At the Personal Democracy Forum in NYC last week, I enjoyed a fascinating interview-turned-rambunctious-conversation with Marc Smith, a social network sociologist who founded and managed the Community Technologies Group at Microsoft, and is currently the Chief Social Scientist at the Connected Action Consulting Group. This wasn’t the first time I’d communicated with Marc; we’d corresponded via email more than two years ago about AURA, a social shopping app he was working on at Microsoft which I was writing about in my master’s thesis on emerging uses of social media for consumer governance. So it was a great pleasure to finally meet Marc in person (which his gregariousness makes more than worthwhile). Following is a synthesis of our encounter.
Because he explicitly specializes in online collaboration, I started by asking Marc about the difference between online and offline collaboration, whether he thinks this difference is becoming less meaningful, and whether he foresees the emergence of another typology (or multiple typologies) for categorizing types of collaboration. He first responded with what many of us who engage in any type of collaboration well know: that online collaboration is increasingly part of offline collaboration, which – in response to the second part of my question – does make this difference somewhat less meaningful. But he continued with what many of us are also aware of: that there are meaningful differences between online and offline collaboration. For instance, online enables global distribution, discoverability, and reproducibility at negligible or no cost, and although online is more costly in those three realms, it has its own obvious affordances. “Online means you have to come to the conference, you have to have a booth. It’s the difference between tweeting with hashtags and having to make and put up fliers.” (At that point, he plugged the SMR Foundation, pronounced “Smurf Foundation,” which “charitably builds tools needed to catalyze scholarship to understand social media and collective action in the modern world.” It was as if to say, ‘if you’re interested in this line of questioning, support Smurf.’ I’m sold.)
There are three traditions in sociology for understanding collective action, he continues. First, Elinor Ostrom’s, Robert Axelrod’s, and Peter Kollock’s work on offline and online cooperation; second, the field of interaction sociology; and third, social network analysis. Combine these three together and “bang! You get a lens for understanding what gives on the net.” And combining them for purposes of understanding online collaboration, he adds, “any object in cyberspace that attracts more than one person is a collective good.” Beautifully put.
Marc then distinguishes between two types collaborative activities: ‘give some’ and ‘take some.’ ‘Give some’ asks who will write the Wiki pages, i.e. how resources will be provisioned and what will incentivize collaborative behavior. ‘Take some’ asks who will stop spamming an email list, i.e. how to prevent a tragedy of the commons and moderate anti-collaborative behavior. He acknowledges that the online domain still has limits, such as in bandwidth, storage, and the precious resource of human attention, but emphasizes that “online [allows for] an economy of digital objects that allows us to aggregate vast amounts of data, [which] can create new resources and bring to bear the attention of people worldwide.” Bringing together all of the above, Marc conceives of the difference between online and offline collaboration primarily in terms of scale – the online domain can leverage more data and more people at a greater geographic expanse. Instead of completely conflating or separating online and offline, it’s refreshing to hear an operational model that articulates the differences between the two, and can be used decide when to employ one or the other, or more likely, how to integrate both.
To probe further into how Marc differentiates between types of online collaboration, I confess my qualm with Clay Shirky’s Here Comes Everybody. Given that in 2002, Howard Rheingold’s Smart Mobs introduced the phenomenon of technology-enabled collective action and gave it a name (the name of the book and this blog), I figured Clay’s treatment of the same phenomenon, published 6 years later, would somehow advance the body of theory around it. Instead, IMHO, Here Comes Everybody simply re-introduced the phenomenon of smart mobs but by another name (“organizing without organizations”) and offered a few more recent case studies. (This doesn’t mean that I don’t consider Here Comes Everybody a compelling and well-researched book, which I absolutely do, but that I don’t think it significantly advanced the existing body of theory around technology-enabled collective action. And yes, I recognize that this is somewhat of a blasphemous proposition.) There are instances in the book where Shirky almost advances theory, by distinguishing between organizations with few organizers, in which a small group of individuals coordinate technology-enabled collective action on behalf of a larger group of otherwise unaffiliated individuals, versus organizations with no organizers, in which there are no individuals at the helm, but merely a group of otherwise unaffiliated individuals leveraging technology to engage in collective action. I say almost because this distinction could have evolved into a model of different types of smart mobs and, for example, described which types emerge under which circumstances or predicted how smart mobs change through time.
