Letter to
Ellen DeGeneres
The following was an
assignment for the
Dear
Ellen,
In
your American Express commercial you say: “My life’s ambition: to work with
animals...”
Well,
that’s what I do. I train animals, and it isn’t as hard as you
would think. The kind of training I’m
going to tell you about is a blast for the animal and a gas for the human! Ask
my dog Madison who loves to skateboard; or my Sunshine whose favorite game in
life is fetch-mom-a-beer-from-the-fridge (ok actually, it just sounds better to
say “beer”, she brings me bottled waters and Mountain Dew, but I could have
trained beer if I’d wanted...) Then there is my Toby who politely brings a
guest a tissue when they sneeze (when the guest laughs or gasps in amazement,
Toby exuberantly races around the room in a victory lap), or there is my Emma
who holds her paw up and wags her tail during nail clippings, or my hound Dylan
who will leave the trail of a rabbit in mid-stride when I call and come
bounding back to me. How do I train all of this? Why do my dogs LOVE the training? It’s because of a teaching method called
“clicker training.”
As
a trainer, the best part of clicker training is that it satisfies my “positive,
fun, creative” side as well as my “geeky, scientific” side, and, of course, it
gets results! Clicker training is based
on our scientific knowledge of how organisms learn. It uses the principles of something called
“operant conditioning”... most people vaguely remember that term from some
Psychology class they had in college.
Too bad they didn’t have Sunshine demonstrating the beer retrieval thing
in Psych 101, eh? Students would
remember the concept then! Good clicker
trainers know learning theory and how it relates to training.
To
explain, let’s start with something basic... (I like to call this example “sit
happens”.) Imagine your dog was standing
in front of me now. Pretend I said
“sit”, and he sat. But WHY did he
sit? The answer we often hear is
“because you told him to”, and in a certain way that’s true, but it isn’t
really the REASON he sat.
Clicker
trainers, who understand the theory behind learning, know why. We have to go back to this really old,
now-dead guy named Edward Thorndike.
Thorndike gave us a “law” that he called “the Law of Effect”. (Like the
“law of gravity” but no apples were harmed in the making of this law.) Eddie told us, in a fancy way, that “behavior
is a function of its consequences” and that “behaviors that are reinforced will
repeat themselves.” This is actually
very cool stuff... it’s the reason for
the behaviors all animals do! Think of
it this way. As animals evolve, they
have to learn what it takes to survive; they adapt, right? In evolution, animals keep behaviors that
work in their repertoire and lose those behaviors that don’t. It’s the same principle here. All animals do what works for them. Old Eddie’s fancy law that says “behaviors
that are reinforced will repeat themselves” just means that if you want an
animal to learn a behavior and to keep that behavior in his repertoire, you
have to reinforce it. Make doing that
behavior “work” for that animal! The
reason my dog Dylan leaves the trail of the rabbit is not because I, by nature,
am more interesting to him than a rabbit (though I am rather interesting...) Dylan leaves the rabbit because I made the
word “come” be a predictor of things so wonderful that he leaves the rabbit to
see what surprise I may have in store for him.
Leaving the rabbit and coming to me works for him, bless his heart. (And bless Oscar Mayer, because Liver Sausage
had something to do with it... oh and Karen Pryor, because she introduced me to
the “clicker”, and my mom for her love and support... and, never mind, I’m off
track here.)
So
now that you’ve learned about the impressive Edward Thorndike, take another
stab at my little scenario. Your dog is
standing in front of me. I say “sit” and
he sits. Why does “sit happen?” (hee
hee.) Because in the past, sitting has
“worked” for your dog!!! Traditional
training has placed lots of emphasis on what happens before the behavior. Often
in traditional training, a command is given and the animal is forced into the
desired position. Traditional trainers
say that this teaches the dog to sit. Clicker
trainers say “nope, not necessarily...”
We say that the stuff that comes before the action of sitting, though
important, is not where you get the bang for your training buck. We say that if you want to shift the
probability that an animal is going to give you a desired behavior, you have to
look at what will reinforce that behavior.
What is going to make this behavior work for this animal? We look closely at the animal and what would
be rewarding to him and we use the things he likes as reinforcers. We call this “positive reinforcement.” It’s a clicker trainers bread and
butter.
So,
where does the “clicker” come in? Ahhhh,
now we’re getting to the beauty of it...
this is where we marry Edward Thorndike and Dr. Doolittle (...though if
they lived in
Going
back to “let’s pretend...” (Cue dream music)
Suppose I wanted to train your
dog to sit. First I would teach him that
the click meant a reward is forthcoming.
I would find something that he loves - tiny slices of hotdogs maybe - and
“click” my clicker. I would follow the
sound immediately with the bit of hotdog.
This click/treat pairing usually only has to go on a few times before a
dog will whip his head around when he hears the sound of the click. Now I have
some training choices:
I might just stand around
without asking him to do anything and wait for your dog to sit on his own and
click and treat him every time he does.
After a few repetitions he would figure out that sitting PAYS and I
could introduce the cue. (Traditional
trainers called “sit” a command; clicker trainers call it a “cue”. It’s more than a semantical distinction I
think. The process in which cues and
commands are introduced, the perceptions about how and when they are understood
by the animal, and the reasons they are employed is quite different.)
OR
I might click and treat
tiny movements TOWARD the sit. The microsecond your dog does one of the
behaviors I’m looking for, I would “click” my clicker and swiftly deliver the
hotdog. Then I would wait again for him
to do that behavior again. The instant I
saw that behavior again I would click and treat. After a few times your dog would say “you
mean you like it when I do this?”
Click. In incremental steps I would wait
for behaviors that moved us toward the goal behavior of
place-your-tush-squarely-on-the-ground. When your dog is now enthusiastically
offering the sit, I would introduce the cue.
Whether
you know it or not, in the last six paragraphs I’ve just briefly explained how
clicker trainers influence behavioral probabilities using the principles of
operant conditioning via Thorndike’s Law of Effect; positive reinforcement;
learning theory; antecedents, behaviors, and consequences; the concept of an
event marker or bridging stimuli; a few of the fundamental differences between
traditional training methods and clicker training; capturing; shaping; cueing
basics, and successive approximations.
Sneaky little science geek, aren’t I?
Don’t you wish you took your Psych 101 class from me and Sunshine?
At
this point in the letter I would like you to be nodding enthusiastically and
rushing out to sign up all of your pets for clicker training classes. If you still need convincing I’ll be very sad
- and I haven’t trained Toby to bring me a tissue when I cry… only when I
sneeze. So do me a favor and when you
want to know how to make “sit happen”, look for a clicker trainer.
Patty
McKenzie