Mill's dissatisfaction with the claims of the Dictum de omni et nullo to represent the structure of the syllogistic process arose from his conviction that an adequate principle of syllogism must mention attributes: whether or not to syllogise is to infer, it will, on his view, be a perfectly trivial thought process if it is concerned only with objects and classes. In the previous chapter I have argued that Mill's objections to the Dictum are misconceived and his proposed alternative non-extensional principle of syllogism unsatisfactory. But now we must turn, as Mill did after completing his discussion of the Dictum, to the question of whether syllogism is really a form of inference at all.