I investigate a form of argument which refers to Leibniz's law as its inference ticket (where ‘Leibniz's law’ is understood as the thesis that if x = y, then all properties of x are properties of y, and vice versa). Arguments of this form are often used to establish certain categorial distinctions, e.g., a distinction between kinds and properties, or a distinction between processes and events. I show that there can be deficient arguments of this form, and why. I then argue that the interesting philosophical cases of this argument form are unconvincing, since they cannot be seen as clear cases of its unproblematic variety.