A violation of realism?
It turns out that an event in the future can go back and change the outcome of something that already happened in the past. I’m not kidding. It’s been experimentally demonstrated. Sorry to greet you at the door with a bag of bricks like that.
The “violation of realism” snippet comes from the title of a scientific paper which is a big part of the focus here. It’s not a brand-new paper, and it’s not the only one that describes this kind of thing.
But it’s hard to imagine a more profound outcome. It completely upends our notions of past and future, and of cause and effect. All the Big Questions we ask turn into categorically misplaced nonsense. And I think that’s very refreshing!
But let me rewind just a bit...
Every time I go back home to visit my mom, with my wife and kids, she makes each of us pick out a book at the bookstore, her treat. It’s not like she has a ton of money lying around, but this is the one thing she imposes on you when you visit. She’s an Educator, and always has been. When you get in her zone, you’re going to get educated, whether you like it or not.
We visited her last month, and the book I selected — because you must select one — was Through Two Doors At Once: The Elegant Experiment That Captures the Enigma of Our Quantum Reality by Anil Ananthaswamy, which just came out in paperback. A book comes along every once in a while that sort of changes your trajectory, and this was one of those for me.
By all means read the whole thing if you can, because there’s a lot of fascinating variations in it, but I’m going to try to walk you up to what I felt was the centerpiece of the book, the one result — the “violation of realism” — that made me stare out the window for a while. I may have drooled slightly.
We have to start simply, but that’s OK, because each step of the way is bizarre and amazing in its own right.
Let’s begin with Thomas Young, pictured at left. Young made significant contributions to about a zillion different areas, from medicine to Egyptology. He’s been called “The Last Man Who Knew Everything”. But Young himself felt that his most important contribution was the discovery that light acts like a wave.
In 1801, he did a super-simple experiment that you or I could easily do. He cut two very thin slits into a card and allowed a little bit of sunlight to strike it. Behind the card was a dark wall. You would think you’d simply see two lines on the wall behind the card, but you don’t. You see a pattern like the one pictured at the top of this diary.
That pattern is the same one you’d see if you made waves with water by poking at two points and allowing the waves to ripple out and interfere with each other. When two crests meet, they make a bigger wave, but when a crest and trough meet, they cancel out, like this:
OK, so light acts like a wave. Neato. That doesn’t really surprise people today, but it did in Thomas Young’s day, because the revered Isaac Newton had said that light acts like particles. It would turn out that Newton and Young were both right… ish.
Albert Einstein figured out in 1905 that light comes in little packets that we now call “photons”. You can’t have less light than one photon. They’re so tiny that in full sunlight, about a quadrillion photons hit the head of a pin every second. Einstein found that when a single photon hits a surface, all of its energy is focused in one place, not spread around like some kind of wave. So even though lots of photons might interfere with EACH OTHER and act like waves, a SINGLE photon acts more like a particle. Got it.
OK, then! If we only had some way of shooting photons one at a time through Thomas Young’s double slit, then they couldn’t interfere with each other, so they wouldn’t act like waves. After all, each photon is just an individual little packet, so it’s got to go through one slit or the other.
It would be like a quarterback accuracy drill. We’d have our QB throw balls one at a time through slots, and we’d mark off where they hit. QBs do this kind of thing all the time:
Our QB would throw a bunch of balls, one at a time, and the good throws would hit the back wall behind the two slots. After marking off where all the balls went, we’d get two big stripes on the wall. You wouldn’t get any weird interference pattern. How could you?
These days, it is possible to flick one photon at a time toward a double slit. (Physicists spend a lot of time monkeying around with lasers in basements to make these things possible, OK?) Now we can try Thomas Young’s setup with one photon at a time, instead of having a bazillion photons all at once to interfere with each other. We can record where each one hits, and then look at the pattern they make on the screen. So we’ll end up with two stripes, because we’re basically throwing footballs one at a time, and there’s nothing to interfere with, right?
Wrong! We get the same interference pattern we got when we sent all the photons at once.
The only way to explainify this away is to say each photon did act like a wave and spread out, so that part of it went through each slit. That allowed it to interfere with itself. So far, that doesn’t seem so weird.
But then, when the spread-out photon reached the wall, it suddenly pulled itself together and hit only one spot. It just “picked” one somehow. It didn’t hit the wall like a spread-out wave. If you show where any single photon hit, you get a dot, not any kind of pattern. But when we show all the spots made one at a time by each photon, we DO get a pattern. An interference pattern.
But how could a photon do that? It would be like throwing a handful of sugar at a wall, but when the sugar reached the wall, it would suddenly collect itself into a sugarcube and drop to the floor. Every time you did it, the sugarcube would hit the wall at a different place. How would all the sugar “know” to instantaneously collect itself at one spot, though? Which spot? Plus, how would almost all of it suddenly disappear from one place and reappear at another, traveling at infinite speed?
These are some aspects of what Einstein called “spooky action at a distance”, and he didn’t like it. Not one little bit. But photons don’t really care; they do it anyway.
Well, fine, then. Let’s accept it. Light is weird. It doesn’t have any mass, so it’s just an oddball.
Maybe so, but when we try the double-slit experiment with single electrons, which DO have mass and are indisputably “things”, we get this:
If electrons aren’t “thingy” enough for you, this has also been done with entire molecules, which are GIGANTIC compared to electrons. And yet these molecules show the same kind of behavior.
This means that a real, massive object can be in multiple places at once, until it’s forced to pick one. When we need a logical outcome — like, where on the wall did it hit? — then our object has to “choose”. And so it just … does.
How does it “choose”, though? Nobody knows. Let’s take a minute to simplify this “choosing” demonstration so that the bizarreness of it is crystal clear.
First, we’ll grab a beamsplitter. This is just a surface that splits an incoming light beam. If you look at a window from an angle, it can be a beamsplitter:
With lasers and optics and such, it looks a lot cleaner, with exactly half the light reflected:
Let’s send a single photon through this nice beamsplitter, and put a detector at both possible exits. We’ll see the entire photon come out one way or the other. One of the detectors will go off, but not both. We’ll never get a half-and-half response.
If we keep on doing this, we’ll end up with half the photons going to one detector, and half going to the other. But there’s NO WAY to predict which way any individual photon will “choose” to go. It’s completely random.
Just like in the double-slit experiment, the photon seems to split up and go both ways at once, and it doesn’t “choose” one way or the other until it’s forced to. When a logical outcome is needed — which detector did it hit? — it instantly focuses all of its energy at one spot or the other. It has to “pick” one of the detectors. So it just … does.
“God does not play dice!”, Einstein is supposed to have said, because he didn’t like this random-choice-out-of-nowhere stuff, either. Not one little bit. Once again, photons don’t really care; they do this anyway.
Now back to the double slit for something weirder. If I put a detector beside each slit — that is, if I measure the presence of each photon as it’s passing through — each photon will “choose” one slit or the other, because I asked it to give me a logical, either-or result. Both detectors never go off; only one or the other does.
If I keep on doing this, one photon at a time, then I will finally get my two stripes, like a quarterback throwing footballs, because I’ve forced each photon to dedicate itself entirely to one slit OR the other. Hey, here’s a nice diagram of this someone made:
So a photon, an electron, a molecule, whatever, is really in many places at once until we force it to “choose” one by measuring or observing it, or at least having it interact with something. We’ve established that. Thanks for sticking around with me this far, because now we’re coming to the heart of the matter.
What if there was a way for us to measure the photon without actually doing anything to it? What if we could figure out which slit it went through without affecting it at all? We actually can do this, it turns out, and it’s the basis of that centerpiece experiment I was talking about earlier. The s&!# is about to hit the fan.
You can get a potassium titanyl phosphate crystal to spew out TWO photons at a time by zapping it with a super-focused laser. Not only that, but you can get those two photons to be emitted in two different directions. The really useful thing is that the two photons always end up polarized in opposite directions. If I measure the polarization direction of both photons after they’re emitted, they’ll be exact opposites every single time without fail. That means I can gain information about one of the photons by measuring the other one. Now I just need a way to exploit this and do that double-slit experiment again!
You might own some polarized sunglasses, and if so, you know that when you turn your head back and forth while wearing them, different reflections that come off of surfaces appear and disappear. That’s because polarized sunglasses act as a polarization beamsplitter: light that’s polarized one way makes it through, while light that’s polarized the other way gets reflected off of them.
Here’s your nice laser-optics version of that:
With a polarization beamsplitter, we can make a photon go one way or the other based on its polarization, so now we can point this beamsplitter at a double slit, making the photon go through one slit if it’s polarized one way, and through the other slit if it’s polarized the other way. But remember — if I don’t force the photon to choose one or the other, it won’t. It’ll be polarized both ways at once, and it will go through both slits. (I know, I know, that’s weird, but just go with it.)
We can zap our crystal with a laser and make it emit two photons at a time. We won’t know their polarizations, but we will know that they’ll always be opposites of each other. If we measure the polarization of one, we’ll know the polarization of the other.
Let’s send one of these two photons through our polarization beamsplitter and double slit. We’ll call that photon Bert. Let’s let the other photon go somewhere else and call that one Ernie.
The first time we try this, we’ll let everything be. No measurements. Our Bert photons go through the beamsplitter and double slit one at a time, and we don’t measure them in any way except marking where they hit the wall. We don’t do anything at all with the Ernie photons, so it’s like they don’t even exist. If we keep doing that, as before, we get lots of stripes — an interference pattern — on that back wall that the Bert photons keep hitting. That’s because we didn’t know which way the Bert photons were polarized when they went through the slits, so they didn’t have to choose where they were until they hit the wall.
But now, what if we DO measure each Ernie photon’s polarization? Then we’ll know which way each Bert photon must have gone, without even measuring it. If we do that, this time we DON’T get an interference pattern. We get two stripes! That’s because we gained information about the Bert photons, thanks to Ernie, even though we didn’t touch the Bert photons in any way. We found out their polarizations, and therefore which way they went, just as we did when we put a detector by each slit.
That’s already pretty freakin’ weird, but now we’ve arrived at the main event. And that takes us to the Canary Islands! (Maybe not what you were expecting?)
If we stand on La Palma, the island at left center on the map, we can look across the Atlantic Ocean along that red line and see Mount Teide over on Tenerife, about 90 miles away:
On a moonless night, we can send a photon all the way over there, and usually its polarization won’t change along the way (that is, it won’t hit anything big enough to throw it off). And we have equipment sensitive enough to detect that photon.
And this sets the stage for the big experiment. We can still put each Bert photon through our little beamsplitter and double slit right here on La Palma, while each counterpart Ernie photon flies all the way over to Tenerife, 90 miles away, which takes about half a millisecond. That means we can’t measure Ernie’s polarization until well after Bert has already hit the back wall and “picked” a spot.
But let’s go one further. We won’t even DECIDE whether we want to measure Ernie’s polarization until AFTER Bert has already hit the wall. We’ll let a computer flip a coin. If it’s heads, we’ll measure Ernie’s polarization, but if it’s tails, we won’t. But the computer won’t flip the coin until Bert’s fate is decided, until he’s already hit the wall and therefore already had to decide where to focus himself.
Let’s do this a few thousand times, and we’ll look at the pattern our Bert photons make when we do measure Ernie and when we don’t. For example, let’s say we measured Ernie photons on tries #2, #5, #6, #8, etc. We can look at Bert photons #2, #5, #6, #8, etc. and see what pattern they end up making. Then, separately, we can look at the other Bert photons (#1, #3, #4, #7, etc.), and see what pattern they make.
But remember — each Bert photon had already hit the wall long before we even decided whether to measure the polarization of its matching Ernie photon or not.
So what happens?
When we look at Bert’s pattern for all the times we did measure Ernie (and therefore found out which way Bert went), we get two big fat stripes. No interference. And that’s because we gained information about them, via Ernie, in the future. When those Bert photons hit the wall, they couldn’t have “known” that we were going to measure Ernie, because we hadn’t even decided yet whether we were going to measure Ernie.
