Open Call for Creativity
By Angelica Steinker, M.Ed. C.C.B.C.
The purpose of this article is to encourage trainers to be open to experimenting with various motivational and playful training techniques. The purpose is not to censor the use of punishment, but to evoke thought about its use and to encourage trainers to find alternative methods of teaching their dogs. For purposes of this article, punishment is referring to both positive punishment and negative punishment, and reinforcement is referring to positive reinforcement. Positive punishment is when you add something to the dog's environment that decreases a certain behavior. Negative punishment is when you take away something that the dog wants in order to decrease a certain behavior.
Punishment is Familiar
Some parts of this article may be difficult to read. The subject of punishment is a sensitive topic. It is sensitive because we have all experienced being punished and we have all punished. Because we know that punishment is unpleasant, it is uncomfortable to think about exactly how we use punishment.
All of us know what it is like to receive punishment. Most of us have gotten a speeding or a parking ticket. All of us were punished in some way while we were growing up and our parents taught us proper behavior. Because of these experiences we know that punishing is no fun for the person on the receiving end. Punishment is, however, familiar. As people and trainers we tend to gravitate toward what is familiar to us. It is much easier to use something familiar than to come up with new ideas. The result is that we tend to use what we are familiar with. This may cause us to miss out on other effective training techniques, because we have locked in on what we already know how to do.
Knowing what it feels like to be punished can cause some shame about using punishment in our training or with other people. It is important to be aware of this, and to understand that this shame may make parts of this article difficult to read.
Punishment Begets Punishment
Punishment is popular. In our society we are surrounded by it. Managers use punishment to control employees. Teachers use it to discourage misbehavior in the classroom. The police use it to discourage speeding, and more serious violations of law. Society uses it in an attempt to control criminals. The examples are countless. Even in the most intimate of relationships punishment is rampant.
The thirst for punishment seems to be driven by people's desire to control. The irony is that people truly control very little. The very fact that all of us are almost always at a loss of control is the driving force behind the need for control. The best dog trainers seek perfection. It is this desire for perfection that can cause a desire to totally control their dogs. On the surface this concept of "total control of the dog" sounds pretty good. Trainer gets total obedience and dog gets rewards. However, invariably the dog will make errors. These errors clash with the concept of total control and can pave a road to the use of punishment.
Anyone who as ever owned a dog can tell you many stories about how total control is not possible. There are OTCH dogs--dogs trained to the highest level of obedience--that refused to sit in a certain circumstance. There are dogs that refuse to recall in the presence of squirrels, and many other examples. The desire for and the glitzy illusion of total control turn punishment into a magnet for trainers. Even some world-class trainers state that using shock collars for recall training or dog aggression can be appropriate. Neither problem will be resolved by the use of a shock collar, but the technique is recommended anyway. A dog's recall is the reflection of the relationship between dog and owner, and how much recall training has been done under severe distraction. Treating dog aggression requires positive associations with other dogs, not shock associations with other dogs.
While each individual is responsible for how they choose to train their dog, society has set us up to choose punishment. As young children the conditioning begins in school when errors are marked in red. Adolescents watching movies in a crowded theater will scream "fake" if the special effects fall short of their expectations. We are conditioned to have laser-error-vision. The fact that we are conditioned to hone in on errors sets the stage for punishment. Our laser-error-vision is programmed to see errors, not exceptional behavior. When exceptional behavior goes unnoticed or is taken for granted the opportunity to reinforce it has passed. Let me say that again: When exceptional behavior goes unnoticed or is taken for granted the opportunity to reinforce it has passed. This is a very important point.
Another dynamic intertwined with punishment is blame. When disasters strike or accidents happen, we have been conditioned to locate and punish the individual or entity that is to blame. Blaming is very self-reinforcing. Blaming is fun, because it means that the error is somebody else's fault, so someone else will suffer for the error. Blaming is the underlying force behind punishment. First you blame, then you punish.
In dog training our entire system of competing is based on perfection. In obedience every dog enters the ring with a perfect 200. Errors subtract points and reduce the score. In agility "clean runs" are the goal, and as the sport becomes more competitive this has become "clean runs with tight turns." The perfection competitors seek is elusive, and again, trainers are set up to see errors, assign blame, and punish.
Traditionally a dog is taught attention--to watch the handler--by initially rewarding eye contact and then punishing a dog for looking away as distractions are added. The resulting passive attention and the submissive body posture of dogs trained by this method is not ideal, but no points are taken off for this. The trainer will do well in the show ring despite the unhappiness of the dog. Although unhappiness of the dog can not be proven, submissive body posture and a limp tail wag speak volumes to the dog's state of mind. Despite the dog's poor state of mind he can still win in the ring, so the use of punishment has been reinforcing to the handler.
