The Three Evolutions Refuse to Become One

The Department Standardized Evolution

The Department solved the problem at 08:40.

This was concerning because I had not yet finished describing it.

Chapter 3 had preserved a narrow claim for Triadic Evolution. Its architecture could keep several legitimate analytical cuts from collapsing when specialists combined their work, but it had not shown that the framework's three named evolutions were scientifically equivalent, independently necessary, or even entitled to the same noun. The final sentence of my report had therefore removed a courtesy that the source had so far received without examination: I would no longer call all three processes evolution merely because the book did.

The Department responded by creating Form 31-E, Evolutionary Process Equivalence and Terminological Control. Its purpose was to ensure that any process using the word evolution met one approved definition before the term could be released for field use.

This was methodologically attractive. It was also designed around biology.

Evolution by Natural Selection completed the form almost immediately. Evolution by Organization disputed several fields. Evolution by Extension asked whether model number could be substituted for generation length and whether warranty claims counted as fitness. The Department denied both requests.

At 08:47, the form classified one strand as evolution and two as metaphor.

I nearly accepted the result.

There was an obvious advantage. Evolutionary biology possesses a mature population framework in which inherited variation, reproduction, mutation, drift, gene flow, selection, and changing allele frequencies can be specified and measured. Natural selection has no mission, no designer, and no meeting at which a population agrees to become better adapted. By contrast, institutions can be redesigned deliberately, and artifacts are commonly created by actors who become impatient when variation is left blind.

But the Department had embedded its conclusion in the fields. Cultural-evolution research had spent decades developing formal accounts of socially transmitted variation without pretending that cultural inheritance was Mendelian. Evolutionary economics and organization studies had likewise examined variation, retention, selection, routines, organizational populations, technological lineages, and path dependence while arguing about how far Darwinian concepts could travel. The disagreement was not between science and metaphor. It was partly a disagreement about which properties make an evolutionary account scientifically useful outside genetics.^[2–6]

The source was also more careful than the form. Triadic Evolution explicitly denied that its three strands shared one mechanism. It treated them as interacting strands of cumulative adaptive change with different primary objects, retention routes, timescales, and relations to purpose. It even warned that selection among technologies was not one unified process and that organizational persistence could arise through power, habit, identity, dependency, institutionalization, or deliberate reform rather than natural selection.^[1]

The Department had therefore tested a claim the framework did not make.

I rejected the form.

I reopened the composite consultation with the evolutionary biologist from Chapter 2. As before, the exchange synthesizes recurring positions and objections in the cited literature; no statement is attributed to an identifiable scholar.

“You built a population-genetics form,” she said after I showed her 31-E.

“I was testing whether the three uses of evolution shared one scientific meaning.”

“No. You were testing whether three claimed processes were one biological process. They are not. Your source already says they are not.”

I asked whether biology therefore owned the noun.

“No. Biology owns the burden when the claim is biological. If you use evolution elsewhere, tell me what changes, what survives, how it survives, and what would become scientifically worse if you removed the word.”

The final condition was more severe than the Department's form. A nonbiological use of evolution did not have to imitate genes. It did have to justify why the evolutionary framing constrained inquiry rather than merely dignifying history.

This did not release the other two strands. It made their burden harder. If Evolution by Organization and Evolution by Extension were not Darwinian mechanisms, what made the word evolution do scientific work rather than provide a distinguished accent for change?

Natural Selection Returned a Complete Form

I began with the strand that did not require defense from metaphor.

Evolution by Natural Selection is the biological strand of Triadic Evolution. The source uses the formal name for a process broader than selection alone: inherited population change also includes mutation, recombination, genetic drift, gene flow, and related mechanisms. This is slightly inconvenient terminology, but the book states the inconvenience explicitly. Natural selection is central because it connects heritable differences to unequal reproductive outcomes under specified conditions; it is not the cause of every change in allele frequency.^[1]

Population genetics supplies the discipline that ordinary use of the word evolution often lacks. A population is specified. Variation is inherited. Allele or genotype frequencies can be measured across generations. Natural selection, drift, gene flow, mutation, mating structure, and demographic history can be distinguished. An advantageous-looking trait cannot be declared an adaptation merely because a retrospective story makes it useful. Selection is one mechanism inside evolutionary change, and adaptation is one possible result.^[2]

The distinction between development and evolution is equally important. A person can gain muscle, learn a language, receive a transplant, change diet, acquire immunity through vaccination, or become very tired during a Departmental seminar without biologically evolving during that lifetime. Those changes may alter later selection conditions. They are not themselves changes in inherited population variation.

This gave me a reference case with three properties I wanted the other strands to confront.

