Sunday, June 16, 2019

Book review: “Einstein’s Unfinished Revolution” by Lee Smolin

Einstein’s Unfinished Revolution: The Search for What Lies Beyond the Quantum
By Lee Smolin
Penguin Press (April 9, 2019)

Popular science books cover a spectrum from exposition to speculation. Some writers, like Chad Orzel or Anil Ananthaswamy, stay safely on the side of established science. Others, like Philip Ball in his recent book, keep their opinions to the closing chapter. I would place Max Tegmark’s “Mathematical Universe” and Lee Smolin’s “Trouble With Physics” somewhere in the middle. Then, on the extreme end of speculation, we have authors like Roger Penrose and David Deutsch who use books to put forward ideas in the first place. “Einstein’s Unfinished Revolution” lies on the speculative end of this spectrum.

Lee is very upfront about the purpose of his writing. He is dissatisfied with the current formulation of quantum mechanics. It sacrifices realism, and he thinks this is too much to give up. In the past decades, he has therefore developed his own approach to quantum mechanics, the “ensemble interpretation”. His new book lays out how this ensemble interpretation works and what its benefits are.

Before getting to this, Lee introduces the features of quantum theories (superpositions, entanglement, uncertainty, measurement postulate, etc) and discusses the advantages and disadvantages of the major interpretations of quantum mechanics (Copenhagen, many worlds, pilot wave, collapse models). He deserves applause for also mentioning the Montevideo interpretation and superdeterminism, though clearly he doesn’t like either. I have found his evaluation of these approaches overall balanced and fair.

In the later chapters, Lee comes to his own ideas about quantum mechanics and how these tie together with his other work on quantum gravity. I have not been able to follow all his arguments here, especially not on the matter of non-locality.

Unfortunately, Lee doesn’t discuss his ensemble interpretation half as critically the other approaches. From reading his book you may get away with the impression he has solved all problems. Let me therefore briefly mention the most obvious shortcomings of his approach. (a) To quantify the similarity of two systems you need to define a resolution. (b) This will violate Lorentz-invariance which means it’s hard to make compatible with standard model physics. (c) You better not ask about virtual particles. (d) If a system gets its laws from precedents, where do the first laws come from? Lee tells me that these issues have been discussed in the papers he lists on his website.

As all of Lee’s previous books, this one is well-written and engaging, and if you liked Lee’s earlier books you will probably like this one too. The book has the occasional paragraph that I think will be over many reader’s head, but most of it should be understandable with little or no prior knowledge. I have found this book particularly valuable for spelling out the author’s philosophical stance. You may not agree with Lee, but at least you know where he is coming from.

This book is recommendable for anyone who is dissatisfied with the current formulation of quantum mechanics, or who wants to understand why others are dissatisfied with it. It also serves well as a quick introduction to current research in the foundations of quantum mechanics.

[Disclaimer: free review copy.]


  1. Talking about satisfaction and QM, as a pure non-specialist I've never been able to grasp if quantum superposition was only an abstract concept or a real concrete situation... I get sometimes terrible headaches about the possibility than the abstract can be taken for the concrete thing... The problem is that I don’t necessarily find an answer and then the question remains suspended in my mind during months, or even years ;-)

    1. Superposition is merely a fancy word for 'sum'. The reason you use it is that in quantum mechanics if you have two solutions of the evolution equation, then any (weighted) sum of those is also a solution. That is very really and concretely so, in the hope that this answers your question.

      It's really entanglement that's the relevant concept, not superpositions, but you can't discuss entanglement without first discussing superpositions (at least I don't see how).

    2. Like many quantum mechanical phenomena, whether quantum superposition is real depends on how you look at it. For example, particles can certainly behave as if they were in two places at once in certain situations. However, it's not too hard to argue that this is just an illusion.

    3. Sabine, Peter,
      Thanks for your answers.
      Eventually, if you know some good and readable literature - I mean not too technical - about these questions...

    4. Here a recent video explaining superposition (sum) as solution to a linear equation. Instead of the Schrödinger eq. the heat eq. is used as an example, but this does not matter in this context. (They transform into each other by Wick rotation t → it as described here). All over this should be a “not too technical” account about Fourier and the animations are stunning.

  2. The Montevideo Interpretation of Pullen et al is similar to Penrose's idea. The reduction of a wave function is associated with a metric shift, even if that is very tiny. They are clever ideas IMO, though I question whether they really solve the measurement problem. Penrose claims this is an R-process that absorbs information and in so doing “solves” the measurement problem. As a sort of phenomenology these Penrose-Pullen interpretations are I think potentially useful. After all that is really how we think of things in the Bohr's so called Copenhagen Interpretation.

    To be honest I am not in favor of adulterating QM in various ways. The real question should be “how is there a classical world?” Is classicality a real hard aspect of the universe, or is it just a really good approximation to large N quanta? In either case, how is it these arise? It would appear so far the answers to these questions in part depend on various quantum interpretations.

