The ethics of network topology

Posted on February 23, 2011 - wayback

In a previous post, I argued for why it is important to look at the quality of the transactions that lead to the forces of preferential attachment and hence scale free networks. In this post I want to look at the opposite, that is what you can say about the human quality of a social network, by considering it’s form. This should I believe be an entire research discipline, one I’m actively looking for, but not yet found.

Let’s call this discipline: social anatomy, or network aesthetics, or some such snappy title. Whatever we call it, it would be the study of emergent values, of network topologies. There would be a body of longitudinal studies, in which different topologies were compared, randomly allocated to different groups and the quality of the emergent behaviour at an individual and social level compared. Studies would take place within the same domain, comparing different topologies, and then compared across domains to see if general lessons could be learned regarding the role such topologies have on abstract qualities, such as creativity, or empathy, or aggression for instance.

This research would have a practical focus. It would help us design online communities, social media tools, social institutions and legal frameworks. But to keep things simple, let’s take a look at social media, and (for the purpose of illustration), the media reporting of the current Libyan protests.

@techsoc relates:

Is this not an emergent property, of the topologies of the two network structures – television and twitter? What substance can we give to the observation that the professional behaviour encouraged by hierarchical institutions tends towards the well – bland, and the cold? What sort of structures could preserve a wider range of qualities within the diversity of reporting? It is also surely the case that the potential of social networks does not stop with the direct reporting of individual incidents in all their emotive intensity. We have a network, and are beginning to see social editorial structures emerge, both technical, and cultural. It is certainly possible to imagine, and indeed occasionally experience mediated representations of events that combine the two – direct and emotionally intense on the ground accounts, with objective, impartial or statistical analysis. Networks can serve to address some of the shortcomings of professional reporting, by giving a richer structure and typology to the objective representations, while retaining the immediate presence of direct visceral experience. This can be technically mediated by showing an image, or raw video footage, but is also socially mediated by the fact of knowing that the reporter is not professional, is not being paid, but is perhaps being motivated directly and simply by the need or impulse to tell a story. Tweets are based on a format designed by engineers to exchange technical data over a cellphone network – their social and emotive impact comes not from the format, as much as from the topology of the communications network. Tweets are able to capture the emotional context of news stories, in a way in which a television news studio and network of reporters is unable to. A community and culture of reporting can then grow around this network. Is it not the case that we may be able to look at the topology of such networks and map classes of these topologies to a range of social qualities. It would seem to me that there are a number of qualities; authenticity, empathy, humour fairness to take a few, that appear to suffer in any conventional process of institutionalisation. It is no coincidence that the social sciences, lack good theoretical frameworks to understand these qualities, and that research in these areas is poorly funded. Nor is it a coincidence that these areas are of genuine human importance. The important things are hard, and we have not had ethically viable research tools to investigate these core qualities of human society. Fairness may be related to an appropriate use of reputation in a network, but studies like this are only just scratching the surface. We need longitudinal studies, in order to move an academic discipline towards a real science of social institutions in a networked world, not one stuck in an analysis of historical structures, but an experimental science that enables us to look at appropriate designs for qualities that we seek, not simple side effects of historical happen-stance. In medicine we have a range of instruments that allow us to look at the structure of the human body, and detect signs of structural disease, or signs of healing and good health. When will we have these for the social structures of our institutions?

