Michelson–Morley experiment: Did it disprove the existence of ether?

[thwle]

The usual interpretation of MMX is biased both by the intended purpose addressing the concept of a luminiferous ether and by the subsequently enunciated special theory of relativity.

I wish contributors to this thread would notice my post, quoted above.

 

[harrylin]

In his 1905 paper, he stated, in effect, that it was apparently self-contradictory, but that it wasn't really.

He stated that the two postulates are only apparently contradictory; and that is obviously correct. That has little to do with your interpretation.

Regards,
Harald

 

[harrylin]

I wish contributors to this thread would notice my post, quoted above.

Actually, the intended purpose was to detect the velocity of the Earth relative to the ether. That's clearly explained in the MMX paper to which I provided the link. Do you think that that basic fact is ignored?

 

[ibreakkidsleg]

No! That is a much different question than your first question - and a poorly phrased one at that. One can never prove anything absolutely, particularly for all situations, so it is nonsensical for that to be the conclusion of an experiment. But the MMX did prove (to a reasonable certainty) that light is always perceived to travel at C...in an MMX.

If you mean that, even in a different inertial frame, the speed of light is measured to be c in an MMX, then you are in error. This is because a corpuscular theory of light is consistent with MMX's.

 

[thwle]

Thanks for responding.
Yes, MMX was intended to provide evidence to help decide between emission theory and ether theory (referred to as undulatory theory in the paper). I am not conversant with emission theory, of which Ritz theory is one variant, but I believe these are pretty much out of favor these days.
Trying to reconcile MMX with ether theory many thought it necissary to conceive of ether as fluid ("ether drag") so these day ether theory is also out of favor because it became untenably complex.
My intention was to point out that MMX is simply explained without complexity and in a way that vitiates its value for the intended purpose -- it can be seen as compatible with both emission theory and ether theory, (though for "ether theory" I prefer to say "the existence of an absolute frame of reference"). It also does not support or refute relativity.

 

[ghwellsjr]

In his 1905 paper, he stated, in effect, that it was apparently self-contradictory, but that it wasn't really.

He stated that the two postulates are only apparently contradictory; and that is obviously correct. That has little to do with your interpretation.

Regards,
Harald

What was it about the two postulates that made them only apparently contradictory?

 

[grav-universe]

If you mean that, even in a different inertial frame, the speed of light is measured to be c in an MMX, then you are in error. This is because a corpuscular theory of light is consistent with MMX's.

Yes, but not wave theory. In order to obtain a null result from the MMX with corpuscular theory, light pulses would have to travel with the added motion of the source. However, experiments with the aberration of light pulses tell us the pulses are observed to travel at the same speed in a particular direction regardless of the motion of the source. The combination of MMX type experiments and experiments with aberration show that light travels as both a wave and a particle, and only Relativity relates them properly.

 

[thwle]

Folks, please check your assumptions. Some posts in this thread assert that MMX measured or compared light speed. It did not because it could not -- there was no calibrated distance measure and no calibrated time measure hence no possibility of measuring speed. The closest thing to a clock in the experiment is the light source which has a frequency, to be sure, but that frequency was not used as a time measure. It wasn't even assumed to be constant -- it didn't need to be.
The light beam was split so the frequency was the same for both paths.
The interference pattern did not shift, so the RATIO of the times for the two paths was constant.

The ONLY thing MMX demonstrates is the correspondence of two methods of length measurement (radar ranging and solid measuring stick) through changes of orientation and velocity in an inertial frame of reference.

 

[ghwellsjr]

Some posts in this thread assert that MMX measured or compared light speed. It did not because it could not -- there was no calibrated distance measure and no calibrated time measure hence no possibility of measuring speed.

MMX was designed to measure a speed, the speed of the surface of the Earth relative to the presumed absolute ether rest frame as a fraction of the speed of light. But since the designers of the experiment did not understand that Maxwell's equations are invariant under the Lorentz Transformation, there was no hope of getting anything but a null result.

 

[harrylin]

Thanks for responding.
Yes, MMX was intended to provide evidence to help decide between emission theory and ether theory (referred to as undulatory theory in the paper). I am not conversant with emission theory, of which Ritz theory is one variant, but I believe these are pretty much out of favor these days.

