Peaceful Wolverine Rudolf Schenkel was nineteen years old when he began watching the wolves.
He had grown up in Basel. He finished school there in the spring of 1933, enrolled at the university to read zoology, botany, physics, and mathematics, and — because a person needs a trade — qualified as a gymnastics and sports instructor on the side. In 1934 he started going to the zoo. The Basel Zoological Garden kept wolves then, in an enclosure small enough that the animals could not get away from one another. He watched them for eight years, until the zoo gave up its wolves in 1942, and then he spent three more years writing up what he had seen, under the supervision of the zoo’s director, Heini Hediger.
The paper appeared in 1947, in the first volume of a journal called Behaviour, under the title Ausdrucks-Studien an Wölfen: Gefangenschafts-Beobachtungen. Expression studies on wolves: observations in captivity.
It is a careful piece of work. Schenkel was a serious man and he became, in time, a serious scientist — later a professor at Basel, an authority on rhinoceroses, one of the people who fought to keep the last Javan rhinos in the world alive.
What he described was a society organised by competition. The animals contested position; a top male and a top female held the others in check. For these he coined a term that had not existed before in the study of animal behaviour. Alpha-Tier. The alpha animal.
He described something else as well. He watched the larger wolves attack a single individual, over and over, for long stretches of time. The animal so attacked stopped functioning as a social partner. It was stripped of any capacity to initiate anything. Sometimes the injuries killed it.
And somewhere in the same paper — wolf biologists reading him now point this out; I have not read the German myself — he allowed that a wild pack might well consist of a monogamous pair, their pups, and their yearlings. He was right about that. Nobody picked it up.
The wolves in the enclosure were strangers. They had been gathered from various places — a couple from one zoo, a couple from another. None was related to any other. None could leave.
The problem was not captivity as such. Captive packs made up of a breeding pair and their own offspring behave much as wild wolves do. What Schenkel had in front of him was a group of unrelated adults who could not disperse, held together year after year in a space they could not leave, and what he recorded was what happened to those individuals in that situation.
The idea travelled. It reached the study of dogs, and then dog training, mostly through a book.
L. David Mech did his graduate work on Isle Royale, in Lake Superior, under Durward Allen — the wolves and the moose, the classic predator-prey study. He wrote The Wolf: The Ecology and Behavior of an Endangered Species in 1968 and published it in 1970, and it was, by common consent, the best book on the subject that had ever been written. It won prizes. It made him the person you called if you wanted to know anything about wolves, and eventually the person the American government called when it decided to put wolves back into Yellowstone.
The book relied on Schenkel, as any book of that period would have. Mech has since explained why, plainly and without excuse: When he wrote it, everything the field knew came from watching wolves who were unrelated and confined together.
The alpha left the book and went out into the world, where it positively thrived, and thrives to this day. It reached television. It reached the training manual and the obedience class. It reached the anxious conviction of a man in a suburb who believes he must eat his dinner before his labrador eats hers, and walk through doorways first, and never under any circumstances let her onto the bed, because these are the terms on which power is held. It kept going. There is now an entire vocabulary of human social ranking — alpha, beta, and the rest of the Greek alphabet — that traces back through a bestseller and a doctorate to a group of wolves who could not get away from each other in a yard in Switzerland.
Mech spent thirteen summers on Ellesmere Island, in the Canadian Arctic, watching wolves who had never been collected or confined or introduced. In 1999, he published a paper in the Canadian Journal of Zoology saying that the pack he had described in his own book did not exist. The typical wolf pack, he wrote, is a family. Two parents and their offspring. The deference that runs through it is the deference of children to parents, and the parents lead not because they won but because they bred. Calling such an animal an alpha, he observed, is about as useful as calling a human parent one, or a doe deer: Any parent is dominant to its young, so the word adds nothing.
Then he asked his publisher to stop printing his book.
They kept printing it. He asked again. He wrote it on his website — that the book was still in print, despite his numerous pleas. He explained, in paper after paper, that the concept was wrong and that his own book had helped spread it. The Wolf remained in print until 2022, fifty-two years after publication. In 2008 he noted the saying that new research takes twenty years to sink in, and wondered whether the alpha would take longer.
In subsequent studies of canine social hierarchies, almost none of the people who went and looked found what they went to find.
Aggression was the obvious place to begin. If a group of animals is organised by rank, and rank is held by force, then the animal who does the most fighting should be the animal at the top. The model is intuitive and it is testable, and when Joanne van der Borg tested it in a group of dogs in the Netherlands she had to throw it away.
Aggression ran in every direction. High-ranking dogs were aggressive to low-ranking dogs. Low-ranking dogs were aggressive to high-ranking dogs. As a measure of anything, it was useless. What she found instead — reliable, consistent, unidirectional — was submission. The lowered body. The low tail wag. If dog A shows this to dog B and dog B never shows it to dog A, you have learned something. You have not learned it by watching a fight. You have learned it by watching a greeting.
