Thursday, April 30, 2020

Book Review: “A Philosophical Approach To MOND” by David Merritt

A Philosophical Approach to MOND: Assessing the Milgromian Research Program in Cosmology
By David Merritt
Cambridge University Press (April 30, 2020)

Don’t get put off by the title of the book! Really it should have been called “A Scientific Approach To MOND,” and I am so glad someone wrote it. MOND, to remind you, stands for Modified Newtonian Dynamics, which is the competing hypothesis to dark matter. Dark matter explains a whole bunch of astrophysical observations by positing a new type of matter that makes itself noticeable only through its gravitational pull. MOND instead postulates that the laws of gravity change on galactic scales.

The vast majority of astrophysicists today think erroneously that dark matter has better support in observational evidence, but Merritt cleans up with this myth. Let me emphasize that Merritt is not originally a philosopher by training. He worked for decades in astrophysics before his interest turned to the philosophy of science in recent years. His book is not a verbose pamphlet, as – excuse me – philosophical treatises tend to be, but it’s an in-depth scientific analysis.

What makes Merritt’s book special is that he evaluates the evidence, both for MOND and the standard model of cosmology, according to the most widely accepted criteria put forward by Popper, Zahar, Musgrave, and Carrier. The physicists among you need not despair: Merritt’s book has an excellent (and blissfully short) introduction into the philosophy of science that contains everything you need to know to follow along.

The book is extremely well structured. Merritt first analyses MOND as a phenomenological idea, largely formulated in words, then MOND in the non-relativistic case, then relativistic completions, and then the hybrid theory of dark matter and modified gravity that can be interpreted as a type of superfluid dark matter. In each step, Merritt examines how the theory fares with respect to confirmed predictions and corroboration, which he summarizes in handy tables.

Along the way he cleans up with quite a number of mistakes that you encounter all over the published literature. Yes, this is hugely troubling, and it should indeed trouble you. There is for example the idea that MOND cannot explain the CMB power spectrum when indeed it made a correct prediction for the second peak, whereas dark matter did not. In fact, astrophysicists had to twiddle with the dark matter idea after the measurement to accommodate the new data. Another wrong but wide-spread conviction is that modified gravity has somehow been ruled out by observation on galaxy clusters.

Having said that, Merritt clearly points out that MOND (or its relativistic generalizations) has certain problems, notably the third peak of the CMB is a headache.

The most interesting part of the book, though, is that Merritt demonstrates by many quotations that astrophysicists who prefer dark matter are confusing the predictive power of a theory with the ability of the theory to accommodate new evidence.

I have found this book tremendously useful, though I want to warn you that this is clearly not a popular science book. The book is full with technical detail. However, I believe that the biggest part of it should be understandable for anyone who has an interest in the topic. There are some parts which will be incomprehensible if you don’t at least have an undergrad degree in physics, eg when Merritt goes on about the Lagrangian formulation of the relativistic completions. But I don’t think that these parts are really essential to understand Merritt’s argument.

But. Of course I have a “but”!

I think that Merritt does not pay enough attention to the problem that MOND, because it is non-relativistic, is incompatible with an extremely well-confirmed theory – General Relativity –, and that we have to date no relativistic completion that does not run into other problems with evidence. This means that MOND, simply put, does not live up to the current scientific standard in the field.

Let me be clear that this does not mean that MOND – as an approximation – is wrong. But I believe the lack of a controlled limit to recover General Relativity is the major reason why so many physicists presently reject MOND. I find it somewhat unfair to simply disregard the scientific standard. The standard is there for a reason, and that reason itself is based on evidence, namely: Certain types of theories have proved successful. MOND is not that type of theory, and no one has yet managed to improve it. It only reproduces General Relativity in the cases where we have precision tests by postulating that it does so, not because there is an actual derivation that demonstrates this is consistently possible. This is an extremely non-trivial problem.

This problem is solved by the hybrid version that can be interpreted as superfluid dark matter. In Merritt’s evaluation this option receives mediocre grades. But of course this is because he does not appreciate the need to remove the tension between MOND and general relativity to begin with. Superfluid dark matter does this.

In summary, I think that everyone who has a research interest in astrophysics and cosmology will benefit from reading this book. And I think that physics would much benefit from a similar analysis of inflation and other hypotheses for the early universe, quantum gravity, theories of everything and grand unification, and quantum foundations.

Disclaimer: Free review copy


  1. You write
    "it is incompatible with an extremely well-confirmed theory - General Relativity"
    But e.g. the equivalence principle is not broken by the MOND, or?
    Personally, I don't understand why the equivalence principle has to apply to long distances or low densities. Since the equivalence principle is only a “local theory”?
    If the inertial mass becomes smaller, but the gravitational force does not, then the acceleration would be greater.
    If the gravitational and inertial masses would differ at long distances and low density, then MOND would not be incompatible with general relativity, or?

    1. "But e.g. the equivalence principle is not broken by the MOND, or?"

      Depends on how you define equivalence principle. If I take the equivalence principle to mean general covariance, then MOND trivially violates it.

  2. This sounds like a really good book, not to mention that I am a fan of MOND to begin with. With our endless, dreary weather in the northeast US there's plenty of time for reading.

  3. Violates or Enhance it?
    This is the question that I would like to understand.

  4. I have a very naive, maybe even stupid, question. What is the problem with GR reducing to Newton in the (intermediate) weak-field limit; and then Newton reducing to MOND in the extremely weak field limit? Why do you need to go all the way from GR to MOND in one step, as it were?

    1. Newtonian gravity just does not reduce to MOND. I don't know why you think so. I guess Milgrom would be thrilled if it did, but then if it did that would be in conflict with loads of evidence, so it's not going to happen.

    2. Nobody asked me, but I think you are using different concepts of "reduces to". In mathematical physics, I think that means you have a general set of equations (General Relativity) and in some circumstances some terms of those equations become negligible, and when you drop them out you have a reduced set of equations (Newton's Gravity) which work usefully some of the time.

      In Mond, you have Newton's equations, and then you materially alter them under certain conditions, without any reason for saying the difference in calculations between them is negligible (it isn't).

      As an analogy, you have water which obeys equations of fluid dynamics, and then you have a phase transition to ice, which obeys rigid-body (plus Theory of Elasticity) equations. Those sets of equations are incompatible and do not reduce one to the other as physicists/mathematicians use that term, since one assumes a crystalline structure and the other does not. (For glass, on the other hand ..)

    3. Hi susskind — and you do know that’s a famous name in string theory, yes?

      What I think you were trying to ask is this: What’s wrong with simply using GR for intense gravity fields, Newtonian for average gravity fields, and MOND for very weak gravity fields? Can’t the math be set up to allow the three to fade into each other, with GR at one end handling black holes, Newtonian in the middle handling solar systems, and MOND at the other end handling galaxies?

      The answer, alas, is no.

      You can set up a smooth transition between Newtonian and MOND. That is precisely why it’s called modified Newtonian dynamics! However, in terms of your question, the “fade in” is the other way around: Newtonian gravity pops out of MOND in cases where distances (not field strength per se) are modest, e.g. on the scale of our solar system. For the scale of galaxy clusters and galaxies, though, there simply is no scenario where unaided Newtonian gravity gives result anything remotely like what is actually seen in observations. MOND is just a weird, unique, very arbitrary-seeming addition to Newtonian equations that has gained interest for one reason and one reason only: It works remarkably well for predicting many aspects of what is actually seen. But it’s not perfect, and it definitely does not work for everything.

      Einstein’s GR has a similar relationship with Newtonian gravity, except that in this case the “fade in” depends more on field strength, not distance. The black hole extremum in particular results in vastly different predictions Also, some of the predictions of Newtonian gravity are just flat-out wrong, such as how much gravity bends light paths. Still, it’s not too inaccurate to assert that GR “fades into” Newtonian gravity as field strength (again, versus distance) gets weaker.

      What you have, then, is a classic love triangle: Both GR and MOND are “competing” for the fade-in love of Newtonian gravity… and the competition flatly does not work well, either conceptually or mathematically, though I gather lots of folks have tried. For more on that the Merritt book would be a great resource.

      I hope that helps a bit.

  5. But, what is the type of evidence that precludes Newtonian gravity reducing to MOND in the very weak-field limit?

    And if you are convinced that a MOND-type theory cannot be obtained as an extremely weak-field of Newtonian gravity , in which sense do you expect it could stand as a dynamical theory describing the dynamical evolution of galaxies?

    1. Newtonian gravity just does not reduce to MOND. How do you want to do that? Do you seriously think people wouldn't have noticed if Newtonian gravity gave you an 1/r force instead of a 1/r^2 force? What are you even talking about?

    2. If there is something about mond that is right, then somewhere at the 'edge' of the solar system (~0.1 light year.) There would be a transition region... from ~Newtonian to ~'mond' gravity.
      (where the ~ sign is to indicate that these are (hopefully) approximations to some unifying underlying theory.)

  6. Sabine,

    I bought and downloaded the Kindle version of this excellent book, and can honestly say I've viewed every page... for about 1/4 second each for most.

    But that was enough to get a feel for the contents. It does indeed start off very philosophically, but then rather abruptly shifts to full science mode. I love the precise details, tables, and especially curve-fit data point charts later in the book. Also, little details such as the discussion of the cosmogenesis and star history of lithium-7 are just the kinds of details I delight in finding in a good science book.

    Without mathematical support it's easy for our minds to estimate poorly the levels of improbability associated with various types of data-to-prediction curve fits. Straight-line fits, meh... but complex curves with considerable detail, such as I recall seeing even decades ago (drat... I cannot find the article!) in the first round of renewed interest in Milgrom's idea? Those quickly become astronomically unlikely. So when such curves are generated by extremely minimal hypotheses, attention should be paid, indeed.

    And precisely as you have stated in your review, poor attention to the implications of unexpected complex curve fitting is exactly the misperception that this book is trying to correct. Dark matter theory has for decades been extraordinarily weak on prediction. Far from predicting anything, it has instead seemed as if every time a new observation comes along, the data once again "breaks" the dark matter model in some way. Folks then say hey, no problem, we'll just tweak the software again to make it fit the new data fit again.

    Bleh! If Kepler had taken that approach we'd still be using angels to push planets around, just with more sophisticated software to track how the angels like to dance. In a worst-case version this approach to modeling theory becomes not much more than a lookup function for a nicely organized database of observations. That in fact would often be the more honest ways to explain such theories. Sure, they describe the data -- they are the data! -- but such data-capture models tell you next to nothing about what's really going on at a deeper level.

    The real message is always hidden in the Kolmogorov complexity of the data, in the ability to reduce all of it as if by magic to some insanely compact set of equations (Kepler! and Newton, and Maxwell). Making some rather loopy but very simple modifying assertion -- Milgrom's idea that gravity falls off too slowly, which even he initially thought was a pretty wild idea -- that nonetheless ends up predicting remarkably good curves for phenomena for which there was not yet any data back in 1986 definitely falls into this category, even if it does so in a very strange fashion.

    I have no idea what to make of MOND and its relationship to one of my favorite theories of all time, general relativity... and I love that! It tells me that Mystery still exists, that there is something deep and likely profound that everyone is still missing. Those are the kinds of conundrums that can give rise to entirely new ways of understanding physics, just as happened in the emergence of quantum mechanics from a few nominally "simple" loose ends related to atom electromagnetic emission and absorption.

    P.S.: A side effect of scanning the entire book is that I now know why Sabine took a sudden interest in the Raven Paradox.

    1. Ha, yes, you are right, it was Merritt's book that reminded me of the Raven paradox :)

  7. Sabine,

    Can I clarify something? Is it right to say that MOND can be formulated in a way that is compatible with SR, but not with GR?

