Log in

No account? Create an account
10 July 2014 @ 10:31 am
Quantum questions  
This article provides an explanation I haven't seen before for the difficult-to-wrap-one's-head-around concept that in quantum mechanics it's "observation" that causes the quantum wave form to collapse.

"...if you take a picture of an electron, its probability cloud evaporates and leaves it at one exact place. The light that bounced off the electron to hit your camera forced the electron to appear! The resolution to this troubling idea is that if you leave the light off, no photons hit the electron. The watching camera sees nothing, the electron remains ethereal. The electron will still be forced to resolve in a lit place while watched by a camera with its lens cap on." (emphasis added)

Or, to put it another way, it's not "observation" but "interaction" that collapses the wave form. If a particle isn't interacting with any other particle, it might be anywhere. Only when two particles bounce off each other (or otherwise interact) do their positions and states become clear.

I find this explanation compelling. It makes the whole concept much more comprehensible to me. But it also implies that the whole concept of quantum uncertainty is an entirely theoretical mathematical abstraction, because in reality every particle is interacting with other particles nearly constantly. Even a single hydrogen atom floating in the near-vacuum of interplanetary space is struck by photons from the sun and interacts gravitationally with the planets.

If my understanding is correct, quantum uncertainty never really happens in real life -- all particles are in a constant state of waveform collapse. This makes quantum uncertainty as relevant to the real world as those massless, frictionless ropes we use in physics problems. (By which I mean that it is a useful theory with great predictive power, but doesn't actually describe anything that exists in the real world.)

If you are familiar with quantum theory, does this explanation and its implications match your understanding? If not, can you help me to understand where it differs?

Also, if you are familiar with the history of quantum theory, do you know why the term "observation" was used rather than "interaction"? Because if my understanding of this explanation is correct, the term "interaction" would be a much clearer way of explaining what's happening (and would have avoided a lot of the meaningless woo-woo that's attached itself to the term "quantum physics").
Deldel_c on July 10th, 2014 06:11 pm (UTC)
I r not a physicist, but I think the electron will not be forced to resolve, from where you are, if you don't interact with the photons that are interacting with the electron. Even in a lit room.

This is the basis for the "turtles all the way down" explanations I've seen for the nested Wigner's Friend experiment. Wigner's Friend is where they sweep away the objection that a cat is not a proper observer by saying "Fine, replace the cat with a human, and for ethics' sake, replace killing with something less traumatic." Now, the man in the box clearly experiences a wave function collapse, but we outside the box see a superposition, until we open the box. You can nest this situation, by putting the lab where the box-opening happens in a box: now the box-openers experience a wave-function collapse, but we outside still see a superposition of states.

It's not about whether the electron interacts with a photon, it's whether the electron or any of those photons interact with the system we are talking about, the one that is "observing" a superposition.

I went looking for a link to a competent explanation of the stuff I just wrote there off the top of my head, but for some reason Google insists "multiple" is a perfectly good substitute for "nested", making a proper search beyond my google-fu. This link to a book by a populariser called Colin Bruce insists that such considerations make "many worlds" an unavoidable interpretation, but I think the non-many-worlds people disagree.
Matthew S. Rotundo: Radioactivematthewsrotundo on July 10th, 2014 06:48 pm (UTC)
Also not a physicist, but doesn't the double slit experiment prove you wrong?

The photons in the experiment are traveling through a gravitational field, yet they pass through both slits at once without collapsing the wavefunction, which creates an interference pattern on the opposite wall.

Now, set up a photon detector at one of the slits, and the interference pattern disappears. The observation from the photon detector collapses the wavefuction.

So that seems like a simple example of quantum uncertainty happening in real life. I'm sure there are others.
Dr Plokta: pic#104550077drplokta on July 10th, 2014 08:00 pm (UTC)
My understanding is that the wave function collapses for you when the particle interacts with you. That's why it's called observation and not interaction. Quantum systems of multiple particles interacting with each other but not with you can remain uncollapsed. It's not that the light bounced off your camera that made the wave function collapse, it's that you later looked at your camera and thus interacted with the original particle.

Or in many-words terminology, which is clearer and easier to understand, the wave function never collapses at all, but when it interacts with you, you split into multiple versions, one for each possible outcome.
et in Arcadia egoboo: Sacred Chaoapostle_of_eris on July 10th, 2014 08:37 pm (UTC)
Not only is quantum uncertainty "real", but it's actually possible to put it back after it's gone.
see "quantum eraser"

(I put "real" in quotes because conventional quantum physicists don't use words like that.)
They Didn't Ask Me: maxwells-equationsdr_phil_physics on July 11th, 2014 01:49 am (UTC)
Okay, physicist here. Part of the confusion is based on the very human desire to know what's going on. In this case, what is that darn electron doing? But...why that one? There are a LOT of electrons. There are a LOT of atoms. For every gram of matter, there's an Avogadro's number of nucleons, protons or neutrons. For every gram of protons, there's usually an Avogadro's number of electrons. 6.02 x 10^23, or 602,000,000,000,000,000,000,000 . We cannot possibly locate all the locations of all the electrons in a drop of water or a grain of table salt. The best we can do is determine average behavior, to model and calculate the aggregate.

