A few weeks ago, I met a friend who worked in the USA for 30 years as a master dolphin trainer. I have known that unlike in the case of dogs or horses you can’t use a leash or a bridle on an animal that just swims away. So, I was curious to know about the methods he used to train dolphins.
He answered, “We use what is known as Positive reinforcement, primarily through buckets of fish. It’s our only tool.” Explaining further he said, “It’s not just reward or punishment. These usually come after an act is completed. Reinforcement happens during or immediately upon the conclusion of the behaviour the trainer wants to affect. For example, we blow a whistle once in three hours during daytime to indicate feeding time. That sound becomes a reinforcement by itself. Dolphins work non-stop until they hear the whistle which signals their immediate reward for working three hours”.
Two types
In Applied Behaviour Analysis, there are two categories of reinforcement: positive and negative. A positive reinforcement is something the subject wants, such as, special food, petting or praise. A negative reinforcement is something he wants to avoid – grounding in room for a couple of hours, a small cut from weekly allowance.
Suppose you want someone to telephone you, if he doesn’t call, there isn’t much you can do. You cannot reinforce behaviour that is not already occurring! But, when your friend does call, you act delighted, chances are he will call more often. If, instead, you scold him, you are applying the negative reinforcement of punishment - training him not to call at all.
Relative not absolute
My friend added, “Several years ago at a Zoo in New York city, keepers couldn’t get a particular gorilla out of an indoor cage when it needed cleaning. The gorilla just sat in the doorway; when keepers waved bananas enticingly, the ape ignored them or snatched the food and ran back inside. I pointed out that banana waving was an attempt to reinforce behaviour that hadn’t occurred yet. I suggested to ignore the gorilla but reinforce it with food whenever it did happen to go out by itself. Within a few days, things become better”
Reinforcements are relative, not absolute. It must be something the subject wants. Rain is a positive reinforcement to ducks, negative to cats. Food is not a reinforcement if you’re full. Thus, you need a variety of reinforcements in a given time.
My friend said, at the oceanarium he worked, the killer whales are reinforced in many ways - with fish, with stroking and scratching, with social attention, with toys. Whole shows are run in which the animals never know what behaviour will be reinforced next with what. The “surprises” are what challenge the performers.
Information
A reinforcement is information; it communicates. It tells the subject exactly what it is you like. In coaching athletes or training dancers, the instructor calls out, “Yes!” or “Good!” to mark a movement as it occurs. That truly gives the needed information - not the debriefing later on in the dressing room.
Watching football on TV, I’m often struck at the beautifully timed reinforcements the players receive. When a goal is scored, there’s a frenzy of mutual reinforcement among the players - plus the roar of the crowd.
It’s quite different for onstage actors, with the applause coming long after many dramatic scenes. Reinforcing too late is a real problem. If you tell your wife, “Heavens, you looked great last evening at the party,” she may ask you back: “What’s the matter, don’t I look great now?”
Timing is also important with negative reinforcement, such as, when parents or teachers nag a child. If the nagging doesn’t cease the instant the desired results are achieved, it is not reinforcement and it is not information. In communication theory, it is “noise.”
Variables
Continued reinforcement is needed only in the learning stages. Teaching your son to ride a bicycle may require a stream of “That’s right, steady now!” But your child would think you were crazy if you went on praising him once he had really learned to ride.
To maintain an already learned behaviour, use reinforcement only occasionally.
This is what psychologists call a variable schedule. Random, unpredictable reinforcement is far more effective than a continued, predictable schedule.
My friend puts it succulently. “If every time a dolphin jumps I give it a fish, very quickly the jump becomes as formality as the animal can get away with. If I then stop giving fish, the dolphin quickly stops jumping. But if I reinforce jumps at random, the behaviour is much more strongly maintained. This in turn allows me to selectively reinforce the more vigorous jumps, thus shaping improved performance.”
Real Life
Reinforcements occur all the time in real life, often by coincidence. Find a Rs.500 note in a trash basket, and I challenge you to walk past that basket the next day without looking it over closely.
If, while taking an exam, you happen to be chewing a pencil when you get a right answer, this will reinforce pencil chewing. Many behaviour rituals arise from such accidental reinforcements.
You can also use positive reinforcement on yourself. I once met a lawyer who is an avid squash player and shouts to himself, “Well done old boy. Once more, please!” for every good shot. He says, it has improved his game tremendously since the days when he cursed himself for his errors all the time.
But reinforcing ourselves is something we often neglect. Either it doesn’t occur to us, or we demand too much of ourselves. As a result, we often go on for days without let up, moving from task to task, unthankful to ourselves!
You should reinforce yourself in healthy ways -a walk, a talk with friends, some self-approval. I like the suggestion made by one of my ex-managers: “ I need compliments when I achieve something great. If I go long enough without getting the compliment, I commend myself, and that’s just as good, because at least then I know it’s sincere.”