I’ve reached a stage in my career in which I have started to get letters addressed directly to my from people who state Einstein was wrong and then typically provide 3 to a lot of pages describing, typically with no math, why Einstein was wrong and why their new theory is right.

Here’s the thing. At this stage in the game, even if we do someday find short comings in Einstein’s theories, we won’t be proving his theory wrong, we will be expanding on it. Just as we don’t say Newton’s theories as being wrong, we will not talk of Einstein as being wrong. Newton’s theories were incomplete, but they are correct within the realm of classical physics experiments. We already know Einstein’s theories break in quantum regimens, but they are correct in classical and relativistic regimens. One more theory (at least!) is needed, but all because something isn’t true all the time doesn’t mean it’s wrong.

Example: They sky, as I write this, is not blue. That said, when I say the sky is blue, I am not making a stupid statement – I’m just making a statement that is only true in certain regimens of sunlight and weather.

Einstein’s not wrong.

Got it? Good.

Now, that you know my state of mind, imagine my reaction to the current press being given to this latest claim that relativity had been falsified in the classic regimen (note – the paper is not accepted to any peer reviewed journal). So you’ve seen the coverage in New Scientist, you’ve seen the hubbub on Slashdot. But what does it all mean?

That’s actually a very good question.

If you read the paper, 2 things should stand out: there is no statement of uncertainty, and no data results are stated – they just state rather vaguely: “The measured time delay in both reflection and transmission of the digital pulse is about 100 ps.” Can someone please define about 100ps for me? (ps = picosecond = 10^-12 seconds)

So here is what I understand of their experiment: They took two prisms and shot laser light through them. When the two prisms were touching, the laser light passed straight through both and hit a detector. What they don’t say, but I’m assuming, is the light was pulsed, and they measured the time from some part of the departing pulse to some part of the arriving-at-the-detector pulse. (This is how pretty much all experiments of this type are done.) After sending everything through with the prisms touching, they then pulled the prisms apart. In this new situation, as the light traveled through the first prism and hit the surface between the first prism and the air gap between the two prisms, some of the light went through the prism into the air and some of the light internally reflected back. The reflected light went out the base of the first prism and hit a detector, and the transmitted light went through the air, entered the second prism and then hit a different detector.

Since the two prisms are the same size, and the detectors are the same distance from the detectors, the light that is internally reflected should travel a shorter distance (goes in, reflects, bounces out) than the transmitted light (goes in, goes through air, bounces out). The extra travel distance is the size of the air gap. Their claim is photons traveling both paths hit the detectors at the same time.

Here is where it would be nice to see data, and here is why. First off, I’d like to know, does the travel time stay unchanged as the prism spacing increases? This is really important. There is an effect called the Goos-Haanchen shift (hat tip to Jack Glassman for explaining this one to me) that causes a photon that is reflecting to partially pass through the surface and get laterally shifted. The best way I know how to think of this weirdness is like this: imagine throwing a shoe at a chainlink fence at an angle. If the shoe is rotating and the laces are untied, the heel of the shoe may hit the fence, and the shoe may pivot about the heel until the toe hits, and then fly off at a new angle. While it’s hitting, the laces may fly through the fence. Exactly what happens is going to depend on the size of the chainlink and the size of the shoe. This isn’t a perfect analogy, but it is hard to explain wave packets.

So, there could be weird effects that allow the photon interacting with the surface and the photon passing through the air to both get delayed, and in a table of spacing versus travel time, both paths could have the same travel time, while the travel time increases as the air gap increases. Without seeing a data table I can’t know.

And then there is the little matter of error. They are using light with a frequency of 9.15 GHz ~10GHz. This means there are ~10^-10 seconds between wavepeaks if you are watching a wave go by. Now, they mention that the transmission time of the digital pulse is about 100 ps = 100 * 10^-12 seconds = 10 ^-10 seconds. This means there travel time and the spacing between wave peaks in a continuous wave is suspiciously the same. I want to see that explained and that coincidence designed out of the experiment.

And, what is the resolution they are measuring with anyway? When they say about 100ps, do they mean plus or minus 10 ps or plus or minus 100 ps?

And what about the light pulse size? And how are they measuring time? The shape of a pulse of light as it passes through these different systems can change, and unless they are doing this a single photon at a time, we have no way of knowing if the changes in the pulse shape are effecting things. There is a great analogy over on Cosmic Log that explains it this way: You have a train leaving Boston for New York with 99 cars. As the train leaves the station you start a timer as its middle car, number 50, passes the end of the platform. Now, image that as the train travels, rather than stopping for passengers, it just dropped the last car, and then the next to last car, etc, until only 5 cars remain. If you then stop your travel time clock as the new middle car passes the end of the NY platform – car number 3 – it will appear that the train mysteriously sped up, especially when compared to a train that kept all 99 of its cars (whose clock had to keep going for the extra amount of time it took those extra 47 cars to pass). So, the size of the train, and the size of the light packet, matters when you are triggering on the midpoint.

We don’t know what these folks are triggering on.

I want data. I want it now. (Strangely the Veruca Salt, “I want it now” song just popped into my head, but instead of a golden ticket (which I would accept if you offered it) the voice in my head wanted golden data).

Give it to me. I don’t care how. I want data, and I want it now 🙂

So, the moral – don’t believe a paper that isn’t peer reviewed, that uses the adjective about in front of any of their key results without stating error bars, and that presents no actual data while claiming to be an experimental result.

Be skeptical. The truth is out there, but I think we may need to do some more looking to find it.