Can one say anything, even the most obviously wrong things, to discredit an alternative to the standard model? An incident: cosmology at CalTech

The answer to the question posed in the title is  “Apparently, and sadly, yes.”

In previous contributions we have blogged about sociological problems that arise when attempting to do research in non-standard cosmological frameworks (for example the attempt at closing down “The Dark Matter Crisis”).

Early 2015 an incident occurred which is a contemporary example of this, but which may also possibly be a serious case of scientific misconduct. It appears to be an aggressive act in an attempt to discredit new approaches to cosmology and those working on them. A senior professor at CalTech has expressed, in a public forum, “Take the world’s best courses, online, for free“, directed at students of cosmology, wrong and unacceptable views which are likely to discourage young researchers from studying important theoretical concepts. The statements are derogatory, dismissing and demeaning to those full-time researchers who have been performing research in such fields, and who are without exception very talented physicists.

Prof. George Djorgovski teaches Cosmology at CalTech and his course can be followed by students world wide. In order to dismiss alternatives to the standard cosmological model, he recently used in a public forum (see below) the argument that General Relativity is “conformal”, and that this is “well tested”, while MOND is not. He further writes that “Cosmology tends to attract a certain type of crackpots, and some of them even have PhD’s.” “Some were great scientists, before sinking into the downward spiral” thereby implying, it seems from the context, researchers who work on MOND. He makes other, wrong statements, about MOND.

While there are indeed valid and rational arguments to make about the problems of MOND on large scales and on sub-galactic scales (such as globular clusters), one could seriously wonder whether a respectable institution like CalTech should find it acceptable for someone affiliated with it to make such erroneous statements about physics in a public forum dedicated to an official online lecture.

We remind the reader that a conformal transformation is a transformation preserving the angles but changing the magnitude of the length vectors. While many equations in physics are invariant under conformal transformations, Einstein’s equations are not. If they were, their weak-field limit giving rise to Newtonian dynamics would also be conformally invariant in space-time. Since one of the conformal transformations is the one known as “scaling” (others being related to rotations in space-time), conformal invariance would imply space-time scale invariance. But obviously, Newtonian dynamics is not space-time scale-invariant.

Indeed, assume we seek a trajectory (x,y,z,t): which equations of motion are required such that the trajectory n(x,y,z,t), where n is a number, is also acceptable?

This space-time-scale invariance (Milgrom 2009) actually leads quickly to equations of motion different from Newtonian dynamics, and, remarkably, these strictly imply the baryonic Tully-Fisher relation, flat rotation curves of galaxies as well as the external field effect. The above space-time scale invariance has been noted by Milgrom (2009) to be a new symmetry which may have deep theoretical implications. Milgromian dynamics, or MOND, is a classical framework which contains the Newtonian regime and extends it to the very weak-field regime which is identical to the space-time-scale invariant regime. The interested reader may find additional information in the important review by Famaey & McGaugh (2012) and in Kroupa (2015) as well as in Trippe (2014).

To summarize, it is certainly true that space-time scale-invariance does not imply full space-time conformal invariance. But this is a) of course not a problem, and b) GR is not conformally invariant either.

If on the other hand, Prof. Djorgovski meant that MOND is conformal, while GR is not, this is not true either. And there is certainly nothing “well-tested” about this. So one may be led to conclude that Prof. Djorgovski has either misunderstood some important issues or has a non-scientific agenda when interacting with students, and one could wonder whether a respectable institution such as CalTech ought to accept this, whatever one’s stance on the validity of alternative approaches such as MOND. A rigorously working scientist can only accept objective and evidence-based arguments when testing hypotheses.

Personal opinion ought not to play a role when testing the possible laws of nature. For nature it is irrelevant what opinion someone may have, or how prestigious the institute is where the scientist is opinionating from. A scientist may decide which field to work in, and which tests to perform, but dismissing hypotheses without a rigorous and solid analysis is unscientific behavior.

But perhaps Prof. Djorgovski used the wrong word (“conformal”) but meant “covariant”. GR is covariant, but the original paper by Bekenstein & Milgrom 1984 also explicitly proposed a covariant MOND theory, so this is obviously incorrect too.

