not specifically made. It is asserted, however, that this division is here of no significance. With this last statement I cannot agree. The distinction between impressions and ideas of sensation and those of reflection, is manifestly of some consequence in the discussion which deals with the idea of necessary connection.
The author asserts that the doctrine of belief plays a subordinate part in the Inquiry, as compared with the Treatise, and that the concept of belief is positively changed. We may admit the first statement, and we may also admit that the mode of treatment of the doctrine of belief is changed. But that the doctrine, or the concept, of belief is changed in the Inquiry, is by no means evident. Hume's argument on this subject in the Appendix is in support of his former view, and so also is his summing up of the discussion in the Inquiry. According to the author, the treatment of philosophical relations is omitted in the later work because: (1) in the Treatise Hume had experienced the difficulty of keeping them distinct from the natural relations; and (2) such treatment is now unnecessary, since the philosophical relations served as the basis for the development of the doctrine of the different degrees of certainty; and this doctrine is omitted in the Inquiry because Hume has now a different end in view. Moreover, Hume probably does not now regard his former twofold division of philosophical relations as tenable; for his higher estimation of the fundamental principle of induction would not allow him to say that a relation like causality could be changed without the perceptions, or things, on which the relation is based being changed also. I do not think there is so much difference between the position of the Inquiry and that of the Treatise on this subject as Dr. Brede represents. Although Hume does not give a classification, or distinct treatment, of the philosophical relations in the Inquiry, the existence of these relations is implied in his division of knowledge into relations of ideas and matters of fact. The distinction between natural relations and philosophical relations is also expressly indicated by the twofold definition that is given of cause. Whether relations can change while the perceptions on which they are based remain the same, depends entirely on our definition of the terms. And Hume is undoubtedly right in his contention on this question—from his own point of view—however much he may be logically inconsistent with some of his fundamental principles.
Dr. Brede's treatment of knowledge, or the different degrees of certainty, as presented in the Inquiry, is not quite satisfactory. He asserts that, according to Hume, geometry and arithmetic are equally certain, and geometry is an inductive science. Both statements are open to question. It is true that in section iv Hume speaks of geometry as being an exact science like arithmetic; but in section xii he does not hold this view. And it is also true that, in general, in section xii he regards geometry as an inductive science; but in section iv he expressly states that it is a demonstrative science. It seems impossible to explain or harmonize these contradictory statements. The author also asserts that, in the Inquiry, knowledge of matters of fact—