But when we didn’t measure Ernie, Bert gave us an interference pattern.
We would have gotten an interference pattern for ALL the Bert photons if we hadn’t measured any of the Ernie photons, but for some of them we did measure Ernie, and for those, AND ONLY THOSE, the interference pattern changed into a two-stripe pattern. Because of something that happened in the future.
I don’t mean to keep repeating this, but our measurement of Ernie changed what Bert had already done in the past, from the future.
So…
You and I entered this diary knowing darn well what the “past" and the “future” were, but now we don’t anymore.
You’re welcome.
Ananthaswamy has this to say:
The events at Tenerife, in our usual way of thinking, happen later in time, yet still influence the outcome of measurements at La Palma, even though each measurement at La Palma [Bert] is done and dusted well before the partner photon [Ernie] reaches Tenerife.
Language fails us at this point. Here and there, past and future don’t quite work.
I’m really glad Ananthaswamy came along and did us the service of explaining all of this in his book, because the original paper is practically indecipherable to a non-specialist (i.e., me). It’s called “Violation of local realism with freedom of choice.” I love how nonchalant the authors are. They essentially say, “Welp, looks like what we’ve got here is a violation of realism. So, what’s our next experiment?”. That’s awesome.
Watson and Crick pulled this dum-de-dum routine when they discovered that DNA is a double helix, too:
It has not escaped our notice that the specific pairing we have postulated immediately suggests a possible copying mechanism for the genetic material.
After they get the results of their experiment, they (correctly) decide that there are only two reasonable explanations left for what they have just seen, and they casually lay them out as if they’re soft-serve flavors:
1) Events in the future can go back and change outcomes from the past.
2) By some spectacularly nutty string of coincidences, the heads and tails they generated at random weren’t really random because some common past event tied together the generation of the random numbers and the generation of the photons. (Oh, pleeeease.)
As gobsmackingly life-altering as No. 1 sounds, No. 2 is in the neighborhood of Your Grandma Is An Oboe or Reese Witherspoon Controls The Tides. No. 2 just plain did not happen, people.
What would be REALLY great is if we could do this experiment in space, so that Ernie is able to travel for such a long time that we have a chance to see Bert’s outcome on the screen and understand it, then try to change it after the fact by deciding to measure Ernie. Would we then actually see the past change? Could we manipulate the past from the future at will? Hmm…..
These kinds of experiments have generated all sorts of theories, and the main reason for that is that nobody knows what the f^#% is going on. It’s as if we’ve tapped in to some fundamental parts of the Universe’s software code. We’re solidly in “pay no attention to that man behind the curtain” territory.
Some of these theories are:
- Photons don’t “choose” one outcome or the other; a new Universe is born for each possible outcome, so they ALL happen
- There is no such thing as freedom of choice; those “random” numbers were pre-scripted, just like our whole lives
- There is no reality without consciousness; the act of measuring or observing is what makes reality
- There are other dimensions we can’t see; “spooky action at a distance” isn’t really at a distance
- Time isn’t linear from past to future; that’s all illusory (hey, didn’t ancient Mesoamericans come up with that one already?)
It seems to me that the hidden-dimensions thing is kind of obvious at this point. If I live in a 2-D (flat) world, and you 3-D people pass a donut through my world, I see it as two separate circles that are weirdly able to refocus themselves into one object, like a photon somehow does:
There really do seem to be dimensions that are quite real that we just can’t perceive. Can’t say for sure what goes on in them, but it’s fun to think about.
This future-affecting-the-past thing has thrown me for a real loop, though. I don’t know exactly what to make of it, other than the Universe isn’t at all what I thought it was. And as I alluded to at the beginning, what it means to me is that we’re sorely mistaken about many of the assumptions we implicitly make when we ask the Big Questions. We don’t even know how to pose those questions, because our ideas about fundamental quantities appear to be misguided pretty badly.
I don’t know about you, but I’m oddly comforted by that.