But I was Punished
While we were growing up we all experienced punishment. This leads to pro-punishment thoughts such as, "but I was punished and I turned out okay." People do not turn out okay because of punishment. This is a gross over-simplification of a highly complex process. People turn out okay because they are taught values that are in alignment with most of society's values. People turn out okay because they choose to behave in a decent way. Punished children rarely learn what their parents intended. A child punished by a parent for using a bad word will simply learn not to say bad words in the presence of parents. The true intent of the parent was to prevent all uses of bad words and punishment does not accomplish this. This same punishment dynamic will cause a dog that is punished for soiling the house to misunderstand and avoid soiling in front of the owner. This misunderstanding leads to obvious chaos in terms of housetraining. Since the dog is now still not housetrained this can lead to stronger punishment as the owner mistakenly assumes that the first corrections were too weak. A vicious punishment cycle is born and the dog is still not housetrained.
Punishing is Reinforcing
When a trainer uses punishment it is reinforcing to the trainer in several ways. First, punishment elicits strong responses from dogs so the effect seems powerful to the punisher. The trainer dishing out the punishment feels powerful as a result of punishing the dog and the strong reaction of the dog. In addition, anyone watching is likely to be impressed by the dramatic reactions of the dog. Second, the trainer has made the decision to punish; this is empowering and also reinforcing to the trainer. Being able to make decisions about what type of behavior warrants punishment is an ego trip. Third, if the trainer is feeling frustrated the punishment will have a cathartic effect. The mere act of dishing out punishment will help trainers purge their feelings of frustration.
Punishment is frequently referred to as "the only thing that works." The reason punishment is "the only thing that works" is because the trainer has not chosen to put forth the extra effort required to find the proper reinforcement that would be effective for that particular dog. As mentioned previously, punishment is familiar to all of us. Trainers tend to use techniques they are familiar with rather than come up with new ones. In other words, punishment-training techniques have been practiced more than reinforcement techniques, so trainers are more likely to have better punishment skills than reinforcement skills.
Punishment is reinforcing to the trainer because:
1. It elicits powerful reactions from dogs. This includes a possible secondary gain of impressing spectators.
2. It is an ego gratification (feeling of power).
3. There is a relief from frustration.
4. It does not require any creative thought or problem solving. It is easier for the trainer than using reinforcement.
Even if punishment is not reinforcing to the punisher, it can be reinforcing to the dog. Many dogs will prefer to accept punishment than be ignored. For these dogs punishment will actually strengthen a behavior and increase the likelihood that it will reoccur.
All this ensures that trainers using punishment will never be sure whether they are using it because it truly was the last resort, or because they have a history of being reinforced for using punishment.
Arguments for Punishment
Jerk and praise trainers--trainers who use choke collars and leash corrections--insist that correction-based training is both faster and more effective. However, both punishment and reinforcement require appropriate timing to be effective. For every training scenario that would cause a jerk and praise trainer to opt for punishment, one could also choose to use techniques using reinforcement. Because using punishment is so popular trainers have become good at it. Experienced trainers know what type of punishment will invoke the best results. Consequently a lot of trainers lack the knowledge of how to creatively use reinforcement. It is easier to continue to use the known punishment techniques than it is to dream up reinforcement feasts. This is one reason why some trainers find correction-based training to be quicker. For them it is quicker because they choose not to take the time to learn the reinforcement techniques that would lead to the same results.
Both corrections and reinforcement require appropriate timing. A poorly timed leash correction will be ineffective. For example, during a heeling exercise a dog moves out in front of the handler--this is the incorrect position for the dog. The handler sees the error but accidentally jerks the dog's choke chain when the dog has moved back into to the proper heel position. The dog is now confused, since prior leash corrections told him being in heel position was good, but this correction told him it was bad. A sensitive or independent dog is likely to shut down and stop working after several improperly timed leash corrections. The door is now open to blaming the dog. The confused dog is blamed for having a poor attitude.
Poorly timed reinforcement will also be confusing or meaningless to the dog. If a trainer is using food as reinforcement for the proper heel position, and then mistakenly gives the cookie when the dog is out of position, this will also confuse the dog. The difference is that the incorrectly rewarded dog is less likely to quit working. It is logical to conclude that a mistimed cookie or belly rub is not as unfair as a mistimed leash correction.