First, the changed object and population could be stated. Second, continuity across time had a mechanism that did not depend on the investigator's preferred description. Third, the account could distinguish persistence from improvement. Genetic drift can fix a variant without discovering that it is wise.

The source asked less of Organization and Extension. It required retained organizational change or changed usable capability, not genes, reproduction, or one universal fitness function. This could be legitimate pluralism. It could also be definitional generosity. If the admission conditions became too different, the shared noun would cease to constrain anything.

I therefore replaced the Department's biological form with a weaker test. A candidate evolutionary strand did not have to reproduce biology. It did have to answer five questions: what changed; what made the change persist beyond one event; how later states inherited the changed condition; what mechanism altered or retained variants; and what evidence would distinguish cumulative change from temporary response.

Natural Selection passed easily.

Organization requested clarification of the word inherit.

Organization Failed the Replicator Field

The difficulty with Evolution by Organization was not that social arrangements fail to persist. They persist with impressive determination, including after the people who remember why they were created have left.

Organizational-learning research has long described organizations as history-dependent and routine-based. Experience can be encoded in routines, procedures, interpretive frameworks, records, roles, and memory so that later participants act under conditions produced before they arrived. Those arrangements can survive personnel turnover, be taught, imitated, copied across sites, modified in performance, or abandoned.^[7–9] Cultural-evolution research provides an even broader population account: socially learned behaviors, ideas, skills, institutions, and technologies can be transmitted, vary, spread, drift, and interact with genetic evolution. The field explicitly treats similarities to genetic evolution as useful without denying differences such as horizontal transmission, guided variation, blending, and transformation during transmission.^{[3, 4]}

There was therefore no empirical difficulty establishing social inheritance in the broad sense. The problem was its plurality.

A genetic variant does not usually survive because a committee votes to retain it. An institutional rule can. A routine may persist because it works, because it is required, because its users have forgotten alternatives, because changing it would be expensive, because a powerful actor benefits from it, because everyone else expects it, because the software encodes it, or because the organization has placed it in a manual and trained three generations of employees to regard the manual as ancestral weather.

These mechanisms do not become one process merely because the result persists.

Generalized Darwinian approaches have argued that variation, inheritance or retention, and selection can provide an abstract evolutionary framework for organizations even when variation is intentional and acquired characteristics are transmissible. This program has generated useful work on routines, organizational populations, adaptation, and selection.^[5] Its critics have asked whether the relevant social populations, reproduction relations, and inheritance structures satisfy the ontological conditions that make Darwinian explanation more than analogy, and whether differential retention rather than differential reproduction would require a different theory.^[6]

Organizational ecology supplies one established route for evolutionary analysis of populations of organizations, especially where founding, persistence, structural inertia, and mortality can be studied at the population level. I did not treat that route as equivalent to every case the source placed under Evolution by Organization.^[9]

The dispute was useful precisely because neither side permitted the word evolution to settle itself.

Triadic Evolution took a different route. Evolution by Organization was broader than cultural selection or organizational ecology. It named change in social organization itself: relations, roles, obligations, authority, institutions, routines, norms, culture, collective memory, education, law, and learning. It allowed biological development where social organization was genetically mediated, social learning where behavior was transmitted, and deliberate reorganization where responsible actors changed their arrangements intentionally.^[1]

This breadth made the category easy to populate and difficult to mechanize.

I constructed three cases.

In the first, a hospital revised an emergency handoff after a near miss, trained staff, changed the electronic workflow, and preserved the new process across shifts and subsequent personnel turnover. Organizational learning had occurred. The retention route was reconstructable.

In the second, a committee issued a new rule on Friday, discovered on Monday that nobody had changed conduct, and quietly returned to the previous practice. A document had changed. Organization had not yet acquired a retained new state.

In the third, every member of a work group independently moved away from a leaking pipe. Coordination was visible. No role, routine, obligation, or socially inherited arrangement had changed. The response did not become Evolution by Organization merely because several people behaved similarly.

The category could discriminate these cases, but only by imposing a retention condition. The source already approached this condition when it said that organizational change must alter later collective activity and that a new rule counts only when it changes organized conduct rather than merely being announced.^[1]

I accepted that much.

Then the insects arrived.

The Colony Was Counted Twice

Triadic Evolution had chosen social insects as a limiting case, which was considerate because limiting cases are where definitions stop receiving hospitality.

