 |
||||||||||
|
|
A guide to operant training with positive reinforcement:
The principles of this kind of operant training (aka clicker training) are easy to master. There are only three principles that guide this style of training:
Be fair to your dog In this style of training, we try to look at the world through the dog's eyes. To do this, you'll find it useful to learn a little about dog behavior and how dogs view their world. There are wonderful resources to turn to, in particular three books that embody this approach to dog training:: They say that we punish dogs and children because we can. Although punishment may not be inherently unfair as a training method, I would argue that in practice P+ is unfair to the dog. Because we punish first (good timing, you know!) and think about it later, usually we punish the dog for a mistake that we made such as an ambiguous cue or for something the dog did not understand in the first place, which was, yes, our mistake as trainers. And punishment does harm, often permanent harm to the dog's attitude and even to the dog physically. In contrast, using rewards improperly does no physical or psychological harm to the dog; if you make mistakes you simply slow your dog's learning.
Manage your dog's environment and his reinforcements If you're going to throw out punishment as a way of changing your dog's behavior, does that mean the dog gets to run around like a hoodlum, taking whatever it wants? Absolutely not. In the words of two master trainers:
Dogs and children probably do have a great deal in common. What will be best for them, and what will make them happiest, is NOT to get whatever they want. Nor is it to be told constantly "No! Don't do that!" Rather, they need you to set boundaries, provide structure and safety for them, and to reward them when they make the correct choices.
In this style of training, you manage your dog's environment so that the dog can make the correct choices. Or, you can think of it as preventing the dog from making the wrong choices, if you like.
For example, why would you leave a puppy home alone all day loose in the house if you did not want it to relieve itself on the carpet or chew the antique furniture? The puppy would be safer and you'd be happier if you confine the puppy to a smaller space, leave it a paper or litter box to use for its business, and give it something you approve of to play with and chew.
Another way to view management is that you decide the dog's access to the things it wants. Dogs are fairly simple; they generally want a predictable list of pleasures. Each dog is a little different, but you get to know your dog and what she or he likes. You then become the gatekeeper for the things that your dog wants and in doing so show him how the world works (such as be safe and be polite to others). This view of benevolent leadership (of your dog by you) has replaced the misunderstood concept of "dominance" with the model of a strict but loving parent.
Reward your dog for making the right choices This last step is the essence of shaping your dog's behavior to be whatever you like it to be. Rewarding your dog for being "correct" is essential for this method to work. A dog that is only managed might be acceptable to live with, but he won't understand what you want him to do. If you reward his correct choices, the light bulb will go on for him, and he will not only understand, but he will be an eager participant in your life together.
A reward is anything that your dog likes. Not all rewards will be acceptable to you, of course (like cat poop). But if you pay attention to what your dog likes, you will come up with a long list of buttons to push that will result in your dog doing what you want it to do.
A reward in this kind of training is different than a lure or bribe. A reward is something the dog gets AFTER she has done something correctly. A lure or bribe is something you offer BEFORE she performs a desired behavior, and you do this to show her what you want (a lure) or to convince her that whatever else she had in mind is not as good as what you can provide her (a bribe).
It's too complicated to explain here why the possibility of a reward is SO MUCH MORE POWERFUL than the certainty of a lure or a bribe. For now, just think about how much people like to go to Las Vegas to gamble; there's a lot of behavioral theory involved, but the allure of gambling and operant conditioning are similar <g>.
Here is a list of rewards, along with some of the terminology from behavioral theory:
Secondary or conditioned reinforcer = something associated with a primary reinforcer Where's the clicker come in? Didn't we say up at the top that this is clicker training??? Actually, you can train successfully this way and never use a clicker. That's why it's misleading to call it clicker training. And importantly the clicker is only a small part of what you do under this method.
However, using the clicker opens up an entire world of magic! You can communicate with your dog about 100 times more effectively if you use a clicker to mark the behaviors that will earn your dog its reward. The very best book on the market to explain how this is done is Try it! It can't do any harm and it's SO MUCH FUN. You and your dog will love it. The clicker is an event marker and it's also a conditioned reinforcer because you initially pair the click with a piece of food, or the chance to play with a toy--in other words with a primary reinforcer.
Common conditioned reinforcers that you will likely use are:
|
 |
|
© Catherine Toft 2007 |