When I feel difficulty coming on, I switch to another book I'm writing. When I get back to the problem, my unconscious has solved it.

We mythologists know very well that myths and legends contain borrowings, moral lessons, nature cycles, and a hundred other distorting influences, and we labor to cut them away and get to what might be a kernel of truth. In fact, these same techniques must be applied to the most sober histories, for no one writes the clear and apparent truth-if such a thing can even be said to exist.

The Three Theorems of Psychohistorical Quantitivity: The population under scrutiny is oblivious to the existence of the science of Psychohistory. The time periods dealt with are in the region of 3 generations. The population must be in the billions (±75 billions) for a statistical probability to have a psychohistorical validity.

It is always useful, you see, to subject the past life of reform politicians to rather inquisitive research.

You can prove anything you want by coldly logical reason---if you pick the proper postulates.

And [Asimov]'ll sign anything, hardbacks, softbacks, other people's books, scraps of paper. Inevitably someone handed him a blank check on the occasion when I was there, and he signed that without as much as a waver to his smile — except that he signed: 'Harlan Ellison.

A scientist is as weak and human as any man, but the pursuit of science may ennoble him even against his will.