But if he mistakes the general nature and conditions of the argument when he speaks of it as if it were a mere argument from analogy, his _extension of the analogy_, and the reasonings founded on it, are equally unjustifiable and inconclusive. He forgets that analogy proceeds on a partial resemblance in _some respects_, between things which differ _in other respects_, and that even induction itself requires a perfect resemblance only _in those respects_ on which the inference depends. There may be such a resemblance between the marks of design in nature and in art as to warrant the inference of a contriver in both; and yet _in other respects_ there may be a dissimilarity which cannot in the least affect the validity or the certainty of that inference. It is only when we _extend the analogy_ beyond the inductive point, that the conclusion becomes, in some cases, merely probable, in others altogether doubtful. If we advance a step further than we are warranted to go by obvious and certain analogies, our conclusions must be purely conjectural, and cannot be accepted as inductive inferences. From what we know of this world, and of God's design in it to make Himself known to His intelligent creatures, we may infer, with some measure of probability, that other worlds may also be inhabited by beings capable, like ourselves, of admiring His works, and adoring His infinite perfections; but if we go further, and infer either that all these worlds must _now_ be inhabited, or that the inhabitants must be _in all respects_ constituted as we are, we pass far beyond the point to which our knowledge extends, and enter on the region of mere conjecture. And so when Mr. Holyoake extends the analogy, so as to include not only the marks of design, on which the inductive inference rests, but also the forms of organization, with which in the case of man, intelligence is at presented associated, although not identified, he goes beyond the point at which analogy and induction combine to give a _certain_ conclusion, and introduces a conjectural element, which may well render his own inferences extremely doubtful, but which can have no effect in weakening the grounds of our confidence in the fundamental law, which demands an adequate cause for the marks of design in nature.