Can a theory that speaks only in probabilities be called finished?

That time, that place

In the two years 1925 and 1926, the calculating tools of quantum mechanics were finished. Several people built them in different shapes, and astonishingly they all gave the same answers.

The calculations all worked — hydrogen's spectrum, molecular bonding, electrons in metals.

The trouble was that nobody could agree on what any of it meant.

Why this question

In October 1927 the fifth Solvay Conference met in Brussels on electrons and photons. Twenty-nine people attended; seventeen of them held or would hold a Nobel Prize.

On one side stood Bohr, Heisenberg and Born, arguing that one must abandon the idea that a particle has a definite position and speed before it is measured. Before measurement there is no such value.

On the other stood Einstein. He did not say the theory was wrong; he granted that the numbers came out right. He denied that this could be all there is.

What was found

The argument happened at breakfast rather than in session.

Einstein came down each morning with a thought experiment — an imagined apparatus designed to slip past uncertainty and pin down position and momentum at once. A screen with a slit, a box hung from a spring, a shutter worked by a clock.

By evening Bohr had found the flaw. Once he used Einstein's own general relativity to do it: weigh the box on a spring and the box shifts in the gravitational field, and the clock inside it runs slow.

By the close of the conference most had leaned to Bohr. But nothing had been proved. Each side had demolished the other's thought experiment; neither had shown which way nature actually goes.

A line from a letter Einstein wrote to Born around this time is widely known: that God does not play dice.