You come home. Your dog looks up at you and holds your eyes for a moment longer than seems necessary. You look back. Neither of you says anything, and something in the room feels settled.

There is a chemical reason for that feeling, and it is the same one that runs between a mother and a nursing baby.

Oxytocin is a small hormone made in the brain. It has a lot of jobs, but the ones most people know about are the family ones. It surges during childbirth. It flows during breastfeeding. It rises when a mother and her infant look at each other, and the more it rises in one of them, the more it rises in the other. Scientists call this a positive loop: eye contact boosts oxytocin, oxytocin encourages more eye contact, and the bond tightens with every cycle.

For a long time this was thought of as a within-species thing, a bit of biology that evolved to glue parents to their offspring. Then a group in Japan decided to test it on dogs.

Miho Nagasawa and Takefumi Kikusui at Azabu University, near Tokyo, had already noticed in a 2009 study that owners whose dogs stared at them a lot had more oxytocin in their urine than owners whose dogs did not. In 2015 they went further, and published the results in the journal Science.

Thirty dogs and their owners spent half an hour together in a room. Owners could talk to their dogs and pet them, and the researchers timed how long each dog gazed into its owner’s face. Everybody, human and dog, gave a urine sample before and after. Dogs that gazed for a long time had owners whose oxytocin rose noticeably during the session, and the dogs’ own oxytocin rose too. Short-gaze pairs showed much less change. The longer the eye contact, the bigger the shift, in both directions.

Then came the clever part. In a second experiment, dogs were given a nasal spray of either oxytocin or a saltwater dummy before meeting their owners. Female dogs that got the real hormone gazed at their owners for longer, and those owners in turn showed a rise in their own oxytocin. Push the loop from the dog’s side, and it turned over on the human side.

The team also tried the same thing with wolves. These were not wild animals but wolves raised by people from puppyhood, as tame as a wolf can be, paired with the humans who had raised them.

The wolves almost never held a person’s gaze. A few seconds at most, then they looked away. And their handlers showed no oxytocin bump at all. In wolves, staring at another animal’s face is closer to a challenge than a greeting. Whatever dogs are doing when they look up at us, their closest wild relatives do not do it, even after a lifetime of human company.

That led the researchers to a bold idea. Somewhere in the 15,000 or more years since dogs split from wolves, dogs found a way into a bonding system that had evolved for human parents and children. They did not invent a new signal. They borrowed ours. Two scientists writing alongside the paper put it bluntly: dogs had hijacked the human bonding pathway.

If this is right, it helps explain a lot about living with dogs. Why they watch our faces so closely. Why a look from across the room can feel like a conversation. Why people who have never met a particular dog can still be moved by its eyes in a way that a fox or a wolf does not manage.

It also fits with other findings. Dogs, unlike wolves, look to humans for help when a puzzle stumps them. Dogs have evolved a small extra muscle above the eye that lets them raise their inner brows and produce the look people describe as “puppy eyes”, a movement wolves cannot make. Piece by piece, the picture is of an animal that spent thousands of years being selected for its ability to talk to us with its face.

It is a lovely result, and it is worth saying what it does not show. The study measured oxytocin in urine, not in the brain, and the numbers involved were small: 30 pairs of dogs and owners, and a handful of wolves. Later work has been mixed. Some experiments have found that oxytocin changes how dogs look at people, but the effect depends on the breed, whether the dog is neutered, and the situation. A 2021 study of hand-raised wolves and dogs living in the same packs found that both species preferred the human they were bonded with, and suggested that life experience matters at least as much as domestication in how a canine’s oxytocin behaves.

Nobody has overturned the core finding, but nobody has proved that every dog everywhere is running the same loop, either. It is one strong study with a compelling story attached, and science is still working through the details.

None of that changes what happens at your front door. The dog looks up. You look down. Something old, older than dogs, older than people, hums between you for a moment. Nagasawa’s advice to owners after the study was simple: don’t just give commands. Look at your dog. It may be doing more than you think.