[arXiv:0705.2462] $Lambda$CDM cosmology: how much suppression of credible evidence, and does the model really lead its competitors, using all evidence?

Authors:  Richard Lieu
Abstract:  Astronomy can never be a hard core physics discipline, because the Universe offers no control experiment, i.e. with no independent checks it is bound to be highly ambiguous and degenerate. Thus e.g. while superluminal motion can be explained by Special Relativity. data on the former can never on their own be used to establish the latter. This is why traditionally astrophysicists have been content with (and proud of) their ability to use known physical laws and processes established in the laboratory to explain celestial phenomena. Cosmology is not even astrophysics: all the principal assumptions in this field are unverified (or unverifiable) in the laboratory, and researchers are quite comfortable with inventing unknowns to explain the unknown. How then could, after fifty years of failed attempt in finding dark matter, the fields of dark matter {\it and now} dark energy have become such lofty priorities in astronomy funding, to the detriment of all other branches of astronomy? I demonstrate in this article that while some of is based upon truth, at least just as much of $\Lambda$CDM cosmology has been propped by a paralyzing amount of propaganda which suppress counter evidence and subdue competing models. The recent WMAP3 paper of Spergel et al (2007) will be used as case in point on selective citation. I also show that when all evidence are taken into account, two of the competing models that abolish dark energy and/or dark matter do not trail behind $\Lambda$CDM by much. Given all of the above, I believe astronomy is no longer heading towards a healthy future, unless funding agencies re-think their master plans by backing away from such high a emphasis on groping in the dark.
[PDF]  [PS]  [BibTex]  [Bookmark]

Discussion related to specific recent arXiv papers
Post Reply
Geraint Lewis
Posts: 5
Joined: October 10 2006
Affiliation: University of Sydney
Contact:

[arXiv:0705.2462] $\Lambda$CDM cosmology: how much suppress

Post by Geraint Lewis » May 18 2007

Cosmology is taking a bit of a bashing - does it deserve it?
Richard Lieu
(Submitted on 17 May 2007)

Abstract: Astronomy can never be a hard core physics discipline, because the Universe offers no control experiment, i.e. with no independent checks it is bound to be highly ambiguous and degenerate. Thus e.g. while superluminal motion can be explained by Special Relativity. data on the former can never on their own be used to establish the latter. This is why traditionally astrophysicists have been content with (and proud of) their ability to use known physical laws and processes established in the laboratory to explain celestial phenomena. Cosmology is not even astrophysics: all the principal assumptions in this field are unverified (or unverifiable) in the laboratory, and researchers are quite comfortable with inventing unknowns to explain the unknown. How then could, after fifty years of failed attempt in finding dark matter, the fields of dark matter {\it and now} dark energy have become such lofty priorities in astronomy funding, to the detriment of all other branches of astronomy? I demonstrate in this article that while some of is based upon truth, at least just as much of [tex]\Lambda[/tex]CDM cosmology has been propped by a paralyzing amount of propaganda which suppress counter evidence and subdue competing models. The recent WMAP3 paper of Spergel et al (2007) will be used as case in point on selective citation. I also show that when all evidence are taken into account, two of the competing models that abolish dark energy and/or dark matter do not trail behind [tex]\Lambda[/tex]CDM by much. Given all of the above, I believe astronomy is no longer heading towards a healthy future, unless funding agencies re-think their master plans by backing away from such high a emphasis on groping in the dark.

Tommy Anderberg
Posts: 47
Joined: November 24 2005
Affiliation: independent

Re: [arXiv:0705.2462] $\\\\Lambda$CDM cosmology: how much su

Post by Tommy Anderberg » May 18 2007

Geraint Lewis wrote:Cosmology is taking a bit of a bashing - does it deserve it?
Remarkable. I was thinking I should put together a list of published counterindications to LCDM and maybe post it as a thread under the general Arxiv forum for others to add to, and here Lieu goes and does the work for me - on my birthday! You know that feeling "somebody up there must like me"? :)

Since I started looking seriously at this a couple of years ago, I've definitely noticed a tendency to cherry-pick supporting evidence and ignore/suppress the rest. Cosmology has never been "real physics", so I imagine this kind of thing has always been going on, but a line was crossed with dark energy: that's something which, for a change, has a profound impact on physics, so it's no longer possible for others to just ignore what's going on "over there", in the cosmology nutter's camp. An outsider might say that with the claim of accelerated expansion, cosmology became too important to be left to the cosmologists (is this statement going to make me friends here or what?).

Dark energy is the observational leg of the anthropic landscape. That's truly serious stuff striking at the very core of physics, of what physics can even ultimately hope to be. With that kind of impact comes a new responsibility. To present the LCDM interpretation of the data to other communities as if it were an established fact without alternatives may be in line with Lev Landau's classic quip that "cosmologists are often wrong, but never in doubt", but at this level, it is no longer acceptable. There's definitely something wrong when highly competent HEP theorists react with astonishment when pointed to papers by people like Lieu and Sarkar. What, you mean there are alternatives...?

Besides problematic communication with outsiders, there's something else I've noticed: looking at bios (be afraid, be very, very afraid ;) I see that there are now young people whose entire career, from undergraduate project through doctoral thesis to postdoc projects has been all about one thing only, dark energy. That is not healthy. It means having staked one's entire academic future on a single speculative idea which therefore must be true, or you're off to open an Italian restaurant in Novosibirsk. Hardly conducive to an unbiased stance.

That sort of thing can become self-perpetuating. Give it 5-10 more years and these people will start producing the second generation of dark-energy-only cosmologists...

Thomas Dent
Posts: 26
Joined: November 14 2006
Affiliation: ITP Heidelberg
Contact:

[arXiv:0705.2462] $Lambda$CDM cosmology: how much suppress

Post by Thomas Dent » May 18 2007

This kind of paper seems to me counterproductive, since it mixes crude polemic ('fifty years of failed attempt') and childish name-calling ('cosmology isn't even astrophysics' / 'groping in the dark') plus highly subjective views on the 'right' way of doing astronomy or astrophysics, with what seems to be actual scientific argument on the merits of LambdaCDM. If you believe that many cosmology papers constitute propaganda, and that this is a bad thing, why write something that obviously contains a large percentage of counter-propaganda?

The downgrading of cosmology and astronomy as not 'hard core physics' with 'no independent checks' also seems to me both wrong, and potentially dangerous and self-defeating. Can we use astronomy and cosmology to answer basic and important scientific questions? I would hope the answer would be Yes - and if not, what is the point?

Surely the observational situation is no worse than in high energy physics: every new experiment provides new data (hopefully a lot of it!) and is an opportunity to test models. The road from raw data to conclusions may be every bit as rocky in particle physics as it is in cosmology, since every new experiment has its own, possibly extremely complicated, systematic issues. But still we think it constitutes science.

By the way, the Blanchard et al. model is now considerably complicated, if not yet quite ruled out, by the apparent BAO signal.

astro-ph/0512085 (Blanchard et al.):
Our best-fit E--deS models do possess a baryonic feature at a similar physical scale as the best-fit [tex]\Lambda[/tex]CDM concordance model, but do not fit the new observations as well as the latter. In particular the shape of the correlation function in the range [tex]\sim 10-100 h^{-1}[/tex] Mpc cannot be reproduced properly without violating the CMB angular power spectrum in the multipole range [tex]l \sim 100-1000[/tex]. Thus, the combination of the CMB fluctuations and the shape of the correlation function up to [tex]\sim 100 h^{-1}[/tex]Mpc, if confirmed, does seem to require dark energy for a homogeneous cosmological model based on (adiabatic) inflationary perturbations.
What seems to me missing in the Lieu 'scoresheet' is some measure of how bad each of the 'No's is. I.e. how much pain we have to endure to make the model compatible with the data. Is needing a Hubble constant of 50 km/s/Mpc and no SN1a acceleration and wrong BAO and wrong primordial abundances better or worse than the problems attributed to LambdaCDM?

For the Einstein-de Sitter models to be correct, several different and apparently unconnected types of observation need to be very wrong or wrongly interpreted. For LambdaCDM to be correct, the outstanding claimed problems are all cluster-related. A naive interpretation of this is that we don't understand clusters (cf discussion between Afshordi and Blanchard, http://cosmocoffee.info/viewtopic.php?t=296) - I can live with that for now

As for problems where there is no viable model, like the baryon count, what is the relevance to the clash of models?

Pier Stefano Corasaniti
Posts: 43
Joined: November 11 2004
Affiliation: LUTH, Observatoire de Paris-Meudon
Contact:

[arXiv:0705.2462]

Post by Pier Stefano Corasaniti » May 18 2007

damn it...somebody has already had my idea...http://www.e-novosibirsk.com/eat.php...so I think I have to stick to cosmology then ;0)

Pier-Stefano

Alessandro Melchiorri
Posts: 129
Joined: September 24 2004
Affiliation: University of Rome
Contact:

[arXiv:0705.2462]

Post by Alessandro Melchiorri » May 19 2007

Hi,

I fully agree with Thomas, this paper is just pathetic. 1st april would have been a better submission date. I wonder why the arxiv people are letting these papers to be posted. There is no science: just (bad) politics and whining.

I wonder what kind of argument will come next ? cosmology is against God ? CMB radiation producing cancer ? Dark Energy bad for the environment ?


ciao
Alessandro

Tommy Anderberg
Posts: 47
Joined: November 24 2005
Affiliation: independent

Re: [arXiv:0705.2462] $Lambda$CDM cosmology: how much supp

Post by Tommy Anderberg » May 19 2007

Since I really don't know when it's time to shut up...
Thomas Dent wrote:Surely the observational situation is no worse than in high energy physics
It is obviously much worse. If you see a very unlikely event in a particle collider, you collide a gazillion more particles and see if it happens again, and if so how often. Keep doing it long enough and you can determine whether the theory you're using to compute those likelihoods is falsified or not. If you see a very unlikely feature in the CMB, what do you do? You have only that one CMB. Fluke or physics unaccounted for? No way to tell from astronomical observation alone.
Thomas Dent wrote:By the way, the Blanchard et al. model is now considerably complicated, if not yet quite ruled out, by the apparent BAO signal.
Isn't this why Sarkar has been pounding the table on the assumptions about the inflationary spectrum going into the LCDM fit to the baryonic peak?
Thomas Dent wrote:What seems to me missing in the Lieu 'scoresheet' is some measure of how bad each of the 'No's is. I.e. how much pain we have to endure to make the model compatible with the data.
I agree, though we may disagree about what constitutes pain. ;) To a Taliban like me, the amount of fine-tuning required to make LCDM work is pure torture. That should be reflected in the badness number. How about counting the number of parameters, the right way, decimal by decimal?
Thomas Dent wrote:For LambdaCDM to be correct, the outstanding claimed problems are all cluster-related.
Maybe from a very restricted observational point of view... BTW, what happened to APM 08279+5255 (astro-ph/0509212)?
Thomas Dent wrote:As for problems where there is no viable model, like the baryon count, what is the relevance to the clash of models?
Maybe just a reminder that one should not get too entrenched defending any one of them, because they are all likely to be wrong anyway?
Pier Stefano Corasaniti wrote:damn it...somebody has already had my idea
See, this is why you youngsters need to think more outside the box: how about a Siberian restaurant in Rome? :)

Pier Stefano Corasaniti
Posts: 43
Joined: November 11 2004
Affiliation: LUTH, Observatoire de Paris-Meudon
Contact:

[arXiv:0705.2462]

Post by Pier Stefano Corasaniti » May 19 2007

See, this is why you youngsters need to think more outside the box: how about a Siberian restaurant in Rome? :)


hey thank you so much..that's sound such a great idea...hope you will be my first customer.

Thomas Dent
Posts: 26
Joined: November 14 2006
Affiliation: ITP Heidelberg
Contact:

[arXiv:0705.2462] $Lambda$CDM cosmology: how much suppress

Post by Thomas Dent » May 21 2007

Can anyone explain what this so-called 'fine-tuning' is in LCDM? Something to do with the size of Lambda in Planck units?? But that's not a valid argument unless you have some meaningful prior distribution to argue with. Which is what?
The supposed alternative is that Lambda=0 exactly - by means of ... of ... intervention of magic fairies, I guess, since there is absolutely no good theoretical reason why it should be zero.

It seems to me a remarkably simple model in terms of parameters, you have (along with the usual inflationary cosmology parameter set) simply the density of vacuum energy and the density of CDM, and if it's correct these numbers should be consistently testable by a wide range of observables both now and in the future. (Though perhaps you need the mass and annihilation cross-section of DM in order to deal with some small-scale physics...)

Incidentally I don't count not-understanding clusters as an observational issue. It is clearly a theoretical problem too since one ought to be able to model them in whatever favourite scenario one has. But they happen to be rather complicated things in almost every way. In contrast something like BBN happens when the Universe was (according to most models!) almost perfectly homogeneous and in thermal equilibrium.

Tommy Anderberg
Posts: 47
Joined: November 24 2005
Affiliation: independent

Re: [arXiv:0705.2462] $Lambda$CDM cosmology: how much supp

Post by Tommy Anderberg » May 21 2007

Thomas Dent wrote:Something to do with the size of Lambda in Planck units?
Yes. Lambda is just a constant in GR, but GR is just a classical approximation to something else, which had better be a quantum theory (since the matter contribution to Einstein's equations is ruled by one; try to treat it as an expectation value while keeping the gravitational degrees of freedom classical and you quickly run into weirdness, like no gravitational effects in the z direction from a particle which has gone through a textbook-style Stern-Gerlach experiment separating it into two equal-probability wave packets moving up and down, respectively). In such a theory, Lambda becomes something which you compute, not something which you put in by hand. That computation is naturally done in Planck units, so on dimensional grounds alone you'd expect a finite result to be something on the Planck scale.
Thomas Dent wrote:But that's not a valid argument unless you have some meaningful prior distribution to argue with.
Now you're talking data fitting, not theory.
Thomas Dent wrote:The supposed alternative is that Lambda=0 exactly - by means of ... of ... intervention of magic fairies, I guess, since there is absolutely no good theoretical reason why it should be zero.
Actually, Lambda = 0 is what you get naturally both in SUGRA/strings and LQG. In the latter, people resorted to using deformed symmetry groups (quantum groups) to put in a small Lambda by hand after 1998. In strings, what might look like a universe with a cosmological constant is not the real ground state of the theory, it's just a very long-lived metastable state out of who knows how many possible ones populating the KKLT landscape... which, if you believe in it, is a pretty disastrous concept from the point of view of predictive power.

You need not believe in any of those though. It's enough to note that if Lambda were not very close to 0, we wouldn't be around to discuss it. If it were something of order unity on the Planck scale, we definitely would not be. But that would be the natural size to expect from radiative corrections to any small tree level value, so either you believe that some miraculous cancellation occurs by chance in the correct theory of quantum gravity, making all those corrections balance each other order by order all the way up to the Planck scale, leaving only a tiny little remain for us to observe, or you conclude that there is some symmetry at work which actually enforces Lambda = 0.

Bottom line: a model featuring a tiny but non-zero Lambda may seem remarkably simple at the classical level, but it is quite unnatural at the quantum level.

