What Is Schrödinger’s Cat?

On 19 August 1935, Erwin Schrödinger answered a letter from Albert Einstein with a grim little scenario. A cat, a flask of poison and a speck of radioactive material, sealed together in a steel box. Ninety years on, Schrödinger’s cat is the most famous image in quantum physics, and it is almost always told backwards. Schrödinger was not claiming that a cat can be alive and dead at once. He was arguing that if the theory seemed to say so, something in the way people read the theory had gone badly wrong. Here is what he actually wrote, what he meant, and what happened when physicists started building real cats.

The thought experiment, as Schrödinger wrote it

The cat appears in a single paragraph of a long, three-part paper, published in the German journal Die Naturwissenschaften in November 1935. Its title translates as “The present situation in quantum mechanics.” Schrödinger introduced the cat as a burlesk case, a deliberately ridiculous one. The same paper gave physics its German word for entanglement.

The setup is precise. A cat is shut in a steel chamber together with what Schrödinger called a devilish device, placed out of the cat’s reach. A Geiger counter holds a tiny amount of radioactive material, so little that within an hour one atom is as likely to decay as not. If an atom decays, the counter fires. A relay then releases a hammer, the hammer shatters a small flask of hydrocyanic acid, and the cat dies.

Leave the whole system alone for an hour. Common sense says the cat is simply alive or dead, and we do not yet know which. Quantum mechanics, read literally, says something stranger. The atom is in a superposition of decayed and not decayed, and every link in the chain inherits it. The wavefunction for the whole box, Schrödinger wrote, contains the living and the dead cat “mixed or smeared out in equal parts.” He even added a bracketed apology for the phrasing.

The box Schrödinger described

It started with Einstein’s gunpowder

The cat was born in correspondence. In May 1935 Einstein, Boris Podolsky and Nathan Rosen had published their argument that quantum mechanics was incomplete, and Schrödinger and Einstein began a long exchange of letters about it. Both men had left Germany after the Nazis took power in 1933, and they argued by post.

On 8 August 1935 Einstein offered an example of his own: a charge of gunpowder, unstable enough to have even odds of exploding within a year. On the quantum description, after a year its state would be a blend of exploded and unexploded. In reality, Einstein pointed out, there is nothing in between. Eleven days later Schrödinger replied with the version that stuck. He swapped the gunpowder for a Geiger counter, poison and a cat.

Einstein never let it go. Writing to Schrödinger in December 1950, he said the cat “refuted most elegantly” the idea that the wavefunction gives a complete description of reality. His description also put his own charge of gunpowder into the box alongside the cat. Fifteen years on, the two thought experiments had merged into one.

What Schrödinger was actually arguing

The cat is a reductio ad absurdum: take an assumption, follow it to a conclusion nobody can accept, and the assumption falls. The assumption under attack was that the wavefunction is a literal picture of physical reality.

Schrödinger saw the wavefunction differently. He called it a catalogue of expectations: a record of what we can predict about measurements, not a portrait of what is there. At the scale of atoms nobody can look directly, so a smeared-out description is hard to contradict. The cat shows that the same mathematics does not stay politely inside the atom. A tiny amplifier chain carries atomic uncertainty all the way up to a cat, and at that scale anyone can open the box and look.

He made the point with an analogy about photographs. A blurry photograph of a sharp object is one thing. A sharp photograph of a fog bank is another. The first reflects our limited information; the second reflects a genuinely fuzzy world. Reading the smeared-out wavefunction as a picture of fog, he argued, commits you to a smeared-out cat. Nobody believes in one, so the literal reading has to go.

Three things the cat does not say

It does not say cats can be alive and dead. That was the conclusion Schrödinger used to embarrass a view, not a claim he endorsed. No experiment has ever put anything remotely like a cat into a superposition.

It does not say a human observer is needed. The idea that consciousness collapses the wavefunction is a minority position, historically linked to the physicist Eugene Wigner, and it is not what Schrödinger meant. In his setup the Geiger counter registers the decay long before anyone opens the box. The cat itself is hardly an unobservant bystander.

It does not make opening the box special. Modern physics has a better account of why nobody ever sees a half-dead cat. A cat is a warm, wet object constantly exchanging air molecules, heat and light with its surroundings. Every such exchange carries away information about which branch of the superposition it is in. That process, decoherence, wipes out any detectable interference between “alive” and “dead” almost instantly, long before the lid moves. In practice the environment is always watching.

The real question: where does the quantum world end?

Strip away the cat and a serious problem remains. The equation that governs quantum systems is linear, which means superpositions never break up on their own. Apply the Schrödinger equation to the atom, the counter, the hammer and the cat together, and it predicts a superposition of the whole chain. Yet every measurement anyone has ever made gives one definite result.

Somewhere along the chain, then, the quantum description has to hand over to a classical one. Nobody has ever said exactly where that handover happens, or why. This is the measurement problem, and the cat is its most memorable illustration.

