The working with wide abstractions, which is condemned above, to the entire neglect of the perceptive knowledge from which they are drawn, and which is therefore their permanent and natural controller, was at all times the principal source of the errors of dogmatic philosophy. A science constructed from the mere comparison of conceptions, that is, from general principles, could only be certain if all its principles were synthetical _a priori_, as is the case in mathematics: for only such admit of no exceptions. If, on the other hand, the principles have any empirical content, we must keep this constantly at hand, to control the general principles. For no truths which are in any way drawn from experience are ever unconditionally true. They have therefore only an approximately universal validity; for here there is no rule without an exception. If now I link these principles together by means of the intersection of their concept-spheres, one conception might very easily touch the other precisely where the exception lies. But if this happens even only once in the course of a long train of reasoning, the whole structure is loosed from its foundation and moves in the air. If, for example, I say, “The ruminants have no front incisors,” and apply this and what follows from it to the camel, it all becomes false, for it only holds good of horned ruminants. What Kant calls _das Vernünfteln_, mere abstract reasoning, and so often condemns, is just of this sort. For it consists simply in subsuming conceptions under conceptions, without reference to their origin, and without proof of the correctness and exclusiveness of such subsumption—a method whereby we can arrive by longer or shorter circuits at almost any result we choose to set before us as our goal. Hence this mere abstract reasoning differs only in degree from sophistication strictly so called. But sophistication is in the theoretical sphere exactly what chicanery is in the practical. Yet even Plato himself has very frequently permitted such mere abstract reasoning; and Proclus, as we have already mentioned, has, after the manner of all imitators, carried this fault of his model much further. Dionysius the Areopagite, “_De Divinis Nominibus_,” is also strongly affected with this. But even in the fragments of the Eleatic Melissus we already find distinct examples of such mere abstract reasoning (especially § 2-5 in Brandis’ _Comment. Eleat._) His procedure with the conceptions, which never touch the reality from which they have their content, but, moving in the atmosphere of abstract universality, pass away beyond it, resembles blows which never hit the mark. A good pattern of such mere abstract reasoning is the “_De Diis et Mundo_” of the philosopher Sallustius Büchelchen; especially chaps. 7, 12, and 17. But a perfect gem of philosophical mere abstract reasoning passing into decided sophistication is the following reasoning of the Platonist, Maximus of Tyre, which I shall quote, as it is short: