General Relativity Is At Least 99.95% Right 223
ultracool writes to mention a ScienceDaily piece on compelling proof of general relativity. A team at the University of Manchester have used three years' worth of data on a pair of pulsars as a litmus test, against which they've benchmarked Einstein's theory. From the article: "Though all the independent tests available in the double pulsar system agree with Einstein's theory, the one that gives the most precise result is the time delay, known as the Shapiro Delay, which the signals suffer as they pass through the curved space-time surrounding the two neutron stars. It is close to 90 millionths of a second and the ratio of the observed and predicted values is 1.0001 +/- 0.0005 - a precision of 0.05%. A number of other relativistic effects predicted by Einstein can also be observed. 'We see that, due to its mass, the fabric of space-time around a pulsar is curved. We also see that the pulsar clock runs slower when it is deeper in the gravitational field of its massive companion, an effect known as "time dilation."'"
Re: General Relativity Is At Least 99.95% Right (Score:4, Interesting)
Newton's formulas are not merely missing a few parameters... they involve concepts that simply stop making any logical sense once you get down to very small scales. The idea of a "particle" even existing in a single position, as far as we can tell with modern QM, is completely absurd and meaningless. The concept of an exact momentum is equally so. The "clockwork universe" which contains action at a distance (causal nonlocality) and non-discrete space, time, energy etc (rather than discrete geometry and quanta) is simply so far from the "truth" that our experiments reveal - namely that particles act as if they are in infinite numbers of places at once (or nearly so, given plank limits on spaceitme).
You will, however, find that if you wish to predict the path of a simple artillary shell or design an automobile they are "correct," they have predictive value, specifically because the phenomenon exist within the limits of the model's significance. Taking Relativity into account does nothing but complicate the math to provide a bogus level of significance and Quantum Theory is completely irrelevant.
Back at university, I used these "wrong" theories all the time, as they are useful (if erroneous) abstractions. The problem is that theories are not merely useful for their ability to predict things within the realm of known experience, but also new and different things beyond the current frontiers. Newton's theories, as elegant and beautiful as they are, were long ago surpassed and are now almost useless when it comes to generating new predictions about unobserved phenomena in the universe. The mark of a truly good theory is not that it can compress the set of known expimental results well, but that it can predict entirely new ones, outside the original domain in which it was devised.
Newton was a far smarter man than anyone posting here on slashdot, but like Einstein, he got so very much fundamental very wrong. I think if he lived here today, he'd get new and exciting things wrong (like modern theorists) and that that's a very valuable part of science, but we really shouldn't pretend his theories are anything more than a bunch of mathematical approximations that reference intuitive concepts that have almost no meaning at very small (and possible very large) scales.
You can measure position in quantum physics (Score:2, Interesting)
At least in principle, you can measure position in quantum physics. The particle is temporarily put in a position eigenstate with an exact position eigenvalue associated with it (the momentum is completely indeterminate, however). This only lasts for an instant, however, before the state evolves into a superposition of position eigenstates.
Remember, it is an axiom of quantum mechanics that measuring observables puts the system in an eigenstate of that observable; the eigenvalue corresponds to an exact measurement of that observable. (You will not be in an eigenstate of any observable that commutes with it, and therefore those quantities will not be known exactly — the Heisenberg uncertainty principle.) Of course, you could quibble about our practical ability to put a particle into an exact position eigenstate, as opposed to an eigenstate of an observable merely very similar to the position operator.
Re:99.95% acurate? (Score:3, Interesting)
But regarding your "I think 99.95% is about as close to dead-on-balls-accurate as it gets with our current knowledge of the universe", allow me to take this opportunity to point out that Quantum Electrodynamics (the extension of electromagnitism and quantum mechanics into a quantum field theory) surely is the most accurate theory we have today.
In some circumstances its predictions have been verified to an astounding 14-15 decimal places! (Thats something crazy like 99.9999999999995%).
Of course, the day we combine quantum field theory with Einsteins general theory of relativity, that will be quite something. I for one hope it happens in my lifetime (and plan to go on a month-long rampage of drinking, dancing & singing bad karaoke in the streets if it happens).
But What About the Boomerang Project...? (Score:3, Interesting)
Re:Even Newton is 99.995% right for most stuff (Score:3, Interesting)
Absolutely, and with General Relativity despite its stunning success we know that it must fail at some scale because as a classical theory it simply does not match what we know about space at the very small scale.
The vacuum is a much more active place and while at the long scale it can be described my a nice smooth metric, we already know that it doesn't match up with what we already think we know from field theory. Though even QFT has problems particularly one thats 10^120 orders of magnitude large. These limitations don't make either GR or QFT useless, nor really wrong - incomplete but it is a good description of nature in some regime. So Newton is still useful today partciularly since we still live in a world where his theory still makes very adequately accurate predictions.
Even if GR has its limits its still a very, very powerful theory. What we know about the dark energy seems to indicate that it is a cosmological constant and the univese is asymptotically deSitter space rather than Minkowski. That by itself is one hell of a prediction to be able to make. Theres still a lot of work in GR both theoretically (just take a peek at gr-qc at the arXiv) and experimentally (including this observation, the APOLLO LLR, and Eric Adelberger's group and their beautiful Eot-Wash experiments) - its still being fruitful decades down the line. And just because there will be a new theory that supersedes it there is no way to throw GR out completely - theres just no way to train a physicist without introducing him to Newton's laws at some time. Theres even some merit to studying it for its asthetics - its a pretty theory!
Not good enough for me (Score:4, Interesting)
The measurement is still in the range of first order parametrized post-Newtonian accuracy. What the Donkey Kong that means is that these are the coefficients to the metric that are being tested:
dtaU^2 = (1 - 2 GM/c^2 R + 2 (GM/c^2 R)^2) dt^2
- (1 + 2 GM/c^2 R) dR^2/c^2
- R^2/c^2 dtheta^2
- R^2/c^2 sin^2 theta dphi^2
It is the 5 integers there (1, -2, +2, -1, -2) that are confirmed by this experiment. That is NOT NEWS, because it is not new. Shapiro got the same results. What would be news is if the experiment got to second order parameterized post Newtonian accuracy. I asked Prof. Clifford Will an expert on experimental tests of GR when where the data hunters going to gather that data. He said he knew of no one even discussing it. The reason is that the data must for 2nd order PPN effects must be a million fold more accurate, so we need data that is 99.99995% accurate.
I care a lot about 2nd order PPN tests, since that is were my proposal to unify gravity and EM using a 4D wave equation differs. GR says the metric should go here:
GR:
dtaU^2 = (1 - 2 GM/c^2 R + 2 (GM/c^2 R)^2 -3/2 (GM/c^2 R)^3) dt^2
- (1 + 2 GM/c^2 R + 3/2 (GM/c^2 R)^2) dR^2/c^2
- R^2/c^2 dtheta^2
- R^2/c^2 sin^2 theta dphi^2
GEM (gravity and EM):
dtaU^2 = (1 - 2 GM/c^2 R + 2 (GM/c^2 R)^2 -4/3 (GM/c^2 R)^3) dt^2
- (1 + 2 GM/c^2 R + 2 (GM/c^2 R)^2) dR^2/c^2
- R^2/c^2 dtheta^2
- R^2/c^2 sin^2 theta dphi^2
At first order PPN accuracy, the coefficients (1, -2, 2, -1, -2) are the same. At second order, they are different. That's the data I need. I'll probably be dead before it shows up.
doug