Two stopwatches start from different readings. One ends at 12, the other at 112. Did the same event last longer for the second? The final readings alone cannot tell us.
In physical measurement, time lets us label events and compare the durations between them. A clock does this using a reproducible process and a counter. We need more than the number displayed: the starting point, unit and clock rate matter too. This is a way to test statements about time, not a final answer to every philosophical question about its nature. Einstein Online: connecting time with events and clocks.
An instant and a duration answer different questions
A lamp lights up: an event to which we can assign a place and an instant. It goes out: another event. “When did it light up?” needs a label on a time scale. “How long did it shine?” needs an interval, the duration between start and finish.
A calendar clock helps arrange a meeting; a stopwatch compares two experiments’ durations. We can change the starting label without changing the interval itself. Adding 100 to both readings of one clock leaves their difference unchanged. NIST: time, time interval and time of day.
Compare duration with repetitions
A pendulum returns to a particular position; quartz oscillates in an electrical circuit. A clock counts repetitions and turns them into readings. The process must be sufficiently regular, and its rate known through comparison with a reference. Frequency is the number of complete cycles per unit time; 2 hertz means two cycles per second. NIST: how clocks work.
Compare the three hypothetical counters. They share the same start and end events. The middle one has a different initial reading; the right one counts twice as many cycles per second.
One interval, three counters
Starts at zero
- Frequency
- 2 Hz
- Start and end readings
- 0 → 12
- Duration
- 6 s
A different initial reading
- Frequency
- 2 Hz
- Start and end readings
- 100 → 112
- Duration
- 6 s
Twice as many cycles per second
- Frequency
- 4 Hz
- Start and end readings
- 0 → 24
- Duration
- 6 s
More counted cycles do not by themselves mean a longer duration. Shorter cycles fit more often into the same interval. Likewise, 6000 milliseconds and 6 seconds express one duration, not two different courses of events.
Branch · One calculationOptional · safe to skip
From a count to seconds
For an ideal counter with constant frequency f:
Δt = (ΔN)/(f).
ΔN is the cycle count between events, f is measured in hertz, or (s)^( − 1), and Δt is in seconds. The first two counters add 12 cycles at 2 Hz; the third adds 24 at 4 Hz:
(12)/(2) = (24)/(4) = 6 s.
For the middle counter, 112 − 100 = 12: the initial offset cancels. We assumed an exact constant frequency and aligned start/end counting triggers. Real instruments have frequency errors, finite resolution and triggering uncertainty. NIST: measuring an interval with a counter.
An atomic transition provides a shared second
If everyone chose a different “tick,” comparing laboratory results would be difficult. We need a shared unit. The SI second is currently defined through the frequency of a particular transition between two hyperfine levels of the unperturbed ground state of cesium-133. That frequency has the exact value 9 192 631 770 Hz. A second therefore corresponds to that many periods of the radiation associated with the transition. BIPM: the current definition of the second.
This is not an atomic ball mechanically knocking billions of times. An atomic clock’s oscillator produces electromagnetic radiation. The atoms’ response to its frequency guides tuning near resonance, and the cycles are counted. The atomic transition supplies a reproducible rate reference. Real devices still require perturbations and uncertainty to be accounted for: an exact unit definition does not make every clock perfect. NIST: what actually “ticks” in an atomic clock.
Matching readings does not guarantee matching rates
We can set a clock’s hands to the correct reading without curing its habit of running fast. Two sound clocks may also run at the same rate while starting from different readings. Synchronizing readings and checking frequency are related but different tasks. NIST: time offset, NIST: frequency offset and stability.
Distant clocks introduce another problem: a time message must travel between them. We must account for the signal path and the rule used to compare events in different places. Very precise comparisons also require the clocks’ motion and gravitational conditions. Sharing a unit called a second does not automatically establish one universal clock for all observers. Einstein Online: local clock time and time coordinates.
Branch · Measurement uncertaintyOptional · safe to skip
A clock can run fast very steadily
Stability describes how much the rate varies over a specified interval. It does not guarantee agreement with the required reference. A device may count slightly shorter “seconds” very regularly and accumulate an offset.
To distinguish a constant reading offset from an incorrect rate, compare clocks more than once. If their reading difference grows over time, their rates differ. Report the result with its uncertainty and comparison conditions. NIST: frequency offset and frequency stability.
What does the number on a clock mean?
Predict: between the same events, a counter changes from 100 to 112. We reset its initial reading to zero without changing frequency. Does the interval become shorter? Would simply installing a larger digital display improve the measurement?
No. The first change moves the starting label; the difference remains 12 cycles. The second provides room for more digits but improves neither regularity nor knowledge of frequency. More decimal places do not prove lower uncertainty.
We measure time to compare durations and coordinate events, from repeating an experiment to communication and navigation. When a result is given in seconds, ask: between which events, using which clock, and with what uncertainty? Those details turn a number into a physical measurement. NIST: why precise time and frequency matter.