Look closely at one, and you're ignoring a gazillion others.

For calculation purposes, we don't know what electron A is doing. So the waveform gives the possibilities and that appears to be what electron A is doing. But electron A should be somewhere, and any interaction, say with a photon, changes the situation. The wavefunction collapses.

The double slit experiment looks at what a lot of particles are doing. Shoot just one and a detector can tell where just one went. Repeated measurements show different results. Shoot a lot at once and each electron sees the others -- which affects the results. But you still can't say that every time the 3rd electron is going to take the right rather than the left slit.

Coolest version of this experiment is one I heard at an American Physical Society March Meeting -- probably the largest gathering of physicists each year -- in the early 90s. Using the same lithography method for making the tiny bits in like a Pentium computer chip, they made a tiny double slit and shot hydrogen molecules at them. The patterns of the spread at the target was the classic double slit interference result. They even shot individual helium atoms, what you would think was matter and therefore "solid" and they too made the interference pattern. The wavefunction of the helium atom went through both slits and interfered with itself. Just as a physicist might expect but it was so cool that someone did it.

Two additions to this LONG comment.

-- In the land of ordinary macro objects, none of this applies. Run a Buick into a double slit, and assuming the slits are narrower than the car, you can imagine that getting an interference pattern is not going to happen.

-- Heisenberg's Uncertainty Principle says that you cannot know the position and speed simultaneously at the same high precision. Or energy and time. One or the other. Again, does not apply to Buicks -- we can determine its speed and position at the same time.

Sorry, I teach this stuff to non-majors and thought I'd take a crack at throwing some thoughts out. Hopefully the Kindle Fire miscorrect hasn't ruined my meanings. (grin)

Dr. Phil

Edited at 2014-07-11 01:50 am (UTC)
Dr Plokta: pic#104550077drplokta on July 11th, 2014 06:22 am (UTC)
What you say about the double slit experiment is not true. Even if you send the electrons through the slits one at a time, so that there are no other electrons for them to interact with, you still get the interference pattern.

Edited at 2014-07-11 06:22 am (UTC)
They Didn't Ask Me: smirking-winsletdr_phil_physics on July 11th, 2014 12:17 pm (UTC)
A detector will record an individual event. The pattern developes over multiple events. But those individual events will occur at a particular site with a frequency consistent with the interference pattern. Send only one particle ever and you cannot make a conclusion about the apparent trajectory. It is only through multiple observations that one can determine the pattern.

If you have two wide slits in armored steel in front of a wall and fire a machine gun, the bullets will either go through a slit or be stopped by the steel. Spray the armor with machine gun fire and the bullet patterns in the wall will roughly match the shape of the slits. A shadow pattern. Even if you secure the gun you will get some spread in the shot pattern.

Make the slits too small and the bullets will break up upon impact. That's the classical macro world expectation. That you send particles through small slits and get a non-shadow pattern is counterintuitive.

Dr. Phil

Edited at 2014-07-11 12:18 pm (UTC)
paulshandy on July 12th, 2014 11:44 am (UTC)
Do you enjoy signing off as Dr. Phil? I keep smiling about Oprah's Dr. Phil
They Didn't Ask Me: dr-phil-confusion-2012dr_phil_physics on July 12th, 2014 02:01 pm (UTC)
I was anointed as Dr. Phil in 1989 at Michigan Tech by one of my mentors, Dr. Bob. When The Other Dr. Phil announced he was starting a TV show, I just knew it would be trouble. I do get email, phone calls, letters -- they find the info on my work homepage, right next to the Big type that says I am not the Dr. Phil on television.

Even more amusing, George RR Martin looks like me.

Anyway, I'm the Dr. Phil who has hair and a Ph.D. in Applied Physics.

Dr. Phil
Dr Plokta: pic#104550077drplokta on July 11th, 2014 06:24 am (UTC)
Also, Heisenberg's uncertainty principle does apply to Buicks in principle, it's just that the numbers involved are so small that it's not noticeable.
They Didn't Ask Me: smirking-winsletdr_phil_physics on July 11th, 2014 12:05 pm (UTC)
Absolutely. The equations give delta-x and delta-p values too small to measure. My point is better expressed that Heisenberg is not needed for a Buick. That one of the reasons we have such trouble with things like quantum mechanics and relativity is that we don't have to process such things in our brains in the ordinary macro world.

Dr. Phil
They Didn't Ask Me: smirking-winsletdr_phil_physics on July 11th, 2014 12:26 pm (UTC)
The Correspondence Principle allows that at some level of precision, the classical and modern physics results converge and we don't need to go beyond the classical equations. For special relativity, for example, at the 10% error level, that happens at speeds below about 0.42c. I think I recall that at the 1% level, it's 0.17c.

So for eyeball level estimates, where you cannot tell if a stick is 1 yard or 1 meter long without references -- the 10% level -- you don't really need relativity until about 42% the speed of light. But GPS calculations require both special and general relativity for accuracy.

