I wouldn't say that, no. Perhaps he's not a clicker purist, but he is using a clicker in training, so I'd say that some portion of his training is clicker training. Since you can't be using reinforcement and not be using OC, then he is by definition using OC.

As far as the "keep going signal", there is more than one way to think about it:

"Police Officer Steve White told me of sending his German shepherd patrol dog to search for a thrown object that had landed on top of a six-foot-tall clump of bushes. The dog searched the ground fruitlessly for a long time. Then, when it happened to raise its head, Steve clicked. The dog instantly sniffed the air at head height, alerted to a whiff of the target, and began searching around the area while scenting further upward, even standing on its hind legs to do so. Thus with no further help from Steve, the dog located the object, crashed on top of the bushes, and got it.

Another aspect of Steve's communication with his dog was that Steve used the click as a reinforcer that was not a termination signal; instead it was a "keep going" signal. The click reinforced the upward sniffing and kept the search behavior going, since the lost object had not yet been found..........

Many of the novice clicker trainers I worked with in the 1990s were what author Morgan Spector calls "crossover" trainers........I found that they were all too willing to give clicks but no treats, to the point where the significance of the click was extinguished. It was necessary to stress "One click, one treat" as a general rule, in order to teach people to shape behavior efficiently.

However, there are many situations in real life where some interim reinforcing stimulus can be very useful, as with Steve White's patrol dog. One answer is to use a different reinforcing stimulus to tell the trainee, "That's right, and keep going.". Interestingly, a "keep going" signal does not have to be linked directly with a primary reinforcer. Just start inserting it somewhere before the terminating click, and the learner will soon recongize it as a signal leading toward an eventual reinforcer.........

As with a terminating click, it doesn't matter what kind of stimulus was used: a clicker, a whistle, a shout, or a wave. What counts is the fact that the stimulus was not just hopeful encouragement or cheerleading, which may distract the animal or accidentally reinforce the wrong behavior, but a well-established and precisely used conditioned reinforcer."

Karen Pryor, Don't Shoot the Dog, c. 1984, 1999

K-dog the clicker purist!! LOL is pairing the click to the highly rewarding sensation of thrashing, mashing, killing, and nashing the nasty little prey item.

The click then is like Pavlov's little bell, it stimulates the same highly rewarding sensations in the dog (CC). . .to be used as the trainer sees fit (OC). Marking correct behavior or giving small enticing levels of reinforcement to maintain drive and intensity. . .or motivation.


I think I've got that right?

Lisa wrote:

 Quote:

Interestingly, a "keep going" signal does not have to be linked directly with a primary reinforcer. Just start inserting it somewhere before the terminating click, and the learner will soon recongize it as a signal leading toward an eventual reinforcer.........

Isn't this saying what I am trying to say. You can have a 'keep going' signal, but it is NOT the clicker! Because the 'click' TERMINATES the behavior and means the reinforcer is given (and that may mean the prey drive thing).

The way I am understanding this, is that the dog is not going to be learning by operant conditioning per say. In as much as the dog is always working for the 'click' specifically because that's where the payoff comes in immediately after. The click is not nearly so effective when there is no immediate payoff as far as learning a NEW behavior, sounds more like it's a 'lure' from a distance. So a verbal cue, or the collar with a 'beep' for the correct behavior would be just as effective for giving 'hints' that the dog is on the right track.

From http://www.clickertraining.com/karen/art...at_history_ct_s

 Quote:

The clicker is not just a conditioned reinforcer, a substitute for food. Yes, it is that, and it’s also a bridging stimulus, meaning "Food is coming, but later." And it’s a termination signal: ‘Job’s done, come collect your treat." And, as Ogden Lindsley has dubbed it, it is an event marker. But it is more specific than that, and much more powerful.

And from http://www.clickersolutions.com/articles/2001/ocguide.htm

 Quote:

Reinforcers
It's easy to see how important reinforcers are in Operant Conditioning. We all work for reinforcers every day. I am writing this article partially because of reinforcers. There is a mailing list, I give input, people reply positively to my input, my input on the list increases. You can see how well it works. Dogs are no different.

Skinner defined two types of reinforcers: primary reinforcers and conditioned reinforcers:

Primary reinforcers are something an animal instinctively and inherently finds rewarding. No learning is necessary for these reinforcers to increase the likelihood of a behavior. In dogs, primary reinforcers include food, water, procreation, and sometimes certain predatory behaviors.

Conditioned reinforcers are things that are paired with a primary reinforcer to the point where they have the same meaning as the primary reinforcer. This is done through a process known and classical or respondent conditioning.

In humans, money is a conditioned reinforcer. Money itself is simply a piece of paper -- not intrinsically rewarding. But because it has been paired with primary reinforcers such as food, clothing, and shelter, money has become a conditioned reinforcer.

Often, for pet dog owners, "good dog" or "good girl/boy" has become a conditioned reinforcer because it has been paired often with the giving of a treat. In clicker training, the clicker is a conditioned reinforcer, because it has been paired with a primary reinforcer to the point that the click means the same thing to the dog that the presentation of food would. Conditioned reinforcers are good to use in animal training because they often get a reward/reinforcement to the animal faster than you would be able to with the primary reinforcer.

