multi-centered networks

Duncan Watts is critical about centrality. He cleverly ponders the 'weight' of peripheral players and information brokers not stereotypically believed to wield the most influence. In his mind, it's the interactions of equals more than a predetermined 'hub' that create the innovative and emergent outcomes we all attempt to chase down.

"What if small events percolate through obscure places by happenstance and random encounters, triggering a multitude of individual decisions, each made in the absence of any grand plan, yet aggregating somehow into a momentous event unanticipated by anyone, including the actors themselves?"

This has so many implications for those interested in divining an algorithm for said emergent outcomes. And of course, so many social psychological ideas embedded within... The gist is, be wary of the structure within which you're working before you go imposing or assuming a given structure. In other words, study the group dynamics and realize there may be unique circumstances to account for: particularly strong or weak ties across people and groups, peculiar politics, lack of policies, implicit morays... and furthermore active flux. Networks are rarely static.

Groups are complex beasts. When improperly nurtured, we see things like groupthink, social loafing, deindividuation, conformity, diffusion of responsibility, polarization... Groups gone wild... But groups are important and replete with latent assets; Durkheim, for one, emphasized the importance of such social integration. Much research points to the health benefits of social connections in groups, not to mention the wisdom of crowds.

I'm a big advocate of the proper understanding and mapping of existing groups and networks. Prevent anomie! Figure out a way to take advantage of the rich peripheral activity that Watts rightfully calls our attention to. There are a lot of important voices out there that aren't in the spotlight. Any Gladwellists care to challenge me?