Operant Conditioning vs. Pavlovian Conditioning

This is a discussion thread · 19 replies
1 2 3
Leah:
I just started Jean Donaldson's "Dogs Are From Neptune," and was a bit surprised at what I read. I think I remember Marshall saying he was going to take a break from the ng, but I hope he's here to comment on this.

For dogs with generalized fears, i.e., of men, she recommends Pavlovian (classical?) conditioning. When men appear, the treats and praise come out - no matter what behavior the dog is displaying. Men = open bar. No men = bar closes. Even if the dog is displaying aggression.

I normally have the treats and praise stop when the dog acts out, and instead try to refocus the dog on me.
Can we start a discussion on this? I'm very interested to hear what you guys think. I'm sure the choice between methods must depend on the individual dog and the individual circumstances, but I'd like to hear the pros and cons of both methods.
There is one dog in a basic class, a young Aussie named Mocha, who occasionally acts out at dogs near her. I've not been able to pinpoint the trigger (she's only been to two classes so far), because otherwise she displays friendly, though wary, interest. She's 6 months old and was not socialized. I can refocus her on a dime by simply saying her name in a high, cheery voice, and then asking her for a behavior. I'll have to watch her more carefully - my guess is that she may be acting out whenever food appears. If so, maybe dog close by should = open bar. But I also don't want to reinforce the acting out.

This is a tough one, because there are 10 dogs in the class. I can't always keep my eye on one.
Bubbles, the shih tzu, is the only small dog in this class. Last week she was very frightened of all the dogs and people, and we couldn't even get a sit out of her. I asked mom to bring her to puppy play-time, and she did. By the end of the session, she was playing with the other dogs and visiting the people (she was fearful of strangers prior).
This week, with the confidence she gained in puppy play, she won the quickest sit contest! :}
PetsMart Pet Trainer
My Kids, My Students, My Life:
http://hometown.aol.com/dfrntdrums/myhomepage/index.html Last updated June 27 at 10:00 a.m.
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
shelly:
[nq:1]For dogs with generalized fears, i.e., of men, she recommends Pavlovian (classical?) conditioning. When men appear, the treats and praise ... on the individual dog and the individual circumstances, but I'd like to hear the pros and cons of both methods.[/nq]
this was described in pretty good detail by Lynn K. awhile back (the thread was "Prong Collar Review"). i came away from the discussion with the understanding that behavior modification, which is what i think you're describing, is a matter of changing the dog's hard-wiring. it's not a matter of rewarding the dog for a certain behavior, but a matter of building new associations. Melanie has described the process in pretty thorough detail on numerous occasions, but unfortunately doesn't archive her posts.

shelly (perfectly foul wench) and elliott and harriet http://home.bluemarble.net/~scouvrette
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
Melanie L Chang:
>I normally have the treats and praise stop when the dog acts >out, and instead try to refocus the dog on me.
One of the cardinal rules of behavior mod is that if the dog starts acting out, you're going too far too fast and need to back up.

You want the dog to be aware of the problem stimuli, but not upset enough to be engaging in the undesirable behavior. If the dog's started acting out, it's time to write that situation off as a lost educational opportunity and just end the interaction.
When people don't do this, and push too fast, is when they end up inadvertently reinforcing the behavior they're trying to get rid of.

Melanie Lee Chang > Form ever follows function. Departments of Anthropology and Biology >
University of Pennsylvania > Louis Sullivan (Email Removed) >
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
C. L.:
Well in reply to the subject and not neccessarily the content, Classical Conditioning, that which Pavlov developed, the behavior is not voluntary. Drooling over food is not something the dog can control. Skinner developed operant conditioning, which is basically the same idea but now with voluntary behaviors.
That has nothing to do with what you asked, but since I have no good advice on it I won't butt in. Emotion: smile
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
Tricia9999:
[nq:1]When men appear, the treats and praise come out - no matter what behavior the dog is displaying. Men = open bar. No men = bar closes. Even if the dog is displaying aggression.[/nq]
I sort of doubt this. If you know the dog is aggressive when the man is within
5 feet, you work on 10, 8, 6 feet for a long time - you wait until the dog istruly happy to see the guy at 6 feet before you move to 5 feet. After all, if you wait til the dog displays aggression, you won't have a dog partaking of the bar - too much stress and adrenalin to care about the bar. You've pushed it too far.
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
gryghost:
I went to an aggression seminar put on by Nina Bonderenko of Australia, where
she used treats to "interrupt" the behavior which then allowed her to praise and reward.
In one scenario there was a very large Great Dane/Rotti cross that was lunging, snarling, baring
teeth etc, when someone came into a room. His owner had a pub so this was a problem. They tied
the dog to a wall in a training room and then the trainer would come into the room, dog would
lunge, act nasty and then she would throw a handful of high value treats (beef, chicken etc.), at the dog's
feet and then she'd leave the room. The look on the dog's face was priceless, "what the...?" and he'd eat
the treats. Again she came in, dog lunged, treats thrown, dog perplexed, eats treats. By the third entrance
of the trainer he only woofed. Fourth entrance he woofed with a slight, tentative wag of the tail. Fifth
entrance, dog was lying down wagging tail. Even though you don't see it in her video, she enters
the room each time walking very fast, directly to the dog, eye contact and so on, she wanted the dog
to react to somewhat threatening body language. When the dog settled down then they would repeat the
scenario with Halloween masks on. It was amazing and since then I've used the methods in our
kennel with aggressive dogs including one that came in for training that had bit three people and did
not like women. I gave the dog one day to get used to the kennel and after six or seven interruptions the dog
would wag it's tail and was quiet instead of biting the fence and launching himself at me.
She also showed how she trains for dogs who act aggressive when they see someone with or without a
dog approaching. It was great.
Nina writes articles for Teaching Dogs magazine, their website is www.teachingdogs.com .
If you can get to a seminar when she comes to the states, I'd highly recommend her.
Dawn
"> For dogs with generalized fears, i.e., of men, she recommends Pavlovian
[nq:1](classical?) conditioning. When men appear, the treats and praise comeout - no matter what behavior the dog is displaying. Men ... normally have the treats and praise stop when the dog acts out, andinstead try to refocus the dog on me.[/nq]
= Posted via Newsfeeds.Com, Uncensored Usenet News = http://www.newsfeeds.com - The #1 Newsgroup Service in the World! == Over 100,000 Newsgroups - 19 Different Servers! =
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
Helle H:
[nq:1]When men appear, the treats and praise come out - no matter what behavior the dog is displaying. Men = open bar. No men = bar closes. Even if the dog is displaying aggression.[/nq]I've used a similar approach successfully on many occations, especially when there's barking or growling involved. It's done by dropping a lot of treats on the ground so that the dog never runs out of them. If the dog eats the treats, the approach may work. Dog eats, and then as he relaxes and gets more interested in the treats than in the man he was growling at (or whatever/whoever he is fearful of), you let him finish eating what's left on the ground, and look your way for more treats.

