Forgive me if I have already mentioned this in one of my previous blogs but we worked a little on this yesterday as it reared its head again. In the clip below, you see the boy bouncing backwards and barking at me (yesterday afternoon). In the summer before I adopted him when I took him out, he would do this for the full time we were out, sometimes for 45 minutes. Exhausting for both of us. Extinction happens when a previously reinforced behaviour is now not being reinforced and the learner is expecting reinforcement. The emotion which, as a necessity, accompanies extinction, is frustration. You cannot have extinction without frustration.
Resurgence and spontaneous recovery are terms for when a behaviour which has previously undergone extinction, resurfaces. (Skinner didn’t do us any favours calling it extinction as that suggests dead forever, not so Burrhus, not so). What you are seeing here is spontaneous recovery of a previously reinforced behaviour.
It’s worth repeating, this behaviour used to go on for 45 minutes. During this time he would not take food, would move away when I tried to pet him and was only interested in a toy that I did or did not have. From what I have put togther from my time with Logan, certain envionmental stimuli trigger these bouts of barking. I have identified them as
- high arousal
- the presence of other dogs
- open spaces such as fields or parks (this is the most common one)
- fast movement from me.
Several of these combined would trigger this behaviour in the video, and sometimes one of them would act as the cue. Can you see the problems this has caused? Get them all together and we have the perfect storm.
Many trainers see little problem with using extinction. The main issue I have, and it’s a biggy, is the emotional fallout. He gets frustrated and this feeds his arousal. The more aroused he gets, the less he can think. The less he can think, the less opportunities I have to reinforce other behaviours. A horrible vicious cycle which I do my best to avoid.
Once he started to take food outside, I would then offer him a treat (sausage or cheese) and would toss it for him to catch. I hear the cries of “Dear god man, you are reinforcing the barking!” Am I? Does the food reward serve to reinforce the barking and bouncing? Is he barking because he wants the treat or is he barking because he does not know what else to do given these conditions and this is the best he can come up with? I’d rather risk occasionally reinforcing barking if it means I can provide reinforcement for a few other behaviours, than completely remove reinforcement and have him out his mind through frustration and arousal.
I would start tossing the treat, having him catch it and then lower the movement of the treat (shaping it down) so that I am either delivering it to his mouth or dropping it quietly on the floor so that he collects it and then catches me up for another. I then would give him the option of exploring by sniffing or coming back in for another calmly delivered treat. Shaping his energy levels and arousa down to nice levels.
How do we know if the treat, delivered after the barking reinforces his barking or enables him to calm down? The proof is in the behaviour we see more of. This was a few minutes later.
What I observe at 0.09 – 0.13 in the video is that he is on the verge of another barking episode. Previously, one of the other things I would do when I saw these behaviours following a bout of barking, would be to immediately throw a big load of treats on the ground in front of him. Having previously taught him to search for treats on the ground, I can then use this tactic for two purposes
- reinforce the lower arousal behaviour
- give him something else to do under the same conditions
By doing this, and having built up a long and deep history of searching, this behaviour will then resurge, the searching, not the barking and bouncing. Over time, I can then further shape this to more of what I want.
According to the literature, spontaneous recovery and resurgence are two slightly different things, depending on the conditions they occur. I’m still learning the full differences myself but they are similar enough that it should cause any confusion if I have been mistaken in identifying which is at play. Deeper understanding on my part means I’ll be better able to apply the science to help change his behaviour.
His behaviour is information about how I need to plan my next training session.
Learning never stops. Happy training.