麻豆传媒

TO TRY to explain quantum physics, it鈥檚 always best to start with Einstein. He had two famous complaints about it: how it seems to say that 鈥淕od plays dice with the universe鈥�, and how it appears to allow 鈥渟pooky action at a distance鈥�.

A very simple experiment illustrates both problems. First, imagine sending a single photon of light through a beam splitter, a piece of glass that either reflects it or transmits it with a 50 per cent chance. One of two detectors clicks in each of the two cases. If we run this experiment many times, we find that half of the time detector 1 clicks, and the other half detector 2. But quantum mechanics shows that nothing in the universe can tell us which detector will click any one time. Einstein came from a classical frame of mind where everything is deterministic and predictable: but it seems randomness really is at the heart of quantum mechanics.

But suppose now, instead of detecting the photon with two detectors behind the first beam splitter, you first recombine the light at a second beam splitter, with two detectors now behind here. What鈥檚 remarkable now is that the photon behaves perfectly deterministically. Basically, only detector 1 clicks, never detector 2 (see diagram 鈥淲eirdness, squared鈥�).

How can two random things be put together to give you something 100 per cent predictable? In quantum mechanics, the only way to explain it is to say that the photon actually splits into two at the first beam splitter and takes both routes: it鈥檚 reflected and transmitted simultaneously. When these two possibilities converge on to the second beam splitter, they interfere like water waves: when they go to detector 1, they amplify each other, and when they go to detector 2, they cancel out and make no signal.

We call this state of the photon a quantum superposition. Einstein called it 鈥渟pooky action鈥� because it looks as though the photon can exist in two places. But we鈥檝e tested the effect not just on photons, but also on atoms, subatomic particles and larger particles and molecules, and every time we鈥檝e actually got them to behave this way. Randomness and spooky action at a distance are here to stay.

That鈥檚 where Erwin Schr枚dinger, entanglement and that exciting and stimulating thought experiment we all know and love comes in: Schr枚dinger鈥檚 cat. Imagine I put a cat into the laboratory together with this photon that鈥檚 undergoing interference. If the photon goes one way at the beam splitter, nothing happens. But if it goes the other, it hits a very fragile bottle, releasing poison so the cat dies. If quantum mechanics really describes the whole experiment, and the photon does go both ways simultaneously, then the cat is not either dead or alive; it too must be both things simultaneously.

Genie outside the box

The story doesn鈥檛 stop there. What if there is a physicist observing this experiment 鈥� how would quantum mechanics describe this situation? Well, you would now have a superposition of a physicist who in one branch sees a living cat and in another branch a dead cat. Anyone who interacts with these kinds of superposed states and finds out what state it鈥檚 in has to join in and actually split according to the same two possibilities. This is what quantum mechanics says would happen, if the whole universe really can be described quantum mechanically. If you think small objects like photons and atoms are weird, and can exist in many different states at the same time, then anything that they couple to has also got to become weird.

There鈥檚 a version of the Schr枚dinger cat experiment where you can test this idea 鈥� in principle, because we鈥檙e nowhere near close to doing this in reality yet. Imagine that Alice is the chief experimental physicist outside the lab, and Bob is inside the lab, observing the interferometer and the cat and the two distinct possible outcomes. Alice wants to test whether Bob really sees two alternatives simultaneously, and whether he sees a definitive outcome at some point in the experiment.

So she waits for the photon to go through the first beam splitter and Bob to look at the outcome. Quantum mechanics says that following this point, there is one happy copy of Bob with a live cat and another sad Bob with a dead cat. At this point Alice sneaks a piece of paper under the door so Bob picks it up 鈥� both copies of Bob, if you like. On it, she asks: 鈥渄o you see a definitive state of the cat?鈥�. Note she doesn鈥檛 ask 鈥渄o you see a dead or alive cat?鈥�, because if Alice got the answer to that question, she would join one of these two quantum worlds. What鈥檚 interesting here is that quantum mechanics would suggest that in both branches, Bob must answer the question Alice actually asks with 鈥測es鈥�.

So Bob writes his answer, and then gives Alice the piece of paper back under the door. Only now does Alice complete the interferometer with the second beam splitter. This is tantamount to undoing the whole experiment and returning it to the original state, before the photon split. You recombine the two possibilities, and bring them back into just one universe. That鈥檚 difficult to do in practice, but the conclusion of the thought experiment is that halfway through, if the laws of quantum mechanics hold, Alice can confirm that Bob really is entangled to the cat: there is one branch with dead cat and sad Bob, and another with live cat and happy Bob.

