The Importance of the "emic" Perspective in Information Fusion

In order to conduct "full-spectrum cross-domain operations in volatile, uncertain, complex, and ambiguous environments around the globe"(1), it is important to provide decision makers with the information they need, from traditional and non-traditional sources (open source, social media). The information needs to be fused and presented in a way to maximally support meaning making about not only "what" is happening but also "why". To develop a more complete understanding of a current or future situation or event requires the ability to go beyond the fusion of "data", or even information, to the fusion of perspectives - the "etic", or third person, and "emic", or first person, perspectives. This is something that storytellers do naturally; for example, storytellers will often use first person when there is a strong protagonist or main character that they want the audience to focus on(2). When a reader gets "inside a character's head", they can gain an understanding of their motivations and worldview. Integration or fusion of "emic" information provides important clues/insights critically needed by analysts and decision makers for forecasting behavior and a more nuanced understanding of the situation and threat. AFRL has been engaged in research aimed at enabling meaning making based on the "emic" perspective from discourse (text from a variety of open sources, including social media) for several years. Early research focused on the development of multi-lingual methodologies (Arabic and Pashto), documented in primers transitioned to operational customers, including the National Air and Space Intelligence Center. The methodologies enable the detection and interpretation of the discourse patterns related to social identity (in-group/out-group)(3). Identification of these patterns enables forecasting of events (e.g., violence). Subsequent research developed methodologies to identify and interpret characteristic patterns or themes used to express or detect trust, trustworthiness in Farsi discourse and explored the link/influence between affect expressed in discourse and behaviors(4). Recent research has focused on the development of semiautomatic methods to assess intent based on discourse analysis, resulting in text analytic and forecasting algorithms based on discourse markers related to social identity and integrative complexity(5). "Etic" and "emic" are essentially different ways to view the same thing(6), a "stereoscopic window on the world."(7) In initial "fusion" experiments combining "etic" (events analysis) and "emic" information from discourse and sentiment analysis, the discourse markers were twice as powerful/accurate for forecasting violence as the previous forecasting "gold standard", event analysis. This provides confirmation of the value of integrating/fusing "emic" information and perspectives and provides the motivation for further research.