Most of us have heard of the famous double-slit experiment. Usually it's played out in a lab in seconds. But there's one version, dreamt up by physicist John Archibald Wheeler, that can be played out over much of the galaxy, over millions of years. His thought experiment suggests that we could retroactively determine the fate of ancient photons.
The Double-Slit and the Delayed Choice
The Double-Slit experiment makes an appearance in every high school physics classroom, and for good reason. It gives students their first taste of the truly wild - although many people have explanations for it. Whatever the explanation, the results show us that the world is a stranger place than we have thus far assumed it to be.
A light source, a barrier, and a back wall are all set up. In the barrier are two slits. When a photon comes from the light source, it can pass through the slits. From past experience we know what will happen when a succession of single photons hit the back wall. A pretty little wave pattern will emerge, as if the photons were going through both slits and interfering with themselves the way a wave would. But the photons can't go through both slits. To check which one they go through, we put detectors in both slits. When we let a photon through, we note with smug pleasure that it goes through the left slit. The next goes through the left again, the next through the right, and so on, until we get around a fifty-fifty split between left and right. When we check the back wall, we are dismayed. Instead of the wave pattern, we find a pile of photons on the left and a pile on the right, just as if they were balls lobbed through one slit or the other. Interfering with the photons as they pass through the slits destroys the interference pattern and collapses the wave.
That makes sense, but what if we don't interfere with the photons on their way through the slits? This was the reasoning of John Archibald Wheeler, physicist, eccentric, and world's best reason to call one's child "Archibald." In the mid-twentieth century, he thought of a variation on the Double-Slit experiment. What if along the back wall there were a screen, and behind the screen there were detectors. One detector was focused on the left slit, and one on the right. The screen would stay in place when the photon was released from the light source. It would stay in place as the photon passed through one slit, or the other, or both. Only when the photon had passed through the slits would the screen either be flipped up or not flipped up. The choice to flip the screen wouldn't even be made until then. How could detectors of any kind interfere with something that had already happened?
But, of course, they did. When the screen was down, and no one could detect which slit the photons had come through, there was a wave pattern. When the screen was up, there was a little ping in either detector, and no wave pattern.
The Cosmic Delayed Choice
Wheeler hadn't really expected anything else. The Delayed Choice experiment was meant to make people think about what the Double-Slit experiment meant. To some people, it meant that a choice made in the present could determine events in the past. Although it sounds ridiculous, there is a logic to it. Imagine you flip a switch up and one series of events happen, and flip it down, and a different series of events happen. Those series are absolutely consistent with which way you flip the switch, and they never fail. Doesn't it follow that your flip of the switch was a cause that had an effect?
To show that the Delayed Choice experiment wasn't the result of faulty lab equipment, Wheeler proposed the entire thing on a much, much grander scale. Imagine a distant star emits light. Between that star and the Earth is a galaxy, but instead of blocking the light, it bends light towards the Earth. It bends light in of two different ways. A single photon, going from the star, can take one of two paths, to the left of the galaxy or the right of the galaxy. Suddenly, we have a Double-Slit experiment in space. The photons make their way to Earth, and we can observe them. We can observe either exactly where they come from, or we can neglect to see which side of the galaxy they came around. Again, we have a Double-Slit experiment, and we would get the same results – interference pattern if we don't check the exact origin of the photons, and no interference pattern if we do check the exact origin of the photons.
The interesting thing is, these photons would have made "the choice" between going through one path, the other, or both, millions or billions of years ago. There is no way we could have messed up and measured them as they were coming around the galaxy. And yet, depending on whether we measure them, we will have determined whether they passed through one path, the other, or possibly both. Can we determine, now, events that happened millions of years ago?
Time Travel or Not?
Most physicists think the answer is a resounding "no." No, we cannot kick back with retrocausality. Or, at the very least, the experiment, whether conducted across a lab or across galaxies, doesn't support the idea of time travel.
Instead of thinking of retrocausality, the detractors argue, we should think of a superposition of states. The photon is put in both states when it goes through the slits, or around the galaxies, and stays in both states until it is measured. This, for many, is not a cutesy conceit or an easy way "out" of a complicated situation. It is also not conditional on whether the photon is measured at the moment the photon passes through the slit or millions of years later. The wave collapses, and the particle emerges, when the photon is measured. This idea is supported by any version of the Double-Slit experiment, whether the choice to measure is made before the photon is released or long after it has been put in a superposition of states.
If however, you'd like to do the cosmic Delayed Choice experiment while cackling about your absolute power over space and time, I won't stop you.