Posted on February 15, 2011

In her essay Can “Leaderless Revolutions” Stay Leaderless: Preferential Attachment, Iron Laws and Networks, Zeynep Tufekci describes why we should not take for granted that the “Leaderless Revolution” will be able to resist a natural process of transformation into “hierarchical and ossified networks”. At first sight, there is little to disagree with here. It is clear that while we can be sceptical about the rhetoric of some of these claims, there is little dispute that flat, horizontal and structureless organisations in nature (including human social organisation) are exceedingly rare, and when they occur tend to be replaced quickly. The geek inside all of us can marvel at the structural similarity between emergent networks in systems as diverse as: Protein-Protein interaction networks. Neural networks and memory Sexual partners in humans, which affects the dispersal of sexually transmitted diseases. Many kinds of computer networks, including the World Wide Web. Semantic networks. Airline networks Brazilian Soccer Players Power generation and distribution grids And of course social networks, including collaboration networks The distribution of wealth But our inner geek can lack a little emotional intelligence when it comes to translating these observations into something of value to the human sphere. To pick up on what is really important here, we need to consider more closely the quality of the transactions that lead to these similar emergent structures, and to compare these (and not simply the commonality of topographical structure), with our deeper political or ethical aspirations. First lets take a step back, and take a closer look at the argument, put so eloquently by Zeynep (my emphasis): Preferential attachment means that a network exhibiting this dynamic can quickly transform from a flat, relatively unhierarchical one to a very hierarchical one – unless strong counter-measures are quickly and firmly employed. It is not enough for the network to start out as relatively flat and it is not enough for the current high-influence people to wish it to remain flat, and it is certainly not enough to assume that widespread use of social media will somehow automatically support and sustain flat and diffuse networks. The problem I have with all this analysis is not that it is wrong, but that is misses the point, and does so dangerously. Dangerous because it reinforces the worst possible aspects of both sides of the debate. It misses the point because the real aspiration here, is the desire to replace corrupt hierarchical processes with a richer fairer discursive democracy. It is my assertion here that a great deal of this aspiration is being catalysed by technology, and that scale-free network or not, we should be celebrating the quality (not the form) of this new form of dialogue. It is clearly the case that technology is enabling people to have new and richer forms of two-way political communication, and for these political conversations to scale in ways, and at a speed, that were not possible before. It is also clear that this process has only just begun, but that technology will continue to drive these changes as communication technology (in particular mobile and smart phones) become more ubiquitous, and new social structures emerge that leverage these new capabilities. Scaling these technology-enabled conversations, may well mean forming natural small world or scale-free networks, but who cares? Isn’t that what we want, should we seek to elect a representative or decide together on a policy? We may well for instance decide together on a policy, without there being any clear (or indeed known or knowable) individual leaders. What really counts is how we form such decisions (the qualities of the conversations), and not simply the abstract shape or form of the social or decision making structure. There are more hidden dangers in this way of thinking. The first is to equate the hierarchical structures of classical social organisations, with the more-or-less stable structures we find in scale-free networks. It may be true that the social connections between individuals, or other entities, has a scale-free or small world network structure, but this surely does not equate directly to the perceived deficiencies in the legal organisational structure? It is the latter that is the real target for most of the criticisms, and not simply the rich structure of natural networks founded on real diversity and freedom of opinion. Referring obliquely to both with the term “hierarchical and ossified networks” is not helpful. There are surely stable network structures (that are a result of fluid and flexible transactions), that have almost nothing in common with rigid hierarchical structures? It is also dangerous because the above arguments are a direct and very effective attack on the hopes of the protesters and the intellectuals supporting the idea of new and richer forms of democratic participation. They support a sceptical “best of all possible worlds” view of politics. The stated assumption, is that “strong counter-measures” need to be “quickly and firmly employed”, but why should it be important to flatten this hierarchy, and is this really something that people want and are calling for? There is certainly a general background demand of activists in this area, and that is the desire that all voices are treated equally, and that ideas are allowed to flourish without party or other power structures suppressing them for their own organisational imperatives. However, I think this demand needs looking at more closely than it has been, as it is not as easy to reconcile a naive interpretation of equality, with a true respect for the far more important values of respect for diversity. For now let’s leave this for another post. Lets translate this argument into something more grounded. Around 2004 or thereabouts, I had the privilege of helping to establish the funding and democratic structure for a network or new media artists in Vienna. The aim was for this organisation to function as a (leaderless) but genuinely democratic network, in which the community itself could debate and decide which projects, or groups the city of Vienna should fund. Naturally there was a lot of fierce debate amongst the community, and this forced me to re-evaluate some of my ideals. Vienna was very different from London with regard to arts funding. In the UK there are a wide range of quangos, charities and other intermediaries to which an aspiring artist can apply for funds. In Vienna it is much more likely that funding decisions are made over coffee, discussing the proposal with a city official (who is in turn directly responsible to a local politician). Now don’t get me wrong, I love Vienna, and marvelled at how easy it was to get real work done, simply by sitting in a good café and having meetings (arranged or otherwise), with key people. Beats the hell out of travelling 40 minutes on the tube to have a once in a month meeting with someone in London. However this system is clearly open to a great deal of corruption, and political influence. It was literally a small world, in which if you new the right person socially, you were much more able to attract funding for the projects you favoured. Hence the genuine enthusiasm, by both activists and officials in the city to look at new ways of allocating funding. Phew! I hope I managed to say that without offending too many people :) Now my background was working with social graphs and voting (transitive delegated voting or Liquid Democracy), and also secure digital currencies, and it was from this background that I had come across the mathematics of small works networks, pareto distributions and so forth. It was a clear and present fear of mine at the time, that the ideals of the network would inevitably be corrupted over time by the mathematical consequences of preferential attachment. Call me a geek, but this maths seriously dampened my revolutionary vigour. What would be the point, I asked myself, of replacing this wonderful (if not corrupt), network of cafe meetings, with an equally corrupt network of influence online? The answer to this question is hugely important. It is the key to the vitality of our future organisations. My hope is to explore these issues on this web site – as there are several layers to a deep answer. However one thing stuck me at the time, and that was this. There is much more to the ethical, and effectiveness of the social editorial of an institution than the topology of the dialogue structure. Put more directly, protein networks, sexual relations in Sweden, café conversations between political servants and arts funders, and online discussion fora, may all share a scale free network topology, but we learn little from this with regard to the human consequences for the system as a whole. For one I’d prefer to be part of a network, in which I was free to express my opinion, without coercion, either by peer-pressure or force, based on the perceived quality of the person or idea I am supporting, than to live in a Mafia like society in which family ties, and preferential attachment is used to scale power. The topology of both systems may be equivalent, but not their quality, or the use to which the power of this network topology is used.