Social media provide unique spaces for collaborative discussions, but we all have experienced at some point how hard it can be getting the relevant content from these long threats and sub-threats of comments. Mark Klein, a computer scientist at the MIT Center for Collective Intelligence, addressed this issue in his recent talk at the IIIA-CSIC Seminar. Mark’s strategy to fully capture the good stuff among the noise has crystallized in the MIT Deliberatorium, a technology aiming at organizing complex collaborative discussions, reducing redundancy, and providing more clarity to their participants. More details on how it works can be found here.

The idea is to facilitate a crowdsourced deliberation map where contributors (authors) post an issue, idea, or argument, moderators ensure the deliberation map rules (i.e. avoid undermining editing, refine proposed ideas, etc.) and readers are able to rate and highlight worthy posts. While argument trees have been used for decades in logics and artificial intelligence, the MIT Deliberatorium intentionally keeps a low level of formalization to make it easier for authors to unbundle their contribution into its constituent issues, ideas, and arguments.

Another project on debate mapping and visualization is Debategraph, co-founded by Peter Baldwin and David Price. In Debategraph, contributions need to be broken down into a core set of issues (or questions raised), positions (or answers to the issues), and supportive or opposing arguments for and against these positions. The resulting structure can be visualized as a bubble map (much like a molecular structure) or a tree, which may contain cross-links to connect elements in different maps.

Argument mapping may certainly increase our capacity to identify the main issues at stake, find the best ideas, assess their pros & cons, and engage in creative thinking. In addition, the combination of visualization tools and new deliberation metrics should increase the potential of deliberation maps to structure collective knowledge as it emerges from communities.

But how large can large-scale argumentation be? The number of active users having contributed to deliberative maps in the two projects above remains quite limited so far. Most likely, communities of active domain experts discussing on complex issues won’t typically be very large, but what if thousands of engaged discussants contribute with tens of thousands of posts? How to curb down the costs of participation and the burden on moderators? Ontologies, already used in Debategraph, can be leveraged to decrease the costs of participation for authors, moderators, and readers. However, they might be hard to build and maintain. Which solutions could work best in this context? I’m probably pointing to another complex large-scale deliberation…

Marta Poblet