There is another interesting twist on that, one proposed by my friend and colleague David Deutsch at Oxford. He also invented the idea of quantum computers, and I think you can see why these experiments might make someone think about massive parallel computations, all happening simultaneously but on one and the same device.

Now, as far as each of the Bobs is concerned, there is nothing unusual about their situation: each of them exists in their own alternative world and sees one definitive outcome, a dead or alive cat, and behaves as though the other Bob does not even exist. So armed with her knowledge, Alice can sneak another piece of paper in halfway through this experiment to tell Bob that he鈥檚 actually in two different branches. She can say, I know that each of you doesn鈥檛 know about the existence of the other one, and that you definitely see one state of this cat or another, but I know from my position observing from outside the experiment that you are now split into two different states. But she can鈥檛 know which state Bob鈥檚 in, or she loses her perspective and becomes part of his world.

鈥淨uantum theory suggests that nothing evolves and everything that will happen, already has鈥�

Curiouser and curiouser

You can go further down this rabbit hole. What if there is another observer, call him the Mad Hatter, who is outside Alice鈥檚 laboratory so the he can control Alice, who in turn can control what happens to Bob, who in turn is observing the cat鈥�? You can follow the same logic, and you come to the same conclusion as we did before. If Alice figures out what鈥檚 happening with Bob, she can communicate this to Mad Hatter. She can open her own door towards the Mad Hatter and say look, Bob has made an observation and the cat is alive, but as soon as she does so they鈥檙e all part of that world, even if from the point of view of an ultimate observer, there must be this other world in superposition, where the bad news is that the cat is dead.

This picture of observers observing observers is called Wigner鈥檚 friend, after Eugene Wigner, a quantum physicist who thought very deeply about these issues and asked, what kind of reality does this lead to? The analogy is of a painter who wants to paint, say, a forest 100 per cent faithfully down to the smallest minute detail. When he finishes, he realises that one important bit is missing: himself. So he decides to add himself, but then of course realises that the painter painting the first painter is missing 鈥� and so on ad infinitum.

This is very similar to how, in this picture, a definite reality emerges through a sequence of entanglements that you get from interactions that, we are now assuming, are completely quantum mechanical: between the photon and the poison, the poison and the cat, the cat and Bob, Bob and Alice, Alice and the Mad Hatter and so on. As with the somewhat paranoid painter, it鈥檚 not a completely faithful image of reality 鈥� because there is always another observer missing.

This leads to another fascinating conclusion, one that many of us are actively researching. It is linked to a question that that you frequently get when you talk about this: what does it feel to be in another universe? The interesting thing is that, quantum mechanically, this is the same as asking what would it feel like to exist at another time.

Two physicists called Don Page and William Wootters showed this in the 1980s in a paper whose title is 鈥淓volution without evolution鈥�. What they suggest is that the universe at different times is really just different quantum universes. Nothing really evolves and everything that will happen has already happened: it鈥檚 all sitting simultaneously in this 鈥渂lock鈥� universe that contains all possible things that can happen.In this case, the components of the entangled state are just different times at which the universe exists. This, believe it or not, is a fully consistent way of thinking about quantum mechanics.

Just finally, I鈥檇 like to bring things back to experiments. The reason why doing this Schr枚dinger type of experiment is so difficult in practice is that it relies on the sort of communication between the different observers in the experiment that really only humans can do. You might perhaps substitute Bob, say, with any computer. In fact my impression is that it鈥檚 more likely that a very simple artificial intelligence system will actually be the first to undergo this kind of experiment, in which we will be communicating with it and asking questions, like how do they feel about being in one state or another and things like that.

But the next milestone would be to put something like viruses or bacteria in a superposition state. You might think it鈥檚 something we might like to do with covid-19, if only as a revenge, but it鈥檚 probably too large for the present technology 鈥� you would need something two orders of magnitude smaller in mass. But in principle you can do these experiments with two microbes confined between two mirrors, where a split photon would excite one and not the other. Of course, you cannot communicate with bacteria and ask them whether they feel they鈥檙e in a definitive state of being excited or not, but you can confirm from the light they emit whether they are entangled or not.

It鈥檚 still very hard to do. We鈥檙e trying in Oxford, and there is another group in Vienna run by Markus Arndt that I think is also close to being able to superpose a virus in a number of different spatial locations. It鈥檚 hard to speculate how long it will take, but the race is certainly on. We鈥檙e really at the level where we can test some of these ideas that maybe 100 years ago would would be thought to be completely crazy, maybe impossible, maybe even contradictory. It鈥檚 a very exciting time.

From cosmology to climate change鈥�
Upcoming events include Mark Maslin on coronavirus and climate and Dan Hooper on what happened at the big bang. Details on all events at