The Dominance Controversy
Virtually everyone who started as a dog trainer over 15-20 years ago started out using traditional dog training techniques: similar to those used by Cesar Millan (National Geographic’s The Dog Whisperer). This is how most dogs were trained back then. As a result we have first hand experience as to why and when such punishment-based techniques might work, the pitfalls, and why and when other techniques work better.
Traditional training techniques are based on the idea that we must become the dominant leader and rule our pets the way a wolf would rule a pack. That is, they assume most misbehavior in dogs is due to the dog trying to be dominant and then they employ techniques that they think a wolf (since dogs are seen as having a social structure similar to wolves) would perform in a wolf pack. In order to evaluate whether this reasoning is valid, we must first understand what dominance is.
A. The Definition of Dominance
In animal behavior, dominance is defined as a relationship between individuals that is established through force, aggression and submission in order to establish priority access to all desired resources (food, the opposite sex, preferred resting spots, etc). A relationship is not established until one animal consistently defers to another.
In species where strong hierarchies exist, this hierarchy is important evolutionarily because having a high rank confers a greater ability pass on one’s genes.
For instance, if you put four bulls or roosters together, they will fight and establish a rank order of 1 through 4. The highest-ranking bull or rooster will have access to the most females that are available for mating. The other males may have little chance or much lower chance to mate. In fact in the wild, for many species, most males never get the chance to mate.
Note that dogs evolved for the past 10-15,000 years as scavengers with a promiscuous mating system. That is, free roaming dogs mate with multiple members of the opposite sex. Virtually all males get to mate. Consequently a rigid hierarchy and high rank would not confer a huge advantage compared to, say high rank in a group of bulls or chickens.
This is also very different from wolves, where the pack must cooperate to provide enough resources for the offspring to survive and consequently generally only the highest ranking male and female pack members mate and have puppies.
B. Test Yourself on the Definition of Dominance
- Example 1: Fire Station Puppy
Is this puppy trying to assert higher rank here? »
No. The puppy is not trying to gain higher rank over the people or have priority access to resources. He’s just trying to get attention. Even if they don’t pet or talk to him when he’s sitting on their lap, the act of being in bodily contact with them is a form of attention. If they only petted him when he was sitting on the floor and stood up immediately when he jumped on their lap, then he would instead develop a habit of sitting on the floor.
- Example 2: Conflict Between Cat and Dog
- Why do you think the dog chases the cat? Is he trying to "dominate" it?
- Is the cat aggressive because it’s trying to dominate the dog?
- Why do you think it’s trying to be aggressive?
- Do you think it’s safe to tell owners to restrain a cat as a dog is approaching? »
The dog is chasing the cat because it’s fun. Dogs like to chase prey-like objects. The cat is aggressive because she’s trying to defend herself. Although this cat was calm when Millan held it down, most cats will bite and scratch the person restraining in order to get the person to let go if the approaching dog (or cat) frightens them enough.
I know of one example of a person who’s hospital bill caused by a cat bite in the type of situation depicted in the video amounted to $40,000. If the person who recommended this type of restraint in this situation was a veterinarian or other professional guided by a "practice act" he/she could be held liable for injuries that ensue.
- Example 3: Dog Dominant to Light?
- Is the terminology of dominance used correctly here? Why or why not?
- Is the dog trying to assert higher rank over the light? »
- Example: Dominance is not a personality trait
- These are chickens from my flock that are visiting a friend’s house. The black and white chicken is a rooster and he is ranked highest in my flock. The buff-colored female is the highest-ranking female. She was introduced to my flock recently and she immediately pecked all the females and gained the top spot. Now she randomly pecks at the other hens to reinforce her status throughout the day. Notice that when the rooster meets a large leghorn female, he immediately wants to fight and he loses. Now she is higher ranked than he is and she may randomly peck him and others below her whenever they are close to her.
- Example: Having a stable rank does not mean the situation is peaceful
- With this particular pack of wolves (from National Geographic’s "Arctic Wolves") the aggressive displays over food are high. In other groups the displays may be much more subtle. The implication for humans is that if we use the dominance model for modifying behavior, we may need to continue with threat displays for the rest of the pet’s life.
- Example: David Mech, most published wolf biologist, talks about the term alpha wolf.
