Using the minuses (P- and R-)

FourIsCompany

I'm not sure what this is. Maybe someone can help me. They're trained now, but for a while, I would have them sit at the gate and wait. Then I would open the gate slowly. If anyone broke the wait (which they did), I'd quickly close the gate again. They sit and wait and I open the gate slowly, then release them. Is that -R?

I have done this very thing too.

I think this is a classic use of combined use of P- and R+. When the dog got up, you removed opportunity to go outside that time by shutting the door. Then you started over with a new opportunity, by having them sit again, and when the dog did as you asked, it was reinforced and let out.

Yes, if a dog self-corrects on a prong collar it is R-. But when the dog first hit the end of the leash where he felt the prong collar was P+. With an ecollar, when the shock is given, that is P+.

The ecollar stim is given at the same time as the command- it doesn't STOP ANY behavior, it can't be +P. The stim gets turned off when the dog complies, about as pure -R as you can get; Some people use shocks as +P, true, but I think that's a misuse of the device, except perhaps in snake-aversion training.   Prongs I think use +P, -R, and +R simultaneously, explains why they are so effective.

As to the confusion about withdrawl of attention: be careful. First, you have to be actively paying attention to the dog before you can withdraw it, so if the dog is lying on the bed and you're pottering around ignoring him, that's not -P, that's just you being busy. If you're playing with the dog, and the phone rings, and you bolt off to answer it, that can be -P to the dog and may inhibit whatever behavior he was engaged in at the moment the phone rang. Or, more common, in doggy class you are working your dog and the instructor comes over and you without warning withdraw attention to speak with the instructor, that can have negative consequences. I try to build in "on/off" switches to avoid these things- "Thanks thats enough" and the dog knows he was good but you're just done interacting with him for now.

Even -P requires a lot of skill to use and can have adverse consequences if you mis-time it or mess up.

Kim_MacMillan

I understand what you're getting at now. Firstly, I think dogs, being very intelligent animals, just know. That seems silly and highly unscientific, but what I mean by that is they probably take everything in context into a situation.

mudpuppy

The ecollar stim is given at the same time as the command- it doesn't STOP ANY behavior, it can't be +P.

Are you saying that a shock collar is not +P?  

mudpuppy

Yes, if a dog self-corrects on a prong collar it is R-. But when the dog first hit the end of the leash where he felt the prong collar was P+. With an ecollar, when the shock is given, that is P+.

The ecollar stim is given at the same time as the command- it doesn't STOP ANY behavior, it can't be +P. The stim gets turned off when the dog complies, about as pure -R as you can get; Some people use shocks as +P, true, but I think that's a misuse of the device, except perhaps in snake-aversion training.   Prongs I think use +P, -R, and +R simultaneously, explains why they are so effective.

Yes the shock gets turned off when the dog complies, which is R-. But to use administer the shock *at the same time* as the command is given is unfair ~ it doesn't provide the dog a chance to perform the behavior without aversive. It would condition the dog to the command itself as a marker for P+ much the way clicking while giving a treat conditions the dog to the clicker being a marker for R+. Strikes me as a very unpleasant way to train a dog.

I can see how a prong collar is P+ and R-. How do you get R+ from a prong collar?

corgipower

Yes the shock gets turned off when the dog complies, which is R-.

I'm sorry, but by this logic, I could start whipping my dog with a belt and then when he complied, I'd stop and I could call it -R... instead of positive punishment. I'd really appreciate an explanation on this. How is using a shock collar NOT +P?

FourIsCompany

corgipower

Yes the shock gets turned off when the dog complies, which is R-.

 

I'm sorry, but by this logic, I could start whipping my dog with a belt and then when he complied, I'd stop and I could call it -R... instead of positive punishment. I'd really appreciate an explanation on this. How is using a shock collar NOT +P?

mudpuppy

The ecollar stim is given at the same time as the command- it doesn't STOP ANY behavior, it can't be +P.

If the cue is given at the same time as the stim, I too think that's highly unfair. But then again I think the entire use of ecollars are unfair, so what I do I know? *G*

A collar pop (such as on a choke collar) is defined as P+, and often a collar pop is used for refusal to do a sit. So it follows: Cue - lack of behaviour - collar pop. It is punishing the "refusal", and at the same time is R- for sitting in the future.

So if an e-collar is only used in proofing behaviours (as I think you've mentioned before that's what it "should" be used for), then you only need to proof when the dog doesn't do as you ask (unless you are setting up the dog to experience the shock, even if the dog would do it on cue, which doesn't make any sense to me), you are punishing the refusal, correct?

mudpuppy

The stim gets turned off when the dog complies, about as pure -R as you can get;

This is true. I think we all agree on that. But the application of the stim itself is a P+ if it is used only in proofing.

mudpuppy

Some people use shocks as +P, true, but I think that's a misuse of the device, except perhaps in snake-aversion training.  

Shock is used as P+ by a very wide range of trainers, and they would all claim that it's the proper use of the device too. And some of these are some of the most well-known trainers at that. It's also used in bark collars, which is all P+ and no R-. And your example of snake-aversion teaching. And in most recall teaching with the ecollar its used specifically as a P+.

mudpuppy

Even -P requires a lot of skill to use and can have adverse consequences if you mis-time it or mess up.

I agree, any form of punishment should be carefully thought about before using it.

FourIsCompany

corgipower

I would think it would be P-. You are removing the open gate in order to decrease the likelihood of them breaking the sit.

