Alfie Kohn's head‘Apologize.’ my mom would say to me. ‘What? ‘Say you’re sorry.’ ‘But I’m not.’ ‘Come on, Aaron — just say you’re sorry.’ ‘But that would be a lie! I’m not going to lie.’ ‘Say you’re sorry or [bad thing will happen to you].’ ‘This is absurd! How is forcing me to lie going to solve anything?’

Perhaps it was destined I would find Alfie Kohn. I remember, the first time I heard of his book, I thought its contents were so obvious that I didn’t bother to read it. But Alfie Kohn is far from obvious. In fact, he’s downright radical. And yet there he is, up on Stanford’s stage, a completely packed crowd, cameras all around him (he’ll be putting the talk on DVD).

‘This is the ultimate in behaviorism!’ he screams. ‘Thinking a child is going to feel sorry just because you make him say he’s sorry? The only thing that’s going to do is teach the child to lie about his feelings!’

Kohn paces across the stage, jumping in the air, waving his arms, and screaming for emphasis. He has so much energy he looks like he could burst and what he says is so radical it’s hard for most to swallow, yet he somehow manages to appear (just barely) on this side of sanity.

The obvious and radical thing Alfie Kohn says is this: our system of childrearing — our system of life — is profoundly screwed up. Competition does not work, he explained in his first book (No Contest), it’s less effective, less enjoyable, and less healthy than working alone or cooperating. (Immediately after reading the book I promised myself I would never again play a competitive game. When I tell people this, their response is invariably ‘What other kind of game is there?’ There are many cooperative games, but their total invisibility neatly proves how screwed up our society is.)

Rewards don’t work either, he explained a few years later (Punished By Rewards). Instead of increasing people’s motivation, they destroy it. Once people try to get the goodie, they’re no longer interested in the act unless they get another goodie. So reward kids for being altruistic and they’ll donate less. Reward kids for behaving well and they’ll try to misbehave behind your back.

The thing that makes Alfie Kohn’s books so special, though, is that these aren’t just his claims, backed up by anecdotes. These are the results of copious research, experiment after experiment, study after study. Alfie Kohn manages to collect all the studies together and come to some amazingly powerful conclusions: “not a single controlled study has shown a long-term improvement in the quality of work as a result of any reward system. That would be an astonishing fact were it not for the existence of scores of studies — conducted with adults as well as children, in real workplaces among other venues — that have demonstrated how rewards tend to be not merely ineffective but powerfully counterproductive.” The same is true of competition. Study after study finds it’s powerfully counterproductive; none that it has any long-term benefits.

But today he’s here to talk about his latest book, Unconditional Parenting, which explains how to apply this and other research to raising your kids. ‘The biggest problem in this country is not permissiveness,’ he says. ‘The biggest problem is fear of permissiveness. For every one kid who is allowed to run around the store screaming — and believe me, they annoy me too — a hundred kids are slapped or yelled at or punished or rewarded. Fear of permissiveness leads parents to clamp down on control.’

And, aside from being morally reprehensible, control only hurts children. It makes them rebel wildly or it makes them conform — two sides of the same coin, both hurting kids. Either way, control never educates. No child has ever learned why hurting people is wrong because they were punished for it. No child has ever learned the importance of helping people by being rewarded too.

At base, punishments and rewards and competitions — all the incentive schemes — ignore the fact that there’s a child in there, another human being like you, who can be honest and moral and good if you just give them a chance. But every time you ignore that person and try to take control for it, it withers away a little bit. And that, not children who fail to apologize or are late for school, is the real crime.

posted April 24, 2005 05:26 PM (Education) (25 comments) #


SFP: Come see us
Stanford: The Cynic Returns
Social Class in America
Stanford: Eat the Whales
Alfie Kohn on Incentives and Parenting
Stanford: Seeds of Revolution


Within a few days of each other:

“sleeping the sleep of a man who, whatever his surroundings, knows that at heart he is a capitalist.”

“I promised myself I would never again play a competitive game.”

So I can understand the difference between competition in play and in business; but there’s certainly a common element here. Even if asking you to draw a line would be daft, what’s the dimention? Why is one clearly on the bad side of the line and the other, the good?

posted by Adam Langley at April 24, 2005 06:15 PM #

You’re the second person to ask about it, so I guess I was pretty unclear. The first comment was meant to be ironic. I hate being a capitalist. The motivational posters and signs saying “What would a wealth-maximizing automaton do?” around the office are jokes.

posted by Aaron Swartz at April 24, 2005 06:22 PM #

I’ve heard Alfie Kohn speak, and there’s a lot in what he says.

But in general, I think there’s missing pieces when one tries to apply it.

As best I can express it, purely co-operative systems, while maximal in an ideal sense, seem to have a practical instability that makes them extremely vulnerable to looting from competitive systems. And also that no matter how many times you repeat to some people that they are causing a long-term loss by a short-term gain, they’ll still favor the short-term over the long-term.

“Once people try to get the goodie, they’re no longer interested in the act unless they get another goodie.”

This is actually pretty rational behavior. If you tell someone there’s a market for something, where they can get paid for it, then expect them to work for free, maybe they won’t do it. But I think any conclusion of never create a market, would be a bit too simplistic (maybe it wasn’t right to expect them to work for free in the first place!).

posted by Seth Finkelstein at April 24, 2005 10:16 PM #

The problem is that competitive behaviour wins in the short term. So you either have to be ready to sacrifice/”lose” in order to live out your ideals, or compromise your ideals for the sake of allowing them to live on.

I think this has parallels to Christianity, but that’s just me.

posted by Micah at April 24, 2005 11:06 PM #

I guess I should read the book. As a parent you are in a hard situation. For example, my son hits/throws sand at/says something mean to another boy in the sandbox. I tell him to say he’s sorry. He says “But he started it/did whatever it was first, I’m not sorry.” Now, the other kid’s parent is right there. About all I can do is get my son to say he’s sorry, because he can’t take back his actions or words. In order to continue the social relationship we have with the playmate and his parents, my son and/or I must do something to show we are sorry, and that his behavior is not acceptable. If his insincere appology doesn’t appease the other child and his parent, then I would probably do something along the lines of punishing or repremanding my child. As a parent, you do have to teach your children to fib as part of social training. If someone gives you a present you say thank you very much and try to look pleased even if you don’t like the gift. Etc.

posted by a parent at April 25, 2005 12:15 AM #

If people will be moral and virtuous given the chance, then how do you explain George Bush? What evil controlling force is he exactly rebelling against? Or, what’s this invisible hand that’s controlling his actions? And what’s the invisible hand that controls this invisible hand, etc.? And what does Kohn have to say about the Prisoner’s Dilemma, which isn’t even a competitive game?

I don’t see the constant demand for rewards to be a bad problem. If people need an incentive to tell the truth, then give it to them: punish people who lie, and reward people who tell the truth even when strongly tempted to lie. And why remove the incentive at any time at all? Just leave the incentive there for as long as humanly possible.

But I agree, money doesn’t work all the time. It irks me that when some organizations who want to encourage a particular activity X, the only methods they can come up with are to move money to X, and to drain money away from non-X… dang it, I don’t do everything just for the money!

posted by bi at April 25, 2005 12:25 AM #

Kohn’s sort of walking into the lion’s den by giving that talk at a school like Stanford. By and large the students accepted there are ones who have done extremely well in the various high school competitions (grades, SATs, etc.) and resume-building needed to succeed in the competition to be admitted to such a fine institution. I’d be curious to know more about the audience reaction, questions asked, etc.

Aaron’s response: It was mostly School of Education students and parents and teachers from a nearby progressive school, since those were the groups which co-sponsored the talk. The parents were disppointingly supportive.

posted by mike at April 25, 2005 12:33 AM #

As another person who takes the position that “But I’m not sorry, you’re asking me to lie”, but not a parent myself, I’ll point out just how complex are the implications here:

1) What if he doesn’t want to continue the social relationship? What if the other kid did start it, and he was defending himself?

2) Observe the lesson which can be taught - “Lying when authority figures demand you do leads to a better life, and they’ll punish you if you don’t agree” - that certainly seems a reasonable abstraction (note practice-vs-preach here).

I think we have a short-term vs. long-term here. A lie might defuse the situation in the short term, but the long term implications might be negative.

posted by Seth Finkelstein at April 25, 2005 12:34 AM #

Now, as it relates to raising a child (which I will start doing in a couple of months now), I fundamentally agree with the idea that right and wrong can never be taught with rewards and punishments (my wife is somewhat more sympathetic to rewards than I am, but I don’t think the difference is all that big in practice). Of course, this applies only to “artificial”, or deliberate/manipulative awards. I will still smile when my son does something that warms my heart, and if my son hits the other kids in the playground, I’ll have to take him away from there and explain to him that we can’t go there if he hurts his playmates.

But I don’t entirely understand what this has to do with competition. There is destructive competition, yes. That’s why my wife likes that picture of a child Mozart with “Wanted: CHILD PRODIGY. Has made the life unbearable for thousands of young pianists”, I suppose :-) . And sometimes the competitive elements are best left out. When we play scrabble I won’t keep score, because I think the beauty of the words and patterns outweights, and is sabotaged by, too much competition.

But there’s positive competition, too. I am a big fan of german board games, which are usually (but not always) competitive. The key issue for me is that you don’t need to win for the game to be enjoyable. That’s a large part of what good game design is.

On the other hand, everything from “tag” to Mastermind(c) to playing the piano can be made into a bitter, competitive game if you play with the wrong people. Although certain games are more destructively competitive than others, it seems to me that “Competition doesn’t hurt people, people hurt people” to paraphrase that old slogan.

So I don’t quite understand why you never play competitive games.

posted by Harald Korneliussen at April 25, 2005 02:37 AM #

I am 22 years old. In some respects I still consider myself a child, at least I feel as if I am still overcoming certain aspects of my upbringing when I WAS a child.

For a while now I’ve consciously found my 20s to be a difficult age. I am tortured by anomalies in what I want to be (at least what I feel I want to be at a particular time), what I feel I could be, and what I feel I ought to be.

Every line reiterated from Mr Kohn resounded deeply within my half-child/half-adult heart. I wish I had been brought up with a view for longterm benefits. I wish I had been socialised to be honest, courageous in my truth and unfazed by authority. Learning these things in my 20s is made harder by the behaviours I have to unlearn first.

I am not a parent, and I do not have sophisticated comparisons to make between parenting and workplace behaviour, but it seems to me that most folks would be happier if they were given permission from a young age to be themselves, within boundaries that permit others to also be themselves.

Aaron, I really enjoy reading your writing. Thanks for sharing it.

posted by Davinia Douglas at April 25, 2005 06:11 AM #

Does saying sorry teach the kid that he shouldn’t have done that what he has to say sorry for?

Aaron: No, it teaches him that he can make problems go away by lying,

And doesn’t it also learn him that when he does something wrong, he has to apologize?

Aaron: No, it teaches him he only has to apologize when his mom is around.

As for competition, what about two small groups of kids competing against each other in a race or something? They have to cooperate, but yet they are trying to do their best. I’d say this isn’t bad, unless of course kids are blaming others for losing… that’ll get very destructive indeed.

Aaron: Yes, Kohn’s book reports that having people do things in groups that compete can moderate the negative effects of competition. But that’s only because the kids forget that they’re competing.

posted by Mark Wubben at April 25, 2005 06:33 AM #

It’s not true that capitalism is more successful in the short-run and communism is more successful in the long-run. This hypothesis is both theoretically and empirically disproven.

When I say that, I mean “communism comprised of homo sapien sapien”. Perhaps it would be different with some other agent, as long as that agent could be constrained from acting in its own best interest.

posted by Zooko at April 25, 2005 06:44 AM #

Allow me to post a possibly tangential rant, triggered by a pet peeve of mine.

