Monday, March 06, 2017

The Missing Animal at the IQ Zoo

From Atlas Obscura
comes a nice little article about Marian and Keller Breland's IQ Zoo:

Tourists sailing down the highways toward Hot Springs, Arkansas, in 1955 would have been filled with gleeful anticipation. Numerous resorts and roadside offerings were on offer to sate their recreational lust: They could drop into the Arkansas Alligator Farm and mingle with the toothsome reptiles, ooh and awe at celebrity likenesses at the Josephine Tussaud Wax Museum, or delight in the animated miniatures of Tiny Town. Or they could go to the newly opened I.Q. Zoo and watch Casey the chicken play baseball, a duck play the drums, and a rabbit dunk a basketball, to name just a few oddities.

I.Q. Zoo was the brainchild of a psychologist couple, Marian and Keller Breland, who not too long before had been working alongside the famous psychologist B.F. Skinner to train pigeons to pilot the first “smart bombs” for the United States government.

I have written about Marian and Keller Breland before, but I have not talked about the limits of what they did. More from Atlas Obscura:

As their success grew, so did the pool of animals they trained. They taught a reindeer to operate a printing press, they trained parrots to balance on soccer balls and rollerskate, goats to push baby carriages, and a cow to play the harmonica. They trained cats, raccoons, squirrels and even dolphins. Chickens remained a perennial favorite, they “did math”, walked on tightropes and played tic tac toe with visitors. Under the banner of their business, Animal Behavior Enterprises (ABE), they also sold coin-operated displays that housed trained chickens, and these were scattered throughout the country.

Do you see what is missing?

Marian and Keller Breland  did not train dogs.

Why not?

The answer is that Marian and Keller Breland's pure-positive food-based reward system worked fine when animals were confined, and the animals were being asked to either go with their natural instinct (such as asking a chicken to peck), or do some neutral trick.

But what if you are working "in the wild" with an animal that is a top-end predator?

Can food rewards alone work to stop prey drive?

The answer is too often NO, which is why you never hear of Marian and Keller Breland training dogs, or writing a best-selling book on dog training.

It's not that they didn't try; they knew that was where the money was.

It's that they tried and too often failed, with too many dogs.

Failure using pure positive rewards is pretty common.  As I have noted before, the great evangelist of pure-positive food-based dog training, Karen Pryor, uses an Invisible Fence to contain her own dog, and will not take her own Border Terrier off-lead because she cannot stop it from chasing squirrels.

The point here is basic:  Training prey animals like chickens, pigeons, and raccoons inside captive settings without distractions, is very different from training a top-end predator in an open-field setting with prey distractions.

The good news is that modern e-collars, set at settings so low a human cannot feel them, are enough to break through the Attention Deficit Disorder that plagues some dogs in a natural setting.

Unlike long lines, e-collars can give perfectly timed corrections at a great distance. The result can be seen with my two game-bred terriers, which walk off leash in my local parks, oblivious to squirrels, deer, rabbits, and other dogs.  The dogs are not zapped; in fact they do not even get a tap anymore, as they have been trained to verbal cues and now rarely need any more reinforcement.

Could Keller and Marian Breland control the prey drive of dogs off-lead, in natural settings, with only food rewards?  Apparently not often enough

And to be fair, it's a common enough problem; Rudd Weatherwax never got Lassie busted off of chasing motorcycles.

Dogs are extremely variable animals in terms of size, temperament, drive, and bidability.  Some dogs will "Velcro" to their owners without any training at all, and some are so calm that they will never jump on a person, bolt out the front door, or chase a cat. This is rare enough, however, that an entire profession surrounds the 97% of dogs that are the exceptions.

Can you get basic obedience with rewards-based training?

Absolutely.  In fact rewards-based training is core training for a great deal of what you want to a train a dog to do.

But dogs have epic Attention Deficit Disorder, and the drive to chase a cat, greet another dog, or follow a luscious scent upwind will, at some point, be stronger than the power of the kibble in your bag.

The result is a dog that is terrific, until it isn't.  This is a dog that knows to come, but sometimes "forgets". This is the dog that does a down-stay, but gets up and begins to slowly wander as soon as you turn your back or leave the room.

To get a dog that is truly "bomb proof" you need a "no" signal that is every bit as strong as any "go" signal they will ever receive.

Trying to train a dog with only positive signals is like trying to drive a car without brakes or a reverse gear; it may take you far, but without the confidence, control, or precision required for true safety.  And it will take you a lot longer in terms of time.

This last point is not small.  While a pure-positive dog trainer is charging by the hour, and too often guarantees little or nothing, the IQ Zoo needed to train a large number of animals quickly and economically. Putting prey animals in a Skinner Box with food rewards for specific actions achieved met their business goals; using off-lead predators trained with food rewards alone did not.

So what?  Is there a larger harm here than a million not-to-well trained dogs, and a few billion dollars wasted on a less-efficient method of dog training?

There is.

Every time we make dog training more expensive in terms of time and money, the more unhappiness we guarantee in the world of dogs and dog owners, the more tragedies occur, and the greater the number of dogs that end up in shelters or dead by the side of the road.

Whenever someone tells you they are all about "pure positive" dog training, remember that there is nothing "pure positive" about failure.

Dogs deserve training systems made for them.  To their credit, Keller and Marian Breland never claimed they could train off-leash dogs.  They simply advanced in another direction.


geonni banner said...

Hear, hear!

Miramis said...

Brilliantly written!

Viatecio said...

You hit a point with the note about training predator vs prey animals, but I would also add "domestic vs non" to it.

Conditioning a wolf in a zoo or a dolphin in a tank is done through the same free-shaping, reward-witholding and also reward-driven techniques as many of the prey animals. I mean, if we can train a killer whale to willing present itself for a urine sample and an orangutan to hold out its arm for a blood sample, teaching a dog to LIKE being poked in the butt with a cold thermometer should be peanuts!

The concept of the "hands-off" training techniques about which many "pure-positive" "trainers" crow is based off of these. In essence, we are treating a domestic species like the wild animal kept in a sterile, secure environment and then complaining (or suppressing it, as many owners are forced to do by these "trainers") when the conditioning fails in its many and spectacular ways.

And then, in the midst of all this "hands-off" stuff, we wonder why the dog dislikes being HANDLED or RESTRAINED in any way, and needs further conditioning (such as what I saw from the likes of the Fear-Free originators themselves) on how to allow themselves to be PET and TOUCHED, or how to teach them to INVESTIGATE something.

I'm not sorry. I live with my dogs in both my house and the great outdoors, with all the comforts and dangers therein. My hands are on my dogs all throughout the training process giving reinforcement, guidance and even mild aversion within the context of the lesson at hand. They gain confidence through control of THEMSELVES with minimal input from me once lessons are learned and proofed. We are a team that works in cooperation with each other, although I usually have the final say when something comes up; this point is accepted by all without fuss.

But go on with all the "positive" training that is supposed to change the world and MAKE THINGS BETTER.