Generalization is the process of training a dog to respond to commands with reliability in different places. Because dogs are extremely location oriented in their learning, they must be trained in a number of different places in order to develop reliable responses to cues and commands no matter where they are. The dog that has been taught the recall command “Here” in the backyard, for example, will not obey reliably in the town park with other dogs and distractions until it has been generalized.
To effectively generalize a dog requires a consistent training program with benchmarks for graduating to the next level of learning. It is unrealistic for the owner to expect a dog that does not obey commands in the backyard to be a model hunting companion when the dog’s adrenaline is roaring in the bird coverts. Implementing a logical step-by-step training program based on a building-block system is a much better blueprint for success than a helter-skelter approach.
Avoidance training is required to generalize a dog and develop one that responds with excellence in the field. Avoidance training teaches the dog that by offering a specific behavior in response to a cue or command, it will avoid a negative consequence. The timing of implementing avoidance training is critical; just taking a dog into the field and shouting “Here!” as the dog disappears over the horizon, and then punishing it with an electronic collar can have disastrous consequences. Such an approach is dog training’s equivalent to Russian roulette.
In previous articles I have stressed that the trainer’s objective should be to guarantee the outcome. A core rule of training is that a dog should not be corrected if it does not know what behavior caused the negative consequences. If a dog does not know why it was corrected, it has no solution—as the dog perceives it—to avoid the negative consequence in the future. A dog that does not have a solution to solve a problem will be apprehensive, and an apprehensive dog does not respond to new training with an open mind and is therefore much more difficult to advance to higher levels. Punishing a dog for an unknown reason may also result in an unintended association for the dog. The dog may associate the correction with the place or birds, for example, and become a bolter, a blinker or a dog that won’t hunt.
Using the principles of avoidance training to generalize the dog in the hunting arena begins in yard training. Another foundational rule for teaching obedience is that if the dog does not respond with excellence to a cue or command in the yard, do not give the command in the field. Giving an unlearned command in the field will put the trainer in the position of either letting the dog get away with non-compliance or administering a correction and hoping the dog forgives it and recovers. The first choice is called variable reinforcement, and it is a dog trainer’s albatross. Variable reinforcement results in inconsistent obedience and requires more and more correction. Variable reinforcement creates case-hardened dogs that the owner will still be punishing years later; if a dog has to be constantly corrected, then it is not trained.
Neither approach is correct, so do not give a command in the field if it has not been successfully taught in the yard. When the command is first issued outside of the yard, the process of generalization begins. We need benchmarks in place to guarantee the outcome when we teach the dog that it must comply with known cues and commands in the field, the town park . . . everywhere.
To begin the process of generalization, I first get the desired behavior. Let’s take the recall command, for example. When a dog is coming to me without a command, I reward the dog with a piece of hot dog or cheese. I may be on my knees and a pup comes to me. Whatever it takes to get the dog to come to me without having to say the recall command is the objective. At this stage, the dog does not know what “Here” means, so there is no point in giving the command. I do not use compulsion, such as pulling the dog to me with a check cord. If I use compulsion, the dog will resist and will not be enthusiastic about recalling in early teaching. Also, once the check cord is removed, the dog most likely will blow off the command, setting me up for either letting the dog get away with non- compliance or forcing me to use punishment before the timing is right.
Once the dog figures out that when it comes to me it receives a treat, the dog will begin offering the behavior on its own. Anytime the dog believes that it will be rewarded by offering a specific behavior, it works for the dog and does not force us into a fight. As discussed in “Clicker Talk” (March/April ’08), I begin the process with clicker training in order to identify for the dog the behavior that will result in a reward. A dog that is taught a command with positive reinforcement first, before compulsion or punishment, will be a more enthusiastic student and respond to all training better.
Once the dog is reliably offering the behavior of recalling to me, I say the command “Here” as the dog is coming. Dogs learn by association, and with me saying the command as the dog is recalling, the dog will pair its behavior with the command. When the dog recalls, I reward it. After a number of repetitions of saying, “Here, ” as the dog is coming, I give the command before the dog recalls. When it comes to me, I reward it. If the dog does not recall at this stage, I simply do not reward it. I need to mention that a hungry dog will be more responsive than one that has just been fed. I am not yet ready to correct the dog for not coming other than to penalize it by not rewarding it.
