I still don't get it

June 3, 2011 4:19 AM

Physicists have managed to observe light behaving both as a particle and wave in the same double-slit experimental condition, by means of a new method to weakly observe a particle's momentum. This article in Nature summarizes the results in non-mathematical terms.

The nature article comments: "Intriguingly, the trajectories closely match those predicted by an unconventional interpretation of quantum mechanics known as pilot-wave theory, in which each particle has a well-defined trajectory that takes it through one slit while the associated wave passes through both slits."

posted by leibniz (49 comments total)

Sadly, the Science article is behind that lovely paywall.
posted by sciencegeek at 4:29 AM on June 3, 2011

This is one of those things where, while the words are full of breathless possibility, I'm not sure I'll ever have time to really really dissect it and see if they're saying "Hey, look, we're closer to the Heisenberg limit than those other guys!" or "Oh, by the way, everything you know about physics is wrong.
posted by Kid Charlemagne at 4:31 AM on June 3, 2011

What Deutsch said: "Experiments are only relevant in science when they are crucial tests between at least two good explanatory theories," Deutsch says. "Here, there was only one, namely that the equations of quantum mechanics really do describe reality."
Pilot wave theory and all the other interpretations of quantum mechanics make the same predictions. It seems odd to say the trajectories closely match pilot-wave theory, as if they didn't they wouldn't closely match any quantum mechanical prediction and someone would be getting a Nobel Prize. I'm really not sure what that bit is supposed to be getting at.

Also what Steinberg himself said - "Steinberg stresses that his group's work does not challenge the uncertainty principle, pointing out that the results could, in principle, be predicted with standard quantum mechanics. But, he says, "it is not necessary to interpret the uncertainty principle as rigidly as we are often taught to do", arguing that other interpretations of quantum mechanics, such as the pilot-wave theory, might "help us to think in new ways"."
It seems to me some people weren't understanding quantum mechanics properly, which is hardly a surprise.
posted by edd at 4:33 AM on June 3, 2011

Also, it kind of blows me away that there are people who don't get the difference between collapsing something's eigenstate by observing it (which really doesn't come up much in how we live our day to day lives) and the fact that making observations has an effect on the system.
posted by Kid Charlemagne at 4:37 AM on June 3, 2011 [2 favorites]

I'm confused. How does that blow you away? Half the physics teachers I run across understand it in those terms. I think you need to recalibrate your common physics knowledge-o-meter.
posted by Shutter at 4:46 AM on June 3, 2011

Can someone explain the pilot wave theory a bit more clearly? What is the guiding wave supposed to be exactly- something that connects together all the photons-as-particles? Does it have its own substance, or what?
posted by leibniz at 4:48 AM on June 3, 2011

My layman understanding is that the guiding wave describes the particles over time.
posted by cx at 5:14 AM on June 3, 2011

Doh - I left out a key part of that sentence. It kind of blows me away that there are people reading Nature who don't get it. If you went to school to teach history and then on year two they decide you get to tech physics, you get a bye. I know someone who this sort of thing happened to.
posted by Kid Charlemagne at 5:19 AM on June 3, 2011

Stephen Hawking covers this particle/wave concept in his book: The Grand Design.
posted by bwg at 5:31 AM on June 3, 2011

mushrooms : youth of amerika :: collapsing eigenstates : Nature readers
posted by ryanrs at 5:33 AM on June 3, 2011 [3 favorites]

Interesting. I got into a debate in a thread a while back about whether or not particles had an "actual" position and momentum that simply couldn't be measured, or whether the position and momentum 'didn't exist' at all.

I had the view that they did exist, but you just couldn't measure them together.
posted by delmoi at 5:43 AM on June 3, 2011

Position and momentum exist, but in most quantum systems are smeared out over a range of probabilistic values. Since this experiment appears to be inferring trajectories based on average values, it's explicitly not a challenge to the principle that sub-atomic particles usually can't be classically localized. (Which is a good thing, because otherwise, most molecules will fall apart.)
posted by KirkJobSluder at 6:10 AM on June 3, 2011

To paraphrase Einstein, "Particles and waves are modes in which we think, not conditions in which we exist."
posted by jfuller at 6:13 AM on June 3, 2011 [9 favorites]

delmoi: The wavefunction can't have a single position and momentum at the same time. The two are (give or take a Planck's constant) Fourier transforms of one another. If you give it a single position (a delta function at some value of position) it ends up with a completely spread momentum, and vice-versa. If you require that it has a fairly narrow Gaussian in one, it'll have a fairly broad Gaussian in the other, and so on. Whether the particle has both is a bit of a different question - it's where the pilot wave interpretation differs. The pilot wave itself does not have a definite single momentum and position at the same time, but the particle is not the wave - it is merely guided by it.
At least that's about as good or bad as my understanding gets. I guess there's a similar relationship between the level of my understanding of quantum mechanics and my sanity.
posted by edd at 6:20 AM on June 3, 2011

