Skip to the Main Content

Note:These pages make extensive use of the latest XHTML and CSS Standards. They ought to look great in any standards-compliant modern browser. Unfortunately, they will probably look horrible in older browsers, like Netscape 4.x and IE 4.x. Moreover, many posts use MathML, which is, currently only supported in Mozilla. My best suggestion (and you will thank me when surfing an ever-increasing number of sites on the web which have been crafted to use the new standards) is to upgrade to the latest version of your browser. If that's not possible, consider moving to the Standards-compliant and open-source Mozilla browser.

August 31, 2007

Axis of Evil? Or Axis of Opportunity?

Posted by John Baez

This paper argues that the axes of elliptical galaxies don’t point in random directions. but have a slight tendency to point roughly towards the ‘axis of evil’. The ‘axis of evil’ is a controversial concept in itself: a direction that seems to be picked out by anisotropies in the cosmic microwave background!

Trying to put a more positive spin on the idea, the author calls it the ‘axis of opportunity’:

I thank Daniel Rocha for pointing out this paper to me.

Here’s an elliptical galaxy:

It’s called ESO 325-G004, and it’s about the mass of the Milky Way. It’s in the middle of Abell S0740, a cluster over 450 million light-years away in the direction of the constellation Centaurus.

Longo has analyzed a sample of 200,000 of elliptical galaxies with redshifts ZZ < 0.20.2 from the Sloan Digital Sky Survey. They look round when their spin axes are pointing directly towards us, and like more or less eccentric ellipses otherwise. So, by correlating their observed shapes with their positions in the sky, one can see if their axes tend to point the same way.

The author claims that yes, there’s a statistically significant tendency for them to point towards the direction with right ascension α=202 \alpha = 202^\circ and declination δ=25 \delta = 25^\circ. In fact, he claims the effect is very strong. Interestingly, the quadrupole and octopole moments of the cosmic microwave background radiation (or CMB) seem to pick out roughly similar directions as being important.

Since I’m not an expert on this stuff, I can’t evaluate this paper. So, all I have is a lot of questions.

How convincing is the statistical analysis in this paper? Longo claims a high degree of statistical significance… but what about his method?

(I note that Longo only studies the amount of apparent ellipticity of the galaxies he studies, not the directions these ellipses point. It would be fun to take the directions into account, and see if they confirm his results. But, that’s sort of a separate issue.)

What’s the current conventional wisdom on the ‘axis of evil’? I see only a few papers on it on the arXiv, mainly by a handful of supporters, including Kate Land, Joao Magueio, and the author of this paper, Michael Longo. Magueio has been called the ‘bad boy of cosmology’ for his support of variable speed of light theories, and Longo has proposed an explanation of the axis of evil based on magnetic fields. Neither of these facts inspire confidence in me.

Furthermore, I’ve seen just one paper trying to explain away the axis of evil. Usually when an anomaly is taken seriously, you find more debate in the literature. Maybe I’m not looking in the right places?

On the other hand, even that bastion of reasonableness Sean Carroll has studied cosmological models with a preferred direction — and he doesn’t dismiss the axis of evil off-hand:

The claim that there actually is evidence for a preferred direction in the CMB goes by the clever name of the axis of evil. If one looks closely at the observed anisotropies on the very largest scales, two interesting facts present themselves. First, there is less anisotropy than one would expect, on very large angular scales. Second, and somewhat more controversially, the anisotropy that does exist seems to be oriented along a certain plane in the sky, defining a preferred direction perpendicular to that plane. This preferred direction has been dubbed the “axis of evil”.

Is the axis of evil real? That depends on what one means by “real”. Nobody has a theory that predicts CMB anisotropy directly as a function of position on the sky — rather, theories like inflation probabilistically predict the amplitude of anisotropy on each angular scale.It does seem to be there in the data. On the other hand, maybe it’s just a fluke. But at each scale there are only a fixed number of independent observations one can make, implying an irreducible uncertainty in ones predictions - that was the original definition of cosmic variance, before we re-purposed the phrase. For what it’s worth, the actual plane in the sky defined by the large-scale anisotropy seems to coincide with the ecliptic, the plane in which the various planets orbit the Sun. Many people believe it’s just some local effect, or an artifact of a particular way of reducing data, or just a fluke — to be honest, nobody knows.

