gOD AT THE BOTTOM OF THE GLASS:

Consider the famous "double-slit" experiment. The experimental apparatus consists of a device that sends electrons, one at a time, toward a barrier with a slit in it and, at some distance behind the barrier, a screen that glows wherever an electron strikes it. The journey of each electron can be usefully thought of in two parts. In the first, the electron either hits the barrier and stops, or it passes through the slit. In the second, if the electron does pass through the slit, it continues on to the screen. The flashes seen on the screen line up with the gun and slit, just as we'd expect from a particle fired like a bullet from the electron gun.

But if we now cut another slit in the barrier, it turns out that its mere existence somehow affects the second part of an electron's journey. The screen lights up in unexpected places, not always lined up with either of the slits -- as if, on reaching one slit, an electron checks whether it had the option of going through the other one and, if so, acquires permission to go anywhere it likes. Well, not quite anywhere: Although we can't predict where any particular shot will strike the screen, we can statistically predict the overall results of many shots. Their accumulation produces a pattern that looks like the pattern formed by two waves meeting on the surface of a pond. Waves interfere with one another: When two crests or two troughs meet, they reinforce by making a taller crest or deeper trough; when a crest meets a trough, they cancel and leave the surface undisturbed. In the pattern that accumulates on the screen, bright places correspond to reinforcement, dim places to cancellation.

We rethink. Perhaps, taking the pattern as a clue, an electron is really like a wave, a ripple in some field. When the electron wave reaches the barrier, part of it passes through one slit, part through the other, and the pattern we see results from their interference.

There's an obvious problem: Maybe a stream of electrons can act like a wave (as a stream of water molecules makes up a water wave), but our apparatus sends electrons one at a time. The electron-as-wave model thus requires that firing a single electron causes something to pass through both slits. To check that, we place beside each slit a monitor that will signal when it sees something pass. What we find on firing the gun is that one monitor or the other may signal, but never both; a single electron doesn't go through both slits. Even worse, when the monitors are in place, no interference pattern forms on the screen. This attempt to observe directly how the pattern arose eliminates what we're trying to explain. We have to rethink again.

At which point Copenhagen says: Stop! This is puzzling enough without creating unnecessary difficulties. All we actually observe is where an electron strikes the screen -- or, if the monitors have been installed, which slit it passes through. If we insist on a theory that accounts for the electron's journey -- the purely hypothetical track of locations it passes through on the way to where it's actually seen -- that theory will be forced to account for where it is when we're not looking. Pascual Jordan, an important member of Bohr's circle, cut the Gordian knot: An electron does not have a position until it is observed; the observation is what compels it to assume one. Quantum mechanics makes statistical predictions about where it is more or less likely to be observed.

That move eliminates some awkward questions but sounds uncomfortably like an old joke: The patient lifts his arm and says, "Doc, it hurts when I do this." The doctor responds, "So don't do that." But Jordan's assertion was not gratuitous. The best available theory did not make it possible to refer to the current location of an unobserved electron, yet that did not prevent it from explaining experimental data or making accurate and testable predictions. Further, there seemed to be no obvious way to incorporate such references, and it was widely believed that it would be impossible to do so (about which more later). It seemed natural, if not quite logically obligatory, to take the leap of asserting that there is no such thing as the location of an electron that is not being observed. For many, this hardened into dogma -- that quantum mechanics was a complete and final theory, and attempts to incorporate allegedly missing information were dangerously wrongheaded.

But what is an observation, and what gives it such magical power that it can force a particle to have a location? Is there something special about an observation that distinguishes it from any other physical interaction? Does an observation require an observer? (If so, what was the universe doing before we showed up to observe it?) This constellation of puzzles has come to be called "the measurement problem."

Bohr postulated a distinction between the quantum world and the world of everyday objects. A "classical" object is an object of everyday experience. It has, for example, a definite position and momentum, whether observed or not. A "quantum" object, such as an electron, has a different status; it's an abstraction. Some properties, such as electrical charge, belong to the electron abstraction intrinsically, but others can be said to exist only when they are measured or observed. An observation is an event that occurs when the two worlds interact: A quantum-mechanical measurement takes place at the boundary, when a (very small) quantum object interacts with a (much larger) classical object such as a measuring device in a lab.

Experiments have steadily pushed the boundary outward, having demonstrated the double-slit experiment not only with photons and electrons, but also with atoms and even with large molecules consisting of hundreds of atoms, thus millions of times more massive than electrons. Why shouldn't the same laws of physics apply even to large, classical objects?

Enter Schrödinger's cat, the famous thought experiment beloved by pop-physics expositors and often deployed to wow (and cow) laymen by demonstrating the deep strangeness of quantum mechanics and the mental might of the Scientists who wield it. In fact, Schrödinger offered it as a reductio ad absurdum of the Copenhagen interpretation.

The experiment -- buried in a lengthy 1935 paper, "The Present Situation in Quantum Mechanics" -- asks us to imagine a sealed box with a tiny amount of radioactive substance, a Geiger counter for detecting its decay, a glass jar of cyanide, a mechanism that controls a hammer, and a cat. If the Geiger counter detects that the radioactive substance has decayed, it activates the hammer, which breaks the jar and poisons the cat. Spontaneous decay is a quantum event about whose occurrence we can make only probabilistic predictions. The amount of the radioactive substance is chosen so that the probability of its decaying within one hour is 50 percent. If we leave this sealed box to itself, what can we say about the radioactive substance, and hence the cat, when the hour is up?

If it's literally true, as the Copenhagen interpretation claims, that an unstable atom is in an "indeterminate" state -- neither decayed nor undecayed -- until an act of observation compels it to choose which, then the cat is also in an indeterminate state -- neither alive nor dead -- until we observe it. Someone who takes Copenhagen seriously, Schrödinger says, must say that the cat is neither alive nor dead until we open the box and that, if it is dead, opening the box is what killed it.

Schrödinger describes the thought experiment as a "quite ridiculous case," demonstrating that the attempt to make a principled, radical distinction between a quantum world and a classical world made no sense.

Tweet   @brothersjudd