A cue is not a command. Commands are not commands either. Words do not produce behavior. A cue is a signal to the dog that he will get paid for performing a certain behavior. The cue does not produce the behavior. The behavior is produced by the previous week of sessions with many payments for many repetitions of the behavior. To put the behavior on cue you gradually start preceding the behavior with the cue. Then you begin paying only the repetitions that are preceded by the cue. You stop paying the behavior when it is not preceded by the cue.
In the retriever training world a cue is a trigger to elicit a trained behavior from a dog. Many refer to a cue as a command. We make a very large assumption when we assume that the dog is picking the same cue that we are using for a particular behavior. When the trainer says “sit” and the dog sits, is the dog reacting to the sound “sit” or is he reacting to a visual signal, a consistent movement, probably subconscious, that the trainer makes whenever the trainer says, “sit?”
Consider the fact that dogs communicate almost entirely with the visual signals of a person’s body language. In training dogs preferentially pick a visual cue from the trainerto respond to. Frequently this cue is not the words or sounds emanating from the trainer’s mouth. The dog is really responding to some body movement that the trainer has habitually and subconsciously accompanied with the verbal signal.
Also consider the dog’s 14,000 years of evolution through domestication. Recent research suggests that early humans selected dogs for domestication by picking those animals more attuned to human communicative signals. Research demonstrates that dogs significantly read human behavioral characteristics as subtle as eye movements in the communication process. These factors indicate that dogs prefer a visual to an auditory cue. When you say “sit”, and the dog sits, he is more likely to be responding to your eye movements, or your infinitesimal head movement, or your slight posture change, than he is to your sound.
Trainer Kathy Sdao gave a great example of this tendency during a recent Clicker Expo. It seems a prominent agility trainer and handler had a very talented dog that looked like a sure winner for the national competition. The dog had won several regional contests. At the national started well and then froze at the first turn. He was unable to take a simple directional cue. He was out.
After the competition an analysis of his failure determined that the freezing was caused by the absence of the trainer/handler’s ponytail. The handler usually wore her hair pulled back in a pony tail and subconsciously accompanied her directional cues with a head movement that tossed her ponytail. For the regional competition, she had dressed more formally and had put her hair up. No ponytail. The dog was unable to respond because the cue that the dog had selected and learned was not present.
The most important characteristics of a cue are that it be clear and distinct. If you are not aware of what cue or cues the dog is selecting, then the chore of making that cue or those cues separate and distinct becomes quite difficult.
To add more confusion, here is a piece of Pavlov’s research which I came across recently in Fundamentals of Learning and Motivation by Frank Logan.
One use of the procedure of differentials classical conditioning is to determine how fine a discrimination an animal is capable of. With humans, of course, we can simply ask whether two tones sound alike or different, but with animals, we need to develop some nonverbal response by which they can communicate to the experimenter. Clearly if an animal can learn to respond differentially to two stimuli (cues), he can discriminate between them.
In an effort to obtain such information, Pavlov would first employ differential conditioning with stimuli (cues) that were quite dissimilar, say a very-high pitched tone versus a very low-pitched tone. He would then make them progressively more similar in an attempt to find out where the discrimination broke down indicating that the animal could no longer respond differentially. Interestingly enough, he found that when the stimuli (cues) became very similar, not only did the discrimination break down, so did the dog!
The dog would show obvious signs of fear and anxiety about the experimental situation, so that rather than standing quietly in the stock and salivating when appropriate, he would resist the situation. This behavior carried over to his total behavior. He would huddle in a corner of his living cage, cower at the sight of his familiar handler, refuse to eat regularly, and overreact to the slightest sound or distraction. Pavlov called this result an experimental neurosis because of its apparent similarity to many human neurotic behaviors, and he usually had to send the dog away from the laboratory for a rest cure of tender care in the country.
Considering all the above factors, it appears that is important to pick cues that are separate and distinct and to use those cues in a well-defined manner. All of this should also tell you that it is quite important for the puppy to be trained by only one person. No two people have the same body language, so no matter what the word is that is emanating from the human mouth, two different people are going to be sending two different signals in terms of body language. Pavlov’s experiment demonstrates that lack of consistency of cues has the potential to cause some problems.
You will greatly enhance your if you occasionally check to see what cue the dog is using. Try training with your mouth shut. Give no verbal cues, only gestures. Try it with sunglasses on. That will disguise your eye movements. Try it sitting down instead of standing. You might be surprised to see how often the dog is responding to a signal other than your words.
Because fuzzy cues make it much harder for the dog to behave as desired it is unwise to have two different trainers during the dog’s active learning periods. He will survive it. He will probably even learn in spite of the trainer because of his flexibility, but the learning process will be much more difficult for the dog.
As no two people use the same body language, a dog with two trainers will get inconsistent cues. That is not a recipe for success.
A prompt is an enabler or trigger for a behavior. One example in dog training are a raised hand “traffic cop” gesture to induce a dog to stay. A trainer uses the prompt to produce behavior so he can reinforce it. As the behavior is becoming established the trainer fades the prompt and continues the reinforcement.
Another useful prompt is walking decisively in the direction you want the dog to go when giving a directional cast. The dog has an innate propensity to herd as a legacy from his wolf ancestors. When wolves need to kill a large animal they must operate jointly to bring it down. This joint behavior is basically herding. In the dog it manifests itself in the dog moving the direction that he sees the trainer move.
The first few times the dog is given a directional cast the trainer should take 3 or 4 or 5 decisive steps in the desired direction. Then, as the dog gains proficiency on directional casts, the
prompting steps are faded from 5 to 4 to 1 to 1/2 to a simple lean in the desired direction. The reward comes from the dummy or bird found at the end of the cast or casts.