Notes from Clicker Expo Part 2 (The Four Quadrants revisited)
When I posted my illustration of the 4 quadrants, I received lots of comments about the ‘Negative Reinforcement’ (-R) picture.
Many trainers suggested that -R is not necessarily a bad thing, and that the correct definition should be removing bad stuff to increase a behavior, not delaying bad stuff to increase a behavior.
When I created this illustration, I took my definitions from Cecelie Koste’s Clicker Training 101 seminar at Clicker Expo. She used the words “delay”, “avoid” and “escape” for -R, paired with the same aversives in her +P definition. I thought the pairing was an easier way to understand the concept, but now it’s seems to cause more confusion. :/
An example of when -R is used for good (and not evil) is BAT: we remove the dog from a triggering situation so that he feels SAFE and RELIEVED. In this sense, Negative Reinforcement is a GOOD consequence that doesn’t have to be paired with Positive Punishment. *However – if the dog is already freaking out on a tight leash and if we give leash corrections before dragging him away from the trigger, then this is -R paired with +P.
3/7/11 UPDATE – Here is a modified version.
Another illustration below – the A-B-C concept is something that both Cecelie Koste and Kathy Sdao talked about at ClickerExpo.
As I understand it, the Antecedent is a fancy word for something that happens in the dog’s environment. It can be a word, a signal, an action … It can be a cue or command. It is a trigger (good or bad) that sets off a Behavior.
According to Kathy Sdao, most Traditional dog trainers put too much emphasis on the Antecedents (ie, commands) and not enough on the Consequences.
In Operant Conditioning, to get a behavior, we manipulate the Consequences that come immediately after the behavior. As in BAT, it’s like focusing on What does the dog really want when he does a behavior? Why or how does this behavior work for him? What is the functional reward?
Notes from Kathy Sdao’s DVD seminar:
What makes Carl Lewis run so fast? The starting pistol? Or the gold medal?
Why dogs don’t do what we want them to do:
- They are confused about what we want, or
- They are unpaid
Below is another illustration of the Four Consequences/Quadrants in action. In this case, the Antecedent or verbal cue remains the same (“Boogie!”) and I show that Boogie’s behavior varies depending on whether and how it has been reinforced or punished.
These are real-life Boogie examples… Click on the picture (and again) to see it bigger.
I was taught to only ever use Boogie’s name with Positive Reinforcement, so 90% of the time, Boogie responds gleefully when called. (When I am just chatting to him and don’t have treats, his name is “Pumpkinhead”, “Little Dog”, “Boogaloo” or “Boogiemonster”)
In the Negative Reinforcement picture –> this is the BAT principle in action and the reward is RELIEF. From my own experience, if Boogie is stuck with a stranger for too long and the person is staring at him, leaning over him, rubbing his face, doing all the stuff that they shouldn’t do, Boogie could lunge and bite. Whenever Boogie greets someone, I call him away after a few seconds so there is no risk of him feeling trapped.
Back to BAT and Boogie…
Yesterday I took Boogie out for a long walk around my neighborhood. We passed at least five dogs on the street and in ALL cases, Boogie stopped, looked, and turned towards me. No reactivity at all!!! No stiffness, no hair raised. Boogie was very relaxed and did a lot of sniffing and peeing.
I did my usual BAT routine: When we see a dog, I stop, wait for a polite signal (eye blinks, head turn, softening of posture, etc) , YES! and lead Boogie off in the opposite direction, treat.
Yesterday, a couple of times when the strange dog was on the opposite side of the street, Boogie turned to look at me, (YES!) but he clearly did not want to retreat. He pulled slightly forward on the leash as if to say ” I want to keep going forward”. So we did parallel walking with the other dog still on the other side of the street for several minutes. Boogie didn’t want a treat. He remained happy and relaxed, sniffing here, sniffing there, peeing here, peeing there.
The functional reward = Follow other dog at a safe distance, sniffing and peeing.
I gave him more treats anyway to reinforce this good behavior.