- #176
alxm
Science Advisor
- 1,848
- 9
By analogy to Many-Body perturbation theory, what the virtual-particles-are-real advocates seem to be saying here: It's not the case that there's an exact solution to the Schrödinger equation for a polyelectronic atom or molecule, and a perturbation expansion is not merely a mathematical tool to calculate those mathematically-intractable many-body effects.
Rather, the perturbation series is reality. The way electrons really work is that this many-body problem (which of course exists classically as well), is a force mediated by a bunch of undetectable-but-real 'virtual' interactions (although this concept has absolutely no meaning in the absence of any electrons). Even though you're dealing a non-relativistic, 'standard' QM in the Coulomb gauge, where the coulomb interaction is instantaneous, it really occurs through these 'virtual' interactions. Electrons first interact in pairs, then in triplets, and so forth, and not all at once, because that's what the perturbation series looks like.
I can even draw illustrative diagrams - Goldstone diagrams, according to certain rules, whose topology will tell me what the terms of my series look like:
This 'explains' these 'virtual interactions', since it gives a visual picture of what's going on, physically - it's not just some abstract graph illustrating the mathematics involved. The reason we 'know' this, is because many-body perturbation theory works. It gives the correct result for the true, interacting, system of electrons. It's far from the only way of arriving at the correct result, though. In fact, it's only an approximate method and quite often, it's not the most accurate one. It hasn't predicted anything that can't be predicted by any other method. But - the argument seems to go - it must be describing objective reality because it does work, and how else would you visualize or describe the 'many-body force'?
But why stop there? That's just one example. How about quasi-particles? Or any of the hundreds of other cases in physics where a difficult problem can be made tractable by re-casting it in terms of some fictional-but-easier-to-describe system? Does it make physical sense to say that every periodic function is actually a superposition of plane waves because it can be described that way, mathematically? Or sticking with PT, if you know your stuff, you know that perturbation theory can be applied just as well to classical physics - You can even http://arxiv.org/abs/hep-th/0605061" there as well!
So here's what I'm asking: Do all perturbation expansions describe "real" things? The consequences of that would be absurd and without precedent. But if they don't, why should they enjoy a special status in QFT? The fact that perturbation theory works at describing real-and-observable quantities isn't an argument. Saying you need virtual particles to "explain" things is circular logic - if they're real something is not a full explanation unless they include them. Stating that "X said so." most certainly isn't an argument.
Nobody's denying there are notable physicists who really do believe in the "reality" of virtual particles. It's unfortunate that they assert that opinion as if it were uncontested physical fact in their popular-scientific writings, when it's neither uncontested, nor a matter of actual physics. Worse, they often invoke spurious arguments, such as the Lamb shift or Casimir effect, as 'proof' of their position. And so we have a bunch of people here under the false impression that this is actually physical fact, regurgitating these fallacies and basically appealing to authority. This is stupid. If you're going to debate a philosophical standpoint, you should at least know that you're doing so. And you should present real arguments relevant to the actual debate on the topic, rather than simplistic popular-scientific statements.
http://www.springerlink.com/content/51r27u20u354mh5n/" is an article (from a philosophy of science journal, as it should be) giving an overview of some of the serious arguments for and against virtual particles. The points raised by the virtual-particles-are-real advocates in this thread are notably absent.
Rather, the perturbation series is reality. The way electrons really work is that this many-body problem (which of course exists classically as well), is a force mediated by a bunch of undetectable-but-real 'virtual' interactions (although this concept has absolutely no meaning in the absence of any electrons). Even though you're dealing a non-relativistic, 'standard' QM in the Coulomb gauge, where the coulomb interaction is instantaneous, it really occurs through these 'virtual' interactions. Electrons first interact in pairs, then in triplets, and so forth, and not all at once, because that's what the perturbation series looks like.
I can even draw illustrative diagrams - Goldstone diagrams, according to certain rules, whose topology will tell me what the terms of my series look like:
This 'explains' these 'virtual interactions', since it gives a visual picture of what's going on, physically - it's not just some abstract graph illustrating the mathematics involved. The reason we 'know' this, is because many-body perturbation theory works. It gives the correct result for the true, interacting, system of electrons. It's far from the only way of arriving at the correct result, though. In fact, it's only an approximate method and quite often, it's not the most accurate one. It hasn't predicted anything that can't be predicted by any other method. But - the argument seems to go - it must be describing objective reality because it does work, and how else would you visualize or describe the 'many-body force'?
But why stop there? That's just one example. How about quasi-particles? Or any of the hundreds of other cases in physics where a difficult problem can be made tractable by re-casting it in terms of some fictional-but-easier-to-describe system? Does it make physical sense to say that every periodic function is actually a superposition of plane waves because it can be described that way, mathematically? Or sticking with PT, if you know your stuff, you know that perturbation theory can be applied just as well to classical physics - You can even http://arxiv.org/abs/hep-th/0605061" there as well!
So here's what I'm asking: Do all perturbation expansions describe "real" things? The consequences of that would be absurd and without precedent. But if they don't, why should they enjoy a special status in QFT? The fact that perturbation theory works at describing real-and-observable quantities isn't an argument. Saying you need virtual particles to "explain" things is circular logic - if they're real something is not a full explanation unless they include them. Stating that "X said so." most certainly isn't an argument.
Nobody's denying there are notable physicists who really do believe in the "reality" of virtual particles. It's unfortunate that they assert that opinion as if it were uncontested physical fact in their popular-scientific writings, when it's neither uncontested, nor a matter of actual physics. Worse, they often invoke spurious arguments, such as the Lamb shift or Casimir effect, as 'proof' of their position. And so we have a bunch of people here under the false impression that this is actually physical fact, regurgitating these fallacies and basically appealing to authority. This is stupid. If you're going to debate a philosophical standpoint, you should at least know that you're doing so. And you should present real arguments relevant to the actual debate on the topic, rather than simplistic popular-scientific statements.
http://www.springerlink.com/content/51r27u20u354mh5n/" is an article (from a philosophy of science journal, as it should be) giving an overview of some of the serious arguments for and against virtual particles. The points raised by the virtual-particles-are-real advocates in this thread are notably absent.
Last edited by a moderator: