Why do some narratives suddenly invade public attention, persist for weeks or years, and then disappear, while most never take off? This project will develop a quantitative theory of narrative dynamics inspired by epidemiology and statistical physics. Narratives will be treated as competing contagions spreading through a population with finite attention, so that the success of one narrative necessarily limits the space available to others. Large textual datasets will be used to reconstruct empirical narrative trajectories and calibrate a low-dimensional macroscopic model. A further theoretical objective will be to investigate whether this effective dynamics can be derived from a microscopic interacting-agent model, for instance a Random Field Ising Model (RFIM)-type description combining heterogeneous individual preferences and social imitation.