?? Are you sure you read the paper? They start by explaining how emission theory had been disproved. Here it is again, see the first paragraph:

http://en.wikisource.org/wiki/On_the_Relative_Motion_of_the_Earth_and_the_Luminiferous_Ether

Emission theory was pretty much out of favour then. Their aim was to measure the velocity of the Earth.

Trying to reconcile MMX with ether theory many thought it necissary to conceive of ether as fluid ("ether drag") so these day ether theory is also out of favor because it became untenably complex.

The fluidic dragged ether theory was also already disproved, in part by Michelson's earlier experiments - see paragraph 2! You can read his Fizeau experiment here:

http://en.wikisource.org/wiki/Influence_of_Motion_of_the_Medium_on_the_Velocity_of_Light

My intention was to point out that MMX is simply explained without complexity and in a way that vitiates its value for the intended purpose -- it can be seen as compatible with both emission theory and ether theory, (though for "ether theory" I prefer to say "the existence of an absolute frame of reference"). It also does not support or refute relativity.

Indeed no single experiment addresses all theories. The MMX is part of a package that started with Michelson's Fizeau experiment. For completeness he performed another experiment to demonstrate detection of the rotation of the Earth:

http://adsabs.harvard.edu/abs/1925ApJ...61..140M

Regards,
Harald

 

[harrylin]

[..]
The closest thing to a clock in the experiment is the light source which has a frequency, to be sure, but that frequency was not used as a time measure. It wasn't even assumed to be constant -- it didn't need to be.
The light beam was split so the frequency was the same for both paths. [..]

Actually the frequency did have to be constant: it was used as a time measure. The frequency directly affects the phase shift from which they hoped to measure the velocity of the earth. Based on their assumptions they concluded that "the relative velocity of the Earth and the ether is probably less than one sixth the Earth's orbital velocity, and certainly less than one-fourth".

 

[ibreakkidsleg]

Yes, but not wave theory. In order to obtain a null result from the MMX with corpuscular theory, light pulses would have to travel with the added motion of the source. However, experiments with the aberration of light pulses tell us the pulses are observed to travel at the same speed in a particular direction regardless of the motion of the source. The combination of MMX type experiments and experiments with aberration show that light travels as both a wave and a particle, and only Relativity relates them properly.

Right. MMX is supposed to eliminate preferred-reference-frame theories (ether), and other experiments are supposed to eliminate source-velocity-dependence theories. Although to conclude by process of elimination that ``theory XYZ of light`` is true, one sort of sweeps the problem of unconceived alternatives under the rug.

 

[Buckleymanor]

Folks, please check your assumptions. Some posts in this thread assert that MMX measured or compared light speed. It did not because it could not -- there was no calibrated distance measure and no calibrated time measure hence no possibility of measuring speed. The closest thing to a clock in the experiment is the light source which has a frequency, to be sure, but that frequency was not used as a time measure. It wasn't even assumed to be constant -- it didn't need to be.
The light beam was split so the frequency was the same for both paths.

It did compare light speed as you pointed out.

The light beam was split so the frequency was the same for both paths.
The interference pattern did not shift, so the RATIO of the times for the two paths was constant.

The clock was the frequency and any shift would show that one path of the light took a longer or shorter time to travell a given distance(one of the paths).
So the assumption was that if both beams of light arrived back to the observer in phase the speed was constant throughout the journey that the light took along the two paths.

 

[Dale]

Without assuming Einstein's two postulates you can determine that the metric in a frame using the Einstein synchronization convention in a small region of flat spacetime is of the form:

ds² = -g0²c²dt² + g1²dx² + g2²(dy² + dz²)

Then rather than assuming Einstein's postulates you can simply determine the functions g0, g1, and g2 by experiment. The MMX fixes g1=g2. In other words, the speed of light is isotropic.

See: http://authors.library.caltech.edu/11476/1/ROBrmp49.pdf

 

[roineust]

Hey Dalespam,

The link you attached above, to a 1949 paper, is very interesting in my opinion (the textual part of that paper i mean).

