Skip to content

Hamilton's Rule

Predict when an allele for a costly social behaviour spreads by relocating the accounting from organism to gene: it is favoured whenever rB > C — the relatedness-weighted benefit to relatives exceeds the cost to the actor.

Core Idea

Hamilton's rule is the condition, formalised by W. D. Hamilton in 1964, under which an allele encoding a costly social behaviour will be favoured by natural selection: an altruistic act that imposes a fitness cost C on the actor and delivers a fitness benefit B to a recipient will spread when rB > C, where r is the coefficient of relatedness — the probability, above the population average, that the recipient carries an allele identical by descent to the one driving the behaviour in the actor. The inequality is not merely a metaphor for social intuition; it is a quantitative prediction derived from the gene's-eye accounting of allele-frequency change, and it converts the evolutionary paradox of altruism into a tractable empirical test.

The paradox Hamilton resolved was acute. Darwinian selection acts through differential reproduction of heritable variants: an allele that causes its bearer to sacrifice offspring it would otherwise have produced seemed to be selecting against itself, and therefore could not spread. Hamilton's resolution was to shift the unit of accounting. What selection tracks is the fate of the allele, not the fate of the individual organism carrying it. An allele that reduces its actor's reproductive output by C but increases a relative's by B gains net copies in the population whenever rB — the relatedness-weighted reproductive gain — exceeds C. The individual organism may be altruistic; the allele is, in the gene's-eye accounting, selfish. This reframing, which Richard Dawkins later extended as the gene's-eye view of evolution, dissolves the paradox without invoking any mechanism outside standard population genetics.

The coefficient r is the load-bearing quantity that does the biological work. For diploid organisms, full siblings share r = 0.5 on average; half-siblings r = 0.25; first cousins r = 0.125. The haplodiploid genetics of Hymenoptera — bees, wasps, ants — produce an asymmetry that Hamilton immediately recognised: because the male father is haploid and contributes an identical genome to every daughter, full sisters in a Hymenoptera colony share r = 0.75 of their genomes, while a female shares only r = 0.5 with her own offspring. The arithmetic of Hamilton's rule can therefore favour a female that helps her mother (the queen) produce sisters over one that produces her own offspring directly — a potential genetic explanation for the repeated independent evolution of female-worker sterility and eusociality in the Hymenoptera but its rarity in most diploid taxa.

The rule generates specific and testable predictions that have been verified across vertebrate and invertebrate systems. It predicts that cooperative behaviour should be directed preferentially toward closer relatives; that the threshold relatedness for helping should rise as costs rise or benefits fall; that where relatedness to siblings exceeds relatedness to offspring — as in haplodiploidy — sterile worker castes should be evolutionarily accessible; and that organisms in populations with higher local relatedness (viscous, kin-structured populations) should show more cooperation than those in mixed populations where r among neighbours is close to zero. Paul Sherman's field work on alarm calling in Belding's ground squirrels, which showed that calling was performed preferentially by females with close relatives nearby and reduced when kin were experimentally absent, is a canonical empirical test. Bacterial public-goods systems — where quorum sensing and the secretion of shared metabolic products occur preferentially in clonal patches where r among cells approaches one — extend the framework to microbial sociality, where the gene-level accounting applies as cleanly as it does to animals with complex behaviour.

Hamilton's rule is the operational core of inclusive fitness theory — the enlargement of individual fitness to include fitness effects on relatives weighted by relatedness. In that framework, the rule does not merely explain altruism; it organises the entire space of social evolution, placing altruism, mutualism, selfishness, and spite in a four-cell table defined by the signs of cost to actor and benefit to recipient, with predictions about when each is evolutionarily stable given the r structure of the relevant population.

Structural Signature

Sig role-phrases:

  • the actor — the bearer of a social-behaviour allele who performs the costly act
  • the recipient — the individual who receives the act's fitness effect
  • the cost C — the actor's fitness loss, measured in expected offspring foregone
  • the benefit B — the recipient's fitness gain, measured in expected offspring
  • the relatedness coefficient r — the excess probability, above population baseline, that the recipient carries an identical-by-descent copy of the focal allele
  • the gene's-eye accounting — the relocation of the bookkeeping unit from organism to allele, so success is summed over copies in self plus weighted relatives
  • the spread inequality rB > C — the engineered condition: the allele is favored when the relatedness-weighted benefit exceeds the cost, converting the altruism paradox into a field test
  • the four-cell organization — cost-and-benefit signs partitioning all social behaviour into altruism, mutualism, selfishness, and spite, each with a stability condition read off the population's r structure
  • the haplodiploid special case and non-kin violations — the r = 0.75 sister asymmetry opening worker sterility, and cooperation at r ≈ 0 flagging a substitute (reciprocity, enforcement) rather than refuting the rule

