Tuesday, March 23, 2010

The Complexity of Cooperation

The Complexity of Cooperation Last week when I wrote about the importance of being nice, I was undecided where to place open notebook science (ONS) in terms of the prisoner's dilemma. Since then, I've read Robert Axelrod's follow-up to The Evolution of Cooperation, The Complexity of Cooperation. In this book, Axelrod moves on from the two-person prisoner's dilemma to consider situations much more applicable to the real world - multi-party interactions, significantly without the assumption that the parties involved will always take a rational approach (= noise). He also considers such influences as norms (behavioral expectations and cues within a society or group), and emergent properties of complex systems. Irrational behavior, societal norms, altruism? Suddenly we are into ONS territory.

Unfortunately, as Axelrod's simulations approach reality, the simplicity he was able to bring to The Evolution of Cooperation tends to evaporate. Most of what I know about Nash equilibria I learned from A Beautiful Mind. Nevertheless, Axelrod's landscape theory approach does have valuable things to say to those considering ONS, but on balance, it's more difficult to whole-heartedly recommend that everyone should read this book as I did with The Evolution of Cooperation.

Take home message? ONS is mathematically justifiable, but it's always going to be a risky pathway unless and until societal norms in science change. A good example of this are the ridiculous patent laws, which invalidate ONS discoveries from patent protection. In reality, few organizations have deep enough pockets to establish and defend serious patents, so the beneficial effect of openness is to deprive others of patents through prior publication. I'd count that as a win.