The removal of punishment is a reward. This has been scientifically tested... I can give you the citation if you like.

No argument from me on this Dave. In fact it verifies my contention about Natural law and equal and opposite action reaction laws of life. Avoid the punishment and learn about the praise and its benifit from continuing good behaviour for that reward of praise or food or whatever.

You yourself use the natural law by using Punishment although the absense of the actual act of punishment in the same sentence as Reward. Natural Law.

Also though and I am by no means at all knowledgable in clickers but logic tells me that the absence of a click for a behaviour is lack of the indication of approval and hense means a demeen and do it over again with the manual method which in itself is forced neg no matter how small the effort is you must put out.

You mention trials and a dog being slow to sit and the thought to disobey. Here I am not in agreement.

You, approach this trial from the negetive seeking the positive when a person going into trial should be working from the other end.

What I mean is seeking perfection and accepting nothing else. In other words from the top back.

I myself would never enter a trial seeking to gain points I assume I might lose a few and with a few of those lost points from my own mistakes. Gain/Lose Laws of Nature

I will buy the book as it seems to myself that there is something here worth investigating.... But...

Lets get away from a obed trial or conformation or sport protection and get back into a more realistic location as in most pre arranged trial situations the pattern can be trained to get through something that the dog will not get through without the comfort of the patterned situation.

Your imagined dog is big time working 100% with the aid of a clicker in the day to day obed. Then some distraction takes the dog into a major fight with another dog and not of your dogs will but so be it. The fight goes on and as good owners we now break it up. Your dog learns a fear conditioned response and although it was suppressed prior to the fight it now has learned a new bad behaviour to strike first before getting a licking a second time. OR, Worst case, the dog developes a human aggression problem and becomes a liability but can be re programed. Click or Punishment?

Do you click or punish to prevent this dog from proceeding to the next level.

Just a question?

Does this make sense?

Regards
Jerry

Jerry,

As much as some people would like you to believe it, very few ppeople are going to be able to use a single method, like positive reinforcement, for total dog traning. The use of the positive methods doesn't preclude correction. Some people may be able to manage it that way , but I doubt very many. You can use the clicker/treat and still use correction if required. The author of the book dejure won't say that, it doesn't sell books.

Richard, this is nothing new to me. My reason for wanting to know more about this subject is to simply understand the application.

Truth is a strong majority of my private basic obed training is dirived from folks who tried the Dunbar and now the clicker methods and were still lacking a sense of the better control in a tough situation.

You could say that I love the Clicker and Dunbar set as they pay the bills and more often it is as simple as a correction or two to sort the dog out and gain the trust of the new client.

One of the common responses from the client when they see the positive response to the timed and correct disapline they almost all ask why were they not shown this at the classes.

One of the better lessons I learned when attending a seminar By Daniel Paul from Reims France was that what he was showing us that day was the type of work that gets results for him. He went on to say that for every one of his ways to work a dog there are 1000 more that will get the same result depending on the skill of the dog and your own personal temperment as a person.

As you said Richard...........


You can use the clicker/treat and still use correction if required. The author of the book dejure won't say that, it doesn't sell books.


Yep you are so right, but it does bring more than my fair share of new clients.

God Love Em

Jerry Cudahy

What I've noticed that most clicker training advocates seem to leave out of their training philosophy is the third component of reinforcement theory, which is punishment. I don't think that makes the clicker a useless or ineffective training tool, however. I really think the clicker has a place. To rely upon it as the end-all and be-all of a well trained dog probably won't result in a completely trained dog. I think it's a good place to start, especially with pups where you want to keep aversives to a minimum, and in boosting enthusiasm for the learning process. You can always add aversives when you fade the clicker.

Jerry, I use correction also. But if I have to summarize my philosophy it would be this... when I go into a trial I want the PRIMARY motivation (goal) of the dog to be attaining a positive reward (food, ball etc.)

Ideally... in the back of the dogs mind is also the idea that there will be a price for failure to comply. However, I don't want the dog FOCUSED on punsihment.

In trial, I DO want the dog to hear the command... "sit"... dog sits-- gets no reward but thinks -- maybe I'll get it on the next one if I sit more quickly... and so on. This way... the lack of reinforcement BUILDS motivation to work harder. Somewhat the opposite of what I described earlier.

