Today I want to talk about a peculiar aspect of quantum measurements that you may have heard of. It’s that the measurement does not merely reveal a property that previously existed, but that the act of measuring makes that property real. So when Donald Trump claims that not testing people for COVID means there will be fewer cases, rather than just fewer cases you know about, then that demonstrates his deep knowledge of quantum mechanics.
This special role of the measurement process is an aspect of quantum mechanics that Einstein worried about profoundly. He thought it could not possibly be correct. He reportedly summed up the question by asking whether the moon is there when nobody looks, implying that, certainly, the question is absurd. Common sense says “yes” because what does the moon care if someone looks at it. But quantum mechanics says “no”.
In quantum mechanics, the act of observation has special relevance. As long as you don’t look, you don’t know if something is there or just exactly what its properties are. Quantum mechanics, therefore, requires us to rethink what we even mean by “reality”. And that’s why they say it’s strange and weird and you can’t understand it and so on.
Now, Einstein’s remark about the moon is frequently quoted but it’s somewhat misleading because there are other ways of telling whether the moon is there that do not require looking at it in the sense of actually seeing it with our own eyes. We know that the moon is there, for example, because its gravitational pull causes tides. So the word “looking” actually refers to any act of observation.
You could say, but well, we know that quantum mechanics is a theory that is relevant only for small things, so it does not apply to viruses and certainly not to the moon. But well, it’s not so simple. Think of Schrödinger’s cat.
Erwin Schrödinger’s thought experiment with the cat demonstrates that quantum effects for single particles can have macroscopic consequences. Schrödinger said, let us take a single atom which can undergo nuclear decay. Nuclear decay is a real quantum effect. You cannot predict just exactly when it happens, you can only say it happens with a certain probability in a certain amount of time. Before you measure the decay, according to quantum mechanics, the atom is both decayed and not decayed. Physicists say, it is in a “superposition” of these states. Please watch my earlier video for more details about what superpositions are.
But then, Schrödinger says, you can take the information from the nuclear decay and amplify it. He suggested that the nuclear decay could releases a toxic substance. So if you put the cat in a box with the toxin device triggered by nuclear decay, is the cat alive or is it dead if you have not opened the box?
Well, it seems that the cat is somehow both, dead and alive, just like the atom is both decayed and not decayed. And, oddly enough, getting an answer to the question seems to depend on the very human act of making an observation. It is for this reason that people used to think consciousness has something to do with quantum mechanics.
This was something which confused physicists a lot in the early days of quantum mechanics, but this confusion has luckily been resolved, at least almost. First, we now understand that it is irrelevant whether a person does the observation in quantum mechanics. It could as well be an apparatus. So, consciousness is out of the picture. And we also understand that it is really not the observation that is the relevant part but the measurement itself. Things happen when the particle hits the detector, not when the detector spits out a number.
But that brings up the question what is a measurement in quantum mechanics? A measurement is the very act of amplifying the subtle quantum signal and creating a definite outcome. It happens because if the particle hits the detector it has to interact with a lot of other particles. Once this happens, the quantum effects are destroyed.
And here is the important thing. A measurement is not the only way that the quantum system can interact with many particles. Indeed, most particles interact with other particles all the time, just because there is air and radiation around us and there are constantly particles banging into each other. And this also destroys quantum effects, regardless of whether anyone actually measures any of it.
This process in which many particles lose their quantum effects is called “decoherence” because quantum effects come from the “coherence” of states in a superposition. Coherence just means these state which are in a superposition are all alike. But if the superposition interacts with a lot of other particles, this alikeness is screwed up, and with that the quantum effects disappear.
If you look at the numbers you find that decoherence happens enormously quickly, and it happens more quickly the larger the system and the more it interacts. A few particles in vacuum can maintain their quantum effects for a long time. A cat in a box, however, decoheres so quickly there isn’t even a name for that tiny fraction of a second. For all practical purposes, therefore, you can say that cats do not exist in quantum superpositions. They are either dead or alive. In Schrödinger’s thought experiment, the decoherence actually happens already when the toxin is released, so the superposition is never passed on to the cat to begin with.
Now what’s with viruses? Viruses are not actually that large. In fact, some simple viruses have been brought into quantum superpositions. But these quantum superpositions disappear incredibly quickly. And again, that’s due to decoherence. That’s what makes these experiments so difficult. If it was easy to keep large systems in quantum states, we would already be using quantum computers!
So, to summarize. The moon is there regardless of whether you look, and Schrödinger’s cat is either dead or alive regardless of whether you open the box, because the moon and the cat are both large objects that constantly interact with many other particles. And people either have a virus or they don’t, regardless of whether you actually test them.
Having said that, quantum mechanics has left us with a problem that so far has not been resolved. The problem is that decoherence explains why quantum effects go away in a measurement. But it does not explain how to make sense of the probabilities in quantum mechanics for single particles. Because the probabilities seem to suddenly change once you measure the particle. Before measurement, quantum mechanics may have said it would be in the left detector with 50% probability. After measurement, the probability is either 0% of 100%. And decoherence does not explain how this happens. This is known as the measurement problem in quantum mechanics.
Dr. S.,
ReplyDeleteYou may not be exactly correct about President Trump's deep understanding of QM, although I'd guess the gentleman has as much as Schrödinger's cat.
In debating some other issues, he might have more significant support.
Good Karma to all.
You have deeply insulted cats. Cats have considerable sense of spatial reasoning. They are able to work their way out of things, such as figuring out how to open a latched cage door. Dogs on the other hand have no spatial reasoning; just note that a single dog will get wound up in a chain and not figure if they just retraced their steps they would be free. Dogs though have a deep social intelligence and emotional basis.
DeleteIt is strange how cats become the butt of these sorts of things. Ambrose Bierce said a cat exists for us to kick when affairs went bad. Schrodinger seemed to extend this to the physics world. I mean, we do not talk about Schrodinger's dog.
The Schrodinger cat argument was made to illustrate something that is absurd, the idea cats can be both alive and dead in a superposition.
well, Wolves can cooperate to pull on 2 separate ropes simultaneously, to pull a tray of food toward themselves.
Deletehttps://www.nytimes.com/2017/11/07/science/wolves-dogs-cooperation.html
Unclear whether that implies any spatial reasoning. Perhaps it's no different than 4 wolves instinctively pulling at 4 legs of their prey, to tear apart the hapless victim.
Regarding the last paragraph, haven't you just reasoned a solution for the measurement problem? The time one can maintain a quantum state upon interaction with something (like a detector, or a double slit or anything) is extremely short. In fact, probably a lot shorter then any thinkable detector time-resolution. Hence, decoherence always occurs too quick to observe, and you get 0% or 100% - unless maybe one can build a slowly scaling detector, where a single particle affects really only two others, and so on until its finally measurable.
ReplyDeleteRorokpal,
DeleteThe problem is that the outcome of a decoherence process is *not* 0 or 100%. Instead, it's (generally) something in between, which makes no sense for single particles.
Decoherence establishes decoherence sets, where we can in a coarse grained setting think of the density matrix of our system stripped from off-diagonal terms corresponding tow quantum phases. The remaining diagonal terms just give what amount to classical-like probabilities. This does not give a formalism for predicting an actual outcome. How outcomes of decoherence or measurements happen is unknown.
DeleteThe stripping away of off diagonal terms in a density matrix means that Tr(ρ) = Tr(ρ^n) no longer applies, for Tr(ρ) = sum of probabilities = 1. As a funny sideline, when I think of stripping away off-diagonal terms from the density matrix the Frank Zappa song Montana Dental Floss Ranching comes to mind with the part about waxing them down. Decoherence "waxes" the density matrix.
Sabine, thank you. Your essay provides a precise and beautifully clarifying explanation of how observation works in physics. I especially liked how you constructed a decoherence framework from which to make your final point: the deepest mystery is the source of the final decision that must occur during any given observation. Blaming it all on a naturally occurring quantum version of a random number generator is insufficient, and frankly uncomfortably akin to invoking angels to push planets around in precise orbits. I've written programs to emulate random number generation, and they are anything but trivial. More work, and more skepticism of the casual invocation of magic mathematical functions to "explain" randomness, is needed.
ReplyDeleteCovid-update. With myself I had a prolonged period of fatigue and back in April a funny feeling as if my brain had been rewired. With Don-the-Con t’Rump I figured you had to be kidding when you said this reflects a deep understanding of QM by him. I think there are species of flies that have a longer attention span than t’Rump has. Experiments have been done with Buckminsterfullerenes which show they can have quantum mechanical properties. Many viruses are similarly polytopes in the protein coats, which means that if cooled to supercold temperatures, as was done with Buckminsterfullerenes, they might actually have quantum properties. However, in a warm wet environment of a human body, not a chance.
ReplyDeleteA quantum observer needs to be any conscious entity at all. All that is required for a system to act as a quantum observer is for the it to couple to a pure quantum state so the quantum phase of that pure state is completely transferred to the large N number of mixed quantum states composing this system. This serves as the collapse. Further this quantum system with frequency ν will enter this state of affairs on a time scale T << 1/ν and so the system never executes its quantum oscillations. The quantum properties or phase of a system is taken up into a measurement apparatus. The phase previously a superposition or entanglement phase of a small N-particle system is taken up into a very large N-particle system. This becomes intractable and we can say the wave function has collapsed within a coarse grained setting.
There is no need for a mentally conscious being. A biological system, whether a purring cat or a person is a finite non-zero temperature entity filled with quantum noise due to its thermal properties. As such the Schrödinger cat is not possible. There is no way a cat can be in a superposition of states, or at least not the entire thing. A measurement or decoherent event is what converts a quantum eigenvalue into a classical variable. The old photomultiplier tube used in nuclear and particle physics amplifies the excitation of a grid element to convert the quantum state of that particle into a classical readout. There are pieces of flint with tracks of cosmic rays, usually secondary muons, from millions or even billions of years ago. The rock provided the measurement or decoherence of a quantum state. The geologist who finds the piece did not somehow collapse the flint wave function by looking at it under a microscope. There is no need for there to be any actual awareness or knowing of the measurement or decoherence outcome.
I think it is possible the answer to the measurement problem is there is no answer. A measurement is a sort of quantum self-reference, where quantum states are encoded within qubits of a system. The system → system plus apparatus is a sort of Gödel numbering of quantum states and this has some level of undecidability. There may simply be no possible solution to the measurement problem.
I'm not a physicist, but I like the idea that if gravity is quantized, there is a minimum unit of gravity. Maybe superposition spontaneously collapses if the superposed gravitational fields differ by more than a gravitational unit; i.e. there is only so much "gravitational strain" that superposition can support.
DeleteGravitons can have different wavelengths and thus different energy. The low energy weak interacting gravitons are approximately linear and may form superpositions more readily. High energy gravitons may not. Curvature in a cryptic representation are R ~ ∇Γ + ΓΓ, where the dependency on constants is Γ ~ G/c^2, and Γ varies with energy or momentum as E^2 or p^2, which means terms quadratic in Γ or Γ^2 are going to be smaller and contribute significantly only at high energy.
DeleteThis is where things get a bit odd. For physics should still function at high energy. In fact quantum gravitation, which largely occurs at high energy, should be dual to the source that generates them. In other words
Quantum gravity at UV = Quantum fields at IR.
In some ways this is the Einstein field equation. With the AdS/CFT duality we have nonlocal field theoretic gravitation in the anti-de Sitter (AdS) bulk equivalent to the local quantum field theory on the boundary as the conformal field theory (CFT). The nonlocal bulk q-gravity by some mechanism then must encode quantum information corresponding to the IR local CFT on the boundary.
@ Sabine,
ReplyDeleteYou write: "A measurement is the very act of amplifying the subtle quantum signal and creating a definite outcome. It happens because if the particle hits the detector it has to interact with a lot of other particles. Once this happens, the quantum effects are destroyed."
This is a view from the old quantum theory (1900-1925)! In the new quantum theory (1925+) observables are fundamental and not defined. Your approach doesn't take into account the Kochen-Specker Theorem!
It's not "my approach" it's a summary of decoherence theory, and if you think KS disagrees with it, you misunderstood KS.
Delete@ Sabine,
DeleteYou're right! You were presenting the standard theoretical physicists approach. That approach doesn't require coherence or consistency; these are traded-off for visualizability and computational ease and power. This is proved by the continued lack of satisfaction of even theoretical physicists for the meaning of quantum theory. (For example, Feynman's statement that no one understands quantum theory.) My approach is a synthesis of Bohr and von Neumann without all their confusions.
Bohr and von Neumann, eh? Well good for you, at least it's less tired than the kinship claims to Galileo, Einstein and Feynman.
DeleteQuantum Mechanics is the anthropocentric view of a world where humans possibly mean nothing. People that write computing code used to say that a binary variable is undefined, null, or set to 1 or 0. A variable is undefined when its name is not present in the code script. A variable is null when its name, although present, has not assigned any value, so its content is empty. A variable is set when its name exists and you let it be certain value. In quantum computing, you can set a qubit into certain binary probability, but classically that binary variable is seen just as null, although may be stochastically biased. The probability that the Moon does not exist when you eventually look at it equals the probability that you are dead when the Moon doesn't cast its shadow on you in a summer night. Antianthropocentric models tell us that you are dead when nature can't apply (observation) its laws on you. So you, as a dead, are now disolving yourself (undefining) and homecoming to the Nothing Sea you were into before you were born. An atom in not decayed and not decayed before its decay measurement, it is just its decay state is anthropocentrically null. That atom already exist as an undefined entity, although with a null decay status. The measurement problem becomes a non existing problem if your code is correctly written.
ReplyDelete..
Albert,
DeleteQM does seem anthropocentric in concept, but not the relevance of the application you give.
For me, the value a variable is assigned is usually less significant than its allowable range, particularly when combined probabilistically in groups of variables.
When I write my codes correctly, as (hopefully) when doing Monte Carlo analysis of possible configurations, the measurement problem not only exists, its result, or optimal assignments of all values, is what I try to predict.
Best Regards, Bert
Pat,
DeleteApplying Monte Carlo codes in optimal system design says nothing about QM questions. MC doesn't care if final packet collapse is indeterminate or won't occur without direct human knowledge. MC techniques just use deterministic methods to predict the probable result if there is one. At least, that's been my approach.
Best regards, Bert
Sabine,
ReplyDeleteIn regards to photons have we entirely ruled out there being a basic asymmetry between emission and absorption processes; with emission being the true quantum process and absorption being more of a integral process, perhaps integrating photonic wave energy that has spread out classically with distance or even simultaneously integrating wave energy (of the same frequency) from two separate sources that combine somehow to produce singular emitted photons.
There is no known mechanism to allow photons of a particular frequency to spread out like classical waves and to be reabsorbed over time with emission taking place once the stored energy reaches hv. However there have been some quantum noise experiments documented on-line where asymmetries between emission and absorption have been observed; these may tentatively show that absorbed energy only could have been simultaneously integrated (my interpretation) from more than one source prior to emission.
Are you aware of any recent or historical published theoretical investigations investigating such a possibility?
Some years back I read a biophysics book, Life's Ratchet. It described how cellular machinery, myosin in particular, introduces a time asymmetry which enables cellular metabolism. Myosins carry other chemicals on tracks inside of cells. The chemical storm within a cell can drive reactions forward or backward with more or less equal probability. It is only an asymmetric reaction, one with high forward probability but low backward probability, a toggle reaction, that lets life move forward. In the case of myosins, it is the conversion of ATP to ADP at an effective temperature, according to biophysicists, of about 7000K that ratchets life forward.
ReplyDeleteAt this scale, it is chemical probabilities, not quantum effects that enforce a forward arrow of time, but I'm wondering if this kind and perhaps scale of asymmetry is close to where the quantum world meets the macroscopic world.
P.S. I forgot to point out that in quantum computing, it takes energy to erase information. That ATP->ADP energy release lets the myosin forget its reversible state.
ReplyDeleteSabine,
ReplyDelete“This special role of the measurement process it an aspect of quantum mechanics that Einstein worried about profoundly. He thought it could not possibly be correct. He reportedly summed up the question by asking whether the moon is there when nobody looks, implying that, certainly, the question is absurd.”
Einstein didn’t just “worry” about this, he proved that unless one rejects locality, the measurement results are predetermined. Given the large body of evidence for locality we can conclude that we knew since 1935 that the moon is there when nobody looks.
“Schrödinger says, you can take the information from the nuclear decay and amplify it. He suggested that the nuclear decay could releases a toxic substance. So if you put the cat in a box with the toxin device triggered by nuclear decay, is the cat alive or is it dead if you have not opened the box?”
Just like we “observe” the moon by observing the effects of its gravitational field we also “observe” the cat inside the box by observing the effects of its gravitational field. So, the assumption that you can put large objects such as cats into superpositions by using a box is false. The only superpositions that can be created are those based on the uncertainty limit. And that limit is the same with or without a box. For large objects, the uncertainty in position is practically 0. There is no box that can shield a gravitational field, or even an electric or magnetic field, so all those experiments based on Schrödinger’s cat or Wigner’s friend are undoable, even in principle, so they can teach us nothing.
“The moon is there regardless of whether you look, and Schrödinger’s cat is either dead or alive regardless of whether you open the box, because the moon and the cat are both large objects that constantly interact with many other particles.”
This is also true for an electron. An electron always interacts with all other electrons and quarks in our universe by electric and magnetic fields. There is no way to “switch-off” the field of an electron, there is no way to “isolate” an electron. The reason the so-called “quantum effects” can be noticed for an electron is that, giving its low mass, the associated uncertainty is large.
This is a brilliant reply. If you are right, then SC/WF are not possible even in principle, as they would require the creation of uncertainty "blurriness" in macroscopic objects, which would be impossible. I have looked for papers that discuss the impossibility of SC/WF but can't find any -- can you suggest one? Thanks!
DeleteI am wondering, if in large objects, the moon for example, if the universe, is itself, taking measurements constantly, with gravitational forces, and making the moon real. So the moon comparison of quantum mechanics isn't a real valid comparison.
ReplyDeleteThe problem with this argument is that we don't have a theory of quantum gravity, so no one knows how to make this formally precise.
DeleteRoger Penrose in his Road to Reality states something of this form. Penrose favors the idea of what he calls an R-process, where gravitation demolishes quantum information. His idea is interesting, though I think is is possible that gravitation may not so much demolish quantum information as to convert it to another form. That form might be in the form of IR gravitons or BMS symmetries. Which ever is the case Penrose argues that gravitation couples to a quantum wave to induce a collapse. It is an interesting idea. If this is not fundamental it might have some phenomenology with weak or emergent quantum gravitation.
DeleteI think even short of full quantum gravity, there are some interesting things we can wonder about it. My understanding was that there's an interesting question on whether or not, once we can get the coldness and noise under control to get a system of order Plank mass into a coherent quantum state, its gravitation would cause it to decohere. Though I'm fuzzy on the specifics of what we'd hope to observationally distinguish in this regime - I have a vague sense that we have interesting things to experimentally discover, but it's largely a 'grass is greener with more imminent results in fields I did not go into' sort of feeling.
DeleteQuantum gravitation in the UV limit is in a funny way "self-decohering," and is then not quantum in the standard meaning of the term. See above with my comment to Castaldo ( Dr. A.M. Castaldo1:56 PM, June 29, 2020 ) about some of this. Gravitation is no unitary in the usual meaning because as a gauge-like theory it is not Lorentzian. By some means quantum gravitation is a quantum error correction system if qubits are conserved.
DeleteOh boy, Trump ... I hope he goes a way soon and we do not have to talk about him.
ReplyDeleteBut meanwhile, I found a typo:
"it an aspect" -> "is an aspect".
Ugis,
DeleteThanks for spotting this, I have fixed it!
Ugis,
DeleteBe careful what you hope for.
Best regards to all, Bert
This comment has been removed by the author.
DeleteI'm happy to learn that Trump has such a deep understanding of quantum phenomena. I suppose that gives fodder for long deep conversations with Merkel.
ReplyDeleteSabine wrote: "In quantum mechanics, the act of observation has special relevance."
ReplyDeleteWhy should anybody accept this? If you believe that quantum mechanics is about "measurements", then it cannot be a fundamental theory. John Bell wrote:
"And does not any analysis of measurement require concepts more fundamental than measurement? And should not the fundamental theory be about these more fundamental concepts?"
(Quantum mechanics for cosmologists)
Rather than dismissing QM as not fundamental, shouldn't we search for Bell's more fundamental concepts? The last century witnessed a hunt for ever more fundamental particles. But "particle" is a classical idea (even if shrouded in uncertainty relations). At bottom, couldn't there be something more fundamental than ever smaller particles?
> "explain how to make sense of the probabilities"
Why is it so difficult to accept quantum theory as the statistical theory that it obviously is? Quantum theory is the most successful theory that we have. It is painful to see it badly caricatured as a theory of "quantum particles" with unreal properties, and its statistical character hidden behind an entirely fictitious process: the "collapse of the wave function". I think it is a serious misrepresentation of quantum theory (even if it mostly echoes textbook wisdom!) and a disservice to your audience. There is a deep problem here, but it is not quite where you suspect it to be. Please, carry on!
Best wishes,
Werner
"Why is it so difficult to accept quantum theory as the statistical theory that it obviously is?"
DeleteStatistics of what do you think it is?
@ Werner,
DeleteYou write:
"If you believe that quantum mechanics is about "measurements", then it cannot be a fundamental theory."
This depends upon a metaphysical assumption prioritizing physicalism over perspectivism! But Quantum Theory is best understood as a form of perspectivism not physicalism.
Prof. Edwards,
Deleteof course you can define "measurement" (or "perspective")
as "fundamental". Nice if it helps you to understand quantum theory better. For me it has the opposite effect.
Sabine asked: "Statistics of what do you think [quantum theory] is?"
DeleteStatistics of events, i.e. points in spacetime. Or in a higher-dimensional manifold. For example, in QED each event would not only be characterized by its space and time coordinates, but also by a phase factor (a la Kaluza-Klein). Fields describe correlations between events, and QFT is a machinery for calculating correlation functions.
One must not think of the absorption of a photon as happening in an instant. It is a composite event (a pattern). A glance at the Kubo formula for absorption coefficients indicates that there are two parts separated in time on the order of femtoseconds (for visible light).
[I'm sending this again, because my previous reply, which was sent before the reply to Prof. Edwards, was apparently lost.]
"... we now understand that it is irrelevant whether a person does the observation in quantum mechanics. It could as well be an apparatus."
ReplyDeleteWhen, and how, has this been convincingly shown to be true?
Well then, Jim.
DeleteIf we remove your term 'convincingly ',
and replace it with the
term. 'scientifically',
Then , no.
It hasn't been proved.
.
My question is where did you come up with the term
' We Now understand '
I quoted the blog post.
DeleteI'm sorry that I didn't keep the reference where I can find it, but some years ago I saw an article about a two-slit experiment run completely by computer, with no human observer involved in either detection, choice of whether to detect or not detect which slit a photon passed through, or analysis of results. In fact, the original observations were erased before any human had access to them. According to that article, the computer found the same results that humans do in such experiments.
DeleteOf course, some people will probably conclude the computer was subconsciously controlled by human minds somehow, but the experiment sounded convincing to me at the time I read it. That is, the simplest assumption seemed to be that human minds are the product of physical laws, not the creator of them.
Perhaps with some googling you can find the paper which the article cited, or a similar one. Good luck.
As JimV says: You can now literally do experiments without humans involved. We even have experiments that were designed by robots in the first place (I tell this story in my book). The outcome, well, is the same whether or not a human was in the room. Did you actually expect something different?
DeleteWait a minute. I'm a software guy and I'm raising my hand here. The computer "erased" its results, but somehow we know what they were because we know they're the same as what a human obtained - or would have obtained...?
DeleteAssuming you mean "Lost In Math" I'll re-read the relevant section. But I don't recall being convinced the firat time.
DeleteYou erase the individual slit-choice observations and keep the distribution, which a human will then check. The distribution looks like the one you get when an observer observes at the individual slits.
DeleteI believe I said it erased the original observation data, not the results of those observations. All the computer had to report was that when it didn't try to detect which slot photons went through, it got interference patterns, and when it did, it didn't get interference patterns--without any mind except possibly an omniscient one seeing either the photon detections or the resulting patterns. But please, and as always, don't depend on my feeble memory and do some googling for yourself. Find the paper or a similar one and read it. I could probably do it for you, but why should I? A software guy should be able to do an Internet search.
DeleteWhy believe me or any random person on the Internet? Always look it up for yourself; and then maybe post the link here. No need to thank the person who pointed you toward it--the link would be paying it forward.
I'll have to think about this some more. My initial reaction is that logically, nothing changes; we're just constructing longer and more complex von Neumann chains.
DeleteCould someone point me to a writeup of this experiment? I've Googled on "double slit computer" etc. without success. And I didn't find it in the book's chapter on QM. Admittedly it's tedious to flip through a Kindle book, but "double slit" isn't in the index.
DeleteI just saw the request this morning (July 10), so I spent about about 45 minutes on it. Note: all arguments against a consciousness role are predicated upon a real world. Arguments from "world in the head", "Boltzmann Brains", or "last Thursdayism" are orthogonal to that, except for the Occam's Razor heuristic. (Personally I use Mario's Sharp Rock, or the argument from humility.)
Deletehttps://arxiv.org/abs/1009.2404
Quantum mechanics needs no consciousness (and the other way around)
Shan Yu, Danko Nikolić
Abstract:
It has been suggested that consciousness plays an important role in quantum mechanics as it is necessary for the collapse of wave function during the measurement. Furthermore, this idea has spawned a symmetrical proposal: a possibility that quantum mechanics explains the emergence of consciousness in the brain. Here we formulated several predictions that follow from this hypothetical relationship and that can be empirically tested. Some of the experimental results that are already available suggest falsification of the first hypothesis. Thus, the suggested link between human consciousness and collapse of wave function does not seem viable. We discuss the constraints implied by the existing evidence on the role that the human observer may play for quantum mechanics and the role that quantum mechanics may play in the observer's consciousness.
---------------------------------------
From https://en.wikipedia.org/wiki/Observer_effect_(physics)
"Of course the introduction of the observer must not be misunderstood to imply that some kind of subjective features are to be brought into the description of nature. The observer has, rather, only the function of registering decisions, i.e., processes in space and time, and it does not matter whether the observer is an apparatus or a human being; but the registration, i.e., the transition from the "possible" to the "actual," is absolutely necessary here and cannot be omitted from the interpretation of quantum theory." - Werner Heisenberg, Physics and Philosophy, p. 137
"Was the wave function waiting to jump for thousands of millions of years until a single-celled living creature appeared? Or did it have to wait a little longer for some highly qualified measurer - with a PhD?" -John Stewart Bell, 1981, Quantum Mechanics for Cosmologists.
According to standard quantum mechanics, it is a matter of complete indifference whether the experimenters stay around to watch their experiment, or instead leave the room and delegate observing to an inanimate apparatus which amplifies the microscopic events to macroscopic measurements and records them by a time-irreversible process (Bell, John (2004). Speakable and Unspeakable in Quantum Mechanics
As Richard Feynman put it: "Nature does not know what you are looking at, and she behaves the way she is going to behave whether you bother to take down the data or not." (Feynman, Richard (2015). The Feynman Lectures on Physics, Vol. III.
The Yu and Nicolic paper is incorrect. My paper points out the mistake: https://arxiv.org/pdf/2005.13317.pdf
DeleteAbstract: The assertion by Yu and Nikolic that the delayed choice quantum eraser experiment of Kim et al. empirically falsifies the consciousness-causes-collapse hypothesis of quantum mechanics is based on the unfounded and false assumption that the failure of a quantum wave function to collapse implies the appearance of a visible interference pattern.
Another candidate: Decoherence of matter waves by thermal emission of radiation.
Deletehttps://www.nature.com/articles/nature02276
They do the equivalent of the double-slit experiment with C70 and show that by heating up the molecule, the contrast of the interference fringes gets reduced then disappears, because the emission of a photon by the molecule could give which-way information. (However, this information is not gathered by the experimenters, and hence not known to any conscious entity, barring omniscient ones.)
(I understand though that the solipsism point of view cannot be disproven except heuristically.)
From the abstract:
Large molecules are particularly suitable for the investigation of the quantum–classical transition because they can store much energy in numerous internal degrees of freedom; the internal energy can be converted into thermal radiation and thus induce decoherence. Here we report matter wave interferometer experiments in which C70 molecules lose their quantum behaviour by thermal emission of radiation. We find good quantitative agreement between our experimental observations and microscopic decoherence theory. Decoherence by emission of thermal radiation is a general mechanism that should be relevant to all macroscopic bodies.
Thanks for the link to the Yu/Nikolic paper; unfortunately it's too technical for a layman like myself. I note however that the word "computer" doesn't appear in it.
DeleteSabine,
ReplyDeleteThe more I think the more I am sure that decoherence is just a neat - but very deceptive - way of sweeping the problem under the carpet. The problem is still alive and bites, as you seem to say in the last paragraph.
,,The moon is there regardless of whether you look, and Schrödinger’s cat is either dead or alive regardless of whether you open the box, because the moon and the cat are both large objects that constantly interact with many other particles.''
There is more faith than physics in these statements (as always when we think about really deep problems, nothing wrong with that, even math needs some faith). You can not prove that using only mathematics of decoherence, at the end of the day you still need to make some very strong assumptions, philosophical ones, about consciousness etc.
And even ,,single'' electron constantly interacts with many other particles - virtual ones but still. What is wrong with this counterargument?
Here we have the same problem: what makes a/the particle real? A definite outcome in some mind, or interaction with other ,,real'' particles?
BR,
Wojciech
"So when Donald Trump claims that not testing people for COVID means there will be fewer cases, rather than just fewer cases you know about, then that demonstrates his deep knowledge of quantum mechanics."
ReplyDeleteJust GOLD. Love your sense of humor.