Thomas Dent
Posts: 26
Joined: November 14 2006
Affiliation: ITP Heidelberg
Contact:

Re: [arXiv:0705.2462] $Lambda$CDM cosmology: how much supp

Post by Thomas Dent » May 23 2007

Tommy Anderberg wrote: In such a theory, Lambda becomes something which you compute(...) That computation is naturally done in Planck units, so on dimensional grounds alone you'd expect a finite result to be something on the Planck scale.
Quantum gravity, in other words. Fine - but if you imagine computing the energy scale of inflation, the age of the Universe, the Higgs mass, the cosmological density of matter, indeed almost anything dimensionful with such a theoretical preconception, you expect an answer of about 1 in Planck units. You can't invoke the quantum gravity argument for just one parameter of a cosmological model and ignore the effects on the rest of the framework.
Tommy Anderberg wrote:
Thomas Dent wrote:But that's not a valid argument unless you have some meaningful prior distribution to argue with.
Now you're talking data fitting, not theory.
- I'm talking both! As explained at tedious length elsewhere, the prior distribution can depend both on old data and on theoretical prejudices: for example the idea of eternal inflation and tunnelling in a 'landscape'. Or the idea of an unusual relation between the SUSY-breaking scale and Lambda, a la Tom Banks.
Tommy Anderberg wrote:Lambda = 0 is what you get naturally both in SUGRA/strings and LQG. In the latter, people resorted to using deformed symmetry groups (quantum groups) to put in a small Lambda by hand after 1998. In strings, what might look like a universe with a cosmological constant is not the real ground state of the theory, it's just a very long-lived metastable state(...)
I'm perfectly well aware of the situation in SUGRA / strings : first in SUGRA, one can have a supersymmetric vacuum with negative Lambda; second, the generic ('natural'?) model has broken SUSY, which may have positive, negative or zero Lambda. (It is still debatable whether tunnelling to negative-Lambda states plays any significant role.) Since we don't live in a truly supersymmetric world, SUSY is not that much help, though it does reduce the hierarchy a bit. As for LQG I'm not sure enough is known about Minkowski space in that theory to start worrying about cosmology.
Tommy Anderberg wrote:(...) if Lambda were not very close to 0, we wouldn't be around to discuss it. If it were something of order unity on the Planck scale, we definitely would not be. But that would be the natural size to expect from radiative corrections to any small tree level value, so either you believe that some miraculous cancellation occurs by chance in the correct theory of quantum gravity, making all those corrections balance each other order by order all the way up to the Planck scale, leaving only a tiny little remain for us to observe, or you conclude that there is some symmetry at work which actually enforces Lambda = 0.
I don't follow this paragraph. All I can say is that I think Weinberg's 'anthropic' paper on Lambda is still more or less correct. Therefore if you have a theory, or theoretical prejudice, that gives a prior distribution of Lambda which has a nonvanishing density in the vicinity of zero, then apply a Weinbergian prior, you will end up with an expectation that the observed value should be nonzero and not too large.

Of course a theory that actually gave the correct answer, or even zero, would be almost infinitely preferable, but we don't have one. Comparing an known but crappy model with an unknown but (ex hypothesi) totally wonderful model is magical fairy talk.
Tommy Anderberg wrote:a model featuring a tiny but non-zero Lambda may seem remarkably simple at the classical level, but it is quite unnatural at the quantum level.
A model containing at least 10^78 protons and about 10^10 times as many photons is also pretty unnatural, isn't it? The natural value for these numbers are ... well, who knows.

Anyway, I'm curious to know what Lieu means by 'elite institutions' with 'vested interests', and why these are (as implied) a bad thing. Some universities and departments have better reputations than others, also people prefer to look at one thing rather than another in allocating funding because they find one approach better or worse or more or less interesting than another. How could it be otherwise?

Post Reply