What It Is Not

  • Not a "for the good of the group" or self-sacrifice account. The rule works through gene's-eye accounting: the organism may be altruistic, but the underlying allele gains net copies whenever rB > C, so it is, in the bookkeeping, selfish. The mechanism is ordinary allele-frequency change with the unit of accounting relocated from organism to gene — not group selection, and not a benefit to the species that overrides individual reproduction.
  • Not raw genealogical kinship or overall genome similarity. The coefficient r is the excess probability, above the population baseline, that the recipient carries an identical-by-descent copy of the focal allele — not total shared DNA. Because all conspecifics already share most of their genome, what matters is the relatedness above background at the gene driving the behaviour; reading r as "how much DNA we share" loses the load-bearing quantity.
  • Not refuted by cooperation among non-relatives. Helping at r ≈ 0 is not a counterexample but a diagnostic: the inequality cannot be satisfied by relatedness there, so the observation flags that a non-kin mechanism — reciprocity, enforcement, partner choice — is standing in for r. Those mechanisms are marked off inside the framework as departures requiring a substitute, not as evidence the rule fails.
  • Not a proof that haplodiploidy causes eusociality. The r = 0.75 sister asymmetry makes worker sterility evolutionarily accessible in Hymenoptera — a candidate explanation, not a demonstrated necessity. Eusociality has also evolved in diploid taxa lacking the asymmetry, and the haplodiploidy account is contested; the rule places such systems in the regime where worker castes are reachable, it does not entail that the asymmetry produced them.
  • Not the gene's-eye view in general, nor the Price equation. Hamilton's rule is one celebrated special case of multilevel-accounting of a replicator's success, with copies weighted by their probability of carrying the focal variant. The gene's-eye view is the philosophical frame and the Price equation is the general decomposition; the rule is the specific evolutionary-genetic prediction — its relatedness coefficients, haplodiploid asymmetry, and offspring-fitness currency are the cargo that does not generalize.

Scope of Application

Hamilton's rule lives across the social-evolution subfields of evolutionary biology, ranging over every sexually reproducing system where relatedness can be defined; its reach is bounded by replicators-under-selection, because its load-bearing content is quantitative and genetic — r from pedigree or markers, B and C as offspring-fitness, the inequality derived from allele-frequency change. The multilevel-accounting move that recurs in cultural or computational selection belongs to the parent (the Price-equation decomposition), not the rule. Within the domain it operates across these contexts.

  • Behavioural ecology — predictions about cooperation, alarm calls, food-sharing, allocare, and helping-at-the-nest in vertebrates and invertebrates, with relatedness-targeted helping as the testable signature.
  • Evolution of eusociality — the classical kin-selection account of worker sterility in ants, bees, and wasps, with the haplodiploid r = 0.75 sister asymmetry as Hamilton's headline application.
  • Conflict theory — parent-offspring conflict, sibling rivalry, and intragenomic conflict (imprinted genes, sex-ratio conflict) each treated as the same inequality with different relatedness coefficients on different sides.
  • Microbial social evolution — secreted public goods and social cheating in bacteria studied via r estimated from clonal patch structure, where relatedness among cells approaches one.
  • Non-kin cooperation theory — reciprocal altruism, partner choice, and enforcement are framed within the framework as departures from kin selection requiring non-r mechanisms, flagged whenever cooperation appears at r ≈ 0.

Clarity

Hamilton's rule resolves the sharpest standing paradox in Darwinian theory — how selection, which works through differential reproduction, could ever favour a behaviour that lowers the actor's own reproduction — and the clarifying move is a change in the unit of accounting. By relocating the bookkeeping from the organism to the allele, the rule shows that an act sacrificing the actor's offspring still gains net copies of the underlying allele whenever the relatedness-weighted benefit to relatives, rB, exceeds the cost C. The organism is altruistic; the allele, in the gene's-eye accounting, is not. That single reframing dissolves the apparent contradiction without reaching outside standard population genetics, and it converts "the evolution of altruism" from a verbal puzzle into a quantitative inequality whose three terms — cost, benefit, relatedness — can be measured in the field and the prediction tested.

