If you read my first article about extinction you’ll know that I think it is an overly-hyped suggestion that doesn’t work as advertised to control behavior. Now I will offer the proof of the pudding. A quote from someone who knew that even punishment needs to be reapplied to suppress a behavior. You may be surprised about who acknowledged that fact. It may also caution you about swallowing “what they say” without first using your own knowledge and using a little bit of logic. Before I give you the quote, think about that for a second. Is withholding reinforcement more or less powerful than tangible punishment? People who oppose the use of punishment think so. That is why they have no problem using deprivation to control behavior. (That such deprivation is actually called negative punishment doesn’t resonate with their ideology but is nonetheless true.) so for the moment we will merely concentrate on whether withholding a desired result is a powerful tool. Here’s what B.F. Skinner said about punishment in Science and Human Behavior.

“More recently, the suspicion has also arisen that punishment does not in fact do what it is supposed to do. An immediately effect in reducing a tendency to behave is clear enough, but this may be misleading. The reduction in strength may not be permanent.”

If punishment, the boogeyman of normative hedonists, “does not in fact do what it is supposed to do” because it must be reapplied, then extinction analyzed by the same yard-stick comes up woefully short of its promised benefits. (So does positive reinforcement on the flip-side.) Yet Skinner cautions about all the horrible side effects of punishment. He must have thought it more powerful than extinction. Behavior analysts don’t talk about the horrible side effects of extinction like rebound aggression, hatred, recalcitrance and active resistance. They rarely acknowledge that those things are a direct result of depriving people of expected rewards. Marie Antoinette is known historically for failing to understand how deprivation can lead to violent revolution. One wonders why these easily observed reactions to negative punishment are never held up as objections to things like time-outs or withholding privileges.

The Hole in a Bucket Eventually Empties It:
To achieve extinction, the procedure must be continual and without exception, like a water-tight vessel. If the creature is ever again reinforced for the unacceptable behavior it comes roaring back to life. Pavlov demonstrated that extinguishing a conditioned association does not remove the connection in the dog’s brain. When you again add reinforcement the association comes back at 100% strength. This is called a spontaneous regeneration but that is a mistaken concept. In reality, the association was there the whole time, just as water in a bucket doesn’t disappear because it is contained. The behavior simply wasn’t functional when the reinforcing contingencies were removed. This is a truth of nature. Associations have properties that are not dependent on the whims of academics. Once you learn something can lead to reinforcement it does not go away. It is submerged artificially. If you wish to test this, try to remember the name of an elementary school teacher you liked or hated – but haven’t been “reinforced” for knowing in many years. Any association that is committed to memory will remain, reinforced or not. No one can make me forget Ms. Colignese, my kindergarten teacher, by removing reinforcement. Likewise, no one can remove my ability to fire a field artillery piece though I have not been reinforced for that behavior in almost 40 years. If I had the opportunity to do that, tomorrow, I’d go for it. Why? Because it’s fun. I haven’t done it in 40 years because I have had no opportunity to do it. Put $1,000 as a reward for successfully firing one and I’ll trot out why I was the Field Artilleryman of the Year for the Washington National Guard, a long time ago.

The Ethical and Practical Dilemma:
The real problem with extinction is that it requires constant diligence, takes forever to achieve and can’t be trusted to suppress behavior in the future. Virtually all behavior problems have a time limit on solutions and a limit on diligence. Redirecting a dog will work in all cases all the time – but no one has “all time” to devote to their dog. If you recommend a procedure that is actually the simple by-product of acknowledging that a behavior is unacceptable you need to realize its limitations. There are some behaviors that should never be reinforced. There are some behaviors that should never be punished. There are some behaviors that should always be punished. There are some behaviors that should always be reinforced. Life rarely allows you to live up to those four rules. If you are working with a dog that has been hit by a car and chases sheep, goats, horses and cars you have to find a way to stop the behavior immediately. Pretending that you can remove the sheep and stop the behavior is most often a fatal fantasy. How do you remove the reinforcing effects of the sheep, goats and cars? Lock the dog in a prison. And now another shoe drops – extinction doesn’t work on instinctive behaviors.

Healing the Heeling Heeler:
Take your average Australian Cattle Dog. They are called heelers because they bite the planted heel of the cow they are trying to drive. It comes with the dog. It does not require reinforcement, merely opportunity. Withholding reinforcement will not influence the behavior any more than withholding reinforcement for scent marking will stop it. What if you have such a dog and you wish to stop it from becoming angst ridden when it sees a cat walk by? You cannot remove the angst unless you remove the cat. The behavior does not weaken due to habituation. (If 2.5 years is any indication of how long you have to wait for some diminution of the behavior, don’t hold your breath. ) What if the owner has a cattle dog that was raised with two Bengal cats? What if the owner isn’t going to get rid of the cats but wishes to keep the dog? If you are a traditional behavior analyst or “modern scientific trainer” you have no answer. You cannot provide that service. Reinforcing the dog for “alternate” behavior doesn’t stop it. If you stop reinforcing the alternate behavior the dog reverts to its breeding.

The Ethical Conundrum: Extinction as a practical tool
Now that we have started to examine the limitations of extinction as a go-to method of controlling unacceptable behavior we hit an ethical brick-wall. If you cannot modify that behavior with any success yet accept money from someone to solve that problem you have earned a title not generally desired – fraud. That is the problem with modern training ideology. They claim to do things that simple logic and testing disprove. They attack people for presenting evidence that their methods don’t work. They actively oppose the use of punishment because it stops behavior immediately. They attack people who use method capable of controlling such behavior as using a “short cut”. They cite horrible “side-effects” of punishment without either knowing how to use it or speaking of it in rational terms. They do not sight the horrible primary effects of failing to stop behaviors in a timely fashion – like a Husky clobbered by an F-150 pickup truck. But they take the money anyway. When their methods fail or become burdensome beyond any rational commitment, they blame the victim – the owner who wasn’t willing to lock down their dog 24/7, get rid of the cats or deprive their dog in some other way. How selfish.