A photon walks into a hotel. The bellhop asks, "Where's your luggage?" The photon replies, "I'm traveling light."
posted by exogenous at 6:30 AM on June 3, 2011 [18 favorites]

As far as I can tell, the idea that particles have a discrete momentum and position that's just indeterminate until it interacts with something else has been falsified with chemistry. There's no way you can derive the properties of water from a view that oxygen and hydrogen have little negatively-charged billiard balls flying around. (Actually you can, but the answers you get are empirically false.) Treating electrons as three-dimensional probability fields that hold a partial electron charge, however, does predict real-world values.
posted by KirkJobSluder at 6:51 AM on June 3, 2011 [1 favorite]

KirkJobSluder: I don't think anyone [clueful] would characterise an electron as being a negatively-charged billiard ball flying around. De Broglie-Bohm theories predict the same real-world values - they still use the wavefunction - so if your comment is supposed to argue against them then I'm afraid you've got something of a straw man there.
I'm not exactly a fan of the approach, but it's certainly not easily dismissed compared to the more commonly accepted interpretations.
posted by edd at 7:24 AM on June 3, 2011

The best I can get out of quantum mechanics is that particles really don't exist except insofar as we interact with them (or interact with something that 6 degrees of quantum bacon interacts with the "particles"). Other than that there's just some non-classical causal wave that spreads out "looking" for something to interact with

For my money the delayed choice quantum eraser still wins the prize for best mindfuck quantum experiment
posted by crayz at 7:31 AM on June 3, 2011 [1 favorite]

delmoi: The wavefunction can't have a single position and momentum at the same time. The two are (give or take a Planck's constant) Fourier transforms of one another.

Yeah, I don't really know the mathematical equations for QM, but a normal signal would have both a wave form and a spectrum, even though those are Fourier transforms of each-other.

I'm not saying that electrons or whatever could be mathematically modeled as points with a specific position and momentum that act like classical mass that moves in a straight line until it encounters a force, but rather you could think of it as like something moving on a random walk with some probability function that equals the expected probability distribution. You could say the electron has an average position and an average velocity which you can't measure but does 'exist'
posted by delmoi at 7:31 AM on June 3, 2011

There's no constraint in the math that "particles" like electrons follow continuous paths like planets orbiting a sun. Indeed, tunneling implies that they don't. I prefer to think of electrons as a blob of probability: we know it's in there somewhere, but it could be anywhere in the blob, and have the momentum to match. The blob move continuously, but the electron inside, rattling around like a corn kernal in a baloon, could be anywhere inside.

On the other hand, electrons are indisputably very small in physical dimension, and therefore localizable under certain measurements.
posted by bonehead at 7:36 AM on June 3, 2011

I'm traveling and stuck with a terrible internet connection, so I can't read the Science article. The pop-sci write up makes it sound like someone's taking a perfectly interesting measurement and using it to be all axe-grindy about philosophical interpretations of quantum mechanics that can't be experimentally tested. Science journalists do tend to overhype and are probably misinterpreting what the scientists actually said, but if that's the case, these guys should have known better than to spout off like that. It sounds like they measured the expectation value of the electron trajectory, which is nifty, but not what the article breathlessly reports.

delmoi: In addition to the tunneling example, that interpretation doesn't work because single particles have been observed interfering with their own wavefunction. Hard to do if you're really in one place at all times.
posted by physicsmatt at 7:43 AM on June 3, 2011 [1 favorite]

physicsmatt: I'm probably way out of my depth here but I was kind of imagining, like, a particle 'jumping' around on a random walk, in a way that still satisfied the equations.

It's kind of like in computer science when you talk about a non-deterministic Turing machine, you can imagine it as being in multiple states at once, then branching from each of those states until the set of states contains the halting state, or you can imagine it as making 'random' choices among a group of possibilities, and then analyzing it as if it always made the 'correct' choice to lead to the halting state. Each way of thinking about it results in the same math: If either machine can be proven to be able to reach a halting state in polynomial time, then you have a problem that's in NP.

