When this is transposed into the first figure the minor must be converted, and thus runs: “Some metals are alkalis.” It therefore merely asserts that some metals lie in the sphere “alkalis,” thus [Figure 1], while our actual knowledge is that all alkalis lie in the sphere “metals,” thus [Figure 2]: It follows that if the first figure is to be regarded as the only normal one, in order to think naturally we would have to think less than we know, and to think indefinitely while we know definitely. This assumption has too much against it. Thus in general it must be denied that when we draw inferences in the second and third figures we tacitly convert a proposition. On the contrary, the third, and also the second, figure exhibits just as rational a process of thought as the first. Let us now consider another example of the other class of the third figure, in which the separableness of two predicates is the result; on account of which one premiss must here be negative: No Buddhist believes in a God; Some Buddhists are rational: Therefore some rational beings do not believe in a God. As in the examples given above the _compatibility_ of two properties is the problem of reflection, now their _separableness_ is its problem, which here also must be decided by comparing them with _one_ subject and showing that _one_ of them is present in it without the _other_. Thus the end is directly attained, while by means of the first figure it could only be attained indirectly. For in order to reduce the syllogism to the first figure we must convert the minor, and therefore say: “Some rational beings are Buddhists,” which would be only a faulty expression of its meaning, which really is: “Some Buddhists are yet certainly rational.” As the guiding principle of this figure I therefore give: for the affirmative moods: _Ejusdem rei notœ, modo sit altera universalis, sibi invicem sunt notœ particulares_; and for the negative moods: _Nota rei competens, notœ eidem repugnanti, particulariter repugnat, modo sit altera universalis_. Translated: If two predicates are affirmed of one subject, and at least one of them universally, they are also affirmed of each other particularly; and, on the contrary, they are denied of each other particularly whenever one of them contradicts the subject of which the other is affirmed; provided always that either the contradiction or the affirmation be universal. In the _fourth figure_ the subject of the major has to be compared with the predicate of the minor; but in the conclusion they must both exchange their value and position, so that what was the subject of the major appears as the predicate of the conclusion, and what was the predicate of the minor appears as the subject of the conclusion. By this it becomes apparent that this figure is merely the _first_, wilfully turned upside down, and by no means the expression of a real process of thought natural to the reason.