Heeling is a behavior chain from the earliest lessons since two steps together are a chain. That means that heeling is an exception to the basic rule of only chaining together “known” behaviors, since there is no way around the fact that the dog has to learn the moving part of the heeling skill within the chain. Of course there is a lot of pre-work that you can do to make it easier on the dog but regardless, the day comes when you have to start moving and call it training for heeling. So how does one handle errors in heeling with a dog that is just learning?
I use a cheerful interruptor (CI) with the addition of a stutter step to mark errors, and I help the dog as much as needed. By cheerfully interrupting errors and providing appropriate aides to set the dog up for a higher likelihood of success, I can teach the skill of heeling without the dog practicing long stretches of incorrect behavior.
If a dog is very fluent and experienced at heeling, then a mistake can be handled the same as any error. Mark it (possibly with a traditional NRM or a CI, depending on the dog’s temperament), and then start the heeling chain over. But in this blog, we’ll talk about the young trainee.
Here’s one way I might approach it. First, I teach a fair amount of positional work with lures, aids, and shaping – pocket hands, discs, platforms etc., to get the dog very comfortable with being at my side and looking up when there is no movement. I want the dog to have a pretty good idea of what stationary heel position looks like.
And then I get to walking! To start a less driven dog, I’ll often throw food to get the dog moving in the direction I want and for a more driven dog I’ll reward as close to exact heel position as possible. At this stage there is going to be some shaping going on. I’ll ignore errors within a chain and capture what I like, marking and rewarding as we go.
This is that weird early phase where I allow all sorts of errors in my efforts to get a few good steps in various directions. I want the dog to understand that heel position is the same in movement as it is stationary, just move your feet to stay there! I will probably allow (and even reward) a variety of positional errors if the dog is engaged, focused on me, and making some kind of effort. Once I think the dog has a clue, then I add in my CI with tons of support.
Here’s how it goes after a few days:
I start walking some direction and I want….seven steps! On the fifth step the dog makes an error. Let’s say the dog looked away. I’ll step back quickly – a stutter – and when the dog notices that something changed, I’ll restart immediately. There is no down time. I may or may not be talking to the dog at this stage of training; if I think the dog appreciates some chatter than I’ll talk. It’s ok if the dog only gives me two more steps to the reinforcer; I do not make young learners repeat the entire chain.
After that tiny stutter, I try to re-engage heeling in a manner that allows the dog to be right. I might use a cookie to reengage at their nose level, but I do not give the dog the cookie. Once the dog is back in the game I’ll start up again.
I do not make a young dog repeat the whole chain because most of the time they really aren’t sure what they are doing and I do not want to risk their self-confidence. I just re-work the element. For example, if my seven steps included a right turn at the fifth step and that is where the dog lost attention, then I would stutter backwards to catch the dog’s attention and bring the dog back to me. I often show a cookie, try to get the dog’ whipped” back into heel position by following the cookie lure, and then immediately attempt the right turn again, maybe with a lower hand position to pull the dog around the corner. If the dog makes the corner I will reward immediately.
Here’s a video to give you an idea. This is a young dog that I worked in a seminar; this is her first time trying to move in heel position. She does not know me which makes this more difficult. If she were my own dog it’s not likely that she’d try to keep leaving. Regardless, you’ll get the idea:
I’m generous with rewards as I teach heeling. Some dogs learn very quickly what I want, and their attention span for work allows for extremely rapid progress. Other dogs struggle to maintain strong attention and accuracy within heeling for more than a few steps, so with these dogs we take our time. There is no race.
In the heeling example above, the CI is an combination of stuttering (breaking flow) and stepping in with good cheer and a cookie to re-engage immediately. Dogs quickly learn that if you aren’t heeling forward with flow, then they won’t get a reinforcer. By using a CI to mark errors, the young learner is never left hanging to figure it out on their own. The dog receives lots of support and encouragement, but no free food. It works.
In these past few blogs I’ve tried to introduce you to both classic NRM’s and also CI’s and I’ve explained why I prefer CI’s under most circumstances. In this blog I’ve added the idea of a “stutter” within a chain to re-set the dog and give them another chance to earn reinforcement.