As explained by a high-profile colleague interested in modified gravity theories (who however does not want to be named here, given the quality of Prof. Djorgovski’s statements) below, it is possible that Prof. Djorgovski has been confused by the fact that this first covariant version of MOND proposed by Bekenstein & Milgrom in 1984 involved a conformal transformation between the Einstein metric and the physical metric. This could not reproduce the observed enhancement of lensing, and led Bekenstein to propose a non-conformal relation between the Einstein and physical metric in 2004 (which is not a problem). So, Prof. Djorgovski is likely to have become confused here, leading to his nonsensical sentence. As stated by our colleague below, this again appears to suggest that Prof. Djorgovski may not understand what he is talking about. It would be a rather serious issue for modern cosmology to have ignorant people teaching it to youngsters.

Coming back to Milgromian dynamics, it has proven to be an incredibly rich theoretical approach to understanding the dynamics of galaxies with convincing success. The success in accounting for observations and more importantly in predictions is convincing evidence that Milgromian dynamics needs to be taken very seriously by theoreticians. It is false to claim, as Prof. Djorgovski does, that “epicycles” kept being added to MOND in order “to salvage it”. The classical framework of MOND, written down in Princeton by Prof. Milgrom in 1983, contains one single free parameter a_0 (possibly a new constant of nature, call it Milgrom’s constant, probably related to the properties of the vacuum; is has the value a_0=3.8 pc/Myr^2 approximately, e.g. Kroupa 2015), which is an acceleration, and this parameter can be fixed by one single galactic rotation curve leaving no freedom for further adjustments in other systems. Exploration of how to embed this classical framework into general-relativistic theories do not constitute “adding epicycles” but are important and necessary theoretical and mathematical research at the highest level of intellectual activity (see the comment below by our high-profile colleague and e.g. Zhao & Li 2010).

Indeed, dismissing the possibility that Einstein’s theory of general relativity (GR) may not be correct in the extreme weak-field regime, constitutes an unphysical ideological constitution of the mind in question. It is well known that Einstein’s GR is not unique. It should also be well known that Einstein 1916 put much effort in constraining his geometical interpretation of gravitation to agree with the Newtonion law of universal gravitation in the appropriate limit. But Newton derived this empirical law based on Solar System data only. Even Einstein did not know what galaxies are. Any person claiming that Einstein’s GR is valid on all scales is effectively performing an extrapolation by many order of magnitudes beyond the empirical data which the law was derived from. It is high-school knowledge that such extrapolations are extremely dangerous and are not likley to work. The apparent failure of GR on galactic scales and beyond may thus be the mere break-down of an extrapolation. It may also harald the existence of dark matter particles, which is a resaonable hypothesis a physicist may probe (as done here at great length). But it is not the only hypothesis.

While incredibly successful on galaxy scales, the hardest test of Milgromian dynamics designed until now has come from my (Pavel Kroupa) group in Bonn using globular clusters (Baumgardt, Grebel & Kroupa 2005; see Kroupa 2012 for a discussion). The evidence until now is ambiguous, but Milgromian dynamics appears to be under some stress on these globular star cluster scales.

Another interesting test being followed up now by observational astronomers in Chile has been proposed by Michael Bilek using shell galaxies (Bilek et al. 2015).

World-wide, the interest in Milgromian dynamics is increasing significantly, partly due to its most amazing success in accounting for the properties of galaxies. The increasing interest is shown in the figure below though the rising number of citations of Milgrom’s paper per year.

 

Milgrom1983citations
This chart shows the development of citations to the original research paper by Milgrom (1983). The increase in citations after the year 2004 comes with the break-through by Bekenstein (2004). Source: ADS.
Two independent groups have now created, for the first time ever, Milgromian simulation codes to allow full cosmological computations of galaxy formation and evolution using baryonic physics with feedback and star formation: the publicly available Phantom of Ramses (PoR) code by Lueghausen, Famaey & Kroupa (2015) and the RAYMOND code by Candlish, Smith & Fellhauer (2015). Numerical experiments on galaxy formation and evolution are being started in Concepcion (Chile), Strasbourg (France), Bonn (Germany), St Andrews (Scotland) and other places.

