This blog posting is for our colleagues:
There’s a big issue when positive trainers claim they will never use P+ or R- in training an animal, because you can have every intention in the world to attempt R+ or P- but the dog decides what stimulus is the most salient and whether that stimulus is reinforcing or punishing. I often use the example of how my wife thinks her dinners are R+, but really, I’d rather go to the taco truck. You will go to condition a clicker with a dog and they will begin to respond with fear because it turns out they had sound sensitivity. You will give a dog a meatball and the dog will spit it out and quit working for food because it tasted bad to them. Traditional trainers often knee bump dogs for jumping, some dogs will start jumping more, other dogs will say “knee bumps suck, I’m now scared of you.” So when we get caught up in only discussing the theory of quadrants and don’t examine other factors there is very little room to grow. Classic desensitization is all R-, but there are varying levels of training with regards to the thresholds that make the nuances. In CAT, a dog is brought over reactivity threshold (barking, lunging, crying, etc.) and held there until they stop reacting. This is like swinging a sledgehammer in a china shop. In “BAT” a dog is often brought over their stress threshold (lip licking, yawning, etc.). This is dramatically “nicer” to the dog being trained than with CAT. So to say there is no difference between the two because they are both R-, you have completely lost the forest through the tree. DS/CC is the overarching idea of all these things. DS/CC methods have been evolving for a long time, but there isn’t one method to it, it’s a concept. CAT has been used successfully and those dogs have been desensitized. It was a method developed through scientific research and so there is a lot of data on it. Now you would be hard-pressed to find anyone with a scientific understanding of animal behavior who thinks CAT is humane. The reason though is not because of the presence of R-, because there is R- in DS/CC, it’s because it is an unnecessary and potentially painful amount of stress you have to put on the animal (which is why we call it flooding). DS/CC done correctly still brings a dog within a threshold that they can see their trigger, this immediately activates the amygdala, and so agonistic behavior is already being primed, however ideally they are not displaying any compounding displacement signals (stress signals: yawning, lip licking, etc.) But in order to remove the learned association of that trigger, the dog has to see the trigger and the act of seeing a scary spider starts priming the system. Hebbs law: Neurons that fire together, wire together. Now every dog is going to be different, and some dogs are going to have fear that is of a PTSD magnitude (in which case fluoxetine or some kind of SSRI needs to be introduced to help chemically induce a sensitive period of learning to socialize them to the trigger). BAT receives a lot of criticism which I find to be extremely unfair. BAT may have some flaws. That’s fine, but I’d argue it’s not really for professionals. It’s for people who don’t understand how fear and reactivity works in the brain and they need strict guidelines. I’d much rather have people attempting BAT than CAT, the fact that they are trying anything though is a major positive.
Any system of training has flaws, because the science of training has far too many variables for any system to be complete. The lower we keep stress during learning, the higher the cognitive function of the animal (this is seen across the board in numerous species from people to non-human primates to birds, to dogs, to marine wildlife, etc., high levels of stress inhibits cognitive function and learning). When we are using food, we are plastering a dog’s brain with dopamine, which is building pathways in the medial forebrain bundle (often called the “gambling center”) which is the axis of motivation in animals. The nucleus accumbens is highly sensitive to dopamine, and tells the forebrain to do x,y, or z so that it can receive its dose. Food, sex, water, shelter: anything that is related to fitness or well-being of the animal is going to stimulate the bundle of dopaminergic neurons. But to start drawing lines in the sand about the quadrants is the wrong reason to draw the lines, they are theoretical aspects of learning that every trainer must understand but in my personal opinion I much prefer to analyze training from the type of techniques used: direct confrontation, indirect confrontation, non-confrontation, then watch to see how much stress builds in the animal. If stress is minimal, perfect! Now analyze results, was there a learning event? And if so, did the behavior get reinforced or punished? After that we are not always going to be able to pin-point what was the exact thing that caused the learning event. Was the time-out for counter surfing or mouthing effective because of the removal of me or because the dog really doesn’t like the crate or bathroom? Even Premack argued that reinforcement hierarchies are on a spectrum, and so a massage at one point versus another could actually be punishing. Perhaps the dog has learned so well the association between the word “time-out” and removal that it has a negative association, therefore just saying the word might be punishment in of itself, which would be P+ if the word has a negative CER. There’s a billion neurons bouncing around their little lemon and I refuse to jump in and make declarations of how a dog is interpreting a learning event “exactly”. Anyone who doesn’t know the quadrants I most certainly do not take seriously, but there are other factors involved in behavior modification that go beyond the factors, and so condemning a method because we believe that is how an animal will interpret the learning event is extremely narrow. I avoid confrontational techniques (both direct and indirect) like the plague, don’t get me wrong, and I would never in a million years teach them to a client. But every single time I hear trainers argue about the quadrants, they could accomplish the same thing by talking about the nature of the method: aversive or non-aversive. Aversive or non-aversive is the heart of the quadrants of learning, and we determine the aversive nature of an event by the stress it produces, not just by how we think a dog perceives it.
The Pawsitive Packleader, Inc.