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

A.G. Kelly and his criticism of relativity

  1. Jan 23, 2010 #1

    This may have been discussed, but A.G. Kelly has criticised the Hafele and Keating experiment on the basis of the drift errors associated with the atomic clocks.

    http://www.biochem.szote.u-szeged.hu/astrojan/hafele.htm#figure1 [Broken]

    I know that relativity has been well tested, but did Dr Kelly have have a fair point? It's hard to tell from his site as when it comes to the crunch, he does not describe at all well how H&K analysed their results. A brief investigation would suggest that H&K knew the limitations of their experiment, and acknowledged this in their paper. I also understand that their experiment was not supposed to provide an exact prediction of numbers but demonstrate a more subtle point regarding inertial frames and relativity; viz One plane clock ran fast, one plane clock ran slow, because the relative motion is measured for all three clocks (2 plane, 1 ground) from respect to an inertial frame moving relative to the Earth.

    While it is clear that Dr Kelly has an agenda with his anti-relativity site, I'd like to clarify the precision and conclusions of the H&K experiment from an objective view point, and whether their conclusions were correct. So rather than dismiss this as talking about crack pottery, could we see it as a discussion about the H&K experiment to inform my understanding.


    Last edited by a moderator: May 4, 2017
  2. jcsd
  3. Jan 23, 2010 #2


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Well, no experiment is exact in the sense of having perfect precision, with zero random errors. But the Hafele-Keating experiment was one in which they made a quantitative prediction and verified it to within the error bars of their experiment (which were about 10%). I don't think it's quite right to say that it was a test of inertial frames. The motion of the passenger jets was noninertial. It was a test of time dilation in relativity (gravitational and kinematic).

    The following is from a FAQ I maintain at http://www.lightandmatter.com/cgi-bin/meki?physics/faq .

    Was the famous Hafele-Keating experiment bogus?

    No. Certain internet kooks, including someone named A.G. Kelly, have produced reanalyses of the Hafele-Keating data in an attempt to disprove relativity. This is just silly, because the experiment was reproduced four years later to better precision, and again, to much higher precision, in a 25th-anniversary reenactment. The GPS system depends on general relativity, so any time you use a GPS receiver, you're reproducing relativistic time dilations of the type seen by Hafele and Keating.

    Hafele and Keating, "Around the world atomic clocks:predicted relativistic time gains," Science 177 (1972) 166.

    Hafele and Keating, "Around the world atomic clocks:eek:bserved relativistic time gains". Science 177 (1972) 168.

    Ashby, "Relativity in the Global Positioning System," http://www.livingreviews.org/lrr-2003-1
  4. Jan 23, 2010 #3
    Thank you for the reply, I'll take a look at the FAQs link you sent. Just a few points

    • I'm aware of the GPS prediction, which I'd point to immediately with regard to relativity.

    • When I said exact, I should have said high precision (sorry, not tight in my definition there).

    • My inertial frame point was not explained well - Referring only to kinematic time dilation in the next para.

      I agree, the aircraft are not in inertial frames, neither is the clock on the Earth's surface. To begin with, one would think that both plane clocks should run slow as they are moving with respect to the Earth clock (if one was sat with the Earth clock). This is a common trap when introduced to H&K. However, one clock runs slow, one runs fast (referring to kinematic effect only). My understanding of this was that we need to think of the system in a more robust way to really understand the difference. Namely that if one observes all three clocks from an inertial frame that is moving at VE in an opposite sense to the Earth's rotational velocity (VE), then it explains why one clock on the plane runs slower than the Earth clock, and one runs faster than the Earth clock (this view describes the relative velocities of the planes and Earth clock in a more intuitive way, and a correct way). This was the power of the H&K experiment, it tested the heart of relativity and the concept of relative motion when 'making measurements'. Please excuse any tightness of definition.

    So my question in this respect was, that while H&K did not measure the predictions of relativity with high precision, they made an important step in showing the foundations of relativity were valid. Is this a correct 'interpretation' of the experiment.

    I am ready to sit corrected.

    Kind regards
    Last edited by a moderator: Jan 23, 2010
  5. Jan 23, 2010 #4


    User Avatar
    Science Advisor

    Looks good to me. And you are quite right about the issue with reference frames. The correct way to manage the frames is as you describe and as done in the original Hafele-Keating paper.

    Cheers -- sylas
  6. Jan 23, 2010 #5


    User Avatar
    Science Advisor

    Kelly's criticism is interesting. He says that Hafele-Keating corrected their data, and without this correction, their experiments would not have been consistent with the general relativistic predictions. As far as I can tell, this is true. So were the corrections warranted?

    Kelly gives a link to http://tycho.usno.navy.mil/ptti/1971/Vol%2003_17.pdf [Broken] which gives the details of the corrections. The full time change of each clock would be "general relativity + unintentional bumps", and they wish to isolate only the general relativity contribution. Hafele-Keating were able to compare the rates of the clocks when the had all returned to the same place, when the general relativity effect is zero. Any rate difference between clocks at the same place at the end of the experiment are therefore due to unintentional bumps. So without doubt, some correction of the data was needed before concluding that the data was not consistent with general relativity. Unfortunately, Hafele-Keating did not know when the unintentional bumps occurred, and tried out an assumption that each clock suffered only one bump at the mid-point of its trip in order to estimate the effect due to unintentional bumps. Apparently, this correction produced a result consistent with GR. I suppose the next question is: are their data consistent with all the effect being due to unintentional bumps and zero gravitational effect?
    Last edited by a moderator: May 4, 2017
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook