Comments on: The Golden Rule in the Wild

By: Nick

Nick — Tue, 16 Mar 2010 11:32:20 +0000

Great article.
I have an interest in this kind of thing from an economics background, by I have a tiny understanding from your biology context.
I’ve sometimes wondered whether a species’ nature to co-operate comes not only out of an will for self-survival, but something collectively passed down from a group i.e. organisms survive or evolve collectively based on their nature towards each other and other species – is that one of the conclusions here or just an obvious fact to a biologist?
-my head starts to spin as I’m thinking about the internal game going on: more trusting a species is the more one of its single agents gain by betraying that trust – perhaps as a result for us humans, for example, you have that strange political emotion we all have; to take revenge (or justice) beyond our self-interest, the strange thrill of ganging against a betrayer, etc. as the surviving model of max trust, minimum vulnerability.
Anyway, great article again. I’d love to see a third!

By: Geoff

Geoff — Mon, 08 Mar 2010 16:23:45 +0000

Hi Vorg – I didn’t run the simulations myself, so this is the extent of the data available to me. I would guess that Tit-for-Tat would eventually lose completely and Pavlov would head to 100% as Pavlov is much, much better at dealing with noise. Each strategy will either interact with the other or itself. With noise, Pavlov gets along well with both Tit-for-Tat and itself, avoiding all “death spirals”. Tit-for-Tat cannot get along with itself in a noisy situation however, and I believe this to be too much of a drag on the strategy’s success as a whole.

By: Vorg

Vorg — Mon, 08 Mar 2010 09:39:05 +0000

The simulation of 50 turns doesn’t show any equilibrium. Do the Pavlovs reach 100%, or do they eventually coexist with the Tits-for-Tat in some proportion?

If a coexistence is reached, perhaps it’s similar in concept to the “beehive” of DashingLeech’s comment to your previous post. In the beehive were multiple organisms that could recognize each other: some were “takers” and some were “sacrificers”. When these two types met, the sacrificer would always cooperate and the taker would always defect. When either played against an organism outside the beehive, they would defect. I think this beehive better models the real world, where organisms form “societies”. In a society, a few organisms are the alphas in the societal nucleus, most organisms are surround them in a societal “cytoplasm”, and a few make up the societal “membrane”. The nucleons will usually let those in the nucleus be sacrificed, while the losers in the membrane will stay loyal to the nucleus, no matter how many times they’re betrayed.

Perhaps someone needs to run the simulations of different types of “beehive” structures to see the optimal structure of a “society”.