I know who you know

I've been having a discussion on .talk that has me thinking, not for the first time, about social clustering, or the way in which the people you meet seemingly randomly tend to be already acquainted with other people you know. I used to remark on this in college. When I was an undergrad at Penn, chances are if someone was involved in the honors program, one particular college house, the literary society, or the newspaper, s/he was also involved in at least one other of these organizations. I did two of these things, and because of the way the social dynamics worked out, I knew many many people involved with the other two as well. Nearly all of my friends were involved with at least one of these things, and most of them were involved with two or more.

At graduation, when the whole class was together for the first time since freshman orientation, I found it fascinating to see how the social dynamics of our larger group had played out. It seemed like our class of 2000 some odd people was mainly broken down into clusters of a few hundred people apiece who were loosely connected via some social network.

Now that I'm in a new city where I've only lived for a couple of years and hence don't know as many people, this clustering effect is even more apparent. It's a big place, but the people I'm meeting often tend to know each other, and because I don't know as many people I find the effect even more striking. What seems random isn't. Because you're drawn to people who share your interests, and within each bloc of people sharing a particular interest you're more likely to connect with those who share still others of your interests. The reason it's striking to me here in Seattle is that I haven't been here long enough to feel really ensconced in any particular social cluster... and yet it turns out that I am.

Why does it happen that, of all the runners in Seattle, the one I get to talking to at a race happens to have a graduate degree in linguistics and knows many of the same people that I do? Because I promise, I know from first-hand experience that these sets don't tend to overlap. (There's her, and me, and a professor at Berkeley who runs ultramarathons. That might be all.)

At some level it isn't that strange, because we tend to prolong interactions with people we sense a rapport with, and what is a rapport based on if not shared world view, interests, etc? And by the same token maybe we don't prolong conversations (or find them memorable) when there's nothing to say.

I think this social clustering effect has something to do with the milieus in which we find ourselves tending to attract like-minded people, but that's not all of it. How many thirty-second small talk conversations do most people have in a year? A whole bunch. How many of them do people remember? Very few. The ones you remember, the chance meetings that strike you as noteworthy, are exactly those meetings where you do find something in common, where the conversation lasts longer than a couple of minutes, and where maybe you're inclined to repeat it. It's not just that we only tend to meet people like ourselves, but also that, of all the people we meet, the ones we tend to be going out to lunch with a year later are the ones we had something in common with from the get-go.

I'm not too familiar with the literature on this kind of social clustering, but from what little I have seen I find it really fascinating.