    The only way I can imagine Bohm QM having much to do with quantum gravity is with holography. With a system close to the horizon of a black hole time dilation will make that system on the reduced space of 2 dimensions appear nonrelativistic. Bohm's QM has troubles with ordinary special relativity and it is then harder for me to think how it can be relevant to quantization of general relativity in 3-space plus time.

    1. The real question should be “how is there a classical world?

      I don't understand the reasoning behind this type of question. It seems to suggest that the quantum realm, about which we have the least amount of observational evidence, and none at all pertaining to the actual physical processes on that scale, is somehow more fundamental than the classical realm which is more closely observed and better understood. At best, this seems misguided.

      Even if you could make a sound logical argument that the quantum realm is somehow more fundamental because it is smaller, the lack of detail about processes on that scale should and does mean that no unambiguous relationship between the two scales can be inferred. And I would argue there is no sound logical argument that can be made for the proposition that the smaller the scale, the more fundamental it is.

      That the heuristically derived mathematical formalisms of QM, which are stochastic in nature, can be projected, to some extent, onto the classical scale says nothing about the physics of either scale and everything about the malleability of math formalisms. The idea that somehow physical reality is quantum in nature when we don't have a good model for quantum scale physical processes and how they relate physically to the classical scale seems mindbogglingly obtuse to me.

    2. The quantum domain is fundamental to the structure of atoms, nuclei and particles. The classical world is not so much. How the classical world emerges is strange, and the foundations of the quantum and classical worlds are remarkably different. We have the usual large N type arguments, but we really are not sure how it is there is a classical world.

      Coherent states form a symplectic subset in a Hilbert space of states. So they define a classical-like structure while remaining fundamentally quantum. This is one route, but sort of the opposite of the usual which is the scrambling of quantum states into maximal mixed states.

    3. The quantum domain is fundamental to the structure of atoms, nuclei and particles.

      So you say. But it is that peculiar view that I'm criticizing as illogical. Simply asserting as factual, your belief that the quantum realm, about which little is known, is "fundamental", does not constitute much of a response to that criticism.

      I suppose you might argue in the context of the big bang theory that the quantum realm came first and therefore is "fundamental". But that just means you are deriving your belief about the fundamental status of the quantum realm from your belief in the incoherent big bang creation myth.

      I don't find that line of reasoning to be scientifically or logically compelling. Do you have anything else? What I'm asking, I guess, is for you to explain why you believe the quantum realm is "fundamental"?

    4. Why is the quantum realm somehow more fundamental? I would say the atoms are more nuts and bolts that compose a large object, say the coffee cup I am drinking from. It would be strange for me to say this coffee cup or the coffee are somehow more fundamental than the atoms which compose them. Also the shorter the wavelength the higher the energy, which means more quantum modes and their symmetries are excited.

      A classical system is thought of as having wavelength λ = h/p, the deBroglie equation, but the wavelength is illusory for a standard classical object, say a football in flight, is composed of many quanta that are in destructive interference. So there really is no measurable wavelength. All the quantum states are in a maximal mixture states, or density of states sum_ie^{-E_i/kT}|i><i|.

      I indicated earlier that quantum mechanics is a case of L^2 measure over states. Relativity is also similar and there is a sort of duality of convex sets with 1/p + 1/q = 1 for p = q = 1. A straight up probability system has p = 1 and so the dual is p = ∞ is any deterministic system. Deterministic systems, whether classical mechanics or the theory of Turing machines or Church-Turing thesis computations etc are not measure theories. This is why any measure assigned to them diverges. What role classical systems have is very strange then. Whether classical systems, or equivalently their dual as pure stochastic systems, actually exists is dependent upon one's interpretation of quantum mechanics.

      Now there is a bizarre middle ground, where coherent quantum states, coherent states such as laser states of light, define a symplectic manifold in Hilbert space. This is then a sort of classical-like substructure, and coherent states are condensates of bosons that in the case of gravitons are a massive tensor-sum of states similar to entanglements that then form classical space or spacetime. Fermions do not do this in particular, except when they have low energy (low temperature) wave functions that are similar enough so they then couple to form bosonic like states. Superfluid physics, superconductivity and the Higgs mechanism with Yukawa potentials are of this form. So there is a lot to think about here. Massive N-number of degrees of freedom or modes in quantum physics that are mixed states have an ambiguity in how they define classicality, but the opposite with coherent states define a classical-like substructure. There is a whole lot more I could go into with the algebraic geometry of quantum states. Learn the Jamiołkowski-Chow theorem and isomorphism with densities and entanglements.

      The big bang has a hell of a lot of data to support it. There was a period lasting about 15 minutes (as I recall) where the early universe, starting about 3 minutes after the initial event, was hot enough to fuse hydrogen into helium. The amount of helium predicted matches perfectly with helium abundance. The CMB was predicted and found and …, the data supporting big bang is huge. There is a fair data set that goes further in supporting inflation. Now people can refuse to accept this, which might in time put one in a position of refusing to admit the heliocentric universe or that the surface of Earth is spherical.