    1. David,

      No, by definition. MOND is a non-relativistic limit which means exactly that it's not relativistically invariant and thus incompatible with both GR and SR.

    2. But someone developed a relativistically invariant version of MOND, didn't they? That is what I meant.

    3. There are relativistic completions which have MOND as limiting cases, but, as I said, all the known ones have problems. Why not read the book?

  8. Is there only ONE reasonable MOND so far? And if so, why?

  9. Sabine,
    I completely agree with David Merrit and your enthusiasm, but not with your critic: "MOND, because it is not relativistic, is incompatible with an extremely well confirmed theory - general relativity -, and that to date we have no relativistic completion which does not come up against other problems of proof. This means that MOND, in simple terms, does not live up to the current scientific standard in the field "

    IMO, one should not oppose GR to MOND because MOND is not a theory but a marvelous empirical rule which explains itself several empirical laws of the galaxies, and which remains to be explained by a theory, which indeed will have to integrate the effects of GR, either by modifying GR or by supplementing it with some cosmological effect, or both. It is also the job of this future explanation to theoretically deduce an interpolation rule between the Newtonian and worldly regimes. Not the job of the MOND rule.

    Currently, the MOND rule (from which several empirical rules observed in astronomy are deduced) has somewhat the same status as the Bohr quantization rule for the angular momentum of atomic electrons, which explained Balmer's empirical formula so well but also expected d 'be explained by the MQ (which is itself still awaiting explanation). At that time Bohr's rule contradicted known physics, but had for it to integrate h, the fundamental element of new physics. Likewise MOND, but less clearly, flirts with this incongruous element which calls for a new cosmology: the acceleration of expansion.
    At that time, physicists did not reject Bohr's rule as most of them rejected MOND, sometimes with incredible violence. This is where there is a degradation of the standards of science.

  10. MOND is a “sorta maybe” that I keep on a back burner. To be honest developments do not appear to be leaning in its favor.

    There seem to be some questions people are asking that illustrate some problems. An elementary form of MOND has a modification of gravitation with a potential Φ = -GM/r – gM ln(r). The first of these U = -GM/r is the standard Newtonian potential. The term V = -gM ln(r)is a modifying term and g is the gravity coupling constant associated with that term. The force is given by F = -∇Φ. These potentials can be computed, but in different dimensions of space. The force defines a charge, or in the case of gravitation mass, given by

    q = ∫F∙dA

    where the area A is a Gaussian surface. Appealing to spherical symmetry we can see this is q = 4πGM for the force F ∝ 1/r^2 that divides the r^2 in the area of the Gaussian surface. In 2 spatial dimensions the Gaussian surface is a circle and this requires the force be F ∝ 1/r. It is then not hard to see the potentials.

    If MOND were due to this log potential, where would this have come from? That is where things are odd. Hossenfelder’s suggestion there is a type of phase change between this type of potential and matter is interesting. This suggests maybe something similar to a holographic principle. The phase with the potential might then be a sort of trapped phase, such as seen with pure water that can be liquid below freezing. These fields trapped on the horizon might have been on the horizon of the inflationary cosmos, and with the rapid phase change of reheating it is them plausible a small part of the gravitational field is “trapped” in this inflationary phase on a surface of two dimensions.

    Other questions appear to revolve around general relativity. One can have 2-space plus time general relativity. The BTZ black hole and AdS_3/BTZ-BH correspondence is an example of this sort of physics. With low mass, energy etc the 2-space plus time GR will recover a form of gravity with the F ∝ 1/r for of potential. If gravitation or general relativity has a small aspect of this 2-space plus time phase, then the low energy gravity would be this sort of modified gravitation.

    There are other aspects to MOND, and there are sorts of “form factor” terms people insert. These terms to model galactic structure more fully I can’t comment on too much. At this time I would say that MOND appears to be a bit on a defensive strategy. ΛCDM pretty much dominates the subject of dark matter. If I were to put the contest between ΛCDM and MOND in analogous war history terms, I would say with the American Civil War it is a bit like the siege of Atlanta 1864, where the Confederacy was in trouble, or maybe Napoleon at Leipzig in 1813, yeah he was in trouble. Without some sort of shot in the arm I am not sure if MOND will make much further impact.

  11. Sabine,
    Maybe for the sake of us non-physicists you could comment on where the wiggle comes from that would a theory to make predictions different from the standard model without violating GR. Doesn't GR itself predict the orbital speeds? If so, how could MOND predict different speeds without contradicting GR?

    1. Andrew,

      I am not sure I understand the question. First, MOND is a non-relativistic limit. It will never reproduce GR because it isn't generally covariant. What you are looking for is a theory that in the non-relativistic limit reduces to MOND but at acceleration reduces to GR. Finding such a theory that, in addition, isn't incompatible with evidence already, turns out to be extremely difficult. Despite almost 30 years of trying, no one has found one.

    2. (1) MOND is a empirical rule which cannot reproduce GR (or any theory) just as Bohr's condition could not reproduce classical electrodynamics
      (2) Perhaps the explanation of this rule is not a new gravitational theory but a new cosmological model. A departure of relativistic cosmology (FLRW) which would follow from another GR's interpretation, not from a modified gravity (i have worked in that way but i am just a "philosophe"). In this way the theory which support MOND is not just a modified theory of gravity but a result of a new cosmological effect in another vision of GR.

    3. I was responding to Sabine's apparent major objection to MOND: it "is incompatible with an extremely well-confirmed theory – General Relativity". I was making the simple-minded rejoinder that of course it is incompatible with GR - how else could it make different predictions?

      Perhaps the problem is not that MOND is merely incompatible with GR, but that the incompatibility is too profound?

    4. When I say the two imcompatible, I don't mean that MOND makes different predictions in some parameter ranges, which it obviously does, I mean that we already know that there is no limiting case of MOND in which we will recover General Relativity.

    5. "...there is no limiting case of MOND in which we will recover General Relativity. "

      So in certain domains MOND makes predictions which are known to be wrong? Because it conflicts there with GR which is known to be right?

