Everything is upside down down under, innit? Not only does the water go down the drain in the wrong direction, but also, dogs drive.
At the Auckland SPCA, Kiwi rescued dogs have recently been trained to operate the pedals on a car, and steer the wheel. Rest assured: unlike the Google car, they’re keeping off the main roads. And going reeeeally slow.
This training was done in order to show how smart shelter dogs can be. It’s a gimmick, sure, but it’s also a lesson in the amazing things that can be done by using some of the basic rules of positive reinforcement dog training. These science-based training methods are precisely what you can use to do everything from teaching a puppy “sit” to getting a service dog to push a button in an elevator or get a drink from the fridge. Here are some useful training tips:
Set up your animal for success: The animal trainers belted in their students. Otherwise, they might be wandering around the car, which is not a good thing for a driver to do. By limiting their movement to a fixed area, the trainers improved the likelihood that the dogs would move their paws in the most effective ways to steer, break, shift gears, etc. (Although, I wear a seatbelt and I still can’t shift gears…)
Be clear about what you reward: In the clip that we see of the training that took place before the dogs were allowed on the racetrack, you’ll notice a clicking sound. A clicker is a tool that dog trainers like me often use to pinpoint the exact moment that a correct behavior occurs. The click is always followed by a reward (usually a small edible treat, but the reward might also be the toss of a ball or praise). The noise is different than anything else the dog ever hears, and it acts like the shutter of a camera, precisely capturing a desired behavior. Because the behavior is rewarded, it is more likely that it will happen again. By slowly and incrementally upping the ante, more complex behaviors can be “shaped.” For example: At first, you might click every time the dog’s paw touches the gas pedal, but as he begins touching it more reliably, you’ll start to only click the instances where he touches it with enough force to push it.
Reward one behavior by asking for another behavior: Notice that the trainer doesn’t run over to deliver a treat every time the dog does something correctly. Instead, once he does one thing right, she asks for a new behavior. Is this unfair? Not at all. In order to build complex behavior chains, trainers will often reward one behavior by asking for another behavior — but the latter behavior should be a behavior the dog likes to do. “Sit” is often a behavior that can be used to reward something more difficult. Who doesn’t like to sit!
To become a certified dog trainer under through the Karen Pryor Academy — Pryor is one of the founders of clicker training — I had to show that my dog could do ten behaviors in a row, with one behavior acting as the reward for the preceding one. He only received a reward at the end of the behavior chain. Professional animal trainers refer to this as the Premack Principle, named after University of Pennsylvania psychology Professor David Premack; in his study of monkeys, he found that that more probable behaviors can reinforce less probable behaviors.
Like all animal training principles, Premack has a place in human life too. Next time you have a whole bunch of onerous tasks to accomplish, plan to do the easiest one last; likewise, if you need to memorize something, start by learning the last sentence first. When you’re learning a series of activities, start by perfecting the last step. If you’re teaching your kid to drive, for instance, you might begin with teaching how to park. Or you could just let the dog do it.