to show that it is inherent, but also it must be shown that it is inherent, as genus; but to the subverter it is enough to show it non-inherent, either in a certain or in every individual: still it seems, as in other things, to destroy, is easier than to produce, so in these, subversion, is easier than confirmation.
In the case of accident, we can more easily subvert, than construct the universal, for the confirmer must show that it is present with every, but the subverter need only show it non-inherent in one. On the contrary, it is easier to confirm, than to subvert the particular, as it suffices for the confirmer to show it present with a certain one, but the subverter must show that it is present with none.
It appears also clear why it is the easiest thing of all, to subvert definition, for many things being asserted in it, very many are given; but from the greater number, a syllogism is more quickly made, since it is likely that error should arise in many, more than in few, things. Moreover, it is possible to argue against definition through other things also, since whether the sentence be not appropriate, or whether what is assigned be not genus, or something of those in the definition be non-inherent, the definition will be subverted; but against other things, neither can we assume those arguments which are derived from definitions, nor all others, since those only which belong to accident, are common to all the particulars mentioned. For it is necessary that each of the things stated should be inherent, if however genus is not inherent as property, the genus will not yet be subverted; likewise, also property need not be inherent as genus, nor accident as genus or property, but merely inherent. Wherefore it is impossible to argue from some things to others, except in definition; hence, it is evident that to subvert definition is the easiest thing of all, but to confirm it the hardest, since we must syllogistically infer all those particulars, (viz. that all the
