Crowdsourced knowledge: Peril and promise for conceptual structures research

Abstract

Recent efforts to create natural-language, question-answering systems for the World Wide Web exploit the vast availability of information in Web resources as a knowledge base. Researchers who develop these systems explicitly assume that what is most often repeated in that knowledge base is the truth. Their assumption implicitly relies on a fallacy in classical logic: "proof by assertion" (or proof by repeated assertion). This paper considers the implications-both hazardous and hopeful-of the Most Often Repeated (MOR) assumption, and suggests Peirce's "economy of research" (EOR), in his evolutionary view of logic, as a promising alternative to MOR for truth-finding in the increasing complexity of crowdsourced knowledge. © 2011 Springer-Verlag.