banner

Tuesday, March 23, 2010

Clicking in Shape

Oscar and I have used a variety of training methods over the years. The more I learn about training, the more I lean towards positive reinforcement methods, but there are hundreds of styles out there, and Oscar puts up with most of them pretty well (though he is what is called a "soft" dog, in that he responds very personally to his handler's reactions, and can get his feelings hurt relatively easily).

With Liberty and her fear issues, I have done only positive reinforcement. Her negative behaviors are completely ignored, so long as they don't cause danger (ie: she is distracted from eating dangerous things, not yelled at for trying) and anything positive is rewarded. There are those who don't understand the method that call it "bribery" but it's nearly anything but that, and we had a nice training session the other day that illustrates why that is.

Liberty has a pretty severe case of separation anxiety. That means when I leave the house at all, she goes into a raging panic and is unable to calm herself down at all. It makes sense, she had a pretty horrific past, then came to live with three silly cats, a dog that she adores, and the first person who's ever been consistently kind to her. She's pretty attached to me, and when I go away, it's very, very scary to her. She was crated for a while, but ate a huge hole in the side of her crate, and that became dangerous for her to be in safely without hurting herself. So we've started over with her crate training, with a new crate, but until she's improved a little, I worry less about her doing damage in my house than I do about her hurting herself trying to get out of a crate.

And so Liberty is being clicker trained to love her crate. A clicker is like a little personal United Nations delegate. It means I don't need to speak dog, and she doesn't need to speak English, that little toy will do the translating for us. It lets me tell her "yes, that specific moment right there - that's what will earn you goodies". First we charge the clicker, which just means clicking it, then treating her immediately after, until she learns that when she hears that click, food is coming. Once she's solid on this principle, I can start getting her to do whatever I want, by working in small increments. That's called "shaping a behavior" because you'll start with something closer to the behavior than what you had, and use the clicker to maneuver it into the final behavior. Seeing a dog (or any animal - it'll work on anything, even humans if you're sneaky enough) figure out that they have the power to make that thing click is pretty exciting, for them and you!

This is video of our very first session of clicker training Liberty to lie down and relax in her crate. She did lie down in class the week before, but that was the first work she'd done similar to this, and it didn't have that super scary crate involved. In theory, clicker training can be done without speaking at all, but I'm talking a lot here to explain to you what's going on. This is one session long, but had to be split into two for the sake of upload limits. You can see that she isn't cued, lured or tricked into any behavior - she comes up with the action on her own, and isn't bribed to do anything - she's just rewarded when she does.



3 Comments:

Blogger Walden said...

Loved the videos. It's nice to see how you train her and how it works.

3:40 AM  
Anonymous k said...

That's pretty amazing!

11:17 AM  
Anonymous k said...

So, do you start leaving her in there for short bits of time, while you're in the house? What comes next?

I found a copy of Interweave Knits with that huntingdon castle sweater on it. Weren't you in love with that?

10:34 AM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home

      
Marriage is love.