Positive Training Debates

The other week we were walking along the reservoir and came across one of my favorite dog friends. While I played with the lab-mix and a new fox terrier friend, the people got into a conversation about training.  In a discussion about different training schools the other dog owners expressed disdain for positive training techniques. The quote of that particular day was as follows: 

Dog Owner 1 "What is positive training?"
Dog Owner 2 "It's where you learn not to beat your dog. That positive stuff doesn't really work.

The irony is that every training program I have ever seen for dogs are all based on Operant Conditioning which involves both positive reinforcement, punishment, and a whole lot of other stuff. The differences are in packaging and the tools used to implement the training plan.

B.F. Skinner uses the term Operant Conditioning to describe the effects of the consequences of a particular behavior on the future occurrence of that behavior. The components of operant conditioning are (1) positive reinforcement, (2) negative reinforcement, (3) punishment, and (4) extinction.

Huh? You almost need a doctorate in psychology to understand all this stuff. Good thing my owner has one.

So let's look at the individual terms.

Reinforcement is a consequence that causes a behavior to occur with greater frequency. Punishment is a consequence that causes a behavior to occur with less frequency. Extinction is the lack of any consequence following a behavior. When a previously reinforced behavior is no longer reinforced with positive or negative reinforcements, it leads to a decline in the response.

To fully understand this, there is another level of detail I need to tell you about. Reinforcements can be positive or negative. It's not what you think though.

Reinforcement can be positive or negative. 

A positive reinforcement is the giving of a pleasant even contingent on a behavior with the goal of increasing the likelihood of the behavior in the future. This happens because of the addition of a stimulus immediately following a response.

e.g., Giving me a piece of food when I sit down is an example of positive reinforcement as it increases the likelihood of me sitting the next time I am asked to do so.

A negative reinforcement is the removal of an aversive (unpleasant) event contingent on a behavior with the goal of increasing the likelihood of the behavior in the future.

e.g., Letting me out of my crate (removing a stimulus) the instant I stop crying reinforces me not crying in my crate.

Reinforcements can be primary or secondary. 

A primary positive reinforcer is something that an animal doesn't have to learn to like. I personally am a fan of these little training bits my owner uses. I'm also a huge fan of playing with tug toys. They are very rewarding to me on their own and thus excellent primary reinforcements.

A secondary positive reinforcer is something an animal has to learn to like. The clicker, for example, is a secondary reinforcement. I originally had no idea what the click was all about. However, after my owner spent five minutes doing the following "click, treat, click, treat" I got the picture. I learned that the sound of the click was something pleasurable. I don't always get food with the click now. I get a the food reward on an intermittent basis.

Punishments can also be positive or negative.

A positive punishment is the giving of an aversive event contingent on a behavior with the goal of decreasing the likelihood of the behavior in the future. In other words, a positive punishment is something that is applied to reduce a behavior. Please note that this is about the behavior, not the animal. In operant conditioning punishment is NEVER about the animal. It is ALWAYS about the behavior.

e.g., A dog receives an electric shock from their collar after barking more than three times. The positive punishment is the electric shock because the addition of this stimulus reduces the likelihood of barking more than three times in the future. 

A negative punishment is the removal of a pleasant event contingent on a behavior with the goal of decreasing the likelihood of the behavior in the future.

e.g., A puppy is playing tug with his owner and bites the owners hand. The owner takes the toy away and stops play (removes play when bit) and thus decreases the likelihood of the biting behavior in the future.

Now pay attention, this part is important.

Positive reinforcement is the key to producing desired behaviors. B.F. Skinner (the "father" of operant conditioning) wrote that people "work harder and learn more quickly when rewarded for doing something right rather than when punished for doing something wrong."

Punishment doesn't provide the direction that a reward does. In other words, punishment can let a dog know that behavior is not desirable but it does not teach the dog an alternative behavior. A punishment has to continuously and consistently be applied to control behavior. A behavior that is rewarded on an intermittent basis tends to be the behavior that is most enduring and lasting.