[Footnote 23: Aristot. De Interpr. p. 19, b. 19. [Greek: o(/tan de\ to\ e)/sti tri/ton proskatêgorê=tai, ê)/dê dichô=s le/gontai ai( a)ntithe/seis; le/gô de\ oi(=on _e)/sti di/kaios a)/nthrôpos_; to\ _e)/sti_ tri/ton phêmi\ sugkei=sthai o)/noma ê)\ r(ê=ma e)n tê=| katapha/sei. ô(/ste dia\ tou=to te/ttara e)/stai tau=ta, ô(=n ta\ me\n du/o pro\s tê\n kata/phasin kai\ a)po/phasin e(/xei kata\ to\ stoichou=n ô(s ai( sterê/seis, ta\ de\ du/o, ou)/. [le/gô de\ o(/ti to\ _e)/stin_ ê)\ tô=| dikai/ô| proskei/setai ê)\ tô=| ou) dikai/ô|], ô(/ste kai\ ê( a)po/phasis. te/ttara ou)=n e)/stai. noou=men de\ to\ lego/menon e)k tô=n u(pogegramme/nôn.] In this passage the words which I have enclosed between brackets are altered by Waitz: I shall state presently what I think of his alteration. Following upon these words there ought to be, and it seems from Ammonius (Schol. p. 121, a. 20) that there once was, a scheme or table arranging the four propositions in the order and disposition which we read in the Analytica Priora, I. xlvi. p. 51, b. 37, and which I shall here follow. But no such table now appears in our text; we have only an enumeration of the four propositions, in a different order, and then a reference to the Analytica.] First, let us assume _homo_ as subject. We have then (QUATERNION I.) (A) Est justus homo ... ... ... ... (B) Non est justus homo. (D) Non est non justus homo ... ... ... (C) Est non justus homo. Examining the relation borne by the last two among these four propositions (C and D), to the first two (A and B), the simple affirmative and negative, we see that B is the legitimate negative of A, and D that of C. We farther see that B is a consequence of C, and D a consequence of A, but not _vice versâ_: that is, if C is true, B must certainly be true; but we cannot infer, because B is true, that C must also be true: while, if A is true, D must also be true; but D may perhaps be true, though A be not true. In other words, the relation of D to A and of C to B, is the same as it would be if the privative term _injustus_ were substituted in place of _non justus_; _i.e._ if the proposition C (_Est injustus homo_) be true, the other proposition B (_Non est justus homo_) must certainly be true, but the inference will not hold conversely; while if the proposition A (_Est justus homo_) be true, it must also be true to say D (_Non est injustus homo_), but not _vice versâ_.[24]