There are two polarities of thought in the world of behavioral control. 
That’s a problem. Behavior is not a one-dimensional phenomenon. It’s almost never a choice of positive or negative to solve a single behavior problem, let alone to guide a life. That is because it’s about the human or animal you are trying to teach or help. It’s not about your personal preferences or ideology. It’s not about your laziness or passion that causes you to be skilled with only one end of the spectrum. If you say the words reinforcement and punishment with any emotion, you shouldn’t be offering behavior services. If you use those two poles to guide your efforts you will paint yourself into a corner. Using an either/or mantra may put money in your pocket or elevate your status but it doesn’t help the person or animal who needs the help.
Punishment: The Lethal Apple
If you offer behavior or training services, the word punishment can destroy your business. Anti-punishment zealots abound. They will try to hurt you if you speak about or use punishment. They know that the word has a scientific definition but they have swallowed the bait set by ideologues. They hear ‘punishment’ and automatically think ‘abuse’. No right-thinking person can tolerate abuse so they have no problem trying to destroy abusers. This twisted logic is not an accident. Behavioral scientists, educators, social workers and animal trainers have spent 75 years fusing the words punishment and abuse. Now there is no word to describe environmental influences that cause behavior to decline or stop. That is because it’s not about a natural phenomenon anymore. It’s about what humans intend to do. A human may not intentionally apply aversive control to change the behavior of another. That is portrayed as always causing harm. Somehow an organism can experience pain, fear and physical damage falling off their bike without noticeable behavioral trauma, but spanking destroys them. As long as a human hand has nothing to do with the pain and fear, it’s OK. The problem is that this conclusion, however popular, is a fantasy. To believe it you must ignore reality and scientific evidence.
““One possible side effect of punishment is the production of emotional reactions. Several researchers have demonstrated that punishment with electric shock does not produce lasting emotional results. (Hearst, 1965; Hunt & Brady, 1955) Further, as Azrin and Holz (1966) have pointed out, gross observation seems to indicate that no chronic emotional maladjustment is engendered by a child’s having been burned by touching a radiator or having skinned a knew by falling off a bike. It is fortunate that punishment does not produce strong lasting emotional effects under most circumstances since it would be impossible to eliminate punishment from the natural environment. Although it is true that emotional behaviors are frequently observed following punishment, it is also true that these effects are usually short lived.” Dr. Ron Van Houten, The Effects of Punishment on Human Behavior, Axelrod and Apsche, Academic Press, 1983.
What are the most common assaults on punishment? It causes emotional side effects, fear and pain. These are assumed to be damaging, yet a common observation of the process shows the actual effects of rational punishment are primarily beneficial. When people observe the reckless behavior of adolescent boys they invariably imagine it will end with someone getting hurt. That is a correct projection. The child who does not fear falling or crashing their bike is going to end up terribly damaged. The reason this kind of behavior ends in tragedy is that there were no moderate punishments for unsafe behavior. The skinned knee didn’t teach the kid the proper devotion to safety so he ends up with a shattered arm. (Ironically, it is most often the effect of social positive reinforcement that builds reckless behavior to that end.) Somehow the transitory experience of fear and pain is far more terrible than what happens in their absence. The concept that punishment can be moderated to fit the lesson is absent – though all of the people who oppose punishment have benefitted from its blessings. Step in front of a bus, just once, and you will regret that you were not punished safely for such behavior. Fail to wear safety goggles while using power tools and you will wish someone had punished you for their absence. While obsessed with ‘terrible side-effects’, people who oppose punishment ignore its biological purpose and evolutionary function. Unpunished people die, or worse, kill others.
If that last sentence seems too broad, consider my first point. Life isn’t about all or nothing. It’s not about cherry picking events to prove an ideological position. I simply did what the anti-punishment folks do. I used the worst things that could happen as my example. No, not all unpunished people die, but some do. No, not all unpunished people kill, but some do. That is the intellectual trap. To advocate punishment is as stupid as opposing it unless you include a context. One cannot logically oppose gravity because they fell off a house. One cannot logically advocate gravity as their airplane loses power. Yet that is exactly what behavioral zealots do. They propose that a single polarity – positive reinforcement – is preferable in all cases and conversely, punishment is always abusive.