I think a couple of the earlier comments which mentioned the phrase “short-term” triggered one of my pet peeve reactions.

See, there is a very simple and widely known game theory puzzle which is widely misunderstood. It’s the Prisoner’s Dilemma: two players, each can choose A or B, the payoff matrix is { A/A: 3/3, A/B: 0/5, B/A: 5/0, B/B: 1/1 }.

That’s it. That’s the whole puzzle. Now there is something about it which is counter-intuitive to most humans: if each player acts in their own best interest, then they will both get a suboptimal result.

Humans seem to have a very hard time getting their heads around the truth of this statement. They generally take refuge in one of two evasions:

  1. The players will get a “short-term” benefit from playing the B strategy, but a “long-term” benefit from playing the A strategy. Wrong wrong wrong wrong wrong. I have gotten increasingly grumpy over the years with this one. Look at the %$*! problem again. This is math — there is no room for interpretation or ambiguity here.

  2. Okay, so the theoretical puzzle itself does have this perplexing counter-intuitive consequence, but such a situation never arises in real life, or arises so infrequently that we can safely ignore it.

Evasion #2 deserves more respect than evasion #1. Even though #2 is in fact wrong, it is wrong because of complex reasons, and to understand its wrongness might require either empirical evidence or sophisticated modelling. Whereas evasion #1 is just plainly wrong and indicates an inability to do simple math.

By the way, I first heard evasion #2 from Objectivists — adherents to Ayn Rand’s philosophy.



P.S. Perhaps one reason people fall for evasion #1 is because of the widely discussed and important Iterated Prisoner’s Dilemma. Just restrict your attention to the one-shot Prisoner’s Dilemma long enough to understand that evasion #1 is untrue when applied to the one-shot, and then you can safely think about the Iterated version and about evasion #2. Thank you for you attention.

posted by Zooko at April 25, 2005 07:05 AM #

Zooko that’s a great example (and I assume you’re going with B/B as the only outcome). I think it argues the capitalist case well, I also think that’s why you used it, so I’m going to expand on it.

In the example both actors have perfect information of the consequences of their actions. In the real world you don’t perfectly know the consequences of your actions, and you also don’t know exactly what other people’s preferences are. So total well being arises from each actor acting in what they perceive as the best interest (personal or group), given the provided information — since each they can only act based on provided information.

If you assume each actor will reliably act how their society’s ethics predict, the utilitarians will both go for A/A, but the self-optimizers will both go for B/B.

The problem I see with the utilitarian model in practice (as seen in communism) is it assumes there is accurate information about what makes everyone best off, and in order for this to work in practice there must be an enormous transfer of information (some of which, like personal utility functions, are very hard to codify). It also assumes that everyone has the same ethics.

Maybe if Kohn’s methods were used to raise children everyone would not be subject to any cohersion, and they could cooperate on collective goals. But this raises the information transfer problem, which is neatly solved by prices (for anyone in doubt of this see Hayek: The Use of Knowledge in a Society).

It’s also worth pointing out, since it seems we’re only talking about cohersion in anti-violence contexts, that preventing people from hurting other people is the only cohersion present in a libertarian (perfectly capitalist) society. Now there are other incentives present in our society. For instance, the government steals 15% of my income for social security because not having starving people on the street, and the protection from personal starvation, are considered public goods. My returns should amount to greater 15%. Whether this is true in practice is hard to know, given the information problem, but assuming it was true I’m only being coherced because social security is a public good and I would have an incentive to free-ride on altruists otherwise — so the cohersion actually benefits me, although that’s only an emergent (and long-term) property.

I think there’s a confusion in the economics arguments presented here. So far I’ve seen three economics systems discussed: communism, capitalism, and anarchosyndicalism. Pure capitalism is the only one which doesn’t fall prey to the information problem, but capitalism in practice (with incentives not to hurt other people and to give to social security) does. However, I believe there is less of an information problem in capitalism than in the other two systems, because capitalism has prices as a distributed representation of preferences. Communism in theory has similar results as anarchosyndicalism except that communism employs cohersion and anarchosyndicalism doesn’t, but both have information problems which in my mind keeps them from being practically possible (both are consciously attempting to act in either society or the union’s best interest).

