History of Western Philosophy - Bertrand Russell [402]
At a time when he was more troubled by the arguments of sceptics than he was earlier or later, he wrote a curious work called Dreams of a Ghost-seer, Illustrated by the Dreams of Metaphysics (1766). The 'ghost-seer' is Swedenborg, whose mystical system had been presented to the world in an enormous work of which four copies were sold, three to unknown purchasers and one to Kant. Kant, half serious and half in jest, suggests that Swedenborg's system, which he calls 'fantastic', is perhaps no more so than orthodox metaphysics. He is not, however, wholly contemptuous of Swedenborg. His mystical side, which existed though it did not much appear in his writings, admired Swedenborg, whom he calls 'very sublime'.
Like everybody else at that time, he wrote a treatise on the sublime and the beautiful. Night is sublime, day is beautiful; the sea is sublime, the land is beautiful; man is sublime, woman is beautiful; and so on.
The Encyclopaedia Britannica remarks that 'as he never married, he kept the habits of his studious youth to old age'. I wonder whether the author of this article was a bachelor or a married man.
Kant's most important book is The Critique of Pure Reason (first edition, 1781; second edition, 1787). The purpose of this work is to prove that, although none of our knowledge can transcend experience, it is, nevertheless, in part a priori and not inferred inductively from experience. The part of our knowledge which is a priori embraces, according to him, not only logic, but much that cannot be included in logic or deduced from it. He separates two distinctions which, in Leibniz, are confounded. On the one hand there is the distinction between 'analytic' and 'synthetic' propositions; on the other hand, the distinction between 'a priori' and 'empirical' propositions. Something must be said about each of these distinctions.
An 'analytic' proposition is one in which the predicate is part of the subject; for instance, 'a tall man is a man', or 'an equilateral triangle is a triangle'. Such propositions follow from the law of contradiction; to maintain that a tall man is not a man would be self-contradictory. A 'synthetic' proposition is one that is not analytic. All the propositions that we know only through experience are synthetic. We cannot, by a mere analysis of concepts, discover such truths as 'Tuesday was a wet day' or 'Napoleon was a great general'. But Kant, unlike Leibniz and all other previous philosophers, will not admit the converse, that all synthetic propositions are only known through experience. This brings us to the second of the above distinctions.
An 'empirical' proposition is one which we cannot know except by the help of sense-perception, either our own or that of someone else whose testimony we accept. The facts of history and geography are of this sort; so are the laws of science, whenever our knowledge of their truth depends on observational data. An 'a priori' proposition, on the other hand, is one which, though it may be elicited by experience, is seen, when known, to have a basis other than experience. A child learning arithmetic may be helped by experiencing two marbles and two other marbles, and observing that altogether he is experiencing four marbles. But when he has grasped the general proposition 'two and two are four' he no longer requires confirmation by instances; the proposition has a certainty which induction can never give to a general law. All the propositions of pure mathematics are in this sense a priori.
Hume had proved that the law of causality is not analytic, and had inferred that we could not be certain of its truth. Kant accepted the view that it is synthetic, but nevertheless maintained that it is known a priori. He maintained that arithmetic and geometry are synthetic, but are likewise a priori. He was thus led to formulate his problem in these terms:
How are synthetic judgments a priori possible?
The answer to this question, with its consequences, constitutes the main