Long Tails All the Way Down: Rethinking A-Lists
In his recent post on mainstream newspapers’ gradual warming to Google as a source of traffic to their Websites, Jason points to Matthew Hindman’s claims that descriptions of the Net as ‘democratising’ media participation may be overstated (in Hindman’s book manuscript Voice, Equality, and the Internet), and notes my skeptical stance towards Hindman’s conclusions. Heh. Jason, are you baiting me? 
Seriously: yes, I do have very strong reservations about Hindman’s work. While he’s got access to some useful statistics about the distribution of attention in the blogosphere, and is perhaps worth reading for that reason, I’m very doubtful about what he does with those stats, and why he does it. To some extent, Hindman’s work seems to me to be cut from the same cloth as that of Andrew Keen (though operating at a more apparently scientific level) - which is to say that in the first place, he seems to grab many of the most banal and over-exuberant claims about the blogosphere and unsurprisingly has no trouble debunking them.
In reading his book manuscript, I noted down ’straw man argument’ a good half dozen times - he claims that his work helps to debunk views that the Internet in general and blogs in particular democratise media participation, but pays little attention to exactly what other authors mean when they express such views, and instead substitutes a caricature of their views instead.
To take one example: on page 71, Hindman writes:
In large part, this book is about that Field of Dreams assumption. Over the past decade and a half, the Web has been built. Billions upon billions of documents are now online, and this vastness of content has been used to support claims that Internet audiences must be more widely dispersed than audiences for broadcasting or print. The notion that the Internet is part of a continuing shift from broadcasting to narrowcasting, and that the Web will empower new small-scale producers of content, is a central part of the Internet’s identity in the public mind. From law to public policy, democratic theory to party politics, interest in the Internet has begun from the belief that the Web is “democratizing” the flow of information.
So far, so good - I think many people working in the field would agree with this description; there are many of us interested in the potential for more (and more diverse) groups of users to contribute their own views and ideas to public discussion. That said, the Field of Dreams allusion (which translates to the famous line “if you build it, they will come”, of course) already paints this in a technologically determinist light that ultimately misrepresents what most researchers into user-led content creation actually claim: for most of us, I would suggest, the rise of Web 2.0 represents the social appropriation of Web technologies rather than a technology-driven phenomenon.
Anyway: the way I read it, Hindman goes on from here to translate “democratisation” incorrectly as “equalisation” or “levelling” - and much of what he has to say throughout his manuscript is tainted by this misunderstanding. He notes that his “study finds powerful hierarchies shaping a medium that continues to be celebrated for its ‘openness’” (p. 15), and claims that this, somehow, undermines the potential for the Net to be a democratising medium. By extension, this seems to claim that democracy is incompatible with any sense of hierarchy - but I would argue that this fundamentally misunderstands the way that democracy works.
Democracy, after all, is (ideally) all about identifying the best solutions to current problems, and placing the best and brightest in positions of leadership - so it’s very much about hierarchies, but about hierarchical structures which are determined by public deliberation and popular election rather than by hereditary lineage, military force, or economic power. The distinction between democracies and other forms of social organisation has nothing to do with whether hierarchies are or aren’t present - it has everything to do with how these hierarchies are determined, and how changeable and responsive they are to the will of the people.
If, as Hindman found, there are clear hierarchies of influence and attention in the blogosphere, then, this does not inherently undermine the democratising potential of the Internet, as many scholars describe it. He points to what he calls “‘Googlearchy’: the rule of the most heavily linked” (p. 44), and somehow paints this as anti-democratic - but of course, what determines this hierarchy of visibility in the blogosphere are in fact the linking and browsing patterns of all of us which are tracked by Google’s algorithms: rather than obscure editorial decisions behind closed doors at Google headquarters, it is the Internet population at large which determines what sites are and are not part of the ‘A-list’ of bloggers. Except for the obvious limitation that Internet access itself is unevenly distributed, I can’t think of many more democratic processes.
Ultimately, I think what we need to consider here is the difference between equality and equipotentiality. Democracy doesn’t mean that everybody is equal: yes, in democracy everybody has one vote to cast in elections, but that doesn’t mean that each of us has the same chance to be elected to public office. Instead, we are merely equipotential (as Michel Bauwens has put it, in a slightly different context): we each have the same potential to run for office, but whether we are elected depends in the end on how many other citizens we are able to convince of the quality of our ideas and the sincerity of our convictions. (Other factors, such as our media performance and the amount of cash we are able to spend on campaigning, also interfere here, of course, which is a significant problem for mass-mediated democracies.)
