Wednesday, July 19, 2017

Is compulsion necessary for some dogs?

Not surprisingly, I discuss dog training quite a bit. I have discussions with all different levels of trainer from very skilled and experienced trainers to dog owners just learning about how dogs learn. In talking to people about various methods, I frequently have people tell me that they have a breed of dog that requires some compulsion. By compulsion, I mean the use of physical corrections or force. Sometimes from people that feel that some compulsion is necessary with all dogs and some people that feel that some compulsion is required only for certain breeds of dogs. When these comments are made, the dogs that are being referred to are almost always large working dogs or sporting dogs. Although, to be fair I have heard it said of dogs from nearly all of the groups except possibly the toy dogs. It makes sense as many of the dogs from these groups are large, strong, powerful and sometimes quite self assured. I understand where they are coming from, but I don't necessarily agree with this viewpoint.

Many working and sporting dogs (and even dogs from some of the other groups) are large, powerful and committed to their work. Terriers aren't always big dogs, but they are frequently extremely committed workers who are bred for persistence. While some might consider dedication and perseverance the same as being stubborn or hard headed, I would argue that these are necessary traits that we bred into these dogs in order to do the jobs they were bred to do. How can we, in good conscience breed a dog to be a certain way and then punish them for exhibiting those breed traits? Because some of these dogs can be less sensitive to corrections than other dogs it is taken for granted that they can "take it". And, because dogs are the brilliant, honest and forgiving species that they are, they will take it, but that doesn't mean it is necessary or even that effective actually. In fact, I have seen corrections be only mildly annoying to extremely frightening and not have any impact on the dog's training.

A word about Punishment
In learning theory we know that reinforcement causes a behavior to happen more and punishment causes a behavior to happen less or to stop happening altogether.

Positive Punishment (+P) is when an aversive is added which causes the behavior to stop happening or go down in frequency. Examples of intended +P are squirting an animal with water, hitting an animal, giving a leash correction.

Negative Punishment (-P) is when something that the animal wants is removed and unavailable which causes a behavior to stop happening or go down in frequency. Examples of intended -P turning away from a dog that is jumping up, withholding a treat, time out.

By definition, punishment decreases the future probability of a behavior. In other words, if you punish a behavior, that behavior should not happen in the future. If the behavior is "punished" but continues to occur, it has not been punished out, it has simply been suppressed in that moment. There is a big difference there. You could be attempting to punish a behavior repeatedly, but if that behavior is continuing to happen you have not punished it and really, the dog has learned nothing. That is one reason why I don't like to use positive punishment. Punishment needs to be severe enough that you don't need to do it more than once, twice at the most. I understand that when an unwanted behavior is happening and a dog is corrected and stops doing the behavior it may appear that you have been effective, but unless it affects the future of the behavior, you have done little to change the behavior. For instance, I frequently see people use a squirt bottle to stop a dog from vocalizing. The dog barks, they squirt the dog and the dog stops barking. However, unless the squirting was very aversive, the dog usually will bark again and will stop only when you squirt him or perhaps when you threaten him with squirt bottle in hand.

Before anyone thinks that I have small dogs so I must not understand large working dogs, please know that I work with all different breeds and I use the same methods for all of them. I live with Dachshunds, an independent free thinking breed known for being difficult to train and have lived with several Australian Cattle Dogs, Rottweilers and Chow Chows. In fact, the first dog I trained without compulsion was an Australian Cattle Dog that I got as a puppy and committed to attempting to train him without corrections. And, I did start out using what some call "balanced training" where I combined both physical corrections with reinforcement. I have experienced correcting a behavior and seeing that behavior stop in the moment, but I have also learned that it never changed the behavior long term. At best I was managing much of the time. The prong collar was managing the dog, not actually training them not to pull, it simply stopped pulling in the moment, but when I got decent heeling, for instance, it was the reinforcement that built and maintained those behaviors, not the punishment of pulling.

When training dogs, of course  I always take a dog's breed into consideration along with the breeds natural traits and purpose. I may adjust how I get a behavior in the first place or the reinforcement that I choose, but I do not ever feel I need to use compulsion. The reason that I wanted to write this blog is that some people seem to feel that because of their breed some compulsion is necessary and I am saying, using compulsion is a choice. I am trying to choose my words carefully so that it doesn't come across as a judgement call. I am not trying to judge or say that if someone is using compulsion they are bad, don't love their dogs or are a poor trainer, I am simply saying, you could also do it without compulsion if you wanted to go that route. I don't want people to be under the false impression that if they have a certain breed they cannot train without compulsion. I have met many dogs whose owners use some compulsion and it doesn't always mean that the dog is terrified or doesn't enjoy working or doesn't spending time with the owner, however, sometimes this is the case. What I am suggesting is why not try it without correcting the dog? I am suggesting that people look very closely at their punishment and look at the future of the behavior to measure whether or not you are actually punishing out what you don't want. You may find that simply reinforcing and strengthening your target behavior, which should be an alternate behavior to the one you don't like, may get you where you want to go and could do so with less frustration and possibly even faster.

Because we are on the topic of punishment, there are a few other things to consider.

Punishment Callus
After repeatedly using an aversive an animal will develop a punishment callus. This means that the aversive no longer suppresses or stops the behavior. In fact, it pretty much stops making an impact to the animal at all. This is one reason why people will frequently increase the intensity of the aversive because they are ineffectively using it but seeing no results at all. But remember that if the future of the behavior is not being impacted you are essentially doing nothing except stopping the behavior when it is happening, but not necessarily teaching the dog anything.

The effect on the Punisher
This is something that not everyone knows or understands, but it is important when talking about positive punishment. Punishment is reinforcing to the punisher. Sometimes it is because it "works" in the eyes of the punisher. In other words, the behavior occurs, the animal is corrected and the behavior stops so the punisher feels that it worked. Again, suppressing behavior is different than changing the future outcome of a behavior, but still, that suppression can be reinforcing to the punisher. But, here is the uncomfortable part, issuing punishment can actually be reinforcing regardless of the effectiveness. When someone is frustrated or upset by an animal's behavior, they issue a correction and they feel better. It is almost as if it feels good to hurt them because they are frustrated or upset or angry. It is no different than saying something hurtful to hurt someone because they are upset, except that they are using physical discomfort instead of words.

Both the punishment callus and the punishment being reinforcing to the punisher can become a loop of lots of corrections with little affect at all on the future of the behavior. This is sad because it can cause a lot of frustration on the animal and the person and can affect the relationship between the dog and owner.

So, here is what I am proposing. We know that all the quandrants of operant conditioning can change behavior, but generally it is positive punishment and negative reinforcement and the use of compulsion that creates behavioral fallout and relationship damage. If we know this to be true, why not stick to positive reinforcement and negative punishment? Can it take longer? Sometimes, but generally not much longer because you are building the behavior as soon as you are getting it and the more you reinforce it, the stronger it becomes. Give it a shot! You can always decide to add compulsion later on if you really feel you need to, but hopefully you won't.

All this being said, I would be remiss if I didn't mention that poorly executed positive reinforcement can also create frustration, confusion and a fractured relationship between dog and owner. Any poor training will elicit poor results, but the fact that there is some really poorly executed positive reinforcement training out there is not a good enough reason for me to use compulsion.