Complex Adaptive Reasoning: Knowledge Emergence in the Revelator Game

We introduce the game of complex adaptive reasoning (CAR), as a self-correcting methodology for knowledge emergence in dialogue and narrative contexts of collaboration. Inspired by Holland's complex adaptive systems research (CAS), Revelator's CAR methodology is designed to model the complex logical relations among conjectures (represented in "If ... then" rule form) that players articulate in game plays. The game's agent-based representation of these rule-form plays gives them adaptive capability, making it possible for "argument colonies" to emerge as possible knowledge to be tested. We are experimenting with natural language processing (NLP) query support for controlled natural language (CNL) formulation of game plays, to reveal knowledge formation in the conversational contexts of news blogs and research wikis. In Revelator's framework, knowledge is represented as continually emerging, not simply to be captured but to advance in a recursive process that improves human reasoning skills with appropriate support from technology.