Learning By Association
By Gary Wilkes
Imagine a dog in nature. It has instinctive reactions that jump-start survival. Those instincts are not sufficient to protect the dog in all cases. Learned behaviors are added that modify and augment instincts to give the animal a chance to survive in a complex, often hostile world.
The most basic form of learning is making a simple association between two initially unconnected things. A dog does not naturally know that a groomer trims nails or arriving at the boarding kennel means they won’t see their owners for a while. They figure that out pretty quickly. At the turn of the last century, one man started the scientific research aimed at finding the rules of the process. His name was Ivan Pavlov.
About all anyone knows about Pavlov—if they have ever heard the name—is that he experimented on dogs and made them drool. As drooling isn’t especially valuable to most people, the name isn’t connected with anything important and has less impact than knowing what a stripper brush does or how to scissor a perfect mug on a Schnauzer. Guess what? That is what Pavlov studied—the process of making simple associations. His goal wasn’t to make dogs drool. It was to find a way to track the strength of associations to create predictable and trustworthy knowledge of the process.
The famous drool experiment is simple. Find something that has no meaning to a dog—like the sound of a bell ringing. Offer the meaningless thing and then connect it to something that does mean something—like food. Do that 20–50 times and see what happens. The dog doesn’t just drool when it hears the bell; it gets excited the same way your dog does at the sound of kibble falling into its food bowl. This isn’t rocket science. Everyone knows that animals can make such associations.
What was intriguing to Pavlov is how they make those associations. Now we get a little more complex in our investigation of drooling. You see, there really are rules that, if you obey, will dramatically improve your understanding of the dogs you handle. To give you a justification of why that’s important, just watch your clients interact with their dogs and you’ll start seeing how few of them actually know the rules.
We often hear that trainers have to have good timing. OK. What does that mean, exactly? They imply that associating a behavior with an event like food, or some kind of correction, must be applied at a particular instant to make a behavior stick in the dog’s mind as important. That is true, but doesn’t explain how that happens.
If a dog is 20 feet away and pees in the correct part of the yard, how do you toss a treat with perfect timing? You can’t. That is why we use praise. The praise can be applied at the correct instant that allows the dog to come get its treat. We all know that. That means that having pre-learned signals like ‘good girl’ and ‘no’ make connecting behaviors to consequences practical is a pretty big part of training.
That leads to a big, big question. How do we create a ‘praise’ phrase? How do we connect ‘NO’ to a negative consequence? Words without consequences have no meaning. You can see that every day as you watch people chatter at dogs with the intent of controlling their behavior. “Sit, sit, sit” most often means absolutely nothing to the dog. How do we make these things into solid associations? We dig into Pavlov.
Pavlov’s most basic formula connected the sound of a bell to food. He would pair these things together for about 20–50 repetitions until the sound of the bell, alone, would produce the same amount of saliva as real food. That is the gold standard of associations—the learned signal creates the same reaction as the tangible consequence.
There is a rule that helps the process develop. Ring the bell, then feed the dog. It sounds simple. You will find that it is not. You will want to grab the treat and offer it to the dog at the same time or before you ring the bell. I know that because the process of connecting a training clicker to a treat is identical. After thousands of clients, I know this to be true.
People are focused on getting the treat to the dog’s mouth and make all kinds of signals that tip the dog off to what’s coming. (That’s why dogs know the sound of plastic crinkling—it’s the sound of a hand going into a treat bag.) While this works, it’s inefficient compared to doing it the Pavlovian way —ring the bell (click the clicker), wait a second before you move to get the treat, and then offer the food.
Our next question is obvious. How much time can I wait after I ring the bell to offer the treat? The answer will surprise you. It doesn’t much matter. It is not the time between the bell and treats that is most important. It is the predictability that it will happen. To prove this we only have to look at your front door.
Ask yourself what is the fastest time between when you hear the bell and when the guest enters. It’s never one or two seconds. It can be as much as a minute. Yet all dogs who live in houses with doorbells learn the association—ring the bell, followed by an important event. That is why a dog afraid of the blow dryer starts trembling before the event happens. They know that the bath precedes dryer.
While much of behavioral science means little of value to groomers or daycare workers, some of it clarifies our understanding of behavior. This simple concept that pairing an event with a consequence can refine your control of the dogs you handle.
If you say “Up we go” before you bend over to pick up the dog, it will learn to prepare itself for being lifted. If you grab the dog and then say “Up we go”, you are wasting your breath. Knowing this simple sequence makes the process more predictable and less stressful for more dogs.