The Problems with Punishment

Last week I explained why I use positive reinforcement with the dogs and people I teach. This week I’m diving deeper into punishment and its possible fallout.

With positive reinforcement we teach our dogs what to do and build a reinforcement history so that they continue to choose those behaviors. When we use punishment to train, we stop behaviors whether by adding something aversive like a choke chain or taking away something the dog enjoys like our attention. Although punishment might seem to help in the short term because the behaviors stop, it introduces several potential problems that affect your dog’s health and well-being.

I often use this analogy when describing punishment to my clients: let’s say you've come to my office for a job interview. You walk into the room and as you are about to sit, I say, “no, not that chair.” So you choose another and again I say "no." And then it happens again. And again. By the time you choose the "right" chair (without any help from me), how are you feeling? How well do you think you'll do in the interview? What if you got a shock when you sat in a chair? How soon before you stop trying to sit? Now imagine a dog repeatedly being told no or administered a shock--how soon before they stop trying anything? Punishment often shuts down all behavior and a shutdown dog is not a happy or healthy dog.

Another potential problem with punishment stems from how the dog views it. Just as with reinforcements, the learner decides what is punishing. For some dogs a stern “no” can be traumatizing while for other dogs a choke chain tightening around their neck does nothing to stop them from pulling or a shock collar has to be turned up to dangerous levels before it changes their behavior. For these dogs, using such tools can cause physical harm like collapsed tracheas or serious burns.

Dogs subjected to punishment can develop traumatic associations with things or people in the environment when the punishment occurs. This works just like classical conditioning--whatever goes before the painful or frightening stimulus starts to predict the fear or pain itself. A child walking by while the dog receives a painful or startling leash correction, for example, may start to trigger a reaction from the dog when he sees children if it happens often enough. We work hard to undo the fear-based associations that our dogs develop, so why would we want to create more of them?

I avoid negative reinforcement, too, because it works by removing something that the dog finds aversive to reinforce a behavior--the dog stops pulling and the choke chain relaxes, for instance. So it’s hard to use negative reinforcement to train a dog without also using positive punishment.

As a force-free trainer I cannot in good conscience apply punishment to the dogs and people in my care, including my own dog, Ralphie. And there are decades of evidence to show why we should leave punishment behind.