[8] Analyt. Post. II. iii. p. 90, b. 13: ἱκανὴ δὲ πίστις καὶ ἐκ τῆς ἐπαγωγῆς· οὐδὲν γὰρ πώποτε ὁρισάμενοι ἔγνωμεν, οὔτε τῶν καθ’ αὑτὸ ὑπαρχόντων οὔτε τῶν συμβεβηκότων. ἔτι εἰ ὁ ὁρισμὸς οὐσίας τις γνωρισμός, τὰ γε τοιαῦτα φανερὸν ὅτι οὐκ οὐσίαι.
Again, let us ask, vice versâ, Can everything that is declared by Definition, or indeed anything that is declared by Definition, be known also by Demonstration? Neither is this possible. One and the same cognitum can be known only by one process of cognition. Definitions are the principia from which Demonstration departs; and we have already shown that in going back upon demonstrations, we must stop somewhere, and must recognize some principia undemonstrable.[9] The Definition can never be demonstrated, for it declares only the essence of the subject, and does not predicate anything concerning the subject; whereas Demonstration assumes the essence to be known, and deduces from such assumption an attribute distinct from the essence.[10]
[9] Ibid. b. 18-27.
[10] Ibid. b. 33, seq.: ἔτι πᾶσα ἀπόδειξις τὶ κατά τινος δείκνυσιν, οἷον ὅτι ἔστιν ἢ οὐκ ἔστιν· ἐν δὲ τῷ ὁρισμῷ οὐδὲν ἕτερον ἑτέρου κατηγορεῖται, οἷον οὔτε τὸ ζῷον κατὰ τοῦ δίποδος οὐδὲ τοῦτο κατὰ τοῦ ζῷου — ὁ μὲν οὖν ὁρισμὸς τί ἐστι δηλοῖ, ἡ δὲ ἀπόδειξις ὅτι ἢ ἔστι τόδε κατὰ τοῦδε ἢ οὐκ ἔστιν.
Themistius (p. 71, Speng.) distinguishes the ὁρισμός itself from ἡ πρότασις ἡ τὸν ὁρισμὸν κατηγορούμενον ἔχουσα.
Prosecuting still farther the dialectical and dubitative treatment,[11] Aristotle now proceeds to suggest, that the Essence (that is, the entire Essence or Quiddity), which is declared by Definition, can never be known by Demonstration. To suppose that it could be so known, would be inconsistent with the conditions of the syllogistic proof used in demonstrating. You prove by syllogism, through a middle term, some predicate or attribute; e.g. because A is predicable of all B, and B is predicable of all C, therefore A is predicable of all C. But you cannot prove, through the middle term B, that A is the essence or quiddity of C, unless by assuming in the premisses that B is the essence of C, and that A is the essence of B; accordingly, that the three propositions, AB, BC, AC, are all co-extensive and reciprocate with each other. Here, then, you have assumed as your premisses two essential propositions, AB, BC, in order to prove as an essential proposition the conclusion AC. But this is inadmissible; for your premisses require demonstration as much as your conclusion. You have committed a Petitio Principii;[12] you have assumed in your minor premiss the very point to be demonstrated.
[11] Analyt. Post. II. iv. p. 91, a. 12: ταῦτα μὲν οὖν μέχρι τούτου διηπορήσθω. One would think, by these words, that τὸ διαπορεῖν (or the dubitative treatment) finished here. But the fact is not so: that treatment is continued for four chapters more, to the commencement of ch. viii. p. 93.
[12] Analyt. Post. II. iv. p. 91, a. 12-32: ταῦτα δ’ ἀνάγκη ἀντιστρέφειν· εἰ γὰρ τὸ Α τοῦ Γ ἴδιον, δῆλον ὅτι καὶ τοῦ Β καὶ τοῦτο τοῦ Γ, ὥστε πάντα ἀλλήλων. — λαμβάνει οὖν ὃ δεῖ δεῖξαι· καὶ γὰρ τὸ Β ἔστι τί ἐστιν ἄνθρωπος. Themistius, pp. 72, 73: τὸν ἀποδεικνύντα τὸ τί ἦν εἶναι τοῦ ἀνθρώπου, ἄλλο τι δεῖ προλαβεῖν τοῦ αὐτοῦ τὸ τί ἦν εἶναι. — οὗ γὰρ βούλεται τὸν ὁρισμὸν ἀποδεῖξαι, τούτου προλαμβάνει τινὰ ὁρισμὸν εἶναι χωτὶς ἀποδείξεως.
M. Barthélemy St. Hilaire, notes, p. 205:— “Il faut donc, pour conclure par syllogisme que A est la définition essentielle de C, que A soit la définition essentielle de B, et que B soit lui-même la définition essentielle de C. Mais alors la définition de la chose sera dans le moyen terme lui-même, avant d’être dans la conclusion; en effet, la mineure: B est la définition essentielle de C, donne la définition essentielle de C, sans qu’il soit besoin d’aller jusqu’à la conclusion. Donc la démonstration de l’essence ainsi entendue est absurde.�
If you cannot obtain Definition as the conclusion of syllogistic Demonstration, still less can you obtain it through the method of generic and specific Division; which last method (as has been already shown in the Analytica Priora) is not equal even to the Syllogism in respect of usefulness and efficacy.[13] You cannot in this method distinguish between propositions both true and essential, and propositions true but not essential; you never obtain, by asking questions according to the method of generic subdivision, any premisses from which the conclusion follows by necessity. Yet this is what you ought to obtain for the purpose of Demonstration; for you are not allowed to enunciate the full actual conclusion among the premisses, and require assent to it. Division of a genus into its species will often give useful information, as Induction also will;[14] but neither the one nor the other will be equivalent to a demonstration. A definition obtained only from subdivisions of a genus, may always be challenged, like a syllogism without its middle term.