Moving from Rewards and Punishments to Love and Reason

(New York: Atria, 2005)

Punitive Damages

By Alfie Kohn

… To punish kids, very simply, is to make something unpleasant happen to them -- or prevent them from experiencing something pleasant – usually with the goal of changing their future behavior.  The punisher makes them suffer, in other words, to teach them a lesson.[1]

Fundamental questions about the wisdom of this approach may suggest themselves even before we look at the results of scientific investigations.  For example, it may occur to us to ask How likely is it that intentionally making children unhappy would prove beneficial in the long run?  And:  If punishment is so effective, how come I have to keep doing it to my child over and over?

The available research does nothing to allay such doubts.  The results of a classic parenting study, published in 1957, seemed to catch even the authors by surprise.  After reviewing all the data from their investigation of kindergarteners and their mothers, they reported that “the unhappy effects of punishment have run like a dismal thread through our findings.”  Punishment proved to be counterproductive regardless of whether the parents were using it to stop aggression, excessive dependence, bed-wetting, or something else.  The researchers consistently found that punishment was “ineffectual over the long term as a technique for eliminating the kind of behavior toward which it is directed.”[2]  Newer and better-designed studies have only served to strengthen this conclusion, finding, for example, that parents who “punish[ed] rule-breaking behavior in their children at home often had children who demonstrated higher levels of rule-breaking when away from home.”[3]

By now there is an especially impressive collection of research demonstrating the destructive effects of corporal punishment in particular -- that is, the practice of spanking, slapping, or otherwise causing physical pain as a form of discipline.  The data overwhelmingly show that corporal punishment makes children more aggressive and leads to a variety of other damaging consequences.  (It’s not even clear that it succeeds at getting temporary compliance.)[4]  Hitting children clearly “teaches them a lesson” – and the lesson is that you can get your way with people who are weaker than you are by hurting them. 

I believe the research supports a zero-tolerance policy for spanking, given that it’s unnecessary, unproductive, and potentially very harmful.  But this, too, may be a case where data are not absolutely necessary.  Fundamental values may be enough to justify our opposition.  As sickening as it is that some men hit their wives or girlfriends, it’s arguably even worse for adults to hit children – in any manner and for any reason.

Still, just as the problems with control are not limited to punishment, so the problems with punishment are not limited to the physical kind.   The late sociologist Joan McCord put it well:

If parents and teachers were to substitute non-physical punishments for physical ones, they might avoid teaching children to hit, punch, and kick; yet, they would nevertheless perpetuate the idea that giving pain is a legitimate way to exercise power. . . . The consequences could be no less undermining of compassion and social interests.[5]

The problem, in other words, rests with the idea of forcing children to undergo something unpleasant.  The unpleasantness can consist of physical assault, deprivation of affection or attention, humiliation, isolation, or anything else.

This is worth emphasizing, first of all, because even some writers who firmly oppose corporal punishment seem to take on faith that other sorts of punishment are harmless or even necessary.  (Three shining exceptions, who have written eloquently on the problems with the very idea of punishment, are Thomas Gordon, Haim Ginott, and William Glasser.) 

A number of consultants, meanwhile, have responded to the understandable reluctance of many parents to use punitive tactics by repackaging them as “consequences.”  In some cases, the change is purely semantic, the implication being that a friendlier name will make the same practices less offensive.  But sometimes we’re told that if the punishments are less severe, or “logically” related to the misbehavior, or clearly spelled out in advance, then they’re OK to use – and, indeed, shouldn’t be considered punishments at all. 

I don’t buy it.  More important, I don’t think kids buy it.  While it’s certainly true that a bad thing can be made worse by adding such elements as unpredictability or lack of clarity – or by really overdoing it or being excessively nasty -- these aren’t the main reasons that punishment has the effects it does. 

Announcing how we plan to punish children (“Remember:  if you do x, then I’ll do y to you”) may salve our conscience because we gave them fair warning, but all we’ve really done is threaten them.  We’ve told them in advance exactly how we’ll make them suffer if they fail to obey.  This communicates a message of distrust (“I don’t think you’ll do the right thing without the fear of punishment”), leads kids to think of themselves as complying for extrinsic reasons, and emphasizes their powerlessness.  All the destructive effects predicted by logic, experience, and research are likely to follow regardless of these minor modifications – and regardless of whether we call punishment by a different name.[6]

Sometimes parents are advised to use a time-out instead of spanking their kids -- as though these were the only two options available.  The reality, as we saw in an earlier chapter, is that both of these tactics are punitive.  They differ only with respect to whether children will be made to suffer by physical or emotional means.  If we were forced to choose one over the other, then, sure, time-outs are preferable to spankings.  For that matter, spanking kids is preferable to shooting them, but that’s not much of an argument for spanking.

Another version of what might be called Punishment Lite is known as “natural consequences,” in which parents are invited to discipline by inaction – that is, by refusing to help.  If a child is late for dinner, we’re supposed to let her go hungry.  If she leaves her raincoat at school, we’re supposed to let her get wet the following day.  This is said to teach her to be more punctual, or less forgetful, or whatever.  But the far more powerful lesson that she’s likely to take away is that we could have helped -- but didn’t.  “When you stand by and let bad things happen, your child experiences the twin disappointments that something went wrong and you did not seem to care enough about her to lift a finger to help prevent the mishap.  The ‘natural consequences’ approach is really a form of punishment.”[7]

One of the most striking features of punishment -- any punishment -- is the way it creates a vicious circle for all concerned, very much like what we find with love withdrawal and positive reinforcement.  No matter how many times we’ve watched as the child being punished lashes out in anger or pain, no matter how many times a punitive intervention fails to bring about any improvement (and, more likely, actually makes things worse), we may assume that the only possible response is to punish again – perhaps even upping the ante.  Interestingly, research finds the worst effects aren’t due to the parent’s initial intervention but rather to the use of punishment after the child fails to comply with the first request.  It’s the reactive use of punishment, the choice to employ it once we’ve already locked horns with the child, that proves most worrisome.  Therefore, it’s most important to refrain from punishing precisely when we’re most angry or frustrated.[8]

The more important vicious circle, however, takes place not at the time we confront a child, but over time -- that is, as events play out over many years.  Repeatedly punishing a young child may help to turn him into a defiant adolescent, yet we’re advised to continue, and even intensify, the punishing:  ground the disobedient teenager, cut off his allowance, use our power to make him act responsibly.  The more this strategy fails, the more we assume the problem is with the child, rather than with the strategy itself.  And if we do stop to reconsider what we’re doing, we assume we’ve just been implementing it ineffectively – as opposed to realizing the trouble is with the whole idea of making children suffer to teach them a lesson.  Ginott was absolutely right:  “Misbehavior and punishment are not opposites that cancel each other; on the contrary, they breed and reinforce each other.”[9]

Why Punishment Fails

That punishing kids doesn’t work is very difficult to deny in light of all the available evidence.  Why it doesn’t work is harder to say with certainty.  Nevertheless, we can hazard some guesses.

* It makes people mad.  Like other forms of control, the use of punitive consequences often enrages whoever is on the receiving end, and the experience is doubly painful because he or she is powerless to do anything about it.  What history teaches us about nations echoes what psychology teaches us about individuals:  Given a chance, those who feel like victims may eventually become victimizers.

* It teaches power.  The example that corporal punishment sets for children is violence – that is, the use of force to solve problems.  In fact, though, all punishment teaches something similar.  Children may or may not learn the lesson we had in mind when we punished them (“Don’t do x again”).  But they’ll surely learn that when the most important people in their lives, their role models, have a problem, they try to solve it by using power to make the other person unhappy so he or she will be forced to capitulate.  Punishment not only makes a child angry; it “simultaneously provides him with a model for expressing that hostility outwardly.”[10]  In other words, it teaches that might makes right.

* It eventually loses its effectiveness.   As kids grow older, it becomes harder and harder to find things to do to them that will be sufficiently unpleasant.  (By the same token, it becomes increasingly difficult to find rewards that are sufficiently appealing.)  At some point, your threats begin to sound hollow and your kids just shrug off “You’re grounded!” or  “No allowance for you this week!”  This doesn’t prove that kids are tough or obstinate, nor does it mean that you need help devising more diabolical ways to make them suffer.  Rather, what it suggests is that trying to help kids become good people by punishing them for doing bad things may have been a foolish strategy from the beginning.

Think about it this way:  When young children wonder why they should be nice or resist certain temptations, parents have a choice.  They can draw upon the respect and trust they’ve cultivated by loving their kids unconditionally, using reason and persuasion to explain how doing this thing rather than that thing is likely to affect other people.  Or they can just appeal to naked power:  “If you don’t cut that out, you’ll be punished.” 

The problem with the latter approach is that once your power begins to ebb – and it will -- you’ve got nothing left.  As Thomas Gordon pointed out, “The inevitable result of consistently employing power to control [your] kids when they are young is that [you] never learn how to influence.”  The more you rely on punishment, therefore, “the less real influence you’ll have on their lives.”[11]

It erodes our relationships with our kids.  When we punish, we make it very hard for our children to regard us as caring allies, which is vital for healthy development.  Instead, we become (in their eyes) enforcers to be avoided.  Very young children begin to wrap their minds around the fact that their parents, those huge all-powerful people on whom they are totally dependent, occasionally make them miserable on purpose.  Those giants who hold me and rock me and feed me and kiss away my tears sometimes go out of their way to take away things I like, or make me feel unworthy, or hit me on the backside (even though they keep telling me I’m always supposed to “use my words”).  They tell me they’re acting this way because of something or other that I did, but all I know is now I’m not sure I can trust them or feel completely safe with them.  I’d be pretty stupid to admit to them that I’m angry, or that I did something bad, because I’ve learned that I might be given a time-out or talked to in a voice that has all the love drained out of it or even smacked.  I’d better keep my distance.

* It distracts kids from the important issues.  Suppose a child is told that, because he just punched his brother, he has to go to his room and miss his favorite TV program.  Let’s peek in on him, sitting on his bed.  What do you imagine is going through his mind?  If your guess is that he’s been reflecting on his actions, perhaps saying to himself thoughtfully, “Y’know, now I see that hurting people is wrong” – then, by all means, keep sending your kids to their rooms whenever they misbehave.

If, however, like anyone who has ever spent time with a real child (or has ever been one, for that matter) you find that scenario laughably improbable, then why would you ever impose this – or any other – punishment?  The idea that time-outs are an acceptable form of discipline because they give kids time to think things over is based on an absurdly unrealistic premise.  More generally, punishment doesn’t lead children to focus on what they’ve done, much less on why they did it or what they should have done instead.  Rather, it leads them to think about how mean their parents are and maybe how they’re going to get their revenge (on the kid who got them into trouble).

Above all, they’re likely to focus on the punishment itself:  how unfair it was and how to avoid it next time.  Punishing kids – with the threat that you’ll do so again if they displease you in the future -- is an excellent way to hone their skills at escaping detection.  Tell a child:  “I don’t want to catch you doing that again,” and the child will think, “OK.  Next time you won’t catch me.”  It also sets up a strong incentive to lie.  (By contrast, children who aren’t punished are less afraid of owning up to what they’ve done.)  Yet punitive parents, faced with the predictable dishonesty that accompanies traditional discipline – “I didn’t do it!  It was already broken!” – are likely to respond to this not by questioning their use of punishment but by punishing the child again, this time for lying.

* It makes kids more self-centered.   The word consequences is tossed around a lot, not only as a euphemism for punishment but also as a justification for it – as in “Kids need to learn that there are consequences for their actions.”  But consequences to whom?  The answer given by all punishment is:  to yourself.  A child’s attention is firmly directed to how she personally will be affected by breaking a rule or defying an adult – that is, what consequence she will face if she’s caught. 

When we punish, in other words, we lead children to ask, “What do they (the grown-ups with the power) want me to do, and what will happen to me if I don’t do it?”  Notice that this is a mirror image of the question evoked in a home or classroom in which children are promised a reward for being good:  “What do they want me to do, and what will I get for doing it?”  Both questions are entirely about self-interest.  And both are completely different from what we’d like kids to ask themselves – for example, “What kind of person do I want to be?”

No wonder a pair of researchers, after discovering that punishing children interferes with their moral development, made sense of that finding by pointing out that punishments “direct the child to the consequences of his behavior for the actor, that is, for the child himself.”[12]  The more we rely on punitive consequences, including time-out -- or rewards, including praise -- the less likely children are to consider how their actions affect other people.  (They may, however, become more likely to perform a cost-benefit analysis – that is, to weigh the risks of being caught and punished against the pleasures of doing whatever it is they’re not supposed to do.)

These responses – calculating the risks, figuring out how not to get caught, lying to protect themselves -- make sense from the child’s perspective.  They’re perfectly rational.  What they’re not is moral, and that’s because punishment – all punishment, by its very nature -- impedes moral thinking.  Thus, when defenders of traditional discipline insist that kids are going to face consequences for their behavior when they’re out in the “real world,” the reasonable response would be to ask what sort of adult out there in the real world is dissuaded from unethical behavior only when he, himself, will pay the price (if he’s caught).  Our answer would have to be:  the sort of adult most of us hope our children won’t become.