"when Donald Trump claims that not testing people for COVID means there will be fewer cases,"
DeleteIs it not true there will be fewer _REPORTED_ cases? That's what virtually all news media report, even if it's not explicitly stated.
Epidemiologists will estimate & extrapolate the "number of infections", but majority of TV, web, and print reporters are too severely arithmetic-impaired to understand.
The likely reason he doesn't want to more tests is that fewer positive cases will be reported by the media, yes, I think you understood that correctly. In case you missed the point, people die regardless of whether you know what they died from.
DeleteIn any case, the situation in the USA looks so bad at this point I suspect they'll soon give up testing people other than those who are on an ICU, so they can write Covid in the death certificate.
The situation here is different in every state, due to the current lack of a functional adult as President. Always keep in mind that USA is an acronym.
DeleteSabine, what do you think of the following critique (not mine, I think it's by Earman) of decoherence: it requires a limit in which time t must go to infinite in order to produce the desired result, which doesn't happen for any finite time, no matter how large. It's like the singularities in idealized situations like the Schwarzschild metric, you need more general theorems to prove that in physically meaningful situations the singularities still happen. An analogous theorem seems to be absent in the current mathematical framework of decoherence.
ReplyDeleteAleask,
DeleteThat's just not correct. In the t-> infinity limit you get the off-diagonal elements to strictly zero, but fapp you can't tell exactly zero from 10^-200 anyway, so you don't need t -> infinity. Decoherence is all well and fine, really. There's nothing wrong it it. It just does not solve the measurement problem, as too many people erroneously think. (Though, oddly enough, philosophers tend to understand this point. Not sure how come.)
"measurement does not merely reveal a property that previously existed, but that the act of measuring makes that property real."
ReplyDeleteCan this statement be empirically confirmed? Isn't it just a prejudice?
@ Greg Feild,
DeleteThere's a theorem that the standard quantum logic cannot be embedded in a Boolean Logic. (Corollary of Gleason's Theorem or the Kochen-Specker theorem.) Since QT is empirically highly confirmed, it is reasonable to conclude the above!
Greg Feild wrote: "Can this statement be empirically confirmed?"
DeleteGood question! It depends very much on the interpretation. In the Aspect et al. experiments you can view "polarization" as a property of the detectors as much as a property of the photons.
Pragmatically, I think it better to err on the less spooky side.
Delete@Werner
DeleteCertainly, Aspect et al. showed that local measurement of a (hidden) parameter is not possible. The "measurement" had to be reconstructed in hindsight. However, in some way, the photons were able to reflect the relative position of the detectors. Because of relative, no intrinsic parameter could have been involved. The photons reproduced the expected cos²α statistic (for some people, this was a surprise though). So, in some way every single photon was influenced by the contralateral measurement ( the position of the contralateral polarizer).
Sabine,
ReplyDeletebesides physics, how interesting do you find viruses?
When we talk about Schrödinger’s cat, we should be aware of his motivation to create this thought experiment. It was not Schrödinger’s idea to give a lecture here about quantum mechanics. But his point was that he did not at all like the idea of superposition and he used this picture of a cat experiment to demonstrate how crazy he found this portion of quantum mechanics.
ReplyDeleteAgreed.
DeleteSchrödinger, 1935, Naturwissenschaften: "Die gegenwärtige Situation in der Quantenmechanic," Wikipedia.
I've always wondered about the role of consciousness in QM by Wigner. One place consciousness is used as a foundational concept in reality is in Vedanta. I know that there was quite an upsurge in interest in Indian culture and philosophies in Europe during the 19th Century and it occurs to me whether this is one reason why it occurred to physicists to wonder whether consciousness had a role to play in QM, given that so much of what we took to be foundational in classical physics suddenly looked shaky. According to the Encyclopedia Brittanica, Schrodinger was interested in Vedanta in later life.
ReplyDeleteI think Einstein with his usual acumen got to the root of a profound question in QM, put into focus through the EPR experiment. Given that Bells experiment rules out locality, we're really forced to the conclusion that value-definiteness doesn't hold. This is pretty hard to imagine since everything we think of is value-definite.
Personally, I think in attempting to understand this, it's worth going back to Aristotle where he put forward his theory of potentiality-actuality as his solution to the problem of motion which was put into sharp relief by Zeno.
It's often said, that it was the introduction of infinite series that solved Zenos paradox. But actually, it was already understood by Aristotle - he called it an 'adequate' solution but it doesn't really get to the real roots of the problem of change.
Given, that he ascribed to some sort of ontological indefiniteness, and QM also does, it might be worth looking at his arguments as to why this had to be the case.
If this sounds a little strange, it's worth pointing out that Heisenberg thought along very similar lines. In fact, I think there was a recent article in Nature that probed Heisenberg's take on this, but it didn't really do justice to what Aristotle thought.
"It's often said, that it was the introduction of infinite series that solved Zenos paradox ..."
DeleteIf so, it was Zeno himself who introduced the infinite series, which was part of all his examples. I expect however, that the notion of infinite series was well known to others at the time. His first example, The Arrow, in fact could be used as a proof that the infinite series L/2 + L/4 + L/8 ... sums to L. (It starts with the distance L and subdivides it. Just add the statement that the whole is equal to the sum of its parts and there you are.)
Zeno's issue to me is not whether an infinite sum can have a finite limit, but how any physical summing process in nature can reach that limit, or, how is it possible for nature to get to the end of a series which has no end?
For me, again, the answer to Zeno is the quantum principle, that eventually you get to a smallest quantity which cannot be subdivided. Democritus and his teacher used Zeno's thinking and added a quantum principle to postulate their atomic theory. (Matter can not be subdivided infinitely. Continuous matter is an illusion.) For some reason I see many deprecatory comments about Zeno, but not about Democritus.
(It seems to be one of my missions in life to defend Zeno, for which I will at least pay a double fine in this instance.)
@JimV:
ReplyDeleteIt's an interesting insight to say that Zeno discovered infinite series. I'm not sure Zeno would have put it this way, however. Moreover, Aristotle didn't ascribe the solution by infinite series to Zeno, in fact, he didn't ascribed it to anyone; which suggests as a solution it must have been generally well-known at the time.
Zeno was defending Parmenides rather than any atomic conception of nature. Aristotle wasn't convinced by atomism. Given how we understand atoms now, via quantum theory, it seems his intuition on this was well-founded. I mean here, the lack of value-definiteness, or in his language, potentiality.
Personally, I think Zeno comes in for a lot of stick simply because people think the solution by infinite series solves his so-called paradox, and so dismiss it. Whereas Democritus is associated with the rise of atomic theory in the early 20th Century; and then, wrongly in my opinion, with quantum theory, despite the fact he conceived of motion occurring randomly, his clinamen. That reality, at one level, is somehow value-indefinite, to my mind, strikes me as more basic.
"It's an interesting insight to say that Zeno discovered infinite series."
DeleteI didn't write that or mean that. Zeno certainly constucted examples of infinite series in his examples. That does not prove or suggest that he discovered the concept, any more than Einstein's thought experiments involving trains means that Einstein discovered trains. It does seem to me that the concept that an infinite sum can have a finite limit was already implicit in his example of the Arrow, and therefore does not refute that example.
"Zeno was defending Parmenides rather than any atomic conception of nature."
Who said Zeno was defending any atomic conception? Not me.
I assumed that everyone knew that Zeno was simply pointing out a paradox involving continuous motion. Zeno had never seen a movie or a TV cartoon and didn't think of that way of resolving it, but I firmly believe he was correct to say that the concept of continuous (infinitely divisible) anything raises paradoxes.
Calculus of course assumes infinite divisibility, so my notion of why Zeno is disrespected is that people with a calculus course under their belts believe in continuity implicitly, and since calculus works very well Zeno must have been badly wrong. They do not realize that calculus is applied to many things which are not in fact continuous (heat transfer in a girder, fluid flow, electricity) and is only a good approximation of the underlying reality, not a proof that continuous nature exists. Again, the point that the math of infinite series does not address is how nature physically gets to the end of a series which has no end. (Is nature a mathematician, to manipulate series algebraically as in the Geometric Series proof, or to investigate Taylor Series?) (No mathematican has ever physically summed an infinite series term by term--without stopping somewhere.)
I also thought that people should know that Zeno's thinking was appreciated by Democritus through Democritus's teacher and was an important step in the formulation of Democritus's atomic theory. To me that was a crowning achievement of philosophy, better than anything Aristotle achieved. However, I am not an Aristotle scholar and have heard more about his failures than his successes, so my opinon on that is not important. I am willing to change that opinion upon hearing of things as important as atomic theory which we owe to Aristotle today.
Of course science has developed atomic theory far beyond what Democritus envisioned, and done far, far more work on it, but it helps to have a basic concept in one's mental library.
@JimV:
DeleteYour clarifications are appreciated. One point to bear in mind is that we have much more of what Aristotle wrote than Democritus did. Apparently, Democritus wrote quite a number of books but all we have are maybe a couple of hundred or so fragments, most of which don't deal with atomic theory. What we do know of his work comes from Epicurus and Lucretious - as far as I know. I've actually looked through a book of his fragments but I don't recall him referring to Zeno. Do you possibly have a reference?
I have the same impression as you do with how Aristotle is discussed today outside of specialised scholarship, that is his errors are more often discussed than what he got right. One author that actually discusses his work without at the same time trashing it is Julian Barbour.
I think this is primarily due to the debate that began with Galileo/Copernicus vs Aristotelianism/Christianity. Personally, I think we have gone far enough along this path to appreciate Aristotle for what he said now rather than Co Stanley belittling Aristotelianism for its errors. After all, we don't trash Newton for not discovering electromagnetism, for not finding a mechanism for gravity's action-at-a-distance or for not giving rigorous foundations to calculus.
If I recall correctly, Julian actually credits Aristotle as a precursor to Newton's laws of motions, though I don't think he goes into details. I'd say, that Aristotle's theory of force/change implicitly has Newton's first law of motion, and half of the third law. Actually, I think that after you allow for atoms, it's possible to induce the full third law and the full second law.
I think also, Julian says that we don't know just how Newton came up with his laws. It would be interesting to discover if this is the angle he took, after all, we know he was familiar with both Aristotle and Democritus.
I fully agree with your last point. But I'd add that modern science developed *on* from ancient Greek science. The renaissance was called such because it was a rebirth rather than a birth.
Sabine, I enjoy your blog but am really surprised by your discussions on the measurement problem.
ReplyDeleteProbabilities have 2 interpretations: Frequentist or Bayesian.
If QM probabilities are frequentist, then we must accept the many-world interpretation of QM and its resolution to the measurement problem.
That is, Psi^2 reflects the fractions of worlds where an event was realized.
We both agree that this is non-scientific.
Therefore, we are left with the fact that QM probabilities reflect plausibilities, i.e., they are Bayesian.
All of sudden, if we accept this, there is no spooky-action-at-a-distance
(since probabilities reflect knowledge).
Reductionism is also not necessary as nothing is hiding behind the probabilities. They are fundamental.
What's so hard to understand?
p.s. to all the perpetual-motion mechanics, hidden-variable theorists and internet creepers who mention their teenage daughters in discussions (like PhysicistDave who meets all three criteria--search for word daughter in http://backreaction.blogspot.com/2020/06/physicists-still-lost-in-math.html),
I will ignore all your comments.
Steve,
Delete"If QM probabilities are frequentist, then we must accept the many-world interpretation of QM and its resolution to the measurement problem."
Uhm, no.
"Reductionism is also not necessary as nothing is hiding behind the probabilities. They are fundamental."
The whole concept of Bayesian probabilities only makes sense if someone or something has information about something else. This is just another way of postulating macroscopic concepts. If these are Bayesian probabilities and they are compatible with reductionism, then you should be able to derive them.
You are trying to "solve" the problem by just refusing to define terms.
I think ψ-ontic interpretations are in line with a frequentist interpretation of probability, while ψ-epistemic are more in line with Bayesian statistics. A Copenhagen ideology of QM that a wave function does not exist is parallel to the Bayes theorem where probabilities do not exist independent of an analyst's deductions. The many worlds interpretation is more in line with the existence of probabilities independent of an observers deduction about them.
DeleteSteve Presse,
Delete"Therefore, we are left with the fact that QM probabilities reflect plausibilities, i.e., they are Bayesian.
All of sudden, if we accept this, there is no spooky-action-at-a-distance
(since probabilities reflect knowledge)."
As the EPR experiment proves, QM is either complete and non-local or incomplete (a statistical approximation of a deterministic hidden variable theory). You, on the contrary, claim that by assuming QM is complete (probabilities are fundamental) you save locality. So, your claim is false.
DeleteSabine, you say:
>"The whole concept of Bayesian probabilities only makes sense if someone or something has information about something else. This is just another way of postulating macroscopic concepts. If these are Bayesian probabilities and they are compatible with reductionism, then you should be able to derive them.
You are trying to "solve" the problem by just refusing to define terms. "
Sorry, no. Reductionism is hidden variables. Plain and simple.
We can define terms mathematically rigorously starting from Cox's theorem, but not in an online forum. People can google it if they like.
Andrei, Lawrence,
DeleteLawrence: no working physicist uses words like psi-ontic or psi-epistemic. We do calculations, we don't philosophize ;)
Andrei: You are playing around with the word "complete". QM probabilities are not frequentist (they reflect states of knowledge, cf. Wigner's friend) and QM is local (knowledge does not coincide with supra-luminal communication).
You are not making sense. I explained that waving your hands and yelling "Bayesian" does not solve any problem. In response you say "reductionism is hidden variables" which is (a) nonsense and (b) has nothing to do with my point.
Delete@Steve Presse:
DeletePopper endorsed the frequentist interpretation of probability until he discovered his notion of propensity. He then said propensities are everywhere.
Sabine,
DeleteSteve Presse is a crackpot who thinks that mentioning the word "Bayesian" solves all problems.
Presse gets very abusive towards anyone who explains why Bayesian methods will not do what Presse claims, even though Presse refuses to give any hint at all as to what he has in mind.
And his reference to my daughter is due to the fact that I mentioned that Presse teaches at a "university" that admitted my daughter even though my daughter never finished the application process! Even offered her financial aid. Which is indeed... peculiar.
All the best,
Dave
To non-physicists:
DeleteYou may wonder why Sabine and my views on Bayesianism and QM deserve any more trust than Steve Presse's views.
The answer is that Presse is pushing a view not present in the standard textbooks or the standard literature on the subject. Sabine cannot justify every single thing she mentions here... and she does not need to, because most of it can be found in great detail in the standard textbooks.
Presse, however, is pushing an eccentric personal view and cannot just rely on the standard texts to fill out the details. If he wants to be taken seriously, he owes us an explanation.
The standard QM textbooks generally assume a “frequentist” approach to probability: i.e., do an experiment a billion times and the fraction of the time that you will get the result A is roughly P(A), where P(A) is the probability of A.
That pretty much has to be the case, since, after all, that is what experimenters in QM actually do. QM is intended to predict the results of actual experiments . It does that statistically, which means that when we refer to probability in QM we mean it statistically, i.e., in a frequentist sense.
