What is the difference between reinforcement and punishment?
Essentially, reinforcement reinforces a behavior meaning a horse learns to repeat the behavior.
Punishment is meant to stop a behavior.
Punishment (P+ and P-)
If a horse tries to bite you and you whip him and he never bites you again, then your whipping was punishment. Specifically “Positive Punishment” because you GAVE the horse an AVERSIVE to END a behavior. It might not sound “positive” but the term is used because you ADDED something in the math sense.
The abbreviation for Positive Punishment is P+.
Any infliction of pain or hunger or discomfort is P+ if it succeeds in ending a behavior. If it doesn’t succeed in ending a behavior, it’s just pure pain or discomfort.
Negative Punishment, on the other hand, is when the rider REMOVES something DESIRABLE to END a behavior.
The abbreviation for Negative Punishment is P-.
Please note that a common misconception is that “negative” in this sense means “bad.” Not true. Instead, “negative” is meant in the math sense which is “to subtract.” As in “take away something” to punish.
In human terms, grounding your teenager for misbehavior would be Negative Punishment. An example of Negative Punishment in the horse world might be to remove a horse from its herd-buddy in order to have the horse stop fussing. As you might imagine, Negative Punishment doesn’t have a great track record because the student involved doesn’t always make the connection between the misbehavior and the removal of something it likes.
Both our lives and our horses’ lives are filled with punishments both positive and negative, so you can’t say that punishment isn’t “natural.” Negative punishment runs the gamut from five-minute time-outs to life in jail. Shunning, banning, grounding, and the cold shoulder are all negative punishments. Positive punishments are even more extensive: spanking, kicking, hitting, slapping, yelling, arguing…the list goes on.
Punishment, when applied at the right time in the right way, certainly can end a behavior.
The thing punishment CAN’T do is TEACH the desired behavior.
Let me say that again: on their own, neither Positive Punishment nor Negative Punishment teaches the student what behavior they should have done INSTEAD.
When punished, the student may choose alternate, unwanted behaviors such as escape, avoidance, or striking back. Punishment turns into a win/lose situation instead of cooperative learning. Most important, Punishment usually leads to Unintended Consequences.
The list of Unintended Consequences is as long and varied as there are individuals and situations on earth.
If a child is grounded for getting a bad grade on a test in school, the child might–instead of the preferred behavior of studying–try to cheat on the test the next time.
If a dog is hit on the head when it jumps up on visitors, it may start to shy away from petting.
If a horse is too sharply corrected by the bit, the horse may begin to evade the bit, bolt, rear, or even buck.
One of the things I’ve noticed with horses who are frequently punished is that they often don’t relax around their handlers. That is certainly an Unintended Consequence. People get so used to working with fearful horses that their anxiety becomes accepted as normal. Horses who never feel secure because they’re literally waiting for the next shoe to drop are thought of as “spirited” or “hot”. Handlers don’t ever see them in a calm mood because every time the handler is around, the horse gets nervous again.
Reinforcement (R+ and R-)
Reinforcement also comes in two flavors: Positive (R+) and Negative (R-).
Positive Reinforcement is fairly straightforward: the horse gets something of value after it does a behavior which teaches it to repeat the behavior. Please note that “something of value” needs to be something valuable to the horse, not necessarily what the handler thinks the horse might like. Food, security, and rest are usually prime motivators. Petting, not so much. Studies have shown that human touch either is an acquired taste, or something some horses have learned to put up with.
Negative Reinforcement, like Negative Punishment, doesn’t mean “bad”–it means to “take away” something aversive when the horse does a behavior which teaches it to repeat the behavior. Many horse people simply refer to Pressure/Release when they mean Negative Reinforcement, but (R-) covers more than Pressure/Release.
The classic example of Negative Reinforcement would be pulling on a halter rope to put pressure on the horse’s head to make the horse go forward. Once the horse steps forward, the pressure ceases and the horse learns to avoid or end the pressure by performing the behavior.
In theory, reinforcement should not be as susceptible to Unintended Consequences as punishment is because the rider or handler is focused on teaching a horse what to do instead of leaving that question unanswered.
It’s my experience, however, that Positive Reinforcement has the potential to go wildly out of control with an animal who becomes overly enthusiastic and starts offering all manner of alternative behaviors, especially if the handler is distracted or inexperienced. Successful training with positive reinforcement comes down to managing the circumstances to discourage those alternatives.
In some ways, some types of Negative Reinforcement are a little more controllable. If a horse feels pressure from the bit and moves away from it, the pressure was successful. Much of Natural Horsemanship is based on negative reinforcement. NH trainers argue that a horse’s herd life is heavily based on Negative Reinforcement-based communication–horses nudge each other and drive each other off all the time.
The thing to keep in mind is that if a horse experiences an aversive which is painful in such a manner that the horse connects the pain with the handler, that pain–in the horse’s opinion–becomes Positive Punishment inflicted by the handler. The horse may still perform the desired behavior, thereby ending the negative reinforcement, but it’s reluctance to do so signals that the horse may not forget that perceived punishment and seek to avoid it in the future.
Or, if it’s a mule, hold it against its handler like…for the rest of its life. Mule training tales are full of anecdotes where a mule waited 20 years for the chance to get even.
Horses seem to be more forgiving of pain which they perceive is a random consequence of their environment. But they are incredibly suspicious and can connect pain to completely unconnected spaces or actions which often turns into an Unexpected Consequence.
Example: Horse is being ridden in an indoor arena and snow is sliding off the roof outside. The horse sidesteps at the same moment it hears the snow and the rider promptly corrects the horse by sharply kicking it with the inside leg and bumping the bit. The second time around the ring, the horse anticipates the rider’s kick when it hears more snow on the roof, and from then on, the horse can’t concentrate on the lesson because it’s preoccupied by the sound of the snow. The horse has learned that the sound of snow on the roof causes pain. The rider, who may not have heard the snow, doesn’t know what’s going on and may assume it’s a “bad” horse.