If you impose uniform simultaneous acceleration of the front and back of the train from the train frame (with your infinitely strong coupling), you lose uniform simultaneous acceleration of the front and back of the train in the station frame, which seems to be one of the assumptions that leads you to expect sliding of the wheels.

To elaborate on the comment by @gs : You wrote "The train is then given an acceleration in the negative x′ direction". Your question can't be answered until you tell us what this means. In (say) the frame of the rails, do the front and back accelerate at the same instant (in which case the length of the train remains the same as before in the rail frame)? Or the front before the back? Or the back before the front? [It's fine to assume for simplicity that the accelerations are simultaneous, but we still need to know the order of the instants for different parts of the train.)

Nothing actually contracts, so there is never any friction! That is your big time misunderstanding ;)

Ok thanks. I hadn't given a notification til just now. (Why can't I start a new paragraph?) "The train is then given an acceleration in the negative x′ direction". That came from someone else on Quora. And I take it, it's the exact same thing as in Bell's paradox, but in reverse (as explained on Wikipedia). So, simultaneous acceleration in the frame of the rails. And so the back of the train starts accelerating negatively a little bit earlier than the front. (Got to think later why that would matter by the way, because actually all I'd like to know is if with a scenario like this could give

.. could give friction caused by the "un-shortening" of the length contracted object. Of course it would be nice to know how to calculate, but it's confusing enough to me already. @m4r35n357 Nothing is really "squished" together because of length contraction, I understand that. But lengths does really change .. (don't they?)

No, I'm sorry. The front in S (train's frame) starts slowing down before the back. (I see I made some language errors as well, I should reply when I'm a little bit more awake.)

@SergioMarquina Yes, the lengths really do change - at least, in the physics sense of "really", that is, the universe acts in every way as if the length was the length that you measure. And yes, if the front and back start accelerating simultaneously in the station frame, the front accelerates before the back in the train frame, so we should expect the train to buckle under compressive stress.

Ok thank you. But what's most confusing to me in this case is that in the trains frame the rails contract. So that would mean that when it stops being contracted, because of the train decelerating and stopping, the stresses and forces caused by the "un-shortening" would point the opposite way. That, of course, is impossible. Or wait a minute, in that frame relativity of simultaneity solves it? Uhm. (So confusing.)

@SergioMarquina no, nothing changes, all effects are apparent (think about it: they are different for different observers!), not real!

I'm trying. But the person who gave me the answer I quoted comes from the witer of this book: amazon.com/Manifolds-Groups-Bundles-Spacetime-Dale/dp/…. And has been a professor on different uni's all his life. I can't just ignore that. It's weird. Normally I have no problems with SR, GR I find much harder. Also some other teacher told me I should use kotler moller coordinates. I tend to believe you, but these people I know understand relativity very well. (Also I feel like overthinking this. I'm very good at that.)

BTW that link does not work. Please test links yourself in future!

@m4r35n357 I get the spirit in which you are making the point about length contraction being apparent, but for someone who is new to the subject, I would like to point out that there is apparent (i.e., observer-dependent where observer means a choice of frame of reference) and then there is apparent (i.e., not only observer-dependent but also standpoint-dependent or something even more trivial, e.g., a purely a human illusion). Length contraction is the former kind of apparent. [...]

[...] An example of the latter kind of apparent (i.e., standpoint-dependent) would be the Doppler effect and the even more trivial example of the latter kind of apparent (i.e., purely a human illusion) would be a mirage or something like that.

arxiv.org/abs/physics/9810017 relativistic motion of an arbitrary point of an accelerated rigid rod is discussed for the case when velocity and acceleration are directed along the rod's length.

@DvijD.C. as a learner here myself, I appreciate clear and unambiguous answers, and try to give them myself where I see a misunderstanding. The idea that length contraction causes friction is one of these! BTW I found your comments very hard to parse ;)

@m4r35n357 , we can construct a scenario in which length contraction unambiguously corresponds to friction. Take Bell's space ship paradox, but replace the string with a winch and cable and put a friction break on the winch. We have to frame boost to find true causation, in which case we find the unequal acceleration of the front and back rocket in the winch frame causing the friction, but in the inertial frame we find friction iff length contraction.

@m4r35n357 You are thus wrong when you say, "Nothing actually contracts, so there is never any friction." You are half right when you say that length contraction doesn't cause friction: Lorentz transformations aren't events, they can't cause, only correlate. But physics is an empirical science. If you can measure it it's real, insofar as science can deal with reality. For what it's worth, I don't find Dvij's comments hard to parse, and I don't have any trouble with Sergio's link.

@gs but correspondence is not causation ;) Any "friction" in the Bell paradox(including static apparently!) is caused by the acceleration, and can be calculated in those terms. Any "length contraction" that you identify is merely a derived quantity of this problem, and not a cause of anything, IMO. I could be wrong, but I need more convincing, and your capacity to understand DvijD.C is neither here nor there ;)

@gs I still get "Firefox can’t establish a connection to the server at rads.stackoverflow.com", so you have the advantage over me!

I have a feeling my question is not clear. I apologize for that. But I'm not asking whether length contraction alone can cause friction (though in Bell's spaceship paradox it's due to length contraction in the inertial frame as I understand). But I wonder about whether the "undoing" (the un-shortening) of a length contracted object, because of a deceleration, can cause friction.

@m4r35n357 The link works fine for me.