    5. Lawrence Crowell: But that isn't the "Big Bang", it is evidence consistent with the notion that at some point the universe was about as dense as 3 minutes AFTER the Big Bang. The Inflation before that is speculation requiring the unicorn Inflaton particle, to explain something that can be equally explained by the Big Bounce, which never collapses to a singularity.

    6. It depends upon what you call the big bang. The term big bang was a rather flippant remark made by Hoyle or Bondi in a discussion about cosmology. The Boltzmann constant can be written as k = 8.6×10^{-5} ev/K and so with the equi-partiton theorem E ~ kT the temperature of the universe at the time of EW breaking with E ~ 1TeV was then T ~ 10^{16}K. So the universe was very hot, it was rapidly expanding and this is the big bang. The big bang is where things thermalized at the end of inflation with so called reheating.

      Inflationary cosmology is where the action of a scalar field acted to inflate space. This field is no different from a dilaton field or the scalar field in the Dicke tensor-scalar relativity. It is also if scale invariant a form of conformal transformation that expands space. During this phase the early quantum state of the universe had constant entropy or information in any comoving volume.

      I am not sure why you and I think others keep bringing up the big bounce. Observations illustrate the cosmos is not going to recollapse, but is rather accelerating its expansion. The cosmological constant is some residual field effect of the the inflaton, maybe where it now sits in a physical vacuum configuration removed from the high energy vacuum, or called false vacuum, of the early inflationary universe. So there is no forthcoming “big crunch” coming that will initiate the next expansion in a bounce. There is maybe some prospect for phantom energy with a big rip. After the big rip maybe things recycle, but this is different and phantom energy is really inflation and accelerated cosmic expansion “on steroids.”

    7. Starting with the sitting duck, the big bang (or Lambda-CDM model) provides a ludicrous description of physical reality that bears no resemblance to observations. None. There is no big bang, inflation, substantival space, time and/or spacetime, dark matter, or dark energy in observed (empirical) reality. At best they are model-dependent inferences, at worst, ad hoc parameters added to cover the model's failures.

      The claim that the LCDM model has data to support it is specious, since it is freely parameterizable, allowing it to be fit to whatever the data might be. That math trick is at least as old as Ptolemy. The claim then, that because the model can be fit to observations, it must therefore accurately represent physical reality, is logically fallacious.

      The question of what is or is not fundamental depends a great deal on the meaning given to fundamental.

      Here are three definitions from the Cambridge-English Dictionary:

      1. forming the base, from which everything else develops: 2. more important than anything else: 3. being the most basic or most important thing on which other things depend

      I don't think the quantum realm fits readily into any of those unless you take a strictly reductionist view that smaller means more fundamental, or the big bang motivated argument that the quantum state came first. Even by your example of a cup of coffee, claiming that atoms are fundamental to the cup and the coffee, obfuscates the fact that the same atoms differently configured would not comprise either a cup or coffee, and neither cup nor coffee exist except as potentialities of the atoms. It is the manipulation by humans, of the atoms that is fundamental, not the atoms themselves, which are unlikely to ever spontaneously generate a cup of coffee.

      Most of your discussion comprises an elaboration of your preferred mathematical model of the quantum realm. My problem with that is twofold:

      One, on a very fundamental level mathematics is not physics. As the Ptolemaic example demonstrates, it is possible to take a completely erroneous qualitative description of physical reality and mathematically massage it into agreement with empirical observations.

      Two, the standard mathematical model of the quantum realm has no realistic qualitative model of the underlying physical processes whatsoever. That is why there is a proliferation of scientifically useless metaphysical interpretations of the model. The standard quantum model simply does not provide an accurate or reasonable physical (as opposed to metaphysical) account of quantum processes.

      Your inability or unwillingness to distinguish between empirical reality and your model leads to this kind of assertion, one that constitutes a theoretical statement having no basis in physical reality:

      ...coherent states are condensates of bosons that in the case of gravitons are a massive tensor-sum of states similar to entanglements that then form classical space or spacetime.

      Having no basis in empirical reality (gravitons and substantival spacetime are not observed phenomena) means it has no scientific value, except perhaps as a theoretically motivated hypothesis. In science, hypotheses are only provisional and subject to empirical verification. It is not clear from your discussion that you appreciate your model's extremely limited congruence with physical reality. In other words, you do not seem to realize that your model is not, scientifically speaking, realistic.

    8. Dr. Castaldo,

      The big bounce eliminates the singularity but it is as reality challenged as the standard model of cosmology. It does not rest on empirical observations except for the limited sense in which the cosmological redshift-distance relationship is assumed to be caused by some form of Doppler-shift. That is to say, the "expanding universe" is strictly speaking, an assumption of the model, not a consequence.

    9. bud rap: Inflation doesn't rest on empirical observations either; like dark matter it is invented to explain a mathematical issue; correlations in the CMB that shouldn't exist.

      Running the "expansion" film backward is all fine and good, but it doesn't make logical sense to claim it runs all the way to a singularity; that is an inference that cannot be proven, and to justify it we must invent new physics, an unobservable inflationary field.

