Relativity
-
Day 25: There Is No Now
Week 5 starts today. For four weeks we’ve been poking at how humans organize time: clocks, calendars, time zones, leap seconds, DST, software bugs, political fights. All of it rests on a quiet assumption. There’s an objective fact about what time it is right now, and the only real question is how cleanly we agree on it.
This week I want to break assumptions and today the thing doing the breaking is Einstein.
The starting point is the conclusion, up front. There is no universal “now." Two observers, depending on how they’re moving and where they sit in a gravitational field, will not agree about what time it is at some distant event. Neither one is wrong, it’s just that the universe does not have a “true” clock.
In order to understand how the theory of relativity impacts time, we should start at the beginning.
1905: the relativity of simultaneity
In 1905, in his off-hours from a job at the Swiss patent office, Einstein published On the Electrodynamics of Moving Bodies. Two postulates:
- The laws of physics are the same in all inertial reference frames.
- The speed of light in vacuum is the same in all of them, no matter how the source or observer is moving.
The speed of light being a constant is a given today but what does that really mean?
Light doesn’t add to or subtract from your motion. Turn on your headlights at 30 mph and the light doesn’t travel at c + 30. It travels at c. Point a flashlight forward from a ship moving at half the speed of light, and the beam still goes at c. Not 1.5c. Just c.
You already accept a small version of this. A passing siren rises in pitch as it comes toward you and drops as it goes, the Doppler effect, even though the siren itself never changes. Motion changes what you measure. Relativity pushes that idea somewhere stranger: because light’s speed won’t budge, it isn’t just pitch that shifts with your motion. It’s time.
If you are observing two events that happen simultaneously, your conclusion will depend on how you’re moving relative to them. Another observer also measuring carefully and subtracting the time the light took to reach them will disagree on whether or not the two events actually happened at the same time. Neither you nor they have made a mistake, but rather the motion, the speed of the observer, changes your reference time.
That’s the relativity of simultaneity, and it’s the single most important fact about time in modern physics. There’s no universal clock marking off “now” across space.
A second consequence is time dilation. Clocks moving relative to you tick slower than your own. How much slower is set by a number called the Lorentz factor. At everyday speeds it’s so close to one that you never notice. As you approach the speed of light it balloons, and the moving clock slows dramatically. This isn’t a measurement artifact. Particle accelerators watch it happen every day: unstable particles pushed near c take measurably longer to decay, by exactly the amount the Lorentz factor predicts.
1915: gravity slows clocks too
Ten years later, general relativity explained how gravity warps spacetime. The whole theory is a beast, but what we care about here is a single consequence:
Clocks deeper in a gravitational well run slower than clocks higher up.
A clock at sea level runs slower than one on Everest. A clock on Earth’s surface runs slower than one on the ISS. A clock just outside a black hole’s event horizon runs enormously slower than one far from any star.
Pound and Rebka confirmed this in 1959 with gamma rays sent down a 22.5-meter tower at Harvard. In 1971, Hafele and Keating flew atomic clocks around the world on commercial jets and compared them to clocks at the US Naval Observatory. The flying clocks came back off by amounts matching both special and general relativity at once.
You rely on this every day. GPS satellites sit higher in Earth’s gravity and move fast, and the two relativistic effects don’t cancel: net, their clocks run about 38 microseconds per day fast. Leave it uncorrected and your position drifts by roughly 10 km a day, useless within the hour. The correction is baked into the firmware. I went deep on exactly how this works back in Day 11: What If We Put Clocks in Space?, so I won’t re-derive it here. The point for today is just that relativity isn’t an abstract theory. It’s running in production, in your pocket.
So what can anyone agree on?
So if observers can’t agree on the time at distant events, what can they agree on? Their own clocks. The time that ticks along your own path through spacetime, what physicists call proper time, is the one measurement everyone computes the same way. Past that, the only thing they share is the disagreement itself. Time is relative, and that’s not a glitch to be fixed.
Picture someone on Mars right now. You each have a perfectly good “now” on your own wristwatch, but there’s no shared “now” linking them, and no measurement either of you can run to pin one down. You can agree on a convention and subtract the light delay, but the universe never marks which slice of Mars-time counts as your “now.” Your “now” is just one slice through 4D spacetime, and nothing in the geometry says it’s the right one. That’s the doorway to the block universe, the idea that past, present, and future all exist equally. It’s an interpretation, not settled physics, and I went down that rabbit hole back in Day 3.
The rest of the series
Every problem from the last four weeks, leap seconds, DST, time zones, calendars, assumes there’s an objective “what time is it” we’re trying to nail down. Relativity says that even at the level of fundamental physics, there isn’t one. The closest thing we have is a coordinate convention agreed on by one community of observers, and by its nature, that convention is “limited”.
Tomorrow: we are going to talk about a problem playing out right now on Earth. UT1, time defined by Earth’s rotation, and TAI, time defined by atomic clocks, have been drifting apart since 1972, and the gap is forcing a decision we keep postponing.
Sources
- Special Relativity & Time Dilation: Einstein, Albert. “On the Electrodynamics of Moving Bodies.” Annalen der Physik, 1905.
- Special Relativity (lecture notes): R. Taylor, University of Oxford, Department of Physics — simultaneity, time dilation, and the Lorentz factor.
- The Pound-Rebka Experiment: Pound, R. V.; Rebka Jr. G. A. “Apparent Weight of Photons.” Physical Review Letters, 1960.
- The Hafele-Keating Experiment: Hafele, J. C.; Keating, R. E. “Around-the-World Atomic Clocks.” Science, 1972.
- Relativity in GPS: Ashby, Neil. “Relativity in the Global Positioning System.” Living Reviews in Relativity, 2003.
- The Block Universe (Eternalism): Putnam, Hilary. “Time and Physical Geometry.” The Journal of Philosophy, 1967.
I’d appreciate a follow. You can subscribe with your email below. The emails go out once a week, or you can find me on Mastodon at @[email protected].
-
Day 11: What If We Put Clocks in Space?
In 1977, three years before GPS launched, the engineers building the satellites had to make a decision.
The clocks they were about to put in orbit were going to run faster than the clocks on the ground. By about 38 microseconds per day.
That sounds like nothing, but over 24 hours of GPS operation, an uncorrected clock would put you 11 kilometers off your actual position.
They had two options:
- Adjust the time signal on the ground, applying the correction as the data came back down.
- Pre-tune the clocks on the satellites to run slow by exactly the right amount, so that by the time relativity sped them back up, they’d tick at the right rate.
GPS chose option two.
They built the clocks to run at 10.22999999543 MHz instead of the nominal 10.23 MHz, so that orbital relativity speeds them up to ~10.23 MHz by the time the signal hits your phone.
The correction is baked in.
That’s what putting clocks in space looks like. One decision, and now everyone on Earth gets both navigation and time from the same signal.
This post is about the impact of that decision.
Why Clocks in Orbit Run Faster
Two relativistic effects act on a GPS satellite clock, and they push in opposite directions.
Special relativity slows the satellite clock down because it’s moving fast. General relativity speeds it up because it sits in weaker gravity than the ground. Gravity wins. Net result: the satellite clock gains about 38 microseconds per day.
Sounds like nothing. But uncorrected, that 38 microseconds drifts your GPS position by 11 km in 24 hours. Within a day of launch, GPS would be useless for anything more precise than “are you in the right country.”
This was known before launch. It was tested. It works.
What GPS Time Actually Is
GPS time is its own scale, started at midnight on January 6, 1980, and ticking continuously since. No leap seconds. No time zones.
The relationship to the other scales is fixed and simple:
GPS = TAI − 19 seconds (constant since launch) GPS = UTC + 18 seconds (today)GPS−TAI never changes. GPS−UTC grows every time UTC gets a leap second, and freezes after the 2035 leap-second abolition.
It is, in every meaningful sense, the most accurate clock in your daily life. And you’ve never seen it.
What It’s Used For
GPS time runs almost everything that needs precise timing in modern civilization, but it’s invisible because nobody consumes it directly.
- Finance. US and EU regulators (MiFID II, SEC) require trading firms to timestamp orders to microsecond precision. GPS-disciplined oscillators are how.
- Telecom. Cellular base stations need their carrier frequencies aligned across the network. GPS clocks them. Without GPS, your phone would struggle to hand off between towers.
- Power grid. Phasor Measurement Units monitor the AC waveform across the entire grid, synchronized to GPS. This is how grid operators detect instabilities before they cascade into blackouts.
- Datacenters. Stratum-1 NTP servers are typically GPS-disciplined. Every clock you’ve ever checked on a computer ultimately traces back, through several layers of network sync, to a GPS receiver in someone’s rack.
- Aviation, surveying, autonomous vehicles, drones, scientific instruments, particle physics. Anything built since 1995 that needs accurate timing or positioning, which is essentially everything.
The civilian world runs on GPS time. It just doesn’t admit it.
What It Didn’t Solve
Putting clocks in space solved navigation.
It did not solve timekeeping.
Your watch is still on local time. Your calendar uses civic dates with leap seconds buried in the UTC. You’re reading a clock face anchored to a Roman calendar, a Babylonian 24-hour day, and an Earth rotation that nobody can predict.
GPS time is great if you are a satellite, a financial trader, a power-grid engineer, a fighter jet, or a cell tower.
It is not great if you are trying to know what time to pick up your kid from school.
For that, you still need wall time, which still needs UTC, which still needs leap seconds, which still needs Earth’s wobbling rotation.
We built absurdly precise atomic clocks. We launched them into orbit. We baked relativity corrections into the silicon. We covered the planet in time signals accurate to nanoseconds.
And your meeting is still at 3 PM on Tuesday.
GPS quietly handles the part it needs to handle. But all of this assumes you’re on Earth.
Where This Goes
Earth orbit needs relativity corrections. The Moon needs more. Mars needs different ones still.
The further you get from Earth, the more “GPS-style time” stops being a solution.
Tomorrow: if an hour is an Earth measurement, so how do you tell time on a planet that doesn’t have them?
Sources
- Error analysis for the Global Positioning System — Wikipedia
- GPS time — Wikipedia
- Schriever Space Force Base — Wikipedia
- Phasor measurement unit — Wikipedia
I’d appreciate a follow. You can subscribe with your email below. The emails go out once a week, or you can find me on Mastodon at @[email protected].
Science Infrastructure 30daysoftime Timekeeping Gps Relativity