    6. Andrew,

      Basically, yes, though that's a weird way to put it. It's more that MOND does not make predictions for certain cases because we already know it doesn't apply. Simple example: No one is going to correctly predict the emission of gravitational waves from binary systems with MOND because MOND is a non-relativistic limit. Now you can insist on using it and then say "it makes a wrong prediction" but that would be silly because we already know you should not use it.

      It's the same as proclaiming that "Newtonian gravity is wrong" because you tried to use it to predict time dilation. It's wrong not because the theory itself is wrong, but because you have used it in a regime in which we know it is not a good approximation.

  12. And now I ask you: why are you assuming that I want to reduce anything to MOND, what gave you such an idea? I am just trying to understand the problem from someone who thought hard about it. And I said at the outset that my questioning is naive. I was also assuming that if you write that such a reduction would conflict with loads of other evidence, you would then give us a couple of examples instead of pointlessly ask me back another rhetorical question.

    You don't want to reply, don't publish my comment. It's not a problem; but if you wish to reply do it seriously.

    1. "And now I ask you: why are you assuming that I want to reduce anything to MOND, what gave you such an idea?"

      You did. You first wrote:

      "What is the problem with GR reducing to Newton in the (intermediate) weak-field limit; and then Newton reducing to MOND in the extremely weak field limit?"

      And then again:

      "And if you are convinced that a MOND-type theory cannot be obtained as an extremely weak-field of Newtonian gravity"

      Now you write:

      "You don't want to reply, don't publish my comment. It's not a problem; but if you wish to reply do it seriously."

      I have responded to your questions twice already, which I frankly think is enough.

  13. Jean-Paul errata :"between the Newtonian and worldly regimes" read Mondian regime.
    And of course, ACDM cannot predict DM effect because in his frame any explanation of discrepancy with GR is a "discovery" of DM object. As Terry Bollinger says : "[such "theories"] describe the data -- they are the data!"

  14. I wonder if we fell back into kindergarten...

    OK, I see your level of understanding.

    1. I have spent time trying to understand and politely respond to your ill-informed and badly phrased questions. In response, you insult me. I have no patience for shitty behavior like this. You need not bother submitting further comments, I will no longer approve them. Good bye.

  15. "There is for example the idea that MOND cannot explain the CMB power spectrum when indeed it made a correct prediction for the second peak, whereas dark matter did not."

    Not really. This refers to Stacy McGaugh's prediction of the height of the second peak. Stacy, of course, is a proponent of MOND, but there is no MOND in that prediction in any way, shape, or form. All he assumed was that there is no dark matter. The prediction otherwise uses standard CMB physics, no MOND at all. Of course standard cosmology can explain the CMB spectrum; this is probably the best success story of modern science. But there are 6 or so free parameters (many of which are constrained by non-CMB stuff). The default values for some of the parameters used by some people didn't get the height of the second peak right. Big deal.

    In retrospect, it's a coincidence that McGaugh got the height of the second peak right. However, his simple model goes vastly wrong on the third peak, as he is the first to admit. So, now that we've seen more than two---about seven---peaks, MOND definitely cannot explain the power spectrum of the CMB. Not only that, but a) no-one knows how to so such a calculation in MOND, and b) the one thing one can say is that if one naively extends MOND to handle structure formation, structure forms way too quickly, which even the MOND people see as a problem.

    Maybe Stacy will weigh in here.

    I haven't read this book yet, but have read a similar article by Merritt from a few years ago. The main problem I see is that he attacks a straw-man version of standard cosmology. You didn't mention that; maybe he has toned it down in the book compared to the article. Also, he doesn't apply the same standards. For example, in the article he criticizes dark matter and dark energy as ad-hoc ideas, whereas Milgrom's idea of a change in gravitational attraction below some acceleration is almost the textbook example of an ad-hoc explanation, i.e. one invented solely to fit the data with no theoretical motivation whatsoever (though in the meantime people have tried to come up with explanations for MOND).

    I actually think that there is something to MOND phenomenology, thought it is not clear what the explanation is. It is a problem that the mainstream cosmology community is not as aware of this as they should be (though to be fair most MOND phenomenology has nothing to do with cosmology). However, I fear that they will be put off by strawman attacks, since they are usually an indication that one has no better arguments, though that is definitely not the case with MOND.

    I'm looking forward to reading the book and am interested in seeing whether it is more balanced than his article.

    1. Phillip Helbig:

      Some of the papers on MOND that I’ve read are quite technical, and were hard for me to understand, but I gather the model doesn’t have a ‘seed’ mechanism equivalent to Dark Matter that can produce density variations in the early Universe, that eventually coalesce into galaxies and larger structures. Googling a minute ago, I was quite surprised to see a Discover magazine article from February describing a computer model that claims that structure formation can proceed without Dark Matter, which I’ve linked below.

    2. As mentioned above, many of the points in the book were made earlier by Merritt in an article (linked to above). My main criticism of that article (and, hence, also a major criticism of the book, which in the meantime I have read) is that he attacks a strawman version of ΛCDM.

      Alas, some points are too long to make in a comment box, so I direct readers to my long article on MOND vs. &LambdaCDM (published in the latest issue of The Observatory), a major part of which is a point-by-point rebuttal of Merritt's arguments. I hope that I have provided enough historical context and references to convince the reader that Merritt's arguments against ΛCDM are mostly wrong.

      My main motivation for writing a critique of the article is that I think that MOND has something to offer (though I'm not sure what). MOND phenomenology needs to be investigated more and needs a convincing explanation (whether via something like MOND or something else. However, strawman attacks hurt the cause much more than they help it. I have actually met people who were interested in MOND, maybe even had done some work on it, then moved on, turned off by the over-the-top rhetoric.

    3. Phillip Helbig4:10 AM, December 07, 2020

      "over-the-top rhetoric"
      But No.1 in the OTT rhetoric charts remains claiming that the universe is fine-tuned without a single shred of evidence.

