35 
And the judge will authorise the defence to test that point 
thus: If another hypothesis than A. B.’s guilt, which, as a 
proposition, is naturally feasible, can be even invented, though 
unsupported by any array of proved facts, which may also 
satisfy the facts established before the court, the prosecution 
have failed to establish the guilt of the accused. The 
ingenuity of the lawyers on that side is no less than was 
supposed, and the probability of A. B.’s guilt may remain; but 
it is not proved, and the man must be discharged. 
We also learn that unless the induction be positively demon- 
strative, it must give way in the presence of any adequate, 
intelligent, parole-evidence, affirming a different cause for the 
phenomenon. Another more popular reason supports this 
conclusion. Does one say, ‘‘ The living witness may be dis- 
honest or deceived; but my facts and inductive argument are 
wholly dispassionate, impartial, and valid” ? He forgets that 
his facts also have no. better foundation than the professed 
eye-witnessing of some human witness. Does he say, “ They 
are facts; for I saw them”? He is but a human witness. Or 
if he derives his facts from the observations of others, they 
are mere human witnesses. But the facts are a premise of 
his inductive logic. ‘The inference cannot be more valid than 
its premise. It thus appears that it is wholly unreasonable to 
claim superiority for an induction over testimony, for this is 
as though one should claim that “‘ testimony is stronger than 
testimony.” ‘The only consistent meaning would be the 
arrogant assumption that “my testimony is honest and the 
other’s dishonest.’’ This conclusion, that competent testi- 
mony is superior to any except an absolute, exclusive induc- 
tion, is practically accepted by all sound physicists. Let all 
the facts previously known tend to refer the effect to a 
supposed cause, so that the scientific world is almost prepared 
to accept it as a law; if one competent observer arises, 
testifying to another actual cause for the effect, seen by him 
to produce it in a single case, the other hypothesis is with- 
drawn. For science admits that here is a case which cannot 
be reduced under it. An illustrious instance will be remem- 
bered in the first telescopic examinations of Galileo. He saw 
that the planet Venus was gibbous at a time and in a way she 
would not have been according to the Ptolemaic hypothesis. 
That one observation, with men of true science, made an end 
of the Ptolemaic theory. The only alternatives were to sur- 
render it, or to say that Galileo did not see Venus gibbous at 
that part of her orbit. 
A very important application of these logical principles is 
to the inductions of geologists concerning the mode of forma- 
D2 
