Teach the Controversy About…Dark Matter?

(If you’re not a cosmologist or astrophysicist, you may need a bit of background before I get to the main point. If you already know all this stuff, good on yer–skip down a few paragraphs.)

One of the big mysteries remaining in our current understanding of the universe is “dark matter”—apparently a type of matter that doesn’t interact with light (“invisible” might be a better word than “dark”), doesn’t seem to heat up much, but which also seems to comprise about 83% of all the matter of the universe. By its nature, dark matter is very elusive and hasn’t been detected directly yet. The evidence for its existence is really strong—although it was first postulated to explain the strange rotation of spiral galaxies, the best evidence for it is in galaxy clusters and in the structure of the early universe.

Galaxy clusters are a little easier to explain in brief (I may come back to the cosmological evidence later, but in the meantime see the WMAP website); they are large objects containing many galaxies and lots of hot X-ray emitting gas. The picture shows a particular galaxy cluster known as the “Bullet Cluster”, which is the result of two galaxy clusters colliding. The colors aren’t “real”: they’re a way to combine information from several different types of observation, and that’s key to understanding the image. The red is the X-ray observation of the hot gas, showing a clear shock way (the “bullet”) where the gas in the two clusters collided and heated up. The blue is where most of the mass of the clusters resides—not in the gas, but in the region surrounding the galaxies. Since we can calculate the mass of the stars and gas in the galaxies independently, we can see that there’s a lot of mass that doesn’t show up as gas or stars or anything “normal”. In other words, the blue in the picture is the closest we’ve gotten to “seeing” dark matter, even though all we’re actually seeing is the effect of its mass.

This isn’t to say there aren’t real problems with our current dark matter models. (Note my careful phrasing!) One of the perpetual challenges in galactic astronomy has been to relate the light from a galaxy (called its luminosity), which is mostly from stars, to the mass of the galaxy, which is mostly not stars. I’m not an astrophysicist so I’m a little hazy on the details, but this has been an area of intense interest, and it’s not a fully resolved problem. In some types of galaxies at least, the estimate of mass from the luminosity contains fairly large errors. (I welcome correction from my galactic-astronomer friends and readers.)

Enter this article, which is basically a press release for a new paper by astrophysicist Stacy McGaugh. There is a small but vocal group of astrophysicists who think the best way to deal with problems with dark matter is to do away with the need for it entirely, by modifying the equations of gravity. Called modified Newtonian dynamics, or MOND, the changed equations do a pretty good job of fitting the rotational dynamics of spiral galaxies. In the paper cited, McGaugh has successfully also solved the problem of determining the mass-light relationship in a certain type of galaxy using MOND. All well and good, and potentially very interesting.

But here’s where the problem lies: MOND isn’t an actual theory. While Newtonian gravity has been shown to be the limiting behavior of Einstein’s general relativity, the generalizations of MOND to be compatible with relativity are very difficult to understand and quite complicated to work with. (Despite general relativity’s mathematical complexity, it’s conceptually elegant.) If MOND is correct in the same way Newton’s law of gravity is correct, it needs to be shown to be a limiting case of a more general theory, one that satisfies the same experimental tests general relativity has already passed. Ultimately, MOND is a heuristic model that fits some types of galaxies, but it fails for galaxy clusters and for the universe as a whole.

So, for those of you waiting patiently for my main point for non-scientists: real scientific controversy is often of this sort, where just explaining what’s going on to non-specialists can be challenging. I wouldn’t teach MOND in an introductory astronomy or astrophysics course for that reason, no matter how empirically successful it is in the cases where it works well. If McGaugh and his colleagues manage to solve the obvious problems with MOND such that it seems like a viable alternative, then I might consider it, but having given it a fair study I can’t in good conscience present it to my students as something to take seriously. Our students need more background before they can “make a decision” between Newtonian gravity with dark matter, and MOND without dark matter. That’s part of what education must do.

28 responses to “Teach the Controversy About…Dark Matter?”

  1. […] Science Vs. Pseudoscience A Blog of Science and What Lies Beneath « Teach the Controversy About…Dark Matter? […]

  2. […] Carroll takes on the same press release about MOND and dark matter that I did, but he’s a lot less um… polite in his rejection of MOND. (Our basic conclusions are […]

  3. AThinkingScientist Avatar

    In this article MOND is referred to as not being a theory. That is wrong. It can be described as a theory, check e.g. http://en.wikipedia.org/wiki/Theory :
    “Theories are analytical tools for understanding, explaining, and making predictions about a given subject matter.”

