In 2049, an optical engineer named Ada Voss constructed a telescope that could observe things that had almost happened.

Its operating principle was never adequately explained.

The instrument resembled an ordinary astronomical telescope, except that its eyepiece was made of black glass and its lens had to be cooled to exactly 11 degrees below the temperature of the surrounding air.

The first observation was conducted on April 17.

Ada pointed the telescope toward the Moon.

She saw the Moon, but with an enormous ring system, brilliant and pale, stretching across the sky.

She checked the alignment, adjusted the focus, and looked again.

Now the Moon was gone. In its place was a cloud of dust, slowly coalescing into a second Earth.

After seventeen observations, she concluded that the instrument revealed possible histories of the universe.

At first, the scientific community was delighted.

Researchers observed worlds in which dinosaurs had survived, worlds in which Saturn had no rings, and worlds in which the gravitational constant was very slightly different.

The discoveries were astonishing but ultimately frustrating. There was no way to visit these worlds, communicate with them, or extract anything from them.

Then a graduate student made an unexpected observation.

He pointed the telescope at an empty chair.

In one possible history, the chair was occupied by a woman reading a newspaper.

In another, the chair had never been manufactured.

In a third, it was burning.

The student noticed that some possibilities appeared much more frequently than others.

He began recording them.

After several months, he discovered a remarkable regularity: the frequency with which a possible history appeared seemed to depend on how little the actual world would need to change to produce it.

The telescope was not showing random imaginary worlds.

It was showing nearby worlds.

Scientists began to use the instrument to study contingency. Historians examined events that had narrowly avoided happening. Biologists watched species that had almost evolved. Architects discovered buildings that had nearly been designed.

But the greatest discovery was made accidentally.

Ada had left the telescope pointing at a laboratory wall overnight.

In the morning, she looked through the eyepiece.

She saw the wall.

She looked again.

The wall.

Again.

The wall.

For the first time in the instrument’s history, no alternative appeared.

Ada spent weeks investigating.

Eventually she discovered that certain events were invariant across every accessible possibility. They happened in every nearby world.

The universe, it seemed, contained necessities.

Her first example was rather disappointing: a particular laboratory wall was always painted