Question: Why at the end of page 378, and at the beginning of page 379, the writer claims: " ...The fundamental measurement of one kind of interval is not to be reduced to that of the other with the aid of postulated constancy of the velocity of light, as would for example..." - Without naming the reason for that restriction...?

Thanks,
Roi.

 

[thwle]

MMX was designed to measure a speed, the speed of the surface of the Earth relative to the presumed absolute ether rest frame as a fraction of the speed of light. .

MMX was intended as you say. But it did not because of a faulty assumption.

But since the designers of the experiment did not understand that Maxwell's equations are invariant under the Lorentz Transformation, there was no hope of getting anything but a null result.

And we believe "that Maxwell's equations are invariant under the Lorentz Transformation" because...? (Is it not an interpretation of MMX itself?)

 

[ghwellsjr]

Without assuming Einstein's two postulates you can determine that the metric in a frame using the Einstein synchronization convention in a small region of flat spacetime is of the form:

ds² = -g0²c²dt² + g1²dx² + g2²(dy² + dz²)

Then rather than assuming Einstein's postulates you can simply determine the functions g0, g1, and g2 by experiment. The MMX fixes g1=g2. In other words, the speed of light is isotropic.

See: http://authors.library.caltech.edu/11476/1/ROBrmp49.pdf

Is this paper concerned with both of Einstein's postulates or just the first one?

The Introduction refers to the first postulate (principle of relativity) as a general postulate and the second postulate (the constancy of the one-way speed of light) as a specific postulate.

On page 380, right-hand column, is the following:

"...Einstein's synchronization insures as a matter of definition the equality of the forward and backward velocity along any given line..." (emphasis in the original).

And then later:

"Alternate synchronizations could have been agreed upon...But...they cannot in practice be carried out as they involve a non-operational appeal to the hypothetical rest-system..."

And right after that they summarize the result of MMX:

"M-M: The total time required for light to traverse, in free space, a distance l and to return is independent of its direction."

And finally, in the Conclusion:

"We have with this completed the task of replacing, so far as possible, Einstein's relativity postulate by facts drawn from experience."

So shouldn't you restrict your final statement to "the round-trip speed of light is isotropic"?

 

[thwle]

?? Are you sure you read the paper? They start by explaining how emission theory had been disproved. ... Emission theory was pretty much out of favour then. Their aim was to measure the velocity of the Earth. ...
Regards,
Harald

I referred to the paper as I wrote my post.

Yes, emission theory was out of favor, but not disproved. They expected to demonstrate further against emission theory by detecting the velocity of the Earth, but that depended on assumptions that light traveled at a fixed speed relative to a frame of reference (ether) and (unstated) that measuring stick length is not equivalent to radar ranging (not by that name, of course.)

They failed to detect motion of the earth. It seems to me that shows either that emission theory had some truth to it or that there is a greater correspondence of the two methods of length measurement than they had supposed.

 

[ghwellsjr]

MMX was intended as you say. But it did not because of a faulty assumption.

And what was that faulty assumption?

And we believe "that Maxwell's equations are invariant under the Lorentz Transformation" because...? (Is it not an interpretation of MMX itself?)

No, it has nothing to do with any experiment, it is pure mathematical manipulation. Maxwell's equations are, well, a set of equations. Lorentz's Transformation are a set of equations. When you do the operation of Lorentz's Transformation on Maxwell's equations, the question is, do Maxwell's equations come out the same? And yes they do.

Now if you are asking if the Lorentz Transformation is an interpretation of MMX, it was derived from that and several other experiments, but it was also derived by Einstein purely on theoretical grounds based only on his two postulates.

My point is that if Maxwell had the foresight of Einstein's hindsight, he would not have proposed an experiment to measure the speed of the surface of the Earth through the absolute stationary ether because he would have realized that his equations predicted that it was unmeasureable. But I'm not complaining, Maxwell was the most brilliant scientist of the 19th century in my opinion and hindsight is always better than foresight.

