Capturing and Shaping

This evening I sent an Email to some of my friends, describing the progress the puppies have made during the past week and 1/2 while at my place of employment. It's really a rather impressive list.

- Run up the ramp to load into the truck

- Rest comfortably while on drop chains

- Run up and down ‘open’, steel-grated stairs without hesitation

- “sit” on cue with excellent latency and duration. We are now working on stimulus control.

- “touch” a finger tip, ink pen or other extended object on cue. Currently working on latency

- Sleeping or being quiet in separate compartments in the dog truck

The puppies have met lots of new people, of many different sizes and shapes including several in uniform (I work as a security officer/paramedic as do many of my co-workers). If these guys ever encounter a cop or uniformed ACO they'll probably run right up to them begging treats .

While I continue to reinforce cued "sits" and "touches", I'm started to expand their repertoire a bit. I want to train all three of them to give me a "down" on cue. I'm especially interested in being able to have them perform the "down" behavior while in harness on the gangline. Then, if I have to stop to clear a tangle or deal with some problem on the trail I can put the dogs on a "down", do what I need to do knowing that a dog resting on his or her belly can't bolt without making a large, noticeable movement before slamming into the harness.

When I trained the puppies to "sit" I used a food treat passed in front of their nose and over their head to lure them into position. As the puppy's nose goes up to follow the enticing smell their little butts go down. As soon a butt hit floor I gave a click and treat (C/T). Once they knew the behavior of "sitting" earned them a C/T, it was easy to start shaping duration (time between the behavior and delivery of the c/t) and latency (time between the cue and behavior). Now all of the puppies are working on stimulus control - in other words they are learning to "sit" when they are given the cue, but only when they are given the cue.

I also lured the puppies to train "touch". "Touch" is the cue for the puppy to touch the tip of my finger, the tip of a pen or pencil or any similar object with the nose. Called "targeting", I can then use the target to train other behaviors in the future, such as loose leashing walking, walking near me without a leash at all, 'line out' on the gangline (hold the tug line tight against the resistance of the gang line) and many others.

Although one can certainly lure a sitting down into a down position, I'm trying to avoid luring for this behavior. Instead, I am trying to 'capture' a behavior they already do, and then 'shape' it into the finished version of the cued behavior I want to train. This will help teach the puppy to repeat behaviors that earn them a c/t, and helps me develop and maintain a broader repertoire of training methods and techniques.

To capture the 'down' I watch the puppy, usually out of the corner of my eye. When I see the puppy lying down I click the clicker while the puppy is going down, and toss him or her a treat as soon as I can. To do this it helps to know when the puppy is most likely to lie down on his or her own.

Capella was a good study in canine problem solving as we started the process with her tonight. She dearly loves playing with her squeaky ball. She tosses it around the room, chases it, bats it around like a cat, tries to roll on it (she'll be the one most likely to roll in stinky stuff I think) and eventually lies down with it. Sure enough she went through the whole process and as she lying down with her ball I give her a click and stuffed a Yummy Chummy (TM) into her mouth. Then the fun began as she started trying to figure out what she had done that earned the reward.

She pounced her ball and looked up at me. Nope
She batted her ball across the room, chased it down and looked up at me. Nope, not that either.
She rolled on her ball and looked up at me. Naw, that won't work.
She heaved a big sigh, sagged to floor with her ball, and CLICK.

She repeated the process, only a bit more quickly, finally settling down with the ball to earn "Click".

At that point it was like the little light bulb went off in her head. We did three reps without all the ball play in the middle before it was time to end the session.

I'm sure we'll have to repeat some of that preliminary stuff at the start of the next session, but it won't take long for her to figure out the behavior that earns the reward, and I'm betting that within just a couple of days I'll have her on a cued "down" that I can then shape into a very elegant behavior.