> Punishment and pain are primitive but functional teaching tools, though I
> would not use pain myself, punishment has its uses. When applied properly,
> its called dicipline, and it works for everything from withholding allowance
> for not completeing chores, to fines of corporations. They are not
> fullproof, and often ill conceived from the beginning, but it still stands
> when effectivly applied.
Here's an excerpt from "Don't Shoot the Dog: the New Art of Teaching and Training" by Karen Pryor
One reason punishment doesn't usually work is that it does not coincide with the undesirable behavior; it occurs afterward, and sometimes, as in courts of law, long afterward. The subject may not connect the punishment to his or her previous deeds; animals never do, and people often fail to. (...)
Even if subjects understand what they're being punished for, they cannot mitigate the punishment in the present, if only because they cannot change their actions in the past. You can't do anything now about a bad report card you already received, so a child with a punishing parent must just take the punishment.
(...) Punishment does not teach a child how to achieve a better report card. The most the punisher can hope for is that the child's motivation will change: The child will try to alter future behavior in order to avoid future punishment.
Learning to alter behavior in the future in order to avoid consequences in the future is more than most animals can understand. If a man catches his bird dog and beats it because it has been chasing rabbits, the dog has no way of knowing which particular recent activity is being reprimanded. It may become more fearful of the owner, which may allow the owner, from then on, to call it off when it chases rabbits. The beating in itself will not affect rabbit chasing.
Punishment or the threat of it doesn't help the subject learn how to modify the behavior involved. What the subject does learn, if the behavior is so strongly motivated that the subject _needs_ to continue it (stealing food when hungry, being one of the gang during adolescence), is to try not to get caught. Evasiveness increases rapidly under a punishment regime--a sad situation in a family setting and not so great in society at large either. Also, repeated or severe punishment has some very nasty side effects: fear, anger, resentment, resistance, even hate in the punished one, and sometimes in the punisher, too. Those are mental states not conducive to learning (unless you _want_ the subject to learn fear, anger, and hatred, emotions that are sometimes deliberately established in the training of terrorists).
One reason we keep thinking punishment works is that sometimes the punished behavior stops--if the subject understands which action is being punished, if the motivation for doing the behavior is small, if the fear of future punishment is large, and finally, if the subject can control the behavior in the first place (punishment doesn't cure bed-wetting, for example). A child who is scolded severely the first time he or she crayons on the wall may very well stop defacing the house. A citizen who cheats on his income tax and gets fined for it may not try it again.
Punishment has the best chance of halting a behavior in its track if the behavior is caught early, so that it has not become an established habit, and if punishment itself is a novel experience for the subject, a shock to which the person or animal has not become hardened.
When a punishment does effectively halt a behavior, that sequence of events is very reinforcing for the punisher. The punisher tends then to sally forth confidently to punish again. It always surprises me to witness the great faith that arises, in some individuals, in the effectiveness of punishment. I have seen it exhibited and defended by disciplinarian schoolteachers, bullying athletic coaches, domineering bosses, and well-intentioned parents. Their own punishing behavior may be maintained by a meager handful of successes in a morass of not-so-good results and can persist despite logical evidence to the contrary--despite the presence of other teachers in the same school, other coaches, heads of other businesses, other generals, presidents, or parents who can be seen to be getting results which are just as good or better without using punishment at all.
Punishment often constitutes revenge. The punisher may not really care whether the victim's behavior changes or not; he or she is just getting revenge, sometimes not against the recipient but against society at large. Think of obdurate clerks who with concealed glee delay or prevent you from getting your license, your loan, or your library pass over some minor technicality; you get punished and they get even.
Punishing is also reinforcing for the punisher because it demonstrates and helps to maintain dominance. Until the day when a boy is big enough to his his brutal father back, the father feels dominant, and is in truth the dominant one. This in fact may be the main motivation behind our human tendency to punish: establishing and maintaining dominance. The punisher may be primarily interested not in behavior but in being proved to be of higher status.
Dominance hierarchies and dominance disputes and testing are a fundamental characteristic of all social groups, from flocks of geese to human governments. But perhaps only we humans learn to use punishment primarily to gain for ourselves the reward of being dominant. So think, when you are tempted to punish: Do you want the dog, the child, the spouse, the employee to alter a given behavior? In that case, it's a training problem, and you need to be aware of the limitations of punishment as a training device. Or do you really want revenge? In that case you should seek more wholesome reinforcements for yourself.
Or perhaps you really want the dog, the child, the spouse, the employee, the neighboring nation, and so on to stop disobeying you. In whatever manifestation, do you want the subject to stop going against your superior will and judgment? In that case, it's a dominance dispute, and you're on your own.