Trained to endure or trained to participate?

In the 2016 Clicker Expos Eva Bertilsson, Emelie Johnson-Vegh and I gave a lecture on the topic of giving animals choice and control in their training. For me it is a significant way to refine and optimize how we use clicker training in our daily lives with horses. I would like to describe more about the process, but first a bit of background.

We’ve been taught that classical conditioning and operant conditioning are always part our training. They “team up”, so to speak, when we train. For example, the sound of the click means food is coming and the animal begins to anticipate the arrival of food after it has experienced enough pairings of the sound/food combination. The sound is “classically conditioned” as a predictor of food.

On the other hand, when a behavior is trained with operant conditioning, the animal needs to respond to a given situation by presenting a behavior which, in turn, results in receiving a reinforcer. In other words, the consequence of the behavior (usually food) strengthens the behavior that happened before the food arrived and the animal is more likely to repeat what it did before it got the food.

OK, that’s sort of the “cliff notes” of the process. Now I’d like to share some thoughts on how I see this information interpreted. For example, I hear people talk about wanting to “desensitize” their horse to something, like a plastic bag, tarp or maybe clippers. They go about the process by taking that item and, hopefully in a skilled manner, they introduce that item until the animal no longer perceives it as a threat. Some people try to use flooding to desensitize the animal, and often times this becomes traumatic or harsh. In a more passive manner, the desensitization process can occur over time and presentation of the item in the environment can predict the arrival of good things.

This latter kind of training is very effective when done well. However, the situation I see often with horse owners is that they tend to progress too quickly. The horse is restrained by the halter/lead and the owner will walk directly into the horse’s space and rub the plastic bag on the horse, then try to offset the fear of the item by handing the horse food. I see this often online and I don’t think the handlers/owners realize that they can actually nullify the value of the food as a training tool if they continue to associate it with something the horse really doesn’t like.

Because of the availability of videos on the internet, I know that the visual input of people walking to their horses with a tarp in one hand and food in the other is seen everywhere. In some ways I think this interpretation of Classical Conditioning got started within our horse circles because of our former training backgrounds. We are USED to seeing people hang onto the lead and let the horse “drift” or “move his feet” as an attempt to avoid the scary item. Plus everyone knows that if you release the pressure when the horse is pulling, you increase the likelihood of reinforcing the pulling, so we’re not supposed to let go when they pull. The halter/lead is held tightly so the horse cannot escape, then the clippers are presented and the pressure is released when the horse stops moving, even slightly. Clicker trainers see this and, trying to introduce food into the training environment, they will then give the horse food after the release of pressure. This method is sort of a cross between negative reinforcement and systematic desensitization with a few bites of goodies thrown in for good measure.

There is an alternative way to help an animal become used to the things humans introduce to the environment. One can focus on using Operant Conditioning, and more specifically positive reinforcement. This is the manner in which I do almost all of my training for potentially scary items. A great example of using R+ in training a different response to a situation was shown in a video during a class I took with Dr. Susan Friedman. In the video we saw a lion who was reacting aggressively (loud vocalizations, lunging at the caregiver). Using a target, the keeper was able to train the lion to wait in a down-stay position and earn reinforcement. The lion learned that the down-stay earned reinforcement, henceforth at the onset of the session there was a marked change in the lion’s behavior. It became “calm.” (It stationed near the bars with its nose near the target, no vocalizations and no lunging.) This is an example of training a behavior that produced a safer environment for the handler and involved less stress for the lion. Again, the technique used was based on Operant conditioning, not Classical.

This takes us back to the opening paragraph, and the explanation of an ever more refined approach. I have taken this type of training technique and added some dimensions which give the animal even MORE control over the process, meaning the HORSE is in charge of the pace. I let the horse determine whether or not to proceed. I do this by setting up a series of “start button” behaviors that signal to me that the animal is receptive to a process and is inviting me to continue with whatever I’m doing. Once the horse initiates the signal, I then introduce whatever item or event I want to work with.

I’ve got some video which shows a horse having a fly mask put on. In the video the behavior looks so easy, fluid and well trained. The process of getting there included many steps of allowing the horse to approach the fly mask and those efforts were reinforced with food. This means I was clicking him for any approach to the fly mask, and allowing him to leave if he showed signs of fear. In fact I took him OUT of the environment, away from the mask, many times during the training, so that he could choose to approach the mask again.

I included ways for the horse to refuse the process if he was concerned. This ability to refuse or back away is the aspect that I personally found sped up the process and made the behavior very solid. The HORSE remained in control of how, when and how fast the mask was put on. One might say that this choice “empowers” the horse; it certainly appears that way to me.

Some of the behaviors are complex and involve creating several built in opportunities where the horse can signal that he is ready to continue at various stages in the procedure. An example of this is creating a behavior that allows a dental exam. The sequence looks like this on video and has many steps to the overall behavior.

1. I position myself in front of the horse.

2. When the horse looks at me I put my hand up with my thumb near his upper lip.

3. The horse puts his upper lip on my thumb, which is his signal to me that I can continue.

4. I then spread his lips and can begin to open his mouth.

I’m building duration and variety into this technique. At any point he can stop. I just go back to an earlier level where he can accomplish the goal.

The point is we are not just teaching the horse to ENDURE the process, we are teaching him to actively engage in it, complete with choosing to participate or not. This type of positive reinforcement helps in building every type of behavior, not just ones that contain a possible aversive.

Please take the time to explore the videos and see the subtle differences. Enjoy!

Advertisement