotonergic innervation of the amygdala in bonobos and chimpanzees. Social Cognitive and Affective Neuroscience, 11(3), 413–422. https://doi.org/10.1093/scan/nsv128
  • Mouginot, M., Wilson, M. L., Desai, N., & Surbeck, M. (2024). Differences in expression of male aggression between wild bonobos and chimpanzees. Current Biology, 34(8), 1780–1785.e4. https://doi.org/10.1016/j.cub.2024.02.071
  • Sakamaki, T., et al. (2026). A lethal incident during an intergroup encounter in bonobos. Scientific Reports. https://www.nature.com/articles/s41598-026-40297-w
  • Zerroug, Y., et al. (2025). The association of cortisol and testosterone interaction with inpatient violence: Examining the dual-hormone hypothesis in a psychiatric setting. Aggressive Behavior, 51, e70027. https://doi.org/10.1002/ab.70027

Individual variability, gene-environment interaction, and the sorting dynamic

  • Caspi, A., McClay, J., Moffitt, T. E., Mill, J., Martin, J., Craig, I. W., Taylor, A., & Poulton, R. (2002). Role of genotype in the cycle of violence in maltreated children. Science, 297(5582), 851–854. https://doi.org/10.1126/science.1072290
  • Rhee, S. H., & Waldman, I. D. (2002). Genetic and environmental influences on antisocial behavior: A meta-analysis of twin and adoption studies. Psychological Bulletin, 128(3), 490–529. https://doi.org/10.1037/0033-2909.128.3.490
  • Kim-Cohen, J., Caspi, A., Taylor, A., Williams, B., Newcombe, R., Craig, I. W., & Moffitt, T. E. (2006). MAOA, maltreatment, and gene-environment interaction predicting children’s mental health: New evidence and a meta-analysis. Molecular Psychiatry, 11(10), 903–913. https://doi.org/10.1038/sj.mp.4001851
  • Byrd, A. L., & Manuck, S. B. (2014). MAOA, childhood maltreatment, and antisocial behavior: Meta-analysis of a gene-environment interaction. Biological Psychiatry, 75(1), 9–17. https://doi.org/10.1016/j.biopsych.2013.05.004
  • Gallardo-Pujol, D., Andrés-Pueyo, A., & Maydeu-Olivares, A. (2013). MAOA genotype, social exclusion and aggression: An experimental test of a gene-environment interaction. Genes, Brain and Behavior, 12(1), 140–145. https://doi.org/10.1111/j.1601-183X.2012.00868.x
  • Newman, T. K., Syagailo, Y. V., Barr, C. S., Wendland, J. R., Champoux, M., Graessle, M., Suomi, S. J., Higley, J. D., & Lesch, K.-P. (2005). Monoamine oxidase A gene promoter variation and rearing experience influences aggressive behavior in rhesus monkeys. Biological Psychiatry, 57(2), 167–172. https://doi.org/10.1016/j.biopsych.2004.10.012
  • Burt, S. A. (2009). Are there meaningful etiological differences within antisocial behavior? Results of a meta-analysis. Clinical Psychology Review, 29(2), 163–178. https://doi.org/10.1016/j.cpr.2008.12.004
  • Ficks, C. A., & Waldman, I. D. (2014). Candidate genes for aggression and antisocial behavior: A meta-analysis of association studies of the 5HTTLPR and MAOA-uVNTR. Behavior Genetics, 44(5), 427–444. https://doi.org/10.1007/s10519-014-9661-y
  • Plomin, R., & Bergeman, C. S. (1991). The nature of nurture: Genetic influence on “environmental” measures. Behavioral and Brain Sciences, 14(3), 373–386. https://doi.org/10.1017/S0140525X00070278
  • Jaffee, S. R., & Price, T. S. (2007). Gene-environment correlations: A review of the evidence and implications for prevention of mental illness. Molecular Psychiatry, 12(5), 432–442. https://doi.org/10.1038/sj.mp.4001950
  • Knafo, A., & Jaffee, S. R. (2013). Gene-environment correlation in developmental psychopathology. Development and Psychopathology, 25(1), 1–6. https://doi.org/10.1017/S0954579412000855
  • Burt, S. A., McGue, M., & Iacono, W. G. (2009). Nonshared environmental mediation of the association between deviant peer affiliation and adolescent externalizing behaviors over time: Results from a cross-lagged monozygotic twin differences design. Developmental Psychology, 45(6), 1752–1760. https://doi.org/10.1037/a0016687
  • Beaver, K. M., DeLisi, M., Vaughn, M. G., & Barnes, J. C. (2010). Monoamine oxidase A genotype is associated with gang membership and weapon use. Comprehensive Psychiatry, 51(2), 130–134. https://doi.org/10.1016/j.comppsych.2009.03.010
  • Dishion, T. J., & Tipsord, J. M. (2011). Peer contagion in child and adolescent social and emotional development. Annual Review of Psychology, 62, 189–214. https://doi.org/10.1146/annurev.psych.093008.100412