      And as you say, the red-shift may just look like a Doppler shift, perhaps photons losing energy as a function of distance traveled is caused by some other effect, quantum or otherwise. After all there are still fundamentals we don't understand, about the muon, neutrino mass, and who knows what else after we figure those out.

      I don't have a problem with proposed models, but I think it is premature to make any claims about the beginning of the universe, especially involving new fields/particles, as long as there are mysteries left to resolve in fundamental physics.

      And if the situation remains premature forever, so be it. "We don't know" is a perfectly valid scientific position. Claiming we know without any proof just creates a roadblock for searching for the truth, because those that would search for it are dismissed as fools searching for something already known. Oh and just maybe they are so mentally deranged, they are dangerous fools.

    10. Dr. Castaldo,

      ...I think it is premature to make any claims about the beginning of the universe...

      In that seemingly reasonable statement you have already accepted an assumption of dubious provenance. I would suggest that it is/was premature to assume the cosmos comprises a "universe" with a singular gestation event.

      The assumption, that the cosmos is a coherent, singular, unified entity, to which a universal frame can be attributed, was the basis for Friedmann's solutions to the field equations of General Relativity. The resulting FLRW equations, and the subsequent assumption that the cosmological redshift is indicative of a Doppler-type shift, provide the basis for all modern cosmological models.

      The apparent inability of modern cosmologists to reconsider those assumptions severely limits their ability to even conceive of realistic alternatives to their reality-challenged model(s). The current situation is similar to the pre-Copernican era. You can't get from a geo-centric to helio-centric model by incremental steps. The geo-centric, perfect-circle model had to be abandoned before progress could be made.

      Until the unitary-cosmos and Doppler-shift assumptions are scrapped, modern cosmology will continue to be an empirically baseless compendium of mathematical fantasias. The old computer aphorism applies here: garbage in, garbage out. You can't get a realistic model of the cosmos from fundamentally erroneous assumptions. It is likely just an unrealistic oversimplification to assume that it is possible to have a singular, unitary model of the cosmos.

      As to the cosmological redshift itself, a crude redshift-distance relationship can be calculated from the standard GR equation for a gravitational redshift, by applying it to an expanding spherical wavefront of light emitted by a galaxy. You simply have to plot repeated calculations for cosmologically significant distances by calculating, using a reasonable mass-density estimate, a mass term for the enclosed volume at each increment. Try it; you can do it on a spreadsheet.

    11. Comparisons with the Ptolemaic system are not appropriate. That system was upheld as a matter of theology. The only near comparison is with perturbative QFT as similar to "nests of epicycles" vs nonperturbative QFT that is not fully understood.

      The basic big bang is very heavily supported by data. From galactic red shift to helium abundance and so forth the big bang is fairly well supported by observation and measurement. I can't "make" somebody accept this any more than I can make a creationist accept evolution. Inflation is moderately supported, or at least aspects of the CMB conform to that. It has a way to go. There are also a range of inflaton potential models and debates over eternal inflation vs "one time 'slow roll'" inflation.

      A gravitational red shift from galaxies does not give one cosmic red shift. Yeah, sure --- one can run any model one apes up, but that does not make it terribly valid. You have to make various additional assumptions about the gravity of more distant galaxies. If galaxies further away have greater gravitation, and thus more redshift, that blows apart homogeneity.

    12. bud rap: .. it is/was premature to assume the cosmos comprises a "universe" with a singular gestation event.

      While I agree it is premature to posit a "gestation event", it is not illogical to assume that what we see is what there is; it would take extraordinary evidence (in this universe) to convince me there were any other universes, or cosmos, or whatever you wish to call it.

      It IS logical to presume this; as logical as presuming anything else we have never seen any evidence for, like winged garden fairies with human levels of intelligence, Peter Pan's Tinkerbell, just do not exist, and are figments of human imagination.

      So I disagree with you, a unitary cosmos makes sense. Observations show us the structure of our galaxy and others close by and for as far as we can see the same shapes hold; the same spectra hold, the whole model seems to apply "universally". It doesn't matter if red-shift is due to the expansion of the universe or GR or something else, it acts as a reliable estimate of distance, and our local versions of gravity and chemistry apply for as far as we can measure them. I don't apply the word "universe" to anything else; I imagine there is stuff out there red-shifted beyond our current detecting ability, and I would include that in the "universe". But I presume it follows the same rules we find locally, because we have looked for evidence it can be otherwise, and there isn't any.

      I don't feel any compulsion to declare there was a start of the universe, or a start of time, any more than I feel compelled to posit an end of time. Although it IS a curious subject to discuss, and fun, in the end I don't care much more than I care about the prospects of the USA in Olympic Volleyball. I like volleyball and care a little, but any answer has zero impact on my life, as far as I can tell. But in truth there is nothing to do with such knowledge.

      I do 100% endorse and recommend a healthy daily dose of pointless entertainments, I would not stop anybody from their speculations and might even enjoy following their logic. But in the end fundamental physics is important to me for the practical predictive applications it may have in us manipulating matter, capturing or generating energy, nano-engineering materials, and otherwise becoming better masters of our physical realm. If somebody figures out how the universe began along the way, great, that would be entertaining, but it would never be my goal to find out.