The most important thing you can remember about reinforcers though, is that the animal decides what is reinforcing, not the trainer/owner! Some dogs will work only for particular types of food; others will work for tennis balls. The trainer's job is to find what the dog finds reinforcing and use it.

OK,
First, I'm not a clicker trainer as I define the frequent followers of much of the work done w/ clickers. Simple reason why is I find as a group they loose sight of the fact that the work we do with our dogs is based, or at least should be based, in drive development first. All to often I see "clicker" folk reduce the work to circus tricks. Circus tricks don't hold up under stress.
What I am doing is OC. Everything a dog learns regardless of how can fall under OC. How people have come to misinterpret OC as simply meaning baited into a behavior then rewarded for it with a toy or food baffles me.
It must be remembered that if I light up a dog with an e-collar for chasing a rabbit and it fears rabbits because rabbit chasing burns so it avoids rabbit chasing this falls perfectly into OC.
You will also find that I will intentionally avoid many of the terms used to discribe OC and CC in explaining what I'm doing with a tone or click in the context of detector work. I do this to keep it from becoming training of tricks. If I want a dog to dance in a cute outfit, I'll do clicker training and explain it through the jargon that has developed for it. But, if I want a dog to come into drive, display a behavior triggered that the dog has inherently under a limited set of circumstances, and develope those behaviors to a high degree under controlled circumstances, I avoid confusion by using more discriptive language.
In this case....the cadaver dog.... I am taking a dog that would display these hunting and prey related behaviors and focusing them on a particular set of circumstances. I'm not training the dog to do anything that isn't part of what it already is and does. I'm just enhancing it in a association with the cadaver odors then limiting it to cadaver odors. same is true with drug dogs and bomb dogs. The tone or click is a training aid only. It has a broad number of uses. It can be used to help the dog approximate the behavior in the desired situation, it can be used to sustain the dogs drive satisfaction or expectation of drive satisfaction.
Notice you ain't seen the word REWARD in my discription of the work.
I'll paraphrase a favorite author. A dog is not a whale or a porpus.
Focus on this dilema.....I don't have to teach a dog to chase a rabbit. In fact it can be hell to teach them not to. By much of the theory presented in training books and manuals you would think that we'd have to teach such a thing or it would have to be learned invoking some type of reward system. Yet, so few dogs actually catch the rabbit...and still the behavior, the more it is repeated, regardless of whether the rabbit is caught or not, becomes more and more resiliant and more intense. The behavior itself is fullfilling to the dog...it just feels good. these are the behaviors we have to seek to utilize in the training of the working dog.
If I have to teach the dog to chase it will never be a good rabbit chaser. Sure I can use a reward system to reinforce the rabbit chasing behavior and even make the dog look like a rabbit chaser. But, given some outside influence that results in stress the behavior will break down.
You might wonder what all this has to do with clicker training. Everything in the field of working dogs.

Originally posted by Jenn Kavanaugh:
You can have a 'keep going' signal, but it is NOT the clicker! Because the 'click' TERMINATES the behavior and means the reinforcer is given ...The way I am understanding this, is that the dog is not going to be learning by operant conditioning per say. I agree with Kevin...the dog IS learning thru OC; he's just not learning in accordance with the strict dogmatic practices of purist clicker trainers. We just finished having this discussion over on the ClickerSolutions list - the click is simply a tool and the trainer can use it however they choose to. It can serve as a KGS (keep going signal) or it can signal the end of the requirement. Trainer's choice. Just cuz Karen Pryor advocates only one narrow way to use a tool doesn't mean that's the only way it can be used effectively.

Jenn wrote: "The way I am understanding this, is that the dog is not going to be learning by operant conditioning per say. In as much as the dog is always working for the 'click' specifically because that's where the payoff comes in immediately after."


<img src="/ubbthreads/images/graemlins/laugh.gif" alt="" /> <img src="/ubbthreads/images/graemlins/laugh.gif" alt="" /> <img src="/ubbthreads/images/graemlins/laugh.gif" alt="" />


I love dog training.


Nice post Kevin.

. . .and Lee.

Wondered how long you could stay out of it. <img src="/ubbthreads/images/graemlins/laugh.gif" alt="" />

Wow, so interesting to get the different takes on clicker training! Do you compare the training with whales and porpoises because they are smarter than dogs, so the same training wouldn't work?

I absolutely agree that 'drive' is vital and the building of that drive. But I don't see that it excludes 'traditional' clicker training. My understanding is the click marks the specific exact and precise behavior you wanted, and then the dog get the 'reward' it is working for. Be it a toy, food, or a release to do whatever.

How does the clicker "mark" the exact behavior?

<img src="/ubbthreads/images/graemlins/laugh.gif" alt="" />

Great post Kevin, very well written.

Jenn, focus on what Kevin wrote, "The behavior itself is fullfilling to the dog...it just feels good. these are the behaviors we have to seek to utilize in the training of the working dog."

(In Kevin's rabbit example) Every second the dog is on the trail, is fullfilling the previous second. The behavior itself is reinforcing the drive-goal. Hunt/Search Drive.