June 16, 2009— -- Karen Pryor, a regarded expert in no-punishment animal training, describes all she has learned about teaching animals and what they have taught her in her 30-year career.
In "Reaching the Animal Mind," she advises that people use a cheap, plastic, handheld clicker to reward animals' good behavior and ignore the bad. Using this positive reinforcement system, you can train your dog to find your car keys or your cat to give you a high-five, she says.
She also gives Web addresses so readers can watch video of the training sessions explained in the book.
Read an excerpt below and head to the "GMA" Library for more good reads.
I'm standing at the edge of a dusty road in a little town in South America. A barefoot, grimy boy walks past, a very little boy, maybe between three and four years old. He's eating a bun. Behind him trails a skinny puppy, itself very young.
The boy turns around, sees the dog, and raises a threatening fist. The dog cowers dramatically, cringing to the ground. The boy looks up with a huge, triumphant grin: "I scared the heck out of him, didn't I!" He walks on down the road. The puppy gets up and slinks after him -- and guess what: the boy has forgotten about the bun. He lets it fall, and the puppy grabs it and runs away.
That's how we've dealt with domestic animals ever since we and they evolved together. We treat them like subordinate, stupid human beings. We dominate them. We punish them. We make them do what we want. And they figure out how to get us to do what they want, anyway. Both sides get some benefit out of the system: in this case, food for the skinny puppy, and a rare moment of superiority for a small boy.
Traditionally the person who actually trains animals, beyond these ordinary practices of threatening them one minute and feeding them the next, has always been a special individual. Often it's someone with a "way with animals," a "natural gift." Usually that gift consists of two things: a personal interest in some particular kind of animals (dog trainers train dogs; horse trainers train horses) and a better understanding than the rest of us of the subtle uses of fear and force.
Traditional animal training, the way it's been practiced for millennia, relies largely on force, intimidation, and pain. While traditional trainers may also use praise and rewards, dominating the animal and obtaining control over its behavior are the main goals, and the main tools are fear and pain.
Traditional trainers are abundant among us. Nowadays of course they justify their practices with pseudoscientific explanations about pack leadership and the importance of dominance and of being the alpha animal; but the basic method, in spite of the overlay, is punishment; and people generally accept that approach. Most horse owners still keep whips and spurs in the barn. The walls in pet stores are plastered with choke chains and the aisles lined with shock collars, and people buy them. Maybe you use them yourself. I won't argue with you. Force and intimidation have been working for people since the first dogs hung around the first campfires (or, more likely, around the first garbage dumps).
But that's all obsolete now. Now we have a new way of dealing with animals. Out of real science we've developed a training technology. Like any good technology it's a system that anyone can use. The basics are easy to learn. It works with all animals (and that includes people). It's fast. What used to take months, the traditional way, can now happen in minutes. It's completely benign; punishment and force are never part of the learning system. And it produces real communication between two species.
Erich Klinghammer, a professor at Purdue University, is a well-known ethologist. He is the founder of a research facility in Indiana called Wolf Park. Dr. Klinghammer came across my book Lads Before the Wind, which describes the years in which I worked as head dolphin trainer at a pioneering oceanarium, Sea Life Park, in Hawaii. Klinghammer saw that the technology we used for training dolphins would be useful for managing wolves. He invited me to Wolf Park to show histeam how to do it.
We modern trainers love the chance to work with a new species. Not just one more dog or horse or dolphin, but something we have never trained before. We begin, always, with curiosity: "Who are you? What can you do? Show me." I had never worked with wolves, so of course I said yes.
A few weeks later I fly to Indiana. At Wolf Park, Erich Klinghammer is eager to have me go into the pens and meet some wolves personally, to "experience their boisterousness." This I am not willing to do. Klinghammer is six feet four with a big Germanic bass voice. He walks through the gate into the main pack's enclosure and booms, "Good morning, wolves!" The wolves gather around him, waving their tails and jumping up to greet him: "Good morning, Dr. Klinghammer!" For me, I think it would be "Good morning, breakfast."
Besides, I don't need to be close to a wolf to work the training magic; in fact, both of us are safer and will feel better with a fence between us. This wonderful technology does not depend on my being able to impress or dominate the wolf. Nor does it depend on making friends first, or on having a "good relationship." That's often a happy outcome, but it's not a requirement: the laws of reinforcement will get the job done.
Klinghammer has selected a large male, D'Artagnan, as my learner. That's a typical wolf name; no one calls wolves Pete or Blackie or Pal. D'Artagnan was raised by humans, so he does not know how to get along with other wolves and has to live alone in a pen on the far side of the park. Klinghammer and I jump into a truck with a couple of students and a large can of dry dog food and drive to D'Artagnan's pen. I get out my dolphin trainer's whistle, pick up the can of kibble, and go over to the chain-link fence. Wolves look a lot like dogs in paintings and even in photographs, but in real life they're quite different. For one thing they don't have pointed ears like a German shepherd, but small, round ears, like a bear; for another, they don't smell like dogs, they smell like fur rugs.
D'Artagnan meets me with a spectacular threat display, snarling, snapping, and lunging at the chain-link fence between us. He is about the size of a St. Bernard but with much wider jaws and bigger teeth, especially the bone-crunching carnassials in back, at which I am getting a really good look.
I'm sure this show of aggression is learned behavior. His hackles are not up, his eye whites are not showing; he's not really that upset. However, he has probably discovered he can sometimes make people flinch, or even run away, by being scary; and that must be fun to do.
The first step toward change is to explain to the wolf that when he hears the whistle, food will arrive. I blow the whistle and throw in some kibble. D'Artagnan just goes on snarling and leaping and snapping at my face. The chain-link fence between us suddenly seems flimsy. I don't want to reinforce his behavior by moving away, but it is indeed difficult to just stand there.
A Jeep full of volunteers and students pulls up. The wolf is quiet for an instant, studying the Jeep. I whistle and toss more kibble through the fence right under his nose. "Oh," he says, and vacuums up the food. I whistle and toss kibble again. Again he eats the kibble. Then he looks at me. I do nothing. He turns away. Good! It's a relief to see the back of that wolf.
So far, I've just paired the whistle with the food, to make it a "conditioned reinforcer," a sound that means "food is coming." I'm now going to start using the whistle to identify for the wolf what action he's getting paid for. This will turn the sound into an event marker (usually just called a marker). So I whistle as he moves away, and toss in more treats. He returns and eats again.
Now that the wolf is listening to the whistle, coming back for more food when he hears it, and has been reinforced for moving away, I can begin "shaping" his behavior. Shaping is the technical term for shifting a behavior by reinforcing any moves that happen to occur in the direction you have in mind and ignoring everything else.
In the middle of D'Artagnan's big enclosure, about thirty feet off, stands a small evergreen tree. I tell the watchers, "I'm going to train him to go out and around that tree and come back." Bravado, of course. I'm going to try would be a wiser promise; but even if it only works for part of the way, it's a useful demonstration. Using a marker signal to shape the behavior of going away from the food, in order to earn the food, helps to shut up the skeptics.
I mark each time the wolf turns away from me, timing my whistle to the stepping of his right front paw. Every time he hears the whistle, I toss in a lump or two of kibble. He snatches them up and then, with increasing confidence, turns and starts moving directly away from me again.
I wait for three strides before I give the whistle, then five, then ten. Now D'Artagnan is so confident he grabs up the kibble and actually trots off, still chewing. Each time, I wait to blow the whistle until he's gone farther than before. Now he's going more than halfway to the little tree.
Then, on the next try, he stops before I blow the whistle. Uh-oh. He is new to the game and hasn't had enough experience to know that a missed marker doesn't mean the game is over, but simply means that you should try again. If I give him a whistle while he is standing still, he might develop the behavior of going out just that far and stopping. And if I don't give him a whistle at all, he might quit altogether. I watch him, praying he'll take another step forward so I can reinforce moving, not just standing.
Instead, he turns around and makes eye contact with me. His yellow eyes look into mine with a focus so intense that it feels as if he were seeing right into my brain -- which, in a way, he is. That penetrating stare is literally breathtaking: I hold my breath and look steadily back.
Deciding, I think, that the game is still on, D'Artagnan turns away from me and heads straight off toward the tree again, breaking into a canter for the first time. I blast the whistle to mark that bold decision. His loping stride brings him abreast of the little evergreen. The wolf wheels around the tree and comes back down the other side at a gallop, screeching to a halt right in front of me, making my brag come true. Quit while you're ahead! I push a double handful of kibble through the chain link and leave him to enjoy that jackpot. Thanks, wolf! You saved my neck.
One experience was all D'Artagnan needed. Now, when Klinghammer takes visitors around Wolf Park, he can drive up to D'Artagnan's pen and blow the horn to call him. The wolf comes to the fence, sizes up the situation (Klinghammer, Jeep, whistle, kibble, got it), then turns away, gallops out and around the little evergreen, hears a blast of the whistle, and comes back for his treats. This new skill also ends his aggression display. Training people to give you kibble is much, much more satisfying than the old game of Get the Guest.
That single demonstration with D'Artagnan was also all Klinghammer needed to convert Wolf Park to the new technology. The staff and volunteers began using treats and acoustic markers, either whistles or clickers, to handle and move wolves and to give medical care. Staff ethologist Pat Goodman mitigated her border collie's irritating habit of staring at her incessantly by teaching him to turn his head when she whistled "Dixie" ("Look away, look away..."). Klinghammer, meanwhile, had fun using the same principles to coach a Purdue girls' volleyball team.
That demonstration paid off for me, too. I learned that wolves, or at least this wolf, enjoy a bit of fun: his game of scaring people was on the rough side, but it was a game. Then, that memorably powerful look into my eyes told me something more: compared to dogs, wolves are grown-ups. He was not asking for help, head down, forehead wrinkled, as a dog might: "Is this right? What do you want?" Instead, head high, gaze level, he was assessing me, like a poker player: "Are you in or out?" Judging that I was in, he made his move; and we both won.