Two points...
POINT ONE: The cholera example is even more fascinating when you drill down. The bacteria involved is “Vibrio cholerae”.
http://www.textbookofbacteriology.net/cholera.html
It seems to actually have numerically common non-pathogenic forms and the ones with enterotoxin genes appear to have received them from bacteria targeting viruses. If I understand correctly, the toxin genes are integrated (but dormant) within bacterial genomes and infection by bacteriophage CTX triggers their expression.
http://www.mrc-lmb.cam.ac.uk/genomes/madanm/articles/cholera.htm
POINT TWO: It is probably worth keeping in mind the fundamental attribution error.
http://www.jstor.org/pss/4545312
This point is mostly in response to the focus here on habits and norms. Not to say that someone couldn’t work on those productively, but I suspect environmental effects like “mere proximity” have a lot more influence than would be assumed without consciously factoring it in, even over people in this community. The cholera example comes bundled with the same “context focused” message: the authors cited in the OP mostly focusing not on hand washing but on the design of water purification infrastructure.
I can’t imagine that this crowd is unaware of this sort of thing, but I’m not aware of a better example of the “location location location” message this Google’s results from studying their internal betting markets.
http://googleblog.blogspot.com/2008/01/flow-of-information-at-googleplex.html
For myself, I tend to assume that if changes to my habits are to have any significant influence on me, many of them must be focused around shaping and choosing environments that support the kinds of thinking and living that I want to do. I’m still working on this process for myself and have few unambiguously positive results to report and the negative results are too embarrassing to list and would take a lot of text to describe any in useful detail :-P
For lack of such text I’ll recommend “Lady of Mazes” for it’s exploration of themes around technology, “spacial nearness”, social networks, medium-message distinctions, choice architecture, suggestions systems, personal character, political awareness, value-technology interactions, and having a life that is felt to be meaningful. This book is less accessible than “Accelerando” but, for me, it has had much more staying power.
http://www.amazon.com/Lady-Mazes-Karl-Schroeder/dp/0765312190
http://www.amazon.com/Accelerando-Singularity-Charles-Stross/dp/0441012841
I recently decided to try reading this blog to see what the fuss was and this leapt out at me:
“At that time, there were no Conspiracies, no secret truths; as soon as Eld scientists solved a major problem, they published the solution to the world and each other. Truly scary and confusing open problems would have been in extremely rare supply, and used up the moment they were solved.”
It occurs to me that Mr Yudkowsky is proposing that having science (or scientific fields) incorporate something like a “hidden secret” of the sort mystery religions use would actually be beneficial for science or the world. It’s not the first time I’ve heard the idea connected to him.
Also, I’ve heard interpretations of Noam Chomsky’s early publishing tactics in linguistics described in roughly this way (not publishing enough to replicate his work, letting special people in on the secret who then publish papers based on it, immunizing his theories from disconfirming argument by explaining that critical papers aren’t criticizing the full true theory (the one he was working on but hadn’t yet published to people not specially selected and sworn to secrecy)). I’m not sure if this is is true or not. It’s academic gossip mostly.
But it might explain why linguistics is in such a “pre science” state even now, with many competing paradigms co-existing in the community so that linguists spend much time re-arguing fundamentals and relatively less solving puzzles about language. At the same time, Chomsky’s citation tree is breathtakingly large. Tenuous conclusion on the mystery cult tactic from this example: good for Chomsksy, bad for linguistics?
Other than modeling experiments, it’s hard to even test the theory because the object of study would be scientific communities and it would be difficult and (ahem!) ethically dubious to experiment on them… but the thought is worrisome when bearing in mind that the payoffs of the dynamic are (on first glance) structured like an N-person prisoner’s dilemma with no obvious regulatory agent.
If this is a failure mode of scientific disciplines, it would tend to occur where someone unilaterally broke with the cultural norms of academic science.
EDIT in 2023: To augment the link, with a residual broken one, and a better link to a hopefully more stable archive that helps maintain the reliable infrastructure that supports Bacon’s Project.