Possible to publish a paper that contradicts Einstein's special relativity?

[Algr]

Right. That's half of the process of 'proving' something in scientific sense.
Limiting the theory to an area where it's actually right.

But in the case of Newton, it was Einstein who did this.

 

[Rive]

But in the case of Newton, it was Einstein who did this.

Yes. It was OK for Newton, though: but in this age we are already a bit more careful with 'all creation included' type of theories, so some initial work from the author is a requirement.
Even if the 'edges' are not visible yet, at least to prove that within the limits of the 'old' theory it works at least just like the old one is the basic courtesy.

 

[russ_watters]

@Algr and @Rive I think there is a fair amount of talking past each other here, but as it reads I agree with the statement that started it, and later logic discussing it via comparison to Newton:

Wanting to disprove Einstein is a noble goal... if you can actually do it.
[separate post]
Ask Newton what .8c + .8c is.

You can't disprove anything what's working. That's just the complete misunderstanding of how science works.

IMO, it's fair to speculate that Newton would answer 1.6C and it is fair to call that "wrong". Broader, it's fair to say that Newton's theory "works" in its domain of applicability and in a handwavey way "works" in the domain in which we use it. But when you do that I think you need to acknowledge that Newton didn't know what that domain of applicability was and almost certainly didn't have a clue as to just how limited/wrong his theory was. Scientists only realized it later when their measurements got more accurate. Newton probably would have said his theory was "working", because as far as he knew, it was -- but he'd have been wrong. And he probably believe his theory was The Correct One.

I do not think it is fair to call Relativity an "extension" or "redefinement" (refinement?) of Newton's laws; it's a replacement (for the purpose of theory; practical use is different). The domain of applicability of Newton's laws is very, very small and we mostly use them today in areas known to be outside their domain of applicability because the math is easier and the error is small enough not to worry about. Their domain are very limited, even single-point special cases at best.

Einstein's theory is much more advanced, with much better experimental data/observation to back it up. But that data isn't infinite in its coverage or precision. There exists the possibility, however remote, that somewhere in those error margins, an experimental result will contradict Relativity. And if that happens, Relativity will be "wrong" and "disproven". Maybe not completely, but within the bounds of the experimental result. In that case, a boundary will need to be drawn or a new theory that matches Relativity's results up to that limit but deviates outside that limit will be needed. Yes, it's extremely unlikely, but if the possibility was absolute zero, then there's be no point in pushing the accuracy of the experimental verification any further. The expectation is that Relativity is 100% accurate in its domain, but it is recognized that it isn't 100% proven, nor can it ever be (though perhaps someday scientists will stop bothering to try to disprove it).

But then you need to prove that within the (proven) old frames your new approach is mathematically equivalent with the old theory.

In my opinion, this is a common statement but an over-statement, and contradicts your previous judgement of the velocity addition example: The velocity addition formulas are not mathematically equivalent. Newton's is wrong, full stop. This is binary. The fact that we use it anyway doesn't mean it isn't wrong.

Extending, the statement then assumes that Relativity is completely accurate instead of acknowledging that by nature experiments can only provide very good approximate verification. I don't think it's hairsplitting because, again, if Relativity were accepted to be 100% accurate, there'd be no need for further testing and the...non-mainstreamers...would be right to point out that we're considering something 100% accurate without 100% accurate experimental verification. A new theory must agree with the existing body of evidence, not the existing theory (except where the two theories and the evidence overlap).

All that said, it's likely the OP doesn't know the overlap and so doesn't know their theory actually does contradict existing experimental evidence.

Note: I define "domain of applicability" as the domain where a theory is believed to be accurate/correct. It does not include the domain where it is known to be wrong but we just don't care because it's close enough. In that sense, as far as I know, Newton's laws of motion have no domain of applicability at all. There is no speed where the Galilean velocity addition formula is believed to be accurate. No speed where it and Einstein's formulas give exactly the same result.

 

[russ_watters]

Since the OP hasn't returned I don't feel bad about hijacking the thread, but if need be we can split part of this off into a new thread discussing "What does 'Domain of Applicability' Mean?" I feel like there are three different definitions people use:

1. The domain in which we use it.
2. The domain in which it is proven accurate.
3. The domain in which it is believe to be accurate.

The definition used will impact how one interprets the statement "correct in its domain of applicability".

 

[Dale]

In my opinion, this is a common statement but an over-statement, and contradicts your previous judgement of the velocity addition example: The velocity addition formulas are not mathematically equivalent. Newton's is wrong, full stop. This is binary.