    4. You'll be pleased to hear I've accepted an invitation to discuss finetuning with Luke Barnes which, I am being told, will be recorded so you can all enjoy it. (It's supposed to happen some time in January.)

    5. I'll definitely be watching that. It will be interesting to finally hear what the eminent Dr. Barnes' evidence is that it is physically possible for the universe to be fundamentally different than it has been observed to be. Given he repeatedly failed to provide evidence when asked on this site.

    6. The evidence is still there even if you ignore it.

      Or should I say that ignorance of evidence is evidence of ignorance?

      Yes, the paper is coming along fine, next on the list. First I have to implement some good suggestions by the referees for my fifth single-author refereed-journal paper this year. When you've managed to actually publish something, then people might listen to you.

    7. Phillip Helbig2:41 AM, December 08, 2020

      "fifth single-author refereed-journal paper this year"
      Nobody wants to work with you? ;)

      " When you've managed to actually publish something"
      Ouch! Like you, Luke Barnes and Geraint Lewis, I have never come up with any ideas worth publishing.

      Anyway, the truth is the truth. There is no evidence that it is physically possible for the fundamental laws and constants to be any different than observed. FT is pure assumption, however much Bible Boy dresses it up in Bayesian Statistics.

      And that new CUP book, "Fine-tuning in the Physical Universe", I guarantee you that it is just one example after another of "fine-tuning" in models not reality, noticed after the fact. No-one ever has discovered anything in Physics based on the concept of fine-tuning. It's impossible. Yes, we can work out the probability of a lottery win precisely because the situation is simple, but working out how likely a gas giant is in a solar system, say, based on the laws of Physics. Impossible.

      They can be counted that's all, and afterwards Martin Rees can publish a book and appear all over the telly claiming "Look! There aren't very many. Fine-tuning!"

    8. Sabine Hossenfelder12:20 AM, December 08, 2020

      "You'll be pleased to hear I've accepted an invitation to discuss finetuning with Luke Barnes"

      You know Crazy Luke has published a raft of Bayesian theory testing BS funded by the Templeton Foundation, don't you?

      This is the kind of nonsense he will be throwing at you:

      The only way to deal with this lunatic is to take apart all the papers he has published on fine-tuning and list up all the specific unjustified assumptions; get these criticisms published; and, then ask him to his face why he's making these specific unjustified assumptions.

      Otherwise, he will just claim what he claims in the linked paper:
      "If the fundamental theories of modern physics cannot in principle justify a normalizable prior distribution over their free parameters, then we have bigger problems than fine-tuning. "

      i.e. Crazy Luke gets to say Baby Jesus' Papa made the universe or all Physical theories are problematic.

      In particular, your usual argument of saying there is no measure to allow probabilities to be calculated will not wash. Barnes is a lying crank but he is a well-funded, lying crank who can churn out manure by the wheelbarrow-full.

    9. Sabine Hossenfelder12:20 AM, December 08, 2020

      Here is another blatant lie that Bible Boy and his partner in scientific corruption, Geraint Lewis, have put out there recently and which has been quoted on religious looney sites:

      A plane of points representing values of the fine structure constant and strong force not known to be physically possible, except for the few pixels labelled "You are here".

      Geraint Lewis has discovered, like Barnes, that being a bent Physicist is more profitable than being an unsuccessful one.

      Why don't you ask our pie-loving, supernatural mythologist, Crazy Luke, why he's spreading lies like this?

    10. I don't think he's lying, I think he doesn't know the difference between science and religion. The problem's more common than you may think.

    11. Organisms are well adapted to their environmental niche. Religious people see that as evidence of design by a creator. Scientists explain it via evolution. A hundred years ago, Evans would have been denying the fact that organisms are adapted to their environment because religious people used that fact to support a wrong idea. In a larger sense, the competence of the speaker has no bearing on the truth of what he says. The biggest fool can say that the Sun rose this morning. That doesn't make it dark outside. Evans's rants against the fact that Barnes is religious demonstrates that he doesn't understand this basic principle of logic. Presumably it is to detract attention from the fact that he has no scientific arguments to back up his case.

      What one can debate is whether there is evidence for fine-tuning. Many highly respected physicists have published extensive claims that there is. I am in good company. Yes, one can have a different opinion, but needs to present evidence for it, rather than Evans's "it just ain't so".

    12. Sabine Hossenfelder12:37 AM, December 09, 2020

      Ask Crazy Luke where on that chart of fine structure constant vs. strong force has been shown to be physically possible by the standards of physics, or even the standards of commonsense, or even the standards of not being a complete liar.

      It's only the pixel or two corresponding to actual measurements in the 1 observed universe and their error bars, but just watch him lie through his teeth about it.

      You are not going to be entering into a discussion with a scientist interested in the truth, but a 5th columnist mental case determined to hang onto his delusions and armed with Bayesian theory testing.

      I would prepare yourself for an experience that will be as infuriating as discussions with a climate change denier in a MAGA cap.

      Good luck!

    13. Phillip Helbig1:32 AM, December 09, 2020

      When will you ever write anything relevant? We're not talking about evolution and arguments by authority are not evidence.

      There is zero evidence that the fundamental laws and constants of Physics can be any different than observed.

      Against my better judgement, I read Crazy Luke's fine-tuning paper and sure enough it was a pile of assumptions about naturalness, and about being able to assume probability distributions about theories.

      Crazy Luke is claiming, based on this bunch of unjustified assumptions, that he can calculate the probability that among all the various physically possible universes (mmmm, there's just 1 we know of actually) he can calculate that the chances of life are extremely small.

      They are the ravings of a complete mental lunatic who is in desperate need of a lobotomy.

      Is this what you think is physical evidence of fine-tuning???
      Assume that other values of fundamental constants are possible for no reason and then assume the probability distribution so you can get the answer you want??

      It's unscientific drivel.

