The multiplicity of human relations, and how online social networks don’t support that

The following came up during Bernie Hogan’s talk at the WebSci summer school. I touch on other aspects of the talk here, but this topic deserves a post of its own.

This got a touch long: there’s a summary paragraph at the bottom if you’re in a hurry!

How we (don’t) understand our own networks

Hogan reports that when people are asked about clusters in their online networks, they are very able to articulate them — for instance, “friends from Oxford”, “colleagues from my last job”, “university buddies”. However, they find it very, very hard to see how those clusters overlap, imagining that they are non-overlapping worlds when that is not true.

Interesting issues arise from this: how do we present ourselves to networks which contain multiple clusters? If we present to the “lowest common denominator” (the people we trust the least), we rapidly find that all we can really share are baby and holiday photos (not even those if we’re pulling a sickie!). (This paper elaborates.)

So, we can’t share sensitive information with our networks… except we do, don’t we?

And we often just post publicly, not realising (or caring about) the visibility of our updates — not realising (or caring) that they are persistent and linked. For instance, here are searches of Twitter and Facebook for updates with the words “I hate my boss.”

We lose context in online spaces

One tweet during Hogan’s talk remarked “sometimes it’s valid to have multiple social identities, which we don’t want to disclose.”

I disagree: it’s always valid to have multiple social identities (and why shouldn’t they be private?). It’s what it means to be human: no one presents the same behaviour to their parents, their colleagues, their buddies and their kids.

We have no good mechanisms for managing our clusters of contacts... running to new, sparsely populated networks doesn't count!

This is a problem on Facebook, of course: you are ‘friends’ with people from multiple arenas, so how do you handle it? Some attendees described deleting their Facebook accounts because the overhead of handling different clusters of people was just too high.

So, we have issues with broadcasting inappropriate information. The flip side of the problem is receiving it: we see baby pictures from work colleagues, updates about wild nights out from family members, meaningless work-related statuses from friends. Privacy is about more than keeping secrets, it’s about managing information and making it contextually appropriate.

Restoring our context online (I don’t see a lot)

Google+ begins to try and address this with Circles (see my initial impressions), but I’m unconvinced Circles will work. We have many, many people in our networks: aside from the fact that human relationships cannot be easily dropped into buckets, people often aren’t good at categorisation, and maintaining Circles will be a real pain. (A partial solution would be to let people create publicly-available Circles, for example a “WebSci summer school” circle… that would help, but I don’t know if it would be enough.)

Google+ tells us which circles we’re posting to, but Hogan thinks it isn’t made crystal clear. Offline, of course, we have a wealth of cues for understanding our current context. Still, as mentioned by Marc Smith, things are changing: the role of ‘speaker’ and ‘audience’ is crystal clear during talks at this summer school, but the talks are broadcast online, tweeted and blogged. How well do we adjust to that?

In summary

So, we struggle to understand our own social networks, and we fail to share appropriate information with clusters within those networks. People are talking about the problem, and social network providers are attempting to respond… but the people who’re talking are an academic/technical minority, and the responses are very ‘engineered’, failing to acknowledge the complexity of human relations. I don’t have any silver bullets, but this is a tough problem and I think we need to work together as a community — academics and industry, sociologists and engineers — to address this.

Will we see progress in the area come WebSci’12? I hope so.