Now in principle, I’m a fan of giving credit where credit is due. Citations not only assign appropriate credit, but they also allow the reader to trace back the antecedents and provide a roadmap for the intellectual history of the piece of work. The citation system also inspired the PageRank algorithm that underlies our always-on-hand Google search engine (http://infolab.stanford.edu/~backrub/google.html). At the heart of both are social systems for building a shared and constantly evolving edifice of knowledge in a distributed fashion. In addition to credit, they also lend credence and authority -- through an automated algorithm for PageRank, through the social-reputational mechanism for academia.
Yet, as I inserted citation after citation in-line to this monolith (it really took on this Sisyphean flavor), it felt….if not archaic, then at least dated. The number of citations drastically reduced the readability of the paper, making it hard to glean whatever original insight existed in the sea of parentheticals. It felt like, at times, I was citing what any reasonable person would call ‘common sense.’ Perhaps a hundred years ago, it wasn’t common sense, but as common sense enters the collective consciousness, when are we allowed to stop giving credit? Never?
There is this friction between the needs of the piece of work being created and at hand, and the debt we owe to the works that have come before and their authors. This is a friction that comes especially to the fore in the clash of traditional copyright and digitization (for those interested, Larry Lessig, in his excellent book Free Culture, offers a brilliant, lucid and forward-looking (if not exactly balanced) discussion of these issues). It is representative of an even larger war between our principled sense of fairness and the pragmatic reality of results, outcomes, phenomena, and emblematic of the breakdown in conversation between the “two cultures” (http://en.wikipedia.org/wiki/The_Two_Cultures).
What struck me about academic citations is that they seem to be structured around attribution, which offers value to a few individuals, rather than authority, which potentially offers value to many. Citations can lend authority in two ways – substantiation (here is where I got this information) and reputation (this paper is cited elsewhere by practitioners with relevant judgment). For people immersed in the relevant body of work, this authority is quite clear. It is, however, challenging for non-practitioners to read a paper and understand its relevance and authority (and for many online resources as well, which suffer from the added suspicion associated with being readily edited). This problem cannot be underestimated, as the dizzying array of information sources becomes ever more nauseating.
Can we make these authority mechanisms more explicit, especially the reputational mechanism? Is it possible to provide simple, cogent, trusted signals of relevance and authority, without resorting to elite curatorship? For a given piece of work, can I dynamically understand the relevance of what I am absorbing – before and during reading, without distracting from the process of reading? Can I access in a simple way the whole history of this piece of work, from whence it came and what later constructs used it as a building block? I focus on long-form text, because despite the advances and advantages of other forms of media, we still haven’t found a better way to manipulate complex concepts than semantic language. If we can do this for academic publications, perhaps we can take the system and generalize it to solve for at least one dimension of the sticky problem of ‘filter failure.’