

Can't be done.
You will encounter
someone who sees something you don't. You may encounter a Feynman.
If you teach, you'll occasionally encounter a student who sees something
you are missing, even in your own field. Occasionally Feynman looked
absurd to someone. ("Feynman wasn't a misogynist; he could be rude
to anyone," was a recent comment in Physics Today.)
However, we can make a list
of useful rules to keep our silliness at a minimum. Here's a start:
Don't
use science to "prove" some farout hypothesis when you can't use the same
science to solve problems it was originally built to solve.
For example, don't use quantum mechanics or special relativity to prove
Uri Geller can mentally bend spoons if you can't successfully apply quantum
mechanics to S = k ln W to calculate entropy or use E=mc² to calculate
the free energy yield of a nuclear explosion without knowing the anticipated
change in rest mass.
Be
able to recognize the distinction between a representation and what it
represents.
For example: the independent variables in an algebraic equation represent
sets of numbers that get operated upon as the equation describes.
Words, like "energy" and "implication" represent sets of experiences, observations,
concepts, constructs, mental models, etc, which get woven into theories
that help us anticipate outcomes of actions we might take. Money
represents value, but is not the same thing.
Learn
to use statistical reasoning.
For example, become able to see how "expectation value" predicts return
in gambling, and learn how to assure that you win in the long run.
Become able to see how random fluctuations affect, microscopically, numbers
that can, nevertheless, be predicted macroscopically.
Acquire
a sense of "selfdeception."
For example, become aware of biases to believe, and how they can distort
apparent
statistics.
Acquire resistance to belief in anything that has little more substance
than its desirability. Acquire faith in reason.
Don't
shirk hard work.
For example, before starting a journey of understanding of elementary physics,
understand acceleration to at least the depth necessary to "see" the direction
of acceleration of the ball as it bounces off the surfaces of the handball
court. An understanding of the simple "p" test of statistical significance
is necessary before we state with certainty that the boardcertified internist
was wrong, and the naturopath was right when he claimed that the juices
in the root of
Conium maculatum cured our intractible indigestion.
Understanding comes best from experience. Socrates understood, through
experience, how Conium maculatum cures intractable indigestion.
Permanently. Conium maculatum is the poison hemlock by which
Socrates was executed.
Learn,
as Socrates understood we must, by discovery.
For example, work on those puzzles, such as Wason's
card selection, vos Savant's door selection, bouncing ball, etc, long and
hard before giving up and looking up the answer. Discuss with others
who also haven't yet "seen." Once you look it up, you have lost most
of your opportunity. Learning is not your goal; "seeing" is. 