I actually do not think Presse is smart enough to grasp this. If he did, he would realize that he needs to explain how a Bayesian approach fits with textbook (i.e., experimentally tested) quantum mechanics. (Answer: it does not.)
Steve Presse,
Delete"You are playing around with the word "complete"."
OK, let me define the word "complete". If QM is complete then there is no need for hidden variables. If it is not complete, it has to be completed by something not already present in the formalism, that "something" being those hidden variables.
EPR proved that the only way to explain locally the EPR experiment is to accept that QM is not complete.
Let me repeat the argument here (I've changed the argument to the Bohm version with spin):
At two distant locations (A and B) you can measure the spin, say on X direction of an entangled pair. QM predicts that:
P1: If you measure the X-spin at A you can predict with certainty (probability 1) the X-spin at B.
Let’s exclude non-locality:
P2: the X-spin at B is not determined by the measurement at A.
From P1 and P2 it follows that there was something at B that determined the result of the B measurement. EPR named that “something” an “element of reality”. So:
P3: There is an element of reality at B that determines the measurement result at B.
You may observe that there is no other logical option available (unless you think that it’s by pure luck that we manage to always predict with certainty the position at B, which is rather absurd).
Let's continue:
From P1 and P3 it follows:
P4: there was an element of reality at A that determined the measurement at A.
This is because once we’ve established that the measurement at B was fixed (P3) it’s impossible that the measurement at A could have been different, right?
OK, so P3 and P4 lead to:
C: The X-spin of both particles, at A and B were determined before any measurement took place (deterministic realism).
Please point out how your concept of probabilities (as states of knowledge) is supposed to refute the above argument.
Hi Dave,
DeleteThanks for the head's up. I find it peculiar how Bayesianism seems to have become a miracle fix for supposedly everything.
As an addendum to the above remark on Poppers interpretation of probabilities as propensities, what then should we make of the frequentist interpretation? We should see it, in my opinion, as an experiment to establish the quantitative value of a propensity and which we call a probability.
DeleteIn the kinetic theory of gases probabilities are not poblematic at all. Why is quantum theory different? Taking a pragmatic stance, also in quantum theory can people calculate averages and expectation values, and be happy about agreement with experiment. It is only a small tribe of stubborn physicists who keep quibbling about "ontic" or "epistemic" Psi, and whether the true meaning of probability is frequentist or Bayesian. :-)
DeleteThe source of the confusion is the mantra that "every physical system is described by a wave function", which in some minds has changed to the mistaken belief that "a wave function represents an individual system". While in classical statistical mechanics a system is represented by a point in phase space and averages are taken by integrating over phase space, in quantum theory a trace needs to be taken. And it is deceptive to think of each |ket> as corresponding to an element of the statistical ensemble. But in order to arrive at something measurable, each ket needs to be combined with a bra. And they have the opposite time-dependence, or no time-dendence at all - in the Heisenberg picture. There is no collapse of the wave function in the Heisenberg picture. And neither in the Schrödinger picture, since it is mathematically equivalent. The measurement problem is based on a misunderstanding of the role of the wave function in the formalism. It is a shame that these debates seem to be endless.
Sabine Hossenfelder4:31 AM, July 04, 2020
Delete"I find it peculiar how Bayesianism seems to have become a miracle fix for supposedly everything. "
It's what Luke Barnes uses to disguise his fine-tuning fantasies, too - the favourite tool of cranks everywhere.
"This claim has been challenged on the grounds that the relevant probability measure cannot be justified because it cannot be normalized, and so small probabilities cannot be inferred. We show how fine-tuning can be formulated within the context of Bayesian theory testing (or \emph{model selection}) in the physical sciences."
Re the above discussion on locality:
DeleteThe principle of locality is supposed to mean that an object is directly influenced only by its immediate surroundings. I suppose that this is equivalent to saying that information is only exchanged during particle or other interactions because there is little distance between the objects: is that correct? But if space is not fundamental, i.e. if distance is a derived category of information, then why shouldn’t at least some primitive information relationships be completely independent of distance?
"We show how fine-tuning can be formulated within the context of Bayesian theory ..."-Dr. Barnes.
DeleteSuch an attempt was made, among other arguments, in a guest post on Dr. Sean Carroll's blog by a respected, theistic physicist (not Dr. Barnes), sometime within the last ten years. Bayesian experts such as Bill Jeffries (if I remember correctly) demolished the argument in my opinion.
An example was made of a machine that spits out a ping-pong ball with a number on it. The ball comes from a source of six million balls numbered one to six million, but it is not known how many of these balls were loaded into the machine. The issue was, given that the number 42 comes out of the machine on its first and only trial, which of these is more likely: a) that all six million balls were loaded into the machine; b) that only balls 1-100 were loaded into the machine; or c) that only the ball numbered 42 was loaded into the machine.
The answer is a bit counter-intuitive which somewhat excuses the theists who miscalculate it in support of their innate convictions: unless the separate hypotheses were predicted before the ball was issued, the probability of getting ball 42 was one in six million in all cases, and so strengthens none of them. Making up a hypothesis after seeing the ball and without further trials does not work in rigorous Bayesian analysis.
Any mathematical discipline can be misused and probably will be. The arguments were quite polite and free of genetic fallacies, however, so some progress was made.
Lorraine asked:
Delete>But if space is not fundamental, i.e. if distance is a derived category of information, then why shouldn’t at least some primitive information relationships be completely independent of distance?
Lorraine, lots of physicists have thought of that: that is basically the point of "ER=EPR."
The problem, of course, is that we physicists want to know the details: give us the exact mechanisms, the exact equations that make it work, and, more than anything, give us something that can somehow be tested against alternatives via observations.
No one has succeeded in doing that. To take a concrete example, ER=EPR seems to rest on an extrapolation of the AdS/CFT correspondence, and no one really quite knows what the AdS/CFT correspondence really means in general: i.e., the real world is not AdS space).
So, what you suggest remains an intriguing thought that has occurred to numerous physicists with no one (yet) able to actually work it out in concrete detail.
All the best,
Dave
Thanks Dave, for clearing that up.
DeleteI thought there must be some such reason.
Werner wrote:
Delete>The source of the confusion is the mantra that "every physical system is described by a wave function", which in some minds has changed to the mistaken belief that "a wave function represents an individual system".
Werner, you keep mentioning this as if it has some relevance to the quantum-measurement problem: you have not shown that it has any relevance, and it does not have any relevance.
There are two basic problems in quantum measurement:
A) Bell's theorem. Bell's theorem can be and normally is proven by taking certain very general, plausible locality assumptions and showing that they lead to certain conclusions concerning statistical correlations between outcomes of widely separated experimental measurements.
No need to refer to the wavefunction at all. And normally the proof of Bell's theorem does not refer to the wavefucntion.
It happens that experimental measurements disagree with the conclusion of Bell's theorem. Again, no need to mention the wavefunction at all!
Your obsession over whether the wavefunction refers to a single system is therefore irrelevant.
B) The second problem in quantum measurement is the fact that the standard textbook presentation has certain axioms that relate to “measurement,” separate from the deterministic, linear evolution exhibited by Schrödinger's equation: all the stuff about eigenvalues, Hermitian operators, etc. That is odd, because anything that could be called a measurement apparatus should also be described by quantum mechanics: there should not be any need for additional “measurement axioms.”
Again, the standard textbook presentation makes no claim as to whether the wavefunction refers to an individual system or an ensemble. The textbooks just give a recipe that works to correctly calculate experimental results.
Your obsession with whether the wavefunction applies to individual systems is again irrelevant.
So, why am I bothering to critique Presse's crackpottery? Different issue.
Bayesian analysis is about people's opinions, their subjective beliefs. But QM is about >what actually happens experimentally, not about people's subjective beliefs.
Presse is trying to replace science by subjective belief. I prefer to keep science separate from theology.
PhysicistDave wrote: "no need to refer to the wavefunction at all"
DeleteGreat. So it seems we've found something we agree on. Wavefunctions are not the essence of quantum theory - let's talk about operators then.
> "axioms that relate to 'measurement', separate from the deterministic, linear evolution"
> "there should not be any need for additional 'measurement axioms'"
But you are conceding that these axioms are a necessary part of the formalism. Without them you'd have nothing that relates to the real world. So "measurement" is necessary to break unitary evolution, for something "really" to happen, rather than virtually. Unless I've misunderstood every one Sabine's posts on the measurement problem, this is where the "wavefunction" needs to collapse. It leaves me puzzled that you deny the relevance of wavefunction collapse to the measurement problem. Do you believe in wavefunction collapse or not?
> "Bell's theorem [...] is proven by taking very general, plausible locality assumptions" ...
Yes, I agree that Bell's inequalities are derived under almost compelling assumptions. But the experiments show that at least one of these assumptions must be wrong. There are quite a few to choose from. It could be the idea that something travels from the source to the detectors.
@ Werner,
DeleteThere are no axioms in quantum theory relating to how one does experiments! That requires the wisdom and experience of experimenters. The axioms of quantum theory deal with the structure of the quantum logic. See "Geometry of Quantum Theory", V.S. Varadarajan for the most complete discussion of the mathematics. (But don't listen to his interpretation!)
Werner,
Delete"Yes, I agree that Bell's inequalities are derived under almost compelling assumptions. But the experiments show that at least one of these assumptions must be wrong. There are quite a few to choose from. It could be the idea that something travels from the source to the detectors."
There is no assumption in Bell's theorem that refers to "something" traveling from source to detectors. There are two main assumptions:
1. locality
2. statistical independence (the hidden variable is statistically independent of the measurement choice.
You need to be careful here because Bell's theorem is proposed as a refinement of the EPR argument. The conclusion of EPR was that QM is either non-local, or, if it is local, it has to be incomplete (an approximation of a hidden variable theory). Bell's theorem concludes that the only way to retain locality is to deny the statistical independence assumption.
Prof. Edwards,
Deleteaxiomatization may be a nice pastime for a mathematician, which I am not. I think it should be applied only to theories that are fully understood, and where the basic concepts have crystallized. "Measurement" is certainly no such concept.
Andrei,
Deletehave you ever heard of *implicit* assumptions? There may be many more than two.
Werner,
Delete"have you ever heard of *implicit* assumptions? There may be many more than two."
Yes, but "the idea that something travels from the source to the detectors" is not one of them.
Andrei,
Deleteto anybody who knows Bell's theorem it is obvious that you haven't understood it.
Don't bother to reply. I won't waste my time even to look at your comments.
Bye,
Werner
Werner wrote to me:
Delete>PhysicistDave wrote: "no need to refer to the wavefunction at all"
>[Werner] Great. So it seems we've found something we agree on. Wavefunctions are not the essence of quantum theory - let's talk about operators then.
Well, I guess I should have made clearer the context: I wrote that in the context of proving Bell's theorem – you need no reference to the wavefunction to prove Bell's theorem.
Werner also wrote:
>>[Dave] there should not be any need for additional 'measurement axioms'"
>[Werner] But you are conceding that these axioms are a necessary part of the formalism. Without them you'd have nothing that relates to the real world. So "measurement" is necessary to break unitary evolution, for something "really" to happen, rather than virtually.
No, I am merely conceding that the measurement axioms are part of the standard textbook formalism. Maybe there is some other approach that gets rid of them.
Bohmian mechanics does, you know, although Bohmian mechanics has some unappealing characteristics (notably, the way it handles relativity is really kludgy).
And MWI also claims to, though I think there are some insurmountable technical problems with MWI.
Werner asked me:
>Do you believe in wavefunction collapse or not?
I believe it is part of the standard calculational techniques we teach students and that those techniques work “for all practical purposes.” As to whether it really happens, I do not know.
Werner also wrote:
>Yes, I agree that Bell's inequalities are derived under almost compelling assumptions. But the experiments show that at least one of these assumptions must be wrong. There are quite a few to choose from. It could be the idea that something travels from the source to the detectors.
Yes, of course one of the assumptions must be false. But are you implying that there is not really electromagnetic radiation going from the calcium atom to the detector? Maybe you can make that work via, for example, Feynman-Wheeler theory. I am not sure it saves you from the violation of Bell's theorem, though.
Anyway, I am just saying that your obsession with whether the wavefunction refers to a single particle or to an ensemble is not relevant to these issues. I know that this was a big subject of debate back in the 1930s: one way of expressing Einstein's concerns was to say that the wavefunction gave incomplete information because it referred to an ensemble not each individual particle.
But Einstein's hope that we could have a quantum theory that was both local and realist was dashed by the experiments that demonstrated the violation of Bell's inequality, so that particular debate does not seem to me to be relevant today.
The problem remains to show which (possibly implicit) assumption in Bell's theorem is violated and exactly how that leads to a violation of Bell's inequality.
No one has yet done that in a way that convinces most physicists, I'm afraid.
Dave
PhysicistDave wrote: "you need no reference to the wavefunction to prove Bell's theorem"
DeleteOf course not. Bell's theorem is intriguing, but it is not the heart of the measurement problem. I wasn't interested in proving Bell's theorem, because experiments show that it is irrelevant, at least in this universe.
> "the measurement axioms are part of the standard textbook formalism. Maybe there is some other approach that gets rid of them."
After my first encounter with measurement in QM, I was convinced that Born's rule must be derivable from the Schrödinger equation, and I studied a paper that seemed to show just that. It took me a long time to realize that Born's rule had slipped in right at the beginning with what the authors set out to calculate. It is an independent component of the formalism, and attempts to derive it are a bit like the old efforts to prove Euclid's fifth axiom: at the same time impossible and pointless. We know that the formalism works. So where's the measurement problem?
> "as to whether [wavefunction collapse] really happens, I do not know"
Lame. A measurement is supposed to bring about a definite result, and a collapse would only happen to the individual system that is being measured. If there is no wavefunction collapse, then I'm still in the dark where you see the measurement problem.
> "But are you implying that there is not really electromagnetic radiation going from the calcium atom to the detector?"
I consider myself a realist, and of course there is a sense in which electromagnetic radiation goes from the calcium atom to the detectors. But we cannot expect classical concepts to remain valid to the very smallest scales of space and time. The classical world was filled with continuous fluids and solids with average properties like density, velocity, pressure, density ... Quantum physics has exposed the underlying graininess that is really present. I'm convinced that electromagnetic radiation is discontinuous not only in space, but also in time. I'd rather give up locality (in space and time) than realism.
Concerning graininess I can relate an anecdote that is perhaps as scaring as it is amusing. Back in 1986 (after the Chernobyl desaster) colleagues were on the lawn outside with their gamma-ray detectors. They expected to hear individual clicks, but what they heard instead was a permanent whistle! Our immediate sense perceptions are not always a reliable guide to the underlying reality. And familiar figures of thought can lead us astray. Talk of traveling photons may help to describe the EPR/B experiment, but this kind of "explanation" raises many more questions than it answers if we enquire into the "properties" of the photons and the superlumnial communication between them. It is more reasonable then to put up with a mere description of the observed events, and accept correlations as a fact of nature. In my view this is the real message to take home from the EPR/B experiment.