So, maybe that’s the conventional wisdom — “nobody knows”.

Posted at August 31, 2007 10:11 AM UTC

TrackBack URL for this Entry:

9 Comments & 0 Trackbacks

Re: Axis of Evil? Or Axis of Opportunity?

Interesting. I’ll read that paper, thank you.

Back in my MSc dissertation, I studied the center of the Shapley Concentration, the cluster of galaxies A3558. Then, I noticed something very curious, that I have reported in my paper:

Evidence of Substructure in the Cluster of Galaxies A3558
Dantas et al.; Astrophys.J. 485 (1997) 447.

I copy here the relevant part:

we qualitatively note a preferential alignment at the 45 deg position angle of a diverse set of features in this cluster, ranging from small to large scales. An application of the Lee-statistics (Fitchett 1988) to this cluster also indicates that the 45 degrees direction has a greater probability of bimodality. This alignment coincides with (a) the major axis PA of the dominant galaxy, (b) the major axis PA of the isocontours of the central core (using AK/WT maps as well as a X-ray countour map), (c) a marginal velocity gradient direction across the central bimodal substructures, and (d) the direction of alignment of major subclumps, namely the bimodal core and SC1327-312. This alignment seems to persist even beyond the analysed field. A qualitative inspection of an isopleth map of a 32 deg \times 32 deg region around A3558 (see Figure 2 of Raychaudhury et al.1991), also indicates an alignment of all major clusters in the Hydra-Centaurus region spanning from position angles of 40 deg to 60 deg. Considering the physical dimensions involved, it is quite surprising that the average alignment of major clusters in the Shapley Concentration is consistent with the direction of alignment of several substructures within A3558 and with the major axis position angle of its dominant galaxy. We find that these small to large scale coincidental associations (namely, cD major axis match to the general clustering alignment) can be taken as an observational evidence of an anisotropic merger scenario as for instance the one proposed by West (1994a). (Page 16 of my paper).

I cannot tell you now whether such an alignment coincides with the Axis. It has been a long time since I was directly involved with studying clusters of galaxies.


Posted by: Christine Dantas on August 31, 2007 1:43 PM | Permalink | Reply to this

Re: Axis of Evil? Or Axis of Opportunity?

Usually when an anomaly is taken seriously, you find more debate in the literature. Maybe Im not looking in the right places?

Well, I would expand your search. Look for WMAP anomalies, peculiarities, quadrupole, problem, statistical anisotropy, etc. There are certainly many more papers on this problem than those that you have found.

I think this site is extremely interesting for exposing some techniques that were used to identify the anomalies in WMAP. There are other techniques as well. But I am not expert on this matter.

Posted by: Christine Dantas on August 31, 2007 2:37 PM | Permalink | Reply to this

Re: Axis of Evil? Or Axis of Opportunity?

I am an astronomer who passes through here to mostly get confused about category theory.

I have not read the paper in detail, but I will post a few comments about this.

First, the authors sample does not consist of elliptical galaxies. Instead, it consists of red galaxies. In the field, roughly 1/3 of all red galaxies are spirals. Of the remaining 2/3, probably half of those are not ellipticals, but lenticular galaxies. Lenticular’s, also called S0’s, are disk dominated galaxies with strong rotation support, but no spiral arms and little or no ongoing star-formation. This author is not the first person to make this mistake, but it does not inspire confidence.

Second, the author picks, as the semi-major and semi-minor axis measurement, the results of fitting a model to the data by the SDSS photometric pipeline. He decided to use the fits of an exponential model. Elliptical and bulge dominated galaxies (like most red spirals and lenticulars) are not well fit by exponential models.

Third, from a measurement statement, there is a real apples and oranges comparison here. The typical seeing for a SDSS image is around 1 or so seconds of arc. This is the angular scale at which useful measurements can be made. In physical terms, that is 400 pc, 1300 light years for a galaxy at a redshift of .02. At the other end of the range, the author picks 0.2 as the redshift limit. The same angle subtends 3300 pc, or ~11000 light years.
For comparison, a typical elliptical galaxy has a half-light radius of 2-3000 pc. The problem becomes what do the ellipticities mean when one is sampling a very different part of the galaxy’s surface brightness profile. In addition, galaxies will appear smaller at that distance and, because the earth’s atmosphere blurs the image with a smoothing filter, one has less information for reconstructing the underlying ellipticity. The pipeline models this blurring, and tries to remove the effect of it from the model fits to the data, but there are limits to how much can be recovered.

The statistics look ok, from looking at the figure. I have certainly published rattier looking data.

It would be interesting to compare the position the author finds with the scan directions of the SDSS. The SDSS does not take single images, but instead using a technique called drift-scanning which moves the built-up charge from a given object across the detector at the same rate the object is moving across the sky. One can imagine how this might cause a small bias in the ellipticity measurements.

As for what Christine Dantas has found, others have found similar work, particularly Mike West, as she references. The interpretation is that the cluster of galaxies are dynamically young. They have relaxation timescales measured in gigayears and are still assembling today, so much of the initial formation process is still imprinted in the cluster, and in the galaxies inside the cluster.

Posted by: Brad Holden on September 1, 2007 1:54 AM | Permalink | Reply to this

Re: Axis of Evil? Or Axis of Opportunity?

Brad Holden wrote:

I am an astronomer who passes through here to mostly get confused about category theory.

Cool! I hope we’re doing our job — confusing you about category theory.

Seriously: if you have some questions, feel free to ask ‘em. These days I’m usually too busy trying to follow what Urs is saying to write expository posts, but I used to do a lot of that over on sci.physics.research.

First, the authors sample does not consist of elliptical galaxies. Instead, it consists of red galaxies. In the field, roughly 1/3 of all red galaxies are spirals. Of the remaining 2/3, probably half of those are not ellipticals, but lenticular galaxies.

Hmm! Yes, the author explains how he made the ‘cut’, and he says: “Elliptical galaxies tend to have mostly old, red stars. Therefore a fairly pure sample can be selected on the basis of their color alone.”

I hadn’t noticed the slip in logic there. It’s a bit like saying all black geese are black, so you can select them from other objects on the basis of their color alone.

I wonder if a contamination of the population by spirals and ellipticals could make there seem to be anisotropy when there really wasn’t.

All the other points you make are also very interesting. I wish it were feasible to correct for some of these effects, or conduct other, better-designed searches for signs of anisotropy. Sounds like the high-risk, potentially-high-payoff work that the FQXi would like to support.

Posted by: John Baez on September 1, 2007 10:41 AM | Permalink | Reply to this

Re: Axis of Evil? Or Axis of Opportunity?

If our good hosts at the Café haven’t confused you enough, I’ve been compiling a list of category-theory references available free on the Web. John Armstrong continues to write many brief essays on the subject, but I’ve rather fallen behind on cataloging them.

Posted by: Blake Stacey on September 1, 2007 3:28 PM | Permalink | Reply to this

Absolute Reference Frame? Re: Axis of Evil? Or Axis of Opportunity?

A “preferred direction” in space strikes at the heart of the evolution on the philosophy of mathematical physics from the Absolute Reference Frame of Newton et al to the relative inertial reference plane of Einstein et al, with a nod to Mach along the way.

This, and the “hole in space” a billion light years across are significant discoveries about our cosmos, and about the physical substrate for any theory of space and time.

Leading experts from WMAP and COBE whom I’ve asked at conferences agreed that the “axis of evil” (however re-named) seems to be real, and not an observational artifact. What it really means may take as long to unravel as the accelerating expansion of the cosmos.

Posted by: Jonathan Vos Post on September 1, 2007 5:30 PM | Permalink | Reply to this

Re: Axis of Evil? Or Axis of Opportunity?

