guyy wrote:Yes, but, if everyone thought that way, how would we get new theories? Special relativity itself was a ridiculous overturning of basic physics ideas (mainly, that time is absolute) based on simple thought experiments (if light travels at a universally constant speed, what happens if you're moving relative to a photon?). Just because an idea is uninformed doesn't mean it's wrong. Probably wrong, but not definitely wrong.
If it is wrong because it's uninformed, it's better to just point the person to the probably-already-existing counterargument than to just say "you don't know enough, so you must be wrong."
Edit: I'm guessing FCN's paradox above would disappear if you actually used Lorentz transformations to find the time in each frame; with the racecar going backwards, the time may not change in the way you might expect. I don't really have time to check that, though.
Mostly we get new theories because we know our old theories are wrong or incomplete.
Like special relativity came about because physicists realised that there was something the matter with Maxwell's theory of electromagnetism, especially with relation to the velocity of propagation of light. This is why the Lorentz transformations were developed, for instance - they're the transformations under which the Maxwell theory is covariant. It took someone like Einstein to put together the puzzle pieces and realise that electromagnetism made a lot more sense if you threw away your old ideas of space and time, but it grew very organically from the incompleteness of the existing theory.
Similarly quantum mechanics came out of several other inconsistencies in the known ideas of physics. For instance, that a classical atom would be unstable because of bremsstrahlung radiation, or the explanation of the photoelectric effect.
The real trouble with today's physics is that why our theories are broken and incomplete are a lot harder to understand. Like that many possible quantum fields are unrenormalisible, or that you need a Higgs effect to bring mass terms into the Standard model (which is something to do with spinor fields... exams this year are going to be fun!). Sure, you might stumble across a simple thought problem that happens to fix all these problems, but it's pretty unlikely.
Incidentally, for fun with people who don't like relativity, see: http://groups.google.com/group/sci.physics.relativity/