[amrit (OP)]

Farsight gravitational red shift shows light velocity c is constant. Only frequency changes. If c is not constant connection betwen SR and GR is broken.

This isn't true I'm afraid, amrit. The frequency of the light isn't relevant. Start with two synchronised light clocks, leave one up in space, take the other down to the planet for a while taking care to avoid radial length contraction, then take it back up to space. The two clocks are no longer synchronised. Yes, the local speed of light is always measured to be 299,792,458 metres per second, but those two different clock readings tell you that 299,792,458 metres per second up in space is not the same as 299,792,458 metres per second down on the planet. People say the difference is because of "time dilation", but you and I know that clocks clock up motion rather than "the flow of time". Hence we know that a light clock clocks up the motion of light. Those two different readings are crystal-clear evidence that down on the planet, the light goes slower.  

Farsight I do not buy that.
yours amrit

[graham.d]

Amrit, "proper time" has a specific definition to be just that of an observer comoving with the moving frame. Your link does not work properly by the way.

I appreciate English is not your first language, but however hard I try, I cannot understand the sense of what you are saying. Perhaps you could show me where my maths is at fault?

[Farsight]

We'll have to agree to differ, amrit. Sorry I couldn't help.

[graham.d]

Farsight, I understand Einstein's view on this but note he was uncomfortable with this view. And I take your point about the clocks being "flat" which I take to mean parallel to a circumference as opposed to in-line and parallel to a radius. I don't think you can conclude that this removes the difference that may occur between remote length measurements although I admit that the maths is sufficiently hard that I could not show it would produce the same quantitative result. For example you could imagine a distant observer receiving signals from the mirrors at each end of the apparatus which were set up to send beams outward exactly parallel, but because they are in a divergent field they will follow geodesics that would have them diverge such that when they reached the distant observer they would show a bigger gap between the mirrors. Now it is very hard to say that this would yield the same value as the simpler calculation for the mirrors in line, but is, nonetheless, possible. The maths to calculate this is a bit hard.

I will give this a little more thought. A tenet of GR is that the all observers should agree on the spacetime interval. I am not sure that it is necessarily possible to resolve the meaning in term of just space and just time in this case. As the modern view is, as you say, that lightspeed is constant, how would these physicist view this scenario?

[imatfaal]

A light pulse going around lots of coils of fibre optic might be a practical photon clock too. With enough coils the time delay can be significant and measurable with low errors.

3 coils at 90 degrees to each other would nicely (after serious number crunching) allow dilation from varying gravitational potential and dilation from relative velocities to be separated - I think...

[JP]

Actually, I'm not sure that there's any contradiction between the idea that the speed of light might not seem constant over large distances and the speed of light being constant locally in a vacuum.  From what I understand of it, you can think of it like traveling on the earth.  Say it takes me 4 hours to travel from Paris to Los Angeles.  I could calculate my average speed by taking the distance between them in 3D space (which is the length straight line through the globe divided by 4 hours) or I could measure my speed at every instant as I fly around the globe and average that.  The path through the globe will be far shorter than the actual speed I was traveling.  I think it's a similar case with GR and the speed of light.  If you look at lengthy paths and just try to calculate the speed of light based on what you'd expect if space-time were flat and the light moves at c, you find that it takes longer.  However, if you look at the light at every instant, it's moving at the speed c--it just takes a curved path, so the total time is longer.

Or put another way, if you look locally enough so that you're calculating instantaneous speed, space-time is flat and SR holds.  It's when you look over large enough regions of curved space-times that you get these deviations.  (I guess even if you're comparing two flat regions--which is what this discussion is about--measurements can differ because space-time can be stretched, but not curved.  Using the rulers and clocks of one region to measure the other will also cause the speed of light to appear to be slowed.)  In fact, I think you can go further and say that if the speed of light wasn't constant locally, the geometrical interpretation of general relativity would fail.  

Is there an equally valid model that allows the speed of light to vary locally and produces identical results (at least within the regimes we've tested)?  I don't know.  

[graham.d]

Paris to LA in 4 hours - I didn't think the Ramjet was flying yet!

It is interesting to think of two regions that are both essentially flat regions of Minkowski space where they have different degrees of Stretched Space-Time. There would seem to be no way of determining the degree of "Stretchedness" from inside such a region because all the physical laws would work fine in both regions. So if one region is open to investigation from the other (as in the case postulated previously) it seems odd that the observed behaviour would depend on the spacetime between the two regions. I think this is what you are saying JP. I could easily imagine that you could measure the temporal and spatial difference but it is hard to believe that these would be different depending on the route your information traveled to get to you.

I think that Farsight was alluding to the idea that the velocity of light being a tensor field (rather than a scalar field) in having different radial and tangential values in a schwartzchild metric. It still seems odd that the intervening space would affect what should be scalar measurements in flat space though. Any ideas? I may be missing some key points here.

[Farsight]

Farsight, I understand Einstein's view on this but note he was uncomfortable with this view.

It's an interesting one Graham. One day when the dust has settled somebody will write an authorative narrative on the long and tortuous trail, and it will be fascinating reading.

And I take your point about the clocks being "flat" which I take to mean parallel to a circumference as opposed to in-line and parallel to a radius.

Thanks, sorry if I didn't make it clear.   

I don't think you can conclude that this removes the difference that may occur between remote length measurements although I admit that the maths is sufficiently hard that I could not show it would produce the same quantitative result.

I think you can actually, very simply. If you go along with GR radial length contraction, and Einstein, and the Occam's razor direct evidence that says the light is moving slower rather than "time runs slower", you can see it. The metre is defined as "the distance travelled by light in a complete vacuum in 1⁄299,792,458th of a second". Hence the bigger second and the slower light cancel each other out. The metre is unchanged.   

For example you could imagine a distant observer receiving signals from the mirrors at each end of the apparatus which were set up to send beams outward exactly parallel, but because they are in a divergent field they will follow geodesics that would have them diverge such that when they reached the distant observer they would show a bigger gap between the mirrors. Now it is very hard to say that this would yield the same value as the simpler calculation for the mirrors in line, but is, nonetheless, possible. The maths to calculate this is a bit hard.

Sorry, I'm not keen on this because it seems to go against the grain of general relativity regardless of which interpretation one prefers. 

I will give this a little more thought. A tenet of GR is that the all observers should agree on the spacetime interval. I am not sure that it is necessarily possible to resolve the meaning in term of just space and just time in this case. As the modern view is, as you say, that lightspeed is constant, how would these physicist view this scenario?

I'm not sure graham. But I'd say that to fix the speed of light as an absolute constant, one has to elevate time to something above and beyond the experimental and observational evidence, and say "time runs slower". The trouble is, when you look at a clock, you don't see time running at all. All you see is cogs moving, or a crystal oscillating, or a caesium atom oscillating, or light moving. It always comes down to motion one way or another. Clocks clock up motion, that's all they ever do, that's why in The Foundation of the General Theory of relativity Einstein gave us the equations of motion. And when the clock runs slower, it isn't really time running slower, it's the motion going slower. Have a google on "Farsight" and "Time Explained" for my take on this.

[Farsight]

..I think that Farsight was alluding to the idea that the velocity of light being a tensor field (rather than a scalar field) in having different radial and tangential values in a schwartzchild metric.

I wasn't, I was just trying to keep things simple and focus on what we actually observe, which is two different readings from two identical parallel-mirror light-clocks at different locations where the gμv gravitational potential is different. 

It still seems odd that the intervening space would affect what should be scalar measurements in flat space though. Any ideas? I may be missing some key points here.

The intervening space isn't relevant. You could run the clocks for years and see the discrepancy grow ever larger. Redshift and signal time becomes increasingly inconsequential. Like you say, the spacetime is flat at both locations. There is no discernible curvature out in space or at the centre of the earth, because there is no detectable gμv gradient. The different measurements are because the space itself is different at those two locations.

[Geezer]

An oscillator produces microwave energy that causes the caesium to fluoresce. The detector measures the amount of fluorescence. Maximum fluorescence (photon emissions I suppose) is achieved when the microwave energy is tuned to 9.1xxx GHz. The clock makes very small adjustments to the oscillator to maintain maximum photon emission.

Noted Geezer, but remember that this is defining the second. Hertz is cycles per second, so we're finding the maximum then defining the frequency to be 9.192631770 GHz.

The clock is really comparing the natural oscillation of the caesium atom with the frequency of a microwave resonator and adjusting the resonator to match the frequency of the caesium. So, it's not really measuring a property of light at all.

The oscillation is a hyperfine transition, and electron spin-flip. The event is electromagnetic, as is the emitted light. If the gravitational potential is lower, all electromagnetic phenomena occur at a slower rate. We call it time dilation, but that rather misses the obvious fact that electromagnetic spin flips are happening slower and the emitted light is moving slower too.

Er, well the second is defined as the interval for a certain number of the atomic events. The number was chosen to line up with more traditional definitions for the second. As you point out, the fundamental "tick" is an electromagnetic event with the atom.

The light output could be delayed in a very variable fashion, but it would have no effect on the frequency of the microwave resonator. The light output is only sampled periodically to make very infrequent and very minute adjustments to the frequency of the microwave resonator. I believe the detector only responds to the number of photons that reach it. I'm not sure how they know to adjust up or down because I would assume that the light output will diminish if the resonator is slow or fast relative to the caesium events.

Being pedantic, you can't really say that the events are happening at a "slower rate", because that seems to suggest that time itself is invariant.

[graham.d]

Both spaces are flat Minkowski space and I think we are assuming that the physical laws are the same, as measured locally, at each location and that it would be expected that the gravitational potential is purely relative and has no meaning in absolute terms. Are these correct assumptions to work with? If, then, the only difference is that space is "stretched" more in one region than the other then this "scaling" must apply to (x,y,x,-ct) equally mustn't it? If not what coordinate transformation would apply? In SR the basis for the Lorentz transformation is the constancy of c to all observers. Is the only effect of gravitational potential (Φ) that c is a function of Φ and that this would be the basis of any coordinate transformation?

This seems reasonable but is it all consistent with observation or gedanken experiments?

[JP]

Paris to LA in 4 hours - I didn't think the Ramjet was flying yet!

Heh.  I'm riding in my photon jet.  It would be faster, but it slowed down a bit because it's so close to the earth.
 []

[Geezer]

Heh.  I'm riding in my photon jet.  It would be faster, but it slowed down a bit because it's so close to the earth.
 []

Yes, but I'm sure you'll arrive in no time at all.

[amrit (OP)]

Amrit, "proper time" has a specific definition to be just that of an observer comoving with the moving frame. Your link does not work properly by the way.

I appreciate English is not your first language, but however hard I try, I cannot understand the sense of what you are saying. Perhaps you could show me where my maths is at fault?

Graham with your math all is fine. Just be aware clocks tick in space only and not in time. Time dilatation means that clocks run slower in a timeless space. You think in math terms, think in physical terms and you will discover universe is timeless. In the universe there is only motion, time belongs to the mind. Time is a mind frame through which we experience motion, read ma article below:
http://www.vetrnica.net/index.php?option=com_content&view=article&id=21&Itemid=23
yours amrit

[Geezer]

In the universe there is only motion, time belongs to the mind. Time is a mind frame through which we experience motion, read ma article below:

Amrit,

I have tried to persuade the molecules that make up my body that time is only a state of mind, but they won't listen.

Isn't it high time we move this topic to New Theories?

[amrit (OP)]

NO. This is not new theory. This is about right understanding of SR. We live in space only. X4 is not time, X4 is spatial too.

PS
missing understanding is that in SR velocity of clocks in all inertial systems is invariant of the observers. Observer has no influence on velocity of clocks. Clocks run with the same velocity for all observers.

[Geezer]

Clocks run with the same velocity for all observers.

Yes. I completely agree with you there. As I see it, all things are affected by their local time. That means everything from subatomic events to fancy atomic clocks have to obey local time. Consequently, any observation of velocity (which can only be expressed in terms of time) will be the same for all observers.

[JP]

Clocks run with the same velocity for all observers.

Yes. I completely agree with you there. As I see it, all things are affected by their local time. That means everything from subatomic events to fancy atomic clocks have to obey local time. Consequently, any observation of velocity (which can only be expressed in terms of time) will be the same for all observers.

I'm going to disagree just to be difficult.  But seriously, clocks don't measure velocity.  The "velocity of a clock" is how fast it goes when you throw it.  Clocks measure time.  If clocks only measured velocity, they couldn't count time, since velocity only tells you the ratio of distance to time.  You need a value for distance and for velocity to get time.  That's how light clocks work: by knowing the distance between the mirrors and using the known value of the speed of light. 

That's also why this theory doesn't agree with special relativity.  The velocity of light might be constant for all clocks, but their distance measurements don't agree, so they can't possibly tick at the same speed.

[Geezer]

I'm going to disagree just to be difficult.  But seriously, clocks don't measure velocity.  The "velocity of a clock" is how fast it goes when you throw it.  Clocks measure time.

Far be it from me to disagree with someone who obviously knows what they are talking about, but clocks don't measure time. The only thing that clocks do is count events. The events may be determined by time, but I'm not sure that's quite the same thing as saying that clocks measure time.

Time measures clocks might be a better way to look at it.

[JP]

I'm going to disagree just to be difficult.  But seriously, clocks don't measure velocity.  The "velocity of a clock" is how fast it goes when you throw it.  Clocks measure time.

Far be it from me to disagree with someone who obviously knows what they are talking about, but clocks don't measure time. The only thing that clocks do is count events. The events may be determined by time, but I'm not sure that's quite the same thing as saying that clocks measure time.

Time measures clocks might be a better way to look at it.

This is somewhat speculative, since I'm not a relativity expert:

I guess maybe since we're stuck using the language of special relativity here, maybe the best any device can do is to measure the interval between two events?  (The interval being Δx²-Δt², where Δx is the spatial separation of the events and Δt is their temporal separation [with c=1]).  You can then get time by relating this interval to physical processes.  If your clock is using a cesium atom at rest, Δx=0, so your measurement of a "tick" is directly getting time out.  If your clock is a light clock, then you're measuring an interval which includes both a Δx and a Δt (I suspect you need to calculate two intervals: one in each direction, as graham.d did).  That's why you need to also use the fact that light speed is constant--which tells you that the interval between the events is zero, which in turn lets you calculate the time of a second from the distance between mirrors. 

---------------------------------
Now that I'm done confusing myself, this part is my real point:

The point I was trying to get across is that clocks spit out "time" as an answer, no matter what they're using as input.  Velocity isn't enough information on its own to give you time as an answer--you need to know a distance measurement as well, which is why you will still get time dilation in a photon clock even with a constant speed of light.

I guess a good question is whether or not a "good" clock (meaning one that measures local time accurately) is in practice any different than measuring time itself.  I don't know the answer, but I've been assuming that it isn't.