In ants, bees, wasps, termites, and other eusocial insects, colony organization can include reproductive division of labor, caste differentiation, communication, task allocation, and coordinated defense or foraging. Genetic, developmental, endocrine, epigenetic, environmental, and social mechanisms contribute in species-specific combinations. Modern work on social insects has made the molecular and developmental basis of social organization increasingly tractable while preserving substantial complexity.^[10]

The source's claim was precise. When heritable biological change alters colony organization, Evolution by Natural Selection identifies the population mechanism through which inherited variation changes. Evolution by Organization identifies the persistent change in colony organization and its effects on collective capability. In this limiting case, the book says, organizational continuity may be reproduced primarily through biological development rather than a separate social-inheritance system. The same event therefore supports two analyses without implying two independent inheritance processes.^[1]

I read the passage several times because it contained both the defense and the objection.

The defense was strong. A genetic account of caste-related variation does not by itself describe the resulting distribution of work, communication, reproductive roles, or colony capability. An analyst can reasonably ask both how inherited biological variation changed and how the social arrangement changed. The two descriptions select different dependent variables.

The objection was stronger than I expected. If the organizational arrangement changes only because a biological inheritance process produces a different social phenotype, what exactly is the second evolutionary process?

A changed object is not necessarily a second mechanism.

Suppose a heritable change in body size alters which workers perform a task, and the new task distribution persists because the developmental program is inherited biologically. The colony organization has changed. But if no socially transmitted practice, retained role structure, learned convention, or other organizational retention route exists independently of that biological development, calling the outcome Evolution by Organization risks counting one evolutionary process twice: once by mechanism and once by phenotype.

The source itself avoided the strongest version of this error by saying explicitly that two analyses do not imply two independent inheritance processes. Yet its formal language still called both analyses evolutionary strands.

I filed the problem.

The correction did not eliminate Evolution by Organization. It divided its cases.

Where social learning, teaching, roles, routines, institutions, records, law, or other socially constituted mechanisms carry arrangements beyond the originating participant, there is an additional inheritance route. Cultural-evolution research already studies such routes across humans and other animals, and evidence for socially learned traditions is broad even though cumulative culture varies greatly among species.^{[3, 11, 12]} Human institutions add unusually durable symbolic and organizational mechanisms: records, offices, procedures, credentials, legal rules, formal education, and deliberate redesign.

Where no such route exists, the organizational analysis can still be scientifically necessary. It tells us what changed in collective arrangement. But the book must not borrow a second process merely from a second description.

This was the first substantive narrowing produced by Chapter 4.

The Department asked whether one of the three evolutions had therefore been downgraded to a dimension.

I replied that I had not yet decided which noun was being downgraded.

The Department created a temporary category labeled DIMENSIONAL EVOLUTION.

I asked it not to.

Extension Changed During the Definition

Evolution by Extension produced a different problem because artifacts are unusually cooperative with historical narratives.

A stone tool has predecessors. A ship has design generations. Software has versions. Technical standards spread, compete, stabilize, and sometimes disappear. Components are recombined. Designs are copied across firms and populations. Infrastructure makes later choices depend on earlier ones. Repair and maintenance preserve capabilities that would otherwise vanish. In cultural-evolution research, technology is often treated as a major domain of cumulative culture, and evolutionary approaches can model populations of artifacts, transmission, adoption, variation, and selection.^{[13, 14]} Evolutionary economics likewise treats technological change as path-dependent, population-like, and shaped by routines, selection environments, and cumulative knowledge.

This established that evolutionary analysis of technology was not a decorative invention by Triadic Evolution.

It did not establish a separate Evolution by Extension.

The source defined an extension by function and association with an actor. A capability-bearing physical, digital, or managed ecological means enlarges what an organism or responsibility-bearing actor can perceive, remember, calculate, construct, move, withstand, communicate, or reach. The strand changes such means through creation, adoption, combination, modification, maintenance, transfer, and abandonment. A prototype that demonstrates possibility but is never adopted or retained may not matter evolutionarily at the relevant institutional or population scale. A software update that changes files but not usable capability may likewise be a technical event without a retained extension-mediated adaptation.^[1]

This was already more discriminating than a chronology of inventions.

The difficulty was dependence on Organization. Artifacts do not usually decide to reproduce. Designs are copied because people teach, imitate, manufacture, regulate, finance, standardize, mandate, purchase, repair, or steal them. A technical lineage may persist because an interface became dominant, because installed infrastructure makes replacement expensive, because a regulator requires compatibility, or because a design remains better for a particular task. The mechanism of persistence often runs through social organization even when the object retained is technical.

This does not make the technical object unreal. It makes the boundary between cultural evolution and extension evolution expensive.

Cultural-evolution scholarship commonly includes artifacts and technologies among cultural variants. From that perspective, a cheese strainer, navigation method, or software architecture can be part of cultural inheritance because its design is socially learned and transmitted. Triadic Evolution proposes to separate the socially constituted practice from the capability-bearing means that carries or enables it. The distinction can be useful. It is not the only scientific way to partition the case.^{[3, 13]}

I tested the category with four objects.

