In what sense is QM not understood ?

[Doofy]

In what sense is QM "not understood"?

This is something that I've seen repeated many times, but I'm wondering how accurate it is. I mean, we've got this mathematical framework where we deal with vector spaces, eigenstates, superpositions, mixed states etc. that works to a high degree of accuracy.

Is it just the fact that QM deals with probabilities of measuring final states rather than the 1 input --> 1 output style of classical mechanics that makes people say it's "not understood" ? Is "not understood" just another way of saying "not familiar in terms of everyday human experience" ?

What I wonder about is how the founders of QM figured out that the mathematics we use in QM (operators, bras, kets etc.) was the right thing to use. They didn't just pull it out of thin air, they must have reasoned their way to at least some of it, eg. Schrodinger didn't just get out a pen and write down $H\Psi = i\hbar \frac{d}{dt}\Psi$ out of nowhere. Why isn't that considered "understanding" it?

 

[jtbell]

To me, "not understood" means that there is no generally accepted interpretation for what is "really happening" underneath the probabilistic mathematics of QM. See all the arguments about interpretational / metaphysical / philosophical issues surrounding QM in this forum.

 

[Bill_K]

Is "not understood" just another way of saying "not familiar in terms of everyday human experience" ?

I'd agree with that. Feynman was one person often quoted as saying that he didn't understand quantum mechanics (!) But his idea of "understanding" was, "can I explain it to somebody without using mathematics" and of course he found that difficult to do.

What I wonder about is how the founders of QM figured out that the mathematics we use in QM was the right thing to use.

I'm not a big fan of the historical approach. The founders didn't understand what they were doing. How they came to their conclusions was often a complex process and largely irrelevant. They made many wrong guesses along the way, and some of these guesses have even become immortalized. People often allude to Dirac's hole theory, for example, without mentioning that it was shown to be false by Heisenberg only a few years later.

 

[Doofy]

I'd agree with that. Feynman was one person often quoted as saying that he didn't understand quantum mechanics (!) But his idea of "understanding" was, "can I explain it to somebody without using mathematics" and of course he found that difficult to do.

I'm not a big fan of the historical approach. The founders didn't understand what they were doing. How they came to their conclusions was often a complex process and largely irrelevant. They made many wrong guesses along the way, and some of these guesses have even become immortalized. People often allude to Dirac's hole theory, for example, without mentioning that it was shown to be false by Heisenberg only a few years later.

ah I like to know these things if possible, it gives you a fuller picture. I kind of feel like I won't "understand" QM in any significant way until I know how its starting postulates were dreamt up. There must be some account out there of how they did it that cuts out all the dead-ends and mistakes.

 

[kith]

I agree with Bill that the historical approach is not very good if you want to understand how the postulates of QM can be motivated. A good motivation is contained in Ballentine's book on QM.

 

[Bill_K]

I kind of feel like I won't "understand" QM in any significant way until I know how its starting postulates were dreamt up.

OK, I can see I haven't convinced you. But ask yourself this - do the Feynman lectures try to teach physics by recounting what Dr X and Prof Y once thought years ago, or do they take an entirely fresh and modern viewpoint?

 

[vanhees71]

I agree that the historical approach to learn physics is not very good. Particularly quantum theory is complex enough without all the balast of the early history. Most hindering in understandung modern quantum theory is the Bohr model and "philosophical considerations". The best thing is to first learn the formalism and then think about the socalled "interpretation". I think the best interpretation is the minimal statistical interpretation, very nicely covered in Ballentines book.

However, on the other hand it is very interesting to know about the history of science. Sometimes it indeed helps to dig deeper into the meaning of theories. That's also true for quantum theory after you have come to term with its physics content in terms of the minimal interpretation. A very concise source is the multi-volume work by Mehra and Rechenberg on the history of quantum mechanics.

 

[kith]

A very concise source is the multi-volume work by Mehra and Rechenberg on the history of quantum mechanics.

Have you read all of it?

I read some reviews and got the impression that it would be a good thing to read through when you are retired. ;-) Unfortunately, it is quite expensive.

 

[Ken G]

The answers you have already are excellent, but perhaps we can dig a little deeper into what is "not understood." So far you have your own idea that this can mean "not familiar with everyday experience", but I don't think that quite cuts it, because we don't necessarily expect everything we discover to be familiar to us, our everyday experience is limited. Relativity is a perfect example-- we all know it includes extra elements that break our naive concept of time, and yet you don't hear nearly as much said about "no one understands relativity." Relativity simply involves postulates that go beyond our daily experience-- they don't contradict our daily experience. It's like if our daily experience was a trunk, we're not surprised to discover, with higher precision, that the trunk has an elephant that we were not familiar with attached to it.

We also have that it could mean that there is no accepted interpretation of what is "really happening", below the surface. I think that is getting closer, but as one who never tends to regard physics as a story of "what is really happening", I don't see classical mechanics as providing that story much better than quantum does (why is action minimized? Why do forces produce acceleration? We really don't have much of a sense of what is "really happening" in classical mechanics either). Indeed, some people interpret the backstory of classical mechanics to be quantum mechanics, by invoking the "correspondence principle."

Nor can we say that what is not understood about quantum mechanics is that it admits to multiple interpretations that have no obvious connection. In classical mechanics, we can learn F=ma, or we can learn the Lagrangian, or even the Hamiltonian approaches, and on the surface, these sound about as different as night and day. So what is so special about quantum mechanics that makes people like Bohr and Feynman, who have earned Nobel prizes in that very field, say that no one understands it?

I think it comes from the measurement problem. Quantum mechanics seems to be built, from the ground up, to revolve around an inherent contradiction. Its fundamental dynamical equation is deterministic, yet it is only used to make statistical predictions. On a related note, the states in quantum mechanics always evolve unitarily (so multiple possible observed outcomes are intrinsically included in the state), but the act of measuring them, and connecting to macroscopic instruments, appears to break that unitarity (since only one outcome is perceived, from the superposition of possibilities). The "cause" of that break is not at all clear, and differs substantially in the different interpretations. That is what I think is at the source of what is "not understood"-- not just that we have multiple interpretations, but that the interpretations all have to grapple with what seems like a central contradiction.

So what is not known is whether some future theory will address this contradiction and remove it, or if the contradiction is in some sense supposed to be there-- it is something we were supposed to discover about reality as we advanced. (And in my view, understanding decoherence in no way resolves this contradiction, it merely shifts the focus of when the contradiction is encountered, since all the interpretations can account for decoherence but do so in very different ways.)

 

[skippy1729]

What I wonder about is how the founders of QM figured out that the mathematics we use in QM (operators, bras, kets etc.) was the right thing to use. They didn't just pull it out of thin air, they must have reasoned their way to at least some of it, eg. Schrodinger didn't just get out a pen and write down $H\Psi = i\hbar \frac{d}{dt}\Psi$ out of nowhere. Why isn't that considered "understanding" it?

In a sense they did pull the formalism out of thin air. They tinkered with classical mechanics trying to make it fit their experiments. Relaxing the commutative law of multiplication lead them to q-numbers then to generalized matrices and then to operators and Hilbert space. The founders of QM invented Hilbert space. If you want to see how they did it try "Sources of Quantum Mechanics" edited by B. L. van der Waerden for English translations of many of the important early papers.

Perhaps the fact that there has never been a consensus as to what might be the nature of the underlying physical system could be taken as evidence that Quantum Mechanics really is a fundamental theory. Maybe reductionism stops here.

 

[stevendaryl]

People often allude to Dirac's hole theory, for example, without mentioning that it was shown to be false by Heisenberg only a few years later.

I'm not sure what you are alluding to here. Dirac's hole theory has been abandoned in favor of a theory of electrons&positrons (as opposed to positive and negative energy electrons), but I always thought of that as a reinterpretation of essentially the same theory. In what sense did Heisenberg prove Dirac's hole theory wrong?

 

[Ken G]

Perhaps the fact that there has never been a consensus as to what might be the nature of the underlying physical system could be taken as evidence that Quantum Mechanics really is a fundamental theory. Maybe reductionism stops here.

"Never" is a very short time when it comes to quantum mechanics! Less than one century. I'd say it's far too early to predict the ending point of reductionism, indeed I don't think there is any time limit on keeping that question open. But by way of analogy, note that there was never a consensus as to what might be the nature of the underlying physical system that holds for gravity in either general relativity, or even Newtonian mechanics! The latter has been around for 250 years, with no consensus about its underlying physical system, and now there isn't even a consensus on what theory we should be looking for an underlying physical system for in the first place! The sands of knowledge can be stable for a long time, and still undergo seismic shifts.

 

[nikman]

If reductionism has an ending point you'll find it via QFT, specifically by exploring the fermion minus sign problem and its distressing/disturbing ramifications. Strong fermionic interaction (nothing to do with the "strong" force) has so far proven not to be computable, not even mathematizable.

It's all about how real messy stuff becomes real messy stuff instead of remaining neat coherent wave functions (which are, of course, symbolic formulations). Anyway if a thing depends for its own definition on the definition of what it interacts with and one continuously re-defines the other you've essentially got an irreducible recursion.

 

[Fiziqs]

Speaking as one who happens to agree with the idea that today's physicists don't really understand what they're talking about, puts me in a unique position to answer this question. It seems to me that QM does a fantastic job of predicting what will happen, but it is for the most part clueless when it comes to explaining why it will happen. To me, if you can't explain why something happens, then you probably don't understand what's going on. I can predict with pretty darn good accuracy that the sun will rise in the east tomorrow, but predicting the outcome, and understanding the cause, are two completely different things.

The excuse that QM is simply too complicated, or too far removed from our normal daily experience to be grasped by a layman, is in my view, complete BS. If I can't understand it, it's most likely due to the fact that the person trying to explain it, doesn't understand it either, so has to rely on psuedo-scientific gibberish in an attempt to fain competence. If you can't put it in layman's terms, it's not due to the fact that it's too complex, it's due to the fact that you don't understand it well enough.

I often hear theists describe science as a religion. Such comments invariably come from people who have little understanding of the true nature of either. But in one disturbing way, they are indeed alike, in both, the elite presume themselves to be in a position of understanding, unattainable by the uneducated layman. They assume a position of intellectual superiority. In truth, if we the laymen fail to understand, the fault lies in the inadequacy of the explanation, not some lack of "divine" insight.

 

[robertjford80]

Is it just the fact that QM deals with probabilities of measuring final states rather than the 1 input --> 1 output style of classical mechanics that makes people say it's "not understood" ? Is "not understood" just another way of saying "not familiar in terms of everyday human experience" ?

I think this statement is correct. The charge that no one understands QM is overblown or hyperbolic. It's just that QM is not deterministic. So long as you understand that it is not deterministic and that that is just a brute fact, then there is nothing to misunderstand.

 

[stevendaryl]

I think this statement is correct. The charge that no one understands QM is overblown or hyperbolic. It's just that QM is not deterministic. So long as you understand that it is not deterministic and that that is just a brute fact, then there is nothing to misunderstand.

I would say that it is not just because it is nondeterministic that people say they don't understand quantum mechanics. It's the combination of nondeterminism together with extremely strong correlations that is hard to understand.

In an EPR experiment, you produce a twin pair of spin-1/2 particles. Alice measures the spin of one particle along some axis, and gets +1/2 or -1/2. Bob measures the spin of the other particle along a different axis, and gets +1/2 or -1/2.

The fact that Alice's result is nondeterministic is not hard to understand. But the fact that, in the case where Alice and Bob choose the same axis, they always get opposite result, is hard to understand. If Alice knew what axis Bob was going to choose, and Alice did her measurement a second before Bob, then she would know exactly what result Bob would get. So in that situation, from her point of view, Bob's result isn't nondeterministic--it's completely predictable.

It's the combination of perfect nondeterminism and perfect correlations that is hard to understand about quantum mechanics.

 

[Fredrik]

The theory is of course understood very well. What's not understood is how the things described by the theory correspond to things in reality, especially at times between state preparation and measurement. The theory tells us how to calculate the probabilities of all possible results of all possible measurements, using knowledge of how the system was prepared as input. It doesn't tell us what the system is "really doing" at times between state preparation and measurement, at least not in terms that we can easily understand. In particular, we don't even know if particles have positions or not.

Why isn't that considered "understanding" it?

What's considered "understanding" is of course highly subjective.

 

[stevendaryl]

The theory is of course understood very well. What's not understood is how the things described by the theory correspond to things in reality, especially at times between state preparation and measurement. The theory tells us how to calculate the probabilities of all possible results of all possible measurements, using knowledge of how the system was prepared as input. It doesn't tell us what the system is "really doing" at times between state preparation and measurement, at least not in terms that we can easily understand. In particular, we don't even know if particles have positions or not.

What's considered "understanding" is of course highly subjective.

I would say that there are aspects of the theory that are not understood, either. We have the recipe for using quantum mechanics, which is:

1. Between measurements, the system evolves according to Schrodinger's equation.
2. Measurement of any observable results in an eigenvalue of that observable, with probability computed from the wavefunction.
3. After a measurement, the system is in the eigenstate of the observable corresponding to the eigenvalue measured.

What's really not understood, at a theoretical level, is what constitutes a "measurement". We have a rule of thumb answer, which is that an interaction counts as a measurement if it leaves an irreversible record, such as a photograph, or a bubble in a bubble chamber, or a click in a Geiger counter, etc. But I wouldn't say that there is a very good theoretical understanding of what a measurement is.

 

[harrylin]

I agree with Bill that the historical approach is not very good if you want to understand how the postulates of QM can be motivated. A good motivation is contained in Ballentine's book on QM.

On this I agree with Doofy: I still would feel that SR is "magic" if I had not studied and understood its historical development. Regretfully I don't know much of the historical development of QM and its motivations (and it's still like magic to me).

[..] I can predict with pretty darn good accuracy that the sun will rise in the east tomorrow, but predicting the outcome, and understanding the cause, are two completely different things. [..]

Exactly.

 

[krd]

I agree that the historical approach to learn physics is not very good. Particularly quantum theory is complex enough without all the balast of the early history.

It depends on your learning style. I'm out of college a very long time. And I'm only coming back to my physics now. I find I'm learning much more, and having a much better understanding of the physics by studying the history. It's helping me re-learn my physics. And sometimes I find little tid bits that I would have missed otherwise. And it's sometimes just little tiny ideas, that link other pieces together.

Most hindering in understandung modern quantum theory is the Bohr model and "philosophical considerations". The best thing is to first learn the formalism and then think about the socalled "interpretation". I think the best interpretation is the minimal statistical interpretation, very nicely covered in Ballentines book.

Learning the formulas and learning how to execute them is not enough. I have seen a few instances where professional scientists - who can do all the fancy calculus - have had misunderstandings of the fundamental theory - or have had gaps in their understanding that shouldn't be there. It's bad science, to know all the names, know all the maths, but have misunderstandings of the underlying theory.

I don't know how much the interpretations may change - but physics and chemistry, the way those subjects are taught in schools, the teaching materials probably need to be completely gutted and rebuilt from the ground up. There might be better ways to describe physics and chemistry to make everything fit more coherently.

 

[stevendaryl]

That sounds inconsistent to me; it looks logical to me that the motivation of the postulates should first be understood in the historical context out of which they emerged.

But often theoretical developments happen when someone just starts "playing around" with ideas and with mathematical formalisms, and then noticing that something neat comes out of it. The context for the development of quantum mechanics was the observation that the energy levels of an electron in a hydrogen atom took only discrete values. So various people started looking at various ways that a discrete set of values can be produced.

Bohr's idea was just the ad-hoc rule that the angular momentum of an electron must be an integer multiple of h-bar. (This could be heuristically justified in terms of de Broglie's notion of matter/wave duality--only for certain values of angular momentum would the corresponding "matter wave" be a standing wave.)

Heisenberg noted that discrete eigenvalues pop up in matrix problems. So maybe operators like position, momentum, angular momentum, energy, etc., can be represented by matrices, or generalizations.

