144

16

But this is a vague statement.
Let us see what can be meant by saying that we must "rely on the assumption that what seems to be good reasoning is so".
If it means that, however closely er criticize our reasoning there will remain a bare possiblity of its being bad, I grant that, at once.
We are not ceetain in that purely theoretical and artificial sense, of anything whatever.
For example, we all know that any man is liable to commit an error in addition.
It is therefore barely possible that all men who have added two to two since the beginning of time have by singular coincidence always commited the same blunder.
In the theoretical sense, therefore, this is not certain; and I do not pretend that the prepositions of logic are any more certain than those of arithmetic.
But what utility, I should be glad to know, does this meaning of the word certainty subserve?