Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Fine structure constant probably doesn't vary with direction in space!

  1. Aug 29, 2010 #1

    bcrowell

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    The thread "Fine Structure Constant Varies With Direction in Space!" was locked because it didn't cite papers published in refereed journals. Actually all of this stuff has been published in refereed journals. The list of references below is cut and pasted from http://en.wikipedia.org/wiki/Fine_structure_constant . Quite a few of the papers are also on arxiv.

    My own opinion is that Webb et al. are wrong. Extraordinary claims require extraordinary evidence. Their evidence is statistically significant if you (a) believe their error bars, (b) believe that there were no unidentified systematic errors, and (c) believe that, as claimed by Webb, the Chand group's failure to reproduce the result is due to statistical mistakes by Chand et al., rather than being due to the nonexistence of the purported effect. Even if I believed a, b, and c, I wouldn't consider it statistically significant at the level that would make me believe such an extraordinary claim. It would be interesting to hear whether the Chand group has ever responded to the statistical criticisms. If you buy the idea that the fine structure constant varies over time, then it's actually not much of a leap to believe that it varies spatially as well. If it only varied with time in one frame of reference, it would vary in both time and space in another frame that was moving relative to the first. If it depended on cosmological parameters, I suppose it would be surprising to see an anisotropy that was observable in the frame of our own galaxy, which is more or less moving with the Hubble flow.

    J.K. Webb et al. (2001). "Further Evidence for Cosmological Evolution of the Fine Structure Constant". Physical Review Letters 87 (9): 091301. doi:10.1103/PhysRevLett.87.091301. arXiv:astro-ph/0012539. PMID 11531558.
    M.T. Murphy, J.K. Webb, V.V. Flambaum (2003). "Further evidence for a variable fine-structure constant from Keck/HIRES QSO absorption spectra". Monthly Notices of the Royal Astronomical Society 345: 609. doi:10.1046/j.1365-8711.2003.06970.x.
    H. Chand et al. (2004). "Probing the cosmological variation of the fine-structure constant: Results based on VLT-UVES sample". Astron. Astrophys. 417: 853. doi:10.1051/0004-6361:20035701.
    R. Srianand et al. (2004). "Limits on the Time Variation of the Electromagnetic Fine-Structure Constant in the Low Energy Limit from Absorption Lines in the Spectra of Distant Quasars". Physical Review Letters 92: 121302. doi:10.1103/PhysRevLett.92.121302.
    M.T. Murphy, J. K. Webb, V.V. Flambaum (2007). "Comment on “Limits on the Time Variation of the Electromagnetic Fine-Structure Constant in the Low Energy Limit from Absorption Lines in the Spectra of Distant Quasars”". Physical Review Letters 99: 239001. doi:10.1103/PhysRevLett.99.239001.
    M.T. Murphy, J.K. Webb, V.V. Flambaum (2008). "Revision of VLT/UVES constraints on a varying fine-structure constant". Monthly Notices of the Royal Astronomical Society 384: 1053. doi:10.1111/j.1365-2966.2007.12695.x.
     
  2. jcsd
  3. Sep 1, 2010 #2
    I'm *really* unconvinced by this paper......

    What the paper is saying is that they've been doing all these studies that say that the fine structure constant is changing with time. O.K. Then they take a dataset from a different telescope pointing in a different direction and they find that the fine structure constant is changing in a *different* way. So the explanation they come up with is that the fine structure constant changes with direction, but it seems more likely to me that there is a calibration issue.

    Something that would be interesting is to try to do the analysis of the same object with different telescopes and see if you get the same result.

    Also, something else to look at would be systemic differences in things like deuterium abundance with respect to angle. If you looked in different parts of the sky, and saw different elemental abundances that match the differences in the fine structure constant, then there might be something there.
     
  4. Sep 1, 2010 #3
    Also, finding a cosmological dipole dependency is really problematic from a theory point of view. If you look at the CBM, you do see a dipole variation, but that's the result of the earth's moving.

    You can see the problems if you assume that there really *is* a dipole variation in the fine structure constant. OK, you observe a dipole variation from earth. Let's assume its real. Now ask yourself what it's going to look like from a point that's 10 billion light years in some given direction of earth. Something that looks like a dipole from Earth is not going to look like a dipole from a different part of the universe, so you have to explain why earth is somehow special.

