|
.jpg)
What is this clicker stuff anyway?
Clicker training is one of the training methods routinely used at Positively Canine. It's not a new method, but many people are not familiar with it in dog training. Before it became popular for training pets, it was used to train marine mammals - large, intelligent animals that you can't put a leash on and make them do stuff. It's also used to help zoo personnel do routine care for their animals . Now they can train a hyena to stretch his neck against the chain link fence, and stay there for a blood draw, instead of risking anesthesia. It's even been used to train animals to act as govenment spies back in the cold war. A fascinating watch on the history of clicker training is Bob and Marion Bailey's DVD "Patient Like the Chipmunks"
The clicker, itself, is just a small object which emits a sharp click (much like clicker toys we had as kids). The way we use it is as a marker - both for a behavior and for a reward. It tells the dog "I like what you are doing this very instant", and "you're going to get something good for that." Something good can be anything that is rewarding to that particular dog. Treats, dinner, a walk, a toy, a snuggle, even sniffing a bush. The dog has to tell us what is rewarding. If it's not important to the dog it won't change behavior. When you click, you are "taking a picture" of the exact behavior you want. It's important information for your dog. And it is important to always pair the click with a reward of some sort. Otherwise it becomes just a meaningless sound.
What is Operant Conditioning? The use of the clicker is based in the science of Operant Conditioning (OC). Unlike Classical Conditioning (Pavlov) which works to change involuntary behaviors and emotional states, Operant Conditioning (Skinner) allows the learner to influence his environment and outcomes by offering, or not offering certain voluntary behaviors. Some people will tell you that clicker training "is" Operant Conditioning. In fact, it is only a part - based on one quarter or quadrant - Positive Reinforcement. (I will provide a footnote at the bottom if you want to know more about the total of Operant Conditioning theory.)
Why use the clicker instead of my voice? For some reward based training, (luring with food for example) your voice is probably almost as useful as the clicker. In training where you are basically helping the dog or showing him what you want, you don't always need the precision of the click. The clicker is like a scalpel. Your voice is more like a butter knife. Both cut but they have different uses.
In clicker training, as with the rest of Operant Conditioning, the learner is discovering for himself what works and what doesn't. In that case, it pays to have a very precise way of communicating when the learner is right. Our dogs hear us talk all the time. Mostly when we talk it's not information for them. The clicker is a very unique sound. It never sounds angry or worried or hoarse. It doesn't sound like other things the dog regularly hears. Some research has also suggested that the click is processed in a more primitive part of the brain, and so doesn't have to be "interpreted." It may go through the hippocampus, which is a part of the brain which is linked to long term memory, and also to SEEKING systems, which make us enjoy seeking out new experiences.
Will I always have to use a clicker? I tend to keep them around because I like to occasionally sharpen up a behavior, or teach a new one. But, once I have a behavior "on cue" (I can say the word, the dog consistently does that behavior) I no longer need the absolute precision of a click, and my voice gives enough information.
If one quadrant is good, wouldn't the whole thing be better? Not necessarily. In clicker training, the dog is trying different things to find out what you will reinforce. You can teach more complex things than you can guide your dog into doing. The dog is a willing participant, and looking for a way to earn reward. If you start introducing aversives which he wants to avoid, the dog is now in conflict. He doesn't know if a good thing is coming for the behavior, or if he may be punished for making a mistake.
There are people who call themselves "balanced" trainers who claim to use positives when the dog is learning, and then add aversives in the "proofing" stage. The problem I see with this is that learning is frequently not linear. There may be something in the environment which is confusing or distracting to the dog. New lessons may change the dog's understanding of things you thought he knew. A lack of practice may make a formerly known skill less clear. How do I really know if the dog is confused or being willfully disobedient? The answer is simple. If my dog has a history of reward for doing the right thing, why would he intentionally choose to do the wrong thing? On the other hand, if he's only working to avoid unpleasantness, he'll give me just as much as he thinks he has to. And if he never knows if he's trying to earn reward or avoid an aversive, he may be too stressed and confused to learn as effectively as I'd like.
It's not precisely clicker training, but sometimes I will use removing the opportunity for reward if the dog is doing something I don't want. For instance, if the dog is pulling on his leash, we won't go where he wants to go. Only loose leashes make my feet move in that direction.
Of Alphas and Whisperers There's a whole school of training (recently popularized by some reality TV shows) which tell us that to teach our dogs, we need to make them think we ARE dogs. There are all sorts of superstitious beliefs and strange behaviors connected to this theory like spitting on the dog's food to show him that the "Alpha" got to eat out of his bowl first.
The fact is, dogs know we aren't dogs. We don't look like dogs, and when we try to "act" like dogs, we do so clumsily - without the subtle body language and timing a dog naturally has. And the things we want from a dog are not the things another dog wants. Any "training" that takes place within a group of dogs is about social behavior. I've never seen a dog cue another dog to sit . And many of the things we DO ask of dogs (come quickly to me in a direct line, give me sustained eye contact and stay very close to me) would probably get his clock cleaned if he did it to a higher status dog.
Unless I want my dog keeping his distance, avoiding my gaze, etc., there's no reason I would act like an animal who would expect those responses. Many ethologists believe there is no such animal as an "alpha dog." Dogs, like all social animals, have some interest in status. But that status is fluid, not absolute. And the higher status animals generally don't seek out and engage in battles of wills with others. They don't need to. The animals who challenge others and make a point of big displays are generally ambitious youngsters who might want to test others to see if they can "move up". Generally, they get the answer "no" from more mature and established animals. When we try to imitate dogs, without a true knowledge of their motives, without the ability to really communicate in dog as a "first language" we can come off as socially inept bullies. It's not a wonder that some people get bitten for trying this strategy. It's a wonder that dogs are so tolerant that so few people get themselves "put in their place" for this sort of behavior.
I've spent three decades studying dogs and trying to learn their language. But I have to admit, I'll never really be "fluent" enough to communicate totally as a dog (and since I want different things, why would I?) I also have dogs who seem to have a farily big vocabulary of "English" words. But I'm going to understand that I can not ever expect it to be their "first language." It seems to me that an obvious and useful compromise is for us both to learn a new and common language that is available through clicker training.
copyright 2010
Sandra Case BFA, MEd, CPDT-KA
* Footnote. Operant conditioning has 4 quadrants made up of two different sets of variables.
Positive means something is added. Negative means something is taken away.
Reinforcement makes a behavior more likely to occur. Punishment makes a behavior less likely to occur.
A lot of confusion occurs because people attach emotional meaning to the terms - positive = good, negative =bad. reinforcement is something kind and punishment is mean. When talking about OC, it's best to think of them mathmatically.
Positive Reinforcement is adding something to make behavior more likely to occur . Positive Punishment is adding something that makes the learner less likely to repeat a behavior. Negative Reinforcement is getting the dog to understand that they can get something to stop by giving a behavior. Negative Punishment is taking away something a learner wants to make a behavior less likely. Then there is Extinction - when a behavior doesn't do anything, the learner will probably give it up.
References:
http://www.behavior1.com/
http://www.clickertraining.com/
Video: Patient Like the Chipmunks by Bob and Marion Bailey
Books: Don't Shoot The Dog by Karen Pryor
Reaching the Animal Mind by Karen Pryor
Photos on this website are copyright Positively Canine LLC and the photographer. They may not be reproduced without permission.
|
