Space and Time in Special Relativity

Special relativity is, before it is a theory of anything else, a theory of space and time together. Its lasting philosophical shock is not the odd effects (moving clocks running slow, moving rods shrinking) but the revision of the concepts those effects presuppose: it dismantles the Newtonian picture of a single absolute time flowing uniformly for everyone and of a single absolute space extended the same for all, and replaces both with something at once more operational and more geometric. This remark asks the questions the theory forces on us — what is time? and how is time measured?, and in exact parallel what is space? and how is space measured? — and follows the answers to where they turn metaphysical. It reads the Special Relativity pages philosophically; the systematic treatment of the metaphysical questions it raises is the Philosophy of Space and Time section.

The guiding tension: in Newtonian physics both time and the extension of space are givens — absolute backdrops against which everything happens, the same for every observer. In special relativity neither is given; both are constructed, and what they are constructed from — clocks, rods, light signals, a convention of simultaneity — turns out to leave no room for a universal "now" or for a frame-independent length. A striking asymmetry sharpens the point: Newtonian physics had already made position and velocity relative, but kept length and simultaneity absolute; special relativity's genuinely new move is to relativize those too, so that spatial extension itself, and not merely spatial location, ceases to be an absolute. Units: $c$ is the speed of light, $\beta =v/c$, $\gamma =(1-{\beta }^2{)}^{-1/2}$; the metric is mostly-minus, matching the physics pages.

Four lenses run through what follows, and it helps to name them at the outset. The operational question — what procedure would actually assign a time or a length? — is answered by clocks, rods, light signals, and the SI units built on them. The mathematical question — what is the abstract structure? — is answered by Minkowski spacetime itself: a four-dimensional space with a signature-$(+,-,-,-)$ invariant interval, its Poincaré symmetry group, and the quantities that group leaves fixed (its full development lives on the Special Relativity pages). The conceptual question — what, then, are time and space, and how do the two concepts hang together? — yields time as the path-dependent length of a worldline, space as a frame-relative slice, and the pair as co-defined facets of one geometry. The metaphysical question — what does this show about what is real? — asks whether the loss of a universal "now" means that the passage of time is unreal and the future already settled.

The four are layered, not independent. The mathematical structure is the shared, interpretation-neutral core the others act on: the operational lens anchors it to rods and clocks, the conceptual lens interprets it, and the metaphysical lens is underdetermined by it — the very same formalism supports both Einstein's (block universe) and Lorentz's (hidden preferred frame) readings, and the wager that the structure simply is the ontology, collapsing the interpretive lenses into the mathematical one, is structural realism (see substantivalism and relationism). Operational facts in turn force the conceptual revision, and the conceptual result constrains but does not settle the metaphysical one. Accordingly this remark pursues the operational and conceptual together, question by question — using the mathematics only as their vehicle — reaches the metaphysical only at the end, at the collapse of the "now," and there hands it on to the Philosophy of Space and Time section, where it belongs.

From absolute space and time to operational space and time

Newton stated the classical view of time with unusual candor, and gave space a twin:

Absolute, true, and mathematical time, of itself, and from its own nature, flows equably without relation to anything external. … Absolute space, in its own nature, without relation to anything external, remains always similar and immovable.

On this picture time and space are containers: they exist independently of clocks and rods and events, are the same everywhere and for everyone, and ground absolute facts — which distant events are simultaneous, and how long a rod "really" is. A clock or a ruler does not define time or length; it merely tracks, more or less accurately, quantities that would be there anyway.

Einstein's method in the 1905 postulates is the reverse, and it is the philosophical hinge of the whole theory. He refuses to take time or space as given and asks instead: what operation would you actually perform to assign a time, or a length, to something? His answer for time — "the time of an event is the reading of a clock at the place of the event" — and its spatial counterpart — "the length of a body is the distance between simultaneous locations of its two ends" — replace the metaphysical containers with operational definitions. Time is what a clock reads; length is what a rod marks off between simultaneously-located ends. Everything strange in the theory follows from taking this operational reduction seriously and refusing to smuggle the containers back in.

What is time? Time as what a clock measures

If time is what a clock reads, then the fundamental temporal quantity is not a coordinate but proper time — the time elapsed along a particular worldline, as recorded by a clock carried along it. Its definition is purely geometric (spacetime):

$$c^{2}d{\tau }^2=d{s}^2=c^{2}d{t}^2-d{x}^2-d{y}^2-d{z}^2,\tau =\int d\tau =\int dt\sqrt{1-v^2/c^2}.$$

Three features of this formula rewrite the concept of time.

• Time is local, not global. Proper time is a property of a worldline, not of the universe. There is no single number "the time" that all clocks share; each clock accumulates its own elapsed time along its own path. "How much time has passed?" has no answer until you specify whose clock, along which route.
• Time is path-dependent. Proper time is the length of a timelike worldline in the Minkowski metric — literally an arc-length integral. And just as two roads between two cities can have different lengths, two worldlines between the same pair of events accumulate different proper times. This is the twin paradox, stripped of paradox: the twin who takes the "longer" (accelerated) route through spacetime is younger on return, not because anything went wrong with her clock but because elapsed time just is the length of the path, and her path was shorter in proper time. Time behaves like distance, not like a universal parameter.
• The inertial worldline is the longest. Of all timelike paths between two events, the straight (unaccelerated) one maximizes proper time. This inverts the Euclidean intuition (there the straight line is shortest) and is a direct consequence of the minus signs in the metric. Ageing is, quite precisely, a measure of how straight your history through spacetime is.

So the answer to "what is time?" in special relativity is: time is the proper length of a worldline — a quantity that clocks measure, that is real and frame-independent for each given worldline, but that is emphatically not a single quantity shared across the world.

What is space? Space as a frame-relative slice of spacetime

The symmetric question about space has a symmetric — and equally deflationary — answer. What we call "space" is a slice of spacetime at an instant: the set of all events an observer counts as happening now, a three-dimensional cross-section of the four-dimensional whole. But "happening now" is exactly simultaneity, and simultaneity is frame-relative. So the very act of carving "space" out of spacetime already smuggles in a choice of frame:

• Space is not a container but a projection. Just as "time" turned out to be one frame's way of parametrizing worldlines, "space" is one frame's way of slicing spacetime into simultaneity hyperplanes. Different inertial observers slice the same spacetime along differently tilted planes (Minkowski diagram), so they inhabit different spaces — different sets of events counted as coexisting. There is no unique, observer-independent decomposition of spacetime into "space" + "time."
• Spatial extension is no longer absolute. In Newton's picture the distance between two points, and the length of a rod, are the same for everyone. In special relativity they are not: a rod's measured length depends on the frame, because measuring it requires locating its ends simultaneously, and simultaneity is relative. The invariant analog of proper time is proper length — the length of a body in its own rest frame, or equivalently the spacelike-interval magnitude between the rest-frame-simultaneous locations of its ends. As with time, the objective quantity attaches to a definite structure (the rest frame / the interval), not to a universal "space."
• What is genuinely new here. It is worth being exact about SR's novelty, because part of the relativity of space is old news. Newtonian mechanics had already made position and velocity relative — absolute rest and absolute location are undetectable under Galilean invariance, a point Leibniz pressed against Newton long before. What Galilean physics kept absolute was length and simultaneity: Galilean transformations preserve both. Special relativity's genuinely new contribution about space is therefore not "location is relative" but that length itself becomes frame-dependent, and that the space/time split is frame-dependent — spatial extension, not merely spatial position, loses its absolute status.

So the answer to "what is space?" mirrors the answer for time: space is a frame-relative slice of spacetime, with proper length — not any universal distance — as the quantity that survives frame-independently.

How is time measured? Clocks, light, and the light clock

Because time is operational, how we measure it is not a side issue but part of what time is. Two ingredients do all the work.

A clock is any device with a reproducible periodic process — a pendulum, an atomic transition, or, in the theory's favorite idealization, a light clock: a photon bouncing between two mirrors, one "tick" per round trip. The light clock is decisive because it lets one derive time dilation from the constancy of $c$ alone, with no algebra (kinematics): set the clock moving transverse to the photon's path, and the photon must travel a longer diagonal while the mirror advances, so — since its speed is still $c$ — the moving clock ticks less often. Pythagoras on the diagonal gives $\Delta t=\gamma \Delta \tau$ immediately. Moving clocks run slow is not a defect of clocks but a theorem about how proper time accumulates, and it holds for every clock (else you could tell absolute motion by comparing two kinds of clock, violating the principle of relativity).

That time dilation is reciprocal — each observer sees the other's clock run slow — looks contradictory only if one still believes in absolute simultaneity. It dissolves the moment one accepts that the two observers are comparing different pairs of events, sliced by different planes of simultaneity (kinematics). The reciprocity is the surest sign that the frame-dependence of time is genuine and not a measurement artifact.

