except according to opinion, but these may be predicated of others, nor can singulars be predicated of others, but others of them. It appears however that those which are intermediate, are capable in both ways (of demonstration), for they may be predicated of others, and others of them, and arguments and speculations are almost all conversant with these.
Still it is requisite to assume the propositions about each thing thus:—In the first place, the subject, (by hypothesis,) the definitions, and such peculiarities as exist of the thing; next, whatever things are consequent to the thing, and which the thing follows; lastly, such as cannot be in it; those however which it cannot be in are not to be assumed, because of the conversion of the negative. We must also distinguish in the consequents what things belong to "what a thing is," what are predicated as properties, and what as accidents; also of these, those which are (predicated) according to opinion, and those, according to truth; for the greater number any one has of these, the quicker will he light upon a conclusion, and the more true they are, the more will he demonstrate. We must too select not those which are consequent to a certain one, but those which follow the whole thing, e.g. not what follows a certain man, but what follows every man, for a syllogism consists of universal propositions. If therefore a proposition is indefinite, it is doubtful whether it is universal, but when it is definite, this is manifest. So also we must select those things the whole of which a thing follows, for the reason given above, but the whole consequent itself need not be assumed to follow; I say for instance, (it must not be assumed) that every "animal" is consequent to "man," or every science to music, but only that they are simply consequent, as we set forth, for the other is useless and impossible, as that "every man" is "every animal," or that "justice is every thing good." To whatever (subject) a consequent is attached, the sign "every" is added; when however the sub-