Schrodinger noted that discrete eigenvalues pop up in solutions to differential equations, so maybe there is some kind of function associated with the electron that satisfies a differential equation that produces eigenvalues corresponding to the observed energy levels.

These ideas were important, but they were really along the lines of guesses. It is really barking up the wrong tree to look to these founders for answers about the true meaning of quantum mechanics. Heisenberg had no more idea about the implications of noncommuting observables than anybody else did. Schrodinger had no more idea about the true meaning of the wave function than anybody else did. They were motivated by wanting to get discrete values for observables. I don't think that there was anything deeper involved. So that's the sense in which the historical point of view is of limited usefulness--the founders don't necessarily understand the theory any better than anybody else.

 

[Fredrik]

I think that is getting closer, but as one who never tends to regard physics as a story of "what is really happening", I don't see classical mechanics as providing that story much better than quantum does (why is action minimized? Why do forces produce acceleration? We really don't have much of a sense of what is "really happening" in classical mechanics either).

I don't share this view. Those questions are really about why things are happening, rather than about what is happening. When we solve the equation of motion of a two-body gravitational system, we find elliptical orbits, and no one doubts that it makes sense to say that an elliptical orbit is an approximate description of what the first object is "doing" near the other.

In classical mechanics, we can learn F=ma, or we can learn the Lagrangian, or even the Hamiltonian approaches, and on the surface, these sound about as different as night and day.

The way I see it, non-relativistic classical theories are all defined in a framework defined by Galilean spacetime. The Newtonian, Lagrangian and Hamiltonian approaches are just three different ways to consistently add matter to an empty spacetime. A specific theory in that framework is defined by its equations of motion. One way to find a new theory in this framework is to simply guess an equation of motion. (Actually, that is the Newtonian approach). The other approaches are just ways to eliminate the worst guesses. So I don't find it surprising that these approaches don't tell us anything about what's actually happening. They're not even part of the theories; they are just tools that help us eliminate the worst candidates for new theories.

So what is so special about quantum mechanics that makes people like Bohr and Feynman, who have earned Nobel prizes in that very field, say that no one understands it?

In my opinion, it's that it assigns non-trivial probabilities (not always 0 or 1) to measurement results even when the state is pure (i.e. when we have maximal information about the preparation procedure). This inevitably raises questions like this: "If the state describes what's happening to the system, and assigns non-zero probabilities to two mutually exclusive measurement results, doesn't that mean that the system is actually doing both of those things?"

 

[harrylin]

But often theoretical developments happen when someone just starts "playing around" with ideas and with mathematical formalisms, and then noticing that something neat comes out of it. The context for the development of quantum mechanics was [..]

Thanks for giving a summary - and I would appreciate to read a detailed, in-depth article superior to textbook summaries. Also thanks for preserving part of my original comment which I completely lost due to an editing mistake. :tongue2:

These ideas were important, but they were really along the lines of guesses. It is really barking up the wrong tree to look to these founders for answers about the true meaning of quantum mechanics. [..]

Likely so; still I expect that there is more to be found in a multitude of opinions and approaches of people who didn't really understand it, than in a single opinion of someone who also doesn't really understand it.

 

[Fredrik]

I would say that there are aspects of the theory that are not understood, either.

I agree. That's why I wrote "very well" instead of "completely". There are certainly mathematical theorems left to be proved, and I don't think we have a perfect understanding of how theories of interacting matter are to be defined in the framework of quantum mechanics. (I think it's perfectly understood in the case of non-interacting particle theories. I'm not sure about the case of interacting quantum field theories. I'm pretty sure that we don't have the answer when gravity is involved).

What's really not understood, at a theoretical level, is what constitutes a "measurement". We have a rule of thumb answer, which is that an interaction counts as a measurement if it leaves an irreversible record, such as a photograph, or a bubble in a bubble chamber, or a click in a Geiger counter, etc. But I wouldn't say that there is a very good theoretical understanding of what a measurement is.

