Chapter 2 LEARNING and TRAINING FUNDAMENTALS
Total Page:16
File Type:pdf, Size:1020Kb
Chapter 2 L EARNING AND T RAINING F UNDAMEN T A L S My Crossover Journal I was presenting an evening program in England. In the audience was the late John Holmes, whose early dog training books profoundly influenced my training philosophy. Speaking about the constructive movement away from physically harsh training methods, I referred to an unintended and undesirable consequence of this shift: that many dog owners and trainers equate “positive” with “permissive.” I saw John nodding his head in agreement when I mentioned the growing reluctance to say or mean “no,” the apparent aversion to discipline, and the increasing tolerance of undesirable, even unacceptable behav- ior. In his no-nonsense, candid manner, John agreed, saying, “It used to be if a dog grumbled at being told to get off the bed, the owner would just take him off the bed. Now, if a dog grumbles at being asked to move, the owner sleeps on the couch and calls a behaviorist in the morning.” “These days, if a dog grumbles at being asked to move, the owner sleeps on the couch and calls a behaviorist in the morning.” 15 THE THINKING DOG 2 — LEARNING AND TRAINING FUNDAMENTALS Everyone who trains a dog wants to use a method that gets results as quickly, effi- ciently, and successfully as possible. By understanding how learning happens—be it dogs, people, horses, cows, cats—trainers are able to make well-reasoned, informed decisions in selecting a fair, effective approach. Three basic “methods” Despite the impression created by the hundreds of dog training books, all dog training fits into one or more of three basic methodologies: compulsion-praise training (C-P), lure-reward training (L-R), and clicker training (CT). (I use the term clicker training here, but one could call this behavior-marker training.) L-R training utilizes a food treat or other object to entice the dog into position. With C-P training the trainer places the dog in position followed by verbal praise or a food reward. In addition to physical placement, C-P training may also include using a collar correction for non- compliance. Clicker training uses a sound to mark a behavior which is then reinforced. You’ll find a more detailed description of each methodology in Chapter 4. Putting clicker training aside for the moment, every other training “method” uses the mechanics, exercises, and fundamentals of lure-reward and compulsion-praise methods, or a mixture and combination of the two. For example, a method might start with luring to lay the training foundation, then utilize a collar correction, for non-compliance. Or a method might use physical placement for some behaviors, and luring for others. No matter what training method you use, your dog’s learning follows “laws”—prin- ciples that influence learned behavior. Over the last decade there has been a sea change in the dog-training community resulting from a growing awareness of and attention to these laws. By understanding how dogs learn, a trainer can objectively consider different approaches that might improve upon, and may be “nicer” than previous methods. Such was my epiphany in each of my training incarnations: from Koehler, to “Motivational,” to clicker training (see Introduction). Regardless of your training history, it is helpful to understand the laws of learning, and how they apply both to your previous method(s) and to clicker training. Armed with this knowledge, crossover trainers are able to make intelligent, objective decisions about their dogs’ training. So let’s briefly examine the science of learning, the basis of all methods of training. (See Resources for references to a more in-depth explora- tion.) How dogs learn When your dog grumbles at you and you end up sleeping on the couch, as John Holmes referred to, is your dog learning something? Of course he is. He’s learning that you move away when he grumbles. Moving away reinforced grumbling: cause and effect; behavior and consequence. The axiom of learning theory is Edward Thorndike’s Law of Effect (published in 1911), which says behaviors just prior to a pleasant event are more likely to be repeated; behaviors just prior to unpleasant events are more likely to diminish. Put another way: • Behaviors that have pleasurable consequences tend to strengthen. • Behaviors resulting in undesirable consequences tend to weaken. 