      "Many highly respected physicists have published extensive claims that there is. I am in good company. "

      Claims with *zero* evidence.
      Martin Rees is the UK Astronomer Royal. He's also a moron. That's your good company - halfwits and morons.

      You have still failed to provide any evidence of universal fine-tuning. You have written 3 volumes-worth of irrelevant comments but provided no evidence. Still waiting for the evidence....

    14. Phillip Helbig1:32 AM, December 09, 2020

      Tell me a single point on this mapping of values of the fine structure constant against values of the strong force, that is outside the "You are here" circle, that is in anyway known to be physically possible:

      You can't. Because there are none.

      Because it is the gormless speculation of gormless morons.

      Luke Barnes and Geraint Lewis should join the circus - they are complete clowns. Now that is an empirical fact.

    15. Phillip Helbig1:32 AM, December 09, 2020

      " rather than Evans's "it just ain't so"."

      How many times do I have to teach you the same one step of logic? I am not saying "it just ain't so". I am saying, or rather the physical record is telling us, that there is no evidence of fine-tuning.

      Do you understand the difference? I've told you this simple point 100 times already.

      "one can have a different opinion"
      No, you can't. Physics is about evidence, not free speech. And you have provided no evidence.

    16. Sabine Hossenfelder12:20 AM, December 08, 2020

      Another trick Bible Boy will try on you, as he did in an online discussion with Sean Carroll, is question why he shouldn't oh-so-innocently research speculations about fine-tuning and multiverses.

      Fine, but you might want to point out he shouldn't be making announcements to newspapers about pure speculation, aka lying:

      According to Crazy Luke:

      “There are a whole lot of properties – like say how heavy an electron is – we don’t know why they are that weight. And if you change them, even slightly, the universe becomes too simple to make life.”

      and further down ....

      "That’s bad news for the multiverse, because it means there’s nothing particularly special about our universe"

      The universe *is* fine-tuned for life, but a multiverse is unlikely according to Crazy's simulations. It must have been God wot done it. QED.

      He essentially called the Sydney Morning Herald to announce to them that he had proved the existence of God by running a noddy computer simulation.

      As mad as a March hare.

    17. As I have mentioned before, a religious scientist who believed in fine-tuning made a guest post on Dr. Sean Carroll's blog a few years ago, and one of his arguments was a Bayesian one. An expert on Bayesian statistical analysis, I think his name was Bill Jeffries, disagreed in the comments, saying that Bayesian analysis doesn't work that way. You have to formulate a hypothesis before seeing the data, not formulate the hypothesis after and claim Bayesian support.

      An example was given which might be interesting to pose to Dr. Barnes. A machine spits out a numbered ping-pong ball when its lever is pulled. The machine has a theoretical capacity of 1000 balls, and each ball is numbered with three digits, 000 to 999. The lever is pulled and ball number 042 comes out. What is the Bayesian probability estimate for each of the following hypotheses: 1) all 1000 balls, numbered 000 to 999, were loaded into the machine; 2) only the first 100 balls, 000 to 099, were loaded into the machine; or 3) only ball 042 was loaded into the machine.

      As I understand it, the event has the Bayesian probability of one in a thousand, under each hypothesis, so the hypotheses cannot be distinguished by Bayesian reasoning, since the hypotheses were made after the result, and would have been adapted to any ball which came out, from 000 to 999. Only if one of the hypotheses was made before the lever was pulled could Bayesian analysis distinguish it. I wonder if Dr. Barnes appreciates this, as his hypothesis was certainly made after the event (universe) occurred.

    18. JimV11:00 PM, December 09, 2020

      He addresses similar questions in the paper here, but I should warn you this paper has nothing to do with Physics and everything to do with mental delusion:

      The problem is obvious in the paper's conclusion:

      " we can take the deepest known laws that describe how the physical universe works and look for ways to vary them. A particularly tractable way is to vary their free parameters. The set of solutions to those laws, usually represented by the set of possible initial conditions, is precisely the set of scenarios that the law deems “physically possible”. "

      He is confusing models with reality. Just because a theory has free parameters in it, doesn't mean all variations of those parameters represent "physically possible" reality. The models and theories are simply tools whose predictions have to be confirmed by observation.

      It's the same problem with the chart showing the strong force plotted against the fine structure constant. Almost all that chart, everything apart from the dot showing observationally confirmed reality, is not known to have any relevance to physically possible reality.

      When the Templeton Foundation is giving you hundreds of thousands of dollars to deliberately corrupt science, you can churn out quite a few papers, and still have plenty of dosh left over for delicious pies.

    19. Crazy Luke: "We turned it up to 300 and still did not kill the universe."

      '...the right amount (of dark energy) was considered good evidence for the multiverse. But after running thousands of hours of supercomputer simulations..Dr Barnes' team found that wasn’t the case at all.

      “We turned it up to 300 and still did not kill the universe. We tried, but we still did not see that star obliterating universe.”'

      Forget Spinal Tap's "these go to eleven", Crazy Luke "turned it up to 300 and still did not kill the universe."

      And he then actually called a major newspaper to announce that he had "turned it up to 300", strongly hinting that this proves baby Jesus' papa made the universe.

      Forget 300. This nutjob is off the scale insane.

      Ask Crazy about this, Dr. H. And ask him if he has ever considered psychiatric counselling. It's time for an intervention.

    20. Phillip Helbig1:32 AM, December 09, 2020

      "The biggest fool can say that the Sun rose this morning. That doesn't make it dark outside."

      Errmm, no. It would make it light outside. Keep working on that one.

      For any empirically confirmed Physics theory, one can outline the theory and its evidence and then link to published papers describing the theory and empirical evidence in detail. I could do it for E=mc^2, say.

      So do this for universal fine-tuning. If you can't do this (which you won't be able to because there is no evidence), yet you insist on believing in universal fine-tuning then you are simply insane.

      Put up or shut up.