So yes using cohersion or incentives to change your child’s behavior is based on imperfect assumptions (that people don’t like being hit and that society will be better off if people don’t hit each other), but these seem more definite than the assumption that we can understand the summed preferences of a collective and act on its behalf.

posted by Jeremiah Rogers at April 25, 2005 11:36 AM #

Jeremiah Rogers: The optimal outcome is 3/3. In a capitalist society where everyone acts in his own interests, the most likely outcome will be 1/1. It’s crystal clear that rational self-interest in this game doesn’t produce the optimal outcome. How then can the Prisoner’s Dilemma be an argument for pure capitalism?

posted by bi at April 25, 2005 01:30 PM #

bi: With respect to the perfect information Prisoner’s Dilemma you’re correct, capitalism is clearly not the best system. The best system in this example is one which analyzes the information and makes the best collective choice (which as you said is 3/3).

I argued practice it is impossible to know the optimal outcome in this situation because the information is imperfect, and cited F.A. Hayek’s paper which illustrates that the best way to collectively gather knowledge in a society is through prices.

posted by Jeremiah Rogers at April 25, 2005 02:31 PM #

The Stanford page gives several realistic scenarios corresponding to the Prisoner’s Dilemma. Knowing the “prices” of things doesn’t resolve the dilemma in these scenarios by one single bit.

posted by bi at April 25, 2005 02:52 PM #

The payoff values in this grid “{ A/A: 3/3, A/B: 0/5, B/A: 5/0, B/B: 1/1 }”, the numbers 3, 0, 5, 1, are what I mean by prices. I’m not disputing that the prisoner’s dilemma exists, there are numerous examples of similar situations (one of which I even provided: social security).

I’m saying that the quality of the information about payoffs (in this example equivalent to prices, preferences, utility, payoffs) is uncertain.

And yes, knowing these prices does resolve the dilemma, it gives you the information you need to realize that there are two maxima here. The actors, and their payoff functions (which determine these values) are the only properties which even allow you to concieve of the dilemma.

I admit I’ve only skimmed the Stanford page. It doesn’t seem to provide more information than is found in Varian’s Microeconomics, Osbourne’s Game Theory or Dixit and Nalebuff’s Thinking Strategically.

Actually, in the cooperative economy the situation just gets remodelled as { A/A: 3/3, A/B: 0/2, B/A: 2/0, B/B: 1/1 } and the dilemma no longer exists.

posted by Jeremiah Rogers at April 25, 2005 04:17 PM #

Zooko: I’m very familiar with the Prisoner’s Dilemma. I didn’t cite it because I was being non-technical, but I’d say that it’s what I meant by mentioning “looting” (== defection). However, short-term vs. long-term is a different, separate issue.

The case of someone saying they think they need to get their kid to tell a lie immediately in order to resolve social tension isn’t really a Prisoner’s Dilemma, more of a short-term gain (social harmony) with possible long-term loss (potentially teaching the kid that lying is a good, socially-approved, way to manipulate people).

posted by Seth Finkelstein at April 25, 2005 11:43 PM #

I’m sorry, but Kohn’s approach has little bearing on the reality of parenting. Yes, to be a good parent you must let your children know you’ll always love them, but there’s still the not insignificant matter of shaping the child’s immature behavior.

No matter how hard I try to persuade my five-year-old daughter to pick her naked barbies off the floor, saying “because it’s the right thing to do” simply won’t cut it. If I tell her that we need to pick things off the floor because otherwise we all trip over them, she’ll do her best to argue the point and say why don’t you walk around it? What does work is telling her she has a choice to pick up her room or not, but if she does, we can then go to the playground. In other words, incentives paired with implicit threat of punishment.