So I'm thinking, it's possible that the electron is randomly jumping around in space, or does the electron actually exist in all those positions as if it were in a density cloud? Presumably, there would be no way to tell, right? Or is it possible to rule one out? I don't know how a particle jumping around could interfere with itself... unless it were also jumping around in time as well :P.
posted by delmoi at 8:04 AM on June 3, 2011

Basically, for QM to be true (which it is for any sane definition of true, it's got the most evidence behind it of any physical theory ever tested), you have to accept *something* really weird. De Broglie-Bohm only avoids biting the bullet on a lack of a "true" position by positing spooky superluminal action at a distance (Bell's theorem shows that any theory that posits a "true" position will have this drawback). At some point you just have to decide which bizarrely counter-intuitive idea to believe or in the words of Feynman "shut up and calculate".
posted by Proofs and Refutations at 8:08 AM on June 3, 2011 [3 favorites]

The best explanation I ever got was in 1st semester Quantum when my professor quoted the Feynman "shut up" line, followed by this: "The problem we have is that up here at our scale, we have particles and we have waves. Down in reality, it turns out that stuff is made of of these things we've never seen before and have zero intuition for. An electron doesn't CHOOSE to sometimes act like a wave and sometimes like a particle. An electron behaves, 100% of the time, like an electron, and the second you try to describe it in terms of something you're more familiar with, you've gotten it at least a little wrong."
posted by range at 8:15 AM on June 3, 2011 [7 favorites]

Does anybody else feel themselves slowly descending into madness while trying to follow this thread?
posted by thsmchnekllsfascists at 8:17 AM on June 3, 2011 [1 favorite]

delmoi: you're not thinking that far off de Broglie-Bohm interpretations but they're quite non-random (although appear so due to hidden variables, as I understand it). The pilot wave precisely guides the particle without it jumping around, and it's the pilot wave's interference that circumvents the issue physicsmatt raised.
The wavefunction is also not a probability distribution - it's a kind of higher level object from which the distribution can be derived.
posted by edd at 8:23 AM on June 3, 2011

I was involved in an earlier debate on these topics here (strangely enough from Stephen Hawking's comments on theology)... Interesting experiment and interesting if overhyped that thinking in terms of trajectories can help one visualize what's going on here. BUT, pilot wave theory is NOT orthodox quantum mechanics. I think it may run into difficulties when people try to generalize if from quantum mechanics (theory of single quantum particles) to quantum field theory (theory that involves the creation and annihilation of quantum particles). There are a wide class of mathematically equivalent ways to view QM, but not all of them mesh with the extensions to QM.
posted by Schmucko at 8:29 AM on June 3, 2011

delmoi: what do you mean by exists? what do you mean by "electron size" or any of the macroscopic concepts we're forcing onto the real quantum world to make our conception of the non-classical reality easier to talk about? I can imagine an electron as a cloud moving in response to the potential: density of the cloud indicates probability density, but what I really want is the wavefunction, which is complex valued and so hard to visualize with a single variable like density. Ignoring this means missing the phase, which is critically important for understanding quantum mechanics. These mental pictures are important to gain some intuition, and it seems that picturing the electron as an "object" that moves around in a path like we're used to tends to miss a lot of important effects that make the world work at the quantum level. You can do it, but you then have to keep a bunch of caveats in mind in order to not miss the fun stuff. Much easier to just to shed the classical picture, in my opinion. But as has been said already in the thread, in the end, we're going to be calculating. So while intuition is important, I better trust my math more than my picture.

thsmchnellsfascists: serious question, why is that? I really would love to know how to make these sorts of things clearer to non-physicists, but part of the problem is that I've been reading about and working with the crazy quantum world for so long that it seems totally normal at this point. So what can we as the experts do in order to communicate the fun stuff more effectively?
posted by physicsmatt at 8:39 AM on June 3, 2011

Ugh, I'm currently way out of sync with US time and going to crash. I'd like to continue the discussion, but being pi out of phase with most people in this conversation isn't going to help. I'll join in again in a few hours.
posted by physicsmatt at 8:49 AM on June 3, 2011

Ugh, I'm currently way out of sync with US time and going to crash. I'd like to continue the discussion, but being pi out of phase with most people in this conversation isn't going to help. I'll join in again in a few hours.

No problem. I like following long threads.

thsmchnellsfascists: serious question, why is that? I really would love to know how to make these sorts of things clearer to non-physicists, but part of the problem is that I've been reading about and working with the crazy quantum world for so long that it seems totally normal at this point. So what can we as the experts do in order to communicate the fun stuff more effectively?

