The primary objective of this research is to test the hypothesis that information in a document, called jargon, can help to identify the background opinions or beliefs of the author. The first phase of the project tests whether an author's opinion can be identified for a text that explicitly addresses an issue about which the author has an opinion. The second attempts to identify the author's background beliefs for an article that deals with some topic in the general domain to which those beliefs are related. The final phase attempts to identify the author's opinion or background system of beliefs for articles that do not directly deal with an issue or topic in the general domain related to the opinion or system of beliefs.
The intellectual merit of the proposal lies in developing more rigorously the linguistic notion of jargon, broadening it to include not only technical terms in domains of activities, but also ideological terms in issue-oriented domains and identifying that ideological jargon in texts that are somewhat removed from the original issue-oriented or activity domain. Although there exist methods for categorizing texts based on information directly available from the text itself (e.g., is it a fact or an opinion, what is the topic), this research goes beyond this to ascertain whether there are properties of the author that are not immediately accessible from the text itself that indicate possible views and beliefs of the author. The immediate impact of the proposed research is the development of a set of tools for identifying jargons and attempting to determine possible attitudes of the authors of texts on the basis of these jargons. If the results are positive, the approach may enable sophisticated text retrieval or web search procedures that allow the search to be constrained by author attitude and belief.
