Vizsladogs, Ltd.

What is Clicker Training?
by Debi Davis


Hi Kay, Delighted to have you post my post to your website! Clicker training works so well with Vizslas. They are extremely responsive to non-coersion, and tend to learn at warp speed when they have choices they can make. I am extremely taken by the breed, and have also found that many of them can make outstanding service dogs. I would not say they were "headstrong" as much as driven by genetics, specifically programmed to kick in certain drives. What I have found is that dogs who are sensitive in general do very well with clicker training, because it invokes their choice, and never pushes them past their limits, to where they become stressed, and learning stops. They can learn so fast! I think they are amazing dogs.

My clicker instructor, Rosemary Besenick, has really turned the AZ field folks on to this type of training, and some of the old timers thought she was totally nuts and could never have a solid field dog this way. Yet her 1 yr old Vizsla-in-training has such incredible control and focus, the old timers were forced to accept that it CAN work. And these are people who would never consider training without an e-collar. Rosemary's Vizsla has never had as much as a leash pop!

They are making real inroads in shaping for field work, and Rosemary and her husband are holding some terrific classes. If you ever get to Tucson, I think you would LOVE attending a class and seeing how well this works with the breed. Her email addy is PerroViz@aol.com . She owns "The Dog Training Company", and opened her own facility a couple of years ago, after leaving the service dog training field. Lots of service dog folks still train at her new facility, however. She has tons of vet referrals now, as she does such good work with agression and severe behavioral problems. But she really, really likes the field work with her Vizslas! Thanks for the lovely note. Debi 


 Clicker Training is a methodology based on using two parts of operant conditioning: Positive Reinforcement, or +R--"Good things hapen" and Negative Punishment or -P--"Good things end."

The philosophy of clicker training is very different from traditional command-based training. Clicker training is based on being "pro-active" rather than re-active. The clicker trainer doesn't coerce nor command nor force the dog to respond to a cue--there is no "or else." We don't wait for an error to happen so we can "correct" it. There is no punishment for "misbehavior." Does this mean it's a free for all? Totally unstructured? Nope.

It means that the clicker trainer works to invoke the dog's own volition. The dog is incrementally "shaped" to perform any given behavior, or chain of behaviors. The dog chooses, the trainer does not force.

So how can this work? Doesn't it mean you then have totally out of control dogs? It would seem so, but the laws of learning don't work that way. The structure is there, but it's just subtle.

The clicker is nothing more than a little plastic box with a metal tab which makes a clicking noise--a bright sounding click--when depressed. The click is like a camera click: it isolates a moment in time and acts as a "marker", giving information to the dog. It's like saying, "YES! Great! Right NOW!"

We use "markers" all the time in our lives. Think of the last time you had your back scratched. When the person doing the scratching got to the really itchiest part, you probably went, "Ahhhh!" and "marked" that itchy spot so that the person could scratch just a bit longer and harder right there. The "Ahhh" was a marker, just as the clicker is a marker.

In dog training, the clicker "marks the moment" you want the dog to remember, and this is followed up by a treat or something the dog really likes, is really motivated by. So why not just chuck food?

Well, you can. And in some instances, this works just fine. But, one can have problems with chucking food. Dogs can easily become "food crazed" and only give the trainer brilliant behaviors when food is present. Also, it's not very accurate for teaching really complicated chains of behaviors, such as the retrieve. It can be hard to get the food to the dog at exactly the right moment--the moment you want the dog to remember.

This is where the clicker works to bypass focus on food. The dog begins to problem solve, to offer behaviors which will "make" you click--and the food is not delivered until after the click--sometimes several seconds after. The dog works for the click, not the food, though the food is the primary motivator. Just that few seconds of delay can make all the difference in performance down the road.

Clicker trainers use liberal doses of punishment! But they use "negative punishment, or -P--the removal of a good thing. When the dog offers a behavior we don't wish to reinforce, we simply ignore the offering, not reinforcing it. An example: dog jumps all over person coming in the door. Person ignores dog, turning back, makes no eye contact, no petting. This is using -P or Negative PUnishment because good things will END until the dog offers a behavior you are looking for. The moment the dog stops jumping, plants 4 feet on the floor, the person clicks and offers a treat, pets dog, gives the dog what it wants the most--ATTENTION. It's still punishment to withhold what the dog wants, but it's not adding an aversive to accomplish the behavior change.