Punishment Begets Counter-Control
Nothing is free. The use of punishment comes with a price. Extreme punishment causes seemingly insane behavior. Several years ago there was a shocking story of a circus elephant that "had gone mad" and attacked its handler and trampled spectators before it was shot to death on the street. Later, in a follow-up story, it was announced that the elephant's trainer had used cruel training techniques. Many examples of excessive punishment causing horrid behavior can be witnessed at any local animal shelter.
Punished animals will tolerate the punishment to a point. This point is the punishment threshold. When the threshold is crossed the animal swings into counter-control. Counter-control is usually aggression. People then control this counter-control with the ultimate punishment of killing the animal. When it comes to control, nothing works better than killing.
Excessive use of reinforcement leads down a road with many forks. The extreme use of reinforcement can lead to many behaviors, depending on the knowledge and timing of the trainer. Some of the resulting behaviors can be unwanted and some wanted. Regardless of where the road leads, there are no known cases of dogs being euthanized as the result of too much praise, or too many poorly timed cookies. Both punishment and reinforcement, at their extremes, do not reflect normal training practices. However, it is important to examine both punishment and reinforcement at the extreme, since some people will make the decision to use extreme measures.
The figure below illustrates the full spectrum of training possibilities. At the far left we have trainers that use mostly punishment and very little reinforcement. At the far right we have trainers using mostly reinforcement and very little punishment.
Bond & Strength of Relationship with Dog
Mostly Punishment --- Even use of both --- Mostly reinforcement
Which end of the spectrum would you prefer others use to teach you new or better behaviors?
As illustrated by the graph above, the greater the use of punishment the more of a negative impact on your bond and your relationship between you and your dog. This is true because no matter how well timed the punishment may be, the dog will still dislike the punishment. Classical conditioning, one of the processes by which dogs learn, dictates that the punishment will be associated with the punisher, hence there is a negative impact on the bond between dog and trainer.
Regardless of what training techniques you choose to use, using both punishment and reinforcement is unavoidable. Even if you are fully intend to train your dog with no reinforcement, many interactions will be reinforcing. Eye contact is reinforcing to dogs, and so is physical touch. Likewise, a kind tone of voice is also reinforcing. The same facts apply to purists who want to train using absolutely no punishment. It is simply not possible. One verbal correction and you have missed your goal. Asking your dog to repeat an exercise can also be punishment. The withholding of a reward is punishment. The point is not to censor punishment, but to meet the challenge of finding and using alternatives.
Both punishment and reinforcement are subject to desensitization. This means that when a trainer resorts to the use of a shock collar, if the collar is used to administer a lot of shocks, the shocks will need to become stronger in order to be effective. Dogs that are shocked will develop a tolerance for being shocked. This is why dog owners who use hand-held remote shock collars to contain their dogs in their yards find themselves increasing the intensity of the shock until even the highest level of shock does not cause the dog to exhibit the desired behavior of staying in the yard. The owner has successfully desensitized the dog to the shock. This desensitization can then lead to abuse. The same process of desensitization can occur with other forms of punishment. A swat becomes a slap, a slap becomes a smack, a smack becomes a kick, and so on. This is not about training--this is about humanity and how we treat each other and our dogs.
If the dog does not get desensitized to the shocks there is a chance that the dog will sensitize to the shock. Studies show that some dogs will desensitize while other dogs sensitize when exposed to the same stimuli (see Excel-erated Learning by Pam Reid for more information). If a dog becomes sensitized to the shocks during the training process the dog is likely to kick into fight-or-flight response because she is pushed over her punishment threshold. This can be dangerous to both the trainer and the dog.
Finally, the use of shock collars will not have a positive effect on your relationship with your dog. Even if the dog only requires one or two shocks in order to stop a behavior, the question remains: how would you want to be trained? Imagine a person working in sales being told that he must sell 1,000 widgets in order to avoid a shock, in contrast to being told that he must sell 1,000 widgets in order to receive a bonus.
Reinforcement is subject to the same dynamic of desensitization. If a trainer always gives the dog the same treat in the same way, the treat will become less effective. This is one reason that if a trainer chooses to train with food the dog should be hungry during training times and the food should be varied. A dog that has free access to food, or toys, will simply not be very motivated. Limiting access to reinforcement used during training will keep that reinforcement desirable and prevent desensitization.
Punishment Means the Dog is at Fault
Logic dictates that fair punishment is the result of mistakes on the
dog's part. Unfair punishment is not logical and therefore does
not warrant being addressed. So, if the handler is teaching the
dog and the dog is learning and a mistake occurs, the dog is punished
because the assumption is that he is being disobedient. However,
the dog's mistake could be caused by many other factors; some of these
Learning has not generalized. What you have taught him is not understood in various contexts.
Cue was not sufficiently proofed -- something else is more interesting to the dog.
Distraction was too overwhelming to the dog. Dog was set up for failure.
Dog does not feel well.
Dog is confused.
Dog has misunderstood.
Training is poor.
Dog was not paying attention. Trainer needs to work on attention.
How is it that we assume that the mistake is disobedience and a deliberate and calculated act? Why assume the worst? Why blame the dog? If the dog is at fault then the trainer is seriously limited in terms of being able to prevent the error from occurring again. This means the trainer lives in fear of the moment when the dog decides to deliberately disobey again. This means that the dog is in charge. The handler can correct only when the disobedience occurs. The result is lots and lots of training time being spent vigilantly watching for the deliberate disobedience to occur so that it can be corrected.
The alternative is so much more pleasant for both trainer and dog. It is to assume that the trainer has made the mistake? That the trainer has caused the dog to err by lack of training, or improper training. If the trainer is at fault, the trainer can easily make the changes required. The trainer changes his behavior rather than blaming the dog. The trainer is in control. This process also assumes that the dog is intelligent enough to learn and intelligent enough to misunderstand or become confused. It is clear that the dog's errors are the trainer's responsibility.
However, the blame game is infinitely more rewarding to play than to actually make changes in training programs, so the jerking continues.
Is Punishment Stronger than Reinforcement?
Punishment elicits dramatic results. A dog that is shocked is highly motivated by the pain she has received to avoid further shocking. A dog that is beaten for messing in the house will go to extreme measures, such as eating her stool, to avoid future beatings. All this evidence suggests that punishment is simply stronger than reinforcement. However, reinforcement wins the strength contest for many reasons. Consider high-energy activities such as running. When do you see a dog running full tilt and going for broke? If punishment is the ultimate motivator why do greyhound tracks bother with the lure of a fake rabbit? True motivation is achieved through reinforcement, not punishment. As stated earlier, how long would people tolerate being forced to work in order to avoid punishment, rather than getting the reinforcement of paychecks?
Regardless of whether you are using punishment or reinforcement, classical conditioning dictates that associations will be formed. If you are punishing your dog, your dog will begin to associate those punishments with you. Likewise, reinforcement will also be associated with you and will increase your bond with the dog. If you choose to train your dog using punishment (i.e. leash corrections), you will always be left wondering whether the dog is really working for you or working to avoid corrections.
The same could be said of play training and food training. However, trainers have the option of using variable reinforcement--reinforcing the dog randomly so he does not come to expect reinforcement. Variable reinforcement only works with reinforcers. Trainers can not pick and choose which mistake will warrant a leash correction and which will not. In order for the dog to learn, the correction must be given each time the mistake occurs. Creativity and out-thinking your dog are more work than a quick pop. If more trainers chose to dream up doggie-reinforcement-fantasies all dogs and trainers would benefit. Dogs are amoral; we are not. Who has the moral obligation to try to do better?
If your dog could speak what she would say to you about how you train her? Putting yourself in the position of he dog can help trainers gain insight regarding their style and the type of training techniques used. If this article has inspired you to try different training techniques and to launch an onslaught of creative training, you may want to prepare for some training speed bumps. New training ideas are always subject to tweaking. Both your personality and the personality of your dog may require that you make adjustments to your new training ideas. By anticipating an adjustment period you will have the perseverance to keep trying. If at first you do not succeed, try, try, try again!
Coersion and its Fallout by Murray Sidman.
Don't Shoot the Dog by Karen Pryor.
So Your Dog is Not Lassie by Betty Fisher and Suzanne Delzio.
Behavior Problems in Dogs by William Campbell.
Train Your Dog the Lazy Way by Andrea Arden.
Culture Clash by Jean Donaldson.
Dogs are from Neptune by Jean Donaldson.
The Man Who Listens to Horses by Monty Roberts
Purely Positive by Sheila Booth.
Excel-erated Learning: How dogs learn and how best to teach them by Pamela J. Reid.
"Of Hostages and Relationships" by Suzanne Clothier, www.flyingdogpress.com.
"The shocking truth about shock collars" Animal Behavior site www.apbc.org.uk/article2.htm.
A version of this article previously appeared in Clean Run magazine, March-April 2001.
Angelica Steinker, M.Ed. is a motivational trainer and owner of Courteous Canine, Inc., a dog training school that specializes in dog behavior, problem solving, clicker training, and agility. Angelica has a Master's degree in counseling from George Mason University and actively uses behavior and learning theories in her work with dogs and their owners.