The rule's second clarifying contribution is to organise the entire space of social behaviour rather than to explain altruism alone. Once cost-to-actor and benefit-to-recipient are the axes, altruism, mutualism, selfishness, and spite fall into a four-cell table defined by their signs, each with a stability condition read off the population's relatedness structure — turning a scattered vocabulary of cooperation and conflict into one accounting scheme. It also makes the load-bearing role of r legible: relatedness becomes a quantity to be measured, not an intuition, so the haplodiploid asymmetry of the Hymenoptera (full sisters at r = 0.75, own offspring at only r = 0.5) becomes a candidate explanation for repeated worker sterility, and parent–offspring conflict becomes a built-in consequence of the differing relatedness each party has to the contested investment. The sharper questions the practitioner can now pose follow directly: not "why do animals cooperate?" but "is rB > C in this system — how related are the parties, how large the benefit against the cost, and where cooperation appears among near-strangers (r ≈ 0), what non-kin mechanism must be standing in for relatedness?" — the last being exactly how reciprocity, enforcement, and partner choice get marked off as departures from the kin-selection account rather than confused with it.

Manages Complexity

Deciding whether a social-behaviour allele will spread is, taken literally, a population-genetic bookkeeping problem of forbidding size: one must track allele frequencies in actor and recipient, conditional on the behaviour occurring, across the kin structure of the whole population. Hamilton's rule collapses that accounting onto three measurable scalars and a single inequality — relatedness r, benefit B, cost C, with rB > C the spread condition — so that the analyst predicts the qualitative fate of cooperation in any system by estimating three quantities (r from pedigree or molecular markers, B and C from fitness assays) rather than re-deriving allele dynamics case by case. That compression is what makes a verbal paradox into a field test, and it scales the disparate-looking cooperation literature onto one axis: helping kin, worker sterility, alarm calling, microbial public goods, and parent–offspring conflict all become the same inequality evaluated at different values of r. The rule further organises the entire space of social behaviour into a four-cell table once cost-to-actor and benefit-to-recipient are the axes — altruism, mutualism, selfishness, spite — each with a stability condition read off the population's relatedness structure, replacing a scattered vocabulary of cooperation and conflict with one accounting scheme. And because the load-bearing parameter is explicit, special cases that would otherwise look mysterious become single substitutions: the haplodiploid asymmetry (sisters at r = 0.75, own offspring at r = 0.5) reads off worker-sterility accessibility directly, and cooperation among near-strangers (r ≈ 0) is flagged as a violation that pins the explanation on some non-kin mechanism standing in for relatedness, rather than absorbed as noise — the analyst locating the anomaly by which term of the inequality fails rather than reconstructing each system from scratch.

Abstract Reasoning

Hamilton's rule licenses a set of inferences that all run on its three measurable scalars — relatedness r, benefit B, cost C — and the single inequality rB > C.

Diagnostic. The rule infers the evolutionary viability of an observed social behaviour from the relatedness structure of the parties, and the move runs from a pattern of helping back to the gene's-eye accounting that sustains it. Observing that a costly act is directed preferentially toward certain individuals, the analyst infers that those individuals are, on average, related closely enough that rB exceeds C — so the targeting of altruism is read as a signature of relatedness, and the prediction is that help should concentrate on closer kin and thin out as r falls. The four-cell table is itself a diagnostic device: from the signs of cost-to-actor and benefit-to-recipient, a behaviour is classified as altruism, mutualism, selfishness, or spite, each carrying its own stability condition read off the population's r structure. And the load-bearing case — cooperation appearing among near-strangers, where r ≈ 0 — is diagnostic precisely because the inequality cannot be satisfied by relatedness, so the observation is read as a pointer that some non-kin mechanism (reciprocity, enforcement, partner choice) is standing in for r.

Interventionist. The inequality specifies exactly which terms move the outcome and in which direction. Raise r — direct the behaviour toward closer relatives, or move into a viscous, kin-structured population where neighbours share more alleles — and the rule predicts more cooperation, because the relatedness-weighted return rises against a fixed cost. Raise the cost C, or lower the benefit B, and the predicted threshold relatedness for helping rises, so cooperation should retreat to ever-closer kin as it becomes more expensive. The most striking interventionist prediction is the experimental one already realised in the field: remove a caller's kin from its neighbourhood and the rule predicts call frequency should fall, because with the recipients no longer related the rB term collapses below C — a causal manipulation of the relatedness term with a directional prediction on the behaviour.