Jonathan writes:

A “preferred direction” in space strikes at the heart of the evolution on the philosophy of mathematical physics …

Right, that’s why it’s interesting. But note this: it’s completely consistent with current physics. Unlike the underlying laws, the observed state of the universe is clearly not invariant under rotations or translations. So, it’s all just a matter of how much anisotropy and inhomogeneity we see.

Just a little, and it’s probably sheer coincidence — or what people now call cosmic variance. Indeed: start with a perfectly homogeneous and isotropic quantum state early on, shortly after the Big Bang. Look later, after gravity has done its thing, and you’ll find the symmetry seems to be broken! There will be inhomogeneities and anisotropies in any of the quantum states you observe, even though their superposition retains all the original symmetry.

(Call it ‘collapse of the wavefunction’, call it ‘Everett branching’ — just don’t tell me what you call it: I don’t give a damn. It’s just what quantum mechanics predicts.)

So, even given maximally symmetric initial conditions, we expect anisotropy and inhomogeneity in the universe we see.

The question is, how much should we expect?

But, the answer to this question is inherently probabilistic. Our theories don’t give a definite answer, just a probability distribution. So, if we see an unexpectedly large amount of anisotropy, it could be a clue that our current theories are wrong. Or, it could just be a fluke — a low-probability occurence.

Which is it? One disturbing possibility is that we’ll never know. But, I’m an optimist: I think we’ll figure this stuff out someday.

However, I’m not up-to-date when it comes to research on this subject! I don’t even know if anyone has tried to compute the chance of seeing as much (or more) anisotropy as we’re actually seeing. Surely someone must have tried, no? What did they find?

Posted by: John Baez on September 2, 2007 10:49 AM | Permalink | Reply to this

Re: Axis of Evil? Or Axis of Opportunity?


An old (April 2007) claim:

Michael Longo of the University of Michigan in Ann Arbor analysed 1660 spiral galaxies from the Sloan Digital Sky Survey and found that the axes of rotation of most galaxies appear to line up with the axis of evil ( According to Longo, the probability of this happening by chance is less than 0.4 per cent. “This suggests the axis is real, and not simply an error in the WMAP data,” he says.

Posted by: Jonathan Vos Post on September 2, 2007 10:34 PM | Permalink | Reply to this

Re: Axis of Evil? Or Axis of Opportunity?

On the other hand, Sean Carroll undercuts the plausibility of probability argument is this context as follows:

[Adolf] Grünbaum’s takedown of this position, which was quite careful and well-informed. But the basic idea is straightforward enough. When we talk about things being ‘natural’ or ‘spontaneous’ we do so on the basis of our experience in this world. This experience equips us with a certain notion of natural – theories are naturally if they are simple and not finely-tuned, configurations are natural if they aren’t inexplicably low-entropy.

But our experience with the world in which we actually live tells us nothing whatsoever about whether certain possible universes are ‘natural’ or not. In particular, nothing in science, logic, or philosophy provides any evidence for the claim that simple universes are ‘preferred’ (whatever that could possibly mean). We only have experience with one universe; there is no ensemble from which it is chosen, on which we could define a measure to quantify degrees of probability. Who is to say whether a universe described by the non-perturbative completion of superstring theory is likelier or less likely than, for example, a universe described by a Rule 110 cellular automaton?

It’s easy to get tricked into thinking that simplicity is somehow preferable. After all, Occam’s Razor exhorts us to stick to simple explanations. But that’s a way to compare different explanations that equivalently account for the same sets of facts; comparing different sets of possible underlying rules for the universe is a different kettle of fish entirely. And, to be honest, it’s true that most working physicists have a hope (or a prejudice) that the principles underlying our universe are in fact pretty simple. But that’s simply an expression of our selfish desire, not a philosophical precondition on the space of possible universes. When it comes to the actual universe, ultimately we’ll just have to take what we get.

Posted by: Jonathan Vos Post on September 3, 2007 12:28 AM | Permalink | Reply to this

Post a New Comment