Surely this increasing activity world-wide is not due to “a certain type of crackpots, and some of them even have PhD’s” (me included with a BSc (hon) from UWA, Perth, a PhD from Cambridge University and habilitation from the University of Kiel as well as receiving a Heisenberg Fellowship, amongst other prizes). It is not so very clear where the crackpots actually are. Prof. George Djorgovski teaches Cosmology at CalTech and his course can be followed by students world wide. Questions may be asked in a forum. Early 2015 a very talented MSc student studying Astronomy and Astrophysics at Charles University, Prague, asked Prof. Djorgovski why he discounts MOND  (here is the MOND_Djorkovski-1 of the discussion, and here is a mond_djorgovski_forum-1screen shot:

 

 

 

 

 

 

Question by a MSc student to Prof. Djorgovski:

In module 7.2 there is short note about the alternavitve explanation of Dark Matter – the MOND. It was the first time I’ve seen such a possibility, so I did some research about it.

1. There is note in the table, that the gravity is modified on large scales, in papers I’ve found about MOND there is wrriten that the non-Newtonian regime should apply not on large space scales but in very weak gravity regime (such as the General relavity in strong gravity regime). Am I correct?

2. Also in the lecture was mentioned that the MOND does not work properly. I tried to find any references, but I did not. Could someone please explain me where is the problem with MOND?

 

The answer by Prof. George Djorgovski:

(In the forum “Take the world’s best courses, online, for free“)

The original formulation of MOND was a purely ad hoc modification of the Newtonian gravity, designed to explain the flat rotation curves, and without any other physical motivation. This made it also predict that galaxy clusters should not exist. More to the point, it was not a conformal theory, and thus in a conflict with the well established (and tested) aspects of the GR. Theoretical proponents of the theory (there are one or two of them) kept adding “epicycles” to it, to salvage it, thus sacrificing any putative elegance to this purported solution, and again, purely in order to save it, and without any other physical motivation.
A very small number (<< 10) of observers keep finding “evidence” that supports MOND, while ignoring any of the problems. Then some other observers point out that this is not the case, and the cycle continues. Most people see it as an exercise in futility.
Why do people persist in such pursuits? I think that this is a matter of psychology, not astrophysics. Cosmology tends to attract a certain type of crackpots, and some of them even have PhD’s. Some were great scientists, before sinking into the downward spiral; the most famous (and most tragic) example was Fred Hoyle, who simply cannot bear the idea that he was wrong about the Steady State cosmology, and he turned what was a brilliant career into becoming an irrelevant crank. Another, lesser, example was Geoff Burbidge, who refused to accept that the quasar redshifts were cosmological, despite an overwhelming and growing evidence, saying how there may be some new physics behind them, but never producing any. There are many more examples, and the proponents of MOND are not nearly as smart as Hoyle or Burbidge were. Once your ego becomes bigger than your ability to be a critical thinker and an honest scientist, so that you cannot admit that you were wrong and move on, it is over.
I should also note that a great majority of theoretical models turn out to be wrong, and simply disappear without a trace – they turn out to be in conflict with some measurements, fail to make good predictions, and that’s that. That is how science works. Sometimes a brilliant, new, original idea does work, or even transforms the physics – e.g., the relativity – by explaining the known facts and by making testable predictions (and surviving those tests). Most do not.
So if you really want to waste your time, go ahead and sift through those 600 papers on arXiv, and make up your own mind, but I think that you could spend your time more productively on other things.

 

Note by P. Kroupa:

Remarkable are the comments by some of the other students, if this is what they are, as evident in the forum. Noteworthy is Stephen Schiff’s addenda: “unscrupolous people”, “quacks”, “own egos or self-delusion” etc. with Prof. Djorgovski replying “Exactly”.

 

A commentary by a high profile colleague who is also an expert on modified gravity:
(given the contents of the text above by Prof. Djorgovski this colleague asked to remain anonymous)

This forum post by Mr. Djorgovski is absolute nonsense. To say that “the original formulation of MOND” was “not a conformal theory (sic)” casts serious doubts that he actually understands what he is talking about. I don’t think anyone could even understand what it is for a theory to be “conformal”… Is GR “conformal”? What does he mean? Does he mean it is conformally invariant? Of course, it is not. So what does he mean, then? Probably one should ask him, but the sad and clear truth is that this statement of Mr. Djorgovski simply does not make any sense whatsoever. But it may surely award him a rather high crackpot index. This is rather ironic, given the rest of his comments, which would probably be best applied to himself.

