Einstein's special relativity rests on two postulates: the laws of physics are identical in all inertial reference frames, and the speed of light in vacuum is the same for all inertial observers regardless of their motion or the motion of the source. These seemingly simple statements force a radical revision of Newtonian notions of absolute time and space. Events that are simultaneous in one frame need not be simultaneous in another, and time and length are no longer invariant quantities.
Start with thought experiments — the classic 'train and lightning' scenario for simultaneity, and a light-clock for time dilation. Construct the argument carefully: if c is constant and finite, something must give. Only introduce the Lorentz factor γ after the conceptual argument is solid.
By the late 19th century, physics faced a quiet crisis. Newtonian mechanics was spectacularly successful, but Maxwell's equations — the theory of electricity and magnetism — predicted that electromagnetic waves travel at a fixed speed c ≈ 3 × 10⁸ m/s. The problem was that Newtonian mechanics said speeds always add: if you run forward on a train, your speed relative to the ground is your speed plus the train's speed. Applied to light, this would mean different observers should measure different values of c depending on their motion. The Michelson-Morley experiment in 1887 tested exactly this and found no variation — c appeared genuinely constant regardless of the direction of measurement or Earth's orbital motion. Something had to give.
Einstein's resolution in 1905 was to take the problem seriously rather than patch it. He elevated two observations to the status of postulates: the laws of physics are the same in all inertial (non-accelerating) reference frames, and the speed of light in vacuum is the same for all inertial observers regardless of the motion of the source or observer. The first postulate is a generalization of Galileo's principle of relativity from mechanics alone to all of physics. The second postulate is the sharp one: c is not just very fast — it is an absolute constant that no observer can outrun or match.
Together, these postulates force a radical conclusion. If two observers moving relative to each other both measure the same c, then their measurements of time and distance cannot be the same. The classic thought experiment is a "light clock" — a photon bouncing between two mirrors. An observer moving relative to the clock sees the photon trace a longer diagonal path, yet must measure the same c; since c is fixed, the time between ticks must appear longer. This is time dilation, one of the first consequences you will derive from the postulates. Simultaneity fails for similar reasons: the relativity of simultaneity follows directly from demanding that both observers measure c for the same light flash.
A crucial point about applicability: special relativity is *always* the correct theory. The reason Newtonian mechanics works for everyday situations is that the relativistic corrections scale with (v/c)², which is negligibly small when v is much less than c. But "negligible" is not "zero." GPS requires relativistic corrections even at orbital speeds that are a tiny fraction of c. The Newtonian world is a mathematical limit (v/c → 0) of the relativistic world, not a separate regime.
The postulates also implicitly tell you that reference frames are abstract coordinate systems — not physical objects or the observers who use them. Two rockets, each moving at constant velocity in different directions, each define a valid inertial frame with equal claim to being "at rest." Neither frame is privileged. This is not just a philosophical nicety; it means every physical prediction must be identical (or consistently transformed) no matter which frame you compute in — which places tight constraints on every equation in relativistic physics.