At Clicker Expo a few years back, Morton Egtvedt (of CANIS clickertraining Academy in Norway) gave a talk and one of his comments was that “Stimulus Control should be under stimulus control.” My ears perked up at this because I had written a long article (it’s on my website) that looked at various aspects of stimulus control and one thing I realized was that in the shaping process, it is often helpful to allow the animal to offer behavior off cue. My conclusion was that it is ok to do this (in some situations) because “being under stimulus control” can be a criteria just like anything else. One of Karen Pryor’s rules of shaping (rule #4) says “When introducing a new criterion, or aspect of the behavioral skill, temporarily relax the old ones.” The advantage to relaxing stimulus control is that it usually creates an environment where the animal is more likely to be creative and experiment. In order to shape behavior, you have to have some variability, so if you allow the animal to do the behavior without being cued, you will see more variations. You can think of it as using the original cued behavior as a starting point to create a new behavior or allowing you to improve some aspect of the original behavior. Once you have reached your goal, you can either put the behavior back under stimulus control with the original cue, or you can add a new cue. This adds another dimension to the idea that stimulus control is dynamic and does not mean that once a cue is under stimulus control it must remain forever under stimulus control. When I was first learning about stimulus control, the emphasis was on putting behavior on cue and the idea was that once you had the behavior on cue, you should not reinforce it if it was offered without being cued. But… as Alex pointed out, we were always wrecking our stimulus control because often the animal’s best efforts happen when the animal offers the behavior. If you click this superior effort, you are allowing uncued behavior. If you don’t click it, you are missing an opportunity to improve an existing behavior. This gets to the point of Morton’s comment which was that if you have really tight stimulus control, then you end up with a dog that acts like a traditionally trained dog and doesn’t offer behavior. What he teaches his dogs is that in some training sessions, they are allowed to offer behaviors without being cued. They learn to differentiate between training sessions where behavior is cued and shaping sessions where he might be working on improving some aspect of the behavior. I can’t remember exactly how he did this (or even if he described it), but I find it’s actually very easy to do. With my horses if I ask for the same behavior a few times in a row, they will start to anticipate a bit. If I allow the anticipation and click it, then they know they can offer behavior. If I reset them and don’t click, then they wait for the cue before doing anything. I suspect there are also context cues and such, but it seems like they figure it out pretty quickly. There’s another piece to this which is the effect being allowed to offer or choose behavior has on the mental state of the horse (and our relationship) during our interactions and that is the subject for tomorrow.
Posted on: Thu, 11 Jul 2013 02:20:21 +0000
Trending Topics
Recently Viewed Topics
© 2015