    MOND is definitely the correct theory for galaxies. So teaching MOND and those aspects of it which are not understood is valuable and necessary, since there may be that one student who would be inspired to develop it further or even revise gravitational theoy altogether.

    1. (Warning: this comment is gonna be technical! Read at your peril.) Perhaps a better way to phrase it is that MOND isn’t a complete theory, since the scale of the acceleration parameter isn’t predicted by the model and can’t be measured independently. In that sense, MOND is less predictive than it is descriptive. Compare to Newton’s law of gravitation: we can measure both the inverse-square property and the strength of the gravitational attraction constant G. MOND has a parameter not derived from anything deeper that can’t be measured directly, but only fit to the data. That by itself doesn’t rule it out, of course: a lot of great theories began as heuristic models like this. However, if MOND works for some types of galaxies but fails for galaxy clusters and the universe as a whole, it can hardly be said to be comprehensive. I definitely think it’s premature to say that MOND is the “correct theory” for galaxies for that reason.

      A student needs to be fairly advanced and sophisticated before he or she is able to tackle the big questions. I don’t advocate hiding the existence of MOND from anyone (or suppressing any controversial idea, correct or not), but that’s not the same as saying it should be taught. The burden of proof is on the researchers working on MOND, and without resolving the issues with the Bullet Cluster and the CMB power spectrum, they haven’t proven their case sufficiently well to convince me–or the wider community–otherwise.

      1. AThinkingScientist Avatar

        No, this is false.

        Newtonian gravity is also empirical, as it does not follow a priori from findamental arguemtns. G is also merely a fitting parameter. In MOND, a_0 can in principle be measured, if we could devise a laboratory in a sufficiently weak-field.

        Notwithstanding this, everyone in the “MOND community” knows that MOND is hinting at something deeper. Just as Newtonian gravitation is superceded by Einstein’s General Relativity (GR). And remember, energy conservation is a critical issue in GR, again hinting at the real possibility that GR is itself not complete.

        Concerning MOND’s supposed failures:

        Clusters work fine in MOND:

        E.g. the Bulet Cluster turns out to be a real problem for LCDM:
        “Bullet Cluster: Chanllenge to LCDM Cosmology” (2010, ApJ):

        This is an excellent example of how hype overstates the meaning of scientific data with the added aftertaste that relevant research results that do not fit a pre-defined concept of reality are simply ignored.

        The Bullet cluster works fine with MOND, in fact better than in LCDM (see beow).

        And yes, in MOND a hot dark matter component may be needed, but this turns out to be nicely consistent with what other clusters also require, and what in fact the CMB also requires, as has been published! In fact, it emerges that a beautifully consistent theory is appearing based on MOND.

        MOND and Bullet Cluster (2006, MNRAS): http://adsabs.harvard.edu/abs/2006MNRAS.371..138A
        Note that the hot dark matter emerges naturally in some ideas on how neutrino masses are generated.

        However, perhaps hot dark matter may not be needed, since most of the normal matter, which is supposed to be around given primordial nucleosynthesis, has mysteriously gone missing in our universe (this is called the unsolved “Missing Baryon Problem”). So Milgrom suggests in research papers that perhaps the dark matter needed in MOND is just this missing baryonic matter.

        Matthew also states that MOND fails on the universe as a whole.

        This is again wrong. What has been done so far, with the very limited resources available to those investigating MOND, suggests that the CMB is just perfectly fine (with hot dark matter, yes – see above) and that structure formation may actually proceed as required by Peeles & Nusser who quite nicely show that LCDM doesn’t really work to make the Local Volume of galaxies (they find one needs to introduce extra forces or some modified gravity): “Nearby galaxies as pointers to a better theory of cosmic evolution” (2010, Nature):

        So it is very important to teach students about MOND (note, that I mean advanced students, not necessarily early undergraduates).

        It is the _young_ people who will solve the problems, not the now established researchers. It has often been like this. Young Einstein himself was not part of the established system when he revolutionised it.

        But there are other ideas about gravity as well.

        E.g. Moffat’s Modified Gravity (MOG, another alternative) and galaxy clusters and specifically the Bullet Cluster: Yes, works without any dark matter:

      2. I agree that it’s not right to assume that GR is the end point of classical gravitation, and that younger researchers are likely to be the ones to resolve the issues. However, it’s one thing to say that the Newton-GR will be superseded at some point in the future and it’s quite another thing to say that MOND-TeVeS is the successor. To fit things like the Bullet Cluster and the CMB power spectrum, MOND needs hot dark matter (as you point out), and requires the mass of the neutrino to be on the edge of what is acceptable experimentally. So, in other words, MOND still needs dark matter—the very problem it was trying to solve! (I haven’t had time to read the first paper you cited (and the link is broken, by the way—here’s one that works), so I can’t assess it yet. I’ll consult with my friends who have actually studied the Bullet Cluster system in some detail.)

