Ana stands on the platform at a small station outside Zurich, watching a train pass through without stopping. Two bulbs are mounted at the front and back of the last car, wired to flash at the exact midpoint of the car's length, timed to fire together in the train's own frame.
Marcus is riding in that last car, seated exactly at the midpoint between the two bulbs. When they flash, the light reaches him at the same instant. He reports what anyone in his position would report: the two flashes were simultaneous. He measured the distance to each bulb, measured the arrival time of each pulse, and found no discrepancy. Nothing on the train's clocks disagrees with him.
Ana, watching from outside, sees something different. In her frame, the midpoint rider is moving toward the light from the front bulb and away from the light from the rear bulb. For both pulses still to reach Marcus together, the rear bulb must have flashed first in Ana's account. Once she works through where the two events occurred and corrects for light's finite speed, that is exactly what she concludes.
Neither of them is wrong. Neither is being sloppy with instruments. The disagreement is not about measurement error, and it is not resolved by one of them stepping outside their frame to check the real answer, because no frame sits outside all frames. Simultaneity, in this setup, is not a fact about the world that observers merely fail to perceive correctly. It is relative to a frame in the way that "to the left" is relative to which way you are facing.
But something does not move. If Ana and Marcus each compute the spacetime interval between the two flash events, using their own measured space and time separations, they arrive at the same number. The interval is built from a fixed combination of the spatial separation and the time separation, and that combination stays put no matter which frame does the measuring. Proper time along a path is invariant in the same way. So is rest mass. So is the ordering of any two events that could causally influence one another. What relativity removes is a single, universal now. What it leaves standing is a structure that every honest frame, given the same events, will compute alike.
This is a narrower claim than the one people often reach for when they invoke relativity outside physics, the idea that since perspective changes the picture, no picture is more correct than another. Jain thought has a name for a similarly narrow claim: anekantavada, the recognition that a thing has many aspects, so that a statement true from one standpoint, one naya, can seem to contradict a statement true from another. The tradition does not conclude that any claim about the object is as good as any other. It asks for statements to be qualified, to name the condition under which they hold, syat, in some respect. A jar viewed as clay is one truth; viewed as a vessel with a particular shape is another; both hold under their own condition, and the discipline lies in naming the condition rather than dropping it.
Ana and Marcus never needed to be told this. They already knew to ask, from where, and they already knew what stayed the same when they compared answers.