Positive reinforcement, negative reinforcement, positive punishment, and negative punishment are all components of operant conditioning. Operant conditioning was a term coined, in 1937, by the American psychologist B. F. Skinner. It is a type of simple learning based on the consequences of acts; the strength of a behavior is modified by reinforcement or punishment.
It is important to note that trainers, such as Konrad Most, a German police/military dog trainer who published “Training Dogs: A Manual” in 1910, were using operant conditioning principles 30 years before Skinner began his research. Some circus trainers were using the same techniques 100 years before that. It is equally important to realize that a great many researchers (W.C. Azrin and N. H. Holz among many others. See references) have subsequently expanded our understanding of reinforcement and challenged some of Skinner’s conclusions and some studies have shown that positive reinforcement and punishment are equally effective in modifying behavior. Research on the effects of positive reinforcement, negative reinforcement, positive punishment, and negative punishment continues to this day. As much as some people may believe, dog training is not an exact science. It is fundamentally based on human psychology, which is far from an exact science. Who can adequately explain a Jeffrey Dahmer? Who can explain the 8-year-old child who can write a complete symphony without ever having learned to play a single musical instrument? The “science of dog training” is constantly in flux.
So what is positive reinforcement? Reinforcement makes a behavior stronger. When it is positive, such as using food, praise, etc., it encourages behavior. The dog becomes more motivated to perform the desired command because of the potential of receiving the treat. The treat represents something being “added” and is known as a reinforcer.
There are two categories of reinforcers: Primary and secondary. Primary reinforcers are reinforcers that are related to biology-food, drink, sexual contact, etc. Secondary reinforcers are reinforcers that can be related to social conditions. Dogs, being social animals, respond well to praise, smiles, toys, attention, pats, etc. Secondary reinforcers have to be paired with primary reinforcers to become reinforcing (Classical conditioning). For example, the sound of a clicker followed by a treat. Eventually, the clicker sound becomes a conditioned (learned) reinforcer.
Reward scheduling is also an important concept when applying positive reinforcement. There are a myriad of reinforcement schedules. The following is a list of some of them:
Variable ratio schedule
New schedules are being researched to this day. The schedule that appears to be most beneficial to dog training (producing the slowest rate of extinction) is variable-ratio reinforcement. This is behavior that is reinforced after an unpredictable number of times. This unpredictability keeps the dog engaged, motivated, and keeps his expectations high.
It should also be noted that during a training session, such as obedience training, the reward should be given immediately after the behavior has been performed and the marker (clicker sound or vocal “yes”) is given.
Negative reinforcement occurs when a behavior is followed by the removal of an aversive stimulus. This causes the behavior to increase. An example of this would be pinching a growling dog’s ear until he stops growling. When the growling stops so does the pinching. The pinching, being taken away, is negative reinforcement. The terms positive and negative, in regards to operant conditioning, do not have “good” or “bad” connotations. They are more of a mathematical expression. When positive is implied it means something has been “added” to modify behavior. When negative is implied something has been subtracted, or removed, or avoided to modify behavior. Both positive and negative reinforcement increase behavior.
What do we mean by punishment? By B. F. Skinner’s definition of punishment, punishment is the reduction of behavior by the application of an unpleasant stimulus-i.e. “positive punishment” or the removal of a pleasant stimulus-i.e. “negative punishment”. An example of positive punishment would be grabbing and shaking a dog by the scruff of his neck for barking. An example of negative punishment would be removing a dog’s favorite toy because he is becoming possessive over it. Skinners’ definition of punishment is restrictive in the sense that if the negative behavior of the subject does not decrease, it is not considered punishment.
The difference between reinforcement and punishment is that reinforcement increases a response whereas punishment is designed to weaken or cause the extinction of behavior. There is a very fine line between some of these concepts.
Negative punishment has its use in dog training. This type of training teaches a dog that there are consequences for his acts. Above I used the example of removing a favorite toy. Another example would be teaching a puppy not to nip when playing. The moment he bites too hard playtime should immediately stop. This is the removal of a pleasant stimulus. Puppies love to play and he will soon come to realize that biting causes playtime to end and eventually he will learn that biting is not in his best interest and the behavior will be eliminated. A mother dog teaches her pups not to bite in a much more straightforward and quick manner. She uses positive punishment. When the puppy nips too hard she immediately nips the puppy in return. It’s like touching a hot stove. The puppy doesn’t forget the lesson his mother just gave him. In the real world of animal behavior, animals use positive punishment all the time. As human trainers, we should avoid using positive punishment because we lack the behavioral instincts to use it effectively. When humans use positive punishment many problems are likely to develop. Fear is one of the most common behaviors that develop when positive punishment is used. A fearful dog is very likely to turn into an aggressive dog (fear-based aggression). Punishment (as humans apply it) does not teach the dog what you want him to do, it only teaches him what not to do. So it is useless as a teaching aid.
In addition, while the behavior that you are punishing will cease during, and for a time after the punishment is given, in the long term, the behavior is very likely to return.