Werner
Werner wrote to me:
Delete>>[Dave] "as to whether [wavefunction collapse] really happens, I do not know"
>[Werner] Lame.
Perhaps. But true.
I do not know whether wavefunction collapse occurs and neither do you.
The difference is that I am willing to admit that fact, lame or not.
Werner also wrote:
>If there is no wavefunction collapse, then I'm still in the dark where you see the measurement problem.
Yes, you are indeed “in the dark.”
Nonetheless, I and Sabine and Steve Weinberg and numerous others have explained this again and again and again.
All of our measuring apparatuses should, in principle, be describable by quantum mechanics. It should not be necessary to have separate “measurement axioms” to describe these apparatuses, which are themselves just physical systems.
That is weird. Something is wrong here. The Bohmians and the MWI guys have proposed possible solutions, but most of us find those solutions to be inadequate.
It's not that hard to grasp.
Unless you like being “in the dark.”
Werner also wrote:
>I'm convinced that electromagnetic radiation is discontinuous not only in space, but also in time. I'd rather give up locality (in space and time) than realism.
Fine. Then you should embrace Bohmian mechanics, which is non-local. Most physicists find that unacceptable.
Werner also wrote:
>Our immediate sense perceptions are not always a reliable guide to the underlying reality.
For some reason, most physicists think that a fundamental theory of physics should not require discussion of human “sense perceptions.” We do not think we humans are central to the existence of physical reality.
Perhaps you think we are wrong.
Dave
PhysicistDave wrote: "I do not know whether wavefunction collapse occurs and neither do you."
DeleteI'm inclined to read this as a confession that you don't know what the measurement problem is. My question was not about a physical fact, but your personal belief. You carefully misread my question and refused to answer.
> "The difference is that I am willing to admit that fact"
Feynman was honest when he wrote: "I cannot define the real problem, therefore I suspect there's no real problem, but I'm not sure there's no real problem."
By contrast, you pretend to know what the measurement problem is (see your earlier comment, items A and B), but refuse to take a stand on the question whether the formalism really entails "wavefunction collapse". That's pretentious.
> "Steve Weinberg and numerous others have explained this again and again and again."
Lacking physical arguments you appeal to authority. Shall we pit Freeman Dyson against Steve Weinberg? Silly.
I'd love to read a reply having some actual substance. More probably you'll just get upset and (yet again) make a fool of yourself.
ReplyDeleteHere are interpretations of quantum processes that do not convince me; for example, when a coin is thrown the result is probabilistic; but if I toss two coins at the same time and opposite results always come out with the coins being separated enough not to influence each other; so the phenomenon is not probabilistic, there is something deeper in the system that favors the result. The other problem is the statistical one; for example, if I create a drawing of a cat using points with a pencil; When I made the first point it was part of the plan even though there was no cat drawn yet.
Hi,
ReplyDeleteSorry to be so naive, but if we consider the Schroedinger box as being completely isolated from the outer world, will it not permit the cat (any system that is inside) to stay coherent?
and thanks for the blog =)
Ignat asked:
Delete>Sorry to be so naive, but if we consider the Schroedinger box as being completely isolated from the outer world, will it not permit the cat (any system that is inside) to stay coherent?
Yeah, that is indeed the point. In practice, you almost certainly cannot get that degree of isolation, and Schrödinger almost certainly knew that.
But, in principle and if you take the quantum-mechanical formalism completely seriously, then, yes, the cat can be in a superposition of |alive> and |dead>.
Which of course Schrödinger, and almost everyone else, recognized to be absurd.
Part of the reason that debate on the foundations of quantum mechanics can get so rancorous is that some people fail to grasp the distinction between “for all practical purposes” considerations versus arguments that try to find the limits of the formalism by pushing the formalism to its limits in principle.
For all practical purposes, the cat is not in a superposition due to decoherence.
But, in principle, the cat is in a superposition, which cannot be, so somehow the formalism breaks down.
Of course, the many-worlds guys just accept that the cat and everything else is forever in a superposition. There are, however, serious technical problems with MWI not to mention the little issue of Ockham's razor!
Dave
Ignat,
Delete"Sorry to be so naive, but if we consider the Schroedinger box as being completely isolated from the outer world, will it not permit the cat (any system that is inside) to stay coherent?"
It is impossible, even in principle, to have a box that completely isolates the cat. It is not possible to shield the gravitational field of the cat for example. So, it is not possible, even in principle, to create a cat in a dead/alive superposition.
Schrodinger's cat experiment is ill-conceived and there is nothing that can be learned from it.
Andrew, I am afraid you misinterpret what "in principle' means, and confuse it with 'in practice'.
DeleteThere is nothing that prohibits us consider pure states or isolated systems. Although none of them are actually available in practice, both are possible in principle, and we investigate them.
Thanks @PhysicistDave, for clarification. Just wanted to check if I understand correctly =)
DeleteIgnat,
Delete"I am afraid you misinterpret what "in principle' means, and confuse it with 'in practice"
I don't think so. In principle, we can build a galaxy-sized collider. We cannot, even in principle, assume that there exist a box that shields the gravitational field (or even the electric field). It's not a problem of time or resources. It cannot be done. It conflicts with known physical principles.
@Andrei which physical principal is violated by the idea of an isolated system?
DeleteIgnat,
Delete"which physical principal is violated by the idea of an isolated system?"
Newton's law of universal gravitation. According to this law, the cat inside the box attracts every other object in the universe (including those outside the box) with a force that is directly proportional to the product of their masses and inversely proportional to the square of the distance between their centers. You want a magic box that does not allow this to happen. Such a box cannot exist.
@Adnrei, which physical principle implies that the gravitational filed MUST exist in the universe?
DeleteOw, and you remember of course, that the Newton's law is neither a correct law (we still do not have a complete theory of gravitation), nor a physical principal, but a particular physical force description (and incorrect as mentioned).
DeleteIgnat,
Delete"Ow, and you remember of course, that the Newton's law is neither a correct law (we still do not have a complete theory of gravitation)..."
Using General Relativity to describe the interaction between a cat and some other object of similar size is not required, but nevertheless possible. Newton's gravity is a very good approximation in this case. Do you claim that in GR you can shield the cat from the exterior? How?
And GR is, as far as we know, a complete theory of gravitation. What's incomplete about it?
Ignat,
Delete"which physical principle implies that the gravitational filed MUST exist in the universe?"
It's a fact that the gravitational field exists. I don't know if it is a necessity in all possible universes. Now, if you accept that the Schrodinger's cat experiment is irrelevant for our universe, I'm satisfied.
But my argument applies equally to EM field. If the cat's brain is functioning it produces electric and magnetic fields that also permeate outside the box. And if you eliminate EM fields, together with the gravitational one you end up with no matter, atoms or cats, so no experiment.
Ignat,
DeleteWhat is going on in the discussion between you and Andrei is that Andrei thinks that classical effects such as the infinite reach of the Coulomb field explains many quantum effects: most physicists would not agree.
In one particular case that Andrei specified in detail, I proved, to his satisfaction, that he was mistaken, but he still thinks something else might work.
In the case you and he are debating, as I assume you know, "for all practical purposes" it is indeed impossible to isolate the cat in the real universe.
But, as you and I discussed, "in principle" you can imagine a nearly empty universe with just the cat in the box and then ask Schrödinger's question.
If necessary, imagine a universe in which the gravitational field does not exist. And make the walls of the box perfect reflectors so that no EM radiation can get in or out.
The paradox remains, "in principle" of course.
Which is why so many of us physicists for so many decades have thought that the paradox is trying to tell us something about quantum mechanics.
All the best,
Dave
PhysicistDave,
DeleteYou can reflect EM waves, but not the static electric field associated with charged particles. Just like in the case of gravity every charge inside the box interacts with charges outside. You need a box that makes a charge inside dissappear. I don't think it is possible. So, you need to imagine an universe with no gravity and no electromagnetism, no atoms, and no cats. Such an experiment is simply incoherent. There is nothing you can learn from it.
@PhysicistDave @Andrei Perhaps gravitational effects are not adequate to decohere SC, if the experiment was done out in deep space with zero curvature. However, I think Andrei is correct that SC experiment is impossible, even in principle, in the actual universe. (Saying it is possible in a different universe where only the cat exists is irrelevant to whether it's possible in our universe!)
DeleteThen we are truly lost.
ReplyDeleteYou can skip the math.
- in regards to QMP, .,.
-. (hate to sound 'cliche')
But, if a tree falls in the woods with no one around.
Does it make a sound ?
All Love,
Love Your Work
Sabine,
ReplyDeleteIt is impossible to discuss serious physics on blogs. I have an idea.
If interested, let's contact some journal editors (like Rev Mod Phys? Or Rep Progress Phys Or maybe even PRL if they allow) and write a scholarly (peer-reviewed) paper on where the community stands on the measurement problem. My perspective is not unique and my comments on probabilities as plausibilities are not my own (actually they originate come from Laplace, Bohr, Kolmogorov and many others physicists and mathematicians later).
I want this to be serious with no obfuscation. Actual Math. No white-noise generating armchair crackpots (PhysicistDave & others). Peer review (pending interest from an editor).
Let me now. My email can be found online.
To non-physicists:
DeleteSteve Presse wrote:
>My perspective is not unique and my comments on probabilities as plausibilities are not my own (actually they originate come from Laplace, Bohr, Kolmogorov and many others physicists and mathematicians later).
In case any of you are wondering, no, Presse's "Bayesian" obsession has not been at all common in the quantum physics community, and it is very hard to see how it can be made to fit with textbook QM, which, as I keep pointing out, is necessarily "frequentist," simply because it predicts experimental outcomes, not changes in Presse's beliefs, which is what Bayesianism does.
In fact, Presse is the first guy I have ever heard claim that Bayesianism solves the quantum-measurement problem, and I have been following both Bayesian analysis and the foundations of quantum mechanics for almost exactly fifty years.
Presse also wrote:
>If interested, let's contact some journal editors (like Rev Mod Phys? Or Rep Progress Phys Or maybe even PRL if they allow) and write a scholarly (peer-reviewed) paper on where the community stands on the measurement problem.
There is no consensus in the field, and everyone knows that. Presse seems to think that his Bayesian approach will solve the quantum-measurement problem, which is an unusual perspective. He himself should indeed write it up and try to get it published, no need for help from Sabine, me, or anyone else. Even if he cannot get a journal to publish it, he should be able to get it on the arXiv: if nothing else, he can just publish it on his own webpage.
I don't think he can write it up, because I do not think he has any actual ideas on the subject, just a few slogans.
But he really should try to prove me wrong.
If he can write it up and show that it works, he will crush me like a bug. And even I will enjoy it because I will learn something.
I do not think he will do it.
Well, Chris Fuchs' QBism is basically Copenhagen with Bayesian lipstick. I think it's good as far as interpretations go, but of course it doesn't solve the measurement problem. As I have said many times before, the moment you start talking about someone or something which has information or knowledge about something else, you have to ask "why wasn't I able to derive this from the constituents of the system". Because if you postulate you aren't able to do that, you are implicitly assuming that reductionism breaks down without telling us when and how that's supposed to happen. And if you do not make this assumption, then all the talk about knowledge and information should be superfluous.
DeleteI don't know why people don't appreciate this point more. Quantum mechanics is fine as an emergent theory, in which case you can very well talk about it encoding information and so on. But if you insist it's fundamental, you are also insisting that macroscopic objects require laws that cannot be derived from the laws of their constituents.
(Which is why I think all particle physicists should be strongly opposed to QBism and Copenhagen-ish interpretations. I guess the main reason this isn't the case is that it doesn't really matter for the type of experiment they do.)
Sabine,
DeleteRe “the moment you start talking about someone or something which has information or knowledge about something else, you have to ask "why wasn't I able to derive this from the constituents of the system"”:
So how does a physicist (or anyone else, e.g. Donald Trump) “derive [information or knowledge]” about the world “from the constituents of the system"? I would think that you can’t derive any (right or wrong) information or knowledge about the world “from the constituents of the system" unless you say that (what is represented by) IF, AND, OR, THEN, and ELSE are also “constituents of the system". And IF, AND, OR, THEN, and ELSE can’t be derived from equations and their associated numeric outcomes, because they are two entirely different types of thing.
(I hasten to add that I’m not trying to imply anything about “QBism and Copenhagen-ish interpretations”, which is a different issue.)
Lorraine,
DeleteLook, I have told you previously that you are rather trivially mistaken in thinking that physical theories cannot contain logical operators. Why don't you do us all a favor and read a book on quantum information, thank you.
So Sabine,
DeleteYou are saying that some aspect of the universe is doing (what we would represent as) IF, AND, OR, THEN, and ELSE? I.e. analysing information. Or is it just people and living things doing (what we would represent as) IF, AND, OR, THEN, and ELSE?
QuBism is a ψ-epistemic interpretation. It is a sort of speciation off Copenhagen Interpretation, which itself has lots of varieties. Chris Fuchs is passionately committed to this. It is an example of my advice in general that it is not wise to chase after quantum interpretations. Of course, maybe Chris will put a rabbit out of the hat where QuBism turns out to be “the way,” maybe even a way to solve the measurement problem. I will say in general I disagree with that prospect, but I cannot disprove it.
DeleteIs quantum mechanics emergent in some way? It is possible that QM is an aspect of something more general. I am not committed to any idea this can solve the measurement problem. In fact, I suspect such a generalization may make our ideas of ontology less relevant and cast things in a stranger way. This is seen in part with gravitation. If a quantum state is observed to be approaching a black hole then by the tortoise coordinate t = t0 - 2m ln|r - 2m| the quantum state remains on the stretched horizon and also appears as Hawking radiation away from the BH. What this means for an observer is strange. If an observer makes a Heisenberg microscope measurement of fields near the stretched horizon she may find the field or quantum state there, but also she might hone her detector further afield to find the Hawking radiation. There is a loss of locality here, and this also appears to indicate two incommensurate measurements are possible. From this is does not appear some solution to the measurement problem is right at hand.
The AdS/CFT suggests something similar. The interior bulk determined by conformal gravitation in the AdS_n bulk is dual or equivalent to a conformal field theory CFT_{n-1} on the boundary ∂AdS_n, that is a timelike boundary at infinity conformally mapped to a finite distance. The CFT in the standard quantum field formalism has field locality as oscillators at each point on a spatial surface defined at the boundary. The physics of the bulk is nonlocal, or at least to the extent there is no way to localize fields on a spatial surface. Any attempt to do so fails because any Gaussian-like surface is itself a surface subject to diffeomorphisms. This means such an attempt is not covariant in a proper way. This duality, which is not honestly completely understood, means that nonlocal field theory of the AdS bulk is equivalent to local field theory on the boundary.
The relationship between gravitation and QFT is also in a UV/IR duality:
UV quantum gravitation = IR quantum field sources
which is a way of writing the Einstein field equation. For the UV quantum gravitation the nonlinearity of gravitation or general relativity enters into the picture and in effect acts to decohere quantum gravitation. I see this as a way the nonlocality of quantum mechanics is converted into the nonlocality of gravitation. Spacetime built up from large N-entanglements may then mean the non-Lorentzian gauge-like field of gravitation assumes a Wigner-like quasiprobability distribution, used in coherent states of laser photons, and that for large N the quantum system assume classical-like properties
It is not apparent there is a solution to the measurement problem this way. As I have said the solution to the measurement problem is there is no solution. As with a quantum field interacting with a black hole, quantum mechanics means there is no unique meaning to a spacetime point or an event. In some ways this might mean the strange oddities of QM are not solved and made tame, but rather things are more bizarre. A possible sort of duality between nonlocality of GR or spacetime and the nonlocality of quantum waves might mean the strangeness of QM is a sort of hydra; we cut off one head and two then emerge.
All the universe has ever done and will ever do is execute mathematical functions, which you can, but do not have to, express as logical gates in certain regimes. Of course it is not just people who do that, ask any computer scientists. Have you ever even looked at how one integrates a differential equation?
DeleteSabine,
DeleteYou wrote:
>Well, Chris Fuchs' QBism is basically Copenhagen with Bayesian lipstick. I think it's good as far as interpretations go, but of course it doesn't solve the measurement problem.
Well... a somewhat bizarre and repellent shade of lipstick, I would say!
Have you read Fuchs' interview in 2015 with Amanda Gefter published in Quantamagazine?
The takeaway line for me was:
>“Quantum mechanics,” Fuchs says, “is a law of thought.”
I don't think that is what Bohr et al. had in mind: if nothing else, the laws of QM tell us about the actual statistical results of experiments, not just our thoughts.
The fact that we often refer to those statistical predictions as “probability” and that the word “probability” has many meanings and uses, including the non-statistical sense of “subjective probability,” does not change the fact that textbook QM is not about subjective probability.
Chris certainly has the ability to churn out veritable oceans of verbiage that will inevitably overwhelm his critics: has anyone actually read through in its entirety either part I or part II of his massive ruminations that he has posted on the arXiv?
Reminds me of Karl Barth's never to be finished six-million-word-long Church Dogmatics: both are theology.
I will say though that I admire Chris's talent for self-promotion and his simple chutzpah. If I had that talent, I myself would be famous! (Hmmm.... maybe I should follow his example and take everything I have posted on the Web for the last three decades and make it a book...)
The frightening thing is that such pointless self-promotion is starting to look like the true future of physics.
All the best,
Dave
Dave,
DeleteI understand your misgivings, I think, but, look, the only thing anyone of us ever observes is subjective. Chris is just taking this very, very seriously. I regard realism as a useful way to think about the world, but it's a philosophical stance after all, and therefore on that shouldn't be forced on scientists.
I have actually read big parts of Chris' papers and heard several of his talks.
(but, no, I haven't read the interview you mention)
DeleteI heard Fuchs a number of times. He is extremely committed to QuBism as the only correct interpretation of QM, and in fact he thinks it is not just an interpretation of QM, but the crux of QM. He is a smart guy, and his single devotion to this has enforced his mind into a steel trap against any criticism of this. Unfortunately for him he has not produced a simple result that clearly makes his case. He as produced his massive set of arguments found on Arxiv, that are suggestive, but not conclusive.
DeleteTo be honest it is rather sad. One can compare it to Hugh Everett, who after proposing his interpretation, that lead to many worlds interpretation, found his research career finished. Hugh was treated unjustly, for he did not dogmatically hold to this and did go on to other things. At least Fuchs though has not not plunged into the self-destructive life that Everett did.
I do not think there is any possible proof that QM is either ψ-epistemic or ψ-ontic. I derived an argument for this some years ago, though it does not work for 2-state systems. I suppose I should not be too worried about this, for the Gleason theorem has a similar hole. The same I think occurs if we think of QM decoherence or measurement as a case of qubits encoding qubits in a form of self-reference. My thinking then is fervent upholders of MWI or QuBism or any other interpretation risk wasting their time.
Lawrence Crowell,
Delete"He (Fuchs) as produced his massive set of arguments found on Arxiv, that are suggestive, but not conclusive."
I find his arguments ridiculous. Let's look at this paper:
Notwithstanding Bohr, the Reasons for QBism
https://arxiv.org/pdf/1705.03483.pdf
About the EPR argument he writes:
"Alice starts with an entangled quantum state for a bipartite system, one half of which might be far away. This quantum state refers only to her beliefs about her own potential experiences should she take an action on either or both of these systems. Suppose she takes an action on the system at A. The consequence of her action on A causes her to update her quantum state for B. So what? What does it mean? Does it mean something spooky happened at a distance? No, it only tells of what she should expect should she take an action on B. On the other hand, under the assumption of locality, how could she take an action on B if not to walk over to it? To repeat, QBism is a strictly local account of quantum theory [8], so making an updated quantum-state assignment for B means nothing more than an updating of the agent’s expectations for what might happen to her should she take an action upon system B.
And there goes poof to the alleged incompleteness, at least in the sense that EPR saw it."
So, according to QBism, the only way to establish that a signal has been sent instantly to Mars is to go instantly there yourself. A computer on Mars, programmed to perform a measurement at a certain time is no evidence for Fuchs. Why? It's beyond silly.
Sabine,
DeleteRe Sabine Hossenfelder 11:43 PM, July 05, 2020:
The fact that one needs to take logical steps to “execute mathematical functions” has seemingly never given physicists pause for thought. There seems to be a myth that the equations of physics could work perfectly well without these logical steps. In other words, the logical steps taken by physicists (or others) should be seen as a necessary part of the model. The model implies that the world itself is taking logical steps. However, we never hear about this aspect of the model: we only hear about the equations.
Lorraine,
Delete"There seems to be a myth that the equations of physics could work perfectly well without these logical steps."
No, you have it exactly backwards. All existing laws of nature are based on logical operations. It is hence utter nonsense to claim that somehow physics can't deal with it.
@ Andrei: This is a strange statement, and seems to mean the nonlocality of QM is removed if Alice goes to B, but in order to do that it requires instantaneous travel. Relativity and QM in a funny way reinforce each other.
DeleteI have not followed all the reasoning behind QuBism. In fact to be honest I have followed little. I would say that unobservable nonlocal correlations between quantum states are related to or measured in part by the subjective expectation of the observer. Fuchs equates the two completely, which one can respond with the question, "What makes that so?"
In some ways this leads to a strange inference of QuBism: solipsism. Again this gets into the question raised in this thread of how do we know anything else is conscious. Of course, the problem with solipsism is that if one were to really internalize this you in effect go mad, or at least can go mad. The belief one is the only existing conscious being, or that ultimately there is nothing objectively real outside of one's self is reason to be put away on the funny farm. Of course one guy made claims a bit like this and has been believed by many hundreds of millions of people for nearly 2000 years.
Sabine,
DeleteI never said that physics "can't deal with it". I'm making the point that, apart from your saying it above, physicists never say that "All existing laws of nature are based on logical operations". I.e. physicists never mention the "logical operations" that they need to perform to make the differential equations work: we only ever see the differential equations. These logical operations are a completely separate aspect of the world to the lawful relationships aspect represented within the differential equations. When are physicists going to divulge to the general public that (what is represented by) IF, AND, OR, THEN, and ELSE are needed to make the models of the world work, and are therefore needed to make the actual world work?
Lorraine,
Delete"I never said that physics "can't deal with it" [...] These logical operations are a completely separate aspect of the world"
No, they are not. How the heck do you think computers work? Also, you are disagreeing with yourself.
Lorraine has a point here: The clash between empiricism and idealism. From an empiristic viewpoint, principles/axioms and theorems of physics are a (man) constructed model to explain phenomena, but they can't be proven in reality, they are possibly even unnecessary. E.g. Hume, who was an empiricist, critized Newton heavily for his theory of motion.
DeleteA physicist who thinks that his theorems and further derivations describe the reality "as it is" (nothing wrong with it, by the way...) is an idealist.
Sixte,
DeleteI don't see anyone here claiming that physical theories describe the world "as it is" which is to me an utterly meaningless phrase. Physical theories describe observations, hopefully accurately. And in any case, this is besides the point. It is bluntly wrong to claim that logical operators are supposedly something that cannot be accommodated within currently used physical law. As I said above, open any textbook on quantum information to see that this is plainly nonsense.
Sabine,
DeleteHere is the Quanta magazine interview with Fuchs.
You wrote:
>I understand your misgivings, I think, but, look, the only thing anyone of us ever observes is subjective.
Sabine, are you aware of the mid-twentieth-century philosopher W. T. Stace? He wrote an interesting book, The Theory of Knowledge and Existence, back around 1930 in which he tried to deal with experience as the only true basis for reality. The logical positivists had a similar perspective, of course, but I think did not carry it out as systematically as Stace.
I am not unsympathetic to this sort of “phenomenalist” approach: as you say, all of our observations are necessarily subjective. A case can be made that showing how to move from phenomenalism to realism is the basic problem of philosophy.
But this is not what Chris Fuchs is trying to do or claims to be doing. Chris is, after all, claiming to do something much more specific: explain QM as being really based on Bayesian probability.
And the problem is that no sensible experimentalist cares about some theorist's or his own subjective probabilities when he is measuring the statistics in, say, the double-slit experiment or an experiment to test Bell's theorem. He cares about the actual statistics. And that is what the textbooks teach us physicists to calculate: the asymptotic statistics if you take enough samples in the experiment (of course, we also learn how to deal with situations where we cannot take an effectively infinite number of data points).
I just cannot find any place in which Fuchs or his pals actually deal with this. There are reams and reams of words – Chris is certainly a very chatty and indeed verbose guy.
But where on earth does any of it actually connect with quantum physics – aside from just declaring dogmatically that “probability” means “subjective probability” and that therefore “probability” in QM just means Bayesain probability, despite the fact that even pretty stupid students realize that we use “probability” in the frequentist sense, i.e., long-term statistics?
As far as I can tell, this is just as dumb as saying that Einstein proved “Everything is relative!” which is just stupid verbal manipulation.
Sabine, can you or anyone point me to any place where Chris's QBism amounts to anything else than just a silly verbal trick based on equivocating about the “true” meaning of the word “probability”?
What Chris is actually doing all seems like pretty cheap verbal nonsense and nothing more.
All the best,
Dave
Lawrence Crowell,
Delete"In some ways this leads to a strange inference of QuBism: solipsism."
Yes, this is my conclusion as well. In fact solipsism seems to be the only way for a non-realist theory to bypass the EPR argument and make the claim that it is local. The problem is that once you deny the existence of an external world containing events in space-time the concept of locality becomes meaningless.
Of course, QBists deny solipsism, they agree that an external world exist, but QM does not describe that world. I'm fine with this, but then why are they making claims about locality? If your theory does not describe the world how do you know if the world is local or not?
But the QBist confusion does not stop here. Fuchs claims that the instrument used to perform a measurement should be seen as part of the agent making the measurement. In the same paper:
https://arxiv.org/pdf/1705.03483.pdf
we read (page 10-11):
"The second level of personalism appears with the meaning of a quantum-measurement outcome.
On this, QBism holds with Pauli (and against Bohr) that a measurement apparatus must be understood as an extension of the agent himself, not something foreign and separate. A quantum measurement device is like a prosthetic hand, and the outcome of a measurement is an unpredictable, undetermined “experience” shared between the agent and the external system."
OK, let's assume, for the sake of the argument that Fuchs is right. Let's now repeat the EPR experiment, but let Alice controll both A and B measurement devices using some long wires. In this case, both polarizers or Stern-Gerlach devices are part of Alice, so, she is in fact at both locations. As a result, even the ridiculous argument that Alice does not experience the result at B unless she goes there fails! So, QBism cannot provide a local account of the EPR experiment.
Sabine,
DeleteHow the heck do YOU think computers work? From what I’ve observed, the majority of physicists DON'T know how computers work. They should be embarrassed, but they are just oblivious. Seemingly these physicists are making the very basic mistake of thinking that genuine IF, AND, OR, THEN, and ELSE could be done with opaque, man-made symbols i.e. they mistake symbols for genuine information; these physicists don’t even know what symbols are anyway. Computers are not doing genuine “logical operations”; computers merely move opaque, man-made symbols about, according to a plan devised by computer programmers. It’s the computer programmers who take logical steps, not the computer program.
The equations of physics are just symbols on pieces of paper. The equations of physics don’t work to model the world without the logical steps taken by physicists (or others). So the logical steps taken by physicists (or others) are an essential part of any model of the world.
Lorraine,
DeleteComputers are made of particles. I'm a particle physicists. Hence I know how a computer works. How does it work? It's executing the laws of nature.
"Computers are not doing genuine “logical operations”; computers merely move opaque, man-made symbols about, according to a plan devised by computer programmers. It’s the computer programmers who take logical steps, not the computer program."
This is nonsense. Logical operations are performed by the processor as it is executing a program. If you do not want to count what a computer does as a logical operation you have simply defined a completely meaningless concept of logical operation. Also, you clearly still have not opened a textbook on quantum information.
QBism avoids nonlocality by saying all probabilities are about the expectations or beliefs of the observer. In this way they avoid the reality criterion of EPR , where without in any way disturbing a system, a value of a physical quantity can be found with certainty (prob = 1), then an element of reality corresponds to that quantity. If they reject any idea of objective probability, even for probability = 1, then they claim they avoid the criteria of the whole EPR argument. I see this as a sort of word game, and the massive tome of emails by Fuchs on the Arxiv amount to a burial of arguments by massive amounts of verbiage.
DeleteThe argument over QBism has a funny self-contradiction. Is QBism real? QBism rejects any concept of reality to QM, which means that if QBism is true or real in some fundamental way, then does this contradict the central concept of QBism there is nothing to reality outside the subjective consciousness of the observer? If we reject any objectivity to unit probability, where in some way Bayesian statistics converges to Boolean logic, then can even classical reality can be rejected?
I what way is Bayesian statistics preferable to frequentism? Frequentism assumes an objective sample space. Boltzmann physics relies on the same with macrostates as groups of sufficiently similar microstates on a coarse grained level. If in a certain system, say one that is maybe sufficiently small, we have a complete knowledge of the system to know the sample space, then as I see it Bayesian and frequentist statistics would give absolutely the same output. From that perspective I would say on a practical level there is nothing distinguishing the two. BQists would say this is false because even prob = 1 is not objective, which in some ways puts them in a position similar to Dostoyevsky’s underground man. In the case of Boltzmann statistical physics I would also say that any errors or deviations that do happen are under the logarithms of the theory so tiny as to be ignorable. At some point in physics we must shake that sort of thing off and say something is “such and such” FAPP and ignore these “shaving down to a point” sorts of arguments.
In physics there is something to be said for, “If it swims like a duck, waddles like a duck and quacks like a duck, them it really must be a duck.”
