< October 2005 >
2 3 4 5 6 7 8
Mon, 03 Oct 2005:

I think a little context is necessary before I embark on this rant. You could stick around and figure out who (Freud isn't) said the infamous 'is bunk' as was liberally quoted in Brave New World. On the other hand, you could just read on. This should be educational even if you disagree with me.

I re-re-read Axelrod's Evolution of Co-operation over the weekend. The book talks about why humans in a social situation co-operate instead of systematically screwing others. The book also deals with how simple selfish evolution makes sure that helping others would be helping themselves. In short, it explains why you might help someone instead of giving him the finger and exactly when you are likely to do the opposite.

For co-operation to evolve, there are two things necessary.

  • Induvidual recognition
  • Long term memory
First of all, you have to recognize the other induvidual you are dealing with, secondly you have to remember the result or experience during your last encounter with the particular induvidual. These two are the basic blocks needed to build co-operation into a group of such induviduals (remember this when you reach the ending paragraphs).

The next rule of the game is easily visible in a simple game called Prisoner's Dillema. The game is very simple, there are two players who can each chose to DEFECT or COOPERATE. If both players co-operate they win more than if one had defected. Defections from both sides gives more rewards than if your opponent defected while you were co-operating. Let's put some numbers on that.

Co-op  | Co-op  ==  200  |  200
Defect | Co-op  ==  100  | -200
Co-op  | Defect == -200  |  100
Defect | Defect ==   50  |   50

So eventhough the two players are likely to win more if they co-operated, they stand to lose if the other person defects against him. Basic probability theory states that to avoid losing you have to always defect. Notice that this is a Non-Zero Sum game, both players can win at the expense of the banker (or me - who's paying out/collecting points). But the simple mistrust between the two players might prevent them from co-operating in a single game of Prisoner's Dilemma.

Now, let's take this up a notch. We're having multiple rounds of the game with the same two players. It is worth noting that these two players are held incommunicado and do not know what the other person is likely to play next. But if we tell these guys that we're having a 100 round game of Prisoner's Dilemma, 100th game is the same as the single version. Since they already are going to defect in the last, it virtually becomes a game of 99 rounds and so forth. Essentially if they had any idea of when the game's going to end, it's virtually a single game in terms of analysis.

Now, what if they have no idea when the game is going to end ?. Then they won't be able to predict behavior of the opponent in any manner. This is exactly when co-operation can be built up. A defection will or can be responded with retaliation by defection from other side, in a following move. Sure, you can continue defecting and play safe, but that's a dead end strategy in the entire game plan. On the other hand co-operation from the opponent starts of a similar cycle where trust is built up, until one of you decides to defect again.

To actually test this theory Axelrod called for programs which have a given method which will return the next move when fed with the current move of the other player (which is also a program). These programs played against each other for an unspecified but constant number of runs. Guess which program got the most number of points ?. It was a simple Tit-for-Tat program (5 lines !!!) which won the tournament. The program just had a return opponent_move; which survived against a variety of opponents and even in a population of its own.

Ok, so my point was that the most successful form of co-operation features a threat of retaliation. So once there is no threat of retaliation, your opponent is likely to defect. This is essentially what I meant by the last move in the known round situation.

Scarily, this is very closely followed in a lot of places as stated. For example, look at what happened during Katrina in New Orleans. There will be a small (or even large) population who do not commit crimes (do not DEFECT) due to the deterrence of retaliation (execution for murders, siezure of property for fraud). This group of induviduals do not DEFECT because they have no idea when the game ends and co-operating right now seems to be the more profitable option. Make the game of one round only (when threat of death is near for example), these cheerful souls will happily defect and leave the co-operating population poorer and feeling foolish by some. If that wasn't bad enough, the only way to survive in a one round game is to defect yourself as well. Before you know everyone around is in I got mine. So, f*** you ! mode. The foundations of society viz Co-operation has broken down.

Ever since basic society evolved from purely selfish motives, we humans have fought against the basic evolutionary urges - kill your neighbour and covet his wife, steal his money and buy better food for your children, leave him hungry if you find him starving - all of which improves your immediate evolutionary future. Thanks to the basic threat of retaliation, we have stopped short of mutual destruction - but we have gone further to build something into our conscious brain which enforces wholly imaginary retaliation from inside when violated. In short, we have invented Guilt.

You only feel guilty for what you think you've done wrong. Integrity is that one bit which actually defines what you think is wrong and what is not. Cheating in an exam is wrong for some, while it's absolutely OK for some others (watch Lisa gets an A). It is a purely personal standpoint and then there are gray areas.

Finally, let me get to the point. How do you keep lazy people honest and productive members of society ?. You make sure that it takes more effort than just a guilty conscience to steal/cheat and on top of that pile consequences to weed out the really stupid who finally do attempt. You put locks on your stuff so that it doesn't get stolen. It is a sort of arms race which selects better and more careful owners while breeding a new generation of better theives as well. A better mousetrap always ensures a better mouse as a by-product - fewer of them in the short term maybe, but only the fittest survive. Evolution is simply beautiful.

Morover, to select for better owners the owner has to have some say in the strategy he follows - in other words, Responsibility for screw-ups, Credit for successes. In other words, induviduality. Unfortunately this is not where my bit of world is headed - You will be assimilated, Resistance is futile.

Most of the world has little choice on whom it welcomes in. Citizens are born into citizenhood, not confirmed after an integrity check up after the age of majority. The place where this particular problem doesn't apply is in the corporate world. Any company hiring, has a face-to-face interview where the person's teamwork and personality are reviewed. Unfortunately, the little time that was spent with the person might have been about whether the chap gets paid 5 lakhs or 5.2, instead of examining his priorities in life - God, Family, Career, Money. Most clubs faced with a similar situation goes for a peer review rather depend on the evaluation of a single induvidual. Black-balling is a simple strategy to ensure that the people being admitted into a club are of compatible, at least superficially, to the majority. The essential problem with majority votes being that people are sheep - most people don't actually have any strong preferences. They don't protest, they don't speak out, they just sit tight, think as a mob and irrationally hope it all gets better.

Silence is often Golden - fate of those who speak out are often worse. Herman Goering has so insightfully commented on human nature - All you have to do is to convince that they are being attacked and these new steps are for their protection - denouncing any opposers as being unpatriotic. Think of the PATRIOT act in US and the TERRORIST act in the UK. Why exactly would I oppose the PATRIOT act ?. Unless I wasn't ... You get the idea, don't you ?.

I will not approve of any law based on an assumption of crime and prevention of the afore-mentioned crime. I will suffer it, if only for the sake of my daily bread, shelter and survival. I will not celebrate it as a positive act of prevention. I will not enjoy it, nor will I want to afflict it on one of my own without adequate warning. I will not neglect the rules, but neither will I ignore them. There is much evil in this world worse than the indifference of a good man.

These are my principles. These are not for sale. They are not negotiable. They are what I didn't carry from my cradle, but what I will carry to my grave. They were not mine to create, but are mine to destroy. They are why I said No, they are why I said Yes. They shall live on, long past I'm gone, to serve as a beacon or warning, I know not what. Call them what you want - Religion, Truth, Love, Philosophy - but they are.

* the sleep of the Just and rest of the Valorous

posted at: 02:19 | path: /rants | permalink | Tags: