8 ms·
I read Feynman's QED back in high school and was fascinated by it, but always had a lingering question. Sure the math works out by summing all possible paths,
by crazygringo 4y ago
I read Feynman's QED back in high school and was fascinated by it, but always had a lingering question.
Sure the math works out by summing all possible paths, and then most of them usually cancel each other out, and so you're left with the ones that mostly contribute to the actual probabilities, as complicated as those can still be.
But I always suspected... surely that just has to be one mathematical way of looking at it. That at some point we'd discover that turns out to be mathematically identical to something else that would be a bit more local, a bit more intuitive, a bit more straightforward. That never even had to deal in the first place with the infinity of paths that all wind up cancelling each other out anyways.
So now that I'm on a forum with smart people, the kind I didn't have access to back in high school... has there been any progress towards something like this at all? Are there any directions for this? Or is there a conceptual reason I'm missing why this fundamentally isn't possible, why the sum of all possible paths is the only way to account for certain behaviors?
- whatshisface 4y agoYou are never going to get the locality you're hoping for, because of Bell's theorem.
- elteto 4y agoI think the current state is very much “the formulas work, don’t get too philosophical with them though”. IANAP.
- andreareina 4y agoThat's more about Copenaghen vs Many Worlds vs Pilot Wave... The infinite path integral is heavy mathematical machinery so it's avoided if possible.
- evanb 4y agoThe path integral is used for computational approaches all the time.
- zeroonetwothree 4y agoPhysics is a model of the way the universe works. There's no reason to be sure that it does work the way that we model it. Practically speaking though it doesn't really matter as long as the model gives us accurate predictions. At that point you are entering the realm of philosophy.
- inciampati 4y agoA model can become a cage. Although the QM "infinite trajectories" interpretation is fine at explaining what happens in a quantum system from the perspective of an observer. But it doesn't provide any intuition about how the systems really behave. There may be ways to get access to more realistic interpretations. If we had them they might help us work with quantum systems.
- WastingMyTime89 4y agoWe, humans, are limited in our understanding of the worlds by what we can perceive and conceive. What does it means to understand how a system really behave when we are so limited? That’s an interesting metaphysical question but we have left the realm of sciences by asking it. Sciences care about what can be experienced. The aim of sciences is to accurately predict the results of experiments. A model doesn’t have to be simple or intuitive, just accurate. Obviously everyone would like a simpler model predicting the same things but not because it would be more real, just because simpler is nicer.
- qqtt 4y agoFeynman's QED talks pretty extensively about this as well. An intuitive mental model that humans can picture in their heads is a nice to have, not a requirement. Understanding how a system "really works" is also not really a goal of physics - the science is about predicting physical phenomena - and we generally choose whatever is the simplest equation we have currently available that does that. If you want to define the bar for explaining how a "system really works" as being able to give you an intuitive human-centric mental model you can fully understand and capture in your head - well, you are moving the goal posts over what physics really is and quantum physics has already left you far behind.
- rthomas6 4y agoPilot wave theory Note: I don't have a good grasp of quantum physics
- inciampati 4y agoYes, this always bothered me as well. It happens to annoy some physicists too, so there has been a resurgence of interest in alternative (less magical?) interpretations of quantum mechanics in the past decade. Perhaps the most striking is experimental evidence that particles follow "surreal" trajectories: https://doi.org/10.1126/sciadv.1501466 https://doi.org/10.1126/sciadv.1501466 There are other very interesting, and intuitive, features of this line of modeling. I think it's better to have a realistic mental model of the microworld even if the overall QM predictions aren't (obviously, yet) affected. Maybe there's more here than immediately obvious to the mainstream physics community.
- saboot 4y agoIn the past 10 or so years, a combinatoric approach to particle interactions has shown surprisingly good results. Simple algebraic calculations have been able to give answers that requires hundreds of pages of math through the feynman diagram approach. This is the amplituhedron, led by Nima Arkani-Hamed, here's an article also by quanta magazine about it https://www.quantamagazine.org/physicists-discover-geometry-underlying-particle-physics-20130917/ https://www.quantamagazine.org/physicists-discover-geometry-... In addition to giving exact simple calculations to particle scattering amplitudes, the most important aspect is that it does so in a way that does not require notions of space or time. Leading to the notion that there are more fundamental aspects to reality and space and time may be emergent phenomena.
- Rattled 4y agoThis is the third formulation I've come across that sounds like this with space and time emerging from particle interactions. The other two are from Stephen Wolfram and Carlo Rovelli. Are the formulations related do you know? I've struggled to get beyond their popular science descriptions to the actual theories, I find descriptions like the amplituhedron being 'a multidimensional jewel' distracting, if you have any suggestions where to start for a deeper understanding of this one I'd be interested.
- DebtDeflation 4y agoSpacetime is emergent in both string theory and loop quantum gravity. Article contrasting how it works in each: https://www.scientificamerican.com/article/what-is-spacetime-really-made-of/ https://www.scientificamerican.com/article/what-is-spacetime...
- danbruc 4y agoYou can find quite a few lectures by Nima Arkani-Hamed on YouTube including some on the amplituhedron. But they are mostly not gentle introductions and require some knowledge of physics to follow along. I think this one [1] provides a lot of background and explains the motivation. [1] https://www.youtube.com/watch?v=mF_vqDt-ETU https://www.youtube.com/watch?v=mF_vqDt-ETU
- lmm 4y agoThe only progress I would say is more and more people realising it's a waste of time. Why should we expect reality on a scale very different to everyday experience to behave like everyday experience? (And frankly I've never found it particularly counterintuitive. It's just a wave bro. It behaves like any other wave)
- nomel 4y agoFlash storage uses QM [1]. Who knows what other advancements could be made with a better grasp of reality, at these scales. It doesn’t seem like a waste of time. 1. https://www.linkedin.com/pulse/contributions-quantum-mechanics-flash-memory-anurag-john https://www.linkedin.com/pulse/contributions-quantum-mechani...
- lmm 4y agoGrasping reality is better done by grappling with it directly than by trying to force it into your everyday intuitions. I'd bet the people who made flash work were many-worldsers.
- wyager 4y agoThe path integral formulation is possibly the most local theory you could come up with - it's incredibly simple and does not require any nonlocal effects! The relativistic lagrangian is even more elegant than the classical one - it's just mass! What kind of theory would be more intuitive and straightforward?
- jiggawatts 4y agoThis! 99% of modern physics was developed for pen & paper mathematics, not numeric simulations. Even when computers became readily available it was seen as somehow "less than" symbolic solutions. More practically, special-case solutions that were tractable symbolically require far too much computer power if approached from a fully general "local" simulation perspective. QED is essentially a Monte Carlo simulation, just like ray tracing in computer graphics. The difference is that QED will correctly model diffraction (and other more complex physical effects) that the simplified graphics-only model doesn't bother with. Both QED and raytraced computer graphics rely on a mathematical dualism where a wavefront can be tracked either as a surface OR as a vector perpendicular to the wavefront. The whole "summing up but mostly cancelling" is just accounting for the wavefronts. The use of complex numbers is mostly also just a method for tracking the cyclic nature of wavefronts. If you skip the "waves" part, it becomes traditional computer graphics. Less accurate, but essentially the same. Computer graphics could also simulate light transport by using a 3D volume, which is roughly speaking how some diffuse light simulations work in real-time games like the Unreal engine. The problem is that to get decent accuracy, you need a lot of memory to store that 3D volume. For low-resolution scene lighting this is fine. Monte Carlo enables a solution to be built up incrementally with much less memory required, needing only a 2D accumulator in the image plane instead of a 3D simulation in object space. This allows a vastly more complex and detailed simulation to be calculated, which is important for the tiny wavelength of physical light instead of some non-diffracting idealisation of it. For a physics problem the memory requirements of any non-trivial system would be gigantic, that's why everyone just ignores this approach even though it makes a lot more intuitive sense.
- consilient 4y ago> that's why everyone just ignores this approach even though it makes a lot more intuitive sense. - No one ignores the correspondence between QFTs and stochastic processes (which is what I assume you're talking about, since the claim that QED is just Monte Carlo is extremely false if taken literally), any halfway decent intro QFT class will cover it. - It only "makes more intuitive sense" because you've papered over all the gory details. There's very little intuitive or natural about analytic continuation in C^{\infty}, for instance.
- pithon 4y agoYou can take an amplitude-vs-time representation of a signal and via a Fourier Transform instead represent the same information as a sum of weighted complex exponentials in the frequency domain. It works mathematically, sure- but does that mean that- physically, at the core of existence- every RF, acoustic, seismic, or financial data ripple is actually a bunch of sines and cosines which are getting summed together to create the real phenomena?
- dsego 4y agoI've read that we can typically reconstruct any complex wave form from other periodic waves. So it doesn't seem like sines are something special. But apparently, mass on a spring behaves in a sinusoidal motion https://youtu.be/n2y7n6jw5d0?t=968 https://youtu.be/n2y7n6jw5d0?t=968 And sines are also an approximation of harmonic motion of a pendulum. https://youtu.be/p_di4Zn4wz4?t=370 https://youtu.be/p_di4Zn4wz4?t=370 Many physical systems resonate or oscillate produce quasi-sinusoidal motion. https://ccrma.stanford.edu/~jos/mdft/Why_Sinusoids_Important.html https://ccrma.stanford.edu/~jos/mdft/Why_Sinusoids_Important...
- cjfd 4y agoThe summing over all paths is basically already there when you do the two slit experiment. In that case you sum over two possible paths. So it is already right there in the basis of quantum mechanics.
- gpsx 4y agoWhen you look at quantum mechanics from Schroedinger's equation point of view, it is very simple. The implications of it are not, which is true of many simple systems. What is difficult in quantum mechanics is trying to understand how those equations have anything to do with what we see. The path integral interpretation actually does give us an intuitive picture of why we see classical physics. If you zoom out, the stationary point of the Lagrangian is all that survives, with other things mostly canceling out as you mentioned. This explains why in lagrangian mechanics of clasical physics we are told, without reason, that the equations of motion are the stationary point of the Lagragian. You don't say this, but I think people think quantum mechanics is weird just because it is different from what they see. People also had trouble with the idea of us moving around the sun, one reason being because it looked like everything was going around us. Of course, if we did go around the sun, it would still look like things went around us. In the end though neither is really true. The sun in the center is just a better model. Likewise, quantum mechanics is probably just a better model to "reality" than classical physics.
- sfmike 4y agoWelcome to philosophy
- froeb 4y agoIt isn't the only way. I would argue that the "sum over paths" is more of a convenient language for categorizing QFTs than an actual statement of physical reality. Path integrals have the benefit that they can be easily connected to correlation functions by a Wick rotation, are naturally invariant under relativity, and work (kinda) well with gauge theories (well better than any other way we know). However, in practice they are not used directly. Instead, the path integral is either treated perturbatively giving you Feynman diagrams, or by working out the Hamiltonian prescription (probably closest to what you are thinking of), or by using Monte Carlo approaches (there are probably other approaches I don't know of too). All of these approaches are easier to calculate with, but are a less natural language for abstractly describing a QFT.
- evanb 4y ago"Monte Carlo approaches" like lattice QCD do use the path integral directly!
- froeb 4y agoAlright fair enough, I'm not an expert on how those work. I was under the impression that there is quite a bit of work that has to be done to turn a path integral into something amenable to Monte Carlo though.
- evanb 4y agoWell it depends on what you mean by "quite a bit of work"! Wick rotate, focus on a finite volume of spacetime, and discretize the spacetime. That's all. The first step changes a difficult-to-sample distribution exp( i S ) into an easy-to-sample exp( – S ). The other two steps keep you from needing infinite RAM. The discretization and finite-volume approximations may be removed by doing calculations with different spacings and volumes and extrapolating. The Wick rotation is more subtle, and typically means only some observables are accessible. To study the path integral with real time causes a very difficult sign problem [essentially: you need an exponentially large number of samples in order to get cancellations under control], which is why real-time quantum dynamics is an exciting potential application of quantum computers.
- Jach 4y agoI'm tempted to go check, I read that too around the end of HS or very start of college, but isn't QED the one that starts off with a story about counting physical beans in jars to do familiar arithmetic operations like summing and multiplying? Then the rest of QED is presented as describing how QED works using a similar counting-beans approach that everyone can understand vs. the more sophisticated methods that working physicists use that require a few years of undergrad training at least. That both end up having infinities at the core is an important part of being able to accurately (mathematically) describe nature, like counting beans vs. just using a calculator both involve numbers at the core to accurately describe to other humans what's going on -- and numbers can seemingly be "canceled out" when removing beans from a jar! (Or perhaps adding a different color bean does the canceling.) Without thinking about it further I suppose this falls under the umbrella question of whether reality is fundamentally continuous (as we assume when we model things with calculus) or discrete.
- prof-dr-ir 4y agoTheoretical physicists are, on the whole, not a silly bunch and there do exist various "mathematically identical" ways to look at quantum field theory. But by the very definition of "mathematically identical" one can find the path integrals somewhere in these alternative descriptions anyway. Contrary to your stated desire, I actually think that the path integral formulation is in some sense the most "intuitive" and "local". I apologize for not taking the time to explain that here, but let me instead try to convince you that "locality" is not at all obvious in quantum field theory. One intuitively non-local process is the following. If I start with a state with one electron at the origin at time 0, what is the probability that I measure an electron one lightyear away one minute later? Naively this probability is zero, because the electron that you started with cannot move faster than light. But the actual (and measured, in a less idealized setting) probability is not zero, since vacuum fluctuations can create an electron exactly one lightyear away and you end up measuring that one. That, a physicist would say, is perfectly cromulent with locality of the underlying theory. I hope this gives a glimpse of how subtle causality (and hence locality) really is, and why at least some believe that path integrals may actually be the most "intuitive" description available.
- Y_Y 4y ago> Theoretical physicists are, on the whole, not a silly bunch ... at least until the workday ends. My own feeling is that it's obviously local if you consider that there really are virtual particles taking the non-critical paths. I know this is a bit interpretation-dependent though. GP may be interested to read e.g. https://en.m.wikipedia.org/wiki/Double-slit_experiment#Interpretations_of_the_experiment https://en.m.wikipedia.org/wiki/Double-slit_experiment#Inter... Of course the simple way around this issue is the best: "shut up and calculate".
- GoblinSlayer 4y agoThat's like looking at distant synchronized clocks and concluding they communicate faster than light. Your wish to make it look nonlocal isn't supported by physics.
- bmacho 4y ago> So now that I'm on a forum with smart people, the kind I didn't have access to back in high school... has there been any progress towards something like this at all? Are there any directions for this? Or is there a conceptual reason I'm missing why this fundamentally isn't possible, why the sum of all possible paths is the only way to account for certain behaviors? It certainly not the only way. One easy trick: you put it under multiple layers of definition, and at the lowest layer you just do the integral without calling anything "paths" or so. (Equivalent theories can be seen as bastardization of each other, that's why none of them says anything meaningful about ontology.)
- 323 4y agoWhy wouldn't summing all possibilities be the simpler solution? Thing about Monte Carlo simulations - it's simpler to just try all possibilities than to figure out what is the formula of the correct result. Neural networks also work like that - you randomly do things and propagate errors with simple rules instead of figuring out the complex formula that would give you the answer. There seems to be a lot of power in summing up all the things. If our universe is a computation, our practical experience seems to suggest that a Monte Carlo like simulation of it would be more efficient than a closed-form formula.
- kzrdude 4y agoSumming all the things is similar to just a big "OR" right? It lets the important signal through (you don't know which it is) and the other ones just have to be suppressed or suppressed enough to not interfere. For example all the other ones might as well be as if stochastic, so on average you get it right.
- shanebellone 4y ago"it's simpler to just try all possibilities than to figure out what is the formula of the correct result." Isn't this the premise of quantum computing too?
- oneoff786 4y agoThe canceling out part is the reason quantum computing works. We basically just avoid computing a bunch of bad answers making the actual problem faster to solve. So it has some observable examples.
- ladyattis 4y agoI think the fact that the oddest approaches work aren't exactly a measure of how the particles actually work in physics. I think it proves that math isn't the language of the universe no matter how much mathematicians and physicists want to say it is. It just proves we're good at modeling but not good enough to actually know what we're seeing/measuring (a natural limit to our knowledge). I don't know why this position is considered controversial or out of the mainstream when it seems to be the logical answer.
- naasking 4y agoI don't see why it's logical to conclude that difficulties modeling hard problems means we can't model them using math. These difficulties are not uncommon in the history of science, and we've eventually solved them all before with math, and we should expect such problems to become more and more difficult as the low hanging fruit has been plucked. I see literally no reason to jump to the conclusion that the core problem is trying to use math at all.
- ladyattis 4y ago>I don't see why it's logical to conclude that difficulties modeling hard problems means we can't model them using math. That's not what I'm stating though. I'm stating that the math involved is a model but it isn't ever going to be identical to the thing being modeled. Meaning that math isn't the "language of nature" as we don't have a direct means to truly comprehend it (direct realism has a whole has been discarded by philosophers for a long time now). >I see literally no reason to jump to the conclusion that the core problem is trying to use math at all. Again, that's not what I said, please read my post again.
- naasking 4y ago> I'm stating that the math involved is a model but it isn't ever going to be identical to the thing being modeled. I don't know what "identical" means in this context. Either a mathematical model can capture all of the information in the system, or it can't. We know that we can reproduce a function on a long enough timeline just by observing its outputs via Solomonoff Induction. The only escape hatch here is if reality has incomputable features. There's no evidence of this at this time. That's why it's confusing that you would go from "we have persistent hard problems" to "mathematical models can't exactly correspond to reality".
- daniel-cussen 4y ago[dead]
- naasking 4y agoMany-Worlds probably fits what you're looking for, that conforms with orthodox quantum mechanics. How to bootstrap probability seems like the only open question there. If you want something less orthodox, then superdeterminism is seeing a bit of a revival. For instance, Tim Palmer's Invariant Set Theory (Invariant Set postulate on Wikipedia).
- GoblinSlayer 4y agoBorn rule postulates that probabilities exist at sample size 1, which is not supported by reality: in reality probabilities are unobservable at sample size 1, which better corresponds to MWI than copenhagen. MWI shoves it in your face why probabilities are unobservable at sample size 1.
- philipov 4y agoIt kind of feels like the epicycles that were used to make geocentrism work. Sure, they make the math work out correctly, but the equations for the motion of the planets become needlessly more complex than if you just take the sun to be the center.
- tim333 4y ago>is there a conceptual reason I'm missing why this fundamentally isn't possible, why the sum of all possible paths is the only way It's not so much a conceptual reason it isn't possible so much as that's just the way the universe is. If you have your particle going from A to B, changing things at some possible other path C simply does change the way it arrives at B, like it or not. Feynman's take was: >If you will simply admit that maybe [nature] does behave like this, you will find her a delightful, entrancing thing. Do not keep saying to yourself, if you can possible avoid it, "But how can it be like that?" because you will get 'down the drain', into a blind alley from which nobody has escaped. Nobody knows how it can be like that. Personally though I actually like pondering how it can be like that as an interesting puzzle. But it's the way the universe is, not a conceptual thing. My guess as to why is that reality at the most basic level is more like a path integral than like objects whizzing around.
- kanzenryu2 4y agoThis is probably the best explanation you're going to find https://www.lesswrong.com/posts/apbcLXz5zB7PXfgg2/an-intuitive-explanation-of-quantum-mechanics https://www.lesswrong.com/posts/apbcLXz5zB7PXfgg2/an-intuiti...
- GoblinSlayer 4y ago>the sum of all possible paths is the only way to account for certain behaviors? It isn't? Also you put too much hope into mathematics, it only calculates, nothing more.