How is space measured? Rods, radar, and length contraction

Because length is operational, how we measure it is again part of what space is — and the method inherits the frame-dependence of simultaneity at its very first step. To measure the length of a moving rod you must mark the positions of its two ends at the same time and take the distance between the marks; mark them at different times and the rod will have moved, giving a meaningless answer. But "at the same time" is frame-relative, so different frames locate the ends by different simultaneity slices and get different lengths. Two operational procedures make this concrete:

• Rigid rods laid end to end define distance within a frame — the direct analog of clocks for time.
• Radar ranging defines distance by light: bounce a pulse off the far point and set the distance to $\frac{1}{2}c\Delta t_{\text{round-trip}}$. This uses only the constancy of $c$ and a local clock, and it exposes that the spatial metric, like distant simultaneity, rests on the same light-signal conventions — measuring a one-way distance already presupposes synchronized clocks at the two ends.

It is worth heading off a natural worry: doesn't laying out that lattice of rulers itself require synchronizing them? It requires coordinating them, but not in the temporal sense — rulers need calibration, not synchronization. A lattice at rest is static: nothing moves while you build it, so it needs no clocks at all. You assemble the rods end to end and check the grid is rigid at leisure, and you certify that they share a unit by comparison — bringing rods together and superposing them (spatial congruence). Synchronization enters only through the clocks hung at the lattice points, which the moving-rod measurement needs in order to fix "the two ends at the same time." The division of labor is clean: positions come from the (congruent) rulers, simultaneity from the (synchronized) clocks, and the length of a moving body needs both.

Yet the worry points at something real. To compare two rulers that are far apart one must transport one to the other and superpose — and whether a rod keeps its length under transport cannot be independently verified; it must be stipulated (that no undetectable "universal force" stretches every rod alike). So the spatial metric rests on a congruence convention in exactly the way distant time rests on a synchronization convention:

Within a single inertial frame, with ordinary rods and no differential forces, both conventions are so natural that we simply "read off" a length without noticing them. The philosophical fine print is that neither grid is handed to us by nature: the temporal grid needs the conventionality of simultaneity, and the spatial grid the twin conventionality of geometry (Reichenbach's and Grünbaum's "metrical amorphousness of the continuum") — an amorphousness the temporal continuum turns out to share, since a clock's own ticks need a convention of their own, as the next-but-one section shows.

The headline consequence is length contraction: a body of rest length $L_0$, measured in a frame in which it moves at $v$ along its length, comes out shorter,

$$L=\frac{L_0}{\gamma }\le L_0,$$

with transverse lengths unchanged. Crucially, this is not a physical squeezing or a force crushing the rod — nothing happens to the rod at all. It is a direct consequence of relativity of simultaneity: the two frames disagree about which pair of end-events counts as "the ends located at the same time," and so they measure different spatial gaps between the same worldlines (kinematics). Like time dilation, length contraction is reciprocal — each observer measures the other's rods as shortened — which would be contradictory under absolute simultaneity and is perfectly consistent once simultaneity is frame-relative. Length contraction and time dilation are thus not two independent oddities but two faces of the single fact that moving frames slice spacetime along tilted planes of simultaneity.

A note on the SI second and metre

None of this is a philosopher's idealization: the modern International System of Units defines the second and the metre in exactly the operational, relativistic terms of this remark — it has, in effect, turned special relativity's lesson into official metrology.

• The second is a proper-time definition. Since 1967 (with the value fixed exactly in the 2019 SI) the second is $9192631770$ periods of the caesium-133 hyperfine transition — a periodic process at one place, i.e. a clock. It therefore defines proper time, the quantity attached to a single worldline, and needs no synchronization or distant simultaneity: one caesium atom at rest ticks out its own seconds locally. The frame- and potential-dependence is not evaded but displayed — two caesium clocks in relative motion or at different gravitational potentials tick at genuinely different rates (optical clocks now resolve a height difference of centimetres). So the SI second is a proper second along each clock's path; any global time scale (TAI, or Terrestrial Time, defined to tick at the rate of a clock on the geoid) must combine clocks with explicit relativistic corrections and a synchronization convention — the coordinate-time step made real.
• The metre is now light-travel-time. Since 1983 the metre is the distance light travels in vacuum in $1/299792458$ of a second, which fixes $c$ by definition. This is the radar operationalization made official: length is reduced to time and light, $\text{length}=c\times \text{time}$. The congruence convention is thereby implemented as "equal length $=$ equal light-travel-time," and because $c$ is now defined, one can no longer measure it — an old "speed of light measurement" becomes a calibration of the metre. But fixing $c$ pins only the two-way speed; the one-way speed (isotropy) stays conventional, so realizing the metre over a distance still carries the simultaneity convention.

