Why would I, a positive reinforcement trainer, want to talk about punishment? Lots of reasons. Mostly I want people to understand what punishment is, what it looks like, and what happens when we use it.
In Learning Science, the scientific definition of punishment is anything the reduces or eliminates the future occurrence of a specific behavior.
I heard one story of dog that was being trained to paw a bell on a rope to let the owner know that it wanted to go outside. It was a small dog and the bell was one of those pressed ones with four flaps rolled up with slits between the flaps and a clapper of some kind inside. Unfortunately for all concerned, the dog caught a toenail in one of the slits and that was the end of that particular behavior and training plan. Once and done. That is punishment. That punishment, whether through fear or discomfort or even pain, eliminated any future repetition of that behavior of ringing the bell.
What about when dogs are trained with chain or prong collars? One example we talked about in a Chicken Workshop with Bob Bailey and Parvene Farhoody was asking the dog to sit, but it stood or stayed standing instead of sitting on cue. One way a trainer could respond to this is to perform a “leash correction” by yanking upward on the leash causing the collar to tighten around the dog’s neck to force a sit.
The dog might sit from that correction; but, in the meantime, the stand has been punished. That correction might not have been a severe action, but it will have some negative effect upon the act of standing. And I’m sure the trainer will want the dog to stand on cue at some point in its career. That will mean having to spend more time and effort to get the stand back in the repertoire after having it corrected even once. How easy will it be to get the stand reliable if the dog is corrected several times for standing rather than sitting?
An example in the horse world is one that has never made sense to me. And that is when the rider uses backing as punishment for some behavior the horse did that the rider didn’t like or want. This might work and it might not. That would depend on whether or not the undesired behavior actually reduced in frequency in the future.
But what if the “undesired” behavior that was being punished was not wanted in that particular instance only? What if that behavior, whatever it is, is wanted in other contexts? Now the trainer has to work harder to get that behavior back when it’s been punished before. And maybe punished several times.
Let’s say that the trainer wanted a canter depart but got a rushy trot instead and he stopped the horse and backed her up. That might work and the rushy trot disappears. But what if the trainer wanted a faster trot and got canter instead, so stopped the horse and backed her up. Now the trainer risks losing the canter depart when the trainer is going to want that canter depart at some point in the future.
The other thing I’ve seen happen is how the rider asks for the back up. The rider asks for a back up and gets two or three fairly nice, relaxed steps, that are willingly given. But instead of stopping the back up, and then stroking, scratching or even praising the horse, the rider continues to ask for (but it’s really a demand at this point) more backing steps by not releasing the reins at all. The rider pulls harder, more forcefully, maybe with a quick jerk for emphasis. And horse stops moving completely, and puts its head up high or down low to avoid the pressure from the bit. The resistance builds and the fight is on.
The horse did not get any signal that backing up was the right answer. That signal could have been a simple release of the rein pressure to let the horse know that she did the right thing. Then the rider can ask for a couple more backing steps. Stop, release the pressure. Breathe. Then ask for a couple more if necessary.
In this process of not releasing the pressure, the horse has been punished for backing. And backing is exactly what the rider wanted. Now both horse and rider are frustrated, backing has been punished (suppressed or eliminated) from the horse’s repertoire, and resistance and frustration is being built on both sides.
Let’s stop and think about what we’re doing. Let’s learn about Learning Science so we can understand what we’re doing. And why. For both our sakes.