    Also if you think of the time evolution of the fine structure constant, you run into a lot of problems. If the results were real, that would suggest that you have the universe start off with different fine structure constant values in different parts of the universe, and then magically everything converges as you are getting closer to the current time.

    This gets you to the horizon problem. How do different parts of the universe that aren't casually connected manage to coordinate their fine structure constants?
     
  5. Sep 1, 2010 #4
    many people dont know that the fine structure constant is simply the speed of the electron in a bohr atom expressed as a fraction of the speed of light.
     
  6. Sep 1, 2010 #5
    This actually seems to be a hot topic at the moment, on account of recent new data.

    http://www.technologyreview.com/blog/arxiv/25673/" [Broken]

    This is certainly a legitimate area of physics, and is frequently published in peer-reviewed journals. It may be controversial, but it isn't crack-pottery.
     
    Last edited by a moderator: May 4, 2017
  7. Sep 1, 2010 #6
    There is crack pot stuff published in astrophysics journals all the time. At least in astrophysics, the publication philosophy is to err on the side of letting nutty stuff in.
     
  8. Sep 2, 2010 #7

    Chronos

    User Avatar
    Science Advisor
    Gold Member

    I object to that paper because of 'cherry picking'. They select a tiny, skewed data set and attempt to assert 'statistically valid' conclusions. [Oh, the stories I could tell you about that] But, I have no problem with allowing papers like this to be published. It's fun, and who knows - they might be on to something. But I second quaint's sentiments. Journal papers are not gospel. If this paper is seminal and well received, it should get many cites.
     
    Last edited: Sep 2, 2010
  9. Sep 2, 2010 #8
    I think it's a pretty desperate reaction to the data they are getting. They've been trying for the last several years to show that there is some time variation in the fine structure constant. They then use another telescope, and the results from that other telescope show that the fine structure constant evolves in a different direction.

    The obvious result is that there is some observational error that they aren't taking into account. But they pretty then desperately come up with a model that says that the direction of the fine structure constant varies with direction. The problem is that the fact that they get this nice dipole variation very strongly suggests to me that they are seeing something but it's *not* the variation of the fine structure constant.

    Basically, if the fine structure constant is varying with direction then this doesn't work with an isotropic universe (i.e. it means that the universe has a preferred direction). That's fine, but then you should see something *other* than the fine structure constant vary.

    Things are often worse in the social sciences.

    One reason I tend to be polite about this is that 1) I don't want to look like an total fool if it turns out that the fine structure constant is changing 2) I want them to be nice to me if I do a major *whoops* and 3) they have real data.
     
  10. Sep 2, 2010 #9
    Curiously I don't have that much respect for peer review, and I don't care much for the idea that peer-review=good and non-peer-reviewed=bad.

    Part of this is because I've seen what peer review looks like in some other fields (finance, economics, and management) and in those areas I'd consider the peer review system to be seriously dysfunctional, and basically worthless as anything other than a political game.

    One reason that I don't think that astrophysics suffers from the dysfunctions that you see in economics is that astrophysics publication is a lot more "crackpot friendly."
     
  11. Sep 3, 2010 #10

    bcrowell

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I agree with you that it's almost certainly bogus. However, I don't think arguments about anisotropy and the horizon problem add that much to the evidence of bogosity.

    It's true that a dipole would be strange for the reasons you stated, but the whole effect probably doesn't exist in the first place. If it does exist, it's barely significant compared to random and systematic errors, and therefore any attempt to map variation across the celestial sphere is going to be extremely iffy. Since the significance of the whole effect is crap, their claim that it has the specific form of a dipole shouldn't be taken seriously.

    Anisotropy in the laws of physics would be revolutionary, and presumably would have shown up already in laboratory tests of Lorentz invariance. But anisotropy of the physical state of the universe is known to exist at some level; isotropy is just an assumption we make for convenience in constructing models. If the fine-structure constant varies, then it's just one more dynamical field like the electromagnetic field or the gravitational field. If there's a horizon problem for this new dynamical field, then it's no more of a horizon problem than the one that exists for all the other fields. It's true that you'd think that spatial variation of the field would be correlated with something else observable, but to me that's not such a significant issue compared to the fact that the whole thing is a crock.
     
