As in the exposition of syllogistic reasoning by means of
_concept-spheres_ these are presented to the mind under the form of
circles, so in the exposition by means of entire judgments we have to
think these under the form of rods, which, for the purpose of comparison,
are held together now by one end, now by the other. The different ways in
which this can take place give the three figures. Since now every premiss
contains its subject and its predicate, these two conceptions are to be
imagined as situated at the two ends of each rod. The two judgments are
now compared with reference to the two _different_ conceptions in them;
for, as has already been said, the third conception must be the same in
both, and is therefore subject to no comparison, but is that _with which_,
that is, in reference to which, the other two are compared; it is the
_middle_. The latter is accordingly always only the means and not the
chief concern. The two different conceptions, on the other hand, are the
subject of reflection, and to find out their relation to each other by
means of the judgments in which they are contained is the aim of the
syllogism. Therefore the conclusion speaks only of them, not of the
middle, which was only a means, a measuring rod, which we let fall as soon
as it has served its end. Now if this conception which is _identical_ in
both propositions, thus the middle, is the subject of _one_ premiss, the
conception to be compared with it must be the predicate, and conversely.
Here at once is established _a priori_ the possibility of three cases;
either the subject of one premiss is compared with the predicate of the
other, or the subject of the one with the subject of the other, or,
finally, the predicate of the one with the predicate of the other. Hence
arise the three syllogistic figures of Aristotle; the fourth, which was
added somewhat impertinently, is ungenuine and a spurious form. It is
attributed to Galenus, but this rests only on Arabian authority. Each of
the three figures exhibits a perfectly different, correct, and natural
thought-process of the reason in inference.