Chapter 3. The Reward System

In the Reward System, the dog develops a strategy for turning on pleasant events. The system is quite versatile and works exceedingly well with dogs of any age, size, and temperament. It does wonders for dogs that show fear, shyness, aggression, or any other emotional or temperament problem that the owner may feel is disabling; for dogs that need to be motivated to perform; and for those that require confidence building.

Reinforcement in the Reward System

Reinforcement happens when you "tell" the dog that he made a good response, at the moment he made it. Thus, behaviors are strengthened (reinforced) with secondary reinforcers—not primary ones. The voiced sounds, GOOD and OUT, become secondary reinforcers1 when they are paired with food or other primary reinforcers over a series of trials. (See the section, "Reinforcing Sounds" under the heading "The Basic Sounds," below).

Most reinforcing is done with sound: It can be a voiced sound, a clicker2 sound, a finger snap… However, trainers in unusual circumstances use non-sound events as reinforcers. One time, I worked with a student who trained his deaf dog, and, on several other occasions, I helped deaf persons train their dogs—the students use "sign," unique movements of the arm and hands to reinforce behaviors in the training trials.

Reward-System Models

Two basic variations of the reward system are available to the trainer. One is the Assisted-response Model, which you will use mostly in this course. The other variation is the Free-response Model. The latter model is often referred to as the Clicker Method, whenever a handheld clicker is used to tell the dog that he made a correct response, at the precise moment that he made it.

The Free-Response Model

Essentially, you allow your dog to respond freely in the workplace—to do anything he wants. You do not tell him what to do or show him how to respond. You just wait for the behavior that you've designated to be correct to happen. Then, at that moment, you'll tell your dog that a reward will follow.

This model is used sparingly in this course. One instance of the model's use is to get early control of the Down response. It's a fascinating and fun way to train your dog. It's used in this text to show you how a behavior can be brought under control in a wholly different way, still without coercion (See Down Training in Chapter 13).

The procedure is meant to show you, in the most direct way, the relationship between behavior and its consequences. Example: One time, Audrey, a class friend attempted to condition the behavior of a professor while he lectured. The behavior that she tried to get him to do was "gaze in her direction with greater frequency." Whenever the professor looked in her direction, which wasn't too often at first, she'd look up at him and slightly nod her head. It wasn't long before the professor looked in her direction with ever-greater frequency. Then it was time to extinguish the behavior. To do that, Audrey looked down at her notes instead. The professor stopped looking in her direction. She again reestablished the conditioned behavior by "the look and nod." I asked Audry to stop before the professor realized what she was doing.

The Assisted-Response Model

This model is the method of choice for training most performance skills in this course. It's the more traditional way of training a dog. In this model, the dog is "told" to make a response—commands and signals do the telling. Then the dog is "assisted" in making the response, without waiting for the dog to do it on his own. The trainer says GOOD or OUT precisely at the moment the response is completed. The delivery of a reward then follows in a deliberate, unhurried way. You will begin Assisted-response training immediately after the dog learns the meanings of the GOOD and OUT sounds.

This method isn't difficult to perform. The procedures are described in clear and ample detail, and the photographs (may not yet be included in your packet) show you precisely how to do them. Nevertheless, the beginner-trainer often fails to give proper attention to these details. For this reason, beginners should do their training at a participating Dog Training Club, where expert observers can critique their performance and make needed corrections.

Most dog owners want their dog to be obedient and friendly with people and other dogs. To accomplish this, two distinctive training strategies are employed: Game Playing initially teaches the dog the obedience skills; then Social Obedience in these skills is practiced for compliance in the social/real world setting.

"Game playing" is a gentle way of training your dog. In Game-playing, coercion is totally absent. Your dog is allowed to refuse a request to perform or to quit "playing the game" whenever he wishes, without penalty. Game playing always begins with "the call to play a game", continues with the game itself, and ends with a "grand" reward (see "Essentials of Game-playing," below).

The work-seasoned dog becomes very cooperative when he hears the call to play, and is eager to perform any task that you may have in mind. He should then perform with little or no hesitation during the training session. He will persist in the work even under demanding circumstances until the grand reward turns up.

When the dog is first exposed to the structured game, the effort that he must expend to get rewards is quite small—he is only to pay attention to what is going on (see Chapter 5, Getting Started). By simply observing, the dog learns to notice all of the session actions and how they relate to one another.

The dog's Game-playing strategy is to turn on the sounds GOOD and OUT—sounds that are followed by tangible rewards. In Social Obedience training, the dog's strategy is to avoid turning on the sounds NO and TIME, which are predictors of unpleasant consequences. NO is used mostly as a warning sound. TIME tells the dog that he earned a penalty; though compulsive, the earned penalty for social disobedience is done with a light hand. A discussion of compliance work is found in some skill-training sections of this manual.

Prompting

New signals, commands, and other stimulus actions in the training trial are learned by association with actions that follow them. Learning happens when a "new" action is "paired" often with an action that the dog knows something about, and which is already controlling a behavior (response). The "assisting" action, the second-appearing action in the "paired" sequence, is called a "prompt." It's called that because it "helps" the dog to make a correct response every time. By pairing a new signal or command with a prompt for a number of trials, the new signal or command begins to control the same response.

The training trial provides the dog with the required experience that results in a speedy acquisition and correctness of any performance. The trial typically starts with a command or signal that the dog is meant to learn. It's followed quickly by assisting the dog to make the response. In these training trials, the dog is helped with a prompt (assistance) before he has a chance to make the response on his own.

