The Problems with Punishment
This blog is part of a series on training methods and training advice. Other blogs in this series can be found in the links at the bottom.
In the last blog we discussed how, in the training world, the definitions of punishment and reinforcement are more strict than our everyday use of those words. If you're not familiar with these, please check out that blog here:
In the above post, we talked about the "four quadrants" of reinforcement learning. These four quadrants were discovered and popularized in the cold-war era -- an era when father knew best. The scientists that developed these understandings brought a very patriarchal and mechanistic viewpoint to their work. They saw animals as mechanisms -- almost operating by clockwork. The idea was that all animals are essentially the same and that any animal receiving a given stimulus should react in the same way as another.
But we know today that animals have different experiences, different skill-levels and different personalities. In short, animals have feelings. In recognition of this, India has declared dolphins to be "non-human persons", and an Argentinian court similarly declared an Orangutan to be a non-human person with all the rights that go along with being a "person" under the law.
When we are living and working with animals, it's very important to keep these facts in mind. The actions we take with our animals have real repercussions not only on the effectiveness of their training, but on the quality of their lives. That's why I want to discuss some of the issues with punishment training here -- and as we go along, it's also important to note that many of the same concerns apply to using force during training whether it is used as punishment or not.
According to the definition we discussed in the last blog, punishment is unpleasant. It has to be, otherwise it wouldn't discourage the behaviours that led to the punishment. On the other side of the coin, reinforcement is pleasurable. It has to be, otherwise it wouldn't encourage the behaviours that led to it.
How do dogs perceive their training sessions?
Assuming for the moment that a trainer might use both punishment and reinforcement in a hypothetical training session of 100 moves, it's not surprising the animal's experience is determined in large part by the percentage of those moves that resulted in punishment and the percentage of those moves that resulted in reinfocement. We would expect the animal to experience huge difference between a session that was 25% punishment and 75% reinforcement versus one that was 75% punishment and 25% reinforcement.
If sessions are consistent over time, we should expect animals who receive mostly punishment to react very differently to training sessions than those who receive mostly reinforcements. In fact, I've heard a lot of well-respected trainers say things like "positive reinforcement encourages animals to enjoy the training process" and "animals that have been punished tend to shy away during training". I'm sure many predominantly punished animals would rather avoid training all together.
Cesar promotes punishment
The reason I've made this point first is because of the popularity of Cesar Millan's "The Dog Whisperer". Cesar's rhetoric pits the will of the dog against the will of the owner in a fight for dominance and the vast majority of the training techniques he recommends involve force and/or punishment. He makes a point of de-emphasizing any rewards in favour of an argument closer to "the ends justify the means".
As a dog training professional, this argument falls flat for three reasons:
Based on their body language many of the dogs are no happier at the end of his shows and are only reluctantly performing to avoid further punishment,
Every single one of the dogs he works with could achieve the same (or better) results with less force and punishment (and more reinforcement), AND;
It could all be done without framing everything within an adversarial relationship between dogs and their owners.
Timing is an important factor
Timing is another big issue with punishment training. Dogs are active creatures and offer a lot of different behaviours during training. It's easy to mis-time a punishment -- especially for non-professionals. When the wrong instant is punished, the dog fails to learn the intended lesson and instead avoids something irrelevant. For example, if the trainer wants to stop a dog barking by zapping it with an e-collar, then the trainer should zap the dog following each bark. But if the first time, the dog manages to bark and look towards the trainer before the first zap arrives, then what the trainer gets is a dog that will refuse to look towards him, but continues barking. The problem is, it's much harder to train a dog that won't look at you. Many home trainers who try to use "Dog Whisperer" methods end up with these types of problems.
Punishment also takes the initiative away from the dog. Dogs trained with punishment are trained to avoid punishment. When the source of the punishment is absent, the dog no longer needs to avoid those behaviours. This results in dogs who will only behave when the owner is present. And since the the dog is not being trained to do anything, but instead to avoid doing things, as punishment is applied throughout its lifetime, the dog takes less and less of an interest in volunteering new, useful or interesting behaviours.
"Never" punish aggression
The last issue I want to talk about today is aggression. It can become very dangerous to work with a dog who has been punished for disobedience or for exhibiting signs of aggression. Aggression, as distasteful as it may be to humans, is almost a universal language among animals. Most species exhibit agressive gestures, postures, growls and feints in order to communicate with each other.
Many Pooch Perfect cases have involved dogs that were punished after showing these aggressive communications. The problem is, that the underlying causes of the aggression -- i.e. the reason the animal is trying to communicate in the first place -- have not been addressed by the punishment, but the animal still receives that punishment as feedback. As a result, the animal learns to hide its aggressive communications and what you're left with is an animal that gives fewer and fewer warning signs before an all-out attack. These animals can become very dangerous over time as they learn to trust humans less, learn to trust training less and learn to hide more and more aggressive communications as they prepare to attack.
Many people say "dogs don't lie", but I'd add to that based on experience over the years. Instead, you could say, "dogs don't lie, unless people teach them to by punishing aggression".
This is one of the many reasons why Pooch Perfect does not advocate any particular "training method". Instead, we try to identify, eliminate and remediate the underlying causes of problem behaviour and build training programs around that rather than around celebrity mantras or catchphrases.
And this is only a small sampling of the ways punishment training can get you into trouble.
Let us know if you have any questions, comments or additions -- Contact us!