A prototype sensor was built once, demonstrated in a laboratory, and discarded. It increased knowledge about what could be built but never became an operational capability. Calling its brief existence Evolution by Extension at the organizational scale added little.

A second sensor was deployed, maintained, copied, and incorporated into a monitoring practice. Later versions changed sensitivity and operating range while preserving interface compatibility. The extension lineage and its retained capability were clear.

A software application changed version number without altering any function used by the actor. The Department argued that version numbers were explicit generations. I argued that administrative numbering is not descent with modification merely because the dots are hereditary.

A bridge remained unchanged for eighty years while the city reorganized around it. The extension had not changed much, but its persistence had altered later conditions profoundly. Evolution by Extension could not be reduced to novelty or rate. Retention mattered as much as modification.

The strand survived these cases if it was treated as change in retained capability-bearing means, not as a synonym for technical history.

Then I gave it a beaver dam.

The Dam Did Not Settle the Matter

Nonhuman extensions were useful because they removed the excuse of human design.

A beaver dam, bird nest, web, shelter, cache, trail, or tool can alter what an organism is able to do and the conditions encountered by later organisms. Niche-construction research has made the broader point familiar: organisms modify environments, and those modifications can feed back into development and selection. The extended-phenotype tradition provides another route for analyzing effects beyond the organism's body. Triadic Evolution acknowledges both lineages and uses extension more narrowly for a constructed or appropriated capability-bearing means associated with an actor.^[1]

The Department considered the dam decisive evidence for Evolution by Extension because the dam was neither a gene nor a committee.

This was insufficient.

A dam can be biologically generated behavior embodied in durable material. Its effects can persist beyond the individual builder. That persistence can change environments and selection pressures. None of this requires the dam itself to constitute a lineage of evolving extensions. The evolutionary explanation may still reside primarily in biological inheritance plus ecological inheritance produced by niche construction.

The case becomes different when construction changes through experience, social learning, retained modifications, or environmental inheritance that materially alters later construction. Evidence from animal culture shows socially learned traditions in foraging, movement, communication, and tool use across multiple taxa; research on cumulative cultural evolution asks specifically whether modifications are retained and built upon rather than merely repeated.^{[11, 12]} In those cases an additional transmission route can be investigated.

The same rule from the insect case returned in another form.

Persistent external consequence was not enough. I needed to know what was being inherited and by which route.

A dam could therefore be an extension in the architectural sense while its construction history remained biologically evolved, socially learned, individually learned, repeatedly improvised, or some combination. Extension described the capability-bearing object. Evolution by Extension required evidence that the capability-bearing means itself entered a retained process of creation, modification, transfer, maintenance, or abandonment at the selected scope.

This distinction was not present in the Department's form because the Department had already decided that artifacts were metaphors.

It was now present in mine.

Change Was Not Enough

By this point I had produced a predictable bureaucratic outcome: a concept called evolution now required more fields than before I began simplifying it.

The additional fields were nevertheless doing scientific work. They prevented transient change, duplicated description, and persistence without a stated retention route from receiving automatic evolutionary status.

I converted them into an eligibility record. It was not part of Triadic Evolution. It was my operational test of the framework's terminology.

Natural Selection did not require every field in this wording, because population genetics possesses stronger and more specific ones. The record was designed for cross-strand comparison, not as a replacement for disciplinary methods.

Applied to Organization, the record excluded a new policy that never altered conduct, an individual preference that remained private, and a one-time coordinated response with no retained social arrangement. It admitted a revised safety practice preserved through roles, training, routines, and later enactment. It admitted a socially learned tradition when the transmission route could be shown. It did not require the organizational variant to be beneficial. An institution can inherit an error.

Applied to Extension, the record excluded a technical novelty that never became retained capability at the selected scope. It admitted a tool, infrastructure, or software change when the altered capability persisted through maintenance, copying, adoption, standardization, installation, or another reconstructable route. Again, persistence did not prove superiority. An inferior interface can become impressively ancestral.

The non-duplication field created the greatest difficulty. Human cultural and technical change are often entangled because the design of an artifact may be socially learned, while the artifact changes what can be learned next. A written standard is socially authoritative because an institution recognizes it and technically persistent because a medium preserves it. The same historical transition can therefore contain two changed objects and several retention routes.

The rule was not to assign one event to one strand. It was to prevent one route from being counted twice without saying so.

