I’ve been wanting to read this book for a while (after I saw an article by Watts, my interest level was even further heightened), and when I saw it on Elizabeth’s shelf in Oberlin, I started reading it. Fortunately, she was kind enough to let me borrow it when I hadn’t finished before leaving, and I finished it on the long plane ride home.
Watts is one of the scientists exploring the science of networks. The title comes from the legendary small world experiment of Stanley Milgram, where he estimated that every person in the world could be connected by links of six people or less. It’s a fascinating result. And it seems highly improbable.
Watts takes us into the math of networks and explains how certain properties of networks can give rise to the small world phenomenom. Not only that, but he demonstrates how such network models have applications in fields ranging from epidemiology to stock markets to power distribution networks to corporations. Really fascinating stuff. And it’s exciting because they’re still at a very preliminary stage of understanding, so there’s a lot of exploratory work to do.
It was almost annoying to see several ideas I’d been having trouble expressing laid out on the page in front of me. Things like perceiving networks of friends as overlapping clusters with myself as the locus. I’d been thinking of this recently when I was at a wedding of a friend, and there were many folks there who I’d gone to MIT with but with whom I’d lost contact. But because I was still in contact with the groom, I was able to reconnect with all of them.
Another idea that struck a chord was the use of social groups to define a person:
Imagine that instead of individuals in a population choosing each other directly, they simply choose to join a number of groups, or more generally, participate in a number of contexts. The more contexts two people share, the closer they are, and the more likely they are to be connected. Social beings, in other words, never actually start out on a tabula rasa in the same way that the nodes in our previous network models had done, because in real social networks, individuals possess social identities. By belonging to certain groups and playing certain roles, individuals acquire characteristics that make them more or less likely to interact with one another. Social identity, in other words, drives the creation of social networks.
You can see the similarities to the concept of reality coefficients. Or to the Orson Scott Card quote I refer to in this old ramble: “Every person is defined by the communities she belongs to and the ones she doesn’t belong to.” This is one of the reasons I think that a site like tribe.net, with its emphasis on groups, makes more sense than friendster, and its emphasis on individual connections.
I also liked the application of networks to corporate structure. Watts points out that the hierarchical corporate structure of the Industrial Revolution were designed to maximize returns on specialization, where economies of scale were able to support people specializing to an extreme extent. He posits that today’s world, with its fast-changing requirements, no longer supports such economies of scale, and suggests that it’s time for a world of flexible specialization, which promote economies of scope – “Flexible specialization relies on general-purpose machinery and skilled workers to produce a wide range of products in small batches.” The reason I like the discussion of economies of scope so much is that it gives me hope that industry will soon learn to value me as a generalist.
A similar quote is this one:
…it appears that a good strategy for building organizations that are capable of solving complex problems is to train individuals to react to ambiguity by searching through their social networks, rather than forcing them to build and contribute to centrally designed problem-solving tools and databases.
This is the sort of thing which plays to my strengths. I have a very good memory, which seems to work associatively, where input will often trigger my memories of people who might be interested in that input. I also feel that hierarchies are often the worst response to an ambiguous situation, because hierarchies and processes only know how to deal with situations they’ve faced before. So you can see why I like what Watts has to say.
Interesting book. I should probably get my own copy, and go back and re-read it at some point when I’m so distracted. And start to read through some of the other resources that he points to. Argh. So much to learn, so little brainpower.