The argument I’ve been making is largely a practical one.  By any meaningful criteria, punishment simply doesn’t work very well, and it’s not realistic to expect that more punishment (or a different kind) will turn things around.  But how are we to respond to parents who contend that explaining, reasoning, empathizing, and so on can’t have more than a limited impact, so we need to “put some teeth into” what we’re telling kids and “get their attention” by imposing a consequence, too?

To begin with, notice that this claim is based on the assumption that without the addition of some coercive enforcement mechanism children will ignore the most important people in their world.  That’s a hard case to make.  Sure, kids sometimes ignore specific things we tell them, demonstrating a remarkable case of selective hearing when we call them to dinner or ask them to clean up, but that doesn’t mean they’re oblivious to our words and actions.  On the contrary, even the words of the gentlest parent – or perhaps I should say especially of the gentlest parent – carry enormous clout just because of who’s saying them.

Still, could someone argue that threats and punishments command children’s attention in a different way?  Yes, but the way they do so is terribly counterproductive.  The very features of punishment that make it impossible to ignore also make it almost impossible for any good to come out of it.  What’s getting the kids’ attention here is pain, along with the fact that someone on whom they’re dependent has caused that pain.  This is hardly likely to produce the effect that most of us are looking for.  In fact, the effect of punishment is such that it can undermine the benefits of good parenting if the two approaches are combined.[13] 

Some parents rationalize the use of punishment by insisting that they really, truly love their kids.  No doubt this is true.  But it creates a deeply confusing situation for children.  It’s hard for them to sort out why someone who clearly cares for them also makes them suffer from time to time.  It creates the warped idea, which children may carry with them throughout their lives, that causing people pain is part of what it means to love them.  Or else it may simply teach that love is necessarily conditional, that it lasts only for as long as people do exactly what you want.

Another rationalization is that punishment isn’t destructive as long as it’s imposed for a good reason and as long as that reason is explained to the child.  The truth is that explanation doesn’t minimize the bad effects of punishment so much as punishment minimizes the good effects of explanation[14].  Suppose you explain things to your child and try to help her focus on how her actions have made someone else feel.  You say:  “Annie, when you grabbed the Legos away from Jeffrey, you made him sad because now he can’t play with them.”  But what if you’re also in the habit of punishing her for certain offenses?  The benefits of your explanation may well be wiped out.  If Annie knows from experience that you might send her to the time-out chair or do something else unpleasant to her, she’s not thinking about Jeffrey.  She’s just worried about what this will mean for her.  The more anxious she’s learned to become about the possibility of punishment, the less chance that meaningful moral learning will take place.


If you combine everything in this chapter with the discussion in chapter 2, then a larger pattern begins to emerge.  What I’ve described as a “doing to” approach, which encompasses conditional parenting, actually exists on a continuum, with “harsh corporal punishment” on one side, then “milder spankings,” then “other punishments,” then “tangible rewards,” and finally, on the other end, “verbal rewards.” I don’t mean to say that hitting your child and saying “Good job!” are morally equivalent.  But they are conceptually connected.  My concern is with all of these techniques as well as with the assumptions that link them.  In my experience, many parents [and teachers] are less likely to explore the “working with” alternative as long as they think it’s enough just to pick one of the “doing to” options on the right side of this diagram.  That’s why I’ve been spending so much time emphasizing how important it is to reject the whole model.

In effect, I’ve also been challenging a view that might be called “the more, the merrier.”  This is the tendency to dismiss arguments that any specific parenting practice is bad news and ought to be replaced by another.  “Why not do both?” some people ask.  “No reason to throw anything out of your toolbox.  Use everything that works.”

To begin with, we should respond once again:  “Works to do what – and at what cost?”  But the real problem is that different strategies sometimes work at cross purposes.  One may wipe out the positive effects of the other.  You may recall the bit of folk wisdom, confirmed by generations of farmers and grocers, warning that a rotten apple placed in a barrel full of good apples can spoil them.  It would be pushing things to postulate a kind of psychological ethylene released by traditional discipline, analogous to the gas given off by bad fruit.  But it does seem that the quest for optimal results may require us to abandon certain practices rather than simply piling other, better practices on top of them.  We have to eliminate the bad stuff, like punishment and rewards, in order for the good stuff to work.[15]



(Please see Unconditional Parenting for the complete citations)

1.  In some cases, children – and, even more commonly, adults – may be punished without regard to whether the intervention is effective.  The point may be not to change future behavior but to exact retribution.  This evidently motivates some teachers to punish their students (Reyna and Weiner); it’s unclear how many parents resort to punishment with the goal of changing how their children act and how many see punishment as a moral imperative.

2. Sears et al., p. 484.

3. Toner, p. 31.  Likewise, “punitive discipline emerged as a common or shared predictor of all the dimensions of child disruptive behaviors,” reported the multi-university Conduct Problems Prevention Research Group in 2000 (Stormshak et al.; quotation on p. 24).  And from another study, conducted in the Midwest:  Punishment of various sorts “contributed more unique variance to predicting problem-behavior ratings than all demographic predictors combined” (Brenner and Fox; quotation on p. 253.)   Of course, the discovery that punishment is associated with children’s misbehavior could be explained by the possibility that parents with tough kids are more likely to punish them; in other words, punishment may be “pulled” by the child’s actions rather than causing those actions.  Undoubtedly it’s true that the causal arrows point in more than one direction, but by now there’s enough evidence, from studies designed specifically to test this hypothesis, to justify the conclusion that punishment is a cause more than an effect.  For example, see Hoffman 1960, p. 141; Kandel and Wu, p. 112; Cohen and Brook, p. 162; and, for the causative role played by corporal punishment in particular, Straus 2001, chap. 12.  Similarly, while parents may respond more harshly to a toddler who is unusually aggressive, that response is driven in large part by the parent’s pre-existing attitudes about child rearing (Hastings and Rubin; see also Grusec and Mammone).

4. At this writing, the most ambitious summary of the existing research on corporal punishment is a monograph published by Gershoff in 2002.  Of the studies she reviewed that looked at the effects on short-term compliance, three found a positive effect and two did not (p. 547).  (Even those three didn’t prove that corporal punishment was more effective than other methods.)  More important, her metaanalysis of a whopping 88 studies discovered that corporal punishment by parents is associated with “decreased moral internalization, increased child aggression, increased child delinquent and antisocial behavior, decreased quality of relationship between parent and child, decreased child mental health, increased risk of being a victim of physical abuse, increased adult aggression, increased adult criminal and antisocial behavior, decreased adult mental health, and increased risk of abusing own child or spouse” (p. 544).  Also see the work of Murray Straus.

5. McCord 1991, pp. 175-6.

6. I offer a critique of some of the “New Discipline” programs, including “Discipline with Dignity,” “Cooperative Discipline,” “Discipline with Love and Logic,” and the recommendations offered by Rudolf Dreikurs and his followers, in my 1996 book for teachers, Beyond Discipline.  See especially chapter 4:  “Punishment Lite: ‘Consequences’ and Pseudochoice.”

7. Pieper and Pieper, p. 208.  This is not to say that there is no such thing as a true natural consequence.  If we stay up late, we’ll likely be tired in the morning.  If we don’t go shopping, we’ll eventually run out of food.  But these scenarios are very different from, say, a parent’s refusal to heat up dinner for a child who comes home late.  Call that whatever you like:  It’s still a punishment, and one that feels particularly humiliating, at that.  (An accompanying “I told you so” or “It serves you right” or “I hope you’ve learned your lesson” will only serve to make the child feel even worse.)

8. Hoffman 1960.  Needless to say, this is hard to do.  Research (e.g., Ritchie) confirms that parents are more likely to respond punitively during a conflict in which they and their children are locked in a battle of wills than after a single act of noncompliance.

9. Ginott, p. 151.

10. Hoffman 1970a, p. 114.

11. Gordon 1989, pp. 74, 7.

12. Hoffman and Saltzstein, p. 54.

13. For evidence that this is true of love withdrawal, see Hoffman 1970a, pp. 109, 115.

14. For example, see Hoffman 1970a, p. 109.  Straus 2001 (p. 101) makes the additional point that parents who spank but explain why they’re doing so are “teaching the child just what to do and what to say when he or she hits another child.”

15. This same phenomenon shows up in schools with regard to better and worse forms of teaching, as I argued in an article called “Education’s Rotten Apples” (Kohn 2002).


Copyright © 2005 by Alfie Kohn. This article may be downloaded, reproduced, and distributed without permission as long as each copy includes this notice along with citation information (i.e., name of the periodical in which it originally appeared, date of publication, and author's name). Permission must be obtained in order to reprint this article in a published work or in order to offer it for sale in any form. We can be reached through the Contact Us page. -- © Alfie Kohn