p the light away from their own position.

When it is rightly stated, the Agnostic view of "miracles" is, in my
judgment, unassailable. We are _not_ justifiable in the _a priori_
assertion that the order of nature, as experience has revealed it to
us, cannot change. In arguing about the miraculous, the assumption is
illegitimate, because it involves the whole point in dispute.
Furthermore, it is an assumption which takes us beyond the range of
our faculties. Obviously, no amount of past experience can warrant us
in anything more than a correspondingly strong expectation for the
present and future. We find, practically, that expectations, based
upon careful observations of events, are, as a rule, trustworthy. We
should be foolish indeed not to follow the only guide we have through
life. But, for all that, our highest and surest generalisations remain
on the level of justifiable expectations; that is, very high
probabilities. For my part, I am unable to conceive of an intelligence
shaped on the model of that of man, however superior it might be,
which could be any better off than our own in this respect; that is,
which could possess logically justifiable grounds for certainty about
the constancy of the order of things, and therefore be in a position
to declare that such and such events are impossible. Some of the old
mythologies recognised this clearly enough. Beyond and above Zeus and
Odin, there lay the unknown and inscrutable Fate which, one day or
other, would crumple up them and the world they ruled to give place to
a new order of things.

I sincerely hope that I shall not be accused of Pyrrhonism, or of any
desire to weaken the foundations of rational certainty. I have merely
desired to point out that rational certainty is one thing, and talk
about "impossibilities," or "violation of natural laws," another.
Rational certainty rests upon two grounds--the one that the evidence
in favour of a given statement is as good as it can be; the other that
such evidence is plainly insufficient. In the former case, the
statement is to be taken as true, in the latter as untrue; until
something arises to modify the verdict, which, however properly
reached, may always be more or less wrong, the best information being
never complete, and the best reasoning being liable to fallacy.

To quarrel with the uncertainty that besets us in intellectual
affairs, would be about as reasonable as to object to live one's life,
with due thought