This is where you can start expecting the dog to offer behaviors before giving him the treat. In the beginning you don't expect much - maybe nothing more than him being silent. Then you can wait for him to offer a bit more; maybe that he sits down, glances over to the man, take a step toward the man, etc.

At this point I'd reinforce any behavior he offers that isn't aggressive behavior, and I would use no signals or luring at all. My only actions are to click/praise and give him the treats.

The next step is for the dog to do gradually more for a treat - to make it gradually harder. He both needs to learn to approach the man, but also to be able to handle that the man moves around, first a little, then more, look at your dog, takes a step towards your dog, two steps, etc.. Then in different places, different situations, etc. (You will probably need to start the first sessions from scratch, but it doesn't take long before the dog will offer behaviors other than to growl or bark)
I wouldn't worry that the treats are reinforcing the aggressive behavior. In the beginning of the session I outlined, the treats aren't reinforcing anything other than him sniffing the ground and shutting up in order to find them. Then he starts offering other behaviors in order to get the treats from you. The man backing up and away from the dog. or you taking the dog away from the man, will most likely reinforce the aggressive behavior, though.
The only cons I can think of with this method is that you will need a patient helper, an owner who realizes how important it is to use enough treats so that the dog never runs out of them during that first phase, and most importantly; the dog cannot be too stressed out to eat! If the dog doesn't eat, this method is meaningless, unless of course he will play instead.
I'm not completely sure when I'd use the "open bar" method rather than start out from a distance where the dog doesn't act out and work our way gradually closer as the dog deals with the present distance. Obviously, when there's no way you can get far enough away is typically when "open bar" is used, and also when the dog is freaking out from whatever distance you start at. And also when I want the dog to become confident enough to be rewarded by the helper(s) rather than me, for instance in SAR(?) training.
Helle
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
Lynn K.:
[nq:1]Well in reply to the subject and not neccessarily the content, Classical Conditioning, that which Pavlov developed, the behavior is ... something the dog can control. Skinner developed operant conditioning, which is basically the same idea but now with voluntary behaviors.[/nq]
Actually it's very much to the point, and important in determining which approach to use and when. Leah, you posted last week about frightened dogs starting to socialize after a few classes. What you are seeing is classical conditioning - the dogs become conditioned to relaxing in the class situation.
Leah, you need to step back and look at every dog in terms of what you want to see happen. If you want to teach the dog to do a behavior, you use OC. If you want to change how the dog feels (perceives something) you use CC.
FWIW, I disagree with cookie pushing in a CC situation unless time is very short because you may unwittingly set off another trigger. It's more important that nothing bad happens and the general environment is pleasant for the dog to form new associations.
Lynn K.
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
Susan Fraser:
The "open bar" is something I read about on Shirley Chong's clicktrain list quite often. Since I have golden retrievers who have never know a stranger (but licked quite a few ;-) and since there are a few bazillion posts to that list every week, I don't read all of them.
But when Shammie had puppies (almost 12 weeks ago) and SheBop stole one out of the whelping box before they were even all born yet, AND since SheBop has never beenknown to be very tolerant of small puppies, I decided it was time to do a little counter-conditioning regarding her attitude about puppies.

Shammie helped, alerting when SheBop was close enough, and at whatever imaginary line Shammie drew in the sand, I would continually pet and scritch and love on SheBop while she watched the puppies.
At first, Shammie would let her stand in the doorway - scritch scritch. Then a couple days later Shammie let her come within a few feet of the box - ear scrumples. Lean over the edge of the box and sniff - shoulder massage.

By the time the pups were old enough to run in the yard, both Shammie and I totally trusted SheBop with any number of the aggrivating little pupsters.

It's coming back to bite me in the butt, though, now that I need an assertive *** to teach the little voodoo doll we kept from that litter some manners!
Susan Fraser, owned and trained by
BeBop a Lu SheBop SH, Shamma Lamma Ding Dong MH,
Semper Choo Choo Ch'Boogie, and Gris Gris Gumbo Ya Ya http://mypeoplepc.com/members/chinchuba/AuH2OK9s /
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
Show more