I suppose I was implicitly assuming "in an economic manner". You're very right that humans do things that narrow, economic game theory can't account for (e.g. the RAND secretaries playing Nash's "So Long Sucker").

To put my argument in context: consider IRC. An economically motivated BFT system would give the IRCops and channel ops some kind of quantified share of power that they stand to lose if they defect. But doing that, I think, would not just be a difficult programming exercise; it would also undermine the ops' intrinsic motivation and lead to a worse environment.

So the likes of Bitcoin work well if the protocol is designed for (game-theoretically) selfish participants whose interests lie mainly in the number going up. But that's a very narrow niche. Your genealogy site would probably suffer if it had a "number of family names added" counter and the users' status was measured entirely by such a number.