  12. Sep 3, 2010 #11
    One thing that is something that is very useful from a teaching point of view is to compare the paper that says that fine structure constant is changing from the paper that first claimed that the expansion of the universe is accelerating, which is an equally extraordinary claim.

    The reaction I had when I first read the paper was similar to that of a lot of people. This is a nutty idea. It's obvious that what they are seeing is because of ..... Oh wait, on page such and such they should it can't be this. Well then, it must be because of .... Oh wait, they thought of that too .... Well if the universe really is accelerating, then you should see X, and we don't and so..... Oh wait, we actually do see X....

    What the authors of the accelerating universe paper did was to start out with 'odd observations" and they obviously showed to lots of people that trying to come up with plausible explanations other than the accelerating universe, before they were able to convince themselves that this is what they were seeing.

    I'm coming at this from a theorist point of view. If there really is a anisotropic variation in the fine structure constant then we ought to be able to see evidence of that in something totally unrelated. The problem with making claims from one set of observations is that you run the risk that there is some systematic bias that you don't know about. You can get around this by pulling in observations that are totally unrelated.

    To relating this to the accelerating universe. Yes, there *might* be something wrong with type Ia supernova observations, and yes it *might* be something that we haven't thought of. So in that situation, you try to come up with evidence that the universe is accelerating using observations that have nothing to do with supernova Ia observations (say CMB background variations).

    Now if the paper said, we looked at these quasar lines and they show some evidence of the fine structure constant changing, and we also looked at this other thing (and the obvious thing for me is nuclear abundances) and it's also changing in the same places, then that would be interesting. As it is, my explanation for their observations is that there is some systemic bias that they haven't taken into account, and you can't refute that by listing all the possible biases, because a bias that you haven't thought about or that you don't know can still mess up your numbers.

    What you really want is something just totally independent that gives you the same results. If they were seeing evidence of fine structure constant change from X-ray observations it would make it more interesting.

    Also, I should point out that the reason for being careful with these things is that a lot of the big discoveries of science happen when people look for X but find Y.
     
  13. Sep 3, 2010 #12

    turbo

    User Avatar
    Gold Member

    And that is why blanket dismissals of observations are dangerous to good science. I don't know why the WMAP data took so blasted long to be released, but I have an idea that it was due to the unexpected anisotropy that "should not have" been observed. Checking and re-checking for systematic errors and brain-storming for possible sources of the observed anisotropy was most likely the cause of the delays in the release of the early data-sets.

    In this case, the observed anisotropy is very small, and it is observed in some of the oddest most distant outliers in the cosmos - quasars. It may very well due to unmodeled systematic errors or some statistical anomaly. In either case, it would be a good argument for follow-up observations. Cosmology is a mostly theoretical field, but it must explain observations if it is to be a true science (hearkening back to Michael Disney, here).
     
  14. Sep 3, 2010 #13
    It is. But coming up with premature theoretical explanations also is troublesome. If you look for X, you stop looking for Y. Also part of the job of an observationalist is to come up with observations that are so obviously discordant that you can't easily dismiss them.

    No it wasn't. They were just overwhelmed with data.
     
  15. Sep 5, 2010 #14

    Chronos

    User Avatar
    Science Advisor
    Gold Member

    The WMAP dataset was not only huge, but, contaminated by artifacts due to instrument and software issues. It took time to subtract these errors from the final data release. Given NASA was besieged by funding shortfalls during this time the delay was inevitable, not conspiratorial.
     
  16. Sep 7, 2010 #15
    If the initial interpretation of the distances of quasars is seriously erroneous (in other words if their distances are very much closer than originally thought) what does that do to the model presented in the paper(s)? How could you alternatively interpret the anisotropy?
     
  17. Sep 7, 2010 #16

    turbo

    User Avatar
    Gold Member

    Even if the Arp-Burbidge crowd is right (intrinsic redshift), an observed anisotropy that varies with direction would still be a real head-scratcher.
     
  18. Sep 7, 2010 #17
    This is starting to sound very interesting and I suppose it's time to dig into the papers. One thing that might produce an anisotropy, that to my understanding is mathematically completely unparameterized starting from the Lorentz/Heavyside version of Maxwell's Equations and hence rippling up through the Lorentz Transformations and SR, is the velocity change relationship between EM fields and a sink.
     
  19. Sep 7, 2010 #18
    Can string theory accommodate fundamental constants varying in space/time, by allowing the underlying Calabi-Yau manifold to change shape? Or by a dilaton wave which messes up the physics wherever it goes?
     
  20. Sep 7, 2010 #19
    But what, exactly, is an "extraordinary claim"? The truth is that this is a subjective concept, i.e., if any
    specific claim is extraordinary or not is in the eye of the beholder. Even more subjective is the
    concept "extraordinary evidence". This is why the old saying "extraordinary claims need extraordinary
    evidence" is not part of the scientific method, but rather a way of dismissing results that do not fit well
    with mainstream theory. IMO, any talk about "extraordinary claims" and "extraordinary evidence" is
    just pseudo-scepticism that is frequently put forward when one has run out of scientific arguments.

    The claims of an accelerating universe versus a varying fine structure constant is a good example of
    this subjectivity. That is, first notice the fact that the accelerating universe is easy to model within
    mainstream theory without any change of the basic underlying theoretical framework. Second, notice
    that a varying fine structure constant, on the other hand, would violate the Einstein Equivalence
    Principle (EEP), and thus falsify one of the fundamental principles underlying modern gravitational
    theory. This is why, IMO, claims of an accelerating universe should be treated as far less
    extraordinary than claims of a varying fine structure constant. In other words, IMO, claims of an
    accelerating universe were never extraordinary, and neither were the evidence for it. On the other
    hand, IMO, claims of a varying fine structure constant are indeed extraordinary, and the evidence for
    it is weak. In other words, IMO, to claim that the claims of an accelerating universe and a varying fine
    structure constant are equally extraordinary, is an extraordinary claim!

    But some people obviously think every claim is extraordinary that does not agree 100% with the party line at any given time, and this proves my point.
     
  21. Sep 7, 2010 #20
    One thing that I saw with WMAP was "semi-conspiratorial."

    People working on WMAP were extremely tight-lipped about their data, so only a very small select group of people were allowed to touch the original data before general release. What this meant was that they weren't in a position to pull in more people and resources to get the analysis done quickly, because pulling in more people increased the chances that some of the results would have leaked out early. Personally, I don't think there was anything wrong with them doing this.

    This had to also do with funding because the people that had priority access to WMAP data were able to get priority access, because they were willing and able to put in the resources to make WMAP happen.

    There's also the general administrative problem with large projects is that if you state schedules in real time with real delays built in, you aren't going to get funding for them. One way around this involves using "business time" and "business money" which is different from "real time" and "real money."
     
  22. Sep 7, 2010 #21
    It's rather easy to get string theory to change varying fundamental constants. String theory doesn't impose any constraints on the value of fundamental constants, which is why anthropic views of the universe have gotten popular.

    You would need some sort of time and space varying field to get different fundamental constants. The problem with that is that if you've worked out the field strengths and you find that it's centered on earth, then this is very odd.
     
  23. Sep 7, 2010 #22
    I think it *is* part of the scientific method.

    It's a heuristic and a good one. Mainstream theories don't come from nowhere, and there is a vast amount of evidence that people have gone through to get to current theories. If you have something that people find extremely unexpected based on what has previously been known, you need to go through more trouble to demonstrate what is known is wrong.

    Something about science is that getting from raw data to a statement about the universe is something that is quite difficult and error-prone. There are lots of weird things to track down, and if you are claiming something weird, then it's *YOUR* job to convince me.

    And that can be done. The claim that the universe is accelerating is as extraordinary as the claim that the fine structure constant is changing, and personally I think that the original paper that made this claim is required reading for how to make a solid scientific argument for a very weird result.

    No it's not. I can point to the hundreds of theoretical papers on the Los Alamos Preprint server that trying to figure out what's going on. An accelerating universe causes a lot of theoretical problems that people are trying to grapple with. At the *very least* you have to add in "dark energy" and it's possible that this won't work.

    One basic theoretical problem with an accelerating universe is that it makes the period of time we are in "special".