This was close to the source's own insistence that one event may involve all three strands while each retains its primary object and mechanism.^[1] Chapter 4 added a stricter consequence: if the object changes but no distinct mechanism or retention route can be reconstructed, the analyst should record the changed dimension without claiming an independent evolutionary process.

The framework had become harder to use.

This was the first result of the chapter that I trusted.

Three Clocks Refused Synchronization

The next claim appeared empirical enough to damage the framework cleanly.

Triadic Evolution proposes that organizational and extension-mediated adaptation can often proceed on shorter timescales than human genetic adaptation. It also spends unusual effort limiting the proposition. Biological change can be rapid under strong selection. Institutions can remain rigid for centuries. Infrastructure can be locked into standards for decades. A rule issued quickly may never become retained practice; a software release deployed quickly may not alter usable capability. The source therefore requires specified populations, environments, adaptive problems, observables, and comparable periods. If organization or extension is not faster in a stated case, the corresponding timescale proposition fails there.^[1]

This was scientifically responsible and administratively difficult.

The Department requested one number for each strand.

BIOLOGICAL GENERATION LENGTH was available.

ORGANIZATIONAL GENERATION LENGTH produced disagreement over whether a generation meant leadership turnover, employee cohort, budget cycle, policy revision, institutional succession, or the interval between two consultants renaming the same process.

EXTENSION GENERATION LENGTH produced model years, release cycles, replacement intervals, patent families, hardware generations, and one software team that deployed seventeen times during the meeting.

There was no common second hand.

The error was comparing clocks rather than responses. If a hazard appears at time T0, the relevant question is not which strand has the shortest abstract generation. It is how long each available mechanism takes to produce a retained change in the outcome under investigation.

A biological response might be measured as inherited population change relevant to exposure or viability. An organizational response might be measured as a new role, rule, routine, or practice that changes conduct and remains in force. An extension response might be measured as a deployed and maintained change in usable capability. These observables are still not identical, but they can be related to one stated adaptive problem.

I proposed a matched-response design rather than a universal speed ranking. For a specified population and pressure, define the outcome that matters, specify the candidate biological, organizational, and extension-mediated routes, and measure time to a retained response meeting a predeclared criterion. The design could then fail. A slow institution might lose to a biological change in a rapidly reproducing organism. A hardware-dependent response might take longer than an organizational workaround. A policy might be announced in hours and become operational in years.

No numerical relationship among the strands followed in advance.

The source had already conceded most of this. The remaining value of the timescale hypothesis was therefore modest but real: a sociotechnical species possesses several retention mechanisms whose response times can differ dramatically under the same pressure, and the difference should be measured through retained consequences rather than declarations of novelty.

I recorded the result.

The Department asked whether a hypothesis that sometimes fails and sometimes does not was useful.

I replied that this property is generally considered an improvement over one that cannot fail.

The Department marked my response as attitude.

Milk Had Several Histories

I needed one historical case in which all three strands could be made to compete with simpler accounts.

I selected milk.

This was not because milk is inherently triadic. It was because lactase persistence and dairying are among the most developed examples of gene-culture interaction, while archaeology provides unusually direct evidence that material processing technologies mattered. If the three strands were merely three labels attached to a familiar gene-culture story, the case should expose the redundancy.^[15–17]

The biological record was the clearest. In many European populations, the ability to digest lactose into adulthood rose from low prehistoric frequencies to become common under strong selection during the last several thousand years. The relevant variants, population trajectories, and selection models belong to evolutionary genetics. Nothing in Triadic Evolution adds to that mechanism.^[15]

The old simple story was attractive: humans began dairying; milk created a nutritional advantage for adults who could digest lactose; lactase-persistence alleles therefore increased. The broad gene-culture relation remains real, but recent evidence has made the causal history less neat. A large 2022 study combined ancient DNA with roughly seven thousand pottery lipid residues from more than 550 archaeological sites. Milk use was widespread across Europe for millennia, yet variation in milk exploitation did not by itself explain the trajectory of lactase-persistence frequencies better than a model of uniform selection. Population fluctuation, settlement density, and wild-animal exploitation—used as proxies for famine and pathogen exposure—better accounted for the intensity of selection in the authors' modeling. The authors proposed that drinking milk while lactase non-persistent may have become especially costly under famine or high pathogen exposure.^[15]

The case therefore punished retrospective fit. Dairying did not simply enter the environment and instruct a gene to become common.

The organizational record was more difficult to recover because potsherds do not preserve committee minutes. Dairying required recurrent social practices: keeping and managing animals, deciding who milked them, transmitting food practices, allocating products, preserving knowledge, and carrying the practice across generations. Cultural-evolution and gene-culture-coevolution research already treat such socially transmitted behavior as part of the causal environment in which genetic evolution occurs.^{[3, 4]} The archaeological evidence could establish use and processing more strongly than it could reconstruct exact roles, authority, or institutions. I kept that limitation visible.

