4 ms·

Fun fact: The speed of light is only constant by convention. There's a nice creationist paper where they try to twist physics into a pretzel in order to make it

by cygx 7y ago

Fun fact: The speed of light is only constant by convention. There's a nice creationist paper where they try to twist physics into a pretzel in order to make it fit the biblical narrative, using a convention where the speed of infalling light rays is infinite, but the speed of outgoing ones only c/2. Of course, any such attempt will eventually fail due to invariants that are independent of convention, such as the age of the universe as measured by an observer following the Hubble flow. One can't help but be impressed at least a little bit by the ingenuity of it all, though... edit: cf https://en.wikipedia.org/wiki/One-way_speed_of_light https://en.wikipedia.org/wiki/One-way_speed_of_light

burnte 7y ago
"Fun fact: The speed of light is only constant by convention." Well, the speed of light through a non-vacuum isn't the same as any medium will slow it to some extent, but the speed of light in a vacuum is the definition of "c" and as such, that IS a constant and invariant.
- cygx 7y ago
  Because we define the speed of light in terms of spatially homogeneous and isotropic inertial coordinates. We do that because we want "the statements of the natural laws [to be] as simple as possible" (Poincaré), but it's not the only way to do it...
  - 6gvONxR4sf7o 7y ago
    Sure the speed is a function of the metric you use, but physics also tells us what metric is useful.
    - cygx 7y ago
      It's less about the metric but the choice of space/time decomposition.
      - 6gvONxR4sf7o 7y ago
        I'm not talking about a metric over space, but rather a metric over spacetime, if that's what you mean. I would consider that to encode the space/time decomposition. My differential geometry sucks, but I thought you could write general relativity in a coordinate free manner that specifically doesn't rely on strong coordinate choices like you're talking about.
        cygx 7y ago
        I thought you could write general relativity in a coordinate free manner that specifically doesn't rely on strong coordinate choices Indeed. But even before you start messing with the formalism, the metric alone cannot be used to define a speed as 4-velocities are normalized. You need to decompose your tangent space into timelike and spacelike subspaces so you can have a dx and a dt.
  - burnte 7y ago
    You can over-complicate anything, that doesn't mean you should, or that it's useful.
    - cygx 7y ago
      Investigating which axioms of your theory can be weakened whithout affecting its predictions seems worthwhile to me even if it's useless from a practical perspective.
whatshisface 7y ago
>The speed of light is only constant by convention. I don't think that's true. If someone wrote a paper where the speed of light is different in one direction than another, then that's in clear contradiction with experiments. Yes, there are some legitimate searches for anisotropy in flat space, but they have all come up negative. So, I wouldn't say that it's "constant by convention."
- o_p 7y ago
  You can make a model where the speed of light is relative and be totally consistent with all the scientific observations. That model would be more complicated though.
  - whatshisface 7y ago
    I can think of two things that you might be alluding to, models that have a parameter that is named "the speed of light," but that does not actually correspond to the speed of light, and models that predict phenomena beyond what current experiments have detected. In the first case, anything that is in line with experiment must predict a constant speed of light, even if they do it with non-constant parameters, perhaps even parameters named c. In the second case, sure, but they aren't supported by experiments. In any case, the speed of light is constant.
    - cygx 7y ago
      https://en.wikipedia.org/wiki/One-way_speed_of_light#Generalizations_of_Lorentz_transformations_with_anisotropic_one-way_speeds https://en.wikipedia.org/wiki/One-way_speed_of_light#General...
segfaultbuserr 7y ago
I have also seen many cranks in physics forums claim special relativity is largely an artifact due to the convention of Einstein synchronization, which they claim is falsehood, as this convention adopt the equivalence between "one-way" and "two-way" speed of light as an axiom, and special relativity would be the natural consequence. But there's no reason to accept that, without accepting it, all the arguments for time dilation, etc, collapse, thus special relativity is invalid, then they return to non-relativistic classical mechanics and Galileo transformation. Not a mathematician or physicist, many things are not right in those arguments, I don't know what to say.
- cygx 7y ago
  You can safely ignore that: Relativity and classical mechanics make different predictions, and we've done the experiments. But if you travel into the weeds, a constant speed of light is not the only way to arrive at those predictions: In the extreme case, you could make the speed of light infinite in one direction if you slow it down to c/2 in the opposite direction. We don't do that because we like symmetries, and having light jump instantly from the surface of the sun to hit your head, then 'crawl' towards a mirror only to make an instantaneous jump into your eye seems rather weird...
  - segfaultbuserr 7y ago
    > and we've done the experiments. I accept the findings of those experiments, such as [1]. But if you are a crank, you can always come up with conspiracy theories which claim "The Experiment X" back in the 1930s disprove it but is suppressed by the establishment. I used to read those posts as entertainment, but I stopped since then, as it stopped being fun after a month or so. > But if you travel into the weeds, a constant speed of light is not the only way to arrive at those predictions I think many major subjects of natural science have a "catch" like this. What is being written in popular science and most textbooks is merely a caricature of the subject, far from the full picture, but is considered "correct enough" to be a useful mental model (or physical model) for public consumption. For a simple example, an electric circuit is inherently an electromagnetic phenomenon, yet the fairy tale of "the flow of electron" is commonly used, it's an incomplete and flawed model even at DC. We use it because most people, myself included, are incapable of thinking the physical picture of vector calculus and solve PDE mentally, let alone Quantum Field Theory, not to mention the "truth" is only sometimes relevant and useful. Thus, if someone wants to engage in denialism or create FUD, one could just write something that is largely relevant but out-of-context, and say "you are being fooled, everything you know is wrong"! I suspect many pseudoscientific subjects, like the anti-vax movement or crank physics, frequency use this technique (or fell into it). Also, I once saw a LessWrong post that hypothesized: engineers are more likely to become cranks - because they are knowledgeable enough to spot a flaw in a physical model commonly used in the industry, and realize they are being fooled by thinking critically, but often they are completely unaware of the underlying history or context of the physics. I don't have data, but I definitely have seen something like this. I once saw a website by an RF engineer that claims GR is false, and built his own version using the ideas from antenna theory... [1] https://en.wikipedia.org/wiki/Michelson%E2%80%93Morley_experiment#Recent_optical_resonator_experiments https://en.wikipedia.org/wiki/Michelson%E2%80%93Morley_exper...

by cygx 7y ago

burnte 7y ago

"Fun fact: The speed of light is only constant by convention." Well, the speed of light through a non-vacuum isn't the same as any medium will slow it to some extent, but the speed of light in a vacuum is the definition of "c" and as such, that IS a constant and invariant.

cygx 7y ago
Because we define the speed of light in terms of spatially homogeneous and isotropic inertial coordinates. We do that because we want "the statements of the natural laws [to be] as simple as possible" (Poincaré), but it's not the only way to do it...
- 6gvONxR4sf7o 7y ago
  Sure the speed is a function of the metric you use, but physics also tells us what metric is useful.
  - cygx 7y ago
    It's less about the metric but the choice of space/time decomposition.
    - 6gvONxR4sf7o 7y ago
      I'm not talking about a metric over space, but rather a metric over spacetime, if that's what you mean. I would consider that to encode the space/time decomposition. My differential geometry sucks, but I thought you could write general relativity in a coordinate free manner that specifically doesn't rely on strong coordinate choices like you're talking about.
      - cygx 7y ago
        I thought you could write general relativity in a coordinate free manner that specifically doesn't rely on strong coordinate choices Indeed. But even before you start messing with the formalism, the metric alone cannot be used to define a speed as 4-velocities are normalized. You need to decompose your tangent space into timelike and spacelike subspaces so you can have a dx and a dt.
- burnte 7y ago
  You can over-complicate anything, that doesn't mean you should, or that it's useful.
  - cygx 7y ago
    Investigating which axioms of your theory can be weakened whithout affecting its predictions seems worthwhile to me even if it's useless from a practical perspective.

whatshisface 7y ago

>The speed of light is only constant by convention. I don't think that's true. If someone wrote a paper where the speed of light is different in one direction than another, then that's in clear contradiction with experiments. Yes, there are some legitimate searches for anisotropy in flat space, but they have all come up negative. So, I wouldn't say that it's "constant by convention."

o_p 7y ago
You can make a model where the speed of light is relative and be totally consistent with all the scientific observations. That model would be more complicated though.
- whatshisface 7y ago
  I can think of two things that you might be alluding to, models that have a parameter that is named "the speed of light," but that does not actually correspond to the speed of light, and models that predict phenomena beyond what current experiments have detected. In the first case, anything that is in line with experiment must predict a constant speed of light, even if they do it with non-constant parameters, perhaps even parameters named c. In the second case, sure, but they aren't supported by experiments. In any case, the speed of light is constant.
  - cygx 7y ago
    https://en.wikipedia.org/wiki/One-way_speed_of_light#Generalizations_of_Lorentz_transformations_with_anisotropic_one-way_speeds https://en.wikipedia.org/wiki/One-way_speed_of_light#General...

segfaultbuserr 7y ago

I have also seen many cranks in physics forums claim special relativity is largely an artifact due to the convention of Einstein synchronization, which they claim is falsehood, as this convention adopt the equivalence between "one-way" and "two-way" speed of light as an axiom, and special relativity would be the natural consequence. But there's no reason to accept that, without accepting it, all the arguments for time dilation, etc, collapse, thus special relativity is invalid, then they return to non-relativistic classical mechanics and Galileo transformation. Not a mathematician or physicist, many things are not right in those arguments, I don't know what to say.

cygx 7y ago
You can safely ignore that: Relativity and classical mechanics make different predictions, and we've done the experiments. But if you travel into the weeds, a constant speed of light is not the only way to arrive at those predictions: In the extreme case, you could make the speed of light infinite in one direction if you slow it down to c/2 in the opposite direction. We don't do that because we like symmetries, and having light jump instantly from the surface of the sun to hit your head, then 'crawl' towards a mirror only to make an instantaneous jump into your eye seems rather weird...
- segfaultbuserr 7y ago
  > and we've done the experiments. I accept the findings of those experiments, such as [1]. But if you are a crank, you can always come up with conspiracy theories which claim "The Experiment X" back in the 1930s disprove it but is suppressed by the establishment. I used to read those posts as entertainment, but I stopped since then, as it stopped being fun after a month or so. > But if you travel into the weeds, a constant speed of light is not the only way to arrive at those predictions I think many major subjects of natural science have a "catch" like this. What is being written in popular science and most textbooks is merely a caricature of the subject, far from the full picture, but is considered "correct enough" to be a useful mental model (or physical model) for public consumption. For a simple example, an electric circuit is inherently an electromagnetic phenomenon, yet the fairy tale of "the flow of electron" is commonly used, it's an incomplete and flawed model even at DC. We use it because most people, myself included, are incapable of thinking the physical picture of vector calculus and solve PDE mentally, let alone Quantum Field Theory, not to mention the "truth" is only sometimes relevant and useful. Thus, if someone wants to engage in denialism or create FUD, one could just write something that is largely relevant but out-of-context, and say "you are being fooled, everything you know is wrong"! I suspect many pseudoscientific subjects, like the anti-vax movement or crank physics, frequency use this technique (or fell into it). Also, I once saw a LessWrong post that hypothesized: engineers are more likely to become cranks - because they are knowledgeable enough to spot a flaw in a physical model commonly used in the industry, and realize they are being fooled by thinking critically, but often they are completely unaware of the underlying history or context of the physics. I don't have data, but I definitely have seen something like this. I once saw a website by an RF engineer that claims GR is false, and built his own version using the ideas from antenna theory... [1] https://en.wikipedia.org/wiki/Michelson%E2%80%93Morley_experiment#Recent_optical_resonator_experiments https://en.wikipedia.org/wiki/Michelson%E2%80%93Morley_exper...