Ever since the Suzanne Clothier seminar, where she said that being “all positive” is impossible, I’ve been obsessed with how we ought to use consequences in training.
I suppose that Suzanne could criticize my current obsession; I recognize that I may be overthinking this issue, after all. But let’s face it: this is who I am. I think. A lot. I am fascinated by science and research, and I love dog training, so really, it ought to be no surprise to anyone who knows me that I think about this stuff often. So, given all that, I’m going to continue to think about this stuff.
I’ve come to the conclusion that it is, indeed, impossible to use only R+ methods in training. But that doesn’t mean I can’t try. While there may be times that it’s necessary to use some of the other principles of operant conditioning, my personal goal is to remain firmly in the R+ area 99% of the time.
The other night, I went out with some friends. As it always does, our conversation turned to dog training. We were comparing different methods, specifically the various pros and cons of luring, shaping and capturing behaviors. I shared that for the first year, I lured every behavior with Maisy exclusively, but that she seems to learn faster through shaping.
The next morning, on one of the many dog-related email lists I belong to, the incredibly delightful Crystal Salig responded to a thread about teaching the recall by tugging on the leash. She said:
I used to mildly coerce the first part of recall also- just very light pressure on the neck with the leash and released it when the dog started to move- it wasn't harmful, it just wasn't as good as a completely uncoerced recall. I'm still trying to find a less aversive way to stop when a dog is pulling on leash so that it doesn't hurt or surprise the dog as much. When I switched from lure-reward to clicker training, I started to become even less aversive than I was before- seeing how much faster the dog learned if I didn't even lure him let alone use a physical prompt.
This all got me to thinking about my recent venture with Maisy, specifically her sniffing while on walks. Basically, if she stopped to sniff something without having been cued “go sniff,” I’d been telling her “let’s go” and kept walking. If she didn’t keeping walking, the resultant tug on the leash was a natural consequence of her behavior. And while this approach has been working- Maisy knows that “let’s go” means it’s time to stop sniffing and keep walking- I’ve felt bad about doing it.
Crystal Salig’s post helped me understand why the leash tug didn’t sit right with me. Although my action wasn’t inhumane- it’s not physically painful, nor did it appear to be causing stress on Maisy’s part- it did involve coercion, as minor as it is. I commented to my friends that Maisy learns better when she thinks something is her idea, which is why shaping seems to go faster than luring does. Tugging the leash makes Maisy stop sniffing, but it’s not her idea, so while it’s working, it’s slow and frustrating.
This doesn’t mean I think that what I’m doing is wrong. I wouldn’t have tried it if I did. But I do think there has to be a more positive way to teach this. I’m pretty sure that it includes using Premack.
So, here’s the first draft of my plan: I’m going to be vigilant on our walks. I will heavily reinforce the behavior I want- walking in a loose heel. I will also reinforce eye contact, as paying attention to me is incompatible with sniffing. I will pay close attention to her, and when I see the early signs of sniffing (and I think I can pick them out), I’ll call her back to me, ask for some kind of behavior (heeling, a sit, eye contact- it doesn’t really matter what), and then reinforce the behavior by cuing “go sniff.”
I’m quite sure this will not be the final draft of my plan. These things seem to evolve over time, after all. But from now on, all training has to pass the “feel test”- if it doesn’t feel right, then I won’t do it. And in the process, I hope to become 99% positive.