The philosophical world is indebted to Alvin Goldman for a number of reasons, and among them, his defense of the relevance of cognitive science for philosophy of mind. In "Simulating minds", Goldman discusses with great care and subtlety a wide variety of experimental results related to mindreading from cognitive neuroscience, cognitive psychology, social psychology and developmental psychology. No philosopher has done more to display the resourcefulness of mental simulation. I am sympathetic with much of the general direction of Goldman's theory. (...) I agree with him that mindreading is not a single system based on a single mechanism. And I admire his attempt to being together the cognitive neuroscientific discovery of mirror system phenomena and the philosophical account of pretense within a unique theoretical framework of mental simulation. To do so, Goldman distinguishes two types of mindreading, respectively, based on low-level and high-level simulation. Yet, I wonder in what sense they are really two distinct processes. Here, I will confine myself largely to spelling out a series of points that take issue with the distinction between low-level and high-level mindreading. (shrink)