The Panacea: “Take a bite of the apple, my pretty” – The Wicked Witch to Snow White
In the history of mankind, the most evil, terrible, destructive and murderous behavioral effect is positive reinforcement in the absence of punishment. It is responsible for every conqueror, pirate, robber, thief, dictator, serial killer, stock swindler, card cheat, bully, mobster and sadist that ever lived. Lives destroyed at the altar of positive reinforcement numbers in the hundreds of millions. Adolf Hitler is considered evil because he killed 12 million people for his own personal satisfaction. That’s chump-change. Josef Stalin starved six million Ukrainians in three years and millions more over the course of his rule. Chairman Mao starved 14 million to create his “Great Leap Forward.” The Spanish conquest of Mexico killed millions to gain gold and souls for conquistadores and priests. You can examine history and find countless examples of the power of positive reinforcement wiping out civilizations, creating monsters and enslaving the masses. However, the same process creates the petty thief, the liar and the swindler. It runs through the whole spectrum from a simple lie to genocide. Despite this track record, anti-punishment people have a single recommendation for solving all behavior problems – positive reinforcement in the absence of punishment. Every villain in history was reinforced, positively, for their villainess behavior. Despite this track record, positive reinforcement is offered as the shining apple – filled with lethal poison. Positive reinforcement increases any behavior it touches. It has no soul. It has no morality. It is a natural phenomenon – without pity or restraint. When left on its own, it turns Snow White into the Wicked Witch.
Now I shall ask you a simple question that is rarely voiced. Who said one can only use a single polarity? Why is it reinforcement vs. punishment? Why is punishment always styled as the Wicked Witch? Why is positive reinforcement always Snow White? The Queen wasn’t evil until the years of positive reinforcement for being ‘the fairest in the land’ shut off. Wait! Oh, my gosh! Removing expected positive reinforcement can cause aggression! They never mention that. Oh, the horrors! If you think my mock shock is inappropriate consider the scientific conclusion that is never mentioned. Withholding reinforcement can have very destructive consequences.
“The issue of elicited aggression is complicated even further when one considers that elicited aggression is also produced by extinction and some schedules of reinforcement. Azrin, Hutchinson and Hale (1966) found that, when pigeons responding for food reinforcement were put on an extinction schedule, they attacked restrained target birds that were situated at the other end of the chamber. Schedules of reinforcement that involve long periods without reinforcement have also been show to elicit aggression.” Ron Van Houten, The Effects of Punishment on Human Behavior
Again, common observation and science say one thing and current culture contradicts both. (In a bizarre twist, behavior analysts helped create our current cultural opposition to punishment and now claim that the public won’t let them use it.) If you wish to test this, try to take a bone away from a dog you don’t know. Try to take someone’s wallet or a child’s bike. Possession of valued objects creates the behavioral equivalent of potential energy. Just like dropping a rock from a great height, attempting to remove positive reinforcement creates destructive energy. The only real question is how to keep the rock from falling and how to stop the possessive violence. Once again, it’s not a matter of polarities. It’s more complicated than that.
In the world of behavior, humans have designated two polar opposite effects as the means of examining nature. Reinforcement and punishment are defined as opposites – one increases or strengthens, one weakens or stops behavior. These simple terms have been loaded down with so much baggage that they become caricatures in a fantasy – much like the Wicked Witch and Snow White. To use an arbitrary dichotomy to examine behavior is childish. The sweetness of the apple can disguise death. The poison in moderation can be healing medicine. (Warfarin, AKA Coumadin, created as a rat poison, is widely used as a blood thinner to prevent strokes) When you can say reinforcement and punishment with the same lack of passion as you say left and right you will have achieved understanding. Until then, you might as well check out #sevendwarfs on Twitter.
Lots and lots to think about with this one, Gary.
Excellent post. I particularly liked ‘the sweetness of the apple can disguise death. Poison in moderation can be healing medicine.’ You made valid points throughout. A balanced training approach makes the most sense.
Becky J. You nailed the main point. Bias in favor or reinforcement or punishment insures’failing to provide effective training.’