The Assisted-response Method doesn't tell you what kind of prompt (assist) you are to give your dog to help him make the desired response. It only says that the prompt you choose must get the response every time. For example, suppose you want to train your dog to sit whenever you say, SIT. To do that, you could follow the voiced SIT by pressing down gently on the dog's hindquarters, provided that your dog sits whenever you perform the pressing action.

Said again: A prompt can be any trainer-action that already controls the dog's response. That is, it gets the response to happen every time. Pairing such a prompt with a new event helps to get the new event to control the same response.

Testing versus training. When the trainer waits for the dog to make the response on his own before she considers delivering a prompt, it's called testing. Testing is not training: the wait period between the first event and the second of the paired events in testing is too long for the dog to make the connection between them. When the trainer attempts to apply the prompt before the dog has a chance to perform it on his own, it's called training. However, you may do an occasional test trial to inquire whether your dog is learning.

Some experienced trainers do little or no testing during their training sessions. They know that learning has taken place when the dog makes the response, or begins making it, before the trainer has a chance to deliver the "assisting" prompt. Alternatively, the experienced trainer may gradually fade out the amount of help given the dog. For example, following the command to halt in "Halt and Stand-stay" training, the response-assistance at first is a non-jerking full restraint imposed on the dog's lead to halt. In no time at all, the aid (prompt) that's given following the halt command is simply a light tug on the lead.

In passing, prompts are not allowed in Obedience competition. Prompts are considered to be extra commands, and thus incur a scoring penalty at these trials.

Essentials of Game-playing

The Game-playing Format is the blueprint for most of the dog's training and maintenance work. The principle features of the Game-playing Format:

The Call to Play a Game. When you are about to begin a training session, ask the dog if he wants to play a "game." The voiced READY does the asking—said in a normal, questioning tone. Be sure to "ask" before you make any moves to start the game, otherwise the dog may mistakenly begin to use your body movement as a cue that a game is about to be played. You don't want the dog to jump about with anticipation whenever you get up to do something else.

The Game-playing Trials. A trial is defined as a unit of training actions. It often begins with a quiet/still wait period action. The wait can vary from between one and five seconds from one trial to another. Following the wait, your dog may get a command to do something—after which, you would help your dog make the response. At the moment the dog responds, you'd say GOOD (or OUT). The trial ends when you reward the dog. You would then begin another trial with a varying "wait period."

The Training Session. The training session, beginning in Chapter 5, has all of the components of the Game-playing Format: the call to play the game, the training trials, and the grand reward that the dog gets at the last trial of the session. In the early sessions of Game-playing, your dog is not asked to do anything except be aware of what is going on (see Chapter 5). Asking the dog to do something begins in Chapter 6.

The Basic Sounds

The Call to Work Sound. The READY sound signals the dog to work3—it's supposed to get the dog to come to you and begin attending to the game that you are about to play. The dog learns the meaning of the READY sound in several working days. You are likely to see a marked reaction to the sound, when learned.

Reinforcing Sounds. The voiced GOOD and OUT are reinforcers. These sounds have the power to reinforce (strengthen) behavior, because they are associated with rewards. You will be using the GOOD or OUT sounds whenever your dog responds in ways that you want. In one or more days, your dog should "expect" a tasty tidbit to follow every GOOD sound, and "anticipate" a more attractive reward to follow the OUT sound.

Timing and Sequencing of Actions

The sound, READY. If the dog is to use the READY sound to predict when a Game playing session is about to begin, the sound must be the first action to signal a training session. For example, when you are ready to work with your dog, say the sound while you are absorbed in another activity, say, reading a book, watching television, exercising,… The sequence might be: say the READY sound; lay down your book; get up out of your chair; go to the kitchen; get your training foods ready for the coming trials; take all of your training aids with you to the place where you will do the work; begin the session.

Do not start any movement to conduct a training session before you say the sound. Otherwise the dog may learn to use your movement to predict the beginning of a training session, when often, no session is intended.

The command for the dog to do something. The voiced command or signal usually follows a brief wait period (see Part 2 of Chapter 7: "Sit and Rise"). The command must come on before you start any movement to assist the dog to make the response. If you start your assist movements before you voice the command, the command will not likely be learned—the dog will be reacting to your movements, or the dog will fail to respond, as if it doesn't know what to do, and the dog will unfairly get the blame for that.

The reinforcing sound GOOD (or OUT). The reinforcing sound must be timed precisely at the moment the response is made. What constitutes a "good" response by the dog must be clear in your mind before you begin an exercise. For instance, in the early steps of Sit Training, the response is defined as a Sit only when the dog's back-end comes in contact with the ground. That's when you will say the reinforcing sound. Later in Sit Training, you will reinforce the Sit-stay—when the dog had been sitting for a short period of time.

The GOOD sound is used in every trial except the last. On the last trial, the OUT sound replaces the GOOD sound, and a grand reward follows. The dog doesn't know that the last trial is being run until he hears the OUT sound.

Rewarding the Dog. Do not begin any food delivery movements before you say the reinforcing sound. Otherwise the dog will latch onto your movement to predict an impending reward, rather than use the reinforcing sound to do that. If the dog wrongly becomes conditioned to your movement to predict a reward, he will continually look at you for the movement "signal." You will thus be restricting the dog in the kinds of things that he can learn.

top ::: home