PetTranslator.ai

Why Dominance Theory Is Wrong

Dominance theory in dog training came from misinterpreted captive-wolf research that the original researcher publicly retracted.

A dog and human in a positive-reinforcement training moment
By Khabir MughalJanuary 30, 202610 min read

Dominance theory in dog training is based on a misread of captive-wolf research from the 1940s through the 1970s. The researcher whose work popularized the "alpha wolf" — L. David Mech — has spent the last two decades publicly retracting the framing. Modern ethology doesn't support rigid linear hierarchies in wolves or dogs. The American Veterinary Society of Animal Behavior has rejected dominance-based training in two formal position statements. Here's what the actual science says.

Where dominance theory came from — and why it was wrong from the start

The story starts at the Basel Zoo. In 1947, Swiss animal behaviorist Rudolph Schenkel published observations of a captive wolf pack — unrelated wolves, forced into proximity in an enclosure, fighting for resources they couldn't escape. He described a linear status hierarchy maintained through aggression and submission, and named the top male and female the "alpha pair." Schenkel later expanded the framework in his 1967 paper "Submission: Its features and function in the wolf and dog" (American Zoologist 7(2): 319–329).

The methodology problem is obvious in hindsight. Captive, unrelated wolves under forced cohabitation behave nothing like wild wolves. The aggression Schenkel observed was an artifact of the captivity, not a feature of wolf social structure. He extrapolated zoo-stress behavior to the whole species, and from there to dogs.

L. David Mech picked up the framework in his 1970 book The Wolf: Ecology and Behavior of an Endangered Species (University of Minnesota Press). The book became the standard reference for a generation of biologists and, by way of trainers who borrowed the language, the standard reference for dog training. The "alpha wolf" entered popular vocabulary.

Then Mech went and studied wild wolves. Across the 1990s and into the 2000s, working with packs on Ellesmere Island and in Yellowstone, he found something completely different from the Basel zoo. Wild wolf packs are families. The "alpha" pair are the parents. The other wolves in the pack are their offspring from prior years. There is no fighting for rank. There is no forced submission. Mom and dad are in charge because mom and dad are mom and dad — the same way a human parent is "in charge" of a four-year-old. No alpha rolls. No status duels. Just a family group raising the next litter.

Mech published the retraction directly. In "Alpha status, dominance, and division of labor in wolf packs" (Canadian Journal of Zoology 77: 1196–1203, 1999), he wrote that "the typical wolf pack is a family, with the adult parents guiding the activities of the group in a division-of-labor system." He expanded this in Wolves: Behavior, Ecology, and Conservation (co-edited with Luigi Boitani, University of Chicago Press, 2003). On his own site, davemech.org, he has a standing public statement asking authors and trainers to stop citing his 1970 framing and to stop using "alpha wolf" to describe wild wolf packs.

The originator of the alpha framing has spent two decades trying to take it back. Dog trainers who still use it are citing a model the source has retracted.

Why the wolf model wouldn't apply to dogs even if it were right

Even granting the original captive-wolf framework for a moment — it still wouldn't apply to dogs. Dogs are not wolves with collars.

The genetic split between modern dogs and the wolves that became their ancestors happened somewhere between 15,000 and 30,000 years ago, with estimates varying depending on which study and which marker. That's roughly the same evolutionary distance that separates humans from earlier Homo lineages. Domestication selected aggressively for tameness, sociability with humans, and bidirectional human communication. Dogs read human gestures and gaze in ways no other species, including wolves raised by humans from birth, can match.

John Bradshaw, an anthrozoologist at the University of Bristol, lays out the case extensively in In Defence of Dogs (Penguin, 2011, published in the US as Dog Sense). His argument: dogs and wolves diverged behaviorally, cognitively, and socially. Free-ranging dog populations — village dogs in India, Russia, Africa, South America — do not form wolf-like packs. They form loose, flexible, context-dependent affiliative groups with no linear hierarchy. Resource access varies by motivation in the moment, not by some persistent rank.

Importing a misread wolf model into dog training is doubly wrong: the wolf model was a misread to begin with, and dogs aren't wolves anyway.

What actual canine social structure looks like

The ethological literature on free-ranging dogs is consistent. Dogs form affiliative groups with overlapping membership, not bounded packs. Relationships within those groups are dyadic — dog A and dog B have a relationship, dog A and dog C have a different relationship — and don't aggregate into a single ranking. Access to resources (food, resting spots, mates) depends on motivation in that moment, prior history with the specific resource, and current arousal state.

There is no observed analog to "the alpha." There is no fight for status. There is no member of a free-ranging dog group who maintains position through aggression toward subordinates. The structure is flatter, more flexible, and more context-dependent than the wolf-pack myth predicts.

Inside a multi-dog household, the same pattern holds. Dogs work out access to specific things — a particular toy, a specific resting spot, the spot next to their human — through repeated, low-level signaling. The dog who gets the toy isn't "the dominant dog." They're the dog who, for that specific item in that specific context, has the higher motivation and the prior history of access. Switch the item or the context and the access can flip.

What the science says about training methods

The empirical question — does dominance-based aversive training work better, equally, or worse than positive reinforcement — has been answered repeatedly in peer-reviewed research. Aversive methods perform worse on every outcome measured.

Hiby, Rooney, and Bradshaw, in "Dog training methods: their use, effectiveness and interaction with behaviour and welfare" (Animal Welfare 13(1): 63–69, 2004), surveyed 364 dog owners and measured method use against reported behavior problems and obedience scores. Owners using reward-based methods reported higher obedience and fewer behavior problems. Owners using punishment-based methods reported more problem behaviors. The reward-based group also reported their dogs were less fearful overall.

Herron, Shofer, and Reisner, in "Survey of the use and outcome of confrontational and non-confrontational training methods in client-owned dogs showing undesired behaviors" (Journal of Veterinary Behavior 4(2): 47–48, 2009), surveyed owners of dogs already showing behavior problems and asked which methods they'd used and what happened. Confrontational methods produced aggressive responses at high rates: 43% for being hit or kicked, 39% for an alpha roll, 38% for grabbing the jowls and shaking, 31% for a "dominance down." Non-confrontational methods produced near-zero aggressive responses. The methods sold as solving aggression were generating aggression.

This is not a marginal finding. It's the headline result. Aversive training increases the behavior it claims to suppress.

The American Veterinary Society of Animal Behavior issued its first position statement on the use of punishment in 2008. The updated AVSAB Position Statement on Humane Dog Training (2021) is unambiguous: "AVSAB's position is that animal training, behavior prevention strategies, and behavior modification programs should follow the scientifically based guidelines of positive reinforcement, operant conditioning, classical conditioning, desensitization, and counter-conditioning. AVSAB recommends that aversive training methods have a damaging effect on both animal welfare and the human-animal bond."

That's the professional body of veterinary behaviorists in North America. Not an advocacy group. The people who treat behavior cases clinically.

The harm of dominance training

The cost of aversive training shows up in four measurable places.

Aggression responses. Herron et al. (2009) above. Confrontational methods elicit aggressive responses in roughly a third of dogs they're used on. Owners who use aversive methods to "fix" aggression frequently produce more of it.

Stress markers. Studies measuring cortisol, behavioral stress signals (yawning, lip licking, displacement scratching, tail tucking), and learning performance during training sessions consistently find that aversive-trained dogs show more stress markers and learn more slowly than positive-reinforcement-trained dogs. The dogs are not just less happy — they're worse at the actual learning task they're being put through.

Damage to the human-dog relationship. Casey, Naj-Oleari, Campbell, Mendl, and Blackwell, in their work on owner-dog attachment and training method, documented that dogs trained with aversive methods show reduced affiliative behavior toward their owners and altered attachment patterns. The relationship itself takes damage. The same was found in earlier work by Rooney and Cowan ("Training methods and owner-dog interactions: links with dog behaviour and learning ability," Applied Animal Behaviour Science 132: 169–177, 2011).

The rebound effect. Punishment suppresses the surface behavior without addressing the underlying state. A dog who growls at strangers and gets punished for the growl will often stop growling. They will not stop being afraid of strangers. The next signal up the escalation ladder — the bite — has no warning attached. Trainers who "fix" growling with corrections are removing the smoke alarm without putting out the fire.

What works instead — the actual science

The alternative isn't permissiveness. It's behavior change that uses the way mammals actually learn.

Positive reinforcement. Behavior that's followed by something the dog values is more likely to occur again. Mark the desired behavior, deliver the reinforcer, repeat. The mechanism is operant conditioning — the same mechanism every functional training program in the world uses, including the ones for marine mammals, captive primates, and zoo elephants who all weigh more than the trainer.

Counterconditioning and desensitization. For fear-based behavior, change the dog's emotional response to the trigger by pairing low-intensity exposure with high-value reinforcement. Over repeated sessions, the trigger comes to predict good things rather than bad. This is the standard treatment protocol in veterinary behavioral medicine for reactivity, fear aggression, and most anxiety presentations.

LIMA. Least Intrusive, Minimally Aversive. A hierarchical framework adopted by the IAABC and most credentialed behavior consultants. Choose the least invasive effective intervention. Address health and management before training. Use reinforcement-based methods first. Aversives are bottom-of-the-stack, only after every less-invasive option has been tried — and in practice, with skilled trainers, almost never needed.

The AVSAB framework. The 2021 position statement is the working document. Force-free, reward-based, with behavior modification protocols rooted in operant and classical conditioning. This is the standard of care.

How to identify a trainer using dominance theory

Red-flag phrases — if a trainer uses any of these, walk away:

Red-flag tools — these are aversive devices that the AVSAB position statement specifically recommends against:

A trainer using any of these is operating on a model the science has rejected. There are credentialed force-free professionals in nearly every metro area. Switch.

What credentials actually mean

These are the credentials backed by evidence-based curricula and continuing-education requirements:

All of these require demonstrated knowledge of operant and classical conditioning, force-free methodology, and continuing education. None of them include dominance-based or punishment-based content in the curriculum.

If you're working through a complex behavior case — fear-based reactivity, separation distress, resource guarding, intra-household conflict — the IAABC and AVSAB websites both maintain searchable directories.

Try it on your own dog

A useful exercise: pick the behavior in your dog you've been told is "dominance" and re-read it through the body-language framework instead. The dog who "won't let you in the door" is probably aroused and undertrained on door manners. The dog who "guards the bed" is probably resource guarding a safe resting spot and needs counterconditioning, not a correction. The dog who "ignores you on walks" is probably under-reinforced for attention and over-rewarded by the environment. None of these are status problems. All of them are training problems.

PetTranslator.ai reads the body-language signals — ears, eyes, lip line, tail, posture — that tell you what's actually happening in your dog in a given moment. No dominance framework. No alpha-rolling advice. Upload a photo and the AI returns a structured behavioral report grounded in the same framework AVSAB-credentialed behaviorists use.

For the underlying signal framework, see the field guide on dog body language. For the stress markers that aversive methods produce, see signs your dog is stressed. For separation cases specifically, see dog separation anxiety. For a frank assessment of which pet-translator products actually work, see do pet translator apps work.

Sources

Khabir Mughal is the founder of PetTranslator.ai. This article was reviewed against the AVSAB Position Statement on Humane Dog Training (2021) and the primary sources cited above before publication.

Tags#training-science#force-free#avsab#anti-dominance

From our Training Science