 

It might be a case where I'm using several quadrants in conjunction with each other, because there are actually several actions being performed in a series. I reward them for sitting and waiting by opening the gate (+R). But they have to continue to wait to be released. And if they don't, I close the gate (+P)

Or I add the open gate (as a reward) to increase the likelihood of them sitting and waiting. (+R)
Then I add the closed gate to decrease the likelihood of them breaking the sit. (+P) 
Or I remove the open gate to increase the likelihood of them sitting and waiting. (-R)
Or I remove the open gate to decrease the likelihood of them breaking the sit. (-P)

This is where I get confused on which quadrant I'm using.
Do I ask, "What is the behavior I want"? Answer: Sit and wait.
Or do I ask, "What is the behavior I don't want"? Answer: Breaking the wait.

I think there are several events chained together... Closing the gate could actually be an NRM, couldn't it? Saying, "I don't want you to get up just now, but that's an acceptable behavior at another time..." 

I have never thought to delve into the "minuses" this much but I hope to learn more.  

I think you're right... it ISN'T just one quadrant.  I think of it this way - you usually use them in pairs, like a seesaw.  Most of the time, I think most of us use R+ and P-. In your example with Jaia, you withdraw your attention (P-) when he is being "rude" and you reward him with attention when he is calm and polite (R+).

I really like your gate example.  I do a similar thing at the door with our dogs.  It isn't really the gate or door which is the reward for the dogs is it?  It is a degree of freedom - the chance to run and play outdoors. THAT is what they want and that is what is being withdrawn when the gate (or door) closes.  Therefore, I see the gate-closing as P-, because you are withdrawing the reward (the play outdoors). The reward (R+) for holding the sit is being released out of the door, nit just the door being open!  That's the way I see it anyway.

corgipower

I never said the shock wasn't P+. I, as well as you, have both inquired as to how mudpuppy decided that it isn't.

I know, I'm sorry. I shouldn't have used your quote. I was actually agreeing with you. I meant by the logic mudpuppy seems to be using.

FourIsCompany

This is where I get confused on which quadrant I'm using.
Do I ask, "What is the behavior I want"? Answer: Sit and wait.
Or do I ask, "What is the behavior I don't want"? Answer: Breaking the wait.

A brilliantly honest paragraph. And it marks, I think, a difference in outlook or philosphy, even. I ask myself the first question more often than I do the second question. But I also think the second question leads to the first question. "If this is what I don't want, then what is it that I do want?"

 If, instead of thinking in terms of positive or negative, you start thinking of either taking something away (-) or adding something (+), it gets easier to differentiate which quadrant.  A reinforcer is merely something that keeps a behavior happening.  A punishment is something that stops behavior from occurring. 

Brief explanation here: http://www.clickersolutions.com/articles/2001/ocguide.htm 

well, let's discuss ecollars again. Some people use them as a punishment- dog does something they don't like, or dog refuses a command, they zap dog. Dog stops doing that. This has been shown to cause serious adverse consequences to doggy mental health, even in very hard dogs such as protection-trained dogs.

Other people use them as "interruptors", i.e. an intrusive weird sensation to break up the dog's train of thought followed by attention being re-directed to the handler- in fact, you can readily teach a dog to look at you whenever he feels a low-level stim, using the stim as a cue. This is probably one of the more common uses when people only want to use the ecollar for recall training. I have trouble believing it is viewed as a punishment by the dog, since no behaviors are reduced in frequency; instead, the behavior of paying attention to the handler is increased in frequency. Interruptor + negative reinforcement?

I think using an ecollar as +P is kind of silly, they work nicely as negative reinforcement. The ecollar is set very low, such that it is not painful to the dog, just feels mildly unpleasant. Feels like a bug crawling on your skin (to me, anyway). The second the dog does what you want, the stim goes off.  It's not punishment, since no behavior is reduced in frequency in future, no behavior is stopped. It's pure negative reinforcement: the dog complies, and an aversive is removed from the dog. The dog is more likely to perform the behavior again in future. At the same time most folks also use +R to reward the dog for compliance. Most people use the ecollar during proofing of commands; but some people actually use ecollars from day one to teach commands to puppies. It works almost as fast as using a clicker- it's sort of the converse of the clicker, which clicks ON when the dog gets it right, whereas the ecollar clicks OFF when the dog gets it right. And just like the clicker, once the dog grasps the concept you stop using the ecollar. Unlike punishment-trained dogs, dogs taught with ecollars don't seem to "shut down", and seem willing to experiment and offer behaviors. We have a local training center that teaches basic obedience this way, and they produce lovely well-mannered dogs. They refer these dogs to our center if the owners want to pursue sports or higher levels of obedience, and these dogs, unlike punishment-trained dogs, rapidly learn how to "play the clicker" and offer behaviors. The only adverse consequence we've observed is many of these dogs take a little while to develop drive.

mudpuppy

well, let's discuss ecollars again.

Can you give an example of a behavior you might use an ecollar for and detail the sequence of events?

mudpuppy

The ecollar is set very low, such that it is not painful to the dog, just feels mildly unpleasant. Feels like a bug crawling on your skin (to me, anyway). The second the dog does what you want, the stim goes off. 

So... you sit the dog down for a training session and turn the collar on (delivering the stim)? And when he looks at you (or does the behavior you're looking for), you turn it off?

mudpuppy

It works almost as fast as using a clicker- it's sort of the converse of the clicker, which clicks ON when the dog gets it right, whereas the ecollar clicks OFF when the dog gets it right. A

So... instead of silence (as is the default with clicker training) you turn the stim on and wait for the dog to do what you want and then turn it off?