You Don't Understand +R

You simply don't understand the method, if you even have to ask how it can solve aggression.

Yes, we're so dense.[] (note the emoticon, denoting humor) That line reminded me of the line "how much is a Ferrari?" "If you have to ask, you can't afford it."

Anyway, in my limited understanding (no degree, no cert from an institute or accredited school) I totally see the value in having a redirection command to divert aggression, if possible, if you are not close enough to physically control the dog. One thing I have learned recently is to re-direct before a behavior escalates. Plenty of times, I can catch Shadow on the first flinch. Another time, I caught him raising his tail and focusing his ears while looking at the cat and re-directed before he would launch. I learned to watch for those signals and re-direct by watching he who shall remain nameless. I learned not to fight Nature but to work with it. Now, would  a re-direction before launch be considere +R or +P? I'm not sure it's either. It is increased awareness and activity on my part, with no change in my dog. I have a better understanding of his environment and seek to control that, which controls, to some extent, his instinctual reactions. I do agree that aggression is simply another behavior and all dogs have this behavior which is why BSL is so stupid.

The exception I would draw or the clarification I would ask for is that, if in simply asking how +R would treat aggression to a managable state denotes our ignorance and ability to misunderstand or not understand at all, then that is limiting the access of +R to people who have aggressive dogs. If we have to ask, then we don't understand? Then we shall not understand and shall be penalized (inferred as ignorant) for asking. When I was growing up, there were times I was punished without being spanked. I was merely informed for hours on end that I was ignorant ("because I said so. It will be my way and there is no other way") and could aspire to one day be the dirt beneath a worm. That was better? There were times I would have preferred a short spanking. I have a high tolerance for pain. I had to learn to holler so that the spanking would stop. Get it done, be over with, and we can move on.

So, can there be examples of how to divert aggression with +R as well as how to stop aggression with +R in the midst of battle? I know you mentioned DirectStop. It's not been as easy to find around here as I thought it would.
 
eta: To me, using DirectStop would be +P. A physical deterrent, though it did not involve putting a hand on the dog to stop the fight. That comes after the dogs are puking and crying and, while they are distracted from the pain of the DirectStop, you re-leash them and separate. Also, what are the limits of DirectStop? There are some dogs that, in full charge, may not stop, just as there are dogs that will charge past an invisible fence. Some dogs, such as the Great Pyrenees have an extra layer of hide and fur around the neck. This permits them to stomp a coyote to bits even while the coyote has fangs stuck in their neck.

Aggressive behavior is just behavior, thus can be modified.

Quite true. I am using the clicker to modify "guarding" behavior in my rescue dog. But I also had to use physical force when it first appeared. She had never shown food aggression when we first got her, she would even leave her dish when our elderly Dalmation would show interest in her kibble. But I discovered the hard way that she would go after any dogs that showed any interest in my food or my children's food. She went after my brother's Golden when the Golden was too close to the kids during snack time, and she has attacked the same Dal that she will allow to eat her own food without a fuss. Both times, I grabbed her collar and pulled her away. I didn't "alpha roll" her, but she did roll over submissively when I put her in a down stay.

Now, I am using the clicker to train her to stay in her "place" while anyone is eating so the problem can be prevented from occuring. My relationship with her is not "damaged" from the physical correction of pulling her off the other dogs any more than the relationship she has the with Dal and Golden that she physically attacked. They are still play buddies. After the incidents, she came to me with that "I'm sorry" submissive look and licked my hand, just like the other dogs did to her after her attack.

I DO understand positive reinforcement and operant conditioning and use it to prevent a reoccurrance of the behavior, but if it does happen again I won't hesitate to do the same thing and physically pull her away then use a down stay to put her in a submissive position. I won't stand there and watch her put another dog down on the ground until she's ready to go to her place so I can click her. I do believe in ignoring the bad behaviors, but some behaviors do need to be corrected THEN the correct behavior needs to reinforced.

I'm sorry" submissive look and licked my hand, just like the other dogs did to her after her attack.

Actually, that submissive look and lick is appeasement behavior. More like "Please, don't hurt me." It helps keep a dog from getting attacked twice.

I perfectly understand +R, but dogs do learn from many methods. Positive reinforcement is a reward, be it food, attention, toy, click, etc. for performing a response.
 
Positive reinforcement is NOT training in itself, but a tool that is used in different conditioning protocols.
Classical conditioning for instance, uses an unconditioned stimulus paired with a conditioned stimulus to produce a response. Pavlov demonstrated this type of learning with a dog that was shown meat, and when paired with the ring of a bell, eventually the dog salivated only the bell was rung. The sight of meat was not needed.
Operant conditioning, which is also called instrumental learning, was first demonstrated by Thorndike. He placed a cat in a slatted box with food on the outside and the cat had to "learn" how he could get out to the food. Finally he used the instument (a latch string) to open the box door and was able to get to the food. Each time the cat was placed in the box, it took less and less time to open the door, until the cat finally opened it almost immediately when placed in the box. Skinner also demonstrated it with his "Box".
 
Another form of learning is called "Shaping" and this is where most of the +R comes in. We "reward" in baby steps at first until the animal is doing the exact behavior that is desired. By shaping, we are giving them the idea of want and we can then expect a bit more each time the grasp the new desired behavior. For example, a heal. You "reward" for a dog walking on a leash, then when it walks 1 yard from your side, then one foot, then right at your side.
 
Habituation is basically getting an animal "used to" a certain situation or circumstance. This should not be confused with "socialization" as they are different.
 
There are other forms of learning that could come into play, but these are the most widely encountered forms of learning in dogs (with maybe the exception or learned helplessness).
 

Positive reenforcement can ONLY be defined by the impact on behavior.  So if you do something people term "positive" and the behavior does not increase in likelihood of occurance, then positive reenforcement has NOT occurred.  If you do something that people term "negative" and the behavior increases, positive reenforcement has occurred.  Reenforcement means behavior likelihood increases whether it is positive (addition of something) or negative (removal of an aversive).  Punishment in the introduction of something that immediately stops the behavior.  It is all in the timing and the impact on behavior that determines how to label the  consequence.  So if you are trying to establish a new behavior for aggression, and the aggression increases; the aggression has been positively reenforced.  The human value statements on the consequence have no real impact (i.e.  treats are not always positive reenforcement)

What I have learned from operant conditioning is that it is self-rewarding mechanism, such as the cat in the box. The cat learns a behavior and is rewarded. Subsequent tests with the cat in the box show a quicker resolution time. That is, the animal learned from the first time and repeats its successful actions to repeat the reward. Which is certainly different from Pavlov's response, which can be found in anything from cows to humans. Human example, lunch trucks have a horn of a certain pitch and timbre. Some equipment cranes have that same horn. So, it is not uncommon to find yourself looking for the lunch truck when the crane operator toots his horn. The crane driver is amused. So, we're exhibiting pavlovian response and the crane operator is in operant conditioning. He is amused, which is a reward. Toots the horn again for another reward.

People have likened getting a paycheck from work as operant conditioning. I disagree to an extent. Work is a contract. You sell your time and skills for a compensation. That is free trade, or representative barter. But a man will work to a frazzle to provide the things a woman may want if it gets him the chance at enjoying her affections. That's operant conditioning.

At least that's my two cents. Now, how can we use operant conditioning to stop aggression? Rewarding the response to a diversion before escalation would be good. But how would we use operant conditioning in the midst of a full-on fight? That is, how does the dog reward itself for stopping the fight? Is there some method or safe way to simulate a fight and provide the dog with a chance to reward himself out of it? When dogs fight, it is instinctual and based on survival, which might be operant conditioning. The dog still standing gets to live.
 
Let's say we break up the fight by pulling on leashes and separating the two. That might be considered +P but it can be a lifesaver. So, how would one use operant conditioning to get that dog not to fight, especially dogs of a certain color? It imght fight them in a pavlovian response to a bad scene with a previous dog of the same color. So, How would we use operant conditioning to overcome a pavlovian response?
 
When I took Shadow to the vet for a vacc update, he did not want to go in the clinic. I had treats to offer him. He refused, time and again. So much for +R solves all. So, the vet came outside with me. I put him in the truck and then realized that won't work because he is defensive in the truck. So, I took him out of there and we wound up by the cattle gate where they bring in cows and horses. Was Shadow in charge or did I lead him to a comfortable environment? He was comfortable enough there to accept treats from the vet. The shots didn't bother him at all but the other scenes did. So, in the end, I win. I got his shots updated, in spite of his fear of the clinic.

here we go, here's the main question, and it's more of a semantics question than an actual training question. When we discuss +R within this thread...are we talking about the training philosophy i.e. clicker training, operant conditioning or whatever other name it has, or, are we talking about using +R exclusively and no -R, -P and +P?

The way I understand it from a non-pro point of view, and I've said this before, is that EVEN purely possitive methods use -R, and some -P, just no +P. And it is the absense of +P that makes a method purely possitive, again, context, here possitive means "good", not "add".