I don't know. I find that pretty satisfactory actually. Not in the sense that I wouldn't want to have a better understanding of it, but in the sense that I believe that this is the best we will ever be able to do without a better theory to replace QM.

 

[harrylin]

I don't share this view. Those questions are really about why things are happening, rather than about what is happening. When we solve the equation of motion of a two-body gravitational system, we find elliptical orbits, and no one doubts that it makes sense to say that an elliptical orbit is an approximate description of what the first object is "doing" near the other.[..]

That shape was first proposed by Keppler who gave the correct equation first. However, if I correctly recall, I read somewhere that Keppler complained that he did not understand it. Later Newton's theory of gravitation gave a first feeling of understanding of the "why", not just due to equations but due to identifying a physical cause to which those equations relate. But perhaps that is what you meant?

 

[stevendaryl]

I don't know. I find that pretty satisfactory actually. Not in the sense that I wouldn't want to have a better understanding of it, but in the sense that I believe that this is the best we will ever be able to do without a better theory to replace QM.

Yeah, I agree that the recipe for using quantum mechanics, using an informal notion of what counts as measurement, works pretty well, but I wouldn't say that the theory behind it is well understood. In particular, if measurements are themselves interactions (and what else would they be?) then they should themselves be described by quantum mechanics, rather than having a separate rule (wave function collapse to an eigenstate following a measurement).

 

[Ken G]

I don't share this view. Those questions are really about why things are happening, rather than about what is happening. When we solve the equation of motion of a two-body gravitational system, we find elliptical orbits, and no one doubts that it makes sense to say that an elliptical orbit is an approximate description of what the first object is "doing" near the other.

Yes, it is easier to describe what is happening in classical systems, but I would argue that even Kepler did that-- before there even was anything we could call classical mechanics. So we cannot argue that we understand the theory of classical mechanics simply because what we are trying to predict is easier to describe pictorially-- I think when we talk about understanding a theory, what we mean is, understand why that theory provides a good description of the behavior we see, even if the behavior seems weird. The classic example is relativity-- with a few fairly reasonable sounding postulates, we obtain an explanation of very weird behavior, so we say we understand relativity. The postulates don't seem to make any unbelievable claims.

But in the case of quantum mechanics, we have that rift built right into the postulates-- the rift between unitary evolution, and the Born rule. There's just no way to describe that rift without either asserting some physical structure that is completely not in evidence (like a pilot wave, or many worlds), or essentially saying "and then something we can never understand happens" (like Bohr did). So the what is not really that hard to describe (we get interference patterns, we get Bell correlations, etc.), it's just a bit more sophisticated than classical physics (and its elliptical orbits, as you say), and the why is inscrutable as usual-- nether the what or the why seem to be the crux of what is so hard to grasp about quantum mechanics. I think it is the measurement problem, that core inconsistency in the theory, which also spawns all the different interpretations. Those interpretations are weird not because they are different (we always see lots of different sounding interpretations of any theory, like Lorentz aethers and so on), but because of the basic disconnect they are grappling with.

So I don't find it surprising that these approaches don't tell us anything about what's actually happening.

Neither do I, because I don't think physics theories are supposed to tell us that. I don't think that's why we say we don't understand quantum mechanics.

In my opinion, it's that it assigns non-trivial probabilities (not always 0 or 1) to measurement results even when the state is pure (i.e. when we have maximal information about the preparation procedure). This inevitably raises questions like this: "If the state describes what's happening to the system, and assigns non-zero probabilities to two mutually exclusive measurement results, doesn't that mean that the system is actually doing both of those things?"

Exactly, we are in agreement-- it is the measurement problem. Our theory is trying to tell us that multiple outcomes are in some sense "wrapped up" in the same state, yet we never actually see anything but one outcome.

 

[nitsuj]

What's really not understood, at a theoretical level, is what constitutes a "measurement". We have a rule of thumb answer, which is that an interaction counts as a measurement if it leaves an irreversible record, such as a photograph, or a bubble in a bubble chamber, or a click in a Geiger counter, etc. But I wouldn't say that there is a very good theoretical understanding of what a measurement is.

I know very little about SR, and even less about QM. But from what I have learned from SR has me believing QM uncertainty roots into how we define & measure the dimensions (not to suggest there is a "solution"). Your comment above is well said and easy to understand. Tough thing to do for QM concepts.

 

[Fredrik]

That shape was first proposed by Keppler who gave the correct equation first. However, if I correctly recall, I read somewhere that Keppler complained that he did not understand it. Later Newton's theory of gravitation gave a first feeling of understanding of the "why", not just due to equations but due to identifying a physical cause to which those equations relate. But perhaps that is what you meant?

I agree that we don't need a theory as sophisticated as Newton's to get this approximate description of what an object is "doing" while in orbit. I mentioned elliptical orbits as an example of when classical mechanics clearly tells us what an object is doing, to counter the suggestion that classical mechanics doesn't do that. (This was not an attempt to prove Ken G wrong, because it's clear that he and I mean different things by "describe what's happening", and "understand a theory". I only meant to illustrate what sort of thing I have in mind when I'm talking about descriptions of "what's happening").

A statement like "the orbits of planets are ellipses", is a theory by my definitions, because it makes testable predictions about results of experiments. This simple theory is already an approximate description about what's happening to an object in orbit. Newton's theory is a better theory, because it makes more accurate predictions about a wider range of phenomena.

Newton's theory explains why the simple theory works, but it raises a whole new set of "why?" questions. This illustrates another important idea: that the only thing that can explain why a theory works, is a better theory.

 

[Ken G]

There's probably a more general way to think about the issue of "what is a measurement" which cuts deeper into the heart of the problem-- and that is, "what is the role of the physicist in the physics." This is the element that Bohr was so focused on, and many take issue with him for raising such a philosophical issue, but I think his insight is still the crux of the matter. So in these terms, "what we don't understand" about quantum mechanics is "why can't we escape the role of the observer." In all other areas of physics, we can imagine that the observer is just a kind of "fly on the wall", and we don't have to attach any importance at all to the fact that an observation is being carried out. That's exactly what we cannot do in quantum mechanics, and we just don't know why. How we resolve that uncertainty is exactly the role of the various interpretations, but none can produce an unequivocally demonstrable answer-- to put it mildly.

 

[Ken G]

Newton's theory explains why the simple theory works, but it raises a whole new set of "why?" questions. This illustrates another important idea: that the only thing that can explain why a theory works, is a better theory.

That's a key point I don't think a lot of people recognize about a physics theory, no matter how accurate or widely accepted it is: it never tells us "why" nature works the way she does, it only tells us why some previous theory worked as well as it did! To explain why we get the observations we do, we would actually need a theory that described what we are doing when we make an observation, which requires that we can model ourselves, modeling ourselves, and so on. That's why I hold it is never possible to use physics to say "why" we observe what we do, and we should not make that our goal for doing physics. But we'd still like to have theories that give a consistent and complete account that connects nature to the observed result, and that's just what quantum mechanics does not do, without invoking an interpretation that few agree on. I actually see this as a feature of QM, not a bug-- we aren't supposed to be able to map the complete connection between what nature is doing to our observation of it.

 

[Fredrik]

Yes, it is easier to describe what is happening in classical systems, but I would argue that even Kepler did that-- before there even was anything we could call classical mechanics.

Yes, I agree about this part. (More details in my reply to harrylin above).

So we cannot argue that we understand the theory of classical mechanics simply because what we are trying to predict is easier to describe pictorially-- I think when we talk about understanding a theory, what we mean is, understand why that theory provides a good description of the behavior we see, even if the behavior seems weird. The classic example is relativity-- with a few fairly reasonable sounding postulates, we obtain an explanation of very weird behavior, so we say we understand relativity. The postulates don't seem to make any unbelievable claims.

I'm not sure what you're saying here. Is it one of the following things? A) To understand the theory is to understand its mathematics and correspondence rules (the assumptions that tell us how to interpret the mathematics as predictions about results of experiments), or B) To understand the theory is to understand why its predictions are accurate.

If you meant A, then what we need to do before we can say that we understand the theory, is to prove the most relevant theorems, and convince ourselves that we have the right idea about how to perform measurements of the sort the theory makes predictions about. (I would say that we have accomplished this to a satisfactory degree already).

If you meant B, then what we need to do is to find a better theory. (If this is what you meant, then we have very different ideas about what it would mean to understand the theory. I would say that this is actually unrelated to "understanding the theory". It's an entirely different issue).

Hm, you probably meant neither. Maybe you meant C) To understand the theory is to know which things in the purely mathematical part of the theory correspond to things in the real world. This is of course the part that no one understands. So if we define "understand the theory" this way, then we don't understand it. But I don't use this definition. I'm using the one I labeled "A" above.

 

[Fredrik]

There's probably a more general way to think about the issue of "what is a measurement" which cuts deeper into the heart of the problem-- and that is, "what is the role of the physicist in the physics." This is the element that Bohr was so focused on, and many take issue with him for raising such a philosophical issue, but I think his insight is still the crux of the matter. So in these terms, "what we don't understand" about quantum mechanics is "why can't we escape the role of the observer." In all other areas of physics, we can imagine that the observer is just a kind of "fly on the wall", and we don't have to attach any importance at all to the fact that an observation is being carried out. That's exactly what we cannot do in quantum mechanics, and we just don't know why. How we resolve that uncertainty is exactly the role of the various interpretations, but none can produce an unequivocally demonstrable answer-- to put it mildly.

I have come to think about this role of the observer as an essential feature of the concept of "physics". Theories of physics are falsifiable statements about reality. To be falsifiable, a statement must have testable consequences. In other words, we must be able to use it to make predictions about results of measurements. And what is a measurement? It's an interaction between the system and its environment that puts some part of the environment into one of several states that a human observer can interpret as a result of the measurement. Such a state must last long enough for a human to observe it, and be distinguishable from states that correspond to other results. So that part of the environment, the "pointer" that indicates the result, has to behave in a way that will be perceived as classical.

A "classical" theory is a theory that only makes predictions that can be tested without significantly disturbing the system. So maybe we shouldn't be asking why QM is so weird, but instead be asking why there are classical theories that are actually pretty good.

That's a key point I don't think a lot of people recognize about a physics theory, no matter how accurate or widely accepted it is: it never tells us "why" nature works the way she does, it only tells us why some previous theory worked as well as it did! To explain why we get the observations we do, we would actually need a theory that described what we are doing when we make an observation, which requires that we can model ourselves, modeling ourselves, and so on. That's why I hold it is never possible to use physics to say "why" we observe what we do, and we should not make that our goal for doing physics. But we'd still like to have theories that give a consistent and complete account that connects nature to the observed result, and that's just what quantum mechanics does not do, without invoking an interpretation that few agree on. I actually see this as a feature of QM, not a bug-- we aren't supposed to be able to map the complete connection between what nature is doing to our observation of it.

Good post. No objections from me.

 

[krd]

I actually see this as a feature of QM, not a bug-- we aren't supposed to be able to map the complete connection between what nature is doing to our observation of it.

Who says we're not supposed to?

We have to keep asking questions - reformulating things. Maybe, sometime in the future - a few thousand years from now we'll arrive at the end.

Dream.

We're no where near the end. Like at the minute we do not have 3d prints, that can shoot beams and create whatever matter we want - like pressing a button and making a chocolate cake appear out of nothing. I know it sounds like impossible magic. But so would mobile phones have sound to the ancients. Although they did believe their priests could talk to god.

 

[nitsuj]

I actually see this as a feature of QM, not a bug-- we aren't supposed to be able to map the complete connection between what nature is doing to our observation of it.

Interpretation whether it be a "problem" (in context of predecessor theories) or a "feature" of nature at that level. A little poetic

Just really shocked that, for as much SR accounts for everything right down to defining dimensions, that QM doesn't mention them. outside of "hidden" dimensions, multi-universes and the like, QM doesn't seem (at a laymen level) to address what the measurements are exactly, less a classic (SR) description.

Which seems to be a dichotomy bridged by probability. And that doesn't seem to "flow naturally".

Perhaps not something that can be described with geometries,