    13. Dr. Castaldo,

      You misunderstand what I mean when I say the cosmos is not unitary - my fault for not being clearer. I am simply being consistent with Relativity Theory by insisting that there is no universal reference frame that can be wrapped around the cosmos.

      Like you I do not believe that there are other "universes" out there - there is only the cosmos we observe. It is the imposition of a universal frame that renders the standard model of cosmology both nonsensical and absurd.

      You don't believe in Tinkerbell - good for you, neither do I. But if you accept the unitary cosmos model you have to believe in the big bang, inflation, substantival spacetime, dark matter, and dark energy. As is the case with Tinkerbell, there is no empirical evidence to support the existence of those entities and events.

      There is of course, model-dependent inferential evidence, but if that's all you've got, then you are simply engaged in an elaborate game of circular reasoning.

      Like you, I see a cosmos of unknown and likely unknowable extent - that is all the evidence supports. Presuming to impose a "universal" model on the unknown and unknowable is just a scientifically useless exercise in mathematical hubris. As science, the big bang is a miserable failure.

      I do not have an issue with your strictly utilitarian view of science, but I do not share it. My goal is not to master the physical realm but to understand it. I just want to know how it works - I was born that way. I used to think all scientists shared that goal. It turns out I was, at least for this era, quite mistaken. Most do not.

    14. It is difficult to say whether the universe is unitary. In a strict sense spacetime physics is not, for the transformations of quantum operators is hyperbolic and not elliptic. The transformation of the operators a and a^† with rapidity or spacetime transformation is

      b = a cosh(gs) + a^† sinh(gs)
      b^† = a^† cosh(gs) + a sinh(gs),

      where we may think of the g as an acceleration or gravity g^2 = ½ (K^a∇_aK^b)(K^c∇_cK_b). As an exercise it is not hard to show that [a, a^†] = [b, b^†] = 1. This means information is conserved, or in a classical analogue the phase space volume is constant. Things get messy when you couple these operators with those that are elliptic and strictly unitary.

      If information is conserved, which we will generically for now call unitarity, does this mean we can say the entire universe is unitary? That is not easy to address. Quantum entropy is constant with unitarity so we may then state δS/δt = 0 and so thermal energy is constant δE/δt = TδS/δt = 0. The problem is that in general we can't define energy globally on a spatial manifold. Energy is the conjugate of time, and this equation means we can define a “time vector field,” or in general a time Killing vector K^t∂_t. However, that is not always the case. Cosmologies are Petrov type O solutions with no Killing vector fields, at least none defined globally. So was have a problem that both energy and time are defined locally. What is local is hard to specify, maybe it means just within an world pocket embedded in an inflationary manifold.

      If we have unitarity it means we are talking about “open world” quantum mechanics. This is by way of contrast to a closed world quantum physics where the Hilbert space of states is completely defined. This is really a fascinating topic, but I will defer on it for now. However, an open world quantum cosmology is one where the quantum error correction problem is potentially better defined.

    15. It is interesting that Bogoliubov argued just the other way around in his approximation, the basis for Bose-Einstein condensate (BEC). He first assumed [b,b^†]=1 and then got the transformation. (see e.g. here pdf eq. 4.20 - 4.23)
      Somewhere you said that L² theories are central - I absolutely agree. The imaginary i attributed to time in QM and the signature of the metric in SR/GR fit remarkable well. And we need to square the amplitudes otherwise we would run into a paradox as Landé pointed out.
      Looks like that QM and GR are made for each other – when we take QM measurements seriously and make local spacetime patches all the “time”.
      Why all the “time”? Well, how do we accelerate? We have to exchange momentum all the “time” otherwise we are just free falling.

    16. @Reimond: Looks like that QM and GR are made for each other They are made for each other, and I think in a way versions of the same thing. It might be said that QM has elliptic transformations and that GR has hyperbolic. An interesting ground to study are parabolic cylinder functions, which have a sort of "middle ground" aspect to them.

    17. Lawrence,

      Comparisons with the Ptolemaic system are not appropriate. That system was upheld as a matter of theology."

      That's an assertion without any real merit. The big bang system is upheld as a matter of mathematicism - different belief system, same result. The real similarity between the two systems is that both are freely parameterized. Despite erroneous assumptions, both can be massaged into agreement with observations. Which brings us to this:

      "The basic big bang is very heavily supported by data. From galactic red shift to helium abundance and so forth the big bang is fairly well supported by observation and measurement."

      To repeat your "supported by data" claim without engaging the counter-argument presented suggests that you are unable to do so. Here again is the counter-argument:

      The claim that the LCDM model has data to support it is specious, since it is freely parameterizable, allowing the model to be fit to whatever the data might be. That math trick is at least as old as Ptolemy.

      Do you wish to argue that the BB model is not freely parameterizable? That's going to be a tough one given all the free parameters the model has.

