on 14/9/08 4:19 pm, Stephen A. Lawrence at [EMAIL PROTECTED] wrote: > > > Harry Veeder wrote: >> on 14/9/08 8:25 am, Stephen A. Lawrence at [EMAIL PROTECTED] wrote: >> >>> >>> Harry Veeder wrote: >>> >>>>> The most common approach to the problem was to postulate an aether which >>>>> carried the EM waves, and then try to patch things up so that Maxwell's >>>>> equations would still work. This approach had the large advantage that >>>>> it did *not* require reforming the common view of space and time -- >>>>> "aether" was a simple extension of a familiar concept, albeit with some >>>>> peculiar new properties. >>>> Since the aether is not identical with Newton's notion of absolute of >>>> space, >>>> the failure to detect an aether does not invalidate the notion of absolute >>>> space. The difference between an aether and absolute space is very apparent >>>> when light is concieved as a particle, although the preference for the wave >>>> theory of light by the latter half of the 19th century resulted in a >>>> tendency to disregard this important conceptual difference. >>>> >>>> Even without quantum theory, one could still argue that >>>> light is a particle and that Maxwell's equations simply provide a >>>> mathemtical *formalism* for predicting how light particles interact with >>>> matter. >>> Yes, for sure, aether theory is totally different from Newton's particle >>> theory. Aether theory was, if I understand this correctly, conceived to >>> explain the wave nature of light, which was known at that time -- >>> diffraction and interference are hard to explain with particles. >> >> It is hard to explain with particles possessing the property of inertia, >> which was considered mandatory for all particles to possess in the late 19th >> century. However, if inertia is some sort of electromagnetic effect rather >> than a fundamental property it becomes easier to see how such particles >> could produce wave-like effects. What I am suggesting is that classical >> electromagentic waves could be reimagined as being roughly analoguous to >> Debrogile's pilot waves. >> >> >>> QM >>> does it, but QM came much later, of course. >>> >>> >>>>> The trouble was that it's very hard to come up >>>>> with an aether theory in which Maxwell's equations are correct at all >>>>> speeds. If they're *not* correct at all speeds, then experiments should >>>>> show differences depending on the observer's speed. And experiment has >>>>> never turned up such a difference. >>>> I'm still waiting for a one-way light speed measurement. As far as I know >>>> all experiments todate use the absence of interference to infer the >>>> constancy of the speed of light over different frames of reference. >>> It's hard to come up with a one-way lightspeed measurement which doesn't >>> require TWLS=OWLS to get its result. The problem is that for OWLS you >>> need to have spatially separated synchronized clocks; how do you sync >>> them up to start with? >> >> You don't have to. > > I think you do; otherwise how can you know how long the pulse took to > arrive at the target? You know when it started, according to A; you > know when it arrived, according to B. But A and B have two different > clocks -- they must, because they're not in the same place, and one > clock can only be in one location at a time. > > You may argue that it's trivial to synchronize their clocks, but I don't > think you can really argue that it's completely unnecessary. > > >> The problem of synchronsization arises only when one >> assumes that special relativity is correct *before* the experiment is >> performed. A one-way test of absolute of motion should begin with the >> assumption that absolute time is correct. > > But even if it is correct, how do you set all clocks to the "absolute > standard"? What do you use for a time distribution network? > > Remember, you're trying to *measure* OWLS here, so you can't assume > anything about the reversibility of paths before the experiment. > > You can start with colocated synchronized clocks and then separate them, > but if you're not careful you'll find that you're *assuming* SR is > invalid before the experiment, which is just as bad as *assuming* it's > valid!
You only have to know that the pulses depart at the same time and travel the same distance and to note the difference (if any) in their arrival times. Synchronization of clocks is necessary only if you wanted to know the total travel time of each pulse. > >> Even if c is found to be constant >> then you might conclude that absolute space is invalid, but the result does >> necessarily invalidate absolute time. >> >>> The approach I'm aware of which should work is "slow transport" -- you >>> colocate the clocks and sync them, and then you move them *slowly* >>> apart, preferably moving them simultaneously in opposite directions. >>> This should work if you're either in an inertial frame (which is to say, >>> somewhere out in space, not in orbit around anything) or if you start on >>> the equator and move one clock directly north and the other directly south. >> >> I am thinking of something even more straightforward. >> Imagine you are standing x feet from a highway sign with a laser pointer. >> Your friend is in car moving past you at v feet/sec also with a laser >> pointer. >> At the moment he passes you, you both "fire" your laser at the sign. >> He aims for the left side of the sign and you aim for the right side. >> Will his laser reach the sign first or will they both arrive at the same >> time? Surely such fine time measurements are now technically possible. > > I'm not so sure. The difference -- if there is a difference -- is going > to be mighty small. And you have the problem of knowing *when* the > pulses were fired, according to a stationary clock located by the > highway sign -- somehow, you still need to synchronize the clock in the > car with the one on the ground. True, but for the sake of argument let us suppose the clock in the car and the clock on the ground can be situated as close together as we wish so that the problem of clock synchrozination is essentially eliminated. I'll adress the rest of your discussion in my next post. harry > >>> If either clock moves on a line which carries it "forward" or "backward" >>> relative to the Earth's rotation, then the Sagnac effect (which has been >>> experimentally verified) is going to cause trouble with the syncing, and >>> slow transport doesn't help much with that. >>> >>> >>>>> Ultimately, as you say, Einstein chose to chuck the common understanding >>>>> of space and time. Our intuition says that in order to have a wave, >>>>> someTHING must wave. Einstein chucked that overboard, which was a >>>>> significant change. And people have been objecting ever since. The >>>>> only reason special relativity is accepted is that its predictions agree >>>>> with experimental results. >>>>> >>>>> The bind most other theories got caught in was that they needed to agree >>>>> with the outcomes of both the Michelson-Morley experiment (with its null >>>>> result) and the Sagnac experiment (with its non-null result). The >>>>> former is inconsistent with most aether theories, and the latter is >>>>> inconsistent with emission theory. >>>> What is the emission theory? The particle theory of light? >>> "Emission theory" holds that light is particles, and travels at C >>> relative to the *emitter*; it takes the analogy of rifle bullets and >>> carries it to its conclusion. >> >> As I said above being a particle does not necessarily require that you >> posses inertia like a rifle bullet apparently posseses. >> >>> The two big problems with it are >>> >>> ** It conflicts with the Sagnac experiment >>> >>> ** When the emitter is in motion, emission theory predicts a >>> longitudinal /frequency/ redshift which is (nearly) identical to the >>> values predicted by SR, but it predicts *no* /wavelength/ redshift. >>> This results from the fact that the propagation velocity of the light >>> varies with the emitter's velocity, which negates the wavelength shift. >>> Since spectroscopes using diffraction gratings measure the wavelength, >>> rather than the frequency, it requires some rather inelegant hackery to >>> get the theory to produce answers which agree with observation here. >> >> Depending on your perspective one could describe special relativity as >> inelegant hackery which produces answers which argee with experiment. ;-) > > The patchup for emission theory is far worse: It consists of adding > "knobs" to the theory which can be tweaked arbitrarily to make it agree > with whatever result you come up with. > > The problem is that as long as the propagation velocity is equal to the > emitter's velocity *plus* C, there should be no wavelength redshift, and > spectroscopes tell us that there certainly is. So, somehow, we must get > the propagation velocity "adjusted" to match C by the time the receiver > gets the signal. > > To start with, this problem came up with regard to stellar redshifts > measured by Earth-based spectroscopes. The initial "fix" was made by > assuming that the signal took on the velocity of C relative to any > *medium* it traversed. Since the light rays passed through the > atmosphere before getting to the spectroscope, the signal velocity was > assumed to have been "changed" to C relative to the air before the > signal arrived; that neatly accounts for the observed redshift. > > However, we arrive at once at a prediction that spectroscopes above the > atmosphere, which are filled with vacuum and separated from the emitting > star by nothing but vacuum, will *not* show a redshift. When the theory > was first put forth that seemed reasonable. However, Hubble shot that > down -- the redshift, measured in space, is the same as the redshift > measured on the ground. So, emission theory needs another fix. > > The next "fix" is to postulate an "extinction rate" for empty space. > This is either due to the presence of rarified gas, or the presence of > some immaterial vortex material (a little like an aether but not quite). > In either case, the consequence is that the velocities of light pulses > passing through interstellar space gradually revert to C relative to the > "empty space" through which they're passing. This can account for the > observed redshifts provided by Hubble's spectrascopes. However, it has > the unpleasant property that one cannot *predict* the "extinction rate" > from first principles (the "extinction rate" is the rate at which the > original velocity of the light beam changes to match that of the > "surrounding space"). Rather, one merely adjusts it to whatever is > necessary to match the observations. Consequently, one eliminates > spectroscopic evidence as a test of the theory; the theory no longer > predicts much of anything about the redshifts one should observe. > > In any case this still leaves the Sagnac experiment, which is a bear to > deal with in emission theory. I don't how to fix the theory to make it > consistent with that result. Sagnac says that, on a rotating ring, the > time to traverse the edge of the ring depends on which direction you go > around the ring. Emission theory, which sets the propagation velocity > equal to C + the emitter's velocity, predicts no such difference, *if* > the experiment is carried out with a hard vacuum in the apparatus. > However, experiments with evacuated rings showed a difference. (Sorry, > I don't have a reference on this one. The effect has been proved many > times over for non-evacuated rings [just google IFOG] but I'm not sure > when, where, or how many times the experiment using vacuum was done.) > > >> >> Harry >> >
