As Easy as Falling off a Bike: What punishment teaches us.

Part of the litany of modern behaviorists and trainers is that punishment doesn’t teach an animal what to do and the implication is that the results of punishment are baaaaad. Terrible. Horrible. Traumatic. Confusing. Evil.

“But just punishing the animal for doing something else does not teach it to sit. At most, punishment only teaches it what not to do.” Murray Sidman Coercion and Its Fallout (1989)

It does not teach the dog WHAT to do.” Pat Miller, Pitfalls of Positive Punishment

“Punishment teaches an animal what you don’t want it to do but fails to teach it what you expect of it.” Valerie Tynes – veterinary behaviorist

In all cases, these people limit the discussion to incorrect applications of punishment by untrained people. As there is no course, text, instructor, practical training, internship or certification in the use of aversive control, they are all blowing smoke. Neither Sidman nor Tynes were ever trained in the use of aversive control by way of their academic studies. They aren’t experts in any practical sense of the word. They got their information elsewhere from unnamed sources – the epitome of anecdotal information. Despite their unwillingness to tell us where they got their knowledge they demand to control the conversation by implying academic imprimatur. Their knowledge is certainly not derived from the successful application of aversive control. Miller says she used to use traditional training methods with punishment  long ago but now doesn’t find a need for it because she is so successful with “positive” methods. That literally means she finds no need to ever stop a behavior immediately or prevent its future occurrence. (Look up the word pica. It is common to dogs. It is most often fatal. Why someone would not suggest stopping a dog from eating inedible objects, I have no idea, but I guess Pat doesn’t see the need. Emergency vets cut dogs open, regularly, because of intestinal blockages. That’s reason enough for me to need to know how to stop the behavior. If you stop it, they live. If not, probably not.) All of these experts have made a simple intellectual blunder. They have confused the application of a behavioral effect with the behavioral effect, itself. As they are not familiar with the correct use of the tool, they attack the tool. I could argue this point by mentioning that punishment stops behavior and it is illogical to suggest that it should teach a behavior. The brakes on your car cannot logically be criticized because they do not propel the car. However, I’ll take the high road and contradict this concept fully. Punishment teaches all kinds of things – if you aren’t an ideologue.


What Punishment Does Do:
Many children learn to ride a bike with training wheels. Training wheels are “positive” because they teach the child to peddle and steer without a risk of falling. Then, one day, the training wheels come off and the child crashes. Ouch. Now we have a problem. The child does not know what to do . Why? We used positive reinforcement, didn’t we? (That is the first hint that the anti-punishment folks haven’t really thought about what they preach. They proclaim that positive reinforcement is a panacea. When if fails, they do not acknowledge the failure the way they always assail punishment for any percieved imperfection. In this case, the selection of a “positive” solution has put the child at risk. The kid thinks he knows how to ride a bike and has no caution. i.e. The fall may be more damaging because the child is uninhibited and has a long history of success. Nobody considers using concurrent but opposite consequences to teach as nature teaches. i.e. Alternate between training wheels and holding the handle bars to help with balance. Gradually lessen your support. ) The child doesn’t know what to do because the training up to that point has been exclusively positive and limited to tangential skills – peddling and steering. The real skills necessary for riding a bike are taught by punishment and negative reinforcement. These skills are keeping your balance, and paying constant attention to keeping your balance. Then you learn to avoid steering into immovable objects – also taught with aversive control. Because the skills are taught with aversive control it is commonly said that once you learn to ride a bike you never forget how. Positive methods, by contrast, cannot create lasting behaviors without almost constant reapplication. Name a behavior that is taught with exclusively positive reinforcement where you can do it forever without retraining or additional reinforcement. Crickets chirping.

To anyone other than an ideologue it should be clear that many skills are created through a concurrent influence of reinforcement AND punishment. It is impossible to “teach” a behavior with punishment because that isn’t what it’s supposed to do. However, it is impossible to teach masterful behavior with exclusively positive reinforcement. (Why would someone wish to be mediocre?) They also ignore the vast shortcomings of exclusively positive training and control – meaning they hold the two behavioral effects to different standards. I can’t say logical standards because there is no logic in anti-punishment ideology. For instance, in some cases exclusively positive control can get somebody killed – like driving a racecar wrecklessly or running a construction crane while not paying close attention. (In case you’re getting a clue, the anti-punishment people exaggerate everything to the level of abuse and trauma while ignoring the tragic results of their ideology. If they were logical they would extrapolate the effects of reinforcement and punishment and then speak in more moderate terms of both.)  The most famous quote associated with an assumption that nothing will go wrong (meaning an absence of negative consequences) is “Challenger, go at throttle up” – seconds later the space shuttle Challenger exploded. As no shuttle had ever exploded before, NASA administrators failed to recognize the potential of “negative” consequences. No one was punished for this decision. Because of a lack of punishment, they lost another shuttle in the same way – they knew that ice falling off during take-off damaged heat-shield tiles – and did nothing about it. One more shuttle and crew destroyed because the NASA administrators were not punished for their former error. Sad.

What anti-punishment people do not realize is that the word “safe” implies performing behaviors in a world that can harm you or someone else. By learning to deal with aversive events we become “safe.” Meaning the “learning” occurs as components taught with reinforcement and punishment are combined into a functional skill. No use of exclusively positive reinforcement can do that. I would never fly in an airplane with someone who had only flown in simulators. I would never use a veterinarian who had never had something go wrong during a surgery. Dogs that have never fallen off a small rock cannot be trusted to go mountaineering or run agility courses. Many agility dogs are injured each year because they are never punished for going at unsafe speeds. The list is endless.

The real issue here is the inability of people to understand that it’s all about developing a safe and efficient working repertoire for our dogs through the fastest, safest and most permanent means possible. Sometimes that means stopping a behavior, cold. We must do that because some behaviors are lethal on a first occurrence – like a dog running across a street to fetch a ball. Does the positive reinforcement associated with chasing a ball “teach” the dog to stay out of the street? Nope. Does the positive reinforcement associated with driving a car prevent you from running a red light? Nope. Does the positive reinforcement from winning in a casino teach you how to put the money in the bank? Nope. To be effective in such situations we must learn lesser behaviors that give us a chance of performing in more complex or dangerous situations. This is an anaology to medical prophylaxis. It is too late as the dog runs into the street to teach a rock-steady recall or “stay”. It is too late to teach an inhibition to ingesting socks once the dog is on the operating table. It is too late to teach NASA administrators to stop a launch because of unsafe conditions after they’ve given the command to launch. The time to teach those things is necessarily before the exposure to potentially deadly consequences. Using a lesser, safe and humane aversive consequence to teach safe, effective performance is the issue. This is no different than vaccinating a puppy for parvo virus before it is exposed to the disease. Yes, you are literally sticking a steel needle into the pup’s body. You are causing pain and sometimes fear by doing this. Apparently that is cruel. 

Neither reinforcement nor punishment can be held to a standard that limits our involvement to teaching new behaviors using only one behavioral effect. The expert ability to inhibit behaviors is vital to keeping our dogs and neighbors safe. The push for all positive training and behavioral control pitches a utopia where we can cure all problems with treats and a dog will never experience anything negative. No rational person would pursue that goal. No logical person would suggest it.

Leave a Reply

Your email address will not be published. Required fields are marked *