Most of the fun parts of physics aren't covered very well in American public schools, and we don't get enough of the math education to start to see what's going on. Couple that with a 10 year old textbook and a bored teacher, making it difficult to get into this stuff as a 17 y/o kid.

A few years back, mefi inspired me to buckle down and try and learn some of this on my own. I started to grok some of the more intuitive things, like treating electrons as a probability cloud instead of discrete objects and the uncertainty principle, but I never was able to make any headway with the math. Without being able to do the calculations, I was stuck. I moved on and have forgotten most of what I learned back then. I remember just enough for this discussion to hurt my head.
posted by thsmchnekllsfascists at 9:12 AM on June 3, 2011

physicsmatt: " So while intuition is important, I better trust my math more than my picture."

Somewhat Off Topic: What about Feynman path integrals? I was under the impression that they're not well defined mathematically. On the other hand they produce useful results rather than gibberish, so they're clearly doing something.

It feels a bit like Newton's (& Leibnitz's) calculus, it gives really good answers despite being in its original formation suspiciously like dividing by zero. Or has some modern day Cauchy found a way to define them rigorously?
posted by Proofs and Refutations at 9:17 AM on June 3, 2011

The same experiment applied to chickens.
posted by brundlefly at 9:22 AM on June 3, 2011 [1 favorite]

Q: How many quantum physicists does it take to change a lightbulb ?
A: One: of course. Two to do it, and -1 to renormalise the wave-function.
posted by BlueHorse at 9:23 AM on June 3, 2011 [9 favorites]

Feynman path integrals are fine in regular quantum mechanics. It's an alternate way of calculating the evolution of the state vector (wave function). For each path you weight by a complex phase that depends on the integral of the action over that path. Intuitively you recover classical mechanics when the phases of the paths far from the one that extremizes the action cancel out.

However, when you get to quantum field theory and path integrals that require renormalization, the mathematical rigor seems to be a problem. Basically you integrate to infinity, but then you step back and realize that the properties of the particle you've put in to your equations are not the "bare" properties. An electron is always measured with a cloud of virtual particles around it. By adjusting the bare quantities you can get answers out of the integral.
posted by Schmucko at 9:40 AM on June 3, 2011

On calculus - nonstandard analysis lets you use infinitesimals and remove them when you're done. I still prefer limits though - as do many others.
posted by edd at 9:42 AM on June 3, 2011

Yeah, I don't really know the mathematical equations for QM, but a normal signal would have both a wave form and a spectrum, even though those are Fourier transforms of each-other.

Yes, but both the waveform and the spectrum are "spread out". You can't give a precise instant at which a tone sounds; it occurs over some period of time, just as a particle's wavefunction occupies some amount of space. And you can't give a single frequency for it, either; it occupies some range of frequencies, just as a particle's momentum does. If you make the sound shorter and shorter (more localized in time) then its spectrum will broaden, fuzz out, spatter— eventually, an ideal zero-duration impulse has equal energy at all frequencies. Conversely, if you want a narrower-bandwidth spectrum, you need to stretch out the sound, making it not vary as much over time; in order to get a perfectly pure tone, you need to make the sound arbitrarily long.

tl;dr: Position and momentum are fourier-transforms of each other; studying QM can actually illuminate the nature of classical waves like music or radio communications.
posted by hattifattener at 11:00 AM on June 3, 2011

jfuller: "To paraphrase Einstein, "Particles and waves are modes in which we think, not conditions in which we exist.""

That's one of the things I thought while I was on shrooms. I was indeed an Amerikan Youth at the time.

(paraphrase? Can you show me the source?)
posted by symbioid at 11:15 AM on June 3, 2011

Also - hasn't this thread kinda completely derailed from the original post? Or am I missing something?
posted by symbioid at 12:57 PM on June 3, 2011

symbioid, the paraphrase was the substitution of the terms "particles and waves" for "space and time".
posted by Schmucko at 3:14 PM on June 3, 2011 [1 favorite]

My ancestry includes a long line of apes who learned that throwing rocks could bring down a bird for lunch: no selective pressure led any of them to contemplate particles as fuzzy probability bundles. Thus I found the "shut up and calculate" quote apt and amusing.

But Wikipedia suggests that it's a line better attributed to David Mermin than to Richard Feynman.
posted by fredludd at 4:47 PM on June 3, 2011

symbiod, I thought the original posts were about a claim of experimental evidence for a particular interpretation of quantum mechanics that's somewhat out of the mainstream. Even the other people interviewed by the reporter disagreed with that claim, and I see nothing in the result that's evidence for anything other than the fact that we can measure expected values of probability distributions (averages).

