Physics is unable to stand on its own feet, but needs a metaphysics on which to support itself, whatever fine airs it may assume towards the latter.
[...] Certainly the whole present condition of all things in the world or in nature must necessarily be capable of explanation from purely physical causes. But such an explanation - supposing one actually succeeded so far as to be able to give it - must always just as necessarily be burdened with two essential imperfections [...] On account of these imperfections, everything so explained would still really remain unexplained.
The first imperfection is that the beginning of the chain of causes and effects that explains everything, in other words, of the connected and continuous changes, can positively never be reached, but, just like the limits of the world in space and time, recedes incessantly and in infinitum.
The second imperfection is that all the efficient causes from which everything is explained always rest on something wholly inexplicable, that is, on the original qualities of things and the natural forces that make their appearance in them.
[...] Accordingly there is not a fragment of clay, however little its value, that is not entirely composed of inexplicable qualities. Therefore these two inevitable defects in every purely physical, i.e., causal, explanation indicate that such an explanation can be only relatively true, and that its whole method and nature cannot be the only, the ultimate and hence sufficient one, in other words, cannot be the method that will ever be able to lead to the satisfactory solution of the difficult riddle of things, and to the true understanding of the world and of existence; but that the physical explanation, in general and as such, still requires one that is metaphysical, which would furnish the key to all its assumptions, but for that very reason would have to follow quite a different path.
The first step to this is that we should bring to distinct consciousness and firmly retain the distinction between the two, that is, the difference between physics and metaphysics. In general this difference rests on the Kantian distinction between phenomena and thing-in-itself.
[...] As for the motion of the projected bullet, so also for the thinking of the brain, a physical explanation in itself must ultimately be possible which would make the latter just as comprehensible as the former.
But the former, which we imagine we understand so perfectly, is at bottom just as obscure to us as the latter; for whatever the inner nature of expansion in space, of impenetrability, mobility, hardness, elasticity, and gravity may be - it remains, after all physical explanations, just as much a mystery as thinking does.
But because in the case of thought the inexplicable stands out most immediately, a jump was at once made here from physics to metaphysics, and a substance of quite a different kind from everything corporeal was hypostatized; a soul was set up in the brain.
Yet if we were not so dull as to be capable of being struck only by the most remarkable phenomenon, we should have to explain digestion by a soul in the stomach, vegetation by a soul in the plant, elective affinity by a soul in the reagents, in fact the falling of a stone by a soul in the stone.
Therefore in the same way, physical explanation everywhere comes across what is metaphysical, and by this is reduced to nought, in other words, ceases to be explanation.
[In materialism we see] the unceasing attempt to set up a system of physics without metaphysics, in other words, a doctrine that would make the phenomenon into the thing-in-itself [...] They endeavour to show that all phenomena are physical, even those of the mind; and rightly so, only they do not see that everything physical is, on the other hand, metaphysical also. Without Kant, however, this is difficult to see, for it presupposes the distinction of the phenomena from the thing-in-itself.
[...] Even behind atheism, in itself absurd and often spiteful, there lies, as its inner meaning and truth that gives it strength, the obscure conception of such an absolute system of physics without metaphysics.
Certainly such a system would necessarily be destructive for ethics, and just as theism has been falsely regarded as inseparable from morality, this is really true only of a system of metaphysics in general, in other words, of the knowledge that the order of nature is not the only and absolute order of things. We can therefore set this up as the necessary credo of all righteous and good men: "I believe in a system of metaphysics."
In this respect it is important and necessary for us to be convinced of the untenable nature of an absolute system of physics, the more so as such a system, namely naturalism proper, is a view that of its own accord and ever anew forces itself on man, and can be done away with only by deeper speculation.
[...] Beginningless and endless causal series, inscrutable fundamental forces, endless space, beginningless time, infinite divisibility of matter, and all this further conditioned by a knowing brain, in which alone it exists just like a dream and without which it vanishes - all these things constitute the labyrinth in which naturalism leads us incessantly round and round.
[...] however great the advances which physics (understood in the wide sense of the ancients) may make, not the smallest step towards metaphysics will be made in this way [...] For such advances will always supplement only knowledge of the phenomenon, whereas metaphysics strives to pass beyond the phenomenal appearance to that which appears [...]
Metaphysics [...] remains immanent, and does not become transcendent; for it never tears itself entirely from experience, but remains the mere interpretation and explanation thereof, as it never speaks of the thing-in-itself otherwise than in its relation to the phenomenon.
The World as Will and Representation, Volume II, p.172-7, 183