Clicker training, because it is so "Pro-active" depends also on developing skills of observation, and working with observable behaviors. We don't worry much about why a dog is doing what it is doing, or trying to psychoanalyze the dog's reasons for doign what it is doing. We merely move from what we see. Dog is jumping on us, we remove the thing the dog wants and give the dog what it wants once he has offered a behavior we want. We don't care if the dog is jumping becuase it's excited, or because it came from a kennel, or because we perceive it may be acting "dominant" or any other reason. We observe the behavior right there in front of us and respond accordingly.

We don't use positive punishment (adding an aversive) for misbehavior. Why? Well, we know that mistakes are a vital, important part of learning anything new. A young child just learning to walk takes many awkward steps, falls, gets back up before he learns to balance, and finally to run. Do we slap that child down for the errors made? of course not! We allow the errors to help the child learn what does and does not work. We encourage the child to make mistakes in order to learn what DOES work.

We could no doubt get a behavior change on the jumping dog by adding an aversive. We can knee the dog in the chest, we can holler, "Bad Dog!", we can throw in a leash jerk, and a whole lot of things. Yep, it will probably work. But why add an aversive if all we want to do is get the information to the dog about what we want? All we need to do is refuse to reinforce behaviors we don't want. By ignoring the jumping, we know that dog will most likely stop the behavior in the next 30 seconds or less, when he is getting no reinforcement for it.

The beauty of this is that the dog figures it out on his own. And one quickly sees that "AHA!" moment when the dog realizes that "Oh, you mean all I have to do is plant my feet on the ground here and I'll get pats and treats and all sorts of good attention???"

Obviously, this example is very simplistic. But it is the principle upon which clicker training works: we mark the moment we want the dog to remember, we reinforce with something the dog wants, and we take away good things when we are offered an inappropriate behavior.

Because clicker training is so based upon observable behavior, one learns to develop a keen skill of observing dog behavior. We read the dog's body language a lot. When we are working with desensitizing a dog exhibiting fear agression, lets say, at the approach of another dog, we would first note at what point the dog began to throw off "calming signals" or displacement behaviors, such as hackles rising, lip licking, tightening of posture, etc. Then we would begin our reinforcement at a point just beyond when the dog begins to stress. If the dog shows signs of stressing at 50 feet from an approaching dog, we begin to reinforce for OTHER behaviors we do want at say, 60 feet. We engage the dog's volition and get the dog's attention, and rapidly reinforce what we want at that point, gradually moving closer to the feared object.

What this allows us to do is to capture the dog's attention without adding an aversive to suppress behavior. When the dog's adrenaline is surging, when he is totally stressed, little learning will happen. So we work from the periphery of that stress, constantly moving INTO it in little steps, small enough steps that the dog can handle it and develop new responses.

Clicker training helps the dog develop problem-solving skills, and to offer new behaviors. This is especially important in service work, where we may have a problem we need the dog to solve, but the dog has never been taught to do that behavior we need. The dog who has been encouraged to continue problem solving, and not just waiting for a cue to perform a behavior--is the dog who can problem solve when you need him to do this the most.

When my friend fell out of her car and her wheelchair rolled away, into a ditch, she cued her dog to go fetch her cell phone, which was in her backpack. She'd never taught the dog to open the backpack. But she had played lots of problem solving games with the dog. The dog first tried to bring her the keys when he couldn't get the phone. Then brought her the jacket hanging over the top of her chair. She ignored these offerings. Finally, the dog went back, sniffed for the phone, figured out how to open the zipper right then and there, and got that phone out and brought it to her.

Same for my Papillon who could not get the wheelchair to me, as it had lodged on a chair leg. He tried all the ways he could to loosen it up, pulling right and left--but it remained stuck. But, thanks to problem

solving games, he finally tried something that DID work--he backed up across the room, took a running jump and leaped into the chair, successfully dislodging the stuck wheel, and then was able to pull the chair the rest of the way to me.

Hope this helps in understanding a little bit about how clicker training works!


Vizsladogs, Ltd.
started
5-21-95 © 1995 - 2006
Last updated 02
/06/06