Thinking about this topic has been interesting for me. I find the Copenhagen-Many Worlds-etc arguments about quantum mechanics to be pretty dull at this stage, so it tends to annoy me, but that's mostly due to repetition (and the fact that no one has made any progress on the problem in forever). I tend to fall back onto the "just calculate" attitude, because: hey it works. But trying to write out this post made me remember that the rules of quantum mechanics can seem pretty arbitrary (not talking about the wave-particle duality, but the assumptions about of Hilbert spaces, etc that lead to the wacky QM results). QM is fantastically successful, but it's not like the rules were handed down from on high: we figured them out after some trial and error. So it's not that people are looking for hidden variables that bothers me so much, it's that they disassemble about the result when the evidence isn't conclusive. So my initial feeling of "comeon guys, this isn't even science" was wrong, it is science, it's just that I don't think they found what they claimed.

I did get off topic when asking how we can do a better job explaining quantum weirdness. That should probably be somewhere else. Thanks for the response thsmchnellsfascists.
posted by physicsmatt at 5:12 PM on June 3, 2011

What I still don't get is people's ideas on quantum physics. Take a look at one of the comments:

I'm not a fan of the idea that an observer changes the way a particle behaves. To me it is like explaining that the moon does not exist unless someone is looking at it. There has to be another explanation...and maybe we already have one, but I haven't heard of it yet.

Surely that's like saying that you don't accept quantum tunneling because it's like explaining that a person can walk through a brick wall as if it didn't exist, despite the fact that we already have machines that rely on this principle in order to work, no walking though walls or disappearing moons required.
posted by Soupisgoodfood at 8:57 PM on June 3, 2011 [2 favorites]

Put a ball inside a cardboard box, close the lid, and set the box down. Once the ball stops moving, you'll know its velocity, but not where it is. If you move the box, you can sort of hear where the ball is for an instant, but you don't know its velocity very well. And it keeps moving and isn't where it was any more. You can guess what the ball's -probably- made of, and a -probable- range of sizes for it.

It's like atom-sized particles are in a much, much smaller box - permanently. Each kind of particle's in a different box. If you can get them to stick to a surface, you can know their position very accurately, but lose the ability to know their mass.

We can't ever really be sure if it's the particles that are so goofy, or if ... because our familiar classical 'yardsticks' for measuring properties are so clumsy ... it's the clumsiness that makes the results so goofy. It just doesn't help to ask what's really going on, because it's impossible to be certain 'in the old way'. The 'uncertainty principle' was an expression of resignation!

When I started to understand that stuff, I really started to appreciate how elegant some of the experiments were back then ... Rutherford shooting alphas at gold foil, Millikan measuring the electron charge.
posted by Twang at 9:02 PM on June 3, 2011

I always loved Millikan's experiment. Though I figure it's only the best experiments that we hear about in classes a hundred years later.
posted by hattifattener at 11:39 PM on June 3, 2011

Twang: You can use classical mechanics to calculate where the ball is in the box at all times
posted by delmoi at 3:42 AM on June 4, 2011

Soupisgoodfood: I'm not sure that's what the commenter is getting at. There's definitely something uncomfortable about having an observer play such an important role without any really good definition of what an observer is (see Wigner's Friend). It's the motivation for objective collapse theories and so on. It doesn't mean the commenter has a problem with the predictions of quantum mechanics. It just means they are uncomfortable with some interpretations of how wavefunction collapse comes about.
posted by edd at 3:54 AM on June 4, 2011

Isn't the scientist doing the experiment the observer? To me it seemed like a baseless comparison, partly because it assumed an interpretation. I've only read a few books on quantum physics, not studied physics or anything. It just seems that when it comes to quantum physics, a lot of people still try to look at it through an out-dated perspective and call it weird or strange.
posted by Soupisgoodfood at 4:15 AM on June 4, 2011

> That's one of the things I thought while I was on shrooms. I was indeed an Amerikan Youth at the time.
>
> (paraphrase? Can you show me the source?)
> posted by symbioid at 2:15 PM on June 3 [+] [!]

Schmucko called out the substitution correctly. (The original substitution that occurred to me, I might as well admit, was "Live and dead cats." I may have been smoking something at the time.)
posted by jfuller at 6:43 AM on June 4, 2011

Here is a scienceblogs entry on the thing. It looks like it has better detail than that bullshit popsci article.
posted by symbioid at 9:58 AM on June 6, 2011