Cover Photo: Hamilton Zoo, New Zealand
Might be thought provoking but that’s what we are all about!
Positive reinforcement training is all about focusing on the behaviours the animal does correctly. Through successive approximation plans we try to assure the animal will not make any mis-movements. Adding a motivation strategy to this makes the animal choose the correct answer on our small given questions. This is all the theoretic talk we talk about but is it really that its not accepted that the animal makes a mistake?
When developing a new show in 2005 in the Netherlands as a young trainer, I worked with 7 Californian sea lions. This was an exciting learning curve for me as a new trainer. Trying to engage and observe as many behaviours as I could I tried to learn and learn and learn. At one point we trained our animals for a crisscross ball jump. The animals kind of knew the ball jump itself but we had to work on the height.
The crisscross worked and the animals didn’t seem to mind swimming towards each other and jumping to a ball but when we had to get the ball higher we came to across small challenge. One of the trainers suggested putting the ball higher, which is an obvious decision, but what they meant was to put the ball so high that the animal was not able to touch it on the first try. Touching the ball gives the animal the reinforcement it wants, this means that the animal understands that when it doesn’t hit the ball it has to speed up. The animal received the signal and went for it and what as predicted, the animal was incorrect. We then asked for the behaviour again and this time the animal hit the ball as hard as it could.
Was it okay for the animal to make a mistake at that moment in time, even if we knew it would happen?
A couple years later working the same animals but this time at another facility one of the trainers explained that we should let animals know what a mistake means.
This made me stop and think because I was always told we want our animals to have a positive experience in training sessions at any given time. Of course we do, but what would happen if an animal doesn’t accept failure? We had one animal named Cleveland, a young female Californian sea lion, who didn’t. When being asked a particular behaviour and Cleveland responded incorrectly she would run to another trainer and bite them in the leg. Animals are not robots and we can’t put the pressure on them to always be good. We solved Cleveland’s problem behaviour by teaching her that it is ok to make a “mistake” as long as you want to participate in the further training.
We believe that we can teach an animal that it’s ok to make a mistake. It’s not only animals we have to teach, us humans have a hard time accepting failure as well. When we have too many failures we can become insecure about situations we are involved in and it can be difficult to make a proper confident decision. We like to talk about the good we do but not wanting to acknowledge the mistakes we make. Just like our animals, it’s okay and common to make mistakes.
Everybody is different and everybody sees things differently. My current role allows me to discover peoples behaviour and I see over time, that the hardest part of training trainers, is to build a strong trust account that they are able to tell me the challenges they have and what mistakes they make. When we can accept we can accept our mistakes and look at the solutions ahead of us, life becomes so much easier.
What if we train the animals the same way?
This is where the LRS could come in. A 3 second pause to show the animal that he is incorrect but still able to potentially get reinforced because of the motivation the animal has to continue the session. This gives the animal an idea of accepting failure as long as you want to keep going.
What about trial and error? The team at Zoospensefull have the privilege of working with many different species during their day. It’s a motivator for us to see people progress with their animals. It’s fascinating to observe the animal when we let them think a bit longer or when the trainer jumps to quickly into redirecting. What we find interesting is when you wait a couple extra seconds and the animal starts guessing what you want it to do. This helps the animal problem solve but also to make the choice of wanting to find out what we want.
The video below shows a neophobic Scarlet macaw being target trained for the first time. Since these birds are incredibly sensitive to new objects it’s often difficult as trainers to know what they will be comfortable with and what requires a little more work. By holding the target stick at a distance, it is up to the bird to choose to approach to get reinforcement. In the video you can see her think about whether she should touch it or not. This helps us gage her comfort level and tells us how quickly she is willing to progress. It also allows us to take large steps and gets her to think more about what actions will result in reinforcement.
The glow and sparkle you see in the animals eyes is amazing when they have it right. This allows the animal to make a mistake, and this allows the animal to try it again on its own. When it doesn’t work, life is good we find something else to make the animal good again.
There are many different ways to go, but we should not only focus on the positive, animals should be allowed to make mistakes for them to understand how to deal with the failure them themselves. Otherwise we have animals who developed learned helplessness quickly when they do not know how to solve the problem themselves.
In Loro Parque I worked with a Killer whale name Adan. The animal was handreared by trainers because the mother didn’t accept him. His training started from the beginning, when it got his bottle. Over the next few years we started to train him for more challenging behaviours. The biggest challenge was that he never was taught to accept failure. What often happened after a couple incorrect responses was he would just swim to the other side of the exhibit and not come back. A classic example of learned helplessness. We solved this by starting to reinforce to accept his failure by reinforcing the LRS slowly but surely exactly like what happened to Cleveland the Californian Sea Lion and Marineland Canada.
Remember every person and every animal is different. Build a proper relationship to discover what is ok for each individual animal. As trainers you need to teach your animals it is okay to make mistakes it’s all part of the learning process!