Has anyone ever used this technique to teach complex behaviors? Our instructor showed us how she uses intermediate bridge and targeting to teach side-stepping, weaving between legs, and service activities (turning on lights, going to get objects, etc.). It was awesome, and totally eye opening. I was thinking that it would be a fun way to teach Theo some things we've been having trouble with, like really precise side-stepping and pivoting.
This potentially gets around a problem I have with free-shaping. Theo can do 5,000 repetitions wrong in a row without changing (that Lab steadiness!), so it can be difficult to capture the right behavior! (And of course I always am fumbling with the clicker at the moment he finally DOES do it right.)
I haven't read the article thoroughly, mostly because it's too confusing to follow (I'm a visual learner - show me), but I think most trainers use bridges and targets when teaching.* I'm not sure if I use an intermediate bridge or not but I use a terminal bridge and* targets.
Lydia, the article *is* confusing, but it was the only thing on the Web I could find. I guess most people who do OC use terminal bridges and targets. It's the intermediate bridge that struck me as so interesting. I still don't completely get it, but it's basically a continuous signal you use to teach your dog that they're going in the right direction, so it eliminates a lot of the noise in capturing and shaping. Our trainer also uses it for reactive dogs.
Most operant trainers (myself included) don't use intermediate bridges that often. The reason why is that we don't like to 'lump', so if we break the behaviour down into small bits (split), then there is no need for an intermediate bridge.
This author gives the example of having a dog touch three targets simultaneously and uses her intermediate bridge to indicate the first two targets touched and her primary bridge for the last one. Touching three targets at one time is a behaviour chain and each step should be tought separately. Most operant trainers would split the behaviour and teach each one separately and then combine the three of them as a behaviour chain. The advantage of splitting and lumping is that there is no need for an intermediate bridge, so there is no need for endless chatter with your dogs (which is something that needlessly needs to be faded).
If you are good at splitting and increasing criteria, there is no need for an intermediate bridge.If your dog is doing 5000 reps of a wrong behaviour then you are either sloppy in your marking or too strict in your criteria. Chances are, you are sloppy in your timing as you said, as most dogs wouldn't do a large number of repetitions if the criteria were too large. Placement of food reward also helps to prompt the correct behaviour without outright luring and sets your dog up for success.This potentially gets around a problem I have with free-shaping. Theo can do 5,000 repetitions wrong in a row without changing (that Lab steadiness!), so it can be difficult to capture the right behavior! (And of course I always am fumbling with the clicker at the moment he finally DOES do it right.)
Most of us do use intermediate bridges of some sort....praise. I use praise to let me dogs know how they are doing on winning the prizes with longer duration behaviours such as heeling or stays. Praise means my dogs are well on their way to earning their click and treat.
To err is human:To forgive, canine."
Thanks, that's really interesting and helpful. Our trainer just uses it for a few odd tasks; I was just impressed at how quickly it helped the dog she was working with. (The issue was getting a dog who was sensitive about its hind end to move that hind end closer to a target.)
As for my horrible timing and its probable effects on Theo: Yes, you're probably right. I seem to be immune to improvement in this arena, but I am trying. It's very frustrating for those of us who are terminally clumsy, and my frustration certainly impedes my progress. (Didn't Skinner say that internal states were irrelevant to science? No wonder behaviorism is dead! )