Warren Sack: Picturing the Public

Algorithms and Interfaces for the Presentation of Public Opinion

Public opinion is something that governments and businesses claim to know and to respond to, but what exactly is ›public opinion‹? Social science research has produced a large variety of methods and technologies for measuring public opinion based on sampling: surveys, interviews, ethnographies, etc. Public discussions on the Internet provide a new opportunity to rethink the methods, techniques and politics of public opinion. Online, there are daily many, huge, email and newsgroup discussions involving thousands of people. These discussions, what we call very large-scale conversations [VLSCs] cover almost any topic. In these many-to-many VLSCs, the so-called ›public‹ is representing itself to itself and itself to others. The availability of these text archives in electronic form, along with recent text parsing and computational analysis tools allows a powerful, rigorous study of this form of communication for the first time. We have been developing new computational technologies that combine insights from conceptual art, linguistic and cultural anthropology, sociology of science and language, and computational linguistics to automatically summarize VLSCs. We hope that this technology, the »Conversation Map,« is useful for political experts, but also for citizen/participants in these new public spaces who want a picture of them ‹selves› that might be useful for community reflection and democratic self-governance. Culturally, we are interested in the new sorts of political strategies and tactics that these VLSCs allow and engender. A demonstration of the tools can currently be found online: