Workshop Proceedings of the 16th International AAAI Conference on Web and Social Media
I argue that automatic detection and understanding of narratives is critical to advancing our fundamental theories of, and ability to model, social media. Understanding the narratives in play has always been a key component to understanding human communication, but this has traditionally been a manual task. Now the huge and growing volume, velocity, and variety of social media content means that we must automate narrative understanding as much as possible if we are to have any hope of effectively tracking and modeling social media dynamics. Importantly, automated narrative understanding is useful not only for understanding the content of communications, but also for assessing the psychological impact of those communications, providing insight into the nature of human groups, and predicting future beliefs and behaviors. As a basis for next steps, I identify a critical common mistake that computationalists often make when approaching narrative, and point out two foundational fields where researchers should look for guidance.