Hello Gary,
I have just discovered this blog, to my relief! I have been experiencing an existential crisis of sorts with regards to training methodologies and am trying to regain my equilibrium in the face of direct dog training experience versus academic and social pressure. I read “the science,” and more importantly – I can understand it. As such, I recognize the limits in samples, methods, and statistical analysis (although this last one is sometimes a challenge!) with regards to the current literature in animal behavioral science. I have had several articles rejected by the leading journals because I questioned the usefulness of the research in real-world dog training. To be clear, I have no qualms about utilizing whichever operant conditioning quadrant is needed to produce effective behavioral changes when the option to continue the behavior (such as bolting out the front door, biting the neighbor, chasing squirrels or cars or bicycles) could result in injury or death to the dog or people. I state this to my students and colleagues even with the possible ramifications of being ostracized or even expelled from therapy animal organizations that REQUIRE the use of R+ only training methods for their member teams. My ethical dilemma however comes when we are considering taught behaviors that do not directly benefit the dog, only the human, e.g. therapy and service work. Do we have the right to require a sentient animal to perform work for us using compulsive techniques (the dog must perform the behavior or receive a correction) when those behaviors are for the sole benefit of the human? Many service dog training agencies continue to use escape/avoidance training techniques, with the legitimate concern for reliability and efficacy across handler experience levels and environmental settings. But does that make it ethical? These are not animals who are facing euthanasia if they can’t demonstrate manners; these are often purpose-bred dogs intensely selected for suitable characteristics that make them amenable to a life spent in constant service to a human (who has generally received about 2 weeks worth of handler training before being sent out into the world with their service animal). I have personally witnessed countless examples of these SD recipients utilizing prong collars and leash corrections poorly, simply because they lack the skills to use them properly. In fairness, I have also seen SD recipients out in the world that did not appear to be mishandling their dog. What I almost NEVER see however is an SD that shows behavioral indications that they are enjoying their work (in contrast to hunting, herding, detection, or bitework). It warms my applied animal behaviorist’s heart to consider a “balanced” approach to training SD’s – use reward to teach behaviors and compulsion to proof them – but the overarching ethical concern then still remains. Due to your extensive experience in training methodologies, I am hoping you might one day address this dilemma! If not, I still thank you warmly for your posts and please know that there ARE those of us out there who recognize the nearly-insurmountable bias that exists in academia today with regards to dog training.
Cheers, Jen
My credentials are as follows: BA in Philosophy (concentration: ethics); MSW (concentration: animal-assisted therapy); MS in applied animal behavior and welfare; Doctoral candidate in human-animal interactions; 10 years kennel experience with German Shepherd Dogs (day to day operations, breeding, whelping, socializing, training [modified Koehler], rehabilitation, co-teaching obedience classes); 2 years mentorship with R+ (but not only) national-level competitor and agility trainer using clicker; 5 years utilizing verbal marker balanced system; 8 years as educator of animal-assisted clinicians (graduate level); 11 years animal-assisted therapy direct service provider; 2 years therapy dog evaluator. Details can be provided privately upon request.
Jen, dogs are a living artifact. They were created to serve mankind. Our ability to use technology to serve our needs has allowed many variations of dogs to exist who have no purpose other than to please humans by their existence. A few dogs retain their original purpose – to serve humans. Work is not always pleasant. Work is often arduous. The test is to examine what behavioral abnormalities come to working dogs and then decide of those abnormalities are harmful. Not speculative harm, but behaviors that can be objectively identified as aberrant behavior.
I have been around therapy and service dogs for more than 30 years. If you compare their lives to dogs that are just pets you do not find the same incidence of aberrant behaviors such as chronic barking, frantic escape, aggression, separation anxiety and serious physical problems like lick granuloma. On a scale of one-too-ten, the sum total of problem behaviors in the service dog community is about a 2 compared to the general population that is about a 5. Which is the greater problem? Is giving a dog a full-spectrum job worse than locking them in a box for life? Does a dog have to be joyous while working to justify teaching it to work? My wife is hearing impaired. I have trained three dogs in a row for her. They lived joyous lives overall. When they worked they worked. I am not troubled by that and neither were they.