    21. Dr. H., "I don't think he's lying,"
      He clearly is.

      In Crazy Luke's crank book, "A Fortunate Universe" and in his main fine-tuning paper he mentions God ("Who" below):

      "the fine-tuning of the universe for intelligent life renders our universe exceed-ingly improbable. There really is something to be explained. Precisely what (or Who) that explanation is, is left as an exercise for the reader."

      The conclusion has to be that he is making assumptions in his unjustified claims about fine-tuning because he thinks they will support his Iron Age primitive superstitions.

      Why don't you ask him?
      Is this ethical to abuse one's scientific credentials and lie like this?

      If he is prepared to write "God" and "Who" in scientific publications, clearly referring to the Christian primitive myth, then he needs to be held to account.

      Does he really think "turning" his noddy simulation "up to 300", shows that a supernatural being "created" the universe.

      I think you should question his sanity. He's clearly a few shrimps short of a barbie.

    22. Dr. H.,

      Also, as Crazy Luke has so far received around $550,000 from the John Templeton Foundation for

      "A Fractured Universe – Fundamental Physics, Symmetry and Life"

      "Galaxy Formation and the Fine-Tuning of the Universe for Intelligent Life"

      Can Crazy define what "life" is exactly and can he tell us a single empirically proven physical fact from this half a million dollars' worth of research, or did he spend most of it on pies?

      Is the reality that he is taking money from the Templeton Foundation because his agenda aligns with theirs to muddy the waters of science with claims made purely on unjustified assumptions, so that they can pretend baby Jesus' papa made the universe?

      There is no doubt that this is the reality.

      You are going to be entering a discussion with a 3rd rate scientist who tells lies in Philosophy journals and in popular "science" books, and only receives large grants and media attention because of his primitive, Iron Age monkey-man superstitions.

      Why not ask him what the Physical definition of "God" is, because he wrote it in a purported Physics book published by the CUP no less?

    23. Sabine Hossenfelder12:20 AM, December 08, 2020

      "You'll be pleased to hear I've accepted an invitation to discuss finetuning with Luke Barnes"

      When is this?
      Remember, the religious just want to be seen to be in an argument with credible physicists like yourself. They don't need to win the argument.

      Unless you get Crazy Luke to admit and say that his various fine-tuning claims are based on unjustified assumptions (naturalness, etc.) or assume the answer dressed up in Bayesian analysis, then he wins.

      Sean Carroll, in his discussion with Crazy Luke on some whacko religious channel, made the key point that Crazy Luke was just making assumptions, but it got buried in an hour of waffle. Crazy Luke wins.

      He must be forced to admit his claims are not backed by any Physical evidence.

  16. Phillip Helbig
    I know that MOND cannot explain all DM effects, but i was not aware of that : " the one thing one can say is that if one naively extends MOND to handle structure formation, structure forms way too quickly, which even the MOND people see as a problem". Could you give a reference ? Thank you

    1. Check out Bob Sanders's book. (Note that Sanders is one of the leading proponents of MOND.)

      Basically, in MOND the gravitational effect is increased at low accelerations (which is why it is an alternative to dark matter for explaining the acceleration above and beyond that due to visible matter). Hence, when fluctuations are small, they are in the MOND regime, and thus grow more quickly than in non-MOND cosmology.

  17. What is your opinion about TeVeS and STVG? What was the impact of GW170817 episode in MOND theories?

    1. GW170817 rules out TeVeS, mainly because TeVeS predicts a much longer Shapiro delay so that the electromagnetic counterpart should have arrived much later, even though the speed of propagation of gravitational waves and electromagnetic radiation in "free space" is the same in TeVeS. This was pointed out by, among others, none other than Bob Sanders, a leading proponent of MOND: Int.J.Mod.Phys.D, 27, 14, (2018) 1847027 (preprint at ârXiv).

    2. Thanks Phillip, apparently STVG can accommodate GW170817 as per

  18. I will for sure as I did for your "Lost in Math"...

  19. Is the "extra" gravity needed to explain the various observations that don't agree with "linearised" GR, uniform or not?
    E.g. do the stars orbiting on the edge of 2 very similar galaxies orbit at similar speeds?
    If so then one would expect a modified gravity solution. If not then one would expect a mixed dark matter and modified gravity solution or possibly just dark matter and GR. Would one not?

  20. One of the problems that I see with dark matter in general is the feeling as the matter is entrained in a Black hole it ought to be producing a signal identifiable as distinct from baryonic matter.
    I wonder if the author addresses dark matter /black hole encounters.

  21. Hi Sabine,
    I came across the interesting work of M. E. McCulloch where there he compares the results of Newton, General Relativity and MoND vs his Quantised Inertia for the wide binaries. The work is named "Testing Newton/GR, MoND and quantised inertia on wide binaries" (Link:

    I do not know (I am new on this Blog) if you have ever written an article or comment about McCulloch's Quantised Inertia before, therefore I would like to ask about your opinion. After MoND, a theory that appears today to receive significant attention (maybe I am wrong) is that of McCulloch.

    What is your view?

  22. Now the Bullet Cluster, and like structures, seem to be impossible to explain in MOND, or any other model, without Dark Matter. Yet, I seem to recall that Sabine stated somewhere, though not on this particular post, that MOND can handle the Bullet Cluster. Or maybe it was on Stacey McGaugh’s website that I read that the Bullet Cluster’s dynamics can be accommodated in MOND. I will have to do some searching. In the meantime, from reading a Forbes article on the Bullet Cluster by Ethan Siegel, I see that there is a modified gravity model that can account for the displacement of the gravitation potential in the Bullet cluster away from the center-of-mass of the system represented by the hot gases in the collision zone. I’m going to read that soon to see how they solve this seemingly intractable problem in a non-dark-Matter framework.

    1. Modified gravity has no problem handling the bullet cluster. Instead, the Bullet Cluster is a problem for LCDM. I explained this here.

