The Sad Truth about Negative Reinforcement

You love your horse. Why in the world would you use negative reinforcement in his training? It sounds horrible!

Yes, it does, which is too bad because negative reinforcement isn’t horrible at all. In fact, you are probably already using it. The worst thing about negative reinforcement is its name. But more on that later.

Here’s the real meaning of negative reinforcement: rewarding a desirable response by removing an aversive (unpleasant) stimulus. Read that sentence again. “Negative” means a stimulus is removed. “Reinforcement” means that the horse sees this as a reward. Negative reinforcement is used to increase the likelihood that a given behavior will be repeated.

In plain old English, it’s what we call pressure and release training. Any time a trainer puts pressure on a horse and releases the pressure when the horse gives an acceptable try, the trainer is using negative reinforcement.

But back to the name. Negative reinforcement is a scientific term coined by American psychologist, B. F. Skinner (1904-1990). Skinner used the labels “positive” and “negative” because they invoked the ideas of adding and subtracting stimuli in response to behavior demonstrated by a subject.

Negative reinforcement has an identity crisis today because the public doesn’t use positive and negative this way. To us, these words are synonymous with good and bad. We say, “She has a positive attitude” or, “The movie got negative reviews.”

It is truly ironic that Skinner’s naming convention made it seem like rewarding the horse was really punishing him. This is an epic fail! It’s also a bit sad. Skinner was absolutely brilliant and is considered the father of Behaviorism, one of the big three learning theories of the 20^th century. But words matter and he was a lousy wordsmith.

Behavior Replacement

Extinction of behavior takes time to work. Today I’ll share some thoughts on speeding up the process through behavior replacement.

First, a quick review: With operant conditioning, “bad” behavior can be modified through punishment (creating an unpleasant consequence) or extinction (eliminating the reinforcement). Punishment is a perfectly good tool in the hands of a confident and experienced trainer, but many horse owners make matters worse through half-hearted or poorly timed responses to their horse’s naughtiness. Extinction may work better for them if they can figure out how the horse is getting rewarded and eliminate that.

Still, it takes a while for unrewarded behavior to die out completely. To speed the process up, get your horse working on something good, or as John Lyons might say, replace the bad behavior with good behavior.

For example, suppose your horse fidgets and paws the ground when he’s tied. Here are some possible responses from you:

1. Give him food to distract him. Result: behavior is reinforced and continues.
2. Scold him or spank him. Punishment requires perfect timing, consistency, and intensity in order to work. Not possible here. Result: the extra attention works as a reinforcer. If you frighten the horse in the process, you’ve created a new problem.
3. Ignore him completely. This is how to use extinction. When there is no reward of any kind, he will eventually stop the behavior.
4. Ignore the behavior, but immediately put the horse to work doing something else, something useful such as backing or sidepassing or sending. Maybe you just focus on getting the horse to move one body part at a time. It doesn’t matter what you do as long as it is positive and you take every opportunity to praise the horse for doing it.

This last option is a way of supercharging the extinction principle by replacing bad behavior with good. Thanks to John Lyons for teaching me this, and for teaching me to always have a positive job ready for my horse to do, just in case I need it.

Operant Conditioning

Operant conditioning, formulated by behaviorist B.F. Skinner, is one of those subjects fraught with misunderstanding, and even those of us who understand it sometimes misspeak, due largely to the uncommon use of common words.

The primary principle of operant conditioning is that when behavior is reinforced, it is more likely to be repeated and when behavior is punished, it is less likely to be repeated. The way we reinforce or punish is by adding or taking away stimuli.

1. Positive Reinforcement means that something pleasant is added to encourage the desired behavior. Example: The horse is given a food treat when he responds correctly.

2. Negative Reinforcement means that something unpleasant is subtracted to encourage the desired behavior. Example: Pressure is released when the horse complies with a request.

3. Positive Punishment means that something unpleasant is added to discourage the undesired behavior. Example: A horse is spanked instantly when he kicks out under saddle.

4. Negative Punishment means that something pleasant is subtracted to discourage unwanted behavior. Example: The trainer withholds a food treat that a horse wants.

These are called the four quadrants of operant conditioning. (Extinction is sometimes called the fifth “quadrant.” I’ll address that another time.)

Unfortunately, keeping all of this straight is next to impossible for most people, including me. The reason is simple. All four of the key words have narrower meanings than we usually give them. Positive means added (rather than good). Negative means subtracted (rather than bad). Reinforcing means rewarding (rather than strengthening) and punishing means deterring (rather than paying back).

These may seem like small distinctions but they can result in big misunderstandings. For example, I often hear trainers use the term "negative reinforcement" to mean creating an unpleasant consequence to a behavior. This, of course, is not what it means in the world of operant conditioning.

My advice is pretty straightforward. Use the principles of operant conditioning but be careful of the terms. They confuse more than they clarify. And when you hear someone else using them, remember, chances are very good that they are using them incorrectly. Better to clamp your hands over your ears and hum the theme to "Gilligan's Island," or at least think about something else until the moment passes. Like maybe lunch. More next time.