Boundary-drawing. The rule draws the line at where relatedness can be defined and the gene's-eye accounting applies — sexually reproducing systems with a measurable r, from vertebrates and insects to clonally structured microbial patches where r among cells approaches one. Within that scope it forces a regime decision the bare observation of cooperation cannot settle: is this a kin-selected behaviour, explicable by rB > C at the system's actual relatedness, or a non-kin one that requires a substitute mechanism because r is too low to carry it? Placing a case on the right side of that boundary is what separates a kin-selection account from reciprocity or enforcement, rather than confusing the two. The haplodiploid asymmetry draws a further boundary inside the kin-selected regime: where the sex-determination genetics make relatedness to siblings (r = 0.75 among Hymenopteran full sisters) exceed relatedness to one's own offspring (r = 0.5), the rule places the system in the regime where sterile worker castes are evolutionarily accessible — a boundary that most diploid taxa, lacking the asymmetry, fall outside.

Predictive / order-of-events. The rule predicts which systems should evolve cooperation before it is observed: high-r groups with favourable B-against-C opportunities are predicted to be where helping, alarm calling, allocare, and worker sterility appear, and low-r mixed populations are predicted to show little. It predicts that kin-recognition machinery should evolve wherever directing aid toward higher-r relatives raises inclusive-fitness returns, because accuracy in identifying kin lets the rB > C condition be met more often. And it builds parent-offspring conflict into the relatedness arithmetic as a structural consequence: because each party values the contested parental investment by its own differing relatedness to the recipients, the rule predicts conflict over investment as a built-in feature of the r structure rather than an aberration, anticipating where cooperation and conflict will coexist within the same family.

Knowledge Transfer

Within evolutionary biology Hamilton's rule transfers as mechanism across every sexually reproducing system where relatedness can be defined, because the cargo is one inequality grounded in gene's-eye allele-frequency accounting — rB > C, with r the excess identity-by-descent between actor and recipient. From behavioural ecology (alarm calls, food-sharing, allocare, helping-at-the-nest) it carries to the evolution of eusociality (the haplodiploid asymmetry as the headline application to worker sterility), to conflict theory (parent-offspring conflict, sibling rivalry, intragenomic conflict, each the same accounting with different r on different sides), and to microbial cooperation (secreted public goods and cheating, with r estimated from clonal patch structure where it approaches one). Across all of these the apparatus carries without translation — the three measurable scalars, the four-cell altruism/mutualism/selfishness/spite table with stability read off the population's r structure, the prediction that helping concentrates on closer kin, and the diagnostic that cooperation among near-strangers (r ≈ 0) flags a non-kin mechanism. Even its apparent counter-examples are handled inside the framework: reciprocal altruism, partner choice, and enforcement are marked off as departures requiring non-r mechanisms, not as refutations. The named applications across clades — eusocial insects, vertebrate cooperators, microbial cheats — are different content domains of one substrate (replicators under selection with measurable relatedness), not structurally distinct substrates, which is exactly why the rule transfers among them intact.

Beyond replicators-under-selection the named rule does not travel, and the honest split is between loose analogy and a genuinely portable parent. Cultural-evolutionary models of nepotism (with a cultural "relatedness" proxy), economic-anthropological models of kin favouritism, and distributed-computing analogies (replicated agents, shared-state replication benefits) do reach for Hamilton's rule, but these are (A) loose analogies that strain the underlying genetics: there is no allele, no identity-by-descent, no fitness measured in offspring, and the load-bearing quantitative content — r from pedigree or markers, B and C from fitness assays, the inequality derived from allele-frequency change — does not transfer. What genuinely recurs cross-domain is the (B) case: the deeper structural commitment the rule rests on — multilevel accounting of a replicator's success, with copies weighted by their probability of carrying the focal variant — is the powerful abstract move, and it is captured at the right generality by the parent, the Price-equation / multilevel-selection decomposition, of which Hamilton's rule is one celebrated special case (the gene's-eye view of Dawkins is the philosophical frame around it). An analyst applying weighted-replicator accounting to cultural transmission or replicated computation is instantiating that general decomposition, not importing Hamilton's rule, whose population-genetic cargo (relatedness coefficients, haplodiploid asymmetry, the offspring-fitness currency) stays home. Stripped of jargon Hamilton's rule is "help relatives in proportion to relatedness and to benefit-against-cost" — a specific evolutionary-genetic prediction, not a substrate-independent pattern; the substrate-independent pattern is the multilevel accounting it instantiates. See Structural Core vs. Domain Accent.

