And that just gets into what you call an empirical fact. I can’t see a muon… but I can see a little light light glowing on my muon detector, or a needle twitching, or whatever. I call seeing that light or needle twitch “detecting a muon”, because I’ve constructed a model of how my detector works, and that model includes particles called muons. Is that really so different from having a model which includes virtual particles, and thus saying that they exist?
I think lattice QCD illustrates the point I’m trying to make, to get the predictions of the theory you have to assume a relationship between the expectation values of the field and the physical properties of the particles, which is not derived from first principles. This is because particles are an empirical fact which cannot be ignored and so the predictions of the theory must be put into those terms. What you don’t have to do though is to put the predictions of lattice QCD in terms of virtual particles, this is because virtual particles are a calculational device which can be discarded when it is not useful.
Virtual particles are real and their existence can be easily detected.
See Scientific American 2006: “Are virtual particles really constantly popping in and out of existence? Or are they merely a mathematical bookkeeping device for quantum mechanics?”
http://www.scientificamerican.com/article/are-virtual-particles-rea/
That article can basically be reduced to: “assuming virtual particles are a prediction of QFT [which is the tacit assumption of the article], then they are real because of the empirical success of QFTs”. The objection though is not that QFTs have been empirically successful, it is that virtual particles are a prediction of QFT, so it misses the point entirely.
The information represented may be more fundamental and in one sense more “real” than the particles.
A line I liked in that link:
(So far little experimental evidence for the so-called holographic principle)
If you knew the exact quantum state of the Universe and assumed that its evolution is governed by a suitable wave equation (e.g. the Schrodinger equation), then you can rewind it to any previous time. This kind of evolution of the wavefunction (i.e. quantum state) is called unitary evolution.
The problem though is with the assumption that the evolution is always governed by a wave equation. In basic QM we’re taught that there is a 2nd type of evolution, which happens when we make a measurement: the so-called collapse of the wavefunction, which takes the form of a random projection (which is non-unitary and irreversible, and the probabilities are governed by a postulate called Born’s rule) onto one of the eigenstates of a measurement operator.
An obvious objection to the idea of measurements being somehow different to other physical processes is that the measurement apparatus is itself just a collection of particles governed by QM. Indeed when we start to think this way we can explain the classical correlation between the state of the measurement apparatus and the state of what it is measuring . However what we still can’t explain without evoking the seemingly supernatural random non-unitary evolution of collapse is why we observe only one particular result out of all possible results for the measurement or why the probabilities of getting a particular result should be governed by Born’s rule (unless we go into the realms of non-local hidden variables).
This is basically the measurement problem of quantum mechanics, and though quantum measurement theory can be put on a more advanced footing, the problem remains and nobody has resolved it in a satisfactory way in the almost 100-year history of quantum mechanics.
Sure we can. After all, we’re just collections of particles governed by QM, too. When we say that we observe one particular result, we just mean that the particles of our brains are in a certain state. But those states, too, can be superimposed.
This is basically many worlds, but that has the problems I specifically mentioned, I.e:
If the actual physical state of is a superposition of the observed state and many other states, why do we only see the observed state?
If all the states we could possibly observe are equally real, then why can they have different probabilities for observation are more specifically why do those probabilities follow the Born rule?
I’m not saying different interpretations don’t have their own problems, or that these questions are unanswerable, but there is at the moment no satisfactory answers to those problems.
To answer in many-worlds terms, the me in this world observes this world’s state, and the me in some other world observes that world’s state. I am entangled with the system I’m observing.
Try again as to me this is not responding to what I stated, is perhaps confirming it.
IF I have am standing outside of something that contains all of the information in the universe (which includes me but cannot include me because I must be outside of the information set), then I can use that something to rewind to any previous time …
Let’s say the Universe branches into two ‘worlds’, A and B. The states of the worlds are respectively |A> and |B> and therefore the state of the Universe |U> =
|U> = c[sub]1[/sub]|A> + c[sub]2[/sub]|B>
where c[sub]1[/sub] and c[sub]2[/sub] are complex coefficients.
However, for example, another perfectly valid way of expressing |U> would be:
|U> = (2c[sub]1[/sub]|A> - c[sub]2[/sub]|B>) + (2c[sub]2[/sub]|B> - c[sub]1[/sub]|A>)
So what is special about the first decomposition of |U> compared to the other infinite possible decompositions? Some say decoherence solves this problem by making these contribution of these alternate ‘alternate worlds’ (sic) to the universal wavefunction decay very quickly, but that is predicated on specific assumptions about measurements that are not always true.
The exact quantum state of the Universe is by definition all the information it is possible to know about the Universe. What I’m pointing out is the potential fly in the ointment in the scheme to rewind our exact simulation of the Universe back in time.
The question really isn’t one of practicality, though, but of what the fundamental theories of physics say on the matter. If it were the case that two different configurations A and B yield the same successor configuration C, then information would be destroyed: even in principle there is no way to ‘rewind’ the evolution to find the prior configuration.
According to our current theories, however, this is not the case: a given prior configuration uniquely yields a given successor, via unitary quantum evolution. The reasons for this are deep, but most intuitively framed by saying that the sum of all probabilities for whatever occurring must always equal one (something always happens), both for prior and successor configurations, which, when translated into the proper mathematical formulation, yields unitarity as a necessary requirement.
However, as Asympotically fat notes, a hitch here is the process of measurement: there, it seems indeed to be the case that different prior states can yield the same successor (i.e. two different quantum states can lead to the same measurement outcome). But this is really only a problem if one fails to include the measuring apparatus into the consideration—thus thinking about an open system, not a closed one. In such a case, it’s no wonder that the information about the system after the measurement—after it has interacted with another system—is not enough to reconstruct its prior state, because it has interacted in a certain way with an external system.
If one then includes the measurement apparatus, again one needs to apply the unitary dynamics, and again one will be able to reconstruct the prior state uniquely. Going then to the extremes, i.e. the whole universe, the prediction of quantum theory is that one can, in principle (if not in practice for obvious reasons) rewind its evolution, since the universe is by definition a closed system.
Of course, this then opens up the question of how we ever observe a definite measurement outcome—which is a question that’s simply open in today’s science.
Moreover, the decoherence-based approach to solving this problem already presupposes a preferred (tensor product) decomposition of the degrees of freedom into system, observer, and environment; but this itself corresponds to a choice of a preferred basis, since one could just as well describe things using an entangled basis that doesn’t show this product structure. So this attempt really is fundamentally circular.