Today we discussed the importance of self-discovery and how lessons will have a greater impact on the dog when he makes his own discover of the behavior connections than if he is made to perform a function using compulsion. From there, the conversation moved to behavior shaping. Enter Chica!
In training we had already discussed the importance of timing, markers and rewards. The clicker was a new addition and I explained how it is used, and the limitations. I like to use a clicker with puppies, when I am teaching a new behavior and with dogs who have negative association with commands and voice inflection. Chica has been exposed to the clicker before. I "charged it" using click, then treat, click, then treat. Chica is a Dutch Shepherd and she understands markers, knows how to go to a place/perch and move around it and loves to play learning games, but she had never done what I asked of her. I placed a cardboard box on the floor. The box was deep, but not very tall so that she could not enter without crouching. The criteria was going to be that she put her head in the box, using shaping. It wasn't truly free shaping because I restricted her on a flexi-leash from going entirely away from me since we were in a large building and we weren't going to wait for hours.
Because I had introduced something novel to the environment, Chica immediately began to interact with it. She looked at it and I clicked and treated. She touched her nose to the top edge and I clicked and treated. She touched it with her nose again and looked at me. Is this what you want? I withheld the reward because I knew if I reinforced that single behavior too many times she would get stuck there and not try other options. She laid down. How about this? She sat. Her tail was wagging and she was trying to figure out what in the heck this box had to do with her. Back to the nose touch and then she lowered her head to the opening. Click/treat. It took all of a couple minutes to accomplish that. She would place her head in the opening, and I would click and bring the reward to the source of the behavior, just was we do with the detection dogs. Being able to watch the exercise helped to bring clarity to the lectures.
Then the criteria went up. Now I wanted her to not only put her head in, but to enter the box. She wasn't too sure about that because she couldn't just walk in. Tiny steps. Head in. Click. Head in again. Click. Waited for more but she backed out and threw behaviors outside the box. Sit, down, stand... how about this? So back to short rewards for attention to the box again. It took a little longer for her to actually enter the box, but ultimately she did.
Karen Pryor, well known for her clicker training, writes about the "101 things to do with a box" games on her web site: http://www.clickertraining.com/node/167
Chica was not fond of the close quarters of the box, but she was willing to try and she was rewarded for her efforts. I then took out a smaller box, open at the top and just shaped having her stick her head in the box. Easy, quick and successful! I use an adaptation of this when teaching the article indication and the head down position on command, as opposed to the method I was first exposed to where the dog is smacked on the head and commanded "such platz!!"
But the biggest success is exposing a new handler to a different way of thinking. Thinking outside the box took us inside the box.