The earliest study of free-living dogs did it the other way. In 1998, S.K. Pal and his colleagues published their observations from Katwa, in West Bengal, and reconstructed the dominance hierarchy of the town’s street dogs from their aggressive encounters. It worked. They found an order, and it was linear, and it held. What Pal could not have known is that his method was measuring the right thing with the wrong instrument, and that when Simona Cafazzo in Rome, and Roberto Bonanni across five Italian packs, and Rebecca Trisko and Barbara Smuts among twenty-four neutered pets at a dog daycare in Michigan came at the same question with better tools, the aggression would fall away and the deference would remain.
The top animal is not the one who wins. It is the one to whom things are conceded.
Rank correlates with age.
Not with size — Bonanni was able to test this directly in two of his packs, and age beat body mass. Not with strength. Not, reliably, with sex. With age. Bonanni found the age ordering in four of his five packs; the fifth did not follow it, and he said so. Trisko and Smuts found older dogs outranking younger ones at a daycare centre where the dogs are neutered and well-fed and have nothing whatever to compete over.
The aggression that maintains this is highly ritualised. Bonanni recorded almost no potentially injurious aggression. Trisko and Smuts recorded conflicts that came to physical contact and never once caused an injury.
Nobody knows why the old are deferred to. It is tempting to call it instinct, and it may be. But the old dog in any group is also the dog everyone has known longest, whose reactions are most predictable, who was already there when you arrived. Age and history are confounded in every study anyone has done. The experiment that would separate them — the old stranger against the young resident — has not, as far as I can find, been run.
They defer to their elders, everywhere we have looked, and we do not know what they are responding to. It doesn’t have easy answers.
Bonanni, working with the Roman dogs, looked at leadership — who walks in front, who goes first toward a threat — and found that it did not simply track rank. The dogs who led were the high-ranking dogs who received submission during greetings. Affectionate submission. The muzzle-lick, the low wag, offered freely and in peace. The high-ranking dogs who received submission only in quarrels did not lead nearly so often.
Two entirely different things have been sharing a word.
There is status that is taken, extracted through the credible threat of cost. And there is status that is given, conceded by others who have decided that this individual is worth following. In our own species we long ago separated these and gave them different names — anthropologists call the second one prestige — and we understand that it is the more durable of the two, and the more powerful, and that the people who confuse them tend to end up with neither.
I will mention here that while I was at Dal, I made the mistake of mentioning Cesar Millan’s name at the dog lab, which is still run by Simon Gadbois (very French), who also teaches Animal Behavior.
You don’t want to mention dominance theory, and Cesar Millan, around Professor Gadbois. It is his strong opinion that such theories have produced nothing but coercive, borderline abusive relationships… Or at least that’s the impression I got.
You see, there’s a lot that observing situationally dependent hierarchical behaviours doesn’t tell you. Relationships, shared history, family ties, environmental stressors, these are all important.
Rank does not tell you who is friends with whom.
It does not tell you what the hierarchy is for. In wolves, rank allocates reproduction: The parents breed and the rest do not, and that is what makes the structure worth having and worth respecting. Anindita Bhadra’s group, studying the free-living dogs of Bengal, report that all the adults in a group have equal mating access. The reproductive monopoly is simply gone. What remains is a linear ordering that predicts who lowers their body first, and predicts, reasonably well, who ends up with the food, and predicts little else.
It does not tell you that these dogs spend more of their social lives interacting with us than with each other. Bhadra’s colleagues mapped it: In urban India, dog-human interactions outnumbered dog-dog interactions, and it was overwhelmingly the humans who started them, for good and for ill.
The pack may not be the unit. We have been drawing the boundary of dog society around the dogs, as if we aren’t right there, in the middle of everything — bigger, smarter, and legally holding the literal leash.
And a general hierarchy is not always there at all.
John Bradshaw, who has argued the case against the dominance construct more persistently than anyone, put nineteen neutered male dogs into an enclosure of a little over a quarter of a hectare — about seven-tenths of an acre — and watched. Some individuals consistently deferred to others. But the relationships did not assemble into a hierarchy. No pyramid, no alpha, no beta.
Trisko and Smuts, whose Michigan daycare study is cited on the other side of this argument, reported that dominance relationships were present in twenty-nine per cent of the dog pairs they observed. Not most. Fewer than a third.
Carri Westgarth, a British researcher, has put the sharpest version of the objection: That the dominance hierarchy may not be a feature of dog social structure at all, but a by-product of human observation. We find hierarchies because we go looking for hierarchies, in a species we have spent several thousand years breeding, handling, and requiring to defer to us. They are one species, yet look at the variation — we did that. Is it plausible that, having manipulated their genetics in virtually every way, there’s still a formal, violence-driven hierarchy installed in their genes?
Laurie Corbett studied a dingo colony in Australia and found what a hierarchy looks like when it has teeth. Rank was determined and maintained by aggression, and it changed hands during the breeding season. Every female in the pack conceived. And the top female killed every puppy born to every other female — animals closely related to her — one litter after another.
Corbett also measured stress hormones. The most stressed animals were not the subordinates. Plasma cortisol was highest in the alpha male and in the very lowest-ranking males, together, at the top and the bottom of the order. I will note, without academic backup for this particular point, that human hierarchies are the same in terms of stress distribution.
Scott Creel found a related inversion in African wild dogs: The dominants carry the higher glucocorticoid load, year-round, every year, which is the reverse of what the social-stress account of reproductive suppression predicts. It is not true everywhere — in wild female baboons the conventional gradient holds, and the subordinates are the stressed ones — but wherever it has been measured in these two canids, high rank is expensive.
The dingoes were captive. This has to be said, in an essay that opens with what captivity did to Schenkel’s wolves. Corbett’s dingoes were not strangers. They were relatives, breeding with relatives, because there was nobody else and nowhere to go. Of a potential forty-three animals in that enclosure, sixteen were alive after three years.
It was not unrelatedness that did it. It was the absence of an exit.
Which means the infanticide is not a fact about dingoes. It is a second demonstration of the thing this essay started with, arriving from a different direction.
And when Friederike Range put wolves and dogs — nine wolves, eight dogs, hand-raised and kept under the same conditions at an Austrian research centre — in front of a shared bowl of food, it was the wolves who tolerated each other. High-ranking wolves absorbed threats from their inferiors without retaliating. The dogs did not. Earlier dog-wolf comparisons had produced contradictory results; hers is the best-controlled to date, on a sample that would fit in a living room, and the gentle, cooperative, tolerant animal that domestication supposedly produced turned out to be the wolf.
None of this is a morality tale. The dingo mother is not a villain, and she is not a lesson about nature “red in tooth and claw.” She is an animal in a box with her relatives and no way out, doing what that produces. Down the road in Bengal her cousins nurse each other’s puppies, and their grandmothers help raise them, and they live under different conditions entirely.
We do this too. We have done all of it, in enclosures of our own making. We do not think it explains us.
The dogs of Moscow
Andrei Poyarkov has been watching the street dogs of Moscow for more than thirty years, from the Severtsov Institute, and what he has produced is not a hierarchy. It is a taxonomy of ways of living.
He sorts the city’s dogs — some thirty-five thousand of them — by how far they have travelled from us, and by how they eat. There are dogs attached to guarded places, building sites and warehouses and depots, who are intensely social toward the particular night watchman who feeds them and indifferent to everyone else. There are dogs who beg, and dogs who scavenge, and dogs who hunt. There are dogs who treat people as a resource, dogs who treat certain people as leaders, and dogs who treat all people as a threat. Over the decades he has watched the population lose the spotted coats, the wagging tails, and the friendliness that distinguish dogs from wolves.
Some of them go down into the Metro. Poyarkov’s account, given to reporters in 2010, is that most are there for the warmth, and that a small fraction seem to have learned to navigate the system — boarding, riding, and leaving at particular stations. Another behaviourist, Andrei Neuronov, has suggested they may recognise the announced names of the stops. It’s never one thing, of course. We know dogs can learn specific sounds, language-irrelevant. It is also smell, light, the movement of the crowd, the time of day. What can be said is that no gene contains the Moscow Metro, and that whatever those dogs are doing, they worked it out where they stood.
The variation is not in dispute.
Van der Borg’s kennel group was, in her word, tolerant, and she described a scale — despotic at one end, egalitarian at the other — along which any group of dogs might in principle be placed. Bonanni measured the steepness of his five packs and found it comparable to that of wolves and of the tolerant primates. The dogs of Bengal share mating access; the dingoes of central Australia enforce a reproductive monopoly. Bhadra’s team found that even within one city, dogs living where the human traffic is heavy behave differently from dogs a few streets away where it is thinner — more begging, more boldness, more exposure to cruelty.
The Bali dogs, an old and largely unmanaged population, are known across the island as guards. Nobody set out to breed them for it, but selection pressure happens with or without intent. The researchers who studied them suspect that centuries of living a particular way alongside a particular people quietly favoured the trait anyway, without anyone intending it.
So: different rules in different places, arriving somehow at the dogs who live there.
Is that culture?
What can be said without leaning is smaller and still large enough. Whatever a dog is, it is not finished at birth. The behavioural vocabulary appears to be species-typical — the low posture, the muzzle-lick, the exaggerated bow that means none of this is real — and it appears early, before an animal could have learned it. But which words a dog says, and to whom, and how often, and whether the group she lives in is a steep one or a level one: None of that is written down anywhere in her. It is worked out. Locally. In company.
Most of the animals I’ve cared for, dogs and cats, I’ve also been part of raising. Like a family, we’ve changed together. They influence me, and I them.
The dog on your sofa left her mother and her littermates at eight weeks.
S.K. Pal, watching puppies grow up on the streets of Katwa, recorded play arriving in stages — social play at around three weeks, mock fighting at five, pseudo-sexual mounting at six — increasing steadily as the litter worked out its own small society, and then dropping away sharply between eight and ten weeks, as though the question had been settled and there were no further need to ask it.
We take them at eight weeks. We take them at the moment the answer is arriving, and before it has arrived, and then we bring them somewhere that has no other dogs in it at all.
Everything your dog has learned about social life since then, she has learned from a species that is not her own — which probably doesn’t matter. Again, morality is a filter we impose.
Since your dog has been with you, she has never been hungry. She has never had to displace a rival to eat. She has never met a dog who could exclude her from anything she needed. Her mother’s temperament shaped her, and her sisters’ roughness shaped her, and eight weeks of the world shaped her, and then you did — every day, for years, whether or not you thought of it as shaping.
Behind that stand thousands of years in which her ancestors were selected to point, to retrieve, to herd, to guard, to run down vermin, to sit in a lap and be warm — and, in the last two centuries or so, sorted into the closed breeding populations we call breeds.
And behind that: She had sisters, raised in the same box, on the same milk, in the same eight weeks, and one is bold and one is watchful and one will not stop talking, and they were never the same.
Individuality is significant. My five cats are seriously distinct. You can’t easily predict by claiming all cats are the same.
So when your dog jumps on the sofa, or goes through the door ahead of you, or puts her head in your lap and sighs: She is a member of a species, and the product of a lineage bred for a purpose, and the result of an upbringing, and — underneath and through all of it — somebody in particular, doing something for her own reasons, which she has, even without language and articulation, and which you can sometimes guess.
She is not making a bid for the throne. There is no throne. There were unrelated wolves in a yard in Basel who could not get away from each other, and a young man who wrote down what he saw, and he saw it correctly, and for nearly eighty years we have been mistaking a cage for a kind of animal.
Sources and caveats
Schenkel, R. (1947). Ausdrucks-Studien an Wölfen: Gefangenschafts-Beobachtungen. Behaviour 1: 81–129. Mech, L.D. (1999). Alpha status, dominance, and division of labor in wolf packs. Canadian Journal of Zoology 77: 1196–1203; and Whatever happened to the term alpha wolf? International Wolf, Winter 2008. Pal, S.K., Ghosh, B., & Roy, S. (1998). Agonistic behaviour of free-ranging dogs in relation to season, sex and age. Applied Animal Behaviour Science 59: 331–348; and Pal, S.K. (2008). Maturation and development of social behaviour during early ontogeny in free-ranging dog puppies in West Bengal. AABS 111: 95–107. Cafazzo, S. et al. (2010). Behavioral Ecology 21: 443–455. Bonanni, R. et al. (2010). Animal Behaviour 79: 981–991; and (2017). Behavioral Ecology 28: 1004–1020. van der Borg, J.A.M. et al. (2015). PLOS ONE 10(8): e0133978. Trisko, R.K., & Smuts, B.B. (2015). Behaviour 152: 677–704. Bradshaw, J.W.S., Blackwell, E.J., & Casey, R.A. (2009). Journal of Veterinary Behavior 4: 135–144. Westgarth, C. (2016). Journal of Veterinary Behavior 11: 99–101. Corbett, L.K. (1988). Ethology 78: 177–198. Creel, S. (2005). Journal of Mammalogy 86: 255–264. Range, F., Ritter, C., & Virányi, Z. (2015). Proceedings of the Royal Society B 282: 20150220. Corrieri, L. et al. (2018). PLOS ONE 13(6): e0197354. Bhattacharjee, D., & Bhadra, A. (2020). Frontiers in Psychology 11: 2153. Paul, M., & Bhadra, A. (2018). PLOS ONE 13(5): e0197328.
Three things here I could not verify to my own satisfaction. Poyarkov’s classification of Moscow’s dogs appears in two versions in the sources available to me, and his primary papers are in Russian; I have described the categories without committing to a count. The Metro-riding is reported by journalists quoting him, not by him in print. And Schenkel’s aside about wild packs comes to me through wolf biologists reading the 1947 paper, not from my own reading of the German… I like German, though. I’d like to learn more of it.
Every study named above is small. Bonanni’s five packs. Range’s seventeen animals. Bradshaw’s nineteen dogs. Trisko and Smuts’s twenty-four. Nobody in this field has the sample size to settle anything, which is worth remembering on both sides of the argument, including mine.