When I crossed over to training with positive reinforcement, I had no idea how much my behavior and even my belief system would need to change. I had to question my faith in some long-held cultural assumptions and learn to rely on scientific observation and analysis.
Crossing over was a lengthy process for me, and even now, after more than 10 years, I occasionally fall back onto old assumptions and behaviors. I wonder sometimes if I am the only one so vulnerable to cultural programming. But a quick look around social media says no, I’m probably not.
There are intellectual, emotional, and cultural barriers to crossing over. For me, certain barriers were so large that they defined whole phases in my thinking and practice about training. I’ll share several of these phases here. Maybe they will be familiar to some of you. And maybe identifying them could be helpful to trainers who meet with resistance or confusion from their clients. Once upon a time, I was that client.
“I Tried R+ and It Didn’t Work”
This was my experience, and it was real—it’s not just something people say to provoke positive reinforcement trainers. I went in and out of this phase, trying and failing several times.
In 2002, when I got my rat terrier, Cricket, I read about positive reinforcement training on the young Internet. I wanted to teach Cricket to walk on a loose leash. I read about the “Be a Tree” method, wherein one stops forward progress whenever the dog pulls. I thought I was trying positive reinforcement training when I tried to be a tree. It sounded elegant and reasonable. But I didn’t know to start indoors, in low distraction. What I had read didn’t mention using food as a reinforcer when the dog was walking nicely, so I didn’t. And I didn’t know about any of the quadrants then, much less recognize the presence of the other three in the protocol as I was practicing it.
I followed the “Be a Tree” protocol as faithfully as I knew how every day for six months, and, not surprisingly, it didn’t work. Cricket would immediately tighten the leash, and I would stop. She would stand there, barking. I would wait until she accidentally loosened the leash. (It often took a while.) Then we would go on, perhaps three more steps, and the process would repeat.
Now I know that she was probably too far over threshold to perceive that the loosening of the leash was connected to being allowed to go forward and that some sort of reinforcement (though not intended by me) was working to maintain the barking. And I know for sure my timing was bad. But since my attempt at the method was unsuccessful, I assumed that positive reinforcement training in general didn’t work. If I had known about learning theory then, I wouldn’t have hesitated to further generalize that learning theory didn’t work either. (Indeed, that was a later phase.)
Isn’t that a little strange? Why would I reject the whole thing rather than consider that my knowledge might be incomplete?
Suppose I had an orthopedic problem and needed surgery. The operation, a well-understood and documented procedure to be performed by a skilled surgeon, had a predicted 90 percent chance of success—but it failed.
My possible responses might include:
• The surgeon did her best, but due to complications I was forewarned about, the method failed
• The surgeon failed
• Western medicine failed
It’s so easy to jump to that third response with dog training. When we are new to training, the idea that it could be based in science is new too. And the science doesn’t always fit well with a lot of what we “know” from living in a punishment-based culture.
“R+ Is Not Practical”
Believe it or not, I failed a second time with loose leash walking, four years later and with a different dog. I was toying with positive reinforcement training again and had read about Premack’s principle, a theory stating that a stronger response will reinforce a weaker response. Since my new dog, Summer, was fixated on squirrels, I decided that running together to a tree where there was a squirrel would be the reward for walking nicely for a few steps and sitting and giving me eye contact. She quickly learned how to “ask” to run to the squirrel. The problem was getting her attention back after that. Also, I had accidentally created the adrenaline-filled, anticipatory stay so prized by some agility competitors. Summer was on pins and needles, then exploded into action when released. But I lacked the skill to get her back, and I saw that the main result was Summer getting more and more hyped up on walks.
I thought, “Well, that worked, but it’s sure not very practical.”
“Force Is Necessary for Dogs with Issues”
Another common phase of crossing over is the period where we believe that positive reinforcement is fine for most dogs/teaching tricks/teaching the basics, but we still need to punish dogs with behavior problems. Yes, I really believed this. I remember one night at an obedience club seeing a dog that was said to be aggressive. I told my friend that I was glad the dog was wearing a prong collar—this was a dog who needed it, and we had to consider the safety of the other dogs. It made intuitive sense to me that tough dogs needed a tougher approach. (I had never watched that show, by the way. And the idea of “being the boss” resonated culturally well before Cesar came along.)
I could think of no option except to suppress the “bad” behavior. I perceived a dog who bit as being “tough and mean,” rather than afraid, as he probably was. And I had no clue that wearing a prong could worsen his fears, as well as exacerbate the risk of aggression against the other dogs.
“My Dog Is Different”
Even after I had grasped the rudiments of operant learning, I figured there must be exceptions to the basic principles. I didn’t understand the breadth and depth of behavior science. I would swear up and down that a certain behavior that one of my dogs performed regularly was not getting reinforced. Or I would search for things that “didn’t work” as predicted. When I thought I’d found one, I had this victorious a-ha feeling: I just knew my dog was different! I had a grand old time going around saying how this didn’t work and that didn’t work.
“I’ve Got It Figured, and It’s Not What They Think!”
A little learning is a dangerous thing, at least for some of us. At some point, I was convinced that I had realized things that almost no one else had. I had it all figured out. And I found support from the iconoclasts: I would glom onto scholarly articles that didn’t say what the people who circulated them thought they did, opinion pieces by critics of behavioral science who didn’t understand it, and arguments that had elements of truth that had been falsely generalized.
I distrusted expertise in behavior analysis and figured that the iconoclast du jour had found a loophole. The trouble was that I didn’t (and still don’t) know the basics well enough to be “proving” that there were exceptions. I’m not saying there’s no nuance to the science; I’m saying that when I perceived something as exceptional I was merely mistaken.
“Not So Fast!”
This is not exactly a phase, but more of a common setback during crossover. I remember an example from a conversation with a force-free trainer friend. We were talking about a talented young agility student. He was taking lessons with a local so-called balanced trainer known for her rough handling. I mentioned this to my friend, and she said, “Oh I hate it when children are taught to hurt their dogs.”
I flinched, bigtime. Even though I had reported this development as bad news, I wasn’t ready to hear blunt language about hurting a dog. I had trained punitively only for a short time, never liked it, and had quit more than a year before. Still, I reflexively defended the rough trainer. Why? I don’t even know.
I Made It Anyway
Despite spending so much time arguing and looking for loopholes, eventually I “got it.” I started seeing how behavior is a map of what is reinforcing. I perceived the fallout of aversives. I learned about competing reinforcers and saw how, when a method failed, one common reason was that there was reinforcement coming from another source. I realized that training involved mechanical skill, and that without it certain methods wouldn’t work well. Failures became easier to analyze. I learned enough about canine body language to see the obvious differences in the demeanors of dogs trained primarily with aversives and those who weren’t.
What Pushed Me Off the Fence?
What helped? Seeing more and more examples of positive reinforcement working. Learning about the theory. Learning about the role emotions play in behavior. Repetition.
But it may well have been agility lessons with an excellent teacher that finally helped me turn the corner. Human neophytes in agility see their dogs go the “wrong” direction or take the “wrong” obstacle again and again. Without the influence of good teachers or other resources, students blame the dog. My teacher challenged my assumptions repeatedly, with gentle but inexorable logic.
If I claimed that my dog “defied” me by taking the tunnel, my teacher would remind me how much I had reinforced tunnel work. If I complained that my dog turned in front of me, my teacher would point out that I had slowed down just before she turned. When I thought my dog took a “random” direction, my teacher would instruct me to look where my own feet were pointing. She showed me over and over that my dog’s seemingly inexplicable behavior was usually a direct response to mine.
Most important, she helped me figure out what would motivate my dog more than the wildlife on the other side of the fence. Seeing my dog’s interest level change from, “This is okay if there is nothing else to do” to “Please, oh please, let’s play again!” showed me the true power of positive reinforcement.
This article was originally published in the Summer 2014 issue of Barks from the Guild under the title “The Crossover Client” and was edited for the magazine by Kiki Yablon.