In some ways there are things to be said for the GRW, Montevideo and Penrose R-collapse interpretations. If quantum information is conserved by being converted to a form of BMS symmetry or charge that escapes out to ℐ^∞ in a quantum gravitation setting, the objective collapse of GRW or R-system of Penrose are perfectly acceptable as phenomenology that ignores this quantum gravitation. This might connect with something like MWI, but it does not have to. In fact, it can have aspects of both ψ-epistemology and ψ-ontology, where conservation of information creates an option for both. There is nothing I see here that is terribly complicated and requires massive tomes of verbiage and megabytes of emails to support.
@ Lawrence,
DeleteOne can view probability as an objective probabilistic disposition.
Sabine,
DeleteHow computers work has got NOTHING to do with the particle physics of circuits, voltages and transistors. No one is interested in circuits, voltages and transistors per se. The power and utility of computers is all about what the transistors/ voltages/ circuits can be made to represent.
“Low” and “high” voltages are not zeroes and ones, but the voltages are made to symbolise zeroes and ones (OR ones and zeroes); and groups of these zero and one symbols are not words, sentences, numbers or equations, but they are made to symbolise words, sentences, numbers or equations. The voltages that symbolise a particular word, for example, are essentially arbitrary: it just depends on the way human beings decide to symbolise things. Essentially, computers are processing symbols; these symbols only mean something from the point of view of human beings; obviously, from the “point of view” of the computer, there are no words, sentences, numbers or equations, and there are not even any zeroes and ones, there are merely “high” and “low” voltages.
When people debate an issue (e.g. do guns protect people or kill people?) using IF, AND, OR, THEN, and ELSE, they know what guns are, they know what people are, and they know what the words “kill” and “protect” mean. But computers only “know” high and low voltages: “logical operations” performed by computers are pretty well meaningless.
Lawrence Crowell,
Delete"QBism avoids nonlocality by saying all probabilities are about the expectations or beliefs of the observer. In this way they avoid the reality criterion of EPR , where without in any way disturbing a system, a value of a physical quantity can be found with certainty (prob = 1), then an element of reality corresponds to that quantity. If they reject any idea of objective probability, even for probability = 1, then they claim they avoid the criteria of the whole EPR argument."
What I find interesting is that QBists accept that QM is not a theory of nature, but a tool to make good bets. But, if QM is not a theory of nature it cannot be a complete theory of nature. So, they do not actually have a problem with EPR. EPR only proves that a complete, non-realistic QM must be non-local.
In regards to probability 1 not being objective, that is silly. We have two logical option:
1. QM is true (its predictions always come out right) - Probability 1 is an absolute certainty, or
2. QM is not true (its predictions sometime fail) - Probability 1 is just a belief, it might be a wrong belief.
If 1, the EPR argument applies, a non-realist fundamental QM must be non-local
If 2, the EPR argument does not apply but QM cannot be a fundamental theory, because it is wrong.
Either way, EPR wins.
"When people debate an issue (e.g. do guns protect people or kill people?) using IF, AND, OR, THEN, and ELSE, they know what guns are, they know what people are, and they know what the words “kill” and “protect” mean. But computers only “know” high and low voltages: “logical operations” performed by computers are pretty well meaningless."--Lorraine Ford
DeleteA) Feral children know none of these things, and seem incapable of learning them, rather like a computer without any social programming. People are trained (programmed, if you will) by their parents and teachers and role models as to what certain words (sounds, or visual patterns on a page) and actions represent. Similarly, computers can be trained to associate symbolic meanings to their voltage, magnetic, and other settings.
B) Thinking done by humans also uses EM/chemical signals and settings in their brains and nervous systems. In fact, computer programming based on models of how neurons and synapses seem to work (neural networks) turn out to be effective ways for computer programs to learn new skills, whose algorithms were not directly programmed into them but developed by trial and error. (Which is how humans developed all their symbols, over hundreds of thousands of years.)
Granted, most application-programmers first determine what algorithms they themselves could use to accomplish a task (such as the first task I was given at GE, to program a method for adding a 60-mil forging envelope to the shape of a vane cross-section) and then tell the computer how to do it step by step, but since there is no magic envolved in thinking, a sufficiently clever programmer who had studied how brains work could in principle program a computer to think for itself. E.g., AlphaGoZero learning how to defeat human Go grandmasters (which the programmers themselves could not do, so they could not tell the computer how to do it step by step, but only how to learn to do it).
The next step in this evolution might be to write programs which let computers add new neural networks to their program, to be trained for new tasks, similar to the way brains grow new synapses. (Not that my idle thoughts will be of any use to experts in the field.)
But would such a computer actually be thinking? Here I apply Dr. Crowell's Duck Principle, along with the question, are humans really thinking?
(Hey, maybe once we develop such computer programs we can use the knowledge to rehabilitate feral children.)
This of course is the same response to your assertions that has been given many times before, repeated here only so that theoretical new readers can see the other point of view. It seems to me the contrary point of view quoted above would require some sort of magic with which to privilege human thinking but no such magic has ever been reliably detected, to my knowledge. I myself have seen too many humans (including myself) screw up too many times to believe such magic likely.
To be somewhat fair to QBism, the Bell theorem just demonstrates that reality, the belief an observable exists when not subjected to an observation, and locality, that an observable can be specified entirely within a region without correlations elsewhere, can not both hold at the same time. We are most familiar with the standard take on this that locality is wrong. As physicists we are most often not willing to abandon the idea that physics is about something that exists in the world. However, we can just as well decide to abandon reality and hold to locality. This is what QBism does, and until we can maybe “wring out better Bell inequality theorems” we are forced to admit oddities such as QBism are admissible.
DeleteThat QBism leads to this sort of possible contradiction, if reality is false, then even QBism is not real and thus we not doing anything better than betting on horses, might mean there is something odd afoot. I find there is something odd about nonlocality as well. Quantum mechanics is a system of dualities, and one of those is between ψ-epistemology and ψ-ontology. I should revisit this work I did back in 2016, but I have an argument that neither of these is provable. It is not possible to prove with QM which of these is the case. This means that both Bohm’s QM and QBism, ψ-ontic and ψ-epistemic, are possible and not provable. I see much the same with the dualism between reality and locality.
With nonlocality there is something odd I have found. I have been working on a sort of equivalence principle that involves quantum entanglements. This states that entanglements between two EPR pairs is the same in flat spacetime as it is in curved spacetime so long as the pairs are on geodesics. This sounds good, until we ask about event horizons. This is where things become a bit strange. If Alice and Bob are on two geodesic paths, and they are also not highly separated, then if they approach a black hole this should still hold. If Alice crosses the horizon a short time interval before Bob, then it makes sense the two should not notice anything in particular. However, if we let Bob be on a more distant frame then Bob will lose contact with Alice and the entanglement is rubbished up by Hawking radiation. For the case where say Alice is on an accelerated frame the entanglement is lost by the Unruh effect, which is a decoherence of quantum information. But with black hole horizon this happens with inertial frames. It also means there is a localization of these quantum states. This then brings in some deep questions about the role of nonlocality in QM and localization of quantum states with gravitation. It also suggests that complete or perfect nonlocality is a bit of a fiction, just as the QBist idea of abandonment of reality.
There are some deep issues here, I think. It might also mean in some ways we are rescued in some ways from the dragon of non-realty. The flip-side of this is we may have to give up the idea of perfect nonlocality as well, which means that pure quantum states are an unattainable idealism.
JimV,
DeleteThere is no debate about computers, there is just the facts about how computers work. I can't even be bothered talking to you, because I've tried at length many, many times before, and you'll never understand it. You, like the people who can’t understand climate change, just haven't got the ability to understand how things work.
To address a few comments here. Dr. Edwards says we can view probability as objective. This is the case, but it does not help much if your data is limited. Rutiger Schack made a good point of this, which is in part what kicked off the whole Bayesian approach to physics and QM. Whether one uses frequentist or Bayesian statistics can depend in ways upon the problem at hand.
DeleteMy basic caveat I have about QBism, and frankly this extends to most interpretations, is none of this has led to anything that permits one to do quantum physics in a new way. The day somebody demonstrates some clear experimental result, even just a phenomenological result that might hold, based on using an interpretation then my ears will pique up. David Deutsche spent years trying to find some clear signature of MWI as the “real interpretation,” but nothing convincing came about.
QBism is not going to change how we do quantum physics, and I see it having no impact on P = ∫ψ*ψd(vol) or that 〈O〉 = ∫ψ*Oψd(vol) and the Born rule. Now if QBism turns out to be the only way to derive the Born rule, then it will win the day. However, Carroll and Sebens have a partial argument for how Born rule is consistent with MWI, where MWI is not consistent with QBism. I then suspect QBism and largely other interpretations will be things that some people will argue over, with tons of words spilling across the internet and publications, but how we actually do quantum physics may not change one jot or tiddle.
I though give a little quarter for Penrose’s R-process, GRW and Montevideo that invokes gravitation. If the loss of quantum information is accounted for with BMS symmetries or gravitational memory that reach ℐ^+ these three interpretations might be wrapped into more complete theory of quantum gravitation. At that point in some ways they are no longer interpretations. The spontaneous collapse of quanta in GRW, which if there are more quantum states occurs more rapidly in a system, might connect with Penrose’s R-process which Penrose connects with metric physics. The superposition of two metric configurations may be sufficiently nonlinear, or evolve towards such, and in Penrose’s R-process this destroys the quantum phase of the wave function and there is a loss of quantum information. This results in the failure of 1 = Tr(ρ) = Tr(ρ^n) for n > 1. This is interpreted as a loss of quantum information, but with gravitation this might mean such information is transformed into another form. I find this a lot more potentially useful and interesting than anything QBism might offer. Spacetime as built from quantum entanglements means on some level this quantum phase is just transferred into a form that is not tractable or practical to observe.
It is interesting that Hugh Everett was pilloried for proposing his interpretation, where of course he was a graduate student and thus “academically expendable.” Fuchs has proposed an interpretation and has managed to acquire a position at UMass. QBism is the only thing that Fuchs has done in physics; he done nothing else to my knowledge. Maybe this speaks to our age, where what gets people into high positions is dependent upon whether it is popular or “hip.’
Oh yeah, I forgot to say this. Lorraine, take a course in digital electronics. I had 3 courses in electronics, and the last was digital. Computers on a chip, gate, flip-flop level are ALL about voltages and currents. In fact you can hook a chip to an current source and oscilloscope and map out the logic of the chip. I did this way back when.
DeleteLawrence,
DeleteYou don’t understand what symbols are, and what symbolic representation is. The transistors/ voltages/ circuits are just a means to an end. I repeat: The power and utility of computers is all about what the transistors/ voltages/ circuits can be made to represent.
You wrote:
Delete"How computers work has got NOTHING to do with the particle physics of circuits, voltages and transistors. No one is interested in circuits, voltages and transistors per se."
I am not sure what to tell you, but it is all about electronics. Contrary to what you say in this post the binary system is on/off = 0/1 and from there these are lexical analyzed into hexadecimal bytes and standard numbers.
Your argument is a bit like saying the power of television is not the electronics, but rather the visual and auditory representations. In one sense, sure, but one can argue a theater does the same. Binary strings are converted into alphanumerical characters or define palate. This done by algorithms that ultimately manipulate logic gates that are circuits.
Lawrence,
DeleteBinary digits are a man-made concept, and transistors, voltages and circuits are used to represent this “binary digit” concept. But there are no genuine physical “binary digits”.
“Binary digits” are represented by higher and lower voltages e.g. -4 to +1 volts or -1 to +3 volts: i.e. there is no genuine physical “0” or “1”. In addition to this, a binary digit (e.g. 1 or true) can be represented by either the higher voltage or the lower voltage: i.e. there is no genuine physical “0” or “1”. “[H]exadecimal bytes and standard numbers” are represented by sets of high and low voltages, but not in the way that you might have expected.
It is not that electronics and computers don’t exist: it’s that from their very foundations, computers/ AIs represent something to human beings that does not exist from the “point of view” of the physical computer: there is no “on/off” or “0/1” from the point of view of the computer; the “true/false”, “on/off” and “0/1”, and everything else that is built out of “binary digits” only exists from the point of view of human beings.
You reject any objective context to mathematics.
DeleteLawrence,
DeleteAre you, or are you not, willing to face the objective fact that binary digits and qubits don't actually physically exist: they only exist from the point of view of human beings?
Qubits = quantum states, and to the extent they exist qubits exist. Maybe better, qubits may be represented by the density matrix ρ = |ψ〉〈ψ|, a Hermitian operator, and are then quantum observables. With ordinary [0, 1] as off and on again I see nothing controversial about saying these physically exist.
DeleteLawrence,
DeleteRe Lawrence Crowell3:58 PM, July 10, 2020:
In a conventional computer, a higher (or lower) voltage can be said to represent “one” or “true” with absolute certainty.
I think this is what is happening in quantum computers: In a quantum computer, with its conventional computer interface, a measurement of a “2-state” quantum object or system results in a conventional computer voltage, where a higher (or lower) voltage represents one of the “2-state” outcomes. The calculated probability of this outcome is a separate, but necessary, item of information. Also, fault/error detection is necessary. The voltages representing the outcome, and the voltages representing the probability of the outcome, and the voltages representing the fault/error detection, are used to determine if the voltage representing the outcome can be said to represent “one” or “true” with an acceptable level of certainty.
The “qubit” idea is an idealised perfect, version of the physical reality found in a quantum computer.
Re Lorraine Ford 11:43 PM, July 11, 2020:
DeleteTo clarify, that last sentence should perhaps have said: The “qubit” idea is an idealised version of the less than ideal physical reality found in a quantum computer (i.e. real "qubits" don't actually exist).
Sabine,
DeleteYour link to the Q-CTRL Quantum Fundamentals videos is misleading because the videos don’t seem to specifically mention the difference between the theoretical qubit and the physical qubit. The online qiskit.org (open-source quantum computing software development) textbook is more realistic:
“Quantum computing requires us to encode information in qubits. Most quantum algorithms developed over the past few decades have assumed that these qubits are perfect: they can be prepared in any state we desire, and be manipulated with complete precision. Qubits that obey these assumptions are often known as logical qubits.
“The last few decades have also seen great advances in finding physical systems that behave as qubits, with better quality qubits being developed all the time. However, the imperfections can never be removed entirely. These qubits will always be much too imprecise to serve directly as logical qubits. Instead, we refer to them as physical qubits.
“In the current era of quantum computing, we seek to use physical qubits despite their imperfections, by designing custom algorithms and using error mitigation effects. For the future era of fault-tolerance, however, we must find ways to build logical qubits from physical qubits…”
Re binary digits and qubits:
DeleteThere are no actual binary digits, but conventional computers can be made to represent the binary digit concept. In a conventional computer, the higher or the lower voltage can represent “one” or “true”: there goes the idea that physical computers are performing actual logical operations. Also, in the real world, numbers [1] do not exist as an array of “binary digits” (more correctly, an array of higher and lower voltages): there goes the idea that physical computers are anything more than a symbolic representation of the real world.
Somewhat similarly there are no actual qubits, but a quantum computer is an attempt to make the qubit concept work. And as with conventional computers, in the real world, numbers [1] do not exist as an array of objects symbolising ones and zeroes (these symbols only appear after the measurement is made).
In other words, there are 3 (or 2) aspects of the world to take account of: 1) physical matter; 2) mind (or maybe matter and mind are the same thing i.e. matter/mind); and 3) man-made physical symbols which only mean something from the point of view of human beings.
……………
1. The numbers that apply to laws of nature are the only genuine numbers: the symbols that human beings write on pieces of paper are not the genuine numbers.
Lorraine,
DeleteI can't remember linking to any video, but trust me when I say that I perfectly well know what's the difference between a logical and a physical qubit. I find it bad terminology for science communication because I know from my own experience that even a lot of *physicists* fail to grasp the relevance of distinguishing both. Hence, I only talk about physical qubits.
Sabine,
DeleteYou didn’t link to it. I made a mistake. Sorry.
However one thinks about quantum interpretations, QM is at least effectively stochastic. In a large enough universe there will be copies of you that due to stochastic effects (whether or not due to QM) will end up evolving differently. Also given all the information you have, the information you don't have yet won't be predetermined even if the ultimate laws of physics are fully deterministic, because the information you have does not locate you within a sector of the universe where the new information you are going to acquire has a definite value.
ReplyDeleteIf it were possible for you to exist on an Earth where dinosaurs had never lived, then in a large enough universe you would have had a copy that lived on such an Earth which would have diverged from you when you first learned about dinosaurs. So, in this sense, dinosaurs only become reality after you read about dinosaurs.
There's a way of looking at some of these quantum issues using some ideas and terminology of Schrodinger's: suppose we reduce spacetime to a "state manifold" at the fundamental level. Schrodinger defined a well-ordered state manifold as one that preserved differences of proximity-- near vs far. He noted that quantum objects would not have a well-ordered state manifold. Now consider something I will refer to as a composite. A composite is some collection of quantum objects. Imagine that we can define a state manifold for composites. As we assemble composites, we should note that at some point we would have a state manifold that IS well-ordered. A well-ordered state manifold is the classical limit. So the interesting question would involve the transition from the quantum state manifold to the classical well-ordered state manifold as we assemble quantum objects into composites. We should note that in some circumstances, of extreme velocity, extreme mass, the well-ordered manifold might cease to be well-ordered, or vice-versa. Again we have the problem of transition, from one type of order to another. The intermediates are of interest rather than the extremes. I hope I haven't abused or misused or misrepresented any of Schrodinger's thoughts. He never proposed representing spacetime as a state manifold. That is my idea.
ReplyDeleteHi Sabine,
ReplyDeleteone last thing. My offer in my last message is meant to be one where we can, hopefully, hear each other spell out precisely and mathematically. It is driven by a curiosity to dig deeper into your position (it is not meant as a junvenile challenge).
Right now, both of our opinions (yours and mine) sound like silly cartoon versions of our (probably defensible) positions and we are being dragged into the mud pit by total internet crackpots/creepers with infinite time.
I have thought about photons and detectors quite a bit,
e.g., my PRX 2020 (which I don't expect anyone to read). And I feel that there are very concrete difficulties we encountered where only understanding probabilities as plausibilities can help move us forward. [For those you outside physics PRX is the second most prestigious journal after Nature Physics. PRL is, sadly, well past its heyday from 20 years ago].
For what it's worth, I am familiar with Fuchs' writing and dislike his philosophizing. I tend to take the shut-up-and-find-the-best-mode-of-calculation approach.
As any paper on "where we stand on the measurement problem" would contain 0 new results, this would need to appear in a review journal (hence, my suggestion for Rev Mod Phys or Rep Prog Phys.)
p.s. I will probably re-surface on your next post on the measurement problem.
Steve
"For those you outside physics PRX is the second most prestigious journal after Nature Physics."
DeleteSays who? And on what grounds?
PRX is an "open-access" journal, which means pay-to-publish. That used to be known as a vanity press. (PRX did waive publication fees at the beginning, but that is not the long-term plan.) Otherwise prestigious journals have recently spawned open-access variants, one example being Philosophical Transactions of the Royal Society, which published a paper by Joy Christian. Go figure.
As measured by impact factor which is directly correlated with referees asking for the moon.
DeletePRX is open access as are most journals now. Why?
Because funding agencies are (justifiably) pressuring journals to make public results paid for by tax payers. Actually, you should know this, coming from Europe where many agencies now actively forbid you from publishing in non-open access.
Hope this helps.
Maybe, I should also add this.
DeleteWhen Nature Physics came out and expected standards higher than PRL, PRL lost its shine. The editors of Physical Review had two choices.
Either expect the Nature Physics standard for PRL (and cause a revolt) or create a new journal which would hold authors to a much higher standard than PRL.
So they opted for the latter and created PRX. Because all new journals know where funding agencies stand on open-access, this is typically their default mode.
If you don't believe me, try this: submit as a pre-inquiry excellent work you have in progress to a PRL vs PRX editors. Night and day.
You'll notice that getting into PRL isn't what it used to be thanks to Nature Physics raising standards.
"As measured by impact factor which is directly correlated with referees asking for the moon."
DeleteIt is well known that impact factors are highly skewed, so while the average number of citations might be high, this is often driven by a few papers, and the median might even be lower than in some other journals.
Of course, judging quality via the impact factor (usually measured in the two years after the article appears), or even the number of citations for the paper itself, is not far removed from judging the quality of music by what is in the hit parade. Take, for example, the paper for which Weinberg was awarded the Nobel Prize, and see how many citations it had in the first ten years.
"PRX is open access as are most journals now. Why?
Because funding agencies are (justifiably) pressuring journals to make public results paid for by tax payers. Actually, you should know this, coming from Europe where many agencies now actively forbid you from publishing in non-open access."
Open-access does not necessarily imply an author-pays-hugely-inflated-fee-to-commercial-publisher model. There are lies, damn lies, and open access. It is sad that with Plan S and so on the commercial publishers made things worse: by transferring the fee from the author of from a library subscription to a higher level, nothing was done about the inflated costs, and authors will notice it even less.
Steve Presse2:30 PM, July 06, 2020
ReplyDeleteWhy don't you just read the book, "Lost in Math"? No need for a journal paper, you just need to learn what Physics actually is, after years of studying it.
Why don't you just read the book, "Lost in Math"? No need for a journal paper, you just need to learn what Physics actually is, after years of studying it.
DeleteMaybe because it has little or nothing to say about the measurement problem?
Mainly because scientific discussion takes place mainly in journals and not in popular-science books?
Thanks Phillip ;)
DeleteSave your valuable time and breath.
Phil,
DeleteThe problem is that Steve Presse has what he considers an important and novel approach to understanding QM that he expects everyone to accept: Presse contemptuously dismisses anyone who rejects it by saying, “Your [rejection of Presse's view] is a silly mistake to make in the 21st century.”
But he won't write it up as a journal article for people to try to understand and critique.
Presse wants to have it three ways: his theory is important and novel, everyone should agree with his theory, but he will not write it up for the literature.
Bizarre.
And in case you think he is just referring people to Chris Fuchs' work, nope. As Presse says above, “For what it's worth, I am familiar with Fuchs' writing and dislike his philosophizing.”
As to Sabine co-authoring a paper with him, she has stated explicitly,
>”The interpretation you suggests is not reductionist. It posits that the wave-function encodes knowledge held by observers and obtained by measurements, but there isn't any such thing as an "observer" or their "knowledge" or a "measurement" in the fundamental theory.”
She does not agree with Presse; of course, it is hard to see how any competent physicist could.
Look at Presse's publication list on the arXiv: long, long ago as an undergrad Steve was a junior co-author on a couple rather weird papers that challenged the accepted approach to the Aspect experiments. But at least those papers seem to say something that is testable.
His other papers seem to be in biophysics, fluid dynamics, stat mech, chem physics, etc., but not the foundations of quantum mechanics. It's fine that he publishes in multiple fields, but I am simply making the point that Presse does not seem to have published his wonderfully brilliant new interpretation of quantum mechanics, at least not on the arXiv.
If Presse really wants us to take him seriously, yes, as you say, “scientific discussion takes place mainly in journals” or at least on the arXiv or at least in some paper that we can find on the Web.
Presse is sure that anyone who engages in rejection of his brilliant theory is guilty of “a silly mistake to make in the 21st century” but Presse won't write up the theory and instead tries to drag Sabine, who disagrees with him, into co-writing a paper!
One word: crackpottery.
Dave
Phillip Helbig1:06 PM, July 07, 2020
DeleteBut the book does have something to say about theoretical physicists confusing mathematical models for physics, which is Steve Presse's fundamental problem. If it's the cure for his illness, what does it matter if it's merely a popular science book? It's the cure for your illness, too, Phillip. And Luke Barnes', and the Astronomer Royal's, and all the physicists involved in the new 20 billion Euro waste-of-time punt at CERN.
Physics is about observations not Bayesian fantasies. Steve Presse hasn't understood this because he doesn't understand what Physics *is*. So what's the point of a journal paper? Why doesn't Steve Presse educate himself on basic Physics instead?
How's that paper on universal fine-tuning coming along, btw? I can't wait to see your evidence that it is physically possible for fundamental constants to take values other than the only ones that they have been measured to be.
You can read God's mind - it's amazing.
And will you be doing another whitewash review for the Observatory, this time of Bible Boy's "Cosmic Revolutionary Handbook"? Another waste of hundreds of thousands of Templeton money on a collection of schoolboy errors. If they were true Christians they'd spend the money helping starving kids. Bible Boy looks as though he's spent half of the funds on pies, too.
Steve Presse12:33 AM, July 08, 2020
"Save your valuable time and breath."
Phillip's valuable time? He's preparing a paper on sci-fi just to try to prove a point, but he will just end up proving my point which is held by all sane people.
You should take the advice given to you, otherwise you're just going to be another mathematically competent but not very wise theoretical physicist getting stuck up a cul-de-sac.
"ow's that paper on universal fine-tuning coming along,"
DeleteIt's risen to the top of the pile. I've published two papers this year, one of them a long review, have two in press, and just finished a fifth, so I am now working on the fine-tuning paper again. There will be interruptions when proofs for the other papers need to be corrected, and for things with shorter deadlines, but there is progress. And this is all in my spare time.
How many papers have you published this year?
Yes, it's about fine-tuning, but it's not about words which you try to put in my mouth.
Steve Presse12:33 AM, July 08, 2020
DeleteIt looks like the claimed necessity for a Bayesian approach was carefully refuted 10 years ago. I'm sure you know the paper, but you're still a member of the cult:
Studies in History and Philosophy of Modern Physics, "A loose and separate certainty: Caves, Fuchs and Schack on quantum probability one" , Allen Stairs
Phillip Helbig1:25 AM, July 09, 2020
DeleteI've published no papers ever, as you know. Maybe I'll publish a paper based on your coming universal fine-tuning paper by adding "not" in every sentence. I'll cite you, of course.
I have discovered academic fraud in Australia, though, funded by the Templeton Foundation and my taxes via those donkeys at CUP.
"but there is progress"
No, there isn't. You have no basis for claiming that it is physically possible for fundamental constants to take values other than the measured ones, so you are still at the starting line.
Why are you doing this? It's crazy. You would be better off spending the time updating your Fortunate Universe review and pointing out that there is no evidence for universal fine-tuning and the multiverse, they're not even known to be scientific questions, and that one of the book's authors is mentally delusional.
If you have discovered fraud, take it to court. Otherwise, shut up.
DeletePhillip Helbig6:12 AM, July 09, 2020
Delete"If you have discovered fraud, take it to court."
Oh, I have discovered fraud. Luke Barnes used his scientific credentials and connections to sneak his crazy Iron Age superstitions into a science book. This is fraudulent and an attempt to corrupt science. It's not actually illegal though. He has made Geraint Lewis and Brian Schmidt look like naive fools.
"Otherwise, shut up."
You mean like you did in your whitewash review of that book of nonsense, "Fortunate Universe"? No, I won't be shutting up about the truth. "Fortunate Universe" presumably had reasonably good sales for a pop sci book given its preface was written by a Nobel Laureate, and is probably the first pop sci book ever to introduce an Iron Age fairy tale as Physics. This is significant. It wasn't that long ago when religious nutters were burning scientists at the stake. I'm sure scientists still suffer in theocracies today. So it's important to point out what lying sociopaths like Luke Barnes are up to. There is no place for fairy tales in science. People like Luke Barnes should seek psychiatric treatment for their mental delusions. Also, I'd rather the CUP didn't spend my taxes on publishing nonsense by halfwits as science. Luke Barnes has been given opportunities on here to defend himself but so far his responses have been to fantasise about killing critics of his nonsensical book and then to tell a blatant lie. Bible Boy is a sociopath.
I see you still can't provide a shred of evidence for universal fine-tuning. The truth is a pain in the neck, isn't it?
"Erwin Schrödinger’s thought experiment with the cat demonstrates that quantum effects for single particles can have macroscopic consequences."
ReplyDeleteYes, but he introduced it to show the absurdity of the Copenhagen interpretation. Like Einstein, also a pioneer of quantum mechanics (he worked in several different areas of quantum mechanics, and got his Nobel Prize for work in one of them), Schrödinger became disillusioned with the Copenhagen interpretation and much of what came after that time. Like Einstein, he spent his later years investigating classical unified field theories and enjoying life with young groupies in an open marriage.
Using the quantum logics approach one can derive all of the standard approach to QT from the single extra axiom that the logic comes from the lattice of closed subspaces of a Hilbert Space of dimension greater than 2. This includes the Born Axiom.
ReplyDeleteI don't think that fine-tuning caused an asteroid to kill the dinosaurs. Such a cause seems too capricious to me.
ReplyDeleteVery interesting post! You said "In Schrödinger’s thought experiment, the decoherence actually happens already when the toxin is released, so the superposition is never passed on to the cat to begin with." Many (MANY) papers have been published, often by philosophers of physics, that assert that if the box (containing SC) is adequately isolated from the rest of the universe, then the cat WILL go into superposition. How do you know this is not the case? How do you know it is not possible to adequately isolate such an experiment? Thank you.
ReplyDeleteAndrew,
DeleteThere's no contradiction in this. In quantum mechanics, you can have decoherence in the box, but not for the box. This is most easily to see if you look at the universe as a whole which can't decohere just because there's nothing to trace out.
Just a thought. We know that we get spectra for different elements when they are heated very hot. Is it possible that there are other invisible orbitals of sorts, for particles. When the particle goes from a high energy orbital to a low orbital, we see electron, muon, or tauon or neutrinos of different flavors.
ReplyDeleteThe father of a friend of mine, who I also know, just died of Covid-19. Two other people I know are in intensive care for Covid.
ReplyDeleteI studied physics. My professor was Lowel Brown in 1974.
ReplyDeleteIn 1994 he published his book, "Quantum Field Theory".
I saw him on TV. The host would not let him use the word "Quark".
He had that same surprised look as when he learned how ignorant and lazy I was, and yet I passing his exclusive little honors class. Guessing my way through college made life easy for me.
The most important thing to guess is the probability of each guess being right.
Right now I am guessing my chances with Covid 19.