I disagree with this. Newton's velocity addition formula is mathematically equivalent to the SR velocity addition formula for $v<<c$. In fact, it is critical that the SR velocity addition formula reduce to the Newtonian formula in the limit $v<<c$ precisely because we have a lot of data in that limit that supports the Newtonian formula.

I define "domain of applicability" as the domain where a theory is believed to be accurate/correct. It does not include the domain where it is known to be wrong but we just don't care because it's close enough.

As far as I know, that is not the usual definition. I believe that the usual definition is that the theory's domain of validity is the domain where there is experimental data which validates the theory. There is a lot of experimental data that supports Newtonian physics, including the Newtonian velocity addition. In that domain Newtonian physics is valid (hence domain of validity). SR, to be valid, must also match the established correct predictions of Newtonian physics in that domain.

Note that the precision of an experiment is an important characteristic of the experiment. So the issue is not "we just don't care because it is close enough" but rather than with a certain experimental precision the theories are indistinguishable. They both agree with the data equally. The domain of validity includes not only the experimental velocity but also the experimental precision.

 

[Rive]

No speed where it and Einstein's formulas give exactly the same result.

It is a bit of a problem to support that claim with experimental data.

 

[Jarvis323]

Newton's velocity addition formula is mathematically equivalent to the SR velocity addition formula for $v<<c$.

This might just be a matter of semantics, but I would say, "experimentally equivalent", yes, "approximately equivalent", yes, but "mathematically equivalent", no.

 

[russ_watters]

It is a bit of a problem to support that claim with experimental data.

Huh? The statement is purely mathematical. It has nothing at all to do with experiment. That's one of the key points of the entire post.

 

[russ_watters]

I disagree with this. Newton's velocity addition formula is mathematically equivalent to the SR velocity addition formula for $v<<c$. In fact, it is critical that the SR velocity addition formula reduce to the Newtonian formula in the limit $v<<c$ precisely because we have a lot of data in that limit that supports the Newtonian formula.

@Jarvis323 accurately describes my position.

As far as I know, that is not the usual definition. I believe that the usual definition is that the theory's domain of validity is the domain where there is experimental data which validates the theory.

That's a problem for me, then. Is there another name for what I describe? To me, the distinction matters because in one case we need to keep looking for a better theory and in the other case we don't.

There is a lot of experimental data that supports Newtonian physics, including the Newtonian velocity addition. In that domain Newtonian physics is valid (hence domain of validity). SR, to be valid, must also match the established correct predictions of Newtonian physics in that domain.

Note that the precision of an experiment is an important characteristic of the experiment. So the issue is not "we just don't care because it is close enough" but rather than with a certain experimental precision the theories are indistinguishable. They both agree with the data equally. The domain of validity includes not only the experimental velocity but also the experimental precision.

Since I'm not a scientist, I may not have the theory/process of how errors are dealt with correct, but my understanding was that error bars are not hard limits, so there is no binary yes or no or necessarily an equality of two theories. Isn't there still debate about whether the original Michelson Morley experiment was accurate enough to claim a null result? E.G., it was zero within the error bars, but the error bars were pretty wide?

It also means that for different experiments, Newton's laws have different domains of applicability. That seems very loose/vague to me. I had assumed the domain of applicability was more specific than that, and even independent of the specific experiment. If I'm using my car odometer and you're using a laser interferometer, we can both reasonably claim different domains of applicability of Newton's Laws.

 

[Dale]

This might just be a matter of semantics, but I would say, "experimentally equivalent", yes, "approximately equivalent", yes, but "mathematically equivalent", no.

$$\lim_{c \rightarrow \infty} \frac{u+v}{1+uv/c^2}=u+v$$ So I stand by “mathematically equivalent to the SR velocity addition formula for $v<<c$”

 

[Rive]

the statement is purely mathematical.

That's nice. But somewhere down the line it is exactly the experiments what turns math into physics.

Without experiments supporting the difference in a specific range all what you can say is that keeping the distinction is nice.

Being nice actually still can be a valid point, but being practical can be one too ... And this has nothing to do with 'right'. Limited - yes, that's the one :)

It also means that for different experiments, Newton's laws have different domains of applicability.

For practical reasons, this stands not just for experiments but even for educational levels too.
Indeed, sometimes it is a mess.

 

[russ_watters]

$$\lim_{c \rightarrow \infty} \frac{u+v}{1+uv/c^2}=u+v$$ So I stand by “mathematically equivalent to the SR velocity addition formula for $v<<c$”

Does adding a condition $v<<c$ really provide the same result as $$\lim_{c \rightarrow \infty}$$ ? I thought $v<<c$ meant "so small the deviation becomes immeasurable."

 

[hutchphd]

Of course this is just semantics perhaps, but I would say, "experimentally equivalent", yes, "approximately equivalent", yes, but "mathematically equivalent", no.

I believe the appropriate term here is "asymptotically equivalent" but I don't really care.
I am more concerned with the larger question vis.

Huh? The statement is purely mathematical. It has nothing at all to do with experiment. That's one of the key points of the entire post.

If you really mean this then we can equally debate the number of angels who fit on the head of a pin. I know that you understand this but the statement itself is far too categorical.

 

[russ_watters]

That's nice. But somewhere down the line it is exactly the experiments what turns math into physics.

I'm an engineer and I recognize that nothing I calculate or measure is exact. It was my understanding that mathematicians deal with exact math and that physicists (scientists) sometimes deal with exact math and sometimes deal with inexact experiments and that the difference matters. I'm honestly baffled that it doesn't appear to be the case - that the line appears to be quite blurry.

Again, the main reason why I think this should matter is that it determines whether there is any need/ value in further theoretical development or experimentation or if science could just stop.

 

[russ_watters]

If you really mean this then we can equally debate the number of angels who fit on the head of a pin. I know that you understand this but the statement itself is far too categorical.

I'm really not following you.

 

[hutchphd]

I'm really not following you.

I'm not saying it very well. I really just mean that the behavior of any mathematical theory can be endlessly debated in the realm where the experimental data is not precise enough to differentiate two theories (angels size for instance). It is the hallmark of physics that experimental data always matters and so I disliked the statement categorically.

 

[russ_watters]

I really just mean that the behavior of any mathematical theory can be endlessly debated in the realm where the experimental data is not precise enough to differentiate two theories (angels size for instance).

Agreed. But the impression I'm geting is that scientists do not believe there is enough room in those error margins for another theory, or believe Relativity is exact/correct -- and flip back and forth between the two positions/descriptions. It's a lot more vague than I thought the way scientists think.

 

[hutchphd]

And actually quite various as well. I think that was more surprising to me. It is why we are humans and not machines I guess.

 

[Jarvis323]

I believe the appropriate term here is "asymptotically equivalent" but I don't really care.
I am more concerned with the larger question vis.

If you really mean this then we can equally debate the number of angels who fit on the head of a pin. I know that you understand this but the statement itself is far too categorical.

Well, I don't know because the much less than symbol is not exactly defined as an asymptotic limit. In this case, if you take a limit, then $v = 0$ is probably the appropriate one I guess. Rather than use the concept of a limit, you could just say for $v=0$.

Anyways, if you take it all the way to the limit, as a way to define the domain of applicability, then you have reduced the domain of applicability to infinitesimally small. And you could come up with all kinds of ridiculous equations that are equivalent at $v=0$.

But I get the point.

 

[PeroK]

$$\lim_{c \rightarrow \infty} \frac{u+v}{1+uv/c^2}=u+v$$ So I stand by “mathematically equivalent to the SR velocity addition formula for $v<<c$”

That's more than stretching a point. If "mathematically equivalent" means anything it means they are the same axiomatically (which they are not). Moreover, one system being a special case of the other does not mean they are equivalent. The geometry of circles is not mathematically equivalent to the geometry of ellipses, even though the circle is a special case of the ellipse.

The Hafele-Keating experiment is evidence that Newtonian physics and Relativity are not equivalent, even within what might be expected to be the domain of applicability of Newtonian physics: airline travel.

 

[Rive]

I'm an engineer and I recognize that nothing I calculate or measure is exact.

With physics it is even worse. It takes a lot of metaphysics (philosophy) to accept that we can't deal with 'reality' (whatever it means). We only have experiments, and theories fitting them. So the whole 'real', and 'right' is a kind of alarm bell, since these are just out of physics.

If you have only experiments, then how can you decide which theory is 'right' if you can't support the distinction (within a range) with experimental data?

Nice - it works. Works - fine. Limited - very good. Useful - even better! Practical - hooray for it!
Right - now, that's very-very suspicious...

 

[sysprog]

<joke>
Just in case you guys were/are still concerned about angel size:
$\ell_{(M,\varphi)}:\Delta^+=\{(x,y)\in \mathbb{R}^n:x<y\}\to \mathbb{A}$ where $\mathbb{A}=\mathrm{[angels]}$
there you go $\dots$
</joke>

 

[DaveE]

Dunning-Kruger comes to mind here.

I have no idea how this relates to the OP's work since we haven't seen it. Even then I probably wouldn't, but y'all might.

I believe that this effect applies to everyone in each knowledge domain. The question is: do we have the self awareness to know where we are on the graph?

“The first principle is that you must not fool yourself — and you are the easiest person to fool.” - R. Feynman

 

[russ_watters]

And actually quite various as well. I think that was more surprising to me.

With physics it is even worse. It takes a lot of metaphysics (philosophy) to accept that we can't deal with 'reality' (whatever it means). We only have experiments, and theories fitting them. So the whole 'real', and 'right' is a kind of alarm bell, since these are just out of physics.

If you have only experiments, then how can you decide which theory is 'right' if you can't support the distinction (within a range) with experimental data?

Well, where the rubber meets the road, I'm not sure I believe it. There's a 8.5 km supercollider near Geneva, not an 8.5 km Michelson interferometer. There's a reason for that. There's a reason we scoff when someone says they have an idea for a new theory that contradicts Relativity. I don't think it's just that the person isn't qualified to make the claim - I think the reason in both cases is that physicists believe Relativity is Correct. Not "correct within its domain of applicability", but Actually Correct.

 

[f95toli]

Agreed. But the impression I'm geting is that scientists do not believe there is enough room in those error margins for another theory, or believe Relativity is exact/correct -- and flip back and forth between the two positions/descriptions. It's a lot more vague than I thought the way scientists think.

I don't think that is true. There are quite a number of scientists who spend their whole career doing more and more accurate measurements to see if they can spot any difference between their measured values and what is predicted by theory. Some of this work is published in high-impact journals (see https://www.nature.com/articles/s41586-020-2964-7 for a recent example.)

We all know that physics isn't "complete"so it is entirely possible that we will one day find a "more complete" theory which also works in situations where existing physics (including SR) isn't applicable or-assuming we one day find an "error"- is able to predict results with higher accuracy.

The key here is realising that these situations would have to be either very exotic OR you are trying to calculate something with a precision which is beyond what we can currently measure. So any new theory would still need to explain existing data.

 

[BWV]

Dunning-Kruger comes to mind here.

I have no idea how this relates to the OP's work since we haven't seen it. Even then I probably wouldn't, but y'all might.

I believe that this effect applies to everyone in each knowledge domain. The question is: do we have the self awareness to know where we are on the graph?

“The first principle is that you must not fool yourself — and you are the easiest person to fool.” - R. Feynman

View attachment 276494

seems like everyone who sees that graph thinks they are an expert on Dunning-Kruger

 

[Vanadium 50]

not an 8.5 km Michelson interferometer

No, that's only 6km, and it's in Tuscany.

However, I think the point you are trying to make is valid. Nobody would say it's worth doing a billion dollar experiment to add one more zero to the null result of Michelson and Morley in lieu of a billion dollars worth of other experiments.

If SR were wrong, would a better Michelson-Morley experiment find it? I think the answer is probably not. You would need some sort of incomplete ether drag, so that the ether wind is a fraction of a meter per second, but then you run into problems with Michelson-Gale-Pearson type measurements.

Trivia question: which Nobel prize winner was a proponent of partial ether drag theories?

 

[Dale]

Does adding a condition $v<<c$ really provide the same result as $$\lim_{c \rightarrow \infty}$$ ? I thought $v<<c$ meant "so small the deviation becomes immeasurable."

I certainly could be understanding it wrong, but I have always understood it as a shorthand for the limit as v/c goes to zero.

In this case since there are two velocities it is equivalent but easier to take the limit as c goes to infinity.

 

[Vanadium 50]

The best Michelson-Morley experiment type-experiment appears to be a 2005 Dusseldorf experiment:
https://arxiv.org/abs/gr-qc/0504109 and https://arxiv.org/abs/physics/0602115

Expressed as ether wind, the ether is moving at less than 9 cm/s (0.2 mph). Somewhere between a sloth and the world's fastest snail. I don't see a huge value in measuring this well enough to move the limit to 1 cm/s. Does anyone?

 

[Dale]

I feel like there are three different definitions people use:

1. The domain in which we use it.
2. The domain in which it is proven accurate.
3. The domain in which it is believe to be accurate.

That is possible. I tend to use 2., but now that I think of it I am not sure that anyone I have read actually defined it clearly. So 2 is simply what I inferred from context and usage.

Since I'm not a scientist, I may not have the theory/process of how errors are dealt with correct, but my understanding was that error bars are not hard limits, so there is no binary yes or no or necessarily an equality of two theories.

You are right. I have an Insights article about Bayesian inference in science. In my opinion that is the best way to evaluate the evidence without resorting to a binary yes no.

But the impression I'm geting is that scientists do not believe there is enough room in those error margins for another theory

I think that there are some beyond the standard model theories that differ from SR but live in those error margins.