CogNarr Ecosystem: Preliminary Thoughts on a Story Graph Meaning Representation

Societies face a host of environmental, social, technical, economic, and other kinds of challenges and problems that must be successfully addressed. Our capacity to do so depends, in part, on the quality and functionality of the deliberation, strategizing, collaborative problem solving, and decision making processes that we engage in. There is need for new tools and approaches that facilitate these processes, especially in the large-group setting where face-to-face interactions are not possible. In that setting, how can each participant's potentially complex, nuanced, and dynamic perspective be heard and understood by the group as a whole? The nascent CogNarr (Cognitive Narrative) ecosystem, proposed in a companion paper, provides a potential answer. Its purpose is to facilitate group cognition, especially in the large-group setting. To accomplish this, the CogNarr computational system must be able to understand the meaning of text, chart, and other kinds of user input, and perform various analytic tasks based on that understanding. A first, key step in the computational pipeline is to convert user input into a meaning representation that is less ambiguous than natural language and that is readable and understandable by both humans and computers. The details of that meaning representation have not yet been specified. This concept paper discusses meaning representations commonly used in natural language inference, provides a desiderata of features for a custom CogNarr meaning representation, and offers some preliminary thoughts on the design of a CogNarr meaning representation and its use in computational pipelines.