"Science" did the trick.

I might take another example out of many hundreds to show what I mean.
You have an authority which is called, where documents are concerned,
"The Best Modern Criticism." "The Best Modern Criticism" decides that
"Tam o' Shanter" was written by a committee of permanent officials of
the Board of Trade, or that Napoleon Bonaparte never existed. As a
matter of fact, the tomfoolery does not usually venture upon ground so
near home, but it talks rubbish just as monstrous about a poem a few
hundred or a few thousand years old, or a great personality a few
hundred or a few thousand years old.

Now if you will look at that phrase "The Best Modern Criticism" you will
see at once that it simply teems with assumption and tautology. But it
does more and worse: it presupposes that an infallible authority must of
its own nature be perpetually wrong.

Even supposing that I have the most "modern" (that is, merely the
latest) criticism to hand, and even supposing that by some omniscience
of mine I can tell which is "the best" (that is, which part of it has
really proved most ample, most painstaking, most general, and most
sincere), even then the phrase fatally condemns me. It is to say that
Wednesday is always infallible as compared with Tuesday, and Thursday as
compared with Wednesday, which is absurd.

The B.M.C. tells me in 1875 that the Song of Roland could have no
origins anterior to the year 1030. But the B.M.C. of 1885 (being a
B.M.C. and nothing more valuable) has a changed opinion. It must change
its opinion, that is the law of its being, since an integral factor in
its value is its modernity. In 1885 B.M.C. tells me that the Song of