 

[thwle]

Actually the frequency did have to be constant: it was used as a time measure. The frequency directly affects the phase shift from which they hoped to measure the velocity of the earth. Based on their assumptions they concluded that "the relative velocity of the Earth and the ether is probably less than one sixth the Earth's orbital velocity, and certainly less than one-fourth".

OK. If there had been a phase shift, a time difference could have been inferred. There was not and would not have been even if frequency changed.

 

[Vanadium 50]

And we believe "that Maxwell's equations are invariant under the Lorentz Transformation" because...?

Because when you apply the Lorentz transformation to Maxwell's equations you recover Maxwell's equations. After many pages of algebra.

 

[thwle]

It did compare light speed as you pointed out.

The clock was the frequency and any shift would show that one path of the light took a longer or shorter time to travell a given distance(one of the paths).
So the assumption was that if both beams of light arrived back to the observer in phase the speed was constant throughout the journey that the light took along the two paths.

Beg pardon! There is a difference between comparing light phase and comparing light speed.

There was neither comparison nor measure of light speed in MMX.

 

[thwle]

Lorentz's Transformation are a set of equations. When you do the operation of Lorentz's Transformation on Maxwell's equations, the question is, do Maxwell's equations come out the same? And yes they do.

Because when you apply the Lorentz transformation to Maxwell's equations you recover Maxwell's equations. After many pages of algebra.

Oops. I defer on that. I read "Lorentz transformation" and reacted without understanding what you said. (my bad.)

 

[thwle]

And what was that faulty assumption?
QUOTE]


The faulty assumption specifically: that length measure by measuring stick and length measure by electromagnetic echo ranging are not equivalent.

MMX does not show that they are equivalent but seems to suggest that they may be.

I would like you to consider the possibility that the spacing of atoms in a solid measuring stick may be determined by forces that are communicated across those spaces at light speed, in which case measuring stick length becomes dependant on echo ranging.

 

[ghwellsjr]

And what was that faulty assumption?

The faulty assumption specifically: that length measure by measuring stick and length measure by electromagnetic echo ranging are not equivalent.

MMX does not show that they are equivalent but seems to suggest that they may be.

I would like you to consider the possibility that the spacing of atoms in a solid measuring stick may be determined by forces that are communicated across those spaces at light speed, in which case measuring stick length becomes dependant on echo ranging.

Where did you get the idea that anybody made the assumption that any two methods for measuring distance would yield different results?

 

[thwle]

Where did you get the idea that anybody made the assumption that any two methods for measuring distance would yield different results?

That IS the MMX.

 

[ghwellsjr]

That IS the MMX.

I suppose your linked paper makes that claim? I tried to understand that paper but gave up. Can you point to a specific quote to support the idea that anybody believed that two different methods to measure a distance would produce different results?

 

[Buckleymanor]

Beg pardon! There is a difference between comparing light phase and comparing light speed.

There was neither comparison nor measure of light speed in MMX.

As you allready stated.

OK. If there had been a phase shift, a time difference could have been inferred.

I can't see why the assumption from this can't be drawn,(there was no phase shift) that the speed of light at the observation point is the same, irrespective of it's speed.
Though you go on to state.

There was not and would not have been even if frequency changed.

Which I am not quite sure about what you are saying.

 

[Dale]

Question: Why at the end of page 378, and at the beginning of page 379, the writer claims: " ...The fundamental measurement of one kind of interval is not to be reduced to that of the other with the aid of postulated constancy of the velocity of light, as would for example..." - Without naming the reason for that restriction...?

He actually did mention the reason for the restriction, but it was rather subtle. In the introduction Robertson said that he wanted to determine "the degree to which postulate can now be replaced by observation". So he could not use light-clocks since those rely on the second postulate.

 

[Dale]

Is this paper concerned with both of Einstein's postulates or just the first one?

Both. By using the form of the metric ds² = -g0²c²dt² + g1²dx² + g2²(dy² + dz²) Robertson is not assuming either postulate.

Instead he is making a general theory without those postulates and then simply looking to experiments to set the undetermined parameters of the metric (or equivalently the coordinate transformation).

