Common myth: Shaping means standing around waiting for the dog to spontaneously do the right behavior. (or worse “free shaping”… don’t get me started…)
No, wait! You know what? Let’s DO get me started!
Somewhere the internet got the idea that in order for training to really be “shaping”, the training session has to be set up to be as inconvenient as possible. The trainer must stand or sit silent, motionless, expressionless, while the dog stands in a sterile vacuum. Food must be delivered in a way that does not facilitate any behavior or encourage movement of any kind, but must also not be delivered TO the dog. And then, like the infinite monkey theorem, the dog, given enough time will randomly produce a Shakespearean sonnet. Or perhaps a haiku. Whatever.
This is bogus. Shaping has nothing to do with “waiting”, and it has nothing to do with starting in a sterile environment or delivering the reinforcement in as inconvenient a manner as possible.
In fact, that’s not part of shaping as a technique at all. It’s just bad dog training.
Good shaping really has two requirements:
1. Reinforcement of successive approximations
2. Voluntary actions from the subject
What this means is, the dog (or human) is initiating the behavior voluntarily and without coercion, and the trainer is reinforcing any behavior that is a piece of or takes a step closer to the desired behavior.
The purpose of a good shaping session is to set the dog up for success by making the desired behavior very likely to occur. So, we design the training session by setting up the environment, placement of props, position of handler, setting criteria, delivery of reinforcement… all to make the behavior (or piece of behavior) we want clean and easy for the dog. When this is done well, the rate of reinforcement is very high and there shouldn’t be much at all in the way of “waiting”.
In fact, when shaping a new behavior, if my dog goes more than 5 seconds without a reinforcer, I need to stop and reevaluate. Usually that means I need to change one of three things:
1. The set up – How I’ve arranged the environment and relative positions of myself, the dog, and any props.
2. My criteria – What exactly I’m clicking.
3. Method of reinforcement – How or where I deliver the reinforcer.
So yes, I absolutely shape quietly. I teach and encourage my students to avoid unnecessary talking during training as well. (As humans we have a strong urge to narrate as we go, which isn’t really helpful here.) I strive to have every sound that I make absolutely relevant to my dog, and I’m very conscious that anything I say or do may become part of the cue for that behavior. And while the dog is still learning the new behavior, extra talking from me is unlikely to be helpful information… it’s much more likely to simply be “noise” that distracts the dog from what he’s doing.
We also need to be conscious of the fact that we are developing our own habits at the same time as we are training the dog. And frankly, the dog’s behavior is infinitely easier for us to change than our own! Especially, since one can only concentrate on EITHER the dog’s behavior OR one’s own… not both at the same time. Easiest to keep mine as clean as possible from the beginning, and have nothing (or very little) to change. For this reason, I tend to encourage changing the environment in some way to make the behavior more likely, rather than changing something about the handler’s movement or position. It’s often easier to fade an inanimate object as a prop than to reduce an undesired hand motion. Hands tend to have minds and agendas all of their own, and upside down flower pots (usually) don’t.
We can also deliver the treat in such a way that it facilitates the behavior we are trying to shape. If you are shaping your dog to turn to the right, it may benefit you both to toss the treat to the right after you have clicked. Doing so does not mean you aren’t shaping anymore. It just means you are being a smart trainer.
So here is an example of a very (very!) realistic training session. Two versions. In the first version, I have set up the session with the box just to the side of the puppy (Baby Rugby!). I’m clicking any interest in the box whatsoever, but because the box is to the side of the path between the puppy and the food, he rarely interacts with it and quickly gets frustrated. In the second version, I bring the box between myself (purveyor of food) and the puppy, and attempt to deliver treats to the far side to start him off on the other side of the box. Then, despite *heavy* unplanned distractions, he can’t help but fall into the box because it’s right in front of him. (Full transparency: there is some substandard clicker mechanics occasionally as I tried to fend off my training assistant.) But even with that, we had MUCH more progress in the same amount of time.
And there is your practical tip for the day: When trying to shape your dog to interact with an object, place the object right between yourself and the dog… with as little daylight as possible. This makes interacting with the object very likely, even if by accident (doesn’t matter, it still counts!), which means you’ll have lots of behavior to click, which means a high rate of reinforcement, which means LEARNING!
So ultimately, shaping just means setting the dog up for success, breaking down the behavior, and making sure that the dog has the choice to perform the behavior (or not), and then making it very pleasant for the dog to perform it! Which is an awful lot like good dog training.
