l sentiment the two kinds of compulsion
differ as life and death. To have no liberty because of being absolutely
in the hands of Almighty God is quite another thing from having no
liberty, as being under the dominion of a dead iron rule. It seems
possible to accept the one and call it an unfathomable mystery; but to
accept the other is to call life a delusion and the moral law a dream.
And in the second place, the doctrine of necessity advanced as a theory
and based on arguments not resting on facts, is a very different
antagonist from the same doctrine advanced as a conclusion of science,
and as deducible from a mass of co-ordinated observations. We may
dismiss the mere theory after showing that it has not substance enough
to hold its ground in ordinary life. We cannot so treat what claims to
be a scientific inference.

The modern examination of the question begins with Hume, who maintains
that the doctrine of liberty and that of necessity are both true and of
course compatible with each other. But his arguments touch only the
broad question whether they are true for practical purposes, not whether
either is true in the strict sense and without exception or
modification. To Kant's system, on the contrary, it was essential that
both doctrines should be true in the strictest sense. Holding that
invariable sequence was a law of Nature known independently of
experience and applicable to all phenomena in the minutest detail, he
could not allow that any act of the human will lay outside the range of
this law. Such an act being a phenomenon must, in his view, be subject
to the law which the constitution of our minds imposed on all phenomena
apparent to us. And yet, on the other hand, holding that the eternal
Moral Law made us responsible for all our acts, he could not but
maintain that in the doing of those acts we must be free. His mode of
reconciling the two opposites amounted to this, that our action
throughout life considered as a whole is free, but that each separate
act considered by itself is bound to the preceding acts by the law of
invariable sequence. We may illustrate this by the familiar instance of
a prism acting on a ray of light. The ray has or may have a colour of
its own before it passes through the prism. The prism spreads it out and
shows a series of colours. The order in which this series is arranged is
determined by the character of the prism acting on the nature of the
ray. The colours when combined give