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

Phase shift in the Michelson-Morley experiment

  1. Mar 10, 2004 #1


    User Avatar
    Science Advisor

    The Michelson-Morley experiment is supposed to be able to detect a time difference between the paths of light in a parallel and in a perpendicular path to the aether wind, both two-way paths of the same length. It is argued that, if such time difference exists, it will be revealed as a phase shift between both reflected light beams.

    Actually a phase difference does not appear, which leads to the conclusion that the aether does not exist and that the speed of light must be constant on both paths.

    Now, in the following paper,


    the author argues that, although a time difference may exist, the MM experiment does not reveal it: no phase shift is produced, since the number of periods of light travelling thought the first path is always the same as the number of periods of light travelling thought the second path (taking into account the doppler effect). The paper concludes that the MM interferometer is ‘blind’ to speed.

    To me the calculations for the number of periods in the parallel and perpendicular paths done in chapter 2.d. seam to be correct (assuming of course that special relativity is invalid, as it is done before performing the MM experiment).

    Is this actually a known problem or may be there are some errors in the paper? Any comments? Thanks.

    Last edited: Mar 10, 2004
  2. jcsd
  3. Mar 12, 2004 #2
    Hellfire - I am surprised that there has not been any comment from the relativity experts on PF - I myself have considered whether the failure to detect a phase shift in the MMx interferometer is being properly interpreted as constant c. Unfortunately, I cannot read the entire article since there are some pdf errors that cause several of the pages to appear blank - is there another source?


  4. Mar 12, 2004 #3

    The author says that his viewpoint and that of SR are mathematically equivalent. If so, his theory and SR always make the same predictions.

    Yogi, the paper loaded fine in my browser. Perhaps you ought to update your reader, or try again.
  5. Mar 12, 2004 #4


    User Avatar
    Science Advisor

    You may try this:


    It is correct that the author says also that both theories are mathematically equivalent. But note that the main point is the claim that the MM experiment does not give any information about time differences and that, therefore, the classical conclusion about constancy of speed of light can not be rationally deduced from this experiment.

  6. Mar 12, 2004 #5
    Well, theory and evidence are not the same. There is just one body of evidence (including not-yet-uncovered evidence), ut there can be an infinity of theories, most of which are easy but wrong as H.L. Mencken might say here.

    To be perhaps overly blunt[b(] , I don't care about different theories if they give the same predictions always. (Exceptions: one is easier to use or think in. Or it makes more philosophical sense. I'm too old to shift to a new philosophical view :D
  7. Mar 12, 2004 #6


    User Avatar
    Science Advisor

    I understand what you mean, but let me insist: this is not the point which I try to bring into discussion.

    I am not interested (now) in questioning the experimental grounds of special relativity and I have also no interest (now) in the alternative theory. Also, it is a marginal matter the fact that both theories may be mathematically equivalent.

    My point is: I just try to understand whether the MM experiment is meaningfull or not. This experiment is described in every textbook as a milestone in modern physics. According to this paper the experiment is senseless due to a quite simple reason. It is hard to belive for me that this can be true, but I am not able to find the errors in the paper.

  8. Mar 12, 2004 #7


    User Avatar
    Staff Emeritus
    Gold Member

    I read most of the paper. I think it has various errors. I wouldn't take it seriously.

    It makes reference to momenta of photons defined as mv (which is already iffy), not noticing that they are massless (and he even uses those momenta in a quotient!). He also tries to start with the Doppler effect (which is included as a third postulate) to explain an effect that may be regarded as a Doppler effect. I don't think there is much to this article.
  9. Mar 13, 2004 #8
    From the standpoint of getting the same answers to already performed experiments, there are several theories that are equivalent - but if the ultimate goal is to model reality, the issue is much more than academic. It would be a devasting blow to the venerable status accorded SR for many years only to find that the experiment upon which it was motivated ...was flawed [even though Einstein dodged the issue of whether he taylored SR to confirm to MMx, evidence has emerged over the years that he was well aware of the experiments and was guided by the null result (actually a nearly null result)]
  10. Mar 13, 2004 #9
    Ok - thanks for the tip - downloaded a new acrobat reader and finally have a complete copy of the article. So if phase is preserved when the receiver and source are comoving, what experiments can be relied upon to validate the constancy of light velocity relative to the inertial observer - what about GPS - do these methods actually measure time directly or is the time derived from phase related data??
  11. Mar 13, 2004 #10


    User Avatar
    Science Advisor

    Could you tell me please whether there are errors in section 2.d. where the number of periods for both paths is calculated?

  12. Mar 15, 2004 #11
    I didn't see any reference to photon momemtum being mv, but rather,
    p = mc.

    Although I haven't formulated a strict forbiddence of such a proceedure, this is exactly what one arrives at if you adopt the termonology of 'relativistic' mass (or 'relativistic' energy) in your derivations.
    Since it has become common proceedure these days, the term p = mc, as ficticious as it seems, has become commonplace.

    Beginning with the relativistic eqn. for total Energy E in terms of momentum, p, and rest mass, m(0):

    E = Sq rt[p^2*c^2 + m(0)^2*c^4]

    Sustituting E = mc^2 on the left and dividing thru by c^2 gives:

    m = sq rt[(p/c)^2 + m(0)^2]
    (m is usually referred to as relativistic mass).

    For photons having no rest mass the second term drops out, leaving:

    m = p/c

    Thus p = mc

    Thus due to their momentum, photons can be said (at least by current physics illiteracy) to have a somewhat effective 'relativistic' mass.

    I don't know how his proof would fair if forced to use another definition of photon momentum.

    Nevertheless, I really like this fellow's thinking and analysis on this subject. However, it seems somewhat inconsistent to use a 'relativistic' mass eqn. a'priori in order to analyse a proceedure which was apparently the basis for the devlopment of such an equation. Maybe I'm wrong and the relativistic equations could have been developed without the assistence of Mich & Mo's experiment.

  13. Mar 16, 2004 #12


    User Avatar

    Staff: Mentor

    His closing paragraph makes a sweeping statement that SR is incorrect. His ideas, in fact, are strikingly similar to David's. He also states that his theory is mathematically equivalent to SR. To do that though, you need to reject time dilation, which he does. The problem is that time dilation is not so easily dismissed.

    Since time dilation is observed to occur exactly the same way in every physical process where the measurement precision to be able to observe it exists, it is illogical to conclude anything but that it is time itself dilating. To conclude anything else requries accepting a massive, coincidence with no exceptions as well as assuming the existence of an absolute time which is inherrently unobservable. Science does not allow such things. A theory that requires less unprovable assumptions is always preferable to one that requires more.
    Last edited: Mar 16, 2004
  14. Mar 16, 2004 #13


    User Avatar

    Which in light of previous discussions here should actually read
    p = E_{R}/c.

    Should be
    [tex]E_{R} = \sqrt{p^{2}c^{2} + m^{2}c^{4}}[/tex]

    Should read simply
    For photons having no mass the second term on the right drops out, leaving:
    E_{R} = pc.

    The term "relativistic mass" is a misnomer and has no place in modern relativity. see-
    Last edited: Jul 14, 2004
  15. Mar 16, 2004 #14
    Isn't the bottom line here based on the notion that - since there is no doppler shift (i.e., no phase displacement for one way measurements) when the source and receiver are co-moving in the same frame, irrespective of the velocity of the frame relative to the medium, MMx will always yield a null result i.e., MMx is not the proper experiment for measuring velocity with respect to the medium.

    That is the way I read his conclusion - if that is in fact so, the development of the over and back equations based upon the transgressing postulates is redundant.

    The author seems to stumble on the issue of time dilation (treating it as an atomic transition delay ... leading to the slowing of atomic clocks. But that doesn't cover all the experiments .. sort of smacks at that point of Lorentz ether theory where time dilation is relagated to a physical process and dimensional change

    Perhaps Einstein got the right result for the wrong reason.
  16. Mar 18, 2004 #15
    OK mentors and other experts - assuming we have a sound wave in the air generated by a train whistle on the engine - will the guy in the cabose be able to determine changes in the speed of the train relative to the atmosphere by measuring phase shift?
  17. Mar 19, 2004 #16
    Of course not. And that is a pretty good analogy, Yogi. I would think since the engine whistle and cabose receiver both change speeds equivilently, there would be no doppler change irrespective of the relative atmospheric speed.