I haven't posted before but the misinformation regarding clicker training is ridiculous. Not just from this board but in general.
Clicker training didn't originate to train dogs. Karen Pryor developed the methods from her work with dolphins. The reason there is no correction is how are you going to physically correct a dolphin or killer whale. Prongs and shock collars don't really work that well. The techniques apply to all sentinent beings and are used to train all types of animals. For those of you who read her book "Don't Shoot the Dog" you will be very suprised at how little of it focuses on canines.
As far as the technique not working with "hard" dogs even a dog of the worst temperment is no match for how dangerous a killer whale or wild animal can be. One of the main ways clicker training is used is with zoo animals. Many zoos have policies limiting human contact with the animals to keep their natural instincts intact. Still the animals have to be worked with. Clicker training works perfectly for this. You are not going to be able to use correction in a situation such as this. Yet with the clicker handlers can work with elephants safely. I have seen this done. The handler is never even on the same side of the fence as the elephant. Yet can get the elephant to place its feet through so they can be cared for. They can have the elephant turn around and place its rear feet through the fence. This is just an example. All this direction comes from using the clicker.
Someone, also, commented that it wouldn't work to perfect behaviors. Actually, the clicker is perfect for this because it isolates the exact moment the dog is in position. Freestlye is a new performance sport that involves dancing with dogs. The behaviors required are outrageously complicated. The majority of dogs are clicker trained. The reason being that everything has to be perfectly timed.
The correct method of using the clicker is to capture a behavior. As one person mentioned you build upon behaviors. Targeting is the best example of this. You give no commands to the animal, the command will be used as a marker once the behavior is learned. Place your target near the dog(the animal should be off leash). Any attention the animal pays to the target is rewarded with a click(at this point the click as a reward has been established by associating the click with a food reward). If the dog looks at the target click. Repeat until it is obvious the dog is repeating the behavior to get the click. Then the dog must progress by moving towards the target, and so on.
As far as research on this method there are numerous studies going back to Pavlov's theory of Operant conditioning.
Clicker trained dogs become thinkers and will begin to work you. I had an experience in agility with a classmate's collie. We don't use clickers at higher levels because the dog has learned all the behaviors and we are the ones now learning strategy. As we were waiting our turns the lady's collie began to enter the weave poles. He made a correct entry and came back to his owner. She rewarded him. He repeated the behavior. This occured a couple of times. Finally, she stopped rewarding. The collie returned to the weave poles and entered again then stopped and look at the handler. The thought process of the dog was obvious. He was trying to figure out what to do to get rewarded again. So he went through the second pole. Bingo reward. I wasn't that familiar with the handler and her dog and I commented he acts like a clicker trained dog. She said he is. We, both, commented that you can always tell. Clicker training places the responsiblity of learning on the dog.
Now that is sounds like I think clicker training is the next best thing since sliced bread, its not. It is only as good and as useful a tool as the person using it. The major problem with clicker training is that it requires a good understanding of the theory behind it. I learned the basics over a two day seminar. The first day my dog didn't even get out of her crate. All we did was take notes. You can't just click and treat and think your dog will learn how to dance. Too many people are using it to train without understanding what they are doing. The clicker is the same as a prong, it only works if you know how to use it.

Skipper

Good post Skipper.

Skipper, your post was very well presented.
Would it be fair to say that the clicker works best with a set pattern?

I am assuming that you will say yes to this and so what about a third party that is themselves different in each routine and the third party acts or behaves in a wide varience of behaviours?

Here is what I am getting at.

I have no problem understanding the concept or how the clicker will work to establish a relationship and behaviour mold between a dog and the handler for anything that involves obed, retrieves, jumps. This is fairly straight forward.

Lets now move into another routine that includes a decoy who has the latitude to do anything the decoy wants. In other words not a set pattern, body positions, demeanur, etc.

Now, the next time that the dog must go through the same set of exercises the next decoy changes and all the same before mention varience of latitude is again changed in the dogs mind.

And again the next time out and so on and so on.

Clicker can it then be applied to compensate for all these individual changes per succeeding routine?

Thanks
Jerry

 Quote:

Originally posted by Jerry Cudahy:
Would it be fair to say that the clicker works best with a set pattern?

No. The clicker reinforces any behavior just like saying "good boy."

In protection, the primary reward comes through the bite... drive satisfaction.

Jerry,

Almost all behavior is what you could call pattern training. It is a response to an enviornmental stimulis. The patterns can be as long or short as you want to make them. The more complex the organism the more flrexability you have in terms of the stimulus. As long as you are working with an organism capable of responding to an enviornmental stimulus you can modify what the organisim does. In simple terms, for a plant modify the direction the light comes from and you can modify the direction the plant grows. For a single celled organism you can stimulate an aversive reaction to a light being turned on by applying a negative stimulus when the light comes on. For a mouse you can teach it to push a lever by providing food or water. By using a combination of the three types of enviornmental reactions to behavior (reward, extinction {neutral}, or negative) you can modify almost any behavior that the organism is capable of. Each of the results of behavior have different levels of effect on behavior. They can be used independantly or in combination to modify any behavior the organism is capable of. Some of the consequences need to be presented earlier, some can happen hours later (taste aversion) some days or weeks later (a pay check, or being fired). The more complex the behavior, or the organism the more flexability you have.

The only limit is the patience and available time of the trainer. Once you can identify what controls the behavior you can use that to modify the behavior. One good thing is that these principles can be used wheather or not you really understand the all the principles involved. In dog training there are basicly two things most people use, reinforcement (positive and negative) and punishment. There is another powerful too that most people don't use (extinctsion) because it is often difficult to use. Since there is no such thing as true altruism, the consequences of any behavior can used to modify any behavior. Part of the problem is that hen you look at behavior, the tendancy is to only evaluate the results at a set point in time. The other thing that needs to be evaluated is the skill of the trainer at the use of the tools they are using. Each person is different and that is what makes the difference in evaluating a method. A skiled trainer working with reinforcement (positive or negative) and punishment can get the same results as someone using only positive reinforcement and extinction.

In termes of the principles of operant conditioning the most powerful tool for eliminating unwanted behavior is extinction, not punishment. This is one of the tools that the positive trainers use. They often don't understand the principle, and don't describe it properly, but they do use it. So the idea that any training uses only positive reinforcement is really a mis-statement. There are other tools involved. If you are interested e-mail me privately and I can recomend a book that may help.