What the Theory Says, and What It Doesn't

Almost everything strange in quantum mechanics is visible in one experiment, and the experiment can be described in a paragraph.

Fire particles at a barrier with two narrow slits in it and let them land on a screen behind. With both slits open you get an interference pattern: bands of many hits separated by bands of almost none, the signature of waves adding and cancelling. Fine: light is a wave, and this was settled in 1801.

Now fire them one at a time, slowly enough that only one particle is ever in the apparatus. Each arrives as a single localised dot; there is nothing wave-like about an individual arrival. But the dots accumulate into the same interference pattern. And it has been done not just with photons but with electrons, neutrons, whole atoms, and molecules of several hundred atoms.

Nothing is interfering with anything else, because there is nothing else in there. Whatever determines where a particle lands depends on both slits being open, for a particle that arrives in one place.

Close one slit and the pattern disappears, as you would expect. Put a detector at the slits to see which one each particle goes through, and the pattern disappears too, even with both open. This is the result usually summarised as "observation changes the outcome", and the summary is where the trouble starts.

The uncertainty principle gets the same treatment. Heisenberg's own first explanation was mechanical: to see an electron you must bounce a photon off it, and the photon kicks it, so measuring position disturbs momentum. It is a vivid story and it is not the modern understanding. The relation follows from position and momentum being what mathematicians call Fourier conjugates, which is the same relationship that holds between the duration of a musical note and the definiteness of its pitch. A very short pulse does not have a well-defined frequency, not because measuring it is hard, but because a brief pulse is built from a spread of frequencies. A wave packet narrow in space is necessarily broad in wavelength, and momentum is wavelength. There is no arrangement of apparatus that escapes it, because the limitation is about what the quantities are.

Then entanglement, which is the most frequently abused result in physics. Prepare two particles in a joint state such that their properties are correlated, separate them, and measure one. The outcome of the second measurement is then constrained by the first, however far apart they are. John Bell showed in 1964 that these correlations are stronger than any theory of pre-existing local properties can produce, and the experiments (refined over decades, closing loophole after loophole, and rewarded with the 2022 Nobel) confirm the violation.

This is genuinely strange and it does not permit faster-than-light signalling. The result you get from your half is random. The correlation only shows up when the two sets of results are brought together and compared, which requires sending them by ordinary means at ordinary speed. The no-communication theorem makes this precise, and it is a theorem, not a hope. Entanglement is a constraint on correlations, not a channel.

What about the transition to the ordinary world? Why does none of this show up in a cricket ball?

The answer that emerged in the 1970s and 80s is decoherence. A quantum system in isolation can maintain the delicate phase relationships that produce interference. A system touching its environment (air molecules, stray photons, thermal vibration) becomes entangled with all of it, and the phase information leaks out into degrees of freedom nobody can track. Interference becomes unobservable, not as a matter of principle but as a matter of overwhelming practice, and for an object of everyday size it happens in a time too short to write down comfortably. This is also why quantum computers are kept isolated and near absolute zero: they are fighting decoherence, and it is the whole engineering problem.

Decoherence disposes of the consciousness stories that attach themselves to this subject. The word "observer" in the early literature meant a measuring apparatus, and decoherence proceeds in a sealed laboratory with nobody watching, as experiments with delayed readout demonstrate directly. Eugene Wigner did briefly favour a consciousness-based interpretation, and abandoned it. The theory is strange enough without the embellishment, and the embellishment replaces a precise difficulty with a vague one.

Because there is a precise difficulty, and decoherence does not solve it. The formalism describes superpositions evolving smoothly; observation always finds exactly one outcome. Decoherence explains why you never see the interference, and it does not explain why this outcome rather than that one. That is the measurement problem, and it is what Schrödinger's cat was an objection to rather than an illustration of: if superposition applies to everything, the mathematics has no principled place to stop.

A century later the competing answers (Copenhagen's refusal to ask, many worlds, pilot-wave theory, objective collapse models) make identical predictions in nearly every accessible case, which is why the argument has lasted and why it is an argument about interpretation rather than a gap in the theory's accuracy. Quantum mechanics is the most precisely confirmed theory in the history of science. Nobody agrees on what it says.