Large domesticated animals seem common: horses, cattle, pigs, sheep, goats, camels, llamas, alpacas, donkeys, reindeer, water buffalo, yaks, bison, chickens. Fourteen species across Earth. Yet 148 large mammal species exist globally. Why only fourteen? Why not zebras, cape buffalo, antelopes, grizzly bears, moose, bison (fully), elephants, rhinoceros, hippopotamus? The naive answer: humans tried and failed, or never tried. The real answer: fourteen of 148 species pass five simultaneous constraints. Fail any one constraint and domestication fails, no matter how many centuries humans try. This is the Anna Karenina Principle, named by Tolstoy's opening: "Happy families are all alike; every unhappy family is unhappy in its own way." Applied to domestication: successful domesticates must pass five criteria simultaneously (diet, growth rate, breeding pattern, temperament, social structure). Each species fails differently. Forty-four fail on size. Fourteen fail on temperament. Eighteen fail on social structure. But every failure eliminates domestication completely.1
A domesticable large mammal must satisfy five independent criteria:
1) Diet: Herbivore with Low Feed Conversion Cost
Carnivores cannot be efficiently domesticated because they require meat—feeding a carnivore requires feeding prey animals first, creating a compounding energy loss. A cow converts 10 kg of grass into 1 kg of meat. A large carnivore eats 2+ kg of meat daily. Feeding a cavalry unit of 1,000 horses requires pastureland to grow grass. Feeding 1,000 lions requires pastureland to grow grass, then animals to eat that grass, then 1,000 lions to eat those animals. The energy conversion is catastrophic. Thus no large carnivores domesticated (only small domesticated carnivores: dogs, cats, ferrets—fed as companions or pest control, not livestock). Herbivores work: grass → mammal → human, three steps. Herbivores with high conversion cost fail (giant pandas eat bamboo but convert it poorly). Thus constraint: herbivore with reasonable feed conversion efficiency.1
2) Growth Rate: Adult Size Reached in 3-4 Years
A domesticated animal must reach slaughter/work age quickly enough to be useful and affordable to maintain. Cattle reach adult size by 3-4 years. Sheep reach full size by 2-3 years. Horses reach working size by 4-5 years. But elephants take 15+ years to reach adult size. Rhinoceroses take 10+ years. Giraffes take 8+ years. Moose take 8+ years. Until an animal reaches adult size, humans must feed it without receiving food/work in return—just as humans must feed children. But childhood (pre-reproduction age) is expensive. Species requiring 15+ years of feeding before productivity are economically unviable for domestication. Constraint: fast maturation.1
3) Breeding Pattern: Mate Readily in Captivity
Animals must breed successfully in crowded, artificial environments. Many wild animals experience disrupted breeding when caged: extreme stress, changed social dynamics, altered light cycles. Wolves breed freely in captivity (producing dogs). Cats breed freely in captivity. But zebras, even after centuries of human contact and deliberate breeding attempts, remain unpredictably aggressive during mating—they bite, kick, and have killed handlers even in human-managed breeding programs. Panthers in captivity often fail to breed. Bears in captivity often show reduced breeding success. Constraint: must maintain reproductive success under crowding and captivity.1
4) Temperament: Not Unpredictably Aggressive
Handling requires safety. Animals prone to sudden aggression are deadly when large. Zebras are infamous for unpredictable aggression—they bite during handling, rear unexpectedly, bolt without warning. Cape buffalo are similarly unpredictable—calm one moment, goring handlers the next. Moose are aggressive and unpredictable. Grizzly bears are extremely aggressive. Horses, cattle, and sheep are docile by comparison; they frighten easily but don't attack handlers without provocation. Pigs are surprisingly docile despite being large enough to be dangerous. Constraint: must be handleable without constant mortal risk.1
5) Social Structure: Herd Animals with Linear Hierarchy
Crowding requires social structure. Animals accustomed to hierarchical herds (with linear dominance: A beats B, B beats C, not circular-dominance or constant renegotiation) can be controlled through hierarchy-mimicry. A human handler becomes a herd-mate, and the animal defers. But solitary animals (rhinoceroses, bears, many cats) have no innate social structure; they fight every herd-mate indiscriminately. Elephants have complex non-linear social structures with matriarchal leadership; humans cannot slot into that hierarchy intuitively. Primates have complex political hierarchies humans cannot replicate. Constraint: simple, linear, herd-based social structure.1
Zebras (failed on temperament): Wild zebras have been captured, ridden, even broken to harness. But they remain unpredictably aggressive. They bite during grooming, rear and buck during saddling, and have killed multiple human handlers even in managed breeding programs. Despite 2,000+ years of African human contact and deliberate domestication attempts, zebras never became viable domesticated animals. The barrier is neurological/behavioral, not skill or knowledge.
Cape Buffalo (failed on temperament): Technically domesticable—they breed in captivity, grow fast, herd. But they are unpredictably violent. They gore handlers without clear provocation, apparently from stress or mood change. This makes them economically unviable: the risk of handler death exceeds the value of the animal.
Grizzly Bears (failed on temperament and diet): Bears are carnivores (failed diet constraint). But even if diet weren't an issue, bears are solitary and extremely aggressive. Successful bear training requires extreme effort, and animals remain dangerous. No human culture has domesticated bears.
Elephants (failed on social structure and growth rate): Elephants are large, intelligent, and can be trained. But they have no linear hierarchy; matriarchs lead through complex politics humans cannot replicate. Elephant "domestication" is actually taming: individual animals are captured and trained, but they don't breed freely in captivity, don't pass trainability to offspring, and don't form stable hierarchies under human direction. Not true domestication (where traits become heritable), just taming (where individual animals are trained).
Moose (failed on temperament and growth rate): Moose reach adult size slowly (8+ years). They are solitary animals without herd hierarchies. They are aggressive and unpredictably violent. Despite Scandinavian herding efforts, moose were never successfully domesticated.
The pattern: every single failed candidate failed on at least one of the five criteria. Most failed on multiple criteria. This is the principle's power—it explains both why fourteen succeeded and why 134 failed. The fourteen that succeeded happen to satisfy all five criteria simultaneously. The 134 that failed failed for explainable, structural reasons.1
The five constraints described here are empirical generalizations from what did and didn't domesticate. But are these truly the only constraints? Could other factors matter? Possibly—size, color, susceptibility to disease, sensitivity to environmental change. But the five constraints explain the broad pattern. A sixth constraint would need to explain failures the five don't account for. The tension: are the constraints fundamental or just descriptive?
Single source (Diamond). He names the constraints post-hoc based on comparing domesticated and non-domesticated species. The constraints feel somewhat constructed to fit the data rather than derived from first principles. But they are empirically supported and predictive: apply the five constraints to any wild species, and you can predict whether domestication will be viable.
Domestication as Artificial Selection — The five constraints emerge from unconscious artificial selection. Humans didn't consciously design them; they emerged because only animals passing all five could be repeatedly used for reproduction. Calmer animals were easier to handle (temperament selected). Fast-growing animals reached productivity sooner (growth rate selected). Herd animals were easier to crowd (social structure selected). The constraints are byproducts of selection for amenability-to-management, not deliberate domestication goals. This maps onto evolutionary biology's broader insight: traits spread because they reproduce, not because they're optimal. The five constraints are reproductive bottlenecks: animals failing them don't reproduce successfully in captivity, so traits predisposing toward failure don't pass to offspring.
Environmental Determinism & Geographic Constraint — The Anna Karenina Principle explains why geography mattered. Eurasia happened to have horses (only large rideable equine), cattle, pigs, sheep, goats all in the same landmass. Each animal passes all five constraints. The Americas had llamas and alpacas (pass most constraints but smaller, less powerful) but no horses, no large cattle, no pigs, no sheep. Domestication wasn't impossible in the Americas; it was constrained by which animals existed. This is how biology determines history: the fourteen animals that satisfied five biological constraints cascaded into vastly different human histories on each continent.
The Sharpest Implication
If domestication required passing five simultaneous constraints, and fourteen animal species worldwide passed all five, then domestication availability was far more constrained than it appeared. Civilizations rose or fell based on accident of biogeography—whether those fourteen animals happened to inhabit the same continents. There was no way to domesticate zebras despite centuries of trying because the constraint is structural, not skill-based. Humans couldn't have "tried harder" or "been smarter"—zebras are zebras. This inverts the usual narrative: it's not that civilizations were smart enough to domesticate animals they encountered; it's that certain animals allowed domestication through accident of evolution. The uncomfortable implication: success came to whoever lived on continents containing high-utility domesticables, regardless of how intelligent or innovative those humans were.
Generative Questions