    If the universe was at critical density, then the parameters of the universe would stay pretty constant over time, so if you picked a random time in the universe, you'll end up with the same numbers. Once you put in an accelerating universe, then it seems weird because then you have to fine tune everything to get the universe that we do see.

    1) No it doesn't since gravity doesn't enter into the fine structure constant, and
    2) I don't have any problem with EEP being wrong. So EEP is wrong, big deal. So is Euclidean geometry, parity, and the time-invariant coprenican principle. If someone came up with a theory that said that EEP was totally wrong, I wouldn't hold that against it strongly.

    This points out the subjectivity of extraordinary claims. Let me just say that when I first heard of someone claiming that the expansion of the universe was accelerating, I was sure that it was just another crackpot group writing some silly paper, and I could think of a dozen places where they could have made a mistake.

    However, the paper itself addressed all of the points that I could think of.

    If every claim is extraordinary then no claim is extraordinary.

    One other thing, you can believe whatever you want, it's convincing other people that's a problem. I have some truly wacky beliefs about how the universe works that I keep to myself. They are fun for discussions at parties, but I'm not going to write a paper on them, or expect anyone other than me to take them seriously because I don't have the evidence or arguments to back them up.

    Also, what exactly is the "party line"? Part of the reason that think the system works, is that I've seen enough crazy and ridiculous ideas become part of the party line, that I don't think that the standards of evidence that people require is bad for astrophysics.
     
  24. Sep 7, 2010 #23
    Also, I don't think that what I'm asking for here is weird or non-constructive. My reaction to the paper is that I'm pretty sure that they are looking at some experimental error or at best something local, and I've stated pretty clearly what would convince me otherwise.

    If the fine structure constant is changing over space, then this will affect things like deuterium abundances, so if you show that deuterium abundances or the CMB systematically varies in the same way as the purported fine structure constant, that eliminates experimental error or something local as an explanation for what is going on. The fact that they get a perfect dipole makes me really suspicious that there is something local going on.

    Also, I've stated my theoretical objections. If someone can show me how you can get an dipole variation in something and *not* have it be a local effect, that would be interesting (and quite useful for things other than this discussion).
     
  25. Sep 7, 2010 #24
    I don't get what you mean.. if something really is spatially varying in some arbitrary manner.. then with the first little bit of data, you would first try to calculate the monopole moment (i.e., their original claim of alpha varying in time) and with more/better data (especially not too limited in direction) you can expect to calculate next the dipole moment (the current claim), and expect to need even more/better data to be able to discern even higher order moments. In particular, if the scale of the spatial fluctuations happens to be much larger than the scale on which the measurement are able to be made then you would always expect to observe a dipole everywhere (except in the very unlikely circumstance that you happen to be located exactly in the middle of a saddle point).

    I'd agree that the paper is written poorly (or, alas, typically). I agree it is about as surprising a claim as accelerating expansion (OS: don't see that the fine structure constant has anything to do with equivalence of gravitation and inertia, though it's true the authors mention EEP too), so the authors would do better to take the time to more clearly repeat the basic physics of their measurements/observations/analysis, and narrate more from the perspective of first trying to defend the orthodox null hypothesis. (In fact, they do the opposite when they mention the 7th pair, hiding it from their initial discussion and then admitting this data-massaging afteward.) The shape of the expectation lobes in fig.1 seems to suggest an obviously inept (well, amateurish) approach to the statistics (it looks as if they did the analysis on R2 rather than S2, and then tranformed the result afterward, which technically is only approximately valid. They also neglect to label the milky way in fig.1).

    But if the data is legit, is there any other interpretation? If they have really double-checked several of these quasar spectra using both different telescopes, and found close agreement, it suggests they really are describing the spectra (rather than telescope artifacts). And if by studying groups of lines they really are able to eliminate redshift (or other intervening distortions), it seems hard not to conclude that physics (e.g., alpha) is different where those quasars are. (Unless.. I don't know, could they get a false result if different elements tend to aggregate at different levels of the gravitational well?) Then statistical correlation between the data from subsequent quasars according to their location in spacetime, that seems hard to explain away. (It'd be nice if we could completely understand the spectra, and even compare isotopic abundances as an independent test of alpha in those regions of the universe, but I take it that's a lot more difficult..)
     