  23. Sabine, thanks for pointing me to your 2017 write-up on the Bullet Cluster, where you emphasize that the actual center of mass of a system can be dislocated from the baryonic mass center in modified gravity models. This is the situation in galactic structures where the radius corresponding to the mean mass center is shifted outward due to the deviation from Newton in the outer galactic burbs. And, it’s in galaxies that MOND has it best track record. Looking at the composite red/blue image of the Bullet Cluster it’s apparent that the bulk of the ionized gas clouds are butted up against their respective galactic clusters where the mass is centered. So, while there is center-of-mass displacement from the Newtonian expectation, it’s perhaps no worse than seen in galaxies.

    I meant to mention the name of the modified gravity model referred to in the Forbes article in my earlier comment. There were actually two papers, which I just noticed this morning. One was authored by John Moffat and Victor Toth, titled “Cosmological Observations in a Modified Theory of Gravity (MOG)”, dated 2013. It was a tour-de-force of mathematical complexity, which I won’t even pretend to say I understood. I did wonder if perhaps they were introducing too many adjustable parameters to make things fit. In my mind the MONDian framework is better with just one parameter. The other paper dealt with MOG’s compatibility with the neutron star merger data from LIGO event GW170817 and the coincident gamma ray burst.

  24. As it was identified a while ago that free fermion e+/- 'pairs' ('99.9% of the universe!) are valid DM candidates/ contributors, and we now find <10 orders more than we thought, why are so many looking for exotic particles or alternatives? Have I missed something Sabine?
    This 'space plasma' surrounds all matter bodies and systems, more so when in in motion, it's n=1 so VERY hard to find with EM, has non-zero G potential, and we find it at <10^18/cm^-3 in 'empty' space near Earth, MORE dense in the '2-fluid' shocks. (mostly along with a lot of 'dust', & some molecular gas). Even Robert Dicke rather anticipated it's effects before SR took over. What am I missing?

  25. What is your opinion about A new relativistic theory for Modified Newtonian Dynamics
    Constantinos Skordis, Tom Zł

  26. I know GR has a good chump of successes on its shoulders, but it has also made a number of wrong predictions:
    - CMB homogeneity ( failed -> had hoc cosmic inflation to the rescue)
    - Galaxy/cluster dynamics ( failed -> dark matter to the rescue, though you can argue he culprit is "only" its Newtonian limit in this case )
    - Rate of cosmic expansion ( failed -> dark energy to the rescue)
    Add to the mix that it predicts the existence of objects which are full of pathologies ( singularities and information paradox for blzck holes for instance), that it is incompatible with quantum mechanics at a fundamental level and you can also very well be tempted to make this bold statement: what if GR was NOT the correct theory of gravity? What if GR "space-time" was a fantasy? A mere mathematical grid/construct with no physical reality? What if all our attempts to keep this paradigm alive is a dead end? What if we really need to restart thinking about gravity through a radical departure from GR and its postulates/representations about the nature of gravity?
    In this case it makes full sense as Merrit suggests not to appreciate the need to remove the tension between MOND and general relativity to begin with, but to embrace this very tension as the beginning of the end for GR as the correct theory of gravity for the aeons to come.

    1. Pablo, IMHO MOND is not a true contender for GR, even with all the flaws you correctly point to GR. Nevertheless MOND has a remarkable track record of fitting observational data regarding galaxy dynamics, is fully compatible with another very successful empirical law, Tully-Fisher relation (more than classical Newtonian theory), made good predictions, impossible to achieve with classical theory…. Unfortunately stops there, you have to compare it with Newton mechanics, because it isn’t relativistic, so can’t tackle more sophisticated things, like gravitational lensing and apparently has some issues dealing with clusters…
      So, I think there are no coincidences, universe is showing us MOND is right, but has to be a tip of an iceberg, to act has seed for a broader theory. There are relativistic extensions of “MOND like” theories, TeVeS for instance, but the apparent near simultaneous detection of gravitational waves and EM emissions from the same source puts a serious constraint to these theories, since most of them predict a reasonable delay between them, the Shapiro delay. One of the survivors from this test is STVG/MOG, from John Moffatt, perhaps is the right way….And a few days ago was presented another relativistic extension, that apparently complies with gravitational waves observational data (
      Still a long way to go, but I feel this is the right direction…

    2. "- CMB homogeneity ( failed -> had hoc cosmic inflation to the rescue)"

      This is the so-called isotropy or horizon problem. Whether it is a problem is not entirely clear. Trivially, it is not a problem for FRW models because those are homogeneous and isotropic by definition. One can ask why the Universe is well described by FRW, but that is a different question. It is certainly not a failing of GR, or of the big-bang theory. (If it somehow contradicts GR, tell me how.) In any case, not something MOND does differently, if at all.

      "- Galaxy/cluster dynamics ( failed -> dark matter to the rescue, though you can argue he culprit is "only" its Newtonian limit in this case )"

      Again, not a failure of GR at all. It is also not clear why you think that the Newtonian limit is relevant here. MOND is an alternative explanation for some types of dark matter, particularly in galaxies. Dark matter in clusters could be all baryonic, and even many MOND enthusiasts admit that dark matter works well in cosmology and MOND doesn't.

      "- Rate of cosmic expansion ( failed -> dark energy to the rescue) "

      Again, no contradiction at all to GR. The cosmological constant was literally in the first paper on relativistic cosmology. True, with a different value, but the idea that dark energy was invented to explain the rate of cosmic expansion is a myth, pure and simple.

  27. Dr Hossenfelder, just finished reading this book. I bought it not long after this blog post. Since I do not have a physics degree, it took me awhile to get through. I had to figure out which bits of the match I could understnad, and which to skip. Super interesting book, and I think your review is fair and accurate.

    I personally remain hopeful for a relativistic completion - because I think that would make sense (yikes - Okay - I think that would remain consistent with beauty). Best wishes

  28. Take a look at RELMOND, in a paper recently published, looks promising..


COMMENTS ON THIS BLOG ARE PERMANENTLY CLOSED. You can join the discussion on Patreon.

Note: Only a member of this blog may post a comment.