Homophily and group stability

  • McPherson, M., Smith-Lovin, L., & Cook, J. M. (2001). Birds of a feather: Homophily in social networks. Annual Review of Sociology, 27, 415–444. https://doi.org/10.1146/annurev.soc.27.1.415
  • Centola, D., & Macy, M. (2007). Complex contagions and the weakness of long ties. American Journal of Sociology, 113(3), 702–734. https://doi.org/10.1086/521848
  • Christakis, N. A., & Fowler, J. H. (2007). The spread of obesity in a large social network over 32 years. New England Journal of Medicine, 357(4), 370–379. https://doi.org/10.1056/NEJMsa066082
  • Christakis, N. A., & Fowler, J. H. (2008). The collective dynamics of smoking in a large social network. New England Journal of Medicine, 358(21), 2249–2258. https://doi.org/10.1056/NEJMsa0706154

Neural mechanism and the limits of awareness

Testosterone and status

  • Mazur, A., & Booth, A. (1998). Testosterone and dominance in men. Behavioral and Brain Sciences, 21(3), 353–363. https://doi.org/10.1017/S0140525X98001228
  • Eisenegger, C., Naef, M., Snozzi, R., Heinrichs, M., & Fehr, E. (2010). Prejudice and truth about the effect of testosterone on human bargaining behaviour. Nature, 463(7279), 356–359. https://doi.org/10.1038/nature08711
  • Eisenegger, C., Haushofer, J., & Fehr, E. (2011). The role of testosterone in social interaction. Trends in Cognitive Sciences, 15(6), 263–271. https://doi.org/10.1016/j.tics.2011.04.008
  • Boksem, M. A. S., Mehta, P. H., Van den Bergh, B., van Son, V., Trautmann, S. T., Roelofs, K., Smidts, A., & Sanfey, A. G. (2013). Testosterone inhibits trust but promotes reciprocity. Psychological Science, 24(11), 2306–2314. https://doi.org/10.1177/0956797613495063
  • Diekhof, E. K., Wittmer, S., & Reimers, L. (2014). Does competition really bring out the worst? Testosterone, social distance and inter-male competition shape parochial altruism in human males. PLOS ONE, 9(7), e98977. https://doi.org/10.1371/journal.pone.0098977
  • Carré, J. M., & Olmstead, N. A. (2015). Social neuroendocrinology of human aggression: Examining the role of competition-induced testosterone dynamics. Neuroscience, 286, 171–186. https://doi.org/10.1016/j.neuroscience.2014.11.029

Limits, replication, and prior literature

  • Dekkers, T. J., van Rentergem, J. A. A., Meijer, B., Popma, A., Wagemaker, E., & Huizenga, H. M. (2019). A meta-analytical evaluation of the dual-hormone hypothesis: Does cortisol moderate the relationship between testosterone and status, dominance, risk taking, aggression, and psychopathy? Neuroscience & Biobehavioral Reviews, 96, 250–271. https://doi.org/10.1016/j.neubiorev.2018.12.004
  • Border, R., Johnson, E. C., Evans, L. M., Smolen, A., Berley, N., Sullivan, P. F., & Keller, M. C. (2019). No support for historical candidate gene or candidate gene-by-interaction hypotheses for major depression across multiple large samples. American Journal of Psychiatry, 176(5), 376–387. https://doi.org/10.1176/appi.ajp.2018.18070881
  • Duncan, L. E., & Keller, M. C. (2011). A critical review of the first 10 years of candidate gene-by-environment interaction research in psychiatry. American Journal of Psychiatry, 168(10), 1041–1049. https://doi.org/10.1176/appi.ajp.2011.11020191
  • Sapolsky, R. M. (2017). Behave: The Biology of Humans at Our Best and Worst. Penguin Press.
  • Mullainathan, S., & Shafir, E. (2013). Scarcity: Why Having Too Little Means So Much. Times Books.
  • Lieberman, D. E. (2013). The Story of the Human Body: Evolution, Health, and Disease. Pantheon.
  • Wilkinson, R. G., & Pickett, K. E. (2009). The Spirit Level: Why Greater Equality Makes Societies Stronger. Bloomsbury.
  • Wilkinson, R. G., & Pickett, K. E. (2018). The Inner Level: How More Equal Societies Reduce Stress, Restore Sanity and Improve Everyone’s Well-Being. Penguin.