In his Topics Aristotle sets up the theory of dialectic, which he seems not to use in his other works. This must be accounted for. Irwin assumes that there is a distinction in dialectic: pure and strong. But in my opinion Aristotle’s theory of dialectic has undergone development from an inchoate one into a revised one: the former is found in Topics while the latter in Rhetoric.
The theory of dialectic in Topics is not full-blown. To explain. Aristotle there asserts that there are four predicables: definition, property, genus, accident and that a set of the predicables, which are mutually exclusive, is exhaustive. However, per se accident does not fall under any of these. Traditionally per se accident has been interpreted to be a sort of property. As Burnes pointed out, however, per se accident does not have to fit the definition of a property. This implies that in Topics the four predicables are not well defined.
Alexander points out that there is an inseparable accident, which is an attribute that belongs to something but is not separate from it. It is noteworthy that the inseparable accident, fitting the first definition of accident given in Topics, does not fit its second definition given there. Aristotle holds that the first definition is equivalent to the second. If this is the case, it will follow that the concept of an inseparable accident contradicts his own idea. Many scholars do not take this fact into account. But, in my opinion, this fact shows that Aristotle has abandoned the theory of dialectic in Topics and introduces a pair of per se and accident in Posterior Analytics instead.
In addition, the dialectic Aristotle discusses in Rhetoric is a revised dialectic. For the explanation of topos has greatly changed. Aristotle distinguishes ‘common’ topos from ‘peculiar’ topos. According to him, the common topos applies for different objects, while the peculiar topos applies for a specific object. This peculiar topos is mentioned in Rhetoric I. 7. Peculiar topos in Rhetoric I. 7 is common to topos in Topics III. On the other hand, the common topos in Rhetoric II 27 is common to the ‘most opportune topos,’ which is referred to in Topics III and VII. The distinction of ‘common’ and ‘peculiar’ is also used in the theory of demonstration. We may say that the theory of demonstration has an influence on the revised dialectic.