Hello Ned,
Glad you carried on the discussion.
I’ll add a few more comment to some of the great points you’ve made. Some of these I agree with you, and of those points where I disagree, I will try to explain the differences.
(a). First, social network connecting the whole planet is actually not a theoretical construct. There are pretty convincing hard data and evidence for this. But I guess that depends on what you mean by a social network. This proved a point that in most cases we need a definition. I guess that is just ingrained in my mathematics training. Because what we can conclude depends on the precise definition. In traditional sociology literatures, social network are often defined to be a collection of entities that are connected by the relationship among them. So a person’s social network includes all his friends, relatives, colleagues, acquaintance, both online and offline. This basically includes everyone that a person knows on a first name basis.
Second, the online social network (e.g. Facebook, LinkedIn) is really just a social graph that represents a part of the whole social network, because it only show a part of the social network by accentuating a particular type of relationships (see Social Network Analysis 101). And these online social graphs certainly that does NOT connect everyone on Earth.
I believe that a rigorous definition is important if you want to do some empirical and theoretical work. Anyone can simply talk about these topics casually, and a general understanding is sufficient to get the idea across without the details. But for deep theoretical abstraction and synthesis based on empirical data, we often need a precise definition in order to make accurate and rigorous predictions. For example, If I were to take your perspective and include the notion of interaction or influence, then I would define something like interaction network or influence network. It is probably more descriptive that way, and we don’t have to qualify which definition of social network we meant and whether our data is generally valid for which definition of social network.
(b). I agree 100% that social network is not static. In fact there is a whole field of knowledge pioneered by CMU professor, Kathleen Carley, called Dynamic Network Analysis (DNA), which combines statistics and time series analysis with social network analysis. Another very interesting topic that I would love to cover in a later post.
I believed what you said is mathematically equivalent to mine. Only under my definition of personal social network, I would simply said that everyone has full control over who he wants to be friend with b/c his personal social network include only those who are directly connected to him. But he may not have control over the 2nd degree network (maybe we can defined this to be his personal interaction network), or the 3rd degree network (maybe we can define this to be his influence network).
(c). I agree 100% with this too. In fact the relationship that I will focus on is the one that are reciprocated. If you are familiar with social network analysis (SNA), the John Doe vs Bill Gates distinction will naturally come out from an eigenvector centrality analysis of the social network. Influence can bleed over to graph neighbors on a the social graph, and this is captured by SNA as I’ve illustrated in an earlier series of blogs on influencers identification, this particular post shows you the result of the computation.
Thank you so much for taking the time to voice your view points. They are definitely valuable and offers a different perspective. I hope we will have opportunities to meet in the real world to discuss some of these interesting topics. Hope to see you back on Lithosphere later.
.
Hello Ned,
Glad you carried on the discussion. I’ll add a few more comment to some of the great points you’ve made.