If there’s any basis in what Kohn says, it’s that overly charged punishment such as slapping or yelling is counterproductive, as is overly rigid control, but we know that.

The last paragraph is pure nonsense, as if your kid is really just a little adult, with the same capacity for reasoning as you do. No, I’m sorry, but kids that age are tremendously impulsive and self-centered, and even the best of them will try to get away with as much BS as they can. And it’s your responsibility as a parent to make sure they don’t grow up behaving that way.

posted by Mike Sierra at April 26, 2005 12:02 AM #

I must agree with Mike Sierra. I once agreed with the “children are little adults” point of view, but now that I’m a parent I have to say it’s patently false. I can’t appeal to my son’s rational sense of fairness and ask him “how would you feel is someone hit you?”, because he’s not really a rational being yet. (Doesn’t mean I still try, though—gotta teach that, too). Learning that hitting (for example) gets (non-violent) punishment is just all they can understand at a young age. You have to do something to stop a child from acting inappropriately in the meantime.

I might add that rational arguments alone don’t seem to work too well for adults, too, as there appears to be a rather large body of laws out there with different punishments attached for breaking them.

“…who can be honest and moral and good if you just give them a chance…”

Honesty and morality (right and wrong) need to be taught just like anything else. Or do you seriously believe that parenting has no infleuence here?

As far as competition goes, I can agree with some of that. I think cooperation is given short shrift these days. However, I don’t believe that competition is inherently bad, either.

posted by Rich Fletcher at April 26, 2005 01:42 AM #

I have a serious problem with the thesis discussed here. Let’s suppose:

One day a child decides to pick up his toys before he goes to bed, without being asked.

As a parent, I would smile and say “good job!”. But you see, that smile and my reaction is a reward for the child.

Let’s say you’re right, the next time if I don’t smile when the child picks up his toys (let’s say I’m in a bad mood that day), he will stop picking up toys: no reward = no picking up toys.

So in order to avoid this problem when a child picks up his toys a new parent should do… what exactly?

Perhaps feign disinterest? But I am happy. By doing anything but showing my happiness I am now lying to my child. So which is worse? Lying to kids or praising them? Sorry, I don’t see a way out of this one.

posted by jg at April 26, 2005 02:19 AM #

Harald Korneliussen:


posted by bi at April 26, 2005 02:28 AM #

jg, what do you do? The same thing you’d do with any normal human being. Say things like “Would you mind picking up those toys? I sometimes accidentally step on them at night and it hurts my foot.” and “Oh, thanks for picking up those toys. I really appreciate it.” (I’m assuming there’s actually a reason for picking up the toys; if not, why are you asking the kid to do it?)

A good rule of thumb is to phrase things in the way you would to a friend or something. If I ever said “Good job!” to a friend for picking things up, they’d probably smack me in the face. (Quite rightly!)

posted by Aaron Swartz at April 26, 2005 03:34 AM #

jg: As I said above, my view is that praising your child for doing the right thing isn’t a reward, or rather it’s not an artificial reward. It saves you from doing it yourself, so of course you’re happy with it, and you should show this. Likewise, if your child is scrupulous with dental hygiene, he’s more pleasant to be with, and you save dental expenses, two great reasons to be happy.

Now another thing entirely would be if you offered a monetary reward, or threatened to punish the child. In both cases you are laying down arbitrary rules of your own invention: if one cleans the room one gets money - completely illogical. If one doesn’t one gets hit by one’s parent - even more unreasonable. The parent is not a reward machine. The parent has a choice to not hit you even if you are bad, and a choice to be nice to you. This is perhaps the most important thing the child has to understand, and one thing most adults don’t understand, that we always have to take responsibility for our own actions, no matter what other people have done.

posted by Harald Korneliussen at April 26, 2005 03:56 AM #

Subscribe to comments on this post.

Add Your Comment

If you don't want to post a comment, you can always send me your thoughts by email.

(used only to send you my reply, never published or spammed)

Remember personal info?

Note: I may edit or delete your comment. (More...)

Aaron Swartz (