No, this new theory does not “cast doubt” on dark matter

A few highlights in the CMB power spectrum. a. The third peak indicates the total matter content of the Universe, both baryonic (ordinary) and dark. Planck has much better data for this peak than WMAP, strengthening the case for dark matter's existence and how much of it there is in the Universe. b. The long tail of smaller peaks is the small-scale fluctuations that gave birth to galaxies at later times. Awesome, right? c. The anomalous temperature fluctuations at the largest scales, first seen by WMAP in 2001, are probably the things getting the most attention today. [Credit: ESA/Planck Collaboration/moi]
A few highlights in the CMB power spectrum. a. The third peak indicates the total matter content of the Universe, both baryonic (ordinary) and dark. Planck has much better data for this peak than WMAP, strengthening the case for dark matter’s existence and how much of it there is in the Universe. b. The long tail of smaller peaks is the small-scale fluctuations that gave birth to galaxies at later times. Awesome, right? c. The anomalous temperature fluctuations at the largest scales, first seen by WMAP in 2001, are probably the things getting the most attention today. [Credit: ESA/Planck Collaboration/moi]

This Huxleyan vision of clean refutation buttresses one of our worst stereotypes about science. We tend to view science as a truth-seeking machine, driven by two forces that winnow error: the new discovery and the crucial experiment — prime generators of those nasty, ugly little facts. Science does, of course, seek truth, and even succeeds reasonably often, so far as we can tell. But science, like all of life, is filled with rich and complex ambiguity. The path to truth is rarely straight marked by a gate of entry that sorts applicants by such relatively simple criteria as age and height….
– Stephen Jay Gould, Eight Little Piggies

Dark matter is one of the most frustrating things in the Universe — at least for those of us who make it our life to study the Universe. Its presence is pervasive, it shapes galaxies, galaxy clusters, and the structure of the cosmos itself. Yet we don’t know what it is, thanks to the fact that it’s entirely invisible: light passes through it, and if it interacts with ordinary matter at all, that interaction is subtle at best. Our experiments to detect dark matter particles directly have either failed or produced ambiguous results.

So, it’s no wonder that some scientists have looked to alternative explanations for the same phenomena that dark matter produces. The OG of that movement is Mordehai Milgrom, who developed a relatively simple modification to Newtonian gravitational dynamics called (wait for it) MOdified Newtonian Dynamics, or MOND. Specifically, MOND was designed as a way to explain the rotation of spiral galaxies without the need for dark matter, a task it’s very good at. Proponents of MOND are a small but very vocal community in the astronomy and astrophysics community.

However, MOND has consistently failed to account for the other phenomena for which dark matter is the standard explanation.

  1. Galaxy clusters, despite their name, are not mostly made of galaxies: most of the mass is X-ray-emitting hot gas and … something else. That something else is what most astronomers call dark matter. MOND requires something like dark matter (extra heavy neutrinos, for example) to explain galaxy cluster dynamics, obviating the original motivation for the theory.
  2. Galaxies and galaxy clusters aren’t arranged randomly. Instead, they clump together, form along filaments, and make very large structures. This large-scale structure of the Universe is described well by a model containing dark matter, but not by MOND.
  3. As I’ve written about extensively, the strongest evidence for dark matter may be from the cosmic microwave background (CMB), the radiation left over from when the Universe became transparent. (Try these posts: “Our weird and wonderful Universe” and  “C is for cosmic microwave background“.) The fluctuations in the CMB allow us to put precise numbers on the major components of the Universe, including the amount of stuff that doesn’t correspond to atoms and other ordinary matter. The image above is a breakdown of the CMB spectrum; the third peak of that spectrum tells us that about 80% of the mass of the Universe is dark matter. MOND is a non-relativistic theory of gravity, so it’s unable to cope with cosmology, which requires something akin to general relativity. That isn’t to say there can’t be a relativistic version of MOND, a point I’ll get back to momentarily.