So, does using a so called "purely possitve method", I guess it depends on the dog, but GENERALLY speaking, I don't see why it wouldn't, just like any other thing, it's probably not a "one size fits all" solution. Does using +R (in the operant conditionin sense of the word) ONLY work? I highly doubt it, mostly because even if you're ignoring the dog for not giving you the correct behavior you're adding -P into the mix (correct me if i'm wrong on the ignoring issue).

I guess what's important at the end of the day s that you use whatever method works for you and your dog, always starting with the "mildest" approach, if that doesn't work, try something else, but the key is to keep an open mind, mostly to thing you don't understand.

He who's name we dare not speak uses operant conditioning. He brings a dog into his gargantuan pack. The dog adjusts its behavior mostly on its own to enjoy the reward of belonging to the pack. +R, in my limited understanding is part of classical conditioning. You reward the behavior you want. -R is removing or withholding the reward in the presence of wrong behavior. Some dogs do not respond to this. They don't give a flying flip if you turn their back, their in charge and they'll do what dogs in charge do. A correction is given, often in a physical form, though not necessarily painful, to curb or stop a behavior. A dog misbehaving off leash is re-leashed and loses his "freedom." I'm not sure of -P, if there is a thing, unless it's euthanasia, which ends all punishment.
 
To me, in my humble ditch digger's opinion, clicker training is fairly classical and quite pavlovian. Pavlov used a bell. People today use a clicker. The dog identfies the sound with something good. So, hit that clicker and watch your dog salivate. I haven't studied the clicker method but I would imagine that you need to reinforce it once in a while with an actual treat so that the dog still identifies it with something yummy.

Hey, my post had the desired effect!!!! Now, we are actually discussing "positive" methods on an intellectual level, not an emotional one.  Great explanations from kennel keeper and mrv. 

Ron, you poor thing.  I wish I could give you your childhood back.  Instead of being willing to trade the tirade for a spank, I wish you could have enjoyed the *encouragement* when you did something right!  If we "catch our dogs doing something right" and reward behavior, it tends to be repeated.  Behavior that is ignored tends to extinguish. Unfortunately, it is possible, and quite likely for newbie owners to accidentally reward unwanted behavior from their dogs.  This is what happens every time you pat a dog that's jumping up on you, or every time you glance at the dog who is whining in its crate.  And, to make matters worse, if you do it intermittently, you make the unwanted behavior even stronger - with variable reinforcement.
Ron, when we use clickers, we do give a treat.  But, the clicker is used when the dog is first learning a new behavior.  We reward often, but later, once the dog understands it, and we have added a cue that the dog responds to (like "lie down"), we begin to space out the rewards (variable reinforcement), which strengthens that behavior.
The clicker just takes the place of saying "yes" or "good boy", and just as those words might predict that the reward is coming, so does the click. 

What I really wanted from this thread was to emphasize the point that what is commonly referred to as +R as a method is more complex than click/treat.  Food is *not* always a reward - a reward is something the dog is willing to work for. 

Also, with regard to "guardy" dogs, I would rather feed my dogs separately than have to damage my relationship with them by physically grabbing at them.  IMO, it's better to teach a solid "leave it".  If my Aussie mix can step over a pile of liver treats on the floor when I say "leave it", so can your dogs be taught to leave a sib's food bowl alone.  But, what's the point?  Just my preference, but I'm not one who insists on making my dogs bear the discomfort that most dogs feel when made to eat in close proximity to others - I let them relax and eat in peace in their own areas.  I *do* insist that my dogs not guard from humans, but that is not accomplished with force either.  In fact, it's been my experience that the more you snatch things away from the dog, or yank him away from the forbidden item, the guardier some dogs get.  Very dangerous, depending on the dog, and not something to be routinely attempted.  I use Jean Donaldson's protocol with great success, and several of my students have done so with some extremely guardy dogs and have had similar success. 

It's a bit dry, but I would suggest that any of you who would like to understand clicker training, get a copy of Karen Pryor's book "Don't Shoot the Dog". 

ORIGINAL: ron2

+R, in my limited understanding is part of classical conditioning. You reward the behavior you want. -R is removing or withholding the reward in the presence of wrong behavior. Some dogs do not respond to this.

To me, in my humble ditch digger's opinion, clicker training is fairly classical and quite pavlovian. Pavlov used a bell. People today use a clicker. The dog identfies the sound with something good. So, hit that clicker and watch your dog salivate. I haven't studied the clicker method but I would imagine that you need to reinforce it once in a while with an actual treat so that the dog still identifies it with something yummy.

 
 
The core difference between operant and classical conditioning is that in operant conditioning, the consequence (e.g. +R)  is contingent (dependent) on behaviour. e.g. sit and get a bit of food, don't sit and don't get a bit of food. With classical conditioning, a stimulus and response are merely associated (usually by occuring at the same time) e.g. with pavlov, ring a bell and present food, dog salivates (an automatic response/reflex) - do this enough times and when you ring a bell (and don't present food), the dog will still salivate. Salivation is the conditioned response, when it occurs after the bell is rung.
 
Clicker training is similar to classical conditioning, in that the click acts as a discriminative stimulus (as in it signals reward), when stimuli are paired with rewards often enough they can often assume reinforcing properties, but this doesn't last very long when it's just the click alone (that's why you need reinforce it occasionally). So a clicker can be considered similar to Pavlov's bell to the extent that it can act a conditioned reinforcer (and thus elicit behaviour).

Nice post, vinia.  I am even more encouraged!!! 

Ron, I forgot to answer you about the Direct Stop.  It's a punisher, but better than sticking your hand between two fighting dogs, or getting attacked by an off leash dog while on a walk.  I can't get it around here, either, so I order online.  Just Google it, you'll find some places to compare prices.

For the lurkers:
-P    negative punisher (operant response followed by removal of an
appetitive stimulus)
+P   positive punisher  (OR followed by an aversive                        stimulus)
- R   negative reinforcer (OR followed by removal of an aversive) strengthens behavior
+R   positive reinforcer  (OR followed by presentation of an appetitive stimulus) strengthens behavior

Premack principle - A high frequency behavior can be used to increase a low frequency behavior.  Let's say you like to watch football on TV (high frequency behavior).  Your spouse tells you if you mow the lawn (low frequency), she will can the tearjerker movie and let you watch the Patriots.  So, you mow the lawn. 

Sometimes the way you put things...lurkers?  Thanks it is hard to follow the dialogue and learn with all the abbreviations.  This helped alot.

Thanks, Anne []

I kind of feel that all this +R -P stuff sounds unnecessarily complicated. Why should we put fancy names to things that sound scientific when we're trying to get the laymen on board? Quite frankly, as soon as I see math-related symbols, my brain curls into a fetal position saying "You said we wouldn't have to do these anymore!". It doesn't help when it's followed with a letter that looks suspiciously like algebra or some kind of spooky equation that describes something I'm not going to understand very well. You might say I've been conditioned through positive punishment to loathe math and anything that looks vaguely like it. [] But even if you take away all the symbols, it's STILL described in an ugly manner. Too much jargon, IMO. I've learnt from talking about a very specialised field of zoology that it's best to avoid jargon wherever possible, and if you have to use it, clearly define it first. Negative punisher, eh? Right, so that involves both an operant response and an appetitive stimulus? What the...? I had to look all those things up because it wasn't intuitive. It irritates me when people, scientists included, invent jargon to describe something that's reasonably simple. Jargon creates chasms between people in the field and people not in the field. And I suspect it often serves to make people in the field feel like they're a lot smarter than they are. It's like people that use big words when a small word would suffice. Having to think hard about what someone is saying is bad communication! Science desperately needs good communication. [sm=2cents.gif]

Okay, now that I've got that out of my system feel free to totally ignore me. I understand why jargon is useful, but that doesn't make me like it. And complicated jargon for companion animals seems a bit snooty to me. Way to alienate the majority of the population, I think. But I'll get off my soapbox. At least there's some kind of understanding of what we're doing when we train our dog and why it works.

Incidentally, I feel it's important to point out that behaviour is not just behaviour. It could and probable is influenced by genetics, which would make it genetics as much as behaviour. Genetics you can't change. If you have a dog that was born freaking aggressive, it's very hard to modify its aggressive behaviour. We had a dog briefly who we couldn't manage because he was aggressive in every way imaginable. The behaviourist that came to evaluate him told us that if he were a larger dog, he'd be recommending having the dog put down. We met his mother, who was delightfully friendly. We met his father, who was a bundle of joy. We brought him up with heaps of socialisation and positive training methods. What happened? He wasn't 6 months old before we had major problems and he'd bitten people. I agree that we could have worked on him and maybe we could have modified his behaviour somewhat, but where do you start when you've already done everything right? In the end, he went back to the breeder and I believe he was placed in a home with someone who had no kids or other animals and was comfortable dealing with his aggression. You can only work with what you've got.

I also agree that most people with seriously aggressive dogs don't have the means or knowledge to help it. We certainly didn't with ours.

I don't really hold with Direct Stop. We don't even have it over here, but I couldn't see myself using it any more than I could see myself using an e-collar or kicking a dog in the nose. The times when I've been charged by an aggressive dog I've been able to save myself and my dog with some quick thinking and instinctive body language. I think it would be too easy to use it when you didn't have to.