On the limits of falsification: models, predictions and postdictions

In this talk, I will argue that by embracing the subjectivity inherent in mathematical modeling, one can make progress at answering substantial questions. I will first layout the philosophy, and then give examples from my prior work that has focused on using evidence accumulation models to understand orthographic processing in both sighted and braille reading. I will describe how to use drift- diffusion model in a principled, yet humble way, and how to connect diffusion model parameters to cognitive processes.