Probably most of the world’s myths are about supernational beings. Second on the list would come myths about dogs. Third on the list are myths specifically about clicker training or progressive reinforcement training as defined in this piece by friend and colleague Emily Larlham.
We frequently talk about dog myths on the dogster blog. Today I’d like to talk about positive training myths, one in particular. Here are just a few of the many myths about positive reinforcement training:
- dogs trained with positive reinforcement are fat
- dogs trained with positive reinforcement exhibit unreliable behaviors
- you will always need to carry food or a clicker
- clicker training works for “easy” dogs but not “real” dogs (?!), tricks but not performance level behaviors, tricks but not behavior problems
- clicker training is a fad
- clicker trainers are “treat slinging weenies” who provide no consequences for undesirable behavior
I could (and probably should) spend a week devoting a blog entry to each of these myths on the dogster B & T blog. For today, I’d like to focus on the last myth, that of clicker training and consequences for undesirable behavior.
First, there are times when the best response is no response at all. Call it reverse psychology if you will, but when a dog is offering undesirable behavior as an attention-seeking device, wouldn’t it make sense that the most effective way to control that behavior would be through manipulating access to attention? Using extinction (the principle of non-punishment, non-reinforcement) is not without side effects. Extinction bursts are common, and can involve a lot of frustration and sometimes even anger and aggression in dogs. However, implemented correctly, this technique can be very effective in treating demand behaviors.
Extinction is by no means the only tool in the clicker trainer’s arsenal for dealing with unwanted behaviors. Sometimes we train alternative, incompatible behaviors instead. My Saint Bernard Monte, for example, used to like to hump people’s legs when anyone would hug in our house. I needed to think of something I wanted him to do instead, a behavior he could not do at the same time as humping. I taught him to go lie on a mat whenever people would hug. Problem solved.
But do clicker trainers ever use aversives? In an effort to avoid getting into conversations about quadrants and technical definitions of punishment, I’ll simplify the matter – do clicker trainers ever expose dogs to situations, stimuli, or experiences which would be stressful to them?
The answer, at least in my experience, is yes.
Take my Saint Bernard puppy, for example. Cuba is a very well-socialized six month old Saint Bernard. He loves other dogs so much that when he sees a new dog on leash, he pulls to the end of his leash, barking, front feet off the ground. No part of this behavior is aggression, it’s simply unrestrained excitement signaling a need for more work on impulse control. (Teenagers, impulsive? Surely you jest!)
When Cuba does this, here is the protocol we follow:
- I say Cuba’s name once, in a happy, neutral tone of voice.
- If Cuba gives me eye contact in response to the cue (within three seconds), I reinforce the eye contact by presenting a hand target in heel position.
- If Cuba targets my hand, the reward is that we move closer to the other dog.
If Cuba does not respond to his name or the signal for the hand target, I begin applying slow, light, and stead pressure through the leash, stepping backwards. I wait until the leash is loose and Cuba starts turning to look at me, cue the hand target behavior, and again reward him by moving closer to the other dog.
While I’d like to say that I only live in the realm of positive reinforcement and negative punishment, in this case I actually used:
- positive punishment – when Cuba failed to respond to his name cue, slow and steady pressure was applied to the leash as a consequence