The deep point is that the SI has made the metre parasitic on the second through the constant $c$ — a postulate of special relativity — so that metrology now embodies the remark's "space and time as one": there is really one primitive standard, a local clock, plus $c$, from which both time and length are generated. The historical arc records the very shift this remark describes — the metre went from a transported platinum–iridium bar (1889: a rigid rod with its congruence convention) through a krypton wavelength (1960) to light-travel-time (1983): from rods to radar. And the split between the convention-light and convention-laden parts survives intact: a single caesium clock, or a there-and-back light flight timed by one clock, is frame-clean and nearly convention-free; it is only distant and global coordination — one-way light speed, world time scales, distant congruence — that carries the conventions.

Do clocks presuppose space?

A sharper worry runs the other way. We have made space lean on time (a metre is a light-travel-time) and on clocks (the moving-rod measurement needs synchronized clocks). But a clock is a periodic physical process — so does defining time by a clock secretly presuppose space after all? It does, and recognizing how reinforces the picture rather than breaking it.

• Every physical clock has spatial content. The light clock is the blatant case: its tick is $2L/c$, defined by a length $L$. But even the caesium standard hides one — its frequency $f=\Delta E/h$ is set by the atom's spatial and quantum structure (the electron wavefunction, the electron–nucleus interaction, ultimately the Bohr radius and the electron mass). A pendulum swings through space, a wave oscillates in space, a decay rate is fixed by nuclear structure. The bare mathematical notion of a period (a phase mod $2\pi$) is space-free, but no physical realization of it is. So "time is what a clock reads" does not reduce time to anything space-free.
• This is co-definition, not a vicious circle. Operationally the clock is a primitive standard — a black box emitting ticks you count; one does not derive the caesium frequency from a prior length but adopts the process as the unit (a coordinative definition). Conceptually, that space appears inside the definition of time — and time, as light-travel-time, inside the definition of space — is exactly what "space and time as one" leads one to expect. In spacetime the point is sharpest: proper time literally is a length, the arc-length $cd\tau =ds$ of a worldline. Neither concept reduces to the other; the failure to find a clean one-way reduction is Minkowski's unification showing through.
• Clocks carry a congruence convention too — isochrony. Comparing two separated rods needed a stipulation that transport preserves length (congruence). Its exact temporal twin: that a clock's successive ticks are equal — isochrony — cannot be checked directly, since two durations never coexist to be superposed. Which process counts as a uniform clock is itself a stipulation, so the temporal continuum is "metrically amorphous" in just the way the spatial one is. The two standards are thus symmetric all the way down:

So the intuition is right, and it completes the remark's symmetry: measuring space needs clocks and a congruence convention, and measuring time needs spatial structure and an isochrony convention. Space and time are interdependent in both directions — which is exactly why special relativity treats them not as two arenas but as one, the theme the invariant sections below make precise.

The collapse of the universal "now"

This is the hinge of the remark: the last operational-and-conceptual result, and the first metaphysical shock.

To assign times to distant events — not just to read a clock at the event's own location — one must synchronize separated clocks, and here the operational method exposes its deepest consequence. Einstein's synchronization: send a light pulse from $A$ at $t_A$, reflect it at $B$, receive it back at $t_A^{\prime }$, and define the reflection at $B$ to be simultaneous with the local time $t_B=\frac{1}{2}(t_A+t_A^{\prime })$ (postulates). This works — but it builds in the assumption that light takes equally long each way, which is itself a stipulation, since measuring the one-way speed of light would already require the synchronized clocks we are trying to establish. Simultaneity at a distance is thus not read off nature but fixed by convention — a point that becomes its own topic in the conventionality of simultaneity.

The result is the theory's defining break with common sense: two events simultaneous in one inertial frame are generally not simultaneous in another,

$$\Delta t^{\prime }=-\gamma \frac{v\Delta x}{c^2},$$

so observers in relative motion carve spacetime into different families of "now"-slices, and none is privileged. Einstein's embankment-and-train illustration (postulates) makes the moral vivid: two people passing each other genuinely disagree about which distant events are happening "now," and neither is mistaken. There is no observer-independent, worldwide present. The Newtonian "now" — the global instant slicing all of reality into settled past and open future — simply has no counterpart in the theory.

This is the single result with the heaviest metaphysical freight, because a metaphysics that privileges the present — presentism, or the idea that the passage of time is an objective advancing of a worldwide "now" — requires exactly the global simultaneity that relativity denies. The Rietdijk–Putnam argument turns this into a case for the tenseless block universe; the counter-moves (frame-relative existence, a hidden preferred foliation) and the full debate belong to the philosophy of space and time.

Space and time as one: the invariant behind the appearances

If time intervals and spatial lengths are each frame-dependent, is nothing objective? Minkowski's answer (1908) is that space and time separately "fade into shadows," but their union is absolute. What every observer agrees on is the spacetime interval:

$$\Delta s^2=c^2\Delta t^2-\Delta x^2-\Delta y^2-\Delta z^2,$$

invariant under all Lorentz transformations. The split of $\Delta s^2$ into "a time part" and "a space part" is like the split of a vector into components under a rotated coordinate system: frame-relative bookkeeping of a single frame-independent object. "Time" and "space" are not two separate arenas but two projections of one four-dimensional structure, and different observers project differently. This is why the theory is a theory of space and time together, and why questions about the nature of time cannot be separated from questions about the ontology of spacetime.

What remains objective, then, is not any observer's time but:

• Proper time $\tau$ along each worldline — a Lorentz scalar, the same in every frame (§ above);
• The causal / light-cone structure — whether two events are timelike, null, or spacelike separated is frame-independent (spacetime). For timelike- and null-separated pairs all frames agree on their order; only for spacelike pairs — the causally disconnected ones — does temporal order become frame-relative, and precisely there no signal can pass, so no paradox arises.

The light cone is thus the true, objective skeleton of time: not a global "now," but a local division at each event into its unambiguous future, its unambiguous past, and a causally-disconnected "elsewhere." Relativity does not abolish temporal order; it relativizes simultaneity while preserving causality, and it is causal structure, not a moving present, that carries whatever is objective about the direction and order of time.

What special relativity settles — and what it leaves open

It is worth being precise about the reach of the physics — what it fixes at the operational and conceptual levels, and what it leaves to the metaphysical — since it is easy to over-read.

What the theory establishes. There is no absolute time and no absolute space; no universal simultaneity, no single global time-rate, and no frame-independent length or space/time split. Time is operationally what clocks measure and space what rods (or radar) measure; proper time and proper length are path- and frame-relative in the sense that no global time or universal distance exists, while the objective invariants are proper time, proper length, the spacetime interval, and the causal structure. These are not interpretations — they are the content of the theory, tested exhaustively (tests: muon lifetimes, Hafele–Keating, the transverse Doppler effect, GPS).

What it leaves to philosophy. The theory strongly constrains but does not by itself settle the further, metaphysical questions:

• Does the absence of a global "now" mean the passage of time is unreal, and that past and future are as real as the present (the block universe, eternalism)? Or can an A-theorist save an objective present by relativizing existence to frames, or by positing a real-but-hidden preferred foliation? — see relativity and the reality of simultaneity.
• Is spacetime itself a substance or a system of relations? Minkowski geometry can be read either way — see substantivalism and relationism.
• Where does time's arrow come from, given that the theory's laws are time-symmetric? Special relativity supplies causal order but not a direction; the direction traces to thermodynamics, not to the Minkowski metric.

These are exactly the questions the Philosophy of Space and Time section takes up. The remark's point is narrower and prior: to show why special relativity makes them unavoidable. Once time is what a clock reads and space is what a rod marks off — and clocks read path-lengths through a geometry with no preferred slicing, while rods measure gaps across frame-dependent planes of simultaneity — the comfortable Newtonian picture of one time flowing everywhere at once and one space extended the same for all is gone, and the metaphysics of space and time has to be rebuilt on the invariants that survive.

Where this sits

This remark reads the Special Relativity pages — the postulates, the interval and proper time, and the kinematic effects — for their answer to what time and space are and how each is measured, and finds a conception on which both are operational, non-absolute, and inseparable: two frame-relative projections of one four-dimensional spacetime. It is the physical entry point to the Philosophy of Space and Time section, where the metaphysical questions it opens — the reality of the present, the block universe, the ontology of space and spacetime, and the direction of time — are pursued in full. It sits beside the section's other remark, Semantics and the Theory of Meaning, as conceptual fine print on a technical body of work.