      "A gravitational red shift from galaxies does not give one cosmic red shift...You have to make various additional assumptions about the gravity of more distant galaxies. If galaxies further away have greater gravitation, and thus more redshift, that blows apart homogeneity."

      Nice straw man argument. I specified that for an expanding spherical wavefront model you have to incorporate "...a mass term for the enclosed volume at each increment". Do you not understand what that means?

  3. I know Lee's thinking on the nature of time, but he simply doesn't get anywhere with it. There is no definitive description. On the other hand, if you take time as emergent from a wavelike field of energy, then you automatically get Lorentz time dilation and even deduce how mass slows time. This does mean a preferred reference frame I'm afraid, the taboo of modern thinking in relativity. If you can get over the shock of entertaining this notion then a whole new view of general relativity opens up which agrees with Einstein's geometric interpretation of space time, but allows space time to be described as a function of time only. The abstract physical geometry disappears. leaving only the energy of time, the energy of this wavelike field. Apply this idea to entanglement, CPT Symmetry, the evolution of the Big Bang and even, yes, the double slit experiment, and you arrive at a consistent, physical and intuitive explanation for all of them and much more. You should read "The Binary Universe" (A Theory of Time), if you are frustrated by our lack of progress in understanding time and seek a description that fits accepted science perfectly as well as clarifying many of today's conundrums in physics.

  4. Good review Dr H. I really enjoyed the book. What are your thoughts on realism? Also, am I correct in guessing that you and Dr S had many great discussions on this in the past? Any stories you can share?

  5. Could you say a little more about the ensemble interpretation and Lorenz invariance? How does one show that a specific theory leads to a violation of Lorenz invariance? What characterizes theories that do so versus theories that don't have this consequence?

    1. Jeff,

      To say the obvious, if you want to know more about Lee's ensemble interpretation, I recommend you read his book. Alternatively, look at some of his papers, link is in my blogpost.

      In Special Relativity (and General relativity) you are free to chose your time and coordinate system. If you postulate a notion of "simultaneity" or "size" you have fixed either time, space, or both. This is permissible only if you demonstrate that this is really derived from a physical system and not an extra postulate. (Eg, a single particle has a rest-frame of course. This doesn't violate special relativity.)

      This in and by itself does not result in a conflict with observation, but the Standard Model (SM) is Lorentz Invariant. This means if you pursue an idea that isn't Lorentz invariant you must at the very least explain how this doesn't result in conflict with observation by failing to reproduce the SM.

      Lorentz invariance is a peculiar symmetry in that it's hard to keep small. If you have any of it, it tends to spill over into all length-regimes, roughly speaking. It's actually remarkable we do not see any Lorentz-invariance violation (at least not so far), I find it very puzzling. But in any case, that's what it is, and if you want to fiddle with quantum mechanics, you better make sure you don't cause any conflict with observation.

    2. Smolin in his ensemble interpretation as I “understand” it has to collect information from “similar” systems all over the universe “simultaneously” - these are his grey and white squirrels - and this ”… would be highly, highly nonlocal."
      One has to take care when using a nonlocal, hidden mechanism that it does not lead to usable superluminal information transfer in our world.
      QFT, well relativistic QFT does this quite clever. It uses Lorentz-invariant Lagrangians in the path integral, but a single path can very well be non-local, off mass shell, i.e. a virtual particle.
      QM(/QFT) is linear, unitary, thus probabilities, energy/momentum are conserved as in Smolin´s energetic causal sets (ECS). I never understood why Smolin just invents new rules like squirrels and ECS and does not accept QM.
      If we assume that GR cannot be quantized as Penrose does, i.e. spacetime cannot be in superposition, but QM particles can, then there must be a threshold how fat a Schrödinger cat can grow. This would result in myriads of measurements - observer independent triggered ones, which is realism (“The moon is there even if …”)
      Newton told us that differentials are linear, now we have another linear element - QM - let us stick it together in a process. Every time a photon hits our retina or heats a stone means a reduction has been triggered and entropy increases. And this is a time that flows towards the future (and not only on Fridays ;-).

    3. You are voicing some of the caveats that have appeared in my head. I am not sure what Smolin is meaning with ensemble. When I think of ensemble I think of a density matrix ρ = sum_φ p[φ] |φ)(φ| for a mixed state or a complete mixture of the form ρ = sum_φ (1/N) |φ)(φ| for N states φ sum to in this stat-mech microcanonical form. I am using parentheses for carrot signs in bra-ket notation because these blogs do not “like” them. Mixed states are formed as a sum of pure states, but we can't determine in any way which pure state the system is in, say analogous to the exact microstate of a stat-mech system, for that would violate quantum contextuality. So building up pure states this way is problematic to say the least.

    4. Reimond wrote, "QFT, well relativistic QFT does this quite clever. It uses Lorentz-invariant Lagrangians in the path integral, but a single path can very well be non-local, off mass shell, i.e. a virtual particle."

      Yeah, and I am always surprised that people do not emphasize this point more strongly. You can create a particle here and annihilate it there where "here" and "there" have a space-like separation.

      (This is obvious by looking at the VEV.)

      It's the pesky 1/sprt(omega) in the relation between the field and the creation and annihilation operators that does it. But you need that factor to get Lorentz covariance of the field right.

      Tom Banks goes into this in detail near the beginning of his Modern Quantum Field Theory: A Concise Introduction. All this would violate signalling locality... except, as Banks explains, introducing antiparticles saves the day.

      I had a chat with Weinberg about this back in the late 1970s, and Steve, at that point, had not thought through the fact that the field does not really create a particle localized to a single point.

      I'd like to see the discussion in Banks' book become the norm in all QFT books: this really bothered me when I studied QFT under Weinberg, and the resolution is very pretty: Lorentz covariance seems to destroy signalling locality, but antiparticles come to the rescue and relativity survives unscathed.

    5. “If we assume that GR cannot be quantized...” So, if we can, anything’s possible; ain’t false premises grand.

    6. PhysicistDave,

      The discussion here in Banks' book is repeated in A. Zee’s QFT book (I.8.14) - unfortunately not on display, and here “antiparticles come to the rescue”.
      I also find that all this non-locality under the hood needs to be much more emphasized.
      Steven Weinberg e.g. here or here seems to be bothered now.

    7. A double slit, a Mach-Zehnder device, an Elitzur-Vaidman bomb test or even a Wheeler´s delayed choice experiment does not lead to cognitive dissonance any more if we fully embrace this “non-locality” (*) in the unitary evolution.
      The trick is a clear separation of the unitary evolution (U) from the reduction (R).
      QM is linear and provides a sneaky mechanism to check out the neighborhood by summing up superpositions (or paths in the path integral) to get the correct probabilities. And this is all what is done in U – there is no particle “flying” (**).
      Is this ”…just an illusion”? I prefer to think of U taking place in a “frozen spacetime”. Why frozen? Well, if we assume that spacetime cannot be in superposition there simply cannot be any backreaction as long as matter is in a superposition.
      In U more and more QM particles get entangled until a threshold is reached and a reduction (R) is triggered. From the superposition of entangled particles one outcome is picked with QM randomness and only then a tiny, tiny backreaction to the spacetime metric can happen.
      If we regard our world as a result of myriads of these tiny steps of U and R all makes sense – it must be that QM is linear (***) and GR is non-linear – this very tension drives the dynamics. These U/R steps take place everywhere and all the “time”.
      Why do QM particles get entangled? Because they are indistinguishable and need to be anti-/symmetrized. And a disentangling step is needed otherwise the whole universe soon would be one entangled mess.(****)

      This is a process of tiny U/R steps where matter and spacetime are interacting all the “time”. This is in contrast to what Smolin has in mind where he just picks time (in some preferred frame) as fundamental and space as illusion.

      And here a recent interview with Tim Maudlin fully embracing “non-locality”.

      (*) “non-locality” in quotes to code spooky-action-at-a-distance, but no faster than light signaling.
      (**) let us take the wave/particle duality for real and clearly separate between U and R.
      (***) QM must be linear/unitary otherwise we would lose stuff. Since it is linear, superpositions (sums) are valid. An entangled state is one that cannot be written as a single product state. In U these entangled states (a sum of products) grow ever bigger until finally in R one entry of the sum is picked. This is e.g. in a double slit experiment the spot where the photon will have transferred its energy to the particles of the screen.
      Nothing is “flying” through these kinds of experiments in U. A classical path only exists for a classical object. But of course, a classical distinguishable object like a squirrel consists of a lot of indistinguishable QM particles and holds together by myriads of U/R steps.
      (****) which indeed many think it is.

  6. I don't know anything yet of the ensembles/squirrels interpretation, but I think the number of interpretations of QM (including those that try to include gravity) could be rivaling the number of Protestant denominations.

    I've been in the path-integral (sum-over-histories) campground for many years now, but who knows what's right.

    1. To misquote Yoda, "there is no right." If all interpretations give the same predictions, they cannot be distinguished by experiment, only by esthetics ... which is easiest to work with? which has fewest extraneous assumptions? which satisfies more of the symmetries that we observe in physics?

    2. So arguing over which of the various interpretations is "right" is like arguing over how many angels can dance on the head of a pin (unless it actually disagrees with the other interpretations in what predictions it makes, in which case it is very likely to be wrong). But often a new interpretation contributes a useful way of thinking about quantum mechanics. So I'm looking forward to reading about the quantum squirrels.

    3. To Smolin's credit he does suggest predictions from his ensemble theory that would distinguish it from at least most others.

    4. I don't see any point or have any interest in criticizing other interpretations. I like the Sorkin/Dowker approach — quantal stochastic processes with "a generalized measure on a space of histories" — basically because my early background is stochastic processes, and I think quantal histories are cool. Other than that, I've got nothing.

  7. Thank you for your review. You have identified what bothered me in the exposition of the concluding chapters of the book. Page 268 "The real ensemble formulation is a relational hidden variable theory" Later saying (pg 269) " us a relational hidden variable theory that is also background independent and which realizes space and locality are emergent."
    This is the aspirational goal of quantum physics in his analysis as I understand it.

  8. @ Peter Shor

    One quantum bit is probably the absolute minimum amount a quantum squirrel can bite. ,)

  9. If anyone cares, I developed, conceptually and a bit programmatically, the Reflective Path integral, where there are histories and (their reversed versions) futures, but is not yet quite mathematical. I did this a year ago, but has a connection maybe to recent news on "mirror matter".

  10. It could be helpful to remember what Albert Einstein wrote on quantum mechanics: [1] "The ψ function is to be understood as a description not of a single system but of a system community [Systemgemeinschaft]. Expressed in raw terms, this is the result: In the statistical interpretation, there is no complete description of the individual system. Cautiously one can say this: The attempt to understand the quantum theoretical description of the individual systems leads to unnatural theoretical interpretations, which immediately become unnecessary if one accepts the view that the description refers to the system as a whole and not to the individual system. The whole approach to avoid 'physical-real' becomes superfluous. [Es wird dann der ganze Eiertanz zur Vermeidung des ‘Physikalisch-Realen’ überflüssig.] However, there is a simple physiological reason why this obvious interpretation is avoided. If statistical quantum theory does not pretend to describe completely the individual system (and its temporal sequence), then it seems inevitable to look elsewhere for a complete description of the individual system. It would be clear from the startt that the elements of such a description within the conceptual scheme of the statistical quantum theory would not be included. With this, one would admit that in principle this scheme can not serve as the basis of theoretical physics.”
    [1] A. Einstein, Out of my later years. Phil Lib. New York 1950 Seite 498

    1. As I indicate in a post in this thread dated yesterday 6/17 there is a problem with trying to identify a simple pure state from a statistical mixture or ensemble. This would violate quantum contextuality.

      As great a giant Einstein was in many ways, he was in my thinking wrong on this. Hawking was great as well, and remember 8 or so years ago he sounded off about rogue aliens coming to Earth. I am not sure what happened with Hawking's brilliant mind there, but I think a few neurons went off the rails. There have been other instances of this with other great people.

      Quantum mechanics hits one with a cognitive dissonance concerning the objective aspects of this world or universe. To be honest I doubt there is any theorem that can tell us decisively whether quantum mechanics is ψ-ontic or ψ-epstemic, though upholders of interpetations on either side of this fence are adamant on their positions.

  11. Sabine,

    On the subject of Smolin's speculations.... Do you have an opinion on his view of a Darwinian multiverse? So far as I understand it, he has argued (in TIME REBORN and elsewhere) that the collapse of a star into a black hole in one universe is the birth of another universe. With universes giving birth to each other one might think we have a new form of the old Hoyle "steady state" theory at a meta level. But not really, because the multiverse Smolin has in mind allows for natural selection. Some universes survive a long time, other's don't. Some of those that last, create another generation of universes, others presumably are barren. So there is a survival of the fittest, where fitness means a set of natural laws that optimally produce black holes.

    Anyway ... do you think he is onto something here and, if so, is it something that would ever be testable?

    1. Smolin's idea is that the universe is optimized for the production of black holes. Clearly our universe would contain more black holes had the LHC been able to produce them, which it did not. So, the LHC has ruled out this idea.

      More seriously, no one has any idea how to make formal sense of the idea that universes are born inside of black holes with somehow random variations on the constants of nature. It's the kind of idea that works well for a popular science book, but in practice doesn't lead to anything useful, so scientists don't pay attention. (Same story of course with Tegmark's Mathematical Universe or Penrose's consciousness stories and so on.)

  12. I'm not a scientist or mathematician. And I don't see the loss of 'realism' as a problem. Or even as a loss; like 'physical' it was always just a word with no meaningful non-circular definition.

    1. Even though we can't 'know' (in my Kantian-esque view, anyway) any theory (or model) is 'true', the escape from 'realism' is weird to me. I made a word: 'simulationitis', meaning 'the belief or feeling that the material world is or could be a simulation, or reduced to information'. I cringe when I see things like "we are lining in a simulation" or "everything is information".

    2. That currently trendy 'simulation' stuff is just techno-Gnosticism, logically indistinguishable from religion. And saying that it's all 'information' is equally vacuous because no one can even define information in the absence of knowledge. These are just the latest attempts to reduce consciousness to an epiphenomenal side effect; and they're empty talking points, without any explanatory power.

      In losing realism, all we give up is the idea that in the absence of conscious experience, reality somehow happens on its own: particles bounce around in a stream of time. But we've known for 100 years that outside of conscious experience there are no 'particles' and no 'now'. Reality is made up of experiences, and what those are "made" of, we can't say. Is that Copenhagenism?



Comment moderation on this blog is turned on.
Submitted comments will only appear after manual approval, which can take up to 24 hours.
Comments posted as "Unknown" go straight to junk. You may have to click on the orange-white blogger icon next to your name to change to a different account.