About the SA Blog Network

Guest Blog

Guest Blog

Commentary invited by editors of Scientific American
Guest Blog HomeAboutContact

Game Theory and the Golden Punishment Rule

The views expressed are those of the author and are not necessarily those of Scientific American.

Email   PrintPrint

Moral sciences are back. Natural laws of ethics, envisioned early in the Enlightenment, can now be studied. Scientists are relearning the wisdom of old traditions by objectively rating their performance. And they’re suggesting improvements: any rule system is weaker without “The Golden Punishment Rule.”

Humans, being social, can’t live without rules. Certain rules work better. Game theory provides “behavioral telescopes” to study them.

The naturalistic fallacy says we can derive no ethical lessons from nature. But without seeking good and evil in nature, we can compare the productivity and sustainability of behavioral rules—and map negative ethical spaces, which are inherently unworkable, and thus inherently bad.

For example: we can compare how ethical traditions do in Prisoner’s Dilemmas against the game’s best strategy, called Tit-For-Tat, which is an “evolutionarily stable strategy.” As Tomas Sedlacek asks: What would Christians do? Or practitioners of any religious or secularly sourced Golden Rule?

The results are clear: Rationalists do worse than the Golden Ruled. And Jewish preferences beat Christian ethics. So-called rationalists, dominated by some dire logic, produce no cooperation and low productivity. Two Golden Ruled players cooperate, thus beating rationalists. But New Testament turning-the-other-cheek is exploitable (as Machiavelli and Nietzsche complained). Old Testament eye-for-an-eye comes closer to Tit-For-Tat, if forgiveness follows (which might be divine, but is also evolutionarily adaptive). But punishment sufficient to ensure that cheating doesn’t pay must also prevent escalating revenge. Hunter gatherers avoid such feuds by delegating the severest punishment to close male kin. A “Golden Punishment Rule,” that mimics Tit-For-Tat, enables cooperation by sustainably preventing exploitation. Similar logic likely applies beyond Prisoner’s Dilemma.

Darwin, being un-Darwinian, said “social instincts…with the aid of active intellectual powers… naturally lead to the golden rule.” Game theory shows that simple rigidly followed rules can create workable cooperation. Evolution is a game theorist, endlessly testing behavioral strategies and naturally selecting the more productive.

Another religious idea can clarify evolutionary thinking. Richard Dawkins’s selfish gene errs by overusing the non-exhaustive binary of selfishness vs. altruism. Dawkins uses and inverts the Christian framing that promotes self-sacrifice and discourages self-benefit. Jewish ethics, however, encourages self-benefit, but warn against the dangers of selfishness. Rabbi Hillel said, “If I am not for myself, then who will be for me? And if I am only for myself, then what am I?” Hillel’s self-and-other frame includes the win-win space in which cooperation can evolve (and which Dawkins initially ignored).

It’s still early in the use of game theory, but it seems the behavioral universe has gravity-like pull towards certain stable high productivity social rules. We should use our “active intellectual powers” to adjust what’s deemed rational, and to more intelligently design our economic and political (once called moral sciences) systems.

Illustration by Julia Suits, The New Yorker Cartoonist & author of The Extraordinary Catalog of Peculiar Inventions.

Previously in this series:

It Is in Our Nature to Be Self-Deficient
Inheriting Second Natures
Our Ruly Nature
It Is in Our Nature to Need Stories
Tools Are in Our Nature
We Fit Nature To Us: Evolutions two way street
Justice Is In Our Nature
Behavioral Telescope Shows How Cooperation Works
Selfish Genes Also Must Cooperate

Jag Bhalla About the Author: Jag Bhalla is an entrepreneur and writer. His current project is Errors We Live By, a series of short exoteric essays exposing errors in the big ideas running our lives, details at His last book was I'm Not Hanging Noodles On Your Ears, a surreptitious science gift book from National Geographic Books, details at It explains his twitter handle @hangingnoodles Follow on Twitter @hangingnoodles.

The views expressed are those of the author and are not necessarily those of Scientific American.

Comments 6 Comments

Add Comment
  1. 1. DustinWyatt 8:31 pm 06/18/2013

    I always get a kick out of someone claiming “rationalists lose in situation X”.

    If following the Golden Rule in situation X leads to better outcomes, than the rational person will adopt the Golden Rule.

    Link to this
  2. 2. davenussbaum 11:14 am 06/19/2013

    Interesting post — I’m really enjoying this series, Jag, thanks!

    It does seem that game theory is developing quickly into a pretty complex tool for analysis, moving from simple prisoner’s dilemma matrices to simulations of very complex networks — and nice modeling of how nature does it. Provides a compelling retort to the comment (above by Dustin Wyatt) that the rational person will always do the thing that leads to the best response — in fact, collective outcomes can emerge from behavior that is not dictated by rationality on the individual level.

    Yesterday I was getting on to the highway in Chicago, and there was a line of cars in front of me. There’s a traffic light that lets one car on at a time, to manage the rate of cars getting on at any one time, especially during rush hour. Then a taxi (of course!) drove up and tried to squeeze into the front of the line, instead of waiting at the end. Naturally, this drives me crazy, although in this situation I was powerless to do anything to punish the violator.

    The reason I mention this, though, is that it can be very interesting how not everyone needs to follow the same set of rules, or respond in the same way, to produce an “optimal” collective outcome. It’s good for the “taxi” who cheats to get punished, in order to eliminate the selfish sort of behavior that harms everyone else. But you don’t necessarily need a rule for how to do so — all you need is enough variability in how much people take offense at rule violation, and their willingness to act on it.

    In other words, as long as there are enough people who get so mad at rule violators that they’re willing to enforce rules at their own expense (an extreme example: somebody to ram into the taxi with their own car) that you will discourage rule violations. That is, the taxi won’t risk cheating because someone may flip out and ram them, which isn’t worth it. But note that most people don’t need to do anything besides quietly disapprove of the taxi’s behavior — as long as there are some people at the “irrational/selfless enforcer” end of the distribution.

    Link to this
  3. 3. Bill_Crofut 2:37 pm 06/19/2013

    Re: “Evolution is a game theorist, endlessly testing behavioral strategies and naturally selecting the more productive.”


    Link to this
  4. 4. hangingnoodles 11:45 am 06/20/2013

    @davenussbaum Thanks. You’re right the, role of punishment is key, and only a few punishers are needed.

    But as I say in my next Scientific American Mind post, such punishment isn’t socially “irrational.” And essentially all our key traits evolved socially and relationally. No view of what’s humanly rational makes sense unless it fully takes that into account.

    Link to this
  5. 5. hangingnoodles 11:47 am 06/20/2013

    @Bill_Crofut – I hope you will forgive my metaphoric personification of “evolution.” I mean that the process of evolution in effect performs game theory experiments. Different behavioral strategies play out against each other and the most productive survive.

    Link to this
  6. 6. Bill_Crofut 9:53 am 06/21/2013


    There’s nothing to forgive, but your response does raise another question for me: Is there an observable example of behavioral strategies playing out against each other?

    Link to this

Add a Comment
You must sign in or register as a member to submit a comment.

More from Scientific American

Email this Article