Understanding Error-Free Learning

Trigger tending her sheep

Error-free learning, is just what it sounds like. It's the concept that through excellent management and good training, we can guide our learners to the solution without the repercussions of making a mistake.

I think there are certainly great examples of where this methodology will go a long way towards avoiding future problems.

I practice error-free learning in housetraining. I set my puppies up so that they are never permitted to make mistakes in the house. While this means that I have to be constantly alert when my puppies are young, and have my puppies moving regularly from outside, to heavily supervised inside, to crate, and out again on a pretty intrusive rotation, It also means my puppies pick things up pretty quickly, and I'm not cleaning my rugs daily.

Management is a large part of error-free learning. By which I mean, we don't give our dogs the opportunity to make mistakes. We dont leave them alone in the kitchen with a T-bone steak on the counter. We don't leave a tub of KFC bones in an unattended trash bag on the floor. Some things are easier to manage than others.

A friend of mine and I disagree on how much error-free learning is required for loose leash walking. She insists that her learners never feel leash pressure. She uses a heavy reinforcement schedule and contained environments to build the concept without the dog ever encountering the end of the leash. This is a very effective method.

I do not do error free learning for loose leash walking. however, I ensure that the ramifications for error (placing pressure on the leash) are low and salient to the dog. By that, I mean, I allow puppies to hit the end of the leash, and when they do, everything stops.If done correctly, this leads to a discrepancy between expectations and consequences. The puppy is cruising along having fun, and the expectation is that this will continue until they stop; however, if everything stops when they hit the end of the leash, that confusion creates an opportunity to figure out how to avoid everything stoppng going forward.

This is not to be confused with the old methods of having a dog on a long line and choke chain and ripping them off their feet when they hit the end. Punishments, like everything else should be tailored to the knowledge, age, and skills of the learner. We don't ask our kids to recite the alphabet and slap them for doing it wrong, especially if we have never even bothered to teach it to them in the first place. Even in the absence of error-free learning, we should be setting our learners up for sucess rather than setting them ip for failure with 'gotcha' moments, and heavy corrections.

Herding cannot be accomplished in an error-free environement. There are too many moving parts. We try to ensure that errors when presented are not too stressful or difficult to fix - think of first learning to drive in a parking lot versus a five-lane high speed highway through a major city.

Recently, Trigger went through what is called a fear period. Most growing puppies have two such periods, and they are a normal part of growing up. These periods can make dogs extrememly sensitive to corrections, either from their people, or from the environment. So, we try, when faced with these periods, to avoid doing things that might create a long-term issue.

Trigger became ultra handler soft. Quitting herding if I so much as said "ahhh". So, we went back to an error-free learning model. We allowed her to continue to herd, and she continued to learn; however, we removed the opportunities for mistakes that would require me to intervene.

After three weeks of this, her fear period ended, and her confidence restored. She is actually more resiliant than she was prior.

Understanding when and how to use error-free learning is a bit of an art. We cannot keep our dogs completely free of making mistakes, any more than we can keep our young driver forever in the parking lot. Rather, we allow our dogs to feel out errors in places where risks are small or of little consequence.