Arguably, the rotation of spiral galaxies is the least compelling evidence for dark matter (even though historically it’s the first to gain traction in the astronomy community), simply because galaxies are messy places. We don’t understand all the dynamics of gas and stars within galaxies, so it’s hard to pinpoint exactly where the dark matter is and how it behaves. However, I don’t want to overstate that problem: many astrophysicists are working hard on galactic dynamics, and most of them haven’t thrown up their hands over dark matter.

Relativistic generalizations

Over the years, several attempts have been made to generalize MOND into a relativistic theory of gravity. The one that caught a lot of eyes when I was in graduate school was TeVeS (which stands for Tensor Vector Scalar, not one of the better theory names out there). TeVeS failed to reproduce the cosmic microwave background as it’s observed, and besides which is a baroque and difficult theory to work with — and I say this as someone who happily worked with baroque and difficult theories back in the day. Arguably, aesthetics are a bad way to judge a theory, but it would suggest an untoward maliciousness on the part of the designer of the Universe. (That’s a joke, for use by physicists only.)

The latest of the relativistic MOND analogs comes out of the University of Salford in England, and was published in Physical Review D. I haven’t had much time to spend with the paper: that sort of thing is important to do properly, and as a full-time writer who needs to eat, I’m not sure I have it in me. (Oh, for a graduate student!) However, my main concern is that the authors are only concerned in the paper with reproducing MOND, which as we’ve already seen isn’t adequate as a dark matter replacement. I’m also a little leery of the central premise, in which our physical Universe is embedded in an abstract higher-dimensional mathematical space, which I’d like to see motivated by something other than “we can get MOND out of it”.

That’s not to say we should dismiss the theory outright. Since it’s relativistic, this new model could conceivably overcome the problems in MOND; I’d like to see a galaxy cluster calculation and early-Universe models for the CMB, as others did for TeVeS. (I’d also like to see a full parameterized post-Newtonain [PPN] analysis.) I’m actually sympathetic to alternative gravity models, though my interests are more in trying to solve dark energy than for dark matter. It’s worth a look, not least since this theory is a lot simpler than TeVeS.

The paper vs. the PR

However — and this is a huge however — the theory’s authors’ piece on The Conversation uses extremely problematic language, from the title on. Reproducing MOND in a relativistic context says nothing about dark matter, either way, yet the piece says that this theory alone “casts doubt” on dark matter’s existence. That to me is a troubling misunderstanding of how science works. A theory, any theory, is only worth something if it tells us something about the Universe. The higher-dimensional relativistic MOND theory may do that, or it may not. This theory reproduces MOND, but so far that’s all it can claim.

Dark matter is frustrating, but the models containing it are extremely successful, from (yes) spiral galaxies up to the structure of the Universe. I will not say (until we discover a dark matter particle) that no modified gravity scheme will ever work. Truly elegant theoretical solutions to difficult problems have appeared in the past, but they’re rare. For any alternative model to “cast doubt” on dark matter’s existence, it would need to mimic dark matter across a wide array of  problems in astrophysics and cosmology. Dark matter as a model, for all its challenges, has an advantage of explaining all those disparate phenomena, including those that were unknown when the model was first proposed.

9 responses to “No, this new theory does not “cast doubt” on dark matter”

  1. In addition to doing a good job of explaining spiral galaxy dynamics, galaxy cluster dynamics, large-scale filamentary structure and the CMB power spectrum, the dark matter hypothesis is also consistent with a lot of gravitational lensing observations. That is, dark matter not only produces (Newtonian) dynamical effects as though it has mass, it also bends light rays as though it has mass. Of course, MOND and related theories have nothing to say about this subject.

  2. Dear Matthew
    Thank you for commenting on our short article in The Conversation. As an author of this, and the discussed paper in Physical Review D, I actually agree with the majority of what you say and how you have said it.
    However, I would be grateful to able to make some comments on the above. Our article and paper introduce a new theory of general relativity (GR) that can inherently preserve the predictive successes of conventional GR, while also providing a new geometrical explanation for anomalous (spiral) galaxy rotation curves. Of course, if at this stage we could also report explanations for every other aspect within, and of, the universe – then we would be very happy authors indeed! It did take Einstein eight years to develop his GR, and there have been many subsequent decades of testing its predictions. On the future of our contribution, we did acknowledge in The Conversation article that we are simply making progress and expressed hope for further applications, saying that “Our theory, while making progress, now poses many more questions about the origin and evolution of galaxies and the universe. But hopefully …”
    We thus agree that any new GR theory needs eventually to be able to explain all these things, and that arguments for introducing dark matter involve more than just consideration of individual galaxy rotation curves. To emphasize this, we tried to point to the bigger picture, though discussion of the Bullet Cluster, in The Conversation article. But this context also demonstrates that it’s simply not that clear cut whether modified gravity or dark matter best explains what is observed. We also discussed why we believe our new GR theory may be an improvement over rival theories such as TeVeS (as quoted above).
    It seems proper to acknowledge claims of successes for modified gravity theories (over dark matter fits) concerning aspects of the universe other than individual galaxy descriptions. There does not seem to be much detail of that above. For example, there is a reasonably-sized body of work (presenting arguments both for and against modified-gravity approaches) cited here: . There are also published comparisons of successes of the two approaches (describing galaxies and other phenomena), such as: . Without getting into too much detail, a minimum should be at least to acknowledge alternative theories and that there is some degree of success in their application.
    On the public relations (PR)/discussion aspects, as you know, there is an enormous amount of public interest in this subject area. The public expect an account of what else may be possible. You seem to say that it is unscientific to suggest that a new discovery or theory can raise any doubt about an existing theory. On the other hand, how scientific would it be to propose that doubts cannot be raised about the existence of dark matter? Specifically, concerning your description “a troubling misunderstanding of how science works”, science tends to develop on an incremental basis and includes an overview of the available evidence. Where do doubts come from, other than new developments?
    On a more minor detail, the central premise of our work is actually the inclusion of an explicit space-time expansion factor within a GR framework, and hence we examined the implications thereof.
    Thanks for the opportunity to comment on the above.
    Best wishes, Graham McDonald

    PS On dwellscho’s comment, relativistic theories do have quite a lot to say about gravitational lensing.

  3. I agree with the main point of this post, that the success of one theory does not imply a failure of another. However, there are a number of items that, as Gould notes, are rather messier.

    Full disclosure: the author of this post will no doubt consider me to be one of the “proponents” of MOND. If by “proponent” it is meant that I have made positive findings about MOND as well as negative ones, then I am guilty as charged. Intellectual honesty compels me to call it both ways, as the data stipulate. If by “proponent” it is meant that I am unreasonable to speak positively of MOND at all – and this often seems to be the intended meaning – then I simply do not fit that description. Nor, in my experience, are the proponents of MOND anywhere near as vocal as its critics, but that is irrelevant to the science.

    The success of MOND surprised me greatly when I first encountered it. It predicted a priori what I observed in low surface brightness galaxies. My own (dark matter based) prediction was falsified, as were all the other dark matter models of the time. Only Milgrom’s MOND got it right in advance of the observation. So what am I suppose to say, that he’s wrong? When a man lies, he murders some part of the world. I would extend this missive to ignoring facts.

    The success of MOND in fitting rotation curves is widely acknowledged. This does not falsify dark matter. It does, however, pose one heck of a fine tuning problem. Think about it: all you need to know to predict the dynamics of a galaxy is the distribution of luminous matter plus the MOND formula. We do not need to know anything about the dynamically dominant dark matter. That’s weird. Worse, a dark matter halo plus a luminous disk could combine in a nearly endless variety of ways. In MOND, a rotation curve can only do exactly one thing. And yet, that’s what they do. That’s not just weird, it is profoundly problematic (ever hear of Occam’s razor?). Asserting that galaxies are complicated just makes it worse, and hardly suffices as a scientific explanation of a well established observational phenomena.

    When MOND first raised its head unexpectedly in my data, I was shocked. Angered, even. How could this stupid theory have its predictions come true when there is so much evidence for dark matter?

    So I made a thorough survey of the literature (ApJ, 499, 66) sure that I’d find some good reason not to believe my own data. And indeed, I found many claims to falsify MOND. Problem was, these were mostly (not all, but mostly) quite biased. Many of the studies could easily have been portrayed as successes of MOND rather than failures. They read like “MOND works here, and here, and here here here, but wait! it is a little off here! It must be wrong!” I realized that if we took an equally harsh attitude to dark matter, we would have abandoned it a very long time ago.

    So we have above repeated some of the problems MOND faces. Let’s review:

    1. Galaxy clusters require some dark matter, even in MOND.
    This is true. I agree. This point appears already in my review with Sanders in 2002 (ARA&A, 40, 263). But lets keep a little perspective. The amplitude of the failure is roughly a factor of two. Often, that’s considered pretty good in astronomy, though I think the problem is genuine. Still, a factor of 2 hardly necessitates the invention of some novel, non-baryonic particle as we require in cosmology. This is not to say this isn’t a serious problem for MOND: it is. But we’re off by a factor of a thousand in the cold dark matter prediction for the number of satellite galaxies around the Milky Way, and lots of people seem to think this is a problem that can be solved. One or the other (or both) of these things is just an example of Gould’s messiness.

    2. It is incorrect to assert that only dark matter models make large scale structures like filaments. People have done MOND simulations; very similar structures grow. This is hardly surprising, since both basically boost the strength of Newtonian gravity. My own prediction is that the cosmic web forms faster than we expect in LCDM, so structures like filaments are not only present, but already in place at higher redshift than we’d otherwise expect.

    3. The CMB is very well fit by LCDM. But the claim that it needs dark matter is based on a fit that assumes General Relativity is correct. Obviously that is not the case if something like MOND is correct. One then has to do the hard work of not only coming up with a relativistic MONDian theory, but also of checking what it predicts for the CMB. These things take time and lots of effort. And as you say, the success of one theory is not perforce a problem for another.

    One does not generalize General Relativity overnight. It took many years for Bekenstein to come up with TeVeS. TeVeS does account for lensing, contrary to what dwellscho asserts. I have no problem with people who make valid criticisms of MOND – I have done so myself – but much of this discussion seems to be occurring at a level too ill-informed to be considered scientific.

  4. see my book “how to understand the true universe” by Dr. Sol Aisenberg
    available from amazon and libraries and some bookstores on request
    it contains essays that remove dark matter and dark enrgy

  5. […] use a less frivolous example, dark matter (which we haven’t detected directly, but which has observable consequences in shaping the structure of the Universe). Telepathy fails as an explanation for physical […]

  6. [Comment removed for failing to follow the commenting guidelines.]

  7. […] And that’s where things can get tough. We know there’s a lot of dark matter in galaxies, since it affects the motion of stars, gas, and dust. Dark matter even affects the path of light through gravitational lensing. However, in the denser regions of galaxies, the precise distribution of dark matter — known as the dark matter profile — is harder to measure. That’s because all the components are more tightly packed close to the galaxy’s center, meaning astronomers have to know a lot about the behavior of the ordinary matter before they can fully extract the dark matter profile. This difficulty has led some researchers to abandon dark matter entirely, despite the overwhelming evidence for its existence in a variety of observations. […]

  8. […] While the details have emerged and changed over the decades, the Big Bang itself has survived. Observations of a variety of objects and phenomena have showed that the Universe is expanding, while also verifying that things were much hotter and denser in the past. The discovery and subsequent measurement of the cosmic microwave background (CMB) helped establish the Big Bang model’s superiority over competing ideas, since it showed the Universe was hot enough in the past to ionize every atom and dense enough that the cosmos was opaque. Similarly, the theory known as Big Bang nucleosynthesis (BBN) predicted the measured abundances of the lightest elements fused in the hot dense environment that constituted the entire Universe during its first few minutes. (These elements are completely separate from dark matter and its “controversies”.) […]

  9. […] physics and astronomy. For example, see my critique of the “warp drive”, my critique of alternatives to dark matter, my critiques of people making every story about Einstein, etc. Even my most recent article in […]

%d bloggers like this: