So here's the science, and then I'll give some examples. Positive reinforcement means adding something to increase the likelihood a behavior will occur in the future, while negative reinforcement means taking something away which will increase the likelihood a behavior will occur in the future. Notice the word negative in there? In the new words I'm suggesting, Positive reinforcement is "Adding to Increase" and negative reinforcement is "taking away to increase." See? Makes way more sense that way.
Examples. Well, Positive reinforcement (or, +R, to make my typing life easier) is an easy one. Rufus the dog is running around. I give the discriminative stimulus (or, the thing I WANT to control the behavior, shorthand as the sD): "Sit!" Rufus sits. I give him a treat. By treating directly after the behavior we are increasing the likelihood that the dog will sit again in the presence of the word sit. Got it?
Now for -R. Rufus has the urge to use the bathroom. He is circling and sniffing and whining at me. I direct his head to ring a bell tied to the doorknob, open the door, and we go out and he goes to the bathroom and feels better. After practicing this a few times, he needs to go. He rings the bell, I open it the door, and out we go. Confusing? The first thing in the chain of command is Rufus' urge to use the bathroom. Much like taking ibuprofen for a headache, Rufus wants to do something to end his urge. The sD is the bell. He learns that by ringing the bell, he is TAKING AWAY the urge to go to the bathroom. So he rings the bell, the urge to eliminate goes away (because he in fact gets to eliminate) and thus the likelihood that the bell will be rung increases. -R is the most difficult of the four options to understand, and actually not utilized much to change behavior. But if that makes at least a little sense, we'll move on.
Time for punishment. It's the same thing as R, only reversed. We want to Add something to decrease a behavior (+P), or take away something to decrease a behavior (-P).
+P: Sally the dog is chewing on the table. We have tried many methods to get her to stop, but introducing reinforcement when she is NOT chewing (my first choice) is not successful. Sally's owner Mary sprays "bitter apple" spray on the table leg. Sally hates the taste and stops chewing. So, we've added bitter apple, and decreased chewing. Adding to decrease behavior.
-P: Sally the dog is at it again. Now she's growling in the presence of her food. We've tried hand feeding, and some other positive methods, but now we need to do something more drastic. We are intrinsically reinforcing the growling by allowing her to eat while growling, so that has to stop. This example will show how -P and +R often work as partners.
Sally growls. We are sure she will not bite, so there is no danger and no need for protective gloves or body blocking. We never want to jump in to growling if it turns into aggression without proper precautions, but we also do not want growling that is not YET aggression to continue.
First, we're going to wait for her to growl. Sounds crazy, I know, but we want the behavior we want to change to be really clear to Sally. As soon as Sally growls, we take away her bowl. She stops growling (in fact she might even whimper or try to chase the food.) This is the -P part of it. We are removing the food to decrease growling. Now of course we want to teach Sally what she SHOULD be doing. When she is quietly sitting, we reintroduce the food and perhaps praise or pat Sally as she eats quietly. This is the +R. We add food so that the quiet behavior is more likely to occur again in the future. We repeat these actions again and again until Sally can eat an entire meal without growling.
Make sense? If not, ask questions, and check out more examples (Google it!). I will be using terms like +/- R and +/-P as I write about behavior and training, so refer back here to get a clear idea of the examples used.
One more post to come, about owning a dog.