When the dog will respond to the command 80 percent of the time in the garage or basement, I have the benchmark to move on to avoidance training. I prefer to do my initial “yard work” inside in a closed area free of distraction. I want to guarantee the outcome and do not want to compete with tweety birds, other dogs or anything that will take the dog’s focus off of learning. When the dog is being reliable 80 percent of the time, I am now ready to correct it for disobeying the command. It is important to understand that at no time in this process have I said the command twice. The objective is not to get the dog to eventually come; the objective is to get the dog to respond with excellence to the cue or command the first time it receives it.
Now when I give the command and the dog blows me off, I say, “No,” and then correct the dog by throwing a rolled-up towel at it. Dogs hate missiles, and this correction will make an impression. Although I am going to incorporate the e-collar into the generalization process, this is not yet the time. The e-collar is not used to teach a command; it is used to teach excellence to a known command. The more a command can be taught with positive reinforcement, the less correction will be needed. Once the dog is responding with excellence to the command and understands why it is being corrected, it will recover and continue giving an effort in school. The dog will choose the desired response. It will be rewarded for complying and corrected for disobeying a known cue/command. This contrast between reward and correction is important. A dog will not achieve excellence solely by being rewarded, nor will it be reliable or stylish if trained with harsh corrections only. Both rewards and corrections are necessary to develop a dog that responds with excellence in an upbeat manner.
Once the dog is reliable and has been introduced to correction, I condition it to the e-collar condition. There is no better tool to generalize a dog in the field than the e-collar. I cannot hit the dog at 50 yards with a rolled-up towel. And I do not want to be guilty of variable reinforcement. I need to have a training tool that I can use if the dog does not respond to my command wherever it is. The e-collar allows me to accomplish this. I do my e-collar conditioning in the area where I have been doing my yard training and do not first use it in the bird field. A this point the dog already has learned that there is a penalty for non-compliance. I give the recall command, and when the dog does not respond, I mark the behavior of disobedience by stimulating the dog with the lowest level it feels. This level is nothing more than a tingle and simply notifies the dog as it is exhibiting the undesired response that this behavior will get it in trouble. I then follow with the correction by stimulating the dog with a higher level that is more impressionable.
The dog already learned identification of behaviors and their consequences in clicker training. In clicker training the desired behavior was identified that would result in a reward. When I first introduced the dog to avoidance training, I notified the dog of an undesired behavior with “No” and then punished it by throwing the towel. The dog will quickly understand that the low-level stimulation is an identifier of the bad behavior and that a meaningful correction will follow. By identifying the unwanted behavior before administering the consequences, I will give the dog time to process the information and understand why it was corrected. The dog will handle the correction better if it is notified and then corrected rather than just corrected.
Once the dog has been introduced properly to avoidance training and the e-collar, it graduates to the outdoors. I will go to the backyard and give the command. If the dog complies, I reward it. If the dog blows me off, I will notify and correct with the e-collar. I will generalize the dog in a number of different areas before going to the bird field. If the dog will not respond without correction at the high school football field, it certainly is not going to comply in the bird field. The last place I generalize the dog is in the bird field— and only after it succeeds in a number of different places without me having to punish it each time for not complying to the command.
The command “Whoa” for pointing breeds and “Sit” or “Hup” (depending on the breed and the owner’s preference) for the flushers and retrievers is taught in the same process as described above. Get the behavior, reward the behavior, introduce the command when the dog is reliably offering the behavior, and reward for success. Once the dog is 80- percent reliable, introduce avoidance training by notifying the dog of bad behavior and then correcting. Generalize the dog by taking it to a number of different places before cementing the command in the bird fields.
A dog is a predator, and it is in the field with its prey. It is not realistic to expect excellence in the field without corrections. Properly introduce the dog to the e-collar so it learns that there are no options to obeying. And never say the command twice.
For more training articles or information on the George Hickox School of Dog Training, visit www.georgehickox.com.