The process of teaching a behavior by getting a critter to avoid something is called aversive control. There are two types of aversive control, punishment and negative reinforcement. Punishment causes behavior to decrease, while negative reinforcement causes behavior to increase.
For example, imagine that your dog Fido chews a couch cushion while you are away from home. When you return and see the damage, you shove the cushion in Fido's face (to show him why he is being punished) and swat him with a rolled up newspaper. The next time you come home, another pillow is damaged, but as soon as you see it, Fido looks "guilty" and ducks under the couch. You drag him out and bop him again, since his guilty look insures that he "knew he did something wrong. "
... If this looks like a good example of punishment, think again. By definition, punishment causes behavior to decrease. In our example, the swats did not cause Fido to stop ripping up pillows, therefore the swats did not punish the act of ripping up pillows. Fido's owner can escalate the intensity of the "punishment" until it injures his dog but will not be able to stop the pillow ripping. That is because he is actually using negative reinforcement.
Things which increase behavior through force, intimidation, fear or avoidance are called negative reinforcers. If you sit on a thumbtack, the pain associated with the tack is a negative reinforcer, which causes you to do a behavior -- "jumping upward." The key difference between a reinforcer and a punisher is that one increases behavior, while the other one decreases behavior. In the case of our couch chewing canine, the swats and scolding did not affect the bad behavior at all. What actually happened was Fido's tendency to hide under the couch look "guilty" increased because of the harsh treatment. Those behaviors were negatively reinforced. [/b]