Examples

Canonical

J. B. S. Haldane is said to have quipped that he would lay down his life for two brothers or eight cousins — the arithmetic that Hamilton later formalized. In a diploid species a full sibling has relatedness r = 0.5 and a first cousin r = 0.125. Take an altruistic allele whose act costs the actor its own life, foregoing C offspring, to save relatives who together gain B offspring. The spread condition rB > C is met at exactly break-even when r·(saved) = (lost): 0.5 × 2 = 1 sibling-equivalent, and 0.125 × 8 = 1 cousin-equivalent. So the allele gains net copies only if the act saves more than two full siblings, or more than eight first cousins — fewer than that, and the relatedness-weighted benefit fails to cover the cost and the allele is selected against.

Mapped back: The dying individual is the actor, the rescued kin the recipients, the foregone offspring the cost C, and the saved offspring the benefit B. The 0.5 and 0.125 are the relatedness coefficient r. The "two brothers / eight cousins" break-even is the spread inequality rB > C evaluated exactly — the reframing that shifts bookkeeping to the gene's-eye accounting and converts a paradox into an arithmetic test.

Applied / In Practice

Paul Sherman's field study of Belding's ground squirrels (Sciurus beldingi) in the Sierra Nevada is a canonical empirical test. These squirrels give alarm calls when a predator approaches — a costly act, since calling draws the predator's attention to the caller. Sherman found the calls were given disproportionately by females, who remain near their birthplace surrounded by close kin, rather than by dispersing males; and calling tracked the presence of relatives rather than mates or offspring alone. Where an individual had close kin nearby, it called; where kin were absent, calling dropped. This is exactly the pattern Hamilton's rule predicts: the costly signal is deployed only when the relatedness-weighted benefit to nearby kin can exceed the personal cost.

Mapped back: The self-endangering alarm call is the actor's cost C; the warned relatives' survival is the benefit B, weighted by the relatedness coefficient r of the surrounding kin. That calling concentrates where kin are present, and thins when they are absent, is the diagnostic signature — the spread inequality rB > C being met only at high r. Sherman's kin-presence contrast is the field realization of the rule's interventionist prediction.

Structural Tensions

T1: Gene's-eye accounting versus organism-level intuition (the altruist is a selfish allele). The rule dissolves the altruism paradox by relocating the bookkeeping from organism to allele: the individual sacrifices offspring, but the allele gains net copies whenever rB > C. That relocation is the whole move, yet it is counterintuitive in a way that invites two misreadings — as a "for the good of the group" account (it is not; it is ordinary allele-frequency change) or as genuine self-sacrifice overriding reproduction (it is not; the gene is, in the accounting, selfish). The tension is that the rule's explanatory power comes precisely from adopting a unit of accounting alien to the organism-level intuition it is explaining, so its correct use requires holding two levels apart that the phenomenon (a self-endangering animal) constantly fuses. Diagnostic: Is the behaviour being explained by net copies of the focal allele across weighted relatives, or has the account slid into organism-level sacrifice or group benefit?

T2: A three-scalar inequality versus the difficulty of measuring its terms (elegance against field operationalization). The rule's great compression is that the fate of a social-behaviour allele reduces to r, B, and C and a single inequality — but two of those scalars are notoriously hard to pin down. B and C are fitness effects measured in expected offspring, requiring assays that isolate the act's marginal reproductive consequence; r is an above-baseline identity-by-descent that pedigree or markers estimate only approximately. The inequality is exact; its inputs are noisy and often unobservable in the field. So the rule can be simultaneously the cleanest prediction in social evolution and, in a given system, nearly untestable without heroic measurement — the analytic tractability and the empirical tractability pull apart. Diagnostic: Are r, B, and C in this system independently measured with enough precision to test rB > C, or is the inequality being asserted from assumed values that make it unfalsifiable here?

T3: A framework that organizes all social behaviour versus its own falsifiability (non-kin cooperation absorbed as a substitute). The four-cell table places altruism, mutualism, selfishness, and spite in one scheme, and cooperation among near-strangers (r ≈ 0) is handled inside the framework as a diagnostic pointer to a non-kin mechanism standing in for relatedness. This comprehensiveness is a strength — apparent counterexamples become located anomalies rather than refutations — but it edges toward the vice of a theory that cannot fail: every observed cooperation is either kin-selected (rB > C) or flagged as reciprocity/enforcement/partner-choice substituting for r. The tension is that the same move which makes the rule exhaustive also insulates it, so its content rests entirely on whether r, B, and C are measured independently rather than inferred backward from the behaviour they are meant to explain. Diagnostic: Is the kin-selection verdict here derived from independently estimated relatedness and fitness, or is r being back-fitted so that whatever cooperation is observed comes out consistent with the rule?

