8 Metaphysics, Old and New: Immanuel Kant, Prolegomena*
Hume’s challenge to speculative metaphysics exerted a strong influence on subsequent philosophy, and was taken up in a systematic way by the famous German philosopher Immanuel Kant, first in his mammoth and highly complex Critique of Pure Reason (1781), and then in the Prolegomena (1783), which was intended as a more popular abstract of that work. The full title of the ‘Prolegomena’ (‘Preamble’ or ‘Preliminary Remarks’) is Prolegomena to any future metaphysics which will be able to present itself as a science (Prolegomena zu einer jeden künftigen Metaphysik die als Wissenschaft wird auftreten können); the question Kant addresses is ‘whether such a thing as metaphysics is even possible at all’.
Hume had convinced Kant that previous attempts to ascend beyond the empirical world and describe the supposed ultimate nature of reality (e.g. the theory of substance proposed by Leibniz (see above, extract 5)) were doomed to failure. Kant condemned such vain aspirations in a famous metaphor: ‘the light dove, cleaving the air in her free flight, might imagine [absurdly] that flight would be easier still in empty space’.1 There is, for Kant, no possible description of the world which can free itself from some reference to experience. But Kant argues that in experiencing the world, the mind is already armed with certain fundamental categories of understanding; these are a priori, but not in the traditional sense of being wholly abstract and independent of experience; rather they constitute the preconditions for all possible experience (see above, Part I, extract 8).
In the following set of extracts Kant argues that we do already possess knowledge that is both a priori and also genuinely informative or synthetic.2 To begin with, he argues that a priori mathematical judgements (e.g. that 7 + 5 = 12) are synthetic, since we cannot arrive at the concept of twelve merely by reflection on the notions of seven and five.3 Second, and more important, he takes up Hume’s challenge with respect to scientific knowledge (Hume had argued, extract 7 above, that we can never know any causal connections a priori). For Kant, the principle that every event has a cause is synthetic; yet it is a priori in the sense of being presupposed by experience (Kant argues elsewhere that we could not even begin to classify sets of perceptions as constituting genuine events unless the mind had the power to interpret the world in terms of causal frameworks).4
So there is, after all, room for a genuine metaphysics – not one which fruitlessly attempts to speculate about what lies beyond experience, but one which instead analyses and systematically lays out all the a priori concepts of the understanding which the mind necessarily employs in processing and interpreting the data of experience (see penultimate paragraph of extracts below). Though Kant does from time to time talk of a hidden world beyond experience, a world of what he calls noumena or ‘Things in Themselves’, he makes it clear that there can be no valid philosophical speculation about what such an independent reality might be like; human reason, when it operates properly, is necessarily confined to phenomena – to the world as experienced. The confusions and pretensions of traditional metaphysics arose from its attempt to describe the supposed nature of reality ‘in itself’; the new ‘critical’ metaphysics, mapping out the necessary preconditions for human experience, can hope to provide instead ‘definite and perfect knowledge’. It has, Kant concludes, the kind of validity and authority which modern astronomy has when compared to the pretensions of fortune-telling or astrology.
The distinction between analytic and synthetic judgements.
Metaphysical knowledge must contain simply judgements a priori; so much is demanded by the speciality of its sources. But judgements, let them have what origin they may, or let them even as regards logical form be constituted as they may, possess a distinction according to their content, by virtue of which they are either simply explanatory and contribute nothing to the content of a cognition, or they are extensive, and enlarge the given cognition; the first may be termed analytic, and the second synthetic judgements.
Analytic judgements say nothing in the predicate but what was already thought in the conception of the subject, though perhaps not so clearly, or with the same degree of consciousness. When I say, all bodies are extended, I do not thereby enlarge my conception of a body in the least, but simply analyse it, inasmuch as extension, although not expressly stated, was already thought in that conception; the judgement is, in other words, analytic. On the other hand, the proposition ‘some bodies are heavy’ contains something in the predicate which was not already thought in the general conception of a body; it enlarges, that is to say, my knowledge, in so far as it adds something to my conception; and must therefore be termed a synthetic judgement.
The common principle of all analytic judgements is the principle of contradiction.
All analytic judgements are based entirely on the principle of contradiction, and are by their nature cognitions a priori, whether the conceptions serving as their matter be empirical or not. For inasmuch as the predicate of an affirmative analytic judgement is previously thought in the conception of the subject, it cannot without contradiction be denied of it; in the same way, its contrary, in a negative analytic judgement, must necessarily be denied of the subject, likewise in accordance with the principle of contradiction. It is thus with the propositions ‘every body is extended’; ‘no body is unextended’. For this reason all analytic propositions are judgements a priori, although their conceptions may be empirical. Let us take as an instance the proposition ‘gold is a yellow metal’. Now, to know this, I require no further experience beyond my conception of gold, which contains the propositions that this body is yellow and a metal, for this constitutes precisely my conception, and therefore I have only to dissect it, without needing to look around for anything elsewhere.
Synthetic judgements demand a principle other than that of contradiction.
There are synthetic judgements a posteriori whose origin is empirical; but there are also others of an a priori certainty that spring from the Understanding and the Reason. But both are alike in this, that they can never have their source solely in the axiom of analysis, viz., the principle of contradiction; they require an altogether different principle, notwithstanding that whatever principle they may be deduced from, they must always conform to the principle of contradiction, for nothing can be opposed to this principle, although not everything can be deduced from it. I will first of all bring synthetic judgements under certain classes.
(1) Judgements of experience are always synthetic. It would be absurd to found an analytic judgement on experience, as it is unnecessary to go beyond my own conception in order to construct the judgement, and therefore the confirmation of experience is unnecessary to it. That a body is extended is a proposition possessing a priori certainty, and no judgement of experience. For before I go to experience I have all the conditions of my judgement already present in the conception, out of which I simply draw the predicate in accordance with the principle of contradiction, and thereby at the same time the necessity of the judgement may be known, a point which experience could never teach me.
(2) Mathematical judgements are in their entirety synthetic. This truth seems hitherto to have altogether escaped the analysts of human Reason; indeed, to be directly opposed to all their suppositions, although it is indisputably certain and very important in its consequences. For, because it was found that the conclusions of mathematicians all proceed according to the principle of contradiction (which the nature of every demonstrative certainty demands), it was concluded that the axioms were also known through the principle of contradiction, which was a great error; for though a synthetic proposition can be viewed in the light of the above principle, it can only be so by presupposing another synthetic proposition from which it is derived, but never by itself.
It must be first of all remarked that essentially mathematical propositions are always a priori,