In the present chapter it is proposed to prove, that—

It is more probable that any law, at the knowledge of which we have arrived by observation, shall be subject to one of those violations which, according to Hume's definition, constitutes a miracle, than that it should not be so subjected.

To show the probability of this, we may be allowed again to revert to the Calculating Engine: and to assume that it is possible to set the machine, so that it shall calculate any algebraic law whatever: and also possible so to arrange it, that at any periods, however remote, the first law shall be interrupted for one or more times, and be superseded by any other law; after which the original law shall again be produced, and no other deviation shall ever take place.

Now, as all laws, which appear to us regular and uniform in their course, and to be subject to no exception, can be calculated by the engine: and as each of these laws may also be calculated by the same machine, subject to any assigned interruption, at distinct and definite periods; each simple law may be interrupted at any point by a portion of any one of all the other simple laws: it follows, that the class of laws subject to interruption is far more extensive than that of laws which are uninterrupted.