    However, to make the analogy complete you would need to travel at constant speed and bounce the engine whistle sound off the cabose and back to the engine while simultaneously sending the sound out perpendicular to the engine motion and back, AND compare the phase shift of THOSE TWO sounds. Would those in the engine hear a difference in the phase of the two sounds??

  18. Mar 20, 2004 #17
    Creator - I appreciate your comments - my question was framed as an attempt to resolve the more complicated question of MMx and other experiments that relied upon null phase shift. It doesn't seem possible that an entire century of physicists could be mislead by experiments that were incapable of measuring the very thing they were designed to detect. The sonic wave analogy would appear dispositive of the issue if it is valid - that is, by reducing the problem to one of sonics, we can actually perform the experiments with not too precise equipment. Moreover, the question as posed deals with the one way velocity - if there is no one way time dilation, there should not be a two way dilation (one way dilation is a first order effect, whereas two way dilation is a second order effect (v/c)^2 ... its the same old question of what does the traveling twin's clock read when he reaches the turn around point). So by breaking the problem down into two segments (a sort of temporal reductionism comprising two separate measurements - one for the outbound wave and one for the return wave) we should be able to posit an answer as to what the more elaborate optical experiments really disclosed.

    Incidentally, this article is not the first to suggest that phase shift experiments are incapable of measuring motion relative to a medium
  19. Mar 21, 2004 #18
    I understand exactly what you had in mind, Yogi.

    Really? What about the concept of geo-centrism. (I can think of many other examples.)
    History has shown that Truth is not held captive by how many fail to recognize it for whatever length of time.

    I wouldn't allow such a prejudice to prevent an objective re-analysis of the DESIGN of the experiment. Surely, it is by encouraging such objective questioning of the commonly accepted (and the associated admission of error) that allows scientific progression.

    In my comments I was merely agreeing with your sonic ANALOGY only as a heuristic means of gaining insight with respect to how it is possible that null phase shift can result even though there is motion through a real medium, without detection of the medium.

    However, I think it is foolish to think that one can dicifer the real effects of M & M experiment simply by 'performing' such a sonic experiment. The two areas, electromagnetic wave propogation & sonic propogation, are two quite independent regimes governed by different mechanics and it would be foolish to think performing an experiment in one area, no matter how simple, will tranfer completely to the other.
    Its not a question of precision, but one of physical principles.

    Again I totally disagree. In decifering experimental physical phenomena, 'reductionism' is far from having 100% validity or reliablity as a method of investigation, especially when trying to switch to different physical regimes!

    Niavely 'reducing' an experiment cannot always be trusted to resolve the issue since many experimental results are far greater than the individual parts. Just try experimental 'reductionism' on a laser and see how far you get. Or...theoretical reductionism in condensed matter, taking into account only the individual particles. How far would you get?

    Last edited: Mar 21, 2004
  20. Mar 21, 2004 #19
    Creator - again your comments are quite informative - I would agree to the extent you
    have interpreted my post. I guess what I was angleing at is the ether analogy - specifically, there is no doubt about a medium for sonics - and the failure to detect the phase shift of the train whistle is concluded as being non-informative as a test for one-way velocity with respect to the medium. I also have big doubts about reductionism in general - in this case however - the reductionism involves breaking the experiment into two separate "one way" experiments. Each is, in this case, more significant,than the over and back combination - for example, one can attach a microphone and amplifier to the cabose and broadcast the received signals back to a receiver on the engine - each thus acts as a separate source - and again if there is no detectable phase shift from front to back, there should not be one from back to front, ergo there should not be one from front to back plus back to front. There is no need to make experiments for orthogonal sonic paths perpendicular to the direction of the train - they are the nominal reference phase required in MMx against which the second order over and back phase phase shift could be measured. In the case of sound a microphone pick up and an oscilloscope with a calibrated sweep would be adequate to detect the phase change.

    Of course - the extension of the result to invalidate MMx does not necessarily follow - the sonic experiment is but a first step -
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?

Similar Discussions: Phase shift in the Michelson-Morley experiment