- Note that this does not mean that hierarchies don’t exist at all. In fact with mixed packs such as with wolves raised in captivity, there are clearly dominance fights which can lead to serious injury. For updated research on dog social systems, refer to (1) MacDonald and Carr, 1995; (2) Boitani, 1995; (3) Bradshaw, 2009.
- Example 1 : Jonesy Wants the Toy
In this video example, my dog (crazy for toys) wants to go after the toy. He knows he only gets to go forward if he’s in heel position. I don’t have to tell him, my actions of waiting make it clear to him.
My goal is to train him to like doing what I want. I don’t nag him or yell at him, or force him, I just make it clear, no reward for undesirable behavior (reward is immediately removed), lots of reward for good behavior (I walk forward immediately when he’s good and in the past have (a) given him lots of food rewards for heeling; (b) rewarded him for heeling for just a few steps before I release him to get the toy).
- In this video does this dog look happy or sad?
- Does he look "resentful?" or just doing it to avoid getting in trouble?
Even though he can’t get to the toy immediately, Jonesy looks happy. His tail is up, his head is up, his attention is focused on me, he’s showing his "play face", and he’s prancing energetically. He is happy because he understands how he can get the toy. He just has trouble containing his excitement. In the past I’ve experimented with punishment to see whether I could use it as a short-cut. Pinch collar corrections cause him to look depressed and subdued - not a great result considering there are already too many things that cause him to become scared and act "depressed". However in other similar contexts, the pinch collar would cause him to become more aroused and aggressive. Any stronger correction could possibly cause injury. Other forms of punishment such as calmly holding him down taught him to avoid me when he thought he was in trouble. You can imagine how difficult it is to catch a Jack Russell Terrier that doesn’t want to be caught!
- Example 2: Counterconditioning a Dog That’s Fear Aggressive to Other Dogs
The goal here is to train the dog to perform alternate desirable behaviors that he enjoys. That way he’ll learn to associate the situation with good things and his underlying emotional state will also change.
- In the end, does he look happy during counterconditioning, or stressed out?
- Does he look "resentful" or just doing it to avoid getting in trouble?
Once I have done the foundation training, he learns that focusing on me is fun. Walks are about playing fun games with me. He looks at me, has a "play face", and his tail is wagging. Then when I work with him around the dog, I continue to keep him in a happy, playful state. So, where the dominance-based (Millan) method is to get the dog to act aggressive and then show him that you’re the boss with an alpha-roll or collar tightening around the neck, the updated method is to keep dogs below their threshold for aggression and train them to perform desirable behaviors using positive reinforcement so that they enjoy the new behavior.
- Example 3: Emily the pitbull
Now lets look at how Cesar Millan would handle a similar case.
- Why is the dog lying on its side at the end of this clip?
- If you could be held liable for any injury to the owner or dog, would you consider this technique safe to the handler or to the dog?
This dog is on its side because it is exhausted. The leash around her neck tightens every time she struggles and cuts off her air supply. This decreases (or helps to decrease) her stamina. At one point when Emily is on her back she flails and bares her teeth aggressively at Cesar Millan as she’s trying to get away. This is probably the first time she’s directed aggression towards a human. Millan controls her and does not let her get up – but most likely if her owners try this they will be bitten. Dogs are smart enough to distinguish which people can physically dominate or control them and which people cannot.
- Example 1: Counterconditioning a Dog Who’s Aggressive for Toenail Trim
Here’s a common example: how many people have seen a dog that doesn’t like its toenails trimmed? I used to sit on these dogs until I learned we could CC them. Meaning, we can pair food with the procedure so that they learn to associate the procedure with good things.
When we do this we have to combine it with desensitization. That is we start the stimulus so low it doesn’t really elicit a response and we pair this with something the dog likes (treats, petting, etc). The goal is the dog never reacts - it is just focused on the treats. Timing is important when we do this, as you’ll see.
Do you think the dog will get better or worse with each visit if we focus on changing the underlying emotional state?
If the owner works on this at home on a daily basis (especially with guidance or coaching from a technician or trainer) they should be able to confidently clip nails themselves or have their vet clip the nails.
- Example 2: Millan works with a dog that hates grooming
Now lets see what Cesar Millan and other old-school trainers would do. He/they use a process called flooding (although sometimes they erroneously think they are using desensitization) where he/they expose the dog to the undesirable condition full force or enough to cause an aggressive or strong fear reaction rather than starting at a level that barely or does not elicit a response.
- If the dog in this videos gives up in the end, and stops lunging, do you think it would be because it feels better?
- Do you think the dog will get better or worse with each visit?
- Given that Millan was bitten in this video, do you think it’s safe for other people to perform? Would you trust them to make a good judgement regarding whether it’s a safe technique if Millan was able to make the safe judgement?
This video depicts what goes on in a lot of hospitals and with many groomers; however, at the vet hospital the staff is generally much more skilled in how they restrain the dog. Generally vets know from experience that a dog handled in this manner remains difficult to handle or gets worse.
- Example: Dog tries to bite when you blow in his face
- This foster dog becomes aggressive when I blow in his face. Then I start to countercondition. Usually I would blow softly enough so he wouldn’t snarl, but for video purposes, I blow hard enough to just make him start to growl/snarl. Watch to see if he gets better or worse.
- Does he get better or worse during the 5 minutes and overall?
- Does he look upset with the blowing at the end? Or fearful or resentful?
This dog improves greatly in just 5 minutes. In the end, he thinks the blowing is a good thing. Note that if I had only given him 5-10 treats or done 5-10 repetitions, my impression would have been that he didn’t get better. In fact, if I went way over his threshold and stimulated him to growl a lot, I might even think I made him worse due to the food (when the worsening would have been due to the poor technique of flooding him instead of desensitizing him).
When performing this technique you must perform enough repetitions to see a change. Ideally you stay just below the threshold that makes the dog fearful or aggressive.
- Example: Conflict between dog and cat.
- We saw this video earlier. While the use of the term dominance is incorrect you may wonder, aren’t the dog and cat better around each other?
- Do you think the dog and cat will behave this well if the owners are not present?
- Do the dog and cat look happy at the end? What would their body posture be if they were happy?
- What body signs indicate that they may not be happy?
- Neither the dog nor the cat are happy, they are just suppressed. The dog is avoiding eye contact with Millan. If he were happy his ears would perk forward, he’d show his play face, his muscles would be relaxed, and he would not be avoiding eye contact as if he’s looking for a way to escape. This cat looks similar to the cat at the vet hospital who’s resigned to the idea that he will be examined whether he likes it or not.
- LiveScience: Critics Challenge ‘Dog Whisperer’ Methods by Lynn Peeples
- KOMO News: Are the Dog Whisperer’s methods harmful? by Joel Moreno
- American Humane Society: ‘Dog Whisperer’ Training Approach More Harmful Than Helpful
- 4PAWS University: The Dog Whisperer Controversy by Lisa Mullinax, CPDT
- San Francisco SPCA: Training Philosophy
- Steve Dale Pet World: He Ought to Call Himself the Dog Screamer by Steve Dale
- Marin Independent Journal: Tails of Marin: Pros and cons of the Cesar phenomenon by Trish King
- Steve Dale Pet World: Cesar Millan Contends ‘I Help The Dogs.’ But Experts Question His Approach by Steve Dale
- SF Gate: The Anti-Cesar Millan by Louise Rafkin
- 4PAWS University: Beyond the "Dominance" Paradigm by Patiricia B. McConnell, PhD
- The New York Times: Pack of Lies by Mark Derr
- Esquire: The Dog Whisperer Should Just Shut Up: Misguided expert of the year by Curtis Pesmen
- Huffington Post: Seven Tips for Preventing Dog Bites in Animal Care Professionals and Dog Lovers by Sophia Yin
- Veterinary Forum: Dominance in Dogs by Sophia Yin
- American Veterinary Society of Animal Behavior: All Four Veterinary Behavior Groups Urge Merial to Stop Partnering with Cesar Millan
- Beyond Cesar Millan
C. Other Facts About Dominance Hierarchies.
In the last several decades field scientists have made many new discoveries about dominance hierarchies. Individual researchers and research groups have each spent hundreds to thousands of hours observing interactions between members in groups and relating the interactions to rank. These studies have spanned many species including macaques, chimpanzees, vervets, baboons, lions, wolves, meerkats, chickens, cattle, goats, and more. Here’s some of what we have learned about hierarchies from all of the research.
1. Dominance is not a personality trait. If you take four individuals, each who is the highest ranked member of its group and put them together in a new group, they will fight and establish a rank order of 1-4 between them. Thus only one of the individuals who was ranked highest in its own group will have highest rank now...
2. Having a stable rank does not mean the situation is peaceful. Rank is maintained through ritualistic aggressive acts on the part of the higher ranked individual and appeasement or deference on the part of the lower ranked individual. In many groups, once rank is obtained, individuals can then forge stronger bonds though mutual grooming or other bonding behaviors. Some alpha individuals have more aggressive styles of maintaining rank while others have a calm, more subtle style. Different species also have different tendencies. For instance macaques monkey are notoriously despotic and aggressive and removal of the high ranked individual for just several hours can result in immediate restructuring (fighting) within the group. Similarly with mixed packs of wolves, such as occurs in captivity, removal of higher ranked wolves can lead a need to re-establish rank once the individual is reintroduced. Note that this model of group structure and leadership is similar to what occurs on shows such as the Sopranos and Madmen.
3. High rank may be short-lived and lost opportunistically. In the wild, a high ranking animal’s alpha status may only last a short time. That is, it will last only as long as the individual can maintain it’s rank though aggressive acts. For instance, a lion’s tenure lasts an average of 2 years, in elephant seals the beach master may only have several years where he can attain high enough rank to mate.
Implication: For humans attempting to use dominance theory to control their dogs, only those humans who are strong enough to attain the high rank can do so with low risk of injury. Once in this spot they must be strong enough to maintain the high status.
Wolves in the wild generally do not gain their high rank by fighting their way to the top. Instead a male and female breed and the pack is a family unit comprised of the parents and the offspring. The parents naturally become the leaders. The offspring naturally follow their lead. As a result of this discovery regarding pack structure, wolf biologists no longer even use the term alpha with wild wolf packs.
4. Wolves in the wild generally do not gain their high rank by fighting their way to the top. Instead a male and female breed and the pack is a family unit comprised of the parents and the offspring. The parents naturally become the leaders. The offspring naturally follow their lead. As a result of this discovery regarding pack structure, wolf biologists no longer even use the term alpha with wild wolf packs.
Implication: The idea that we need to dominate animals (especially dogs) based on our idea of what wolves do in the is based on faulty information on what wolves do in the wild.
D. The Science and Psychology of Behavior Modification
E. Training isn’t JUST about training appropriate behaviors. It’s about addressing the underlying emotional state.
For instance, if an animal is fearful, it doesn’t make sense to treat it by trying to "dominate" it. That won’t bring it out of its fear state.
F. But doesn’t giving food to aggressive dogs make them worse?
Many people think if you give treats when a dog is barking or growling out of fear or other aggressive reasons, you will reward aggressive behavior. Actually, if you change the underlying emotional state you’ll change the outward behavior.
G. Leadership is still important but one can lead with finesse, not force.
One definition of leadership is the ability to influence an individual to perform behavior they would not otherwise perform. By that definition, pet owners do need to develop leadership skills. However we have a choice of leadership style. We can lead by force like a dictator such as Muammar Qaddafi or by providing goals or rewards the follows want, such as Mahatma Ghandi. Schools of marketing and leadership recommend against the dictatorial, coercive style of leadership and encourage methods of leading that motivate humans through positive methods.
A similar approach should be used with animals. Instead of using coercion we can learn to lead like a leader in a dance. When partners dance as a couple, one leads and the other follows. The leader’s job is to decide ahead of time which steps to perform and then guide his partner in a clear manner so that the partner CAN follow. Partners who have to shout out the steps or who yank their follower around don’t make the cut. With animals the approach is similar. If we set rules and have a clear picture of what we want, then we can consistently convey this information to the animal through our body language and perfectly timed rewards. When learning to lead our pets, it is essential to realize that it’s important to reward desired behaviors consistently but it’s equally important to remove rewards for unwanted behavior immediately. This combination is essential if the animal is to learn to behave ONLY in the desired ways. It is also essential to stick to this plan until the good behavior becomes a habit.
H. Positive Does Not Mean Permissive
It’s important to realize that positive does not mean permissive. Regardless of which species you’re dealing with there must be rules and guidelines for behavior.
Rule # 1: The animal should respect your personal space and enter when given permission. That is, dogs should not jump on people unless the humans have solicited jumping, horses and goats should not walk over your toes or butt or push you with their heads, parrots should not fly and land on your head and walk all over you like you’re a jungle gym. Instead pets/animals should ask for permission for your attention and everything else they want by sitting or standing quietly and calmly and focus on you. It’s up to the humans to pay attention and reward this polite behavior.
Rule # 2: The overall goal is to train your pet to respond to your cues/command to, for instance, come, lie down, walk nicely on leash, the first time you ask without your needing to beg or plead or carry treats or have a pinch collar or choke chain on. The response should become a habit.
When is Force (negative reinforcement or positive punishment) appropriate?
Punishment isn’t always inappropriate. It’s just incredibly overused - and in most cases it’s performed incorrectly. As a behaviorist who works with many domesticated and wild species of animals and who started as a traditional dog trainer, I draw from my entire knowledge base to modify behavior in animals. My goal is to use whichever techniques will work best with the least likelihood of side effects in the pet. If that best technique involves a punishment such as, for dogs, a pinch collar "pop" or reprimand, or booby trap of some sort, or even an electronic collar, then I will use it. But it rarely does. Consequently I use the combination of positive reinforcement and negative punishment 99% of the time and use aversives (including verbal) probably 1000 times less than a traditional trainer and relevant rewards 1000 times more. I also occasionally recommend or allow aversives (including verbal "ah") to clients (eg. perhaps to one client once every 2 years). First I educate them regarding the possible adverse effects and difficulties of using the aversive and then let them choose whether they feel comfortable trying the technique. They are supervised and we observe the dog’s body language for negative effects. I also educate them to the fact that some products, such as head collars, flat collars and harness can be aversive to some dogs even though we try to desensitize and counter-condition the dogs to them. These and other products must all be used carefully in order to get the intended positive effect.
With that said, I don’t think that pet owners should use force or punishment. If they decide to do so it should only be after they are proficient at rewarding good behaviors and removing rewards for unwanted behavior until the good behaviors have become a habit. My reasoning here is that:
1. Unless we are intimately aware of how we reward inappropriate behaviors, people will punish bad behaviors when they are thinking about training and reward those same behaviors at other times. They need to first learn to be aware of their actions and how it affects their pet before they earn the privilege of punishing the pet (using force-coercion).
2. Humans tend to fall back on punishment because it requires less brain power to react to a problem than to think about our contribution to the problem and how to proactively prevent the problem instead. In order to make the proactive method a habit or mindset, it’s important to first not have the "punishment" crutch to fall back on. It’s like learning to do math. If you have a calculator, you may never learn how to actually add or multiply yourself.
3. With both positive reinforcement as well as coercion, the timing is the same and owners need to be equally consistent. So if the owner does not have the ability to reward consistently and with the right timing, it’s not likely they will be able to perform the punishment technique well either.
4. Lastly the use of a force as a first-line of treatment for training can cause animals to seem stubborn and willful, when they are actually frustrated, confused, and or have little motivation other than the need to avoid fear and pain to want to perform the behaviors. Again, punishment does not take into consideration the motivation of the animal. And it doesn’t tell them what they should be doing instead, it just tells them what they should not be doing.
I am not against the use of punishment or force altogether, although I do use them about 1/100 or 1/1000 as much as a traditional trainer would, especially when dealing with aggressive or fearful animals. Rather I am saying that coercion techniques are associated with more fallout or adverse effects. To use them we need to know how to employ them effectively and know what ineffective use looks like. And we have to be able to recognize the adverse effects so that we know when coercion is appropriate and when it is not.
I. What to Take Out of Shows Such as "The Dog Whisperer" (dominance-based) and "It’s Me or the Dog" (this show is not dominance-based).
Even though shows such as the "Dog Whisperer" and many others are based on the erroneous understanding of dominance and the need to use force or coercions as the first-line of training for all problems, there are some good recommendations and lessons one can learn from it and other shows.
What to Evaluate:
1. Turn the sound down and watch the animals: Evaluate their body posture. Do dogs look happy in the end and willing to behave the way the owners would like? Or do they look fearful or like they are just behaving as long as they might get punished?
2. Ask yourself, do the techniques appear safe for you or your kids and family members? In cases where the dog acts aggressively to or bites the trainer consider that this could be directed at whoever is performing the technique.
3. What behaviors are owners rewarding? Watch to see how owners reward unwanted behavior. Although "The Dog Whisperer" rarely demonstrates rewarding of appropriate behaviors immediately as they occur (and with a reward the dog wants at that instant) try to come up with ideas on your own regarding behaviors that you would reward. Also keep your eyes open for the rare times that Millan does focus on rewarding appropriate behavior. One more point - to understand the difference between the highly effective positive reinforcement vs. the variably effective "cure", read common questions.
4. How are the owners trained? If the behavior is about the owners then one would expect the owners be trained as much or more than the dogs. Watch to see how or whether the trainer trains the owners. Ask, what does the owner specifically need to do? Is there a clear plan or does it just seem like "movie magic?"
Good Points Gained from Viewing:
1. Exercise: Yes dogs and other pets should get exercise every day. Most people don’t exercise their dogs, cats or other pets enough. Note that exercise is not a substitute for training though. For many dogs it just gets them in shape so that they can misbehave longer! And when exercised, we can accidentally reward unruly behaviors. For instance if we toss the ball for the dog that barks and jumps on you until you throw it, we are actually training overly aroused, unruly behavior. So, even exercise must be done correctly for it to have maximum benefit.
2. Rules and Limits: Yes, all animals need guidelines and limits for behavior in the house. This is for sure. Whether you train pets by punishment and coercion or by rewarding good behaviors and removing rewards for inappropriate ones at exactly the right time is what distinguishes traditional trainers from positive reinforcement trainers taking a scientific approach. Rules are what allow animals in groups to get along. And if we want our pet to be able to accompany us in public, we have even higher expectations for their behavior.
3. Be calm and assertive: Yes, to be a leader that animals want to follow you do have to act like you know what you’re doing. You have to walk like you know which direction you want your dog to follow, and provide visual and verbal cues in a manner that makes it clear to the dog that you know what you want.
4. Training is About the Owners: The owners have rewarded all kinds of unwanted behavior. Watch to see how they reward the inappropriate behaviors. Also, consider, that if the training is about the owners, how much training should the owners have or do they need in order to perform the techniques? Being calm and assertive alone won’t make you a leader though or even good at any particular task or job. To be good at the task you must first develop the skills. For instance Tiger Woods, Yoyo Ma, and every other successful pro athlete or musician are good at what they do because they’ve practiced correctly many hours a day - not just because they are calm and confident.
J. Why Some Dogs Seem to Improve with Force
The purpose of force in most cases is to stop a behavior. Overall, effective force does suppress behavior. Punishment (use of force, coercion to decrease a behavior) can work to permanently stop behavior in instances where it’s used correctly. In many cases where it is metered correctly it serves to suppress behavior only temporarily though. Especially when it is not paired with rewards for appropriate behavior and when the underlying emotional state and cause are not addressed. To the viewer immediate suppression may look like a cure. Watch closely to see if the dogs look like they are happy (ears forward, head and tail up, body relaxed, smile face) to be behaving well, or if they look suppressed and fearful (tail down, tense, avoiding eye contact, ears back or out to the sides, they look sleepy or move slowly, etc).
The old method of training was to assume dogs were trying to be dominant if they didn’t behave. As a result we would use force until they gave in rather than addressing the underlying emotional state or cause (such as fear or rewarding of inappropriate behaviors). We would also focus on punishing the bad behavior when we were also accidentally rewarding the bad behaviors too, and we would forget to reinforce appropriate behaviors at a really high rate so that these behaviors could become a habit.
Updated methods based on the psychology of learning and behavior take into consideration the motivations and emotional state the are driving behavior and factor in the everyday reinforcers—both accidental and incidental, that drive the behavior.
For a more complete discussion of dominance as it pertains to both dogs and other social animals, preview the chapter on Dominance from Low Stress Handling, Restraint and Behavior Modification of Dogs & Cats »
What Others are Saying about Cesar Millan
A Different Type of Perspective
Two of these are parodies. They are included not to mock a specific person, but to provide a different perspective on the techniques.