With that preface, I asked Marc how he distinguishes between different types of online social networks. He replied that answers can be found in Sociology, which studies “the range of different kinds of human associations. There’s multi-hubbed and single-hubbed; there’s temporary and permanent. Social network analysis makes this more of a science and less of [an exercise in] intuition.” This point gives him the perfect segway into NodeXL, an open and free tool that enables social network analysis of email, Twitter, Flickr, Facebook and other network data sets through an Excel spreadsheet (soon to support other formats, see image above). Marc explains that NodeXL is a snapshot camera, and that my question about predicting the evolution of social networks is a film camera (though we agree that NodeXL can make stop-motion animation). Prediction is a major direction for NodeXL’s work, and a hot area of network theory. “Who’s most likely to connect next? Facebook assumes triadic closure,” i.e. if A is friends with B, and B is friends with C, then C should be friends with A, which is incorrect (think competing political candidates with many of the same friends). Conversely, “NodeXL allows for triadic census,” i.e. different patterns of social network ties that don’t assume transitive properties. (Due to this discovery, he adds, merely saying “triadic census” will get you funded nowadays.) NodeXL incorporates the recent advances in social network theory in a way that Facebook does not, clearly a reflection of its architects.
Beyond categorizing types of social networks, it’s possible to categorize the roles people play in them, akin to distinguishing between primary, secondary, and tertiary producers in ecosystems. Marc references a recent article in the Journal of Social Structure that distinguishes between four patterns of connection, and in turn, four roles people play in social networks: 1) discussion starter people (people who start the discussion), 2) discussion people (people who participate in it), 3) question people (people who ask the question), and 4) answer people (people who answer it). The same people can play different roles in the different social networks (or shift their roles within the same social network). There are other meaningful ways roles could be categorized, but this is the kind of operational thinking that people interested in social trends – from sociologists to online advertisers to campaign managers – fiend over. Meanwhile, I’m thinking about how to categorize the types of relationships between people in social networks, again using ecology as a metaphor, akin to the types of relationships between organisms in ecosystems – mutualistic, commensal, and parasitic. By moving from a binary conception of relationships (yes/no, à la Facebook) to a typological one (types of relationships, à la ecology), we’d effectively move from a social web to a semantic social web. I silently wonder whether NodeXL could elaborate a semantic social web, imagining a conversation between Marc Smith and Tim Berners-Lee.
Moving onto the implications of us being able to understand the structure of social networks and our role within them, I ask Marc if he foresees some kind of bio-feedback (socio-feedback?), whereby in becoming aware of our social networks, we become able to transform them. He smiles at this idea. And then laughing but earnestly, he suggests the possibility of looking at the social networks of people we aspire to be like, and using them as a guide. We agree that Howard Rheingolds’s social network (which we’re happily both part of) would make a fantastic map for achieving genuine, humble fame. While the three most important things in real estate are location, location, and location, he continues, the three most important things in social networking are position, position, and position. And NodeXL is the tool that allows you to see (and potentially transform) your position. “What if everyone had their own Hubble telescope?” Answering his own question, he offers, “when you democratize a tool you get an explosion of usage. We accept that, like PowerPoint [was responsible for more good and bad presentations], we’ll be responsible for more good and bad graphs.” He jokes that if graphs are bad enough (i.e. exceeded a measure of fitness), NodeXL will let users know, automatically suggesting that they not show the graph to anyone. I joke back that when graphs are that bad, the NodeXL logo should disappear. Bringing the conversation back home, we acknowledge together the importance of democratizing information and communication technologies, like NodeXL, which is ultimately what we came to discuss here at PDF.