So, in his study of blog ‘A-lists’, I think Hindman is barking up the wrong tree: the mere fact that some blogs are (exponentially) more visible and influential than others is not a sign of undemocratic tendencies, but instead represents a very democratic process of users ‘voting with their browsers’ - nobody, neither Google nor anyone else, forces them to read certain blogs and not others. What’s more important to study, I think, is the extent to which such ‘A-lists’ ossify - that is, how fixed or fluid they are.
If there were some mechanism that prevents intelligent new voices from getting heard in the blogosphere, and from rising towards ‘A-list’ status themselves, then that would be a problem - but I’m not convinced that this is the case. For one, a preliminary study that Lars Kirchhoff, Thomas Nicolai, and I published last year showed that between 2005 and 2006 there was some significant fluctuation in the top blog population (Google PageRank 7 and above) for the 8.8 million blogs hosted at Blogger.com that our study examined; we need to do more number-crunching to look at the details, but what we found seemed to indicate some substantial fluctuation in the membership of Blogger’s ‘A-list’.
Second, many ‘A-list’ bloggers themselves are important distributors of attention - many of them regularly point their readers to blog posts and other material on other sites that they’ve found during their Web travels, and thereby help these other sites gain visibility (for better or for worse, we experienced this ourselves when Tim Blair linked to us a little while ago, for example). In the process, these existing ‘A-listers’ boost the potential for newcomers to rise through the ranks.
Finally - and this is another major problem I have with Hindman’s work - it is inherently misleading to speak of “the ‘A-list’” as if there’s only one. Yes, it is possible to use Google, Technorati, Hitwise, or various other tools to identify the world’s ten most-read blogs, but in itself that’s an excercise of very limited use. In addition to the ‘A-list’, there are also ‘A-lists’ of news bloggers, political bloggers, Australian political bloggers, left- and right-of-centre Australian political bloggers, and so on. The Australian blogosphere may be too small to make finer distinctions meaningful, but in larger countries such as the U.S. we may even find ‘A-lists’ in political blogging for specific geographic areas or fields of interest (libertarian bloggers from Southern California, political philosophy bloggers from New England, etc.), and so on.
Hindman does note that the long tail “power law structure persists even if … sites are broken down into sub-communities”, and concludes that “the structure of political groups on the Web thus may loosely be termed fractal in nature - portions of the community mirror the winners-take-all structure of the whole” (p. 44). In other words, there are multiple more specific, more narrow ‘A-lists’ nested within more generic ‘A-lists’, but Hindman fails to incorporate this observation into his analysis in any meaningful way. To quote him again:
The Internet does provide any citizen a potential audience of billions, in the same way that potentially pigs can fly. In their enthusiasm, many have forgotten to do the math, and that math shows that the odds of hitting it big are vanishingly small. Individually, each of the myriad sources which make up the long tail are insignificant; even together, they remain only a fraction of the content citizens actually see. (p. 86)
Frankly, I think this is wilfully misrepresenting what people are actually saying, and contradicts Hindman’s own observation of the fractal nature of power law structures. It can be proven that individuals can ‘hit it big’ and become influential (take Daily Kos, Russ Kick, Glenn Reynolds, or even Possums Pollytics…) - that’s the core argument in favour of equipotentiality (as opposed to equality, which is what Hindman appears to be talking about). Additionally, of course, ‘hitting it big’ in the present context may mean anything from rising to the ‘A-list’ for your specific field of expertise (say, Australian economic policy) to becoming a globally recognised ‘A-list’ blogger. Indeed, the former may be more meaningful, as it would tend to indicate influence amongst one’s peers in a specialised field of enquiry rather than mere popularity at a general level.
The potential for this rise to influence to happen is significantly greater than that for pigs to fly (and more importantly, it is also greater than the potential for individuals to ‘hit it big’ in more conventional media), and the very mechanisms which promote bloggers into the ‘A-list’ are responsible for boosting this potential. The online system of attention is better geared towards such probabilistic processes than are traditional media systems: its quasi-fractal nature provides the mechanisms for a gradual cascading upwards through the nested long tail systems which even Hindman has identified.
Whether by design or by accident, Hindman’s dogged focus on the global ‘A-list’ - on the one ‘long tail’ graph encompassing all of the blogosphere - obscures this fact, and this means that his description of hierarchies of attention and influence in the blogosphere remains ultimately no more than a first approximation of reality. Instead, I would suggest that it’s long tails all the way down: in essence, a nested, repetitive, fractal structure that repeats the same pattern of a handful of ‘A-listers’ and a long tail of also-rans all the way from the global to the hyperlocal level. My hypothesis - which I hope our blog mapping work will prove - is that there’s far more to hierarchical patterns on the Web than Hindman is prepared to admit: that this hierarchy of influence is multi-layered, that ultimately the blogosphere resembles a (fluid and changeable) heterarchy rather than a simple hierarchy.
I was baiting you Axel, but in the best possible way!