T4: Haplodiploidy as accessible versus as causal (a regime the rule opens is not a mechanism it proves). The r = 0.75 sister asymmetry makes worker sterility evolutionarily accessible in the Hymenoptera, and the rule's arithmetic shows why a female might do better helping her mother produce sisters than reproducing herself. But accessibility is not necessity: eusociality has evolved repeatedly in diploid taxa lacking the asymmetry, and the haplodiploidy explanation is contested. The tension is that the rule's most celebrated application is also its most over-claimed — it places a system in the regime where a trait is reachable without entailing that the asymmetry produced it, and the vividness of the 0.75 figure invites reading a candidate explanation as a demonstrated cause. Diagnostic: Is the claim that haplodiploidy made worker sterility reachable in this clade, or the stronger claim that it caused the eusociality — and does the diploid counter-evidence undercut the latter?

T5: Relatedness as rigorous quantity versus relatedness as routinely misread (the load-bearing term's precision is its trap). The coefficient r is the excess probability, above population baseline, that the recipient carries an identical-by-descent copy of the focal allele — not total shared DNA, of which conspecifics already share nearly all. That precise definition is exactly what makes the rule a quantitative prediction rather than a slogan. But the same precision is what makes it routinely misapplied: analysts plug in genealogical kinship or genome-wide similarity, or import a cultural or economic "relatedness" proxy, and the load-bearing quantity silently changes into something the derivation does not license. The rigor that gives r its power and the ease of substituting a superficially similar number are the same feature. Diagnostic: Is the r in use the above-baseline identity-by-descent at the focal allele, or has it quietly become total shared DNA, genealogical kinship, or a metaphorical similarity that the allele-frequency derivation cannot carry?

T6: Autonomy versus reduction (a named population-genetic prediction or the multilevel-accounting move it instantiates). Hamilton's rule is a specific, celebrated evolutionary-genetic prediction with proprietary cargo — relatedness coefficients, the haplodiploid asymmetry, fitness measured in offspring, the inequality derived from allele-frequency change — and within replicators-under-selection it transfers intact as mechanism across insects, vertebrates, and microbes. But beyond replicators the named rule does not travel: cultural, economic, and computational invocations strain the genetics (no allele, no identity-by-descent, no offspring currency). What genuinely recurs cross-domain is the deeper move it is one special case of — multilevel accounting of a replicator's success, copies weighted by their probability of carrying the focal variant — captured at the right generality by the Price-equation decomposition, with the gene's-eye view as its philosophical frame. Diagnostic: Resolve toward the Price-equation / multilevel-selection parent when carrying weighted-replicator accounting to cultural or computational systems; toward Hamilton's rule when predicting social behaviour in a system with a genuinely measurable relatedness coefficient.

Structural–Framed Character

Hamilton's rule sits toward the structural end of the spectrum but stops short of the pole — best read as mixed-structural, closely parallel to Haldane's sieve and to isostasy: a genuine relational mechanism wearing irreducibly population-genetic vocabulary. Four of the five criteria give it strong structural credentials, and only the vocabulary criterion pins it home.

Its evaluative_weight is nil. rB > C is a neutral spread condition, not a verdict — the rule names when an allele gains net copies, and calls no behaviour right or wrong; even "altruism," "selfishness," and "spite" enter as sign-defined cells of an accounting table, not as moral appraisals. It is not human-practice-bound in any sense: the mechanism runs observer-free — Belding's ground squirrels alarm-call preferentially near kin, Hymenopteran colonies evolve sterile worker castes, bacterial cells secrete public goods in clonal patches, all whether or not a biologist ever writes the inequality; strip away every theorist and relatedness-weighted selection still favours or disfavours the allele. Its institutional_origin is none: Hamilton in 1964 derived a consequence of gene's-eye allele-frequency accounting, he did not legislate it — it is a fact about how selection sums copies over weighted relatives, not an artifact of a survey, agency, or convention. And within its proper range cross-domain reuse is recognition, not import: moving across behavioural ecology, eusociality, conflict theory, and microbial sociality, the same mechanism is recognized intact — the three scalars, the four-cell table, the concentrate-help-on-closer-kin prediction, the r ≈ 0 diagnostic all keep their content, and even the reach to clonal microbial patches is recognition of the identical accounting, not analogy.

What keeps it off the structural pole is vocab_travels, which it fails decisively. The operative vocabulary is irreducibly population-genetic — the relatedness coefficient r as above-baseline identity-by-descent at a focal allele, fitness measured in expected offspring, the haplodiploid r = 0.75 sister asymmetry, the inequality derived from allele-frequency change — and none of it floats free of replicators-under-selection. Beyond that substrate, cultural "relatedness" proxies, economic kin-favouritism models, and distributed-computing analogies borrow the shape while renaming or dropping every load-bearing term (no allele, no identity-by-descent, no offspring currency), so there the import_vs_recognize mark flips from recognition to import-by-analogy — exactly the misreading T5 warns against when analysts plug genome-wide similarity or a metaphorical proxy into r.

The portable structural skeleton is multilevel accounting of a replicator's success — summing copies over carriers, each weighted by its probability of bearing the focal variant. That skeleton is substrate-portable, and it is precisely what Hamilton's rule instantiates from its umbrella — the Price-equation / multilevel-selection decomposition, with the gene's-eye view as its philosophical frame — not what makes "Hamilton's rule" itself travel: the cross-domain reach belongs to that general weighted-replicator accounting, of which the rule is one celebrated special case, while the relatedness coefficients, haplodiploid asymmetry, and offspring-fitness currency stay home. Its character: a structural, evaluatively neutral, recognized-in-nature selection mechanism whose skeleton is the portable multilevel-accounting decomposition it instantiates, but whose distinctive content is stated in population-genetic vocabulary that pins it to its home domain, leaving it mixed-structural rather than a free-floating prime.

Structural Core vs. Domain Accent

This section decides why Hamilton's rule is a domain-specific abstraction and not a prime, and it carries the case for its domain-specificity — there is no separate section for that.

What is skeletal (could lift toward a cross-domain prime). Strip the genetics and a thin relational structure survives: the success of a self-copying entity is scored not on the individual bearer but by summing copies across all carriers, each weighted by its probability of bearing the focal variant — so a variant that lowers its bearer's own reproduction can still gain net copies whenever the weighted return elsewhere exceeds the local loss. The pieces that travel are abstract: a replicator, a bookkeeping unit relocated from the bearer to the copy, a weighting that measures how likely a beneficiary is to carry the same variant, and a net-gain condition summed over the weighted set. That skeleton — multilevel accounting of a replicator's success — is genuinely substrate-portable, which is exactly why the entry names the Price-equation / multilevel-selection decomposition (with Dawkins's gene's-eye view as its philosophical frame) as the parent Hamilton's rule instantiates. But it is the core the rule shares, not what makes the rule distinctive.

What is domain-bound. Almost all the load-bearing content is population-genetics furniture and none of it survives extraction intact: the relatedness coefficient r as the excess, above population baseline, probability of identity-by-descent at a focal allele (not shared DNA, not genealogical kinship); the cost C and benefit B measured specifically in expected offspring; the inequality rB > C derived from allele-frequency change; the haplodiploid r = 0.75 sister asymmetry that opens worker sterility; and the four-cell altruism/mutualism/selfishness/spite table with stability read off a population's r structure. The decisive test: remove the allele and identity-by-descent — the very quantities r is defined on — and there is no rule, only the loose "help those like you in proportion to benefit" resemblance. r is not a portable dial; it is a pedigree-or-marker estimate of a genetic quantity, and its precise definition (T5) is exactly what makes the rule a quantitative prediction rather than a slogan, and exactly what fails to travel when a cultural or economic "relatedness" proxy is substituted.

Why this does not clear the prime bar. A prime is a relational structure whose vocabulary travels and whose cross-domain transfer is recognition of the same mechanism, not analogy. Hamilton's rule's transfer is bimodal. Within replicators-under-selection it travels intact — behavioural ecology, eusociality, conflict theory, and microbial sociality are different content domains of one substrate, so the three scalars, the four-cell table, the concentrate-help-on-closer-kin prediction, and the r ≈ 0 non-kin diagnostic are all recognized, not re-derived, and even apparent counterexamples (reciprocity, enforcement, partner choice) are marked off inside the framework. Beyond replicators the named rule does not travel: cultural nepotism models, economic kin-favouritism, and distributed-computing analogies strain the genetics — no allele, no identity-by-descent, no offspring currency — so there it is analogy, not mechanism. And when the bare structural lesson — weighted-replicator accounting — is needed cross-domain, it is already supplied in more general form by the parent it instantiates: the Price-equation / multilevel-selection decomposition, of which Hamilton's rule is one celebrated special case. The cross-domain reach belongs to that parent; "Hamilton's rule," as named, carries population-genetic baggage — relatedness coefficients, the haplodiploid asymmetry, the offspring-fitness currency — that does not and should not travel.

Relationships to Other Abstractions

Local relationship map for Hamilton's RuleParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Hamilton's RuleDOMAINPrime abstraction: Threshold — is part ofThresholdPRIMEDomain-specific abstraction: Inclusive Fitness — presupposesInclusiveFitnessDOMAINDomain-specific abstraction: Kin selection — is part ofKin selectionDOMAIN

Current abstraction Hamilton's Rule Domain-specific

Parents (2) — more general patterns this builds on

  • Hamilton's Rule presupposes Inclusive Fitness Domain-specific

    Hamilton’s rule presupposes the inclusive-fitness quantity whose marginal change its inequality tests.

  • Hamilton's Rule is part of Threshold Prime

    Hamilton’s rule contains a threshold boundary at rB equals C between favored and disfavored alleles.

Children (1) — more specific cases that build on this

  • Kin selection Domain-specific is part of Hamilton's Rule

    Kin selection contains Hamilton’s rule as its operational spread criterion.

Hierarchy paths (2) — routes to 2 parentless roots

Not to Be Confused With

  • Kin selection / inclusive fitness theory. The broader theoretical framework — the enlargement of individual fitness to include effects on relatives weighted by relatedness — of which Hamilton's rule is the operational core. The framework is the whole edifice (the concept of inclusive fitness, the process of selection acting through relatives); the rule is the specific inequality rB > C that makes the framework quantitative and testable. Tell: are you naming the general idea that selection can act through relatives (kin selection / inclusive fitness), or the precise spread condition that predicts when it does (Hamilton's rule)?

  • Group selection / "for the good of the species." The rejected alternative that explains altruism by benefits accruing to the group or species, overriding individual reproduction. Hamilton's rule explicitly does not work this way: it is ordinary allele-frequency change with the accounting unit relocated to the gene, so the allele is selfish in the bookkeeping even when the organism is altruistic. Tell: is the beneficiary a group whose survival is said to select the trait (group selection), or the focal allele gaining net copies through weighted relatives (Hamilton's rule)?

  • Reciprocal altruism. Cooperation sustained by the expectation of return benefit — "I help you now, you help me later" — rather than by relatedness. It is precisely the non-kin substitute the rule flags whenever cooperation appears at r ≈ 0, where the inequality cannot be satisfied by relatedness. Reciprocal altruism is marked off inside the framework as a departure requiring a non-r mechanism, not as an instance of the rule. Tell: is the cooperation directed at relatives and explained by rB > C (Hamilton's rule), or at unrelated partners and sustained by returned benefit, enforcement, or partner choice (reciprocity)?

  • Green-beard effect. Cooperation directed via a gene (or linked gene set) that simultaneously produces a recognizable marker and the behaviour of helping others bearing that marker — so relatedness at the focal locus is high even between genealogical strangers. It shares the gene's-eye logic but bypasses pedigree: help flows to co-bearers of the specific allele rather than to close kin. Hamilton's rule as usually applied uses genealogical/whole-genome relatedness, whereas the green-beard identifies carriers of the focal allele directly. Tell: is aid targeted by overall relatedness estimated from pedigree or markers (standard Hamilton's rule), or by a directly recognized tag tied to the very allele driving the help (green-beard)?

  • The Price equation / multilevel-selection decomposition (umbrella). The general accounting of a replicator's success — summing copies over carriers, each weighted by its probability of bearing the focal variant — of which Hamilton's rule is one celebrated special case, with Dawkins's gene's-eye view as its philosophical frame. The umbrella is what travels to cultural or computational selection; the rule adds the relatedness coefficient, haplodiploid asymmetry, and offspring-fitness currency that stay home. Tell: strip away the allele, identity-by-descent, and offspring fitness and what remains is bare weighted-replicator accounting — the Price-equation parent, not Hamilton's rule. (Treated fully in a later section.)

Neighborhood in Abstraction Space

Hamilton's Rule sits in a crowded region of the domain-specific corpus (18th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Population Genetics & Kin Selection (10 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-07-12