        Again, let me say that very few people think the situation is fully resolved. After all, dark matter profiles in galaxies are tricky things that don’t look very convincing to many of us, and of course we don’t know what dark matter actually is. I don’t a priori rule out alternative theories of gravity especially as possible resolution to the dark energy problem, but if MOND requires perpetual fine-tuning to fit the data, it doesn’t look promising as a legitimate alternative.

      3. And of course we’ve completely lost any non-specialists who happen to be reading this blog (which is most of the readers!).

    2. Sean Carroll’s blog has an even more scathing takedown of MOND than I do, and he also argues for MOND’s lack of theory status:

      1. Sean Carroll’s “contribution” was crap. He says MOND doesn’t qualify as a theory because it failed to predict/describe something. He conveniently forgets that DM has this same problem in different areas. He’s completely biased.

      2. Everyone acknowledges the problems with the dark matter model with respect to spiral galaxies, which is where MOND is most successful. Sean Carroll is a cosmologist like me, and dark matter is the best explanation for the features in the cosmic microwave background as well as galaxy clusters. MOND is not very successful in either of those regimes, and in fact requires some kind of dark matter to make up the difference.

        It’s very easy to hide behind anonymity and pick on Sean Carroll (not least since I suspect he has never read this blog), but I doubt he is “forgetting” anything. Anonymously saying his work is “crap” when he isn’t able to say anything in response is cowardly, and this isn’t really about his opinion anyway. There may be room to disagree about whether MOND is a theory or not — I call it an “incomplete theory” personally — but ad hominem attacks are not a respectable way to converse about this topic.

  4. AThinkingScientist Avatar

    “And of course we’ve completely lost any non-specialists who happen to be reading this blog (which is most of the readers!).”

    Then please do not make such wrong (but simple) statements as “MOND needs perpetual fine tuning”.

    There is one parameter which can be measured with one galaxy (and a transition function hinting at a deeper theory).

    With this, MOND is an excellent description of the astronomical data, perhaps needing some hot dark matter for galaxy clusters.

    1. But it’s not just a0: as you yourself point out, MOND requires hot dark matter, whose properties must be very specific for MOND to work. The cold dark matter model is far from perfect on the level of galaxies, but the theory of gravity that underlies it doesn’t depend on particle properties beyond its purview.

      1. AThinkingScientist Avatar

        Unfortunately the theory of gravity underlying the cold dark matter model is a theory which:

        1) requires mathematical trick number one: inflation which is not understood at a fundamental level.

        2) requires trick number 2: cold dark matter which is a purely hypothetical addition with all searches having brought null results such that the originally favored cold dark matter particles are already excluded to exist since many years.

        3) requires trick number 3: dark energy which does not fit any other physics (e.g. of the vacuum) and violates energy conservation since the amount of it increases without limit with the accelerating universe.

        All of these unknown additions lead to a model that only describes data that are difficult to observed, and fails catastrophically on scales of 10Mpc and less where we have excellent data. And, most of the normal matter has disappeared since the Big Bang, in this model.

        Thus, we do not have a good theory of the universe.

        MOND is pointing into a new direction, and we would be silly not to take it seriously and not to follow its lead.

      2. Those issues are indeed major and well-known problems—though not necessarily for our current theory of gravitation. I think most cosmologists would disagree that inflation is a major problem for general relativity in any case, and the missing baryonic matter is decidedly not a problem for the gravitational paradigm. You see all these things as problems for the current gravitational paradigm, while most physicists don’t. You may be right for dark energy and (possibly) dark matter! It’s not as though we’re any closer to resolving the problems of dark matter and dark energy. After all, new observations almost always lead to new physics down the road.

        However, MOND is still not a better solution than GR with cold dark matter and dark energy. MOND doesn’t even address dark energy or inflation, and it doesn’t solve the problem of dark matter—MOND still needs dark matter to work for the CMB and galaxy clusters, which weakens its value. TeVeS is a nightmare of a theory which of course doesn’t make it wrong necessarily, just…difficult to see how it’s an improvement over anything else.

        I don’t see the situation in as bleak a light as you do, but either way: we won’t settle it through debate. I obviously won’t convince you and you won’t convince me through argument alone. Apart from personal opinion, it’s not how science works anyway. So, you work on MOND and be welcome. I think the resolution lies elsewhere (and the larger community agrees with me), but ultimately it’s going to be settled by the intersection of specific predictions with the evidence.

      3. AThinkingScientist Avatar


        you are certainly welcome to keep working in the standard framework, and you are right that the issue is not settled.

        But to support an argument by pointing out that the majority of scientists agree with some way of thinking is never a good point to make, because majority thinking has, time and again, been shown to be wrong.

        Examples are: not long ago most believed in the existence of phlogiston http://en.wikipedia.org/wiki/Phlogiston_theory
        and the aether http://en.wikipedia.org/wiki/Luminiferous_aether and the geocentric model http://en.wikipedia.org/wiki/Geocentric_model .

        But we would both agree that deciding which new idea has real potential is very difficult and often nearly impossible, and so science history is littered with stories that arose because bright researchers tried out ideas that were not successful. Doing fundamental research is very risky, but we need to allow for this.

      4. I don’t think you actually read what I wrote. Please do so before commenting—you seem to have completely misunderstood my point of view, or you are being deliberately obtuse. (You’re bringing up phlogiston? really?)

      5. AThinkingScientist Avatar

        I do not quite understand what your last comment is supposed to mean. Are you offended because some time ago phlogiston was being taken seriously by many if not the majority of scientists? This story just shows how science evolves: often new substances are introduced as a first step to solve a problem. Cold dark matter may well be of the same category. My point, which I hope you understand, is that the majority-view argument is inappropriate.

      6. Oh, I’m not offended. If you read what I write, you will find that I’m generally in favor of examining alternative models. Teaching alternatives is generally not wise, especially not something like MOND whose relationship to the available evidence isn’t that good.

        The community isn’t always the best judge of a theory’s correctness, but here’s the deal: if a theory is accepted, it’s usually for good reasons.The community may have a prejudice against MOND, but that doesn’t automatically mean they’re wrong. New theories are proposed daily, and most of them don’t survive in the face of the evidence. The community was wrong about aether, but relativity didn’t replace it just because it was an alternative: relativity solved a problem with the aether model. MOND doesn’t solve enough problems with the current paradigm, and pointing that out doesn’t make me or my colleagues slavish adherents to the status quo.

  5. […] of a physical theory? I was prompted to think about this question by several things: the ongoing conversation about MOND, the proposed alternative to standard gravitational theory, and another conversation on Facebook […]

  6. […] of understanding how general relativity might be modified to solve the dark energy problem. Despite my dislike of MOND, I’m sympathetic to attempts to modify gravity. (I’ve mostly avoided advocating for one […]

  7. […] In a previous post, I talked a little bit about galaxy clusters in the context of that glorious object, the Bullet Cluster. Galaxy clusters are the biggest objects in the universe that are held together by their own gravitation. And they can really be big: the Virgo Cluster has over 2000 galaxies; the Coma Cluster (shown above) has over 1000 galaxies. As a comparison, our Milky Way galaxy is in a group of galaxies called the Local Group (someone was feeling really creative that day), which has over 40 galaxies, only three of which are big enough to be counted as full-size galaxies. Most galaxies seem to be in groups of clusters. […]

  8. […] wrote about the Bullet Cluster in another post, so I’ll send you there for some specifics. The short version: the image you see here is a composite, where the red is the […]

  9. […] talked about the Bullet Cluster in two previous posts, and probably will mention it again — it’s an important object for […]

  10. […] are built of stars, gas, and dark matter, galaxy clusters are comprised of galaxies, gas, and dark matter. The center of the Virgo cluster. Note the giant elliptical galaxy M87 at the […]

  11. […] of spiral galaxies. On the other hand, we know that most of a galaxy’s mass is in the form of dark matter, and based on our understanding of how galaxies form, some astronomers have wondered: could there […]

  12. […] Dark matter has its problems, since we don’t know what it is, but alternative theories of gravity run into their own difficulties when it comes to galaxy clusters. New results may actually rule out some alternative models, though many of the usual caveats apply. […]

  13. […] Strange as dark matter is, it still has the characteristics of matter: it collects together and behaves normally under gravitation. Dark energy doesn’t behave like anything normal: if you compress it, its pressure goes down, while if you let it expand, the pressure goes up. For this reason, many people associate it with vacuum energy: the energy of nothing, since the “more nothing” you have, the more energy you would have in that volume. It’s a good idea, and supported by quantum physics, except that the vacuum energy calculated is far too large compared with dark energy. […]

  14. […] but absent in galaxies, it makes little sense! (Similarly, alternative theories of gravity like modified Newtonian dynamics (MOND) seem still to require at least some dark matter to explain objects like the Bullet Cluster, […]

%d bloggers like this: