Sunday, April 29, 2012

Ant colonies and the prisoner's dilemma; an organizational dilemma


E.O. Wilson, an evolutionary biologist (entomologist) working at Harvard, was recently the focus of a New Yorker article on ants and altruism. Wilson is an ant man. He’s spent a career studying and writing about them. He’s one of the most accomplished scientist in his field and made his name half a century ago when he identified the scent-producing gland in the ant enabling colony members to find their way to food and back home. Altruism is a double-edged sword. A group that acts purely altruistically will be destroyed by its competitors. As an extreme example, a “for profit” enterprise that gives its profits away won’t compete for long. What about at the individual level within a group? A group composed of individuals that work at least partially selflessly will promote the interests of the group before themselves and make the group stronger. At what expense? Most likely individual gain. A group that is composed of individuals acting purely selflessly may eventually be carved up by competitive individuals who join the group. Or, perhaps, the culture of the group will be sufficient to exclude them; the selfish intruders simply leave. An ant colony is a great model for thinking about this: within the colony, the ants work toward a common goal and show extreme amounts of altruism. These colonies, admittedly, are formidable feats of construction; some are enormous. At the colony level, however, altruism is not the driving force. These colonies are expansive, territorial and ferocious; not terribly neighborly. They are certainly not altruistic. So the dynamics that mediate individual relations within a group (balance of altruism versus selfishness) are not the same dynamics that mediate group behavior.

In game theory, the “prisoner’s dilemma” is a great way to model the individual component of this dynamic. Briefly, two individuals are detained for a crime and are not allowed to communicate. They are each offered a deal: if both prisoner A and prisoner B confess, they each get 3 months of jail time. But if one prisoner confesses while the other remains silent, the “defector” who confesses will go free while the silent one gets a year in jail. However, If they both remain silent, they each get one month. What to do? The dilemma for the prisoner is that the optimal outcome for the individual is to “rat out” your partner and go free if your partner doesn’t confess. The optimal outcome for the group (both prisoners) is to keep your mouth shut, expect your partner to do the same, and serve minimal time. But what happens in reality? When the game is simulated, prisoner behavior is to maximize individual gain and confess, yielding a sub-optimal outcome and 3 months of jail time. Within organizations, an individual will “remain silent” and act altruistically if there is something to be gained - but only if other individuals are motivated to behave similarly. The selfless individual within a group who is surrounded by selfish people will be eaten up (like the prisoner who remains silent while the other confesses). How does an organization surmount this obstacle? Clearly, an organization that has a component of altruistic behavior at the individual level will be more competitive at the group level (don’t believe me? remember the ant colonies?). Here’s where organizational leadership and “vision” play a role. You think all “mission statements” are mere window dressing? Wallpaper? I’m referring to  real structural ones – the ones that hold up the building and define the organization (if you’ve never worked for an organization like this, trust me, they do exist). The provenance of this vision is the leadership of the organization. A principal role of an organization’s leadership is to foster an environment where individuals frame their work in terms of the organization’s goals, not the individual’s.