Actually, the problem of the original scalar-tensor theory proposed to reproduce the MOND phenomenology back in 1984 (which is actually what one would now call a “k-essence” scalar-tensor theory) is that it invoked a physical metric (coupled to matter fields in the matter action) which was conformally related to the Einstein metric, and for that reason, while enhancing the dynamical effect (g_00 term of the metric) could not enhance gravitational lensing (through the other space-space diagonal terms) by similar amounts. This is why a disformal transformation, invoking a vector field in addition to the scalar field, was proposed by Jacob Bekenstein 20 years later. This is perhaps what confused Djogorvski. But of course this is not “in a conflict with the well established (and tested) aspects of the GR” (sic). The latter statement relating to a mysterious “conformal” nature of GR, I have still a hard time believing has been written by someone with a PhD in Physics, and not by some random crackpot.

But this so-called TeVeS theory of Bekenstein does have real phenomenological problems, like the fact that without additional non-baryonic matter it has a hard time reproducing the CMB. Much better models in this respect are those recently proposed by Justin Khoury 2014  or Blanchet & Le Tiec 2008 and Bernard & Blanchet 2014.

Regarding his other comments, MOND is obviously not an “ad hoc” modification of gravity, but simply a phenomenological law relating the distribution of baryons to the gravitational field in galaxies. The original Milgrom’s formula is of course not a theory “per se” but a phenomenological law which allows to make predictions on the scale of galaxies. These a priori predictions do work extremely well on these scales, and do of course concern data that were not available back in 1983, which is why it is ridiculous to call it “ad hoc”. Especially so since MOND can be derived from space-time scale-invariance.

Now, the MOND interpretation of these observations is, very generally speaking, just that this fine-tuned relation between baryons and the gravitational field is not a consequence of “gastrophysical” feedback mechanisms (as is usually assumed in the standard dark matter context based on Einsteinian/Newtonian dynamics) but rather a reflexion of something more profound in the Lagrangian of nature, which one usually refers to in the standard context as “dark matter”, and which one also usually conflates with “non-baryonic, mostly collisionless, particles”, which is by no means requested by galaxy-scale data.

It is very true that it is not easy to write a modified action which reproduces this phenomenology, appears natural, and also keeps the most successful aspects of the current standard model such as successes in reproducing the acoustic peaks of the CMB. There are however a few proposed actions which do achieve this such as those proposed by Justin Khoury and Luc Blanchet (see references above), but they still appear a bit unnatural. These should of course just be considered as examples of what kind of Lagrangian can be written to both reproduce the phenomenology of MOND and reproduce the undeniable successes of LCDM on large scales.

Also, to say that MOND predicted galaxy clusters not to exist is of course blatantly wrong. MOND actually leads to galaxy clusters forming more rapidly than in the standard model of cosmology, as has been published years ago. It actually predicted that there should indeed be missing mass there, e.g. in the form of missing baryons such as cold molecular gas clouds, or in the form of hot dark matter with a free-streaming length above galaxies, or that the new degree of freedom in the Lagrangian of nature (see references above in the work of, e.g., Khoury and Blanchet) creating an effective modification of gravity on galaxy scales which is behaving like a collisionless preassureless fluid on these scales, just as it should do to reproduce the angular power spectrum of the CMB.

All of this does of course not mean that “MOND” is right, or in any way a final theory (which it cannot be because it could only come out of a larger theoretical framework), but it is a proof that the criticisms raised by Djorgovski just display ignorance. His comments are, at best, nonsensical.

Until the many challenges to LCDM (see Kroupa 2012; Kroupa 2015) are addressed within the standard model, if they ever can be, it is only a fair scientific endeavor to also consider modifications of the action which could address these issues. That does not prevent people from working on the solutions in the standard context, nor to criticize these alternatives. But when doing so, only rational arguments are admissible. The expressions used by Djorgovski in a public forum are instead completely nonsensical from a physics point of view, demeaning and offensive from a behavioral point of view, and generally unacceptable.

 

Concluding remarks:

The above episode demonstrates that the cosmological research field is broken. It apparently allows its members to teach students the most blatantly wrong contents as long as they are considered to be defending the “mainstream”. It appears that knowledge of basic physical concepts may not seem to be a requirement to teach cosmology at CalTech anymore. This is both pathetic and terrifying.

This example exemplifies the serious sociological forces acting against the few bright and inquisitive minds who, in the true spirit of science, dare to venture outside the dull beaten track followed by most.

 

See the overview of topics in The Dark Matter Crisis.