This post is part of a series of several posts on cues.
When teaching a new behavior, we have control over when we introduce the cue. Many clicker trainers advocate getting the behavior before ever introducing the cue. Here’s a bit about each teaching method.
Introduce the cue, then the behavior.
Many traditional trainers and some clicker trainers teach this way. For instance, the trainer tells the dog to sit. The dog stares at him dumbly, as she has no idea what the word “sit” means. He pushes her into a sit, and gives her a treat. He repeats this a dozen times, until the dog eventually realizes that she can get the treat faster and avoid being pushed around if she just sits as soon as it hears the word “sit.”
This can be an effective way to teach cues. However, by teaching this way, we often end up using quite a bit of negative reinforcement. Even if we are using treats, the animal is still following the cue in order to avoid being physically pushed into a sit. Also, unless care is taken, the animal can get confused or frustrated at the beginning when the cue has no meaning.
Most traditional horse training works this way, mainly because traditional horse training is largely dependent on negative reinforcement and pressure for teaching behaviors. Some trainers like this because they have the pressure and negative reinforcement to back up the cue if the animal does not comply. However, this is an easy way to create an animal who is complying only to avoid further repercussions.
Get the behavior, then add a cue.
This seems counter-intuitive, but it’s actually a very successful way to train behavior. First, we get the animal consistently offering the behavior. With behaviors that occur frequently, we can capture and reward whenever the behavior occurs. For less frequent or novel behaviors, we can lure or shape the animal until we get the behavior.
Then, there’s several ways to add on a cue. A trainer with a good eye can give the cue right when the behavior is starting or right before the behavior starts. Or, if the behavior is occurring frequently, a trainer could alternate between cue and no cue. Give the cue, see the behavior, reinforce. Then let the animal do the behavior and don’t reinforce. Repeat until the animal gets the hang of it. In any case, next we have to extinguish the behavior from occurring off cue and begin to put the behavior under stimulus control.
One advantage of this method is that the meaning of the cue is totally clear. The dog already understands we want it to sit, so by the time we introduce the cue, sit means sit. In the previous method, the sit cue is a bit more ambiguous to the dog. The sit cue could mean sit, or it could mean wait around until the trainer pushes the dog into a sit.