    Last edited: Sep 7, 2010
  26. Sep 8, 2010 #25
    The trouble here is that there are limits to the amount of spatial fluctuations you can have if you are doing cosmology. If your fluctuations are too large in area, then you have to explain how two parts of the universe communicate with each other in the time that universe has been around. The size of any fluctuation is limited by the size of the observable universe and the time that the universe has been around, and if you have any fluctuations that are spatially larger than that, then you have some explaining to do.

    With that data by itself, if I have to chose between "I'm seeing something that is happening in the universe" or "I'm seeing something that is happening on earth" right now it looks like the latter. There are some easy ways of getting around this objection. If they run through the WMAP data and see something weird happening in the same direction or go through X-ray data and see the same thing, then maybe there is something there.

    There are about half a dozen I can think of off the top of my head. The most embarrassing would be some sort of equipment calibration that they didn't take into account (which has happened). There could be some local ISM effect. There could be some selection effect (i.e. you are more likely because of observational limitations to see certain types of quasars in certain parts of the sky). There could be some local gravitational lensing effect. There could be some systemic bias in distance calculations.

    Not convinced. In order to go from raw data to final conclusion, there are about a hundred different steps, any one of which could go wrong. One problem is that a systemic bias that you are not aware of is still a systemic bias. Part of the reason people are skeptical about these sorts of things is personal experience. Pretty much everyone has some story about some great discovery or observation that they had that turned out to be something silly.

    The other thing is that it's possible that you've figured out something amazing, but you aren't seeing it because you've got the wrong explanation. Part of the reason for considering why the observations might be the result of interference in the interstellar medium or intergalactic medium is that you may be seeing some probe for the IGM that no one has ever thought of. It's also quite possible that there is some interesting quasar physics that the authors are missing. There is a lot that we don't know about quasars, the IGM, or even the ISM, and if you see something weird, it's a bad idea to come up with an immediate explanation.

    Part of writing a scientific paper is that you have to write it in a way that convinces people that you won't be retracting it in two or three years because you forgot to take into account the eccentricity of the earth (which happened once). This means going over very carefully what you did and systemically going through every objection that someone can think of. The accelerating universe paper is a excellent example of how to do just that.

    Physics is weird because there is a lot of masochism involved. You have to take your greatest ideas and then grind them into dirt, and then see what survives.

    Me neither. Also I should point out that writing a paper asking (so what would the universe look like if the fine structure constant were varying) would be a fine paper.

    Or a true result. The most obvious implication of the fine structure constant changing that I can think of would be that the rates of nuclear reactions would change, and so if you showed that deuterium abundances are systematically different in different parts of the sky, that would be highly interesting.

    Also WMAP..... If you have different fine structure constants at z=3, then they are going to be very, very different at z=3000, and you should seem some systemic differences in the CMB.

    Or not.... If they had a discussion section, in which they explained why changing the fine structure constant *wouldn't* change the CMB, then I'd be open to that idea. But they haven't which means that they didn't think about it, which leads me to wonder what else they didn't think about.

    The first explanation that people usually give are selection effects. By looking at different parts of the sky, you are seeing different quasars. To give an example of how that could happen. If you point your telescope at one part of the sky, you can get a 12 hour exposure, and if you point it at some other part, you get a five hour exposure. This means that you detection limits are different, and this *WILL* bias your statistics.

    It's not, which is one reason I'm not taking these results too seriously.

    Looking at deuterium abundances is not too difficult. Now there may be some non-obvious reason why you *can't* look at deuterium abundances, but that also should go in the paper. Also the paper is pretty weak as to the observations that should be taken in the future to confirm or refute the reasons.

    Also I'd think about looking at transition lines between spin-orbital states like the 21-cm line. The reason that matters is that if you see a difference there, you don't have to worry about all of the data that happens between the quasar and the earth.

    I'd think seriously about non-statistical tests. For example, if you went out and found *ONE* star that had 100% hydrogen and no-helium, that would be enough to get you the Nobel prize for something. I'm sure that if you think hard enough about what would change if you changed the fine-structure constant, that you could find a "smoking gun" observation that wouldn't require the need for statistics. Personally, I'd think first about finding some ratio in atomic lines that requires no statistical processing.
     
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook