Training, as Bob Bailey likes to say, is a physical skill, a dance between timing, criteria, and rate of reinforcement. Timing is key because reinforcement needs to come right on the heels of a behavior to be associated with it. Criteria need to be set, and constantly adjusted, to an achievable level so that you can get the behavior you need to reinforce. And the rate of reinforcement needs to be high enough to make the whole enterprise worthwhile for the learner.
Criteria and rate of reinforcement have an inverse relationship. When criteria goes up—when things get harder—the rate of reinforcement goes down. So when you’re teaching a new behavior, you want to start out easy. That way you can deliver your well-timed reinforcement at a very, very high rate. Then you want to raise your criteria in steps that are small enough that the rate of reinforcement keeps the learner in the game.
Trainers often refer to keeping criteria achievable as “splitting.” If you want a dog to, say, go to a bed, lie down on it, and stay there while you cook dinner, you’re going to have to break that down. You might first reinforce the dog for investigating the bed, then stepping on it, then standing on it, then sitting on it, then lying down, then staying put, then staying put while you move one foot, then two feet, and so on until he can relax there while you generously salt and pepper a five-pound chuck roast.
Splitting criteria is a prerequisite of the training process called shaping. In shaping, you start with the closest thing the dog can already do, make the dog aware that this behavior is something you really like (via your timing and rate of reinforcement), and when he’s clearly got the idea, you determine (by keen observation) what he might be able to achieve next. Then, accordingly, you shift your criteria. Believe it or not, when criteria are shifted appropriately and the rate of reinforcement is kept high, a tedious-sounding process like the one in the previous paragraph can zip along pretty quickly.
With advanced tasks, it’s fairly easy to understand the need to split. But what about a seemingly simple behavior like “sit”?
Pretty much everyone who owns a dog has taught it to sit, by hook or by crook. Yet many are frustrated because the dog can’t keep its butt on the floor when it counts—for instance, when greeting visitors. That’s because distraction, duration, and distance are all criteria that have to be worked through when training a behavior for actual use in real life.
To keep the rate of reinforcement high (did I mention that this is important?) when introducing a new distraction, we might need to temporarily lower our other criteria. The dog may be lying nicely on the bed while you fiddle with a dry biscuit on the counter. But when you start in with salting the roast, you might temporarily reinforce just sitting on the bed again.
But how do you split a sit?
Well, a dog can’t sit if it doesn’t have all four feet on the floor. So you might start by reinforcing that in the presence of a person a few feet away, then a step closer, then standing right next to you.
A dog also can’t respond to your sit cue if he isn’t paying attention to you. (In fact, if you don’t have his attention, you shouldn’t waste your breath.) But if you’re reinforcing four feet on the floor at a high rate, your dog is very likely going to take note of what you’re up to. When you observe that, you can raise your criteria: Four feet on the floor plus attention is required to get the treat, or permission to meet the person. And when your dog is standing and looking at you as a person approaches, it will soon be easy to invite him to sit.
Standing near you and offering attention are not what most people think of as formal behaviors, but training them as such will go a long way in improving all your dog’s other skills. What is loose leash walking but attentive walking near you? Standing and looking at you is also something very polite that your dog can do when he wants your attention—say, instead of offering every trick you’ve ever taught him and then resorting to barking. (Your responsibility is then to notice that he’s quietly trying to make eye contact and attend to his needs.)
To start, go stand near your dog. Miraculously, your dog is now standing near you! Reinforce him for continuing to do that for one second. Then another second, and then another—remember, a high rate of reinforcement will help him figure out what it is you like.
Do that ten times in a row and then pause for two seconds. Most dogs will look at you as if to say, “Hey, you gonna to do that again?” Now you have the dog looking at you—the behavior we want to reinforce. I like to use a marker signal, like a clicker, to specify my criteria here: as the dog’s eyes hit your face, mark. Then treat below eye level, so the dog has to look down. Wait. He’ll look back up, you’ll mark and treat again, and you’re off. Over time you can increase how long he needs to look at you to meet criteria, around what type of distractions, etc.
Take a clicker and some especially delicious treats on your walk. If your dog walks near you for just a step or two—click and treat. If your dog glances at you—click and treat. When your dog finishes sniffing that other dog’s poop and resumes walking—click and treat. When your dog looks at a person—just looks, nothing more—click and treat.
Soon you may find that you don’t even really care if he sits anymore.