The extension record was materially clearer. Pottery residues provide evidence of milk processing over long periods, and specialized perforated vessels from sixth-millennium BCE northern Europe carry chemical evidence consistent with separating curds from lactose-containing whey. Processing milk into lower-lactose products could preserve and transport food while making dairy products more usable by lactase-non-persistent populations.^[16] Later biomolecular work has likewise identified curd-rich dairy processing in Neolithic ceramics.^[17]

This technical capability complicated the biological story in the opposite direction from the one a simple three-arrow diagram suggested. An extension that reduced lactose exposure could allow dairying to persist among people without lactase persistence. It might therefore buffer the selective pressure associated with fresh milk in some contexts rather than merely intensify it. The interaction depended on diet, pathogens, scarcity, processing, demography, and local practice.

The case could be written as three histories.

One history tracked inherited human variation and population-level selection.

A second tracked socially transmitted subsistence and food practices whose meanings, norms, and organization could persist independently of any one participant.

A third tracked material processing capabilities—vessels, strainers, production methods, storage and transport possibilities—that changed what those practices could accomplish.

Separating the histories prevented one useful correction. Culture was not simply an environmental variable acting upon genes. Material capability altered what the cultural practice could do, while social organization determined how the capability was used and transmitted. The extension could change exposure without becoming the social meaning of dairying. The organization could preserve dairying without becoming the ceramic vessel. Biological selection operated under the resulting conditions without adopting either purpose.

This was a good Triadic reconstruction.

It was not yet a uniquely Triadic explanation.

Gene-culture-coevolution and niche-construction accounts already connect socially transmitted practices, environmental modification, and genetic selection. Cultural-evolution scholarship can include technologies among transmitted cultural variants. Archaeologists routinely integrate artifacts, subsistence practices, and biological evidence without requiring three capitalized strands. The framework could claim no discovery of the coupled process.^{[3, 15]}

Its incremental contribution was narrower. It forced me to keep the socially constituted practice and the capability-bearing medium in separate columns long enough to ask whether they changed by different routes and at different times. In the milk case, that separation mattered because processing technology could alter the biological consequences of the practice without being reducible to the practice itself.

The distinction earned analytical interest.

The word evolution remained on probation.

Selection Was Not a Universal Verb

The most dangerous word after evolution was selection.

Natural selection has a specific population meaning. Organizations also select candidates, projects, standards, suppliers, designs, and strategic options. Markets select in looser senses. Engineers select components. Regulators select permitted conditions. Algorithms select ranked outputs. The grammar encourages unity where the mechanisms may have nothing in common.

Triadic Evolution itself warns against this transfer. A human choosing a component is not performing natural selection merely because both sentences contain select.^[1] The warning had to apply equally to my own analysis.

Generalized Darwinian approaches offer the strongest case for abstraction. They argue that wherever populations of varied entities replicate or are retained differentially, principles of variation, inheritance, and selection can support explanation even when human intentionality affects variation. This can connect organizational ecology, routines, competition, and strategic adaptation without demanding literal genes.^[5] Cultural evolution likewise obtains genuine explanatory leverage from population thinking, transmission biases, drift, migration-like processes, and selection models while explicitly modifying the biological assumptions.^{[3, 4]}

The strongest objection is not that analogy is forbidden. Scientific models are allowed to travel. It is that the traveling terms must continue to constrain explanation.

If an organizational practice persists because law compels it, the mechanism is not illuminated by saying the practice was selected unless the account specifies who or what performed the selection, what alternatives existed, what was retained, and how the resulting population changed. If a technical standard persists because switching costs and network compatibility make replacement prohibitively expensive, calling the standard fit can be a compact summary or a concealment of the actual mechanism. If a rule persists because powerful participants suppress alternatives, survival is not evidence of adaptation to the interests of the organization as a whole.

The same caution applies in biology. Persistence alone does not prove selection; drift and demographic history can produce change without adaptive advantage.^[2] The biological field had already learned the lesson that the social fields were being accused of forgetting.

I therefore refused to require a single selection mechanism for all three strands. Doing so would either exclude purposive organizational and technical change arbitrarily or stretch selection until every form of retention qualified.

Instead, selection became one possible mechanism inside some organizational and extension histories, alongside deliberate evaluation, coercion, imitation, institutionalization, standardization, path dependence, maintenance, regulation, random loss, and other routes. The mechanism had to be named rather than assumed from the strand label.

This made the three evolutions less alike.

It also made them more defensible.

