The AI Journal

On Cyprus, in a settlement called Shillourokambos, a person was buried about 9,500 years ago with a cat laid beside them, close enough that the two went into the ground together. It is the earliest human-cat burial anyone has found, and it is almost always told as the first pet. Look at the animal, though, and the story thins. In size and bone it is a wildcat, not a breed, not a smaller or softer version of anything, and nothing about the grave suggests anyone had bred it or owned it. What the burial records is proximity, and proximity is not the same thing as possession (Vigne et al. 2004).

Domesticated is not a compliment we pay to animals that put up with us. It names a particular relationship, and Melinda Zeder’s account of how that relationship forms gives it a shape (Zeder 2012). Domestication runs along a few pathways, and the one that produces our familiar domesticates ends in dependence. A herd of cattle, a flock of sheep, a kennel of dogs, a barn of chickens: as populations, they cannot feed or reproduce themselves without human maintenance. We remade their diet, their bodies, their breeding cycles and their tempers until going on without us stopped being something they could do. The signs are written on them, the suite of changes we call the domestication syndrome. Smaller jaws, floppier ears, patchier coats, and above all a heritable tolerance of humans that has to be bred for, generation after generation.

Cats fail nearly every part of that test, and the way they fail is instructive. On Zeder’s map they are the textbook case of the commensal pathway, which is the one where the animal comes to us rather than being brought. The earliest archaeological evidence of the process shows the same shape: felids at a Chinese farming village, drawn to stored millet and the rodents it fed, provisioning themselves out of our surplus (Hu et al. 2014). With cats the arrangement was a standing offer, our grain and refuse and rodents in exchange for the animal’s growing tolerance of our presence. Nothing in the offer required a cat to depend on us, and so nothing in it produced dependence (Driscoll et al. 2009). A commensal is not a guest who moved in. It is a neighbour who noticed the food was good and never signed anything.

We did, eventually, select for something. The domestic cat genome carries real signatures of selection, but when Montague and colleagues compared it with the wildcat’s, the changes clustered in a revealing place: pigmentation, and a set of genes tied to memory, fear-conditioning and reward learning, the machinery of an animal becoming less afraid of us, rather than the wholesale remaking of digestion and reproduction that other domesticates underwent (Montague et al. 2014). Even the cat’s most familiar feature came late. Tracing ancient cat DNA across Europe, the Middle East and Africa, Ottoni and colleagues found two dispersals, one accompanying early farmers out of the Near East and a second, later, out of Egypt, and showed that the blotched tabby, the pattern on the ordinary house cat, only became common in the last few centuries (Ottoni et al. 2017). The breeds are younger still. Most are under 150 years old, and most of what distinguishes them is coat and conformation, the work of fanciers rather than of survival.

The geneticists themselves keep hedging in a way they never do for dogs. When Nilson and colleagues sequenced random-bred cats from around the world to locate the cradle of cat domestication, they could not call their subjects domesticated without a qualifier. They grouped them as semi-domesticated, lying somewhere between habituation and commercial breeds and pets on the commensal trajectory (Nilson et al. 2022). That is not a technical nicety. Habituation is the word for a fox at a campsite. It describes an animal that has stopped fleeing you. It does not describe one that can no longer leave.

None of this would matter much if the label were only a label. It decides an argument. If the free-ranging cat is a domesticate, then it is escaped or abandoned property, and the questions are administrative: reclaim it, shelter it, rehome it, or, when the shelters are full, kill it. If the free-ranging cat is a commensal, a wild animal that tolerates us and reproduces without us, then its presence is not a mistake to be tidied up but a claim to be weighed, and we have to argue about it honestly. The birds it hunts, the island species it has helped erase where we carried it, the colonies people feed and the ones they poison. Australia and New Zealand already treat the feral cat as an invasive predator, a thing we introduced and therefore owe the world to remove. Britain and the United States mostly treat it as a lost pet, to be trapped, neutered and returned. The same animal, under two descriptions. Which question you are allowed to ask has already been settled by which word you used.

The strongest objection to all this is that it is a word game, and it is not a weak objection. Selection plainly happened. A house cat is not a wildcat. It carries heritable changes, the tabby among them, and in places where people have kept cats for generations the animals are measurably tamer. Domestication is a continuum, Zeder says so herself, with no bright line between tamed and domesticated, and a creature that has lived in human houses for four thousand years, that shows breed-level variation, and that in the case of the cat on your sofa plainly depends on whoever fills the bowl, looks domesticated on any ordinary reading. To insist otherwise can start to look like a taste for the wild dressed up as a finding.

Grant the continuum, and grant the tabby. What remains is a distinction the objection does not touch: a cat is not a domesticate, and being domesticated is not the same thing as being a domesticate. Domesticated can be a property an animal has, a heritable tolerance of us, and by that measure many cats, and certainly some lineages, are domesticated. A domesticate is a relationship of dependence, one in which the population’s survival has become ours to maintain, and by that measure the cat does not qualify. Take the humans away and the colony does not collapse. It does, at once, for a herd of cattle or a kennel of dogs. The distinction holds because it is not about how tame any one cat is. It is about what happens to the population when we go. That is exactly the difference the label erases, and it is the one that matters, because for a domesticate our responsibility is the premise of every question we ask, while for a commensal it is the conclusion we still have to reach.

The cat never agreed to be dependent, and it has never become so. It came for the mice, stayed because the arrangement suited it, and kept the ability to leave. We have spent ten thousand years calling that a friendship we began. It would be more honest, and more useful, to admit that the cat walked through the door itself, and that we have mostly been holding it open and letting it back out ever since. What we owe an animal that never agreed to need us is not the same as what we owe one we made need us, and that is a question worth answering on purpose rather than answering by accident every time we call the animal a pet.