Dr. Phil
eub on July 11th, 2014 09:13 am (UTC)
By which I mean that it is a useful theory with great predictive power, but doesn't actually describe anything that exists in the real world.

Not sure exactly how you place "predictive power" in opposition to "doesn't actually describe"? QM makes predictions that frequently differ from what non-quantum models make (and QM is right), and it does this by accurately describing e.g. superposition of states in an electron. (Or in a buckyball; the double slit has been run with those.) On the other hand, for lots of systems the predictions do not differ more than infinitesimally.

On terminology: I am no physicist, but as a computer scientist I hear explanations of the quantum computing model in terms of measurements and information, e.g. the Zeilinger quote here:
[T]he superposition of amplitudes ... is only valid if there is no way to know, even in principle, which path the particle took. It is important to realize that this does not imply that an observer actually takes note of what happens. It is sufficient to destroy the interference pattern, if the path information is accessible in principle from the experiment or even if it is dispersed in the environment and beyond any technical possibility to be recovered, but in principle still ‘‘out there.’’

On woo: I am no historian, but it sounds from the writings of some of the early players that they basically imported their philosophical positions into the interpretation of QM. They didn't have a compelling alternative, and some people were fond of the observer business, and near-solipsism is always hard to falsify... so fifty years later people think the physics is talking to them about the power of consciousness.

Nowadays, just within the last ~25 years (man the 90s are getting long ago) people can more or less see quantum decoherence happen, experimentally. Decoherence is not identical with the idea of 'wavefunction collapse', but it's close enough for a programmer's purposes. :)

Edited at 2014-07-11 09:15 am (UTC)
Deldel_c on July 20th, 2014 12:48 pm (UTC)
By coincidence, just today Ethan Siegel's blog tackled this question:

Ask Ethan #46: What is a Quantum Observation?
The observer changes everything, but what does that mean?

...and gives a surprising answer I haven't heard before, but which makes sense now I've heard it. It turns out to be an interaction, but not just any interaction. It must be one with a certain necessary property.
eub on July 24th, 2014 06:45 am (UTC)
I'm glad to see this answer that it's not the Power of Consciousness, but, hm, didn't they end up pretty circular on what precisely that necessary property is?

"It has [...] everything to do with whether you [...] constrain the particle into one particular quantum state or another."

Is that saying "superpositions are collapsed by measurements that collapse superpositions"?

(As a computer scientist what I hear people saying is that superposition is collapsed by measurements that transfer sufficient information out of the particle to determine the particle's quantum state of the particle. But... is that actually a non-circular statement itself? Maybe it is? I think the definition of "information about X" has to bring in some notion of equivalence classes of microconfigurations, and I've never understood statistical thermodynamics at that foundational level either.)
Deldel_c on July 24th, 2014 05:17 pm (UTC)
I think this explanation is similar to the one where they explain Heisenberg's Uncertainty Principle by how a short-wavelength (hence high momentum) photon can pin down the position of an object, at the cost of sending it careening off with unknown momentum. But I've heard physicists say that's only superficially intuitive, a simplistic analogy for classically-minded readers.

Here the writer is saying "For an electron passing through a slit, that means forcing an interaction with a photon that can constrain its position well-enough to be definitively through one slit." But if you use enough photon to narrow the electron down to the slit, you've prevented the electron from possibly passing through the other! If you use so little photon that the electron could have gone through either slit, you aren't using enough to confirm the electron went through the slit. And you can't fix it with slit spacing or anything else; the minimum is like the minimum momentum and position in the Heisenberg thought experiment.

What I'd like to ask a physicist is, can you tune the experiment so that it produces a pattern that is mostly two peaks (particle), but has a little of the wave character (some slight interference)? I'm guessing you can.
eub on July 27th, 2014 08:05 am (UTC)
I'd also love to learn the answer to what happens if you aren't gaining 100% certainty which slit the electron went through. What if you know only with 99% probability, or with 51%? Somewhere along the 50% - 100% range, something must happen!

(To get a "two/one peak mix" pattern by brute force you could have the apparatus just rapidly randomize its behavior whether to take the measurement or not, but that's probably not what you were interested in.)
Deldel_c on July 27th, 2014 08:50 am (UTC)
That's what I was getting at, I expect something other than a snap from wave to particle behaviour.
paulshandy on July 12th, 2014 11:45 am (UTC)
I am going out on a limb and suggesting it is called "observation principle" because it was discovered during an experiment, which is by definition a specialized form of observation.
mcjuliemcjulie on July 12th, 2014 04:05 pm (UTC)
I know quantum uncertainty is usually not assumed to apply in any meaningful way to macro objects, but I believe that spiders have it.

If you see one, and you take your eyes off it (to get a wad of tissue paper or whatever), when you look back to where it was, it's frequently not there anymore. But if you keep staring at it and make your husband get the tissue, it stays put.

Obviously, the instant you stop staring, that spider can be anywhere in the universe. If we can just figure out how they do it, we will have mastered quantum teleportation.
David D. Levinedavidlevine on July 12th, 2014 05:16 pm (UTC)
Thank you to everyone for your comments! This is still extremely counterintuitive and weird, but at least I now understand it a little bit better.