|
My journey to all positive reinforcement
training.
Like quite a few in this generation of "Truly Dog
Friendly" trainers, who use only positive reinforcement training, I
started out as a typical traditional trainer of this era, using a
mixture of corrections and lure-and-reward as the main way of
"teaching" my dog new behaviors.
Unlike some of my trainer friends who made this
transition, however, I did not have a huge epiphany that caused a
sudden transformation to "see the light" and discard corrections
from my tool bag one day. Instead, I experienced a number of
small epiphanies--light bulb moments--and the better I got at
positive reinforcement training, the more I could see that it worked
better than anything I'd tried before and the less I used
corrections.
I began dog training late in life, when I got my
first dog as an adult in 1994 (Kallie--her
story here). I had a lot to learn about dog training and I
immediately enrolled Kallie in a pet dog class, which was even a bit
progressive at the time because the instructor encouraged us to use
food rewards. We had, however, been issued 6' leather leashes
and a choke chain as our foundation training tools.
I came to dog training at this point from a long
career (25 years and counting at that point) as a professional
biologist, and as a scientist with expertise in the field of animal
behavior. I was taken aback often by the degree of
superstitious and anthropomorphic advice that I was given by all of
my various dog training instructors. Usually their remedy for
changing behavior after lure and reward did not work was giving the
dog a whale of a choke chain "correction," sometimes trying to knock
the dog off its feet. I was impressed with how poorly that
method seemed to work (in other words, I couldn't tell any
difference before or after), but I was usually admonished that my
timing was bad, that I was too timid with my corrections (much
better to "get in" a big one and get it over with), and that I
wasn't very good at establishing who was "boss."
Overall, dog training started out as fun, but as
I progressed, it felt worse and worse. When I ended up
training dogs to herd, and was advised to throw heavy objects at
them, and slam them to the ground, I recoiled. I began to
become interested in electronic (remote) shock collars, because they
seemed to be an improvement over these other methods that required I
physically hurt and frighten my dog in crude, imprecise, and often
quite harmful ways (one instructor told about accidentally breaking
a dog's leg...).
In 1995 I decided to learn how to use a shock
collar, and I trained for two years with Jim Dobbs, an experienced
gun dog trainer, who'd worked with Tritronics for years to help them
field test and develop their shock collars. He was (still is)
an excellent trainer. The best part was that he used strictly
operant conditioning (negative reinforcement) in a precise,
unemotional manner and he rejected all of the superstitions I'd
been taught. He used instead well established principles of learning
theory. He also treated each dog as an individual and refused
to consider generalizations about a dog's breed as he approached
training that particular dog. I learned a great deal from him
and considered him to be the best trainer I'd worked with, by far,
until that time. He maintained a broad curiosity about dog
training and generously shared it with his clients. In the
last conversation I had with him, he told me how interested he was
in learning more about clicker training.
Because of Jim, I was able to approach dog
training in a neutral, objective, and scientific manner. I did not
have to swallow implausible superstitions and anthropomorphisms.
Why did I quit training with shock collars?
Mid-way into the two years that I spent learning to use shock
collars, I learned about the clicker and began to experiment with
it. Then I adopted Babe, the most
unreachable dog anyone could imagine. I tried to train her
with a shock collar, once--and it was such an abject failure, I
never tried that again. In fact, I began to understand how
negative reinforcement worked in a deeper way than I'd ever
understood before. When you use R-, the dog must have something
better to compare it to: Jim always said "It's a matter of
pleasure and displeasure" as he explained the use of the shock
collar to each new client. But if you have an animal that gets
absolutely no pleasure at all in being with you, or even in life,
then negative reinforcement cannot work. It's all bad. There's
no "better" alternative for the dog to choose.
Babe forced me to become an expert in positive
reinforcement training, and believe me, it wasn't easy. It was
like trying to train with one arm tied behind my back. Each time I
failed to change her behavior with positive reinforcement, and tried
a quick fix with positive punishment and negative reinforcement, my
experiment blew up on me. Babe was so incredibly difficult to
reach that each setback in the hard-won progress that I'd made with
her was too much to bear. So I stopped reaching for quick
fixes with punishment. Morgan Spector says in his book
Clicker Training for Obedience that the
use of positive punishment is always to make up for trainer
error. With Babe, and many times since then, his words have
rung truer than I'd imagined possible when I first read them.
Eventually, I compared the shock collar and the
clicker head-to-head. They are reinforcement analogs in one
sense, and that is that they are both unemotional and absolutely
precise. The shock collar provides a punishment that goes away
the instant that the dog performs a behavior that the trainer wants
to strengthen, and the clicker marks the instant that the dog
performs a behavior that the trainer wants to strengthen and
promises a reward.
At first, I considered the shock collar and the
clicker to be equally useful tools in my trainer's tool bag.
But as I got better at positive reinforcement training, I began to
compare the effectiveness of the two methods. In another light
bulb moment, I witnessed the stress my dog Megan began to exhibit
when we trained with a shock collar. One day her teeth were
rattling so hard I thought they would break. I remember Morgan
Spector's story of seeing how his sheltie's hair was worn off on one
side of his neck by the jerking of the choke chain, and Pat Miller's
moment of seeing her dog disappear under the porch when it came time
to train. Since then, I witness the diametrical opposite when
the clicker comes out. The dogs are frantic to train.
They are falling all over themselves, eyes glowing and wide with
anticipation. They are offering all of the behaviors in their
repertoire just to get a chance to get their turn to train.
Eventually I found that punishment of all kinds,
whether caused by a shock collar or other method, creates stress in
the dogs and is a blunt instrument for modifying a dog's behavior.
If you are happy with a dog that does nothing, which is what
punishment often creates, then you might not suffer from the side-effects
of its use (your dog still will). But if you want a dog that
performs behaviors willingly and eagerly with precision off leash,
as I do in my participation in dog agility, then there is no use
whatsoever for punishment as a tool. In one of my light
bulb moments, I read Karen Pryor's article on the Poisoned Cue in
the first issue of Teaching Dogs Magazine. In it she explains
the perils of mixing aversives with rewards, of mixing positive and
negative reinforcement.
Then, when you train your first all-positive-reinforcement dog, you realize what you were missing all along.
There is no comparison whatsoever in the results, contrasting the
traditionally trained dog with one trained with all-positive
reinforcement. And, there is no going back.
Read more:
What's wrong with using
punishment?
My current position on shock
collars.
|