The Word Evolution Sent an Invoice

At this stage the framework had survived by surrendering the possibility that its three strands were instances of one generalized mechanism.

This could be interpreted as intellectual retreat. It could also be what the source had claimed from the beginning.

Triadic Evolution did not say that natural selection, social reorganization, and technical development were one process. It said that a sociotechnical species changes through three interacting strands whose primary objects and mechanisms differ. Their unity lies in the changing composite species and in the reciprocal conditions they create for one another.^[1]

The chapter had nevertheless altered what that claim could responsibly mean.

Evolution by Natural Selection retained ordinary biological standards. It required no Triadic permission to exist.

Evolution by Organization remained defensible where a social arrangement changed and the altered condition became available to later collective action through a reconstructable organizational-retention route. This could include social learning, roles, routines, institutions, records, law, or deliberate reconstitution. Where organization changed only as a phenotype carried wholly by biological inheritance, the organizational description remained useful but did not by itself establish a second inheritance process.

Evolution by Extension remained defensible where capability-bearing means changed in a way that was retained, transferred, maintained, reproduced, recombined, or abandoned so that later action inherited a different capability environment. A new object was not enough. A new version number was not enough. A durable artifact could matter even without rapid modification, and the organizational processes governing its adoption had to remain visible rather than being silently relabeled technical selection.

The common word evolution therefore earned only a limited function.

It instructed the investigator to look beyond immediate change and ask what remained, how it remained, what later state inherited the consequence, and how the retained condition altered subsequent possibilities. It encouraged lineage, path dependence, transmission, retention, and interacting timescales to remain visible across biological, organizational, and technical histories.

This was more than rhetoric when those questions changed the reconstruction. It was rhetoric when change received evolutionary status merely because it lasted long enough to sound historical.

The costs were substantial.

The first was borrowed prestige. Evolutionary biology is among Earth's most successful scientific frameworks. Placing institutions and technologies beside natural selection can make their categories appear more explanatory than the evidence warrants.

The second was false symmetry. Natural Selection, Organization, and Extension do not have equivalent populations, inheritance units, generation structures, selection mechanisms, or standards of evidence.

The third was double counting. One biological change can alter organization; one social transmission process can carry a technical design. Different changed objects do not guarantee different evolutionary mechanisms.

The fourth was progress leakage. Human discourse often treats technological evolution as advancement and organizational evolution as improvement. Biological evolution provides no such guarantee, and neither should the other strands. Institutions can inherit cruelty. Extensions can preserve fragility. Capability can increase while resilience decreases.

The fifth was clock abuse. A software release and an allele-frequency shift are not automatically comparable merely because both occur between dates.

The sixth was mechanism concealment. If the strand label replaces the specific process—teaching, institutionalization, network effects, standardization, drift, regulation, maintenance, selection—then the framework has made explanation shorter by removing the explanation.

I placed the gains and costs side by side and did not score them. Chapter 3 had already demonstrated the hazards of invented numerical confidence.

The invoice was real. So was the service.

The Department asked whether I wished to cancel the word evolution and retain three kinds of change.

I considered this seriously.

The evolutionary biologist did not require the same interval.

“Do it,” she said.

I replaced the three labels in my notes. Evolution by Natural Selection became biological evolution. Evolution by Organization became retained organizational change. Evolution by Extension became retained capability change. I then reconstructed the milk case again.

Nothing factual disappeared. Human populations still contained inherited variation affecting adult lactose digestion. Dairying practices were still socially learned, preserved, altered, and transmitted. Pottery, processing methods, containers, fermentation, storage, and transport still changed what forms of milk could be used and how long they remained available. The retained states still altered one another's future conditions. The organization preserved ways of obtaining and using milk. The extensions altered what forms of milk could be consumed and preserved. Those practices and capabilities altered the environment under which biological selection occurred. Biological differences, in turn, changed the consequences of later dietary practices.^[15–17]

The reconstruction had lost a common noun. It had not lost a mechanism.

This was damaging.

I had argued that calling the history merely change was accurate and unhelpful. The biologist pointed out that I had not replaced the strand names with merely change. I had replaced them with more specific descriptions of retained processes.

“Gene-culture coevolution already asks how socially transmitted practices alter selection conditions,” she said. “Niche construction already asks how organism activity changes inherited environments. Cultural evolution asks about transmission and retention. Technological-evolution research asks about lineages, adoption, and path dependence. Which scientific question becomes unavailable if your two capitalized Evolutions become organizational change and capability change?”

I had no defensible answer of the form: this mechanism disappears.

The shared noun was not necessary to identify lactase-persistence selection. It was not necessary to identify socially transmitted dairying. It was not necessary to identify ceramic processing, fermentation, storage, or technological persistence. It was not necessary to show reciprocal causation among them. The literatures I had already cited supplied those possibilities without Triadic permission.^{[3, 4, 13–17]}

I therefore withdrew the strongest defense of the word.

Evolution did not earn its place here as a common causal mechanism. Nor had I shown that it was an indispensable explanatory term. The source's three strands could be scientifically legitimate and still be named more conservatively.

The question that remained was architectural rather than causal: did the shared noun impose a useful discipline on cross-domain inquiry?

Under the eligibility record, an analyst who accepted an evolutionary reading of Organization or Extension could not stop at event-level change. The analyst had to identify succession, a retained state, a route by which later states inherited that condition, and consequences for later action. The common framing also invited comparison among different clocks and asked how one retained process altered the future conditions of the others. Those questions could prevent organizational and technical change from being treated as background while biological change alone received a history.

But the biologist's objection remained. None of those questions belonged uniquely to the noun evolution. An interdisciplinary team could ask them directly. Gene-culture coevolution, niche construction, cultural evolution, organizational learning, evolutionary economics, and technological history already did much of the work in their own vocabularies.^{[3–6, 13–17]}

The difference, if any, was therefore procedural. Triadic Evolution placed an obligation on one investigator to carry three unlike histories in the same record and to ask of each one what persisted into the next relevant state. It made retention and inherited conditions a common comparative demand without asserting that the mechanisms were equivalent.

That could be useful. It was not yet demonstrated to be necessary.

The milk case did not save the word evolution. It showed the work the word was trying to coordinate.

Calling all of the history natural selection remained false. Calling each component an independent Darwinian evolution remained false. Calling the nonbiological parts retained organizational change and retained capability change remained scientifically adequate. Calling them evolutionary strands could be retained as research-program vocabulary only when the framing changed what the analyst compared, what continuity had to be reconstructed, or what evidence was sought.

Where it did not, the noun received no credit.

I allowed the source to keep its term. I stopped treating the term itself as a finding.

The Three Refused to Merge

I returned to Form 31-E and entered all three candidates again.

Evolution by Natural Selection supplied inherited biological populations, specific population mechanisms, and non-intentional differential outcomes.

Evolution by Organization supplied social arrangements whose retention could occur through biological development, social learning, routines, institutions, records, law, authority, practice, and deliberate redesign. The mechanisms were plural, and some cases were not independent of biological inheritance. Where no retained organizational route existed, I refused to infer one from the changed arrangement alone.

Evolution by Extension supplied capability-bearing means whose histories could involve design, copying, adoption, maintenance, recombination, standardization, infrastructure, transfer, and abandonment. The persistence mechanism often depended on organization, but the object of change and the capability consequences could remain analytically distinct.

The form requested one common mechanism.

I entered None required by source claim.

It requested one common unit of inheritance.

I entered None established.

It requested one common fitness measure.

I entered Inappropriate.

It asked what justified the shared classification.

I entered Common research scope, not common mechanism: retained change in distinct primary objects whose altered states become inherited conditions for later sociotechnical action, with mechanisms and evidence stated separately.

The field accepted the sentence. It then warned that the answer was too long for a category.

This was correct.

The three strands had survived by refusing to become one process. Natural Selection did not become purposeful. Organization did not become population genetics. Extension did not become an artifact species. Their interaction remained the framework's claim; mechanistic equivalence did not. The shared noun had survived only as conditional research-program vocabulary, not as an independently demonstrated explanatory necessity.

This result was weaker than a general theory of evolution and also weaker than a demonstration that all three labels were indispensable. It left the framework with a defensible research architecture if analysts obeyed the admission rules: identify the changed object, state the retention route, preserve the specific mechanism, define the scope and time horizon, do not infer an additional evolutionary process merely because another dimension of the event changed, and award the common evolutionary framing no credit when more specific vocabulary performs the same work at lower cost.

The result also created a new vulnerability.

A category becomes scientifically interesting when it excludes something.

Before Chapter 4, Organization and Extension could absorb almost any durable social or technical change if the analyst was generous enough. After the eligibility test, temporary response, one-off novelty, unretained policy, duplicate description, and persistence without a reconstructable route could be refused.

The rejected material accumulated in a folder labeled OTHER.

The Department noticed the folder immediately.

It requested clarification.

I explained that the exclusions did not yet constitute a fourth evolutionary strand. They were simply cases the first three should not absorb.

The Department renamed the folder POSSIBLE FOURTH THING.

I objected that the title prejudged the next investigation.

The Department replied that the folder was only provisional.

I had used the same defense several times.

I accepted the folder.