 

[TrickyDicky]

If the interpretation of Lorentz of the MMX were correct (that the experiment frame was moving wrt the ether absolute rest frame and it was experiencing length contraction and time dilation), it would indeed prevent any experiment from succeding measuring a different speed of light than c regardless the motion of the light source (that is actually the case, to this day no experiment has been able to falsify this claim of special relativity). In this sense Einstein and Lorentz interpretation would reach the same predictions and by Ocam's razor the ether would be superfluous.

But could another approach be used to solve this, since measuring the speed of light seems a dead end, perhaps measuring the doppler shift of the Earth with respect to the vacuum?

 

[thwle]

As you allready stated.

I can't see why the assumption from this that the speed of light at the observation point is the same irrespective of it's speed.
Though you go on to state.

Which I am not quite sure about what you are saying.

What! ?

Let me try again: MMX consists of
(1) a rotatable stone slab on which two orthogonal folded paths are laid out, (two measuring sticks.)
(2) a light source ("argand lamp")
(3) a half silvered mirror at the common endpoint of the two paths to split and recombine the light beam at the beginning and end, respectively, of the round trip along each path.
(4) a telescope to observe the interference pattern in the recombined beam.

(2), (3) and (4) constitute a radar ranging measurement of the measuring sticks, except that rather than a count of wavelengths along the two paths only a count of differences between the two paths is observed. That being zero. It appears that the two methods of length measurment may be equivalent.

I conjecture that the dimensions of a material measuring stick, being dependent on the spacing of its atomic constituents, is dependent upon the forces between those constituents, which (if the forces be propagated or communicated between atoms at light speed) would always establish spacing determined (in effect) by radar ranging.

I hope this clears things up.

thwle

 

[lugita15]

My point is that if Maxwell had the foresight of Einstein's hindsight, he would not have proposed an experiment to measure the speed of the surface of the Earth through the absolute stationary ether because he would have realized that his equations predicted that it was unmeasureable.

I would disagree with that. I don't think Maxwell would have considered it significant that there exists a set of transformations which leave his equations unchanged. In fact, the whole reason he proposed experiments to measure the motion of the aether is because he believed that the aether frame was the only frame in which his equations were exactly correct.

 

[ghwellsjr]

Both. By using the form of the metric ds² = -g0²c²dt² + g1²dx² + g2²(dy² + dz²) Robertson is not assuming either postulate.

Instead he is making a general theory without those postulates and then simply looking to experiments to set the undetermined parameters of the metric (or equivalently the coordinate transformation).

Did I misunderstand the purpose of the paper? I thought it was to change the nature of the first postulate so that it was backed up by experimental evidence and no longer just an unsupported postulate.

 

[ghwellsjr]

That IS the MMX.

I suppose your linked paper makes that claim? I tried to understand that paper but gave up. Can you point to a specific quote to support the idea that anybody believed that two different methods to measure a distance would produce different results?

I must appologize here. You did not provide a link to a paper. I did a search on "emmision theory" earlier and found a paper that I thought you had linked to. Sorry.

But I'm still wondering where you got the idea that anybody at the time of MMX thought that two different ways of measuring a distance would give two different results. Can you provide some evidence for that?

Here is a post of yours:

They expected to demonstrate further against emission theory by detecting the velocity of the Earth, but that depended on assumptions that light traveled at a fixed speed relative to a frame of reference (ether) and (unstated) that measuring stick length is not equivalent to radar ranging (not by that name, of course.)

Here you say that they did not state that they believed that the two methods of measuring length were not equivalent, so if they didn't state it, why do you think they believed it?

And then there's this post:

MMX consists of
(1) a rotatable stone slab on which two orthogonal folded paths are laid out, (two measuring sticks.)
(2) a light source ("argand lamp")
(3) a half silvered mirror at the common endpoint of the two paths to split and recombine the light beam at the beginning and end, respectively, of the round trip along each path.
(4) a telescope to observe the interference pattern in the recombined beam.

(2), (3) and (4) constitute a radar ranging measurement of the measuring sticks, except that rather than a count of wavelengths along the two paths only a count of differences between the two paths is observed. That being zero. It appears that the two methods of length measurment may be equivalent.

What non-null result do you think Michelson and Morley expected?