Wednesday, September 10, 2014

Lack of interpretation consensus is embarrassing

A reader notes that Elliott Tammaro claims to explain Why Current Interpretations of Quantum Mechanics are Deficient in a new paper:
Quantum mechanics under the Copenhagen interpretation is one of the most experimentally well verified formalisms. However, it is known that the interpretation makes explicit reference to external observation or "measurement." One says that the Copenhagen interpretation suffers from the measurement problem. This deficiency of the interpretation excludes it as a viable fundamental formalism and prevents the use of standard quantum mechanics in discussions of quantum cosmology. ...

We argue, and where possible, demonstrate, that all common interpretations have unresolved deficiencies.
This so-called problem really bugs some cosmologists, but not because of any pressing scientific question. Occasionally you hear someone argue that it matters to one falling into a black hole, but that is just a stupid thought experiment.
At this point in time it appears that a stalemate has been reached with regard to the interpretation of quantum mechanics. Surprisingly, despite the roughly ninety years since its conception, there is currently no single widely accepted interpretation. The variety of interpretations has acted to divide the physics community into camps. For example, one might be a “Bohmian” or an “Everettian” or in the “I shut up and calculate” camp. There is virtually no travel between camps, but there is much in the way of campaigning for new recruits. In addition to being a mere inconvenience, we currently stand at the cusp of physics beyond the standard model and it may be that further advancement will demand a deeper understanding of 20th century physics. It is firmly established that string theory, while still the most promising attempt at unification, does not provide any deeper insight into quantum mechanics.
Cosmologist Sean M. Carroll made a similar argument, and was quoted on Why quantum mechanics is an “embarrassment” to science:
The result? Not a single one of the interpretations could even garner a simple majority vote. Ninety years after the theory was first developed, there's still no consensus on what quantum physics actually means. "I’ll go out on a limb to suggest that the results of this poll should be very embarrassing to physicists," wrote cosmologist Sean Carroll.

(On the plus side, the theory turns out to be very, very, very, very accurate in making experimental predictions. So there's that!)
I suspect that some of the excitement over string theory was that it might proved some sort of hidden variables interpretation to make quantum mechanics more philosophically acceptable. No one seems to think that is possible anymore.

Here is the core of the above objection to Copenhagen:
As a result, the collapse of the wavefunction is assumed nonphysical. We find this view untenable. The wavefunction after collapse represents a radically different physical system than before collapse.

Consider a gambler betting on a horse race. Assume she has some (incomplete) data on each horse. Her bets are distributed according to the data. If she is given new information about the horses, her bets will generically be different. Such is the case with wavefunction collapse in QBism. However, the gambler’s bets have no effect on the outcomes of the races, and as such the analogy breaks down.

Finally, the claim that the collapse is a result of the changing knowledge of the observer (agent) contradicts the well verified dictum that knowing the wavefunction of a system represents a state of complete knowledge of system.
This is a strange philosophical objection. He has some preconceptions about reality and observables, and does not like how they match up in quantum mechanics. In short, he wants an ontic wavefunction.

The argument is to say that the theory explains observations very accurately, but is somehow unsatisfying because it does not explain what is really going on. One could make similar objections to relativity, electromagnetism, gravity, and other theories.

The horse race bet is a little strange. Such bets do affect future bets by others, and maybe the race itself if the rider is crooked.

Yes, of course new info may convince us that the system is in a different state. I am not sure the wavefunction represents complete knowledge, but even if it does, a new measurement or new info can change that knowledge.

Bayesianism is a view of probability and statistics that elicits disapproval from others from similarly philosophical reasons. Yes, probability has more than one interpretation. Tammaro just doesn't like the idea of using data to update a probability.

These anti-Copenhagen attitudes are anti-science. Science is all about gaining knowledge from experiments and observations, and also recognizing the limits of what can be done. Quantum mechanics gives a way of quantifying our knowledge about a system. Nothing wrong with that. Some people want the wavefunction to do more than that. Sorry, not possible, as far as we know.

I have argued that QBism is essentially the same as Copenhagen as articulated by Bohr, in spite of Mermin's protests. Tammaro agrees:
QBism shares so much in common with the Copenhagen interpretation that it cannot rightfully be called a distinct interpretation. In particular, it uses a notion of measurement that corresponds precisely to that of the Copenhagen interpretation. No refinement in understanding of the measurement process is introduced. That is, there is no attempt at describing measurement in terms of more fundamental processes.
But Mermin has a new paper arguing that QBism something different, even tho he supports QBism with quotes from Bohr, Heisenberg, and Schroedinger.
Copenhagen, as expounded by Heisenberg and Peierls, holds that quantum states encapsulate “our knowledge”. This has a QBist flavor to it. But it is subject to John Bell’s famous objection: Whose knowledge? Knowledge about what?28 QBism replaces “knowledge” with “belief”. Unlike “knowledge”, which implies something underlying it that is known, “belief” emphasizes a believer, in this case the user of quantum mechanics.

A very important difference of QBism, not only from Copenhagen, but from virtually all other ways of looking at science, is the meaning of probability 1 (or 0).31 In Copenhagen quantum mechanics, an outcome that has probability 1 is enforced by an objective mechanism. This was most succinctly put by Einstein, Podolsky and Rosen,32 though they were, notoriously, no fans of Copenhagen. Probability-1 judgments, they held, were backed up by “elements of physical reality”.
Got that? Sorry, but if that is the big difference, it is a pretty trivial philosophical difference. I think that Mermin is embarrassed that he was confused about the Copenhagen interpretation for so long, and spent much of his career saying foolish things about quantum mechanics.

After writing this, I see that Lumo has some sensible commentary on Mermin's paper:
Mermin writes several paragraphs about this alleged "difference" between QBism and Copenhagen. These paragraphs contain the (bizarre) word "user" many, many times, but he completely avoids the word "observer". But foundational discussions on Copenhagen have always contained many copies of the word "observer". Much of Copenhagen is about the "observer". Mermin has either forgotten about this fact or he is demagogically avoiding the word "observer" because he knows that most readers would then realize that Mermin's claims about the "users" and "observers" who have nothing to do with each other are completely silly. That's particularly the case of statements such as
Science is about the interface between the experience of any particular person and the subset of the world that is external to that particular user.
Holy cow, Copenhagen school's theory of quantum mechanics is all about this interface, too. Mermin has used several new words such as "users" and "agents" in order to make old ideas look new. But even when it comes to this amusing trick, he failed in the case of the word "interface". If you look at Wikipedia's entry describing the Heisenberg cut, you will see that the Heisenberg cut is defined as a "hypothetical interface", by almost exactly identical words that Mermin ascribed to QBism above. There is an observer/user who cares, there has to be an external object/world that he observes, and the theory predicts properties he cares about. ...

At any rate, this sport of spitting on the founders of quantum mechanics who got everything that matters correctly is deeply pathetic. If we lived in a scientific world, they would be celebrated and their priority and their victory in all the intellectual confrontations would be universally known. Instead, we are drowning in the mud of ambiguity and downright hostility, even when we listen to people who should know better.

This is not a scientific world.
Motl is right that the founders of quantum mechanics could be called QBist, but Bell would not be.

My only disagreement with Motl is this:
And indeed, the fundamental thesis of QBism – i.e. Copenhagen – that the probabilistic nature of the wave function is intrinsic means that one shouldn't invent any "mechanisms" that would "transform" the wave function into "what we really observe". No extra "mechanism" is needed.
I agree that no extra mechanism is needed, but not that "the probabilistic nature of the wave function is intrinsic". The wave function can be used to derive probabilities, but I am not sure it is even meaningful to say that the wave function itself is intrinsically probabilistic.

2 comments:

  1. "I agree, no extra mechanism is needed," What mechanism? The probability math mechanism? Math is not a 'mechanism', it can not elide to actuality because it is an abstraction, if it could then reification would be science instead of a beggars wish and a logical fallacy.

    Here is my 'interpretation' of the Copenhagen 'interpretation'.
    If you are given a class average to examine, but have no idea what an individual student is and why they don't all have the same grade, or how the grades are weighted, no knowledge that students are actual (not virtual) 'people' with the ability to think and move and interact, or have any idea of why they have grades, or what they were tested on, or how many students there were to begin with, what do you have in terms of any actual understanding of what produced your 'average' number? Claiming that 'probability' or an assumed randomness pulled out of a vacuum is the 'cause' or mechanism determining your average is ludicrous. It does not even matter if the class 'average' is correctly calculated, since without an understanding of the context in which it was produced it has no meaning.


    Any beginning computer programmer could tell you this, because unlike the bullshit artists that mathematicians and physicists have apparently become, a good programmer actually has to understand the problem since they know what happens when you stack your empty assumptions. You get fired.

    Quantum mechanics needs a new start with an actual foundation free from mystical interpretations no one can seem to agree on, not another coat of concealing paint hiding the fact the old girl lost her dubious charms a long time ago.

    ReplyDelete
  2. Most arguments in this article "Why Current Interpretations of Quantum Mechanics are Deficient" are flawed and mainly reflect the author's naive misunderstandings of the concepts : http://settheory.net/decoherence
    I also wrote my own arguments with references to others arguments against the current main interpretations (Bohm, many-worlds, spontaneous collapse) and argue for the mind makes collapse interpretation, which I express in more details than what I found done elsewhere: http://settheory.net/quantum-interpretations

    ReplyDelete