An article written by Guest Blogger Cynthia Mills
The question of this blog is: do animals learn concepts, and if they do, can we teach them? And by concepts, I mean a sort of short cut, where the learner can apply a principle and then use that principle in many (or all) new situations.
I think we need to go beyond what most dog training instruction books talk about; I think we need to go beyond operant conditioning.
Without a doubt, operant conditioning gave us a great deal of success. It gave us clicker training and the means to predictably get behaviors with completely unrelated (unconditioned) stimuli—we could get a dolphin to jump through a hoop when we whistled for it. We could produce a sequence of behaviors that looked complex. There is no doubt operant conditioning works—we even use it for children when we teach them multiplication tables. The stimulus is the equation: 2 x 4, and the reward is praise or gold stars or good grades.
But learning theory, at least for humans, has gone beyond that. There is now cognitive theory, and constructionist theories—the main difference between these two more modern approaches and operant conditioning is the emphasis on learning being within the mind—that learning goes beyond the repetition of behaviors and into a realm of the imagining brain, one that can imagine the world and test what will happen virtually—to make and test theories without having to go through physical, real world, trial and error.
For sheep dogs the concept can be as simple as the pushing into the flight zone to get stock to move. Or it can be as nuanced as maintaining a distance just far enough behind the sheep to be able to see both eyes of all the sheep, and then work both eyes, leaning out to remind the sheep they can’t go that way, or the other way, as the dog is still there.
I have seen in one of my own dogs the serendipitous moment when she learned that principle of control, and went from a dog who wouldn’t drive to one who loved to drive.
Now you might say that sort of control is instinctive, and I wouldn’t entirely disagree with you. Instead I might insist the motivation to find that control is instinctive, and the sense of having found it—the satisfaction (and dopamine release) the dog feels when they are there is certainly part of the hardware of a well bred sheep dog.
Not unlike the pleasure any one of us might feel when we finally “get” quadratic equations, or iambic pentameter.
The authors of most dog training books might disagree, but they can go beyond scientists to learn about how dogs learn. They can delve into the methods of trainers of working dogs. While trainers don’t do formal experimentation, like double blinded clinical trials, they, if they are good, do try methods and test them, sorting out what works. Most trainers of working dogs have gone long beyond operant conditioning—it’s way more efficient.