Aristotle passes, in the Fifth book of the Topica, to those debates in which the thesis set up declares the predicate as Proprium of the subject.
A Proprium may belong to its subject either per se and semper, or relatively to something else and occasionally or sometimes. It is a proprium per se of man to be an animal by nature tractable. It is a relative proprium of the soul in regard to the body, to exercise command; of the body in regard to the soul, to obey command. It is a proprium semper of a god, to be immortal; it is an occasional proprium (i.e., sometimes) of this or that man, to be walking in the market-place.[202] When the proprium is set out relatively to something else, the debate must involve two questions, and may involve four. Thus, if the thesis affirms that it is a proprium of man relatively to horse (discriminating man from horse) to be by nature two-footed, you may (as opponent) either deny that man is two-footed, or affirm that horse is two-footed; or you may go farther and affirm that man is by nature four-footed, or deny that horse is by nature four-footed. If you can succeed in showing any one of these four, you will have refuted the thesis.[203]
[202] Ibid. V. i. p. 128, b. 14-21. That which Aristotle calls Proprium per se is a proprium of the subject as much relative as what he calls specially the relative Proprium. The Proprium per se discriminates the subject from everything else; the relative Proprium discriminates it from some given correlate.
[203] Topic. V. i. p. 128, b. 22-33.
The Proprium per se discriminates its subject from everything else, and is universally true thereof; the relative Proprium discriminates its subject only from some other assignable subject. The relative Proprium may be either constant and universally true, or true with exceptions — true and applicable in the ordinary course of things: it may be tested through those Loci which have been enumerated as applicable to the Accident. The Proprium per se, and the constant Proprium, have certain Loci of their own, which we shall now indicate. These are the most logical (sensu Aristotelico) or suitable for Dialectic; furnishing the most ample matter for debates.[204]
[204] Ibid. b. 34-p. 129, a. 35. τῶν δ’ ἰδίων ἐστὶ λόγικὰ μάλιστα· &c. He explains presently what he means by λογικά â€” λογικὸν δὲ τοῦτ’ ἐστὶ πρόβλημα, πρὸς ὃ λόγοι γένοιτ’ ἂν καὶ συχνοὶ καὶ καλοί. The distinctions in this paragraph are not very sharply drawn.
Aristotle distinguishes (1) those cases in which the alleged proprium is a true proprium, but is incorrectly or informally set out in the thesis, from those (2) in which it is untruly predicated, or is no proprium at all.
To set out a proprium well, that which is predicated ought to be clearer and better known than the subject of which it is predicated, since the purpose of predicating the proprium is to communicate knowledge.[205] If it be more obscure or less known, you may impugn the thesis as bad in form, or badly set out. Thus, if the thesis declare, as a proprium of fire, that fire is of all things the most like to the soul, this is not well set out, because the essence of the soul is not so well known as the essence of fire. Moreover, the fact that the predicate belongs to the subject, ought to be better known even than the subject itself; for whoever is ignorant that A belongs to B at all, cannot possibly know that A is the proprium of B.[206] Thus, if the thesis declare, as proprium of fire, that it is the first or most universal subject in which it is the nature of soul to be found, the predicate is here doubly unknowable: first, the hearer does not know that the soul is found in fire at all; next, he does not know that fire is the first subject in which soul is found. On the other hand, the respondent will repel your attack if he can show that his proprium is more knowable in both the two above-mentioned ways. If, for example, he declares as thesis, To have sensible perception is the proprium of an animal, here the proprium is both well known in itself, and well known as belonging to the given subject. Accordingly, it is well set out, as far as this condition is concerned.[207]
[205] Ibid. p. 129, b. 7: γνώσεως γὰρ ἕνεκα τὸ ἴδιον ποιούμεθα· διὰ γνωριμωτέρον οὖν ἀποδοτέον· οὕτω γὰρ ἔσται κατανοεῖν ἱκανῶς μᾶλλον.
He repeats the same dictum, substantially, in the next page, p. 130, a. 4: τὸ γὰρ ἴδιον τοῦ μαθεῖν χάριν ἀποδίδοται; and, again, p. 131, a. 1.