Decoherence explains why the handover looks sharp in practice: it hides the superposition beyond any hope of detection. On most readings, it does not explain why one particular outcome occurs. The cat stays a live question precisely because this last step remains open.

Wigner later pushed the chain one link further. Put a friend inside the lab, he suggested, and treat the friend as a quantum system too. Is the friend’s observation a superposition until Wigner asks the result? The question is still generating papers.

How each interpretation answers the cat

Every major interpretation of quantum mechanics has an answer. Each buys it by giving something up.

Copenhagen. Copenhagen treats measuring devices as classical, so the measurement happens at the Geiger counter. The cat was never in a superposition. The price is a line between quantum and classical that the theory never draws precisely.

Many-worlds. The superposition is real and never collapses. The universe branches: in one branch the cat lives, in another it dies, and an observer exists in both. The price is a vast number of unobservable branches.

Pilot-wave theory. The cat is always definitely alive or definitely dead. The wavefunction is real, but it guides particles that always have actual positions. The price is explicit nonlocality.

Objective collapse. These theories modify the Schrödinger equation slightly, so superpositions collapse on their own at a rate that grows with size. An atom can stay superposed for ages; a cat collapses almost immediately. The price is new physics, which experiments are actively hunting for.

Epistemic views, such as QBism. The wavefunction encodes an agent’s expectations, close to Schrödinger’s own “catalogue.” Nothing about the cat is smeared; only our information is. The price is giving up a description of reality that holds for everyone at once.

Then physicists started building cats

For half a century the cat stayed on paper. In the 1980s that began to change. Anthony Leggett and others argued that a large object with many internal parts could, in principle, show genuine quantum behaviour, provided it was isolated well enough from its surroundings. The phrase “Schrödinger cat state” became a technical term for any superposition of two clearly distinct states of something bigger than a single particle.

In 1996 two groups made them. David Wineland’s team in Boulder cooled a single beryllium ion almost to rest, then put it into a superposition of two separated positions, and confirmed it through interference. In Paris, Serge Haroche’s group created a superposition of two distinct states of a microwave field in a cavity. Then they watched decoherence turn it into an ordinary either-or while it happened. The Paris team noted that the cat itself could only ever be a metaphor. Their experiments made the metaphor measurable, and the two leaders shared the 2012 Nobel Prize in Physics.

In 2000, groups in the United States and the Netherlands went bigger. They put electric current in a superconducting loop into a superposition of flowing clockwise and anticlockwise at once, a state involving the coordinated motion of an enormous number of electrons.

From paradox to hardware

The strangest twist is that cat states became useful. Superconducting cavities can now hold superpositions of light containing around a hundred photons, a feat reported by a Yale group in 2013. Those states turned out to be excellent places to store quantum information.

The reason is structural. Flipping a cat qubit from one state to the other means reversing a whole collective wave at once, which the physics makes exponentially unlikely. The physics suppresses one of the two main kinds of error before any error correction runs. That is the idea behind AWS’s Ocelot chip, published in Nature in 2025, which we cover in our roundup of the latest in quantum computing. A thought experiment designed to ridicule the theory now helps protect qubits from noise.

How big can a cat get?

Headlines regularly announce “the biggest Schrödinger’s cat yet.” The honest answer is that it depends on the yardstick, and three different experiments hold three different titles.

The widest. In 2015 a Stanford group put atoms into superpositions of two places about half a metre apart.

The heaviest. In 2023 Yiwen Chu’s group at ETH Zurich prepared cat states of a crystal resonator with an effective mass of 16.2 micrograms, around 10¹⁷ atoms. The catch is in what was superposed. The crystal was not in two places; its atoms were oscillating in two opposite phases at once.

The most macroscopic overall. In January 2026 Markus Arndt’s group in Vienna reported sodium nanoparticles of more than 7,000 atoms, each spread over a distance more than ten times its own diameter. On the standard measure, which combines mass, separation and duration, it beats every previous test by an order of magnitude. It also sets the tightest limit yet on theories where large superpositions collapse on their own.

None of these is anywhere near a cat. But none of them has found the edge of quantum mechanics, either.

Established, contested, unproven

Established. Schrödinger’s cat was a deliberate reductio, written in 1935 in reply to Einstein, aimed at the literal reading of the wavefunction. Superpositions of clearly distinct states, “cat states”, are routine in laboratories, from single ions to 16-microgram crystals to clusters of 7,000 atoms. Decoherence explains why such states are never seen in everyday objects.

Contested. What the cat’s wavefunction actually represents, and so how the paradox is resolved. Copenhagen, many-worlds, pilot-wave, collapse models and epistemic views all answer it, and all remain consistent with every experiment so far. Physicists who agree on every measurement still disagree here.

Unproven. Whether there is a size limit on superposition. If objective-collapse theories are right, somewhere between today’s records and an actual cat, superpositions stop surviving on their own. Each larger experiment narrows where that limit could hide, and none has found it.