As you admit that they might or might not have designed the collision of their balls and its consequences the question arises whether there is any way of ascertaining which of the two conceptions we may form about it is the true one.  Now, let it be remarked that design can never be demonstrated.  Witnessing the act does not make known the design, as we have seen in the case assumed for the basis of the argument.  The word of the actor is not proof; and that source of evidence is excluded from the cases in question.  The only way left, and the only possible way in cases where testimony is out of the question, is to infer the design from the result, or from arrangements which strike us as adapted or intended to produce a certain result, which affords a presumption of design.  The strength of this presumption may be zero, or an even chance, as perhaps it is in the assumed case; but the probability of design will increase with the particularity of the act, the specialty of the arrangement or machinery, and with the number of identical or yet more of similar and analogous instances, until it rises to a moral certainty—­i. e., to a conviction which practically we are as unable to resist as we are to deny the cogency of a mathematical demonstration.  A single instance, or set of instances, of a comparatively simple arrangement might suffice.  For instance, we should not doubt that a pump was designed to raise water by the moving of the handle.  Of course, the conviction is the stronger, or at least the sooner arrived at, where we can imitate the arrangement, and ourselves produce the result at will, as we could with a pump, and also with the billiard-balls.

And here I would suggest that your billiard-table, with the case of collision, answers well to a machine.  In both a result is produced by indirection—­by applying a force out of line of the ultimate direction.  And, as I should feel as confident that a man intended to raise water who was working a pumphandle, as if he were bringing it up in pailfuls from below by means of a ladder, so, after due examination of the billiard-table and its appurtenances, I should probably think it likely that the effect of the rebound was expected and intended no less than that of the immediate impulse.  And a similar inspection of arrangements and results in Nature would raise at least an equal presumption of design.

You allow that the rebound might have been intended, but you require proof that it was.  We agree that a single such instance affords no evidence either way.  But how would it be if you saw the men doing the same thing over and over? and if they varied it by other arrangements of the balls or of the blow, and these were followed by analogous results?  How if you at length discovered a profitable end of the operation, say the winning of a wager?  So in the counterpart case of natural selection:  must we not infer intention from the arrangements and the results?  But I will take another case