16 THE THINKING DOG 2 — LEARNING AND TRAINING FUNDAMENTALS More than just being about your dog’s present behavior, the consequences of this moment are about her future behavior. Any one training session, any one interaction, any one reward you provide or punishment you invoke is less about this moment than it is about the future—how that consequence will affect your dog’s behavior moving forward. Terminology Issues Psychologist B.F. Skinner’s work with incentives (motivation) and behaviors (out- comes) took Thorndike’s axiom to the next level, that the consequences of behavior are what learning is all about. While Skinner, Thorndike, and the other early behavioral psychologists classified how operant learning takes place, they did no one any favors in selecting terminology. They chose familiar words in common usage that have other, unrelated meanings outside the “ivory towers” of academia. Their choices have led to confusion and misunder- standing. They used the symbols + and – to mean add and remove. That’s fine, but rather than calling them “add” and “subtract”, they called them positive and negative. To most of the world positive and negative mean good and bad. Not so in “their” world. Thinking of them as “add” and “subtract” or, “provide” and “remove” makes the consequences easier to understand and apply. Next are the consequences themselves. Reinforcement offers no problem—after all, reinforcement strengthens, even in “our” world. The problem is with the word pun- ishment. In behavioral terms, punishment is not reprisal, payback, or vengeance, as it so frequently means to us. Rather, punishment is defined retroactively by its effect on behavior. If the behavior decreases, the consequence was punishment. With this terminology in mind, let’s investigate the four possible results. The four consequences There are four possible ways that a behavior can be either strengthened or weakened, called behavior effectors. Let’s look at each from the perspective of the trainer’s action and the dog’s perception. Reinforce = strengthen behavior There are two ways that reinforcement increases the likelihood of a behavior being repeated—positive and negative reinforcement: • Positive Reinforcement (R+). The trainer provides (adds) something desirable to the dog to strengthen a behavior. The dog perceives that his behavior results in good things. Examples are rewarding a behavior with a food treat, verbal praise, playing with the dog, or giving him permission to chase a squirrel. • Negative Reinforcement (R-). The trainer’s goal is to strengthen a behavior by removing (subtracting) something the dog perceives as undesirable. From the dog’s perspective bad things stop when he performs the behavior. For example, collar pressure is released when the dog sits, reinforcing sit. An electronic stimulation ends when the dog moves away from the invisible fence boundary, reinforcing boundary training. When the dog stops pulling, pressure on the nose loop of the head halter is relieved, reinforcing loose-leash walking. 17 THE THINKING DOG 2 — LEARNING AND TRAINING FUNDAMENTALS Punish = weaken behavior Punishment involves two possible outcomes that decrease the likelihood of the behav- ior being repeated: positive and negative punishment: • Positive Punishment (P+). The trainer provides (adds) something undesirable to diminish or weaken a behavior. From the dog’s perspective, bad things result. Some examples are a leash pop for pulling, a squirt of water for barking, or the zap of an invisible fence collar for straying over the boundary line. • Negative Punishment (P-). The trainer takes away (subtracts) something desir- able for the purpose of weakening or eliminating a behavior. From the dog’s perspective, good things stop. Examples are a time-out (removing the dog) or leaving the room (removing your attention) to punish attention-seeking behav- ior; punishing non-compliance to a cue by eating the treat yourself or giving it to another dog. By turning her back to him, Shari’s removal of her attention is negative punishment for Siku. My Crossover Journal “Delimishment” On a long car ride across England, Elizabeth Kershaw, a wonder- ful British dog trainer and behaviorist, and I were lamenting the problems caused by the word “punishment.” If only the early behavioral scien- tists had used a different term, there would not be such a hue and cry about employing punishing consequences. After all, punishment per se is not bad. Abuse and mistreatment are, but that’s not what punishment means. Defined by results, punishment doesn’t have to hurt to affect behavior. If the behavior decreases, the consequence was “punishment.” We thought if it were a different word, dog trainers would not be so reluctant to talk about it. We came up with the word “delimish”: a combination of decrease, diminish, limit and elimi- nate. Delimish is a silly word; it made us laugh—the antithesis of the effect of 18 THE THINKING DOG 2 — LEARNING AND TRAINING FUNDAMENTALS “punish.” Since delimishers don’t need to be painful to be effective, nor do they need to be profoundly memorable (eye contact can delimish behavior), we like having a nonsense word to replace the often misused and misunderstood word “punish.” It sometimes seems as if the problem isn’t so much the technique as it is the word. Perhaps trainers can discuss delimishing behavior with greater civility than when talking about “punishment.” If we could discuss delimishing con- sequences unemotionally, maybe it would put an end to delimishing trainers who admit to using delimishment.