****Interrupt - Redirect - Reinforce****
Those three words are the answer if you want to change a dog's behavior using positive reinforcement techniques. When I am training dogs, these three words run through my head, over and over, just like a mantra.
I suppose there might be limits, but honestly, I think I could get any dog to do anything if I followed that mantra. This training makes sense to dogs; they love it, they are happy to change their behavior, and they are never aware they were "wrong." In my world, a dog is never wrong, but I may want something different from him, so I just create that instead.
This twelve-second video will show you how easy this training sequence can be.
The backstory
This is the first time either of these dogs has been off long-lines among this flock. The flock consists of four very flighty sheep, two steady goats, two very young lambs, and a baby goat. This is a zippy group!
The dark pup, Thor, is a steady, slow fellow. He doesn't get provoked by much. The seven-month-old Maremma pup has a different guarding style and speed; they are both excellent at their job, but they do it differently. The two dogs had been living together as a pair for about a week, one field over from this flock. I have done long-line work with both dogs with the flock in their small field.
So, as a trainer, I know some things. I have made sure that both dogs understand what these animals are. They have no need to pursue them to learn about them in this larger field because I gave them that opportunity in a smaller, contained area. The field we are working in today is familiar to them; they just have never shared it with this flock before.
At the time that I did the redirect, the pups had been at liberty with the flock for about three minutes. We started at the back fence coming towards the camera while the flock moved ahead of us. They slowed and stopped along the driveway. The Maremma pup went to investigate them. She approached them slowly, but they cantered away from her; not her fault, this flock is very reactive.
She canters with them, slightly behind them but not chasing them. She just would like to be near them. I watch the speed of the flock, which is increasing. I watch the speed of the pup, and she is steady. I allow her five canter strides because I want to see what she'll decide to do on her own. I would have considered three canter strides, or no canter strides, to be ideal, but she is a young dog in a new situation, so she gets some support from me - not correction, support. I need to create a new behavior: I need to change canter to walk.
I spoke to her brightly, in a singsong fashion. She glanced at me. I used my verbal marker word, quickly followed by the click. The click stops the behavior. As she turned to us to get her promised treat, she was getting tons of verbal praise from my assistant and me. Both of us had chicken in our pouches; I clicked, but my assistant paid. That was the pup's choice.
I'd also like to mention that when training with more than one dog, all the dogs get paid, every time.
The dog that got clicked knows his behavior was marked because I have eye contact with him. Any other dogs nearby that come to get a treat get paid for breathing, basically. Part of the reason this is important is that always paying both dogs means they have no need to compete with each other, for attention or treats. For the second dog, it also means that there is great benefit in staying near humans that are doing training sessions because it is easy and fun, and they always get paid.
We hung out with the pups for thirty seconds or so and then headed towards the flock again, but this time the pup approached them slowly, and she did so for the rest of the session.
Interrupt: Me speaking to her.
Redirect: Her returning to me, giving me eye contact and her attention.
Reinforce: Baked chicken, verbal and tactile praise.
Oh my goodness, how much more interesting we are than cantering behind the flock is. When we move off again, the pup is in a steady, confident frame of mind because, in addition to stopping her body in this sequence, we also gave her an emotional reset.
Correction-based options
A typical correction-based response to this pup's fifth canter stride might be to say, "No!". The pup would probably stop following the flock, but her body language would now be very different, as would her emotional state.
She would lower her body, and cower possibly. Being in this new circumstance with the flock now poses some risk to her, so now maybe it isn't all that happy an event; she is now a worried dog. She might return to her owner, but she would not do so joyfully, as our pup did.
After a time or two of being told "No!" for much of anything, an LGD will learn to avoid the person with the voice and possibly people in general.
Now you have the typical LGD we all have heard so much about; the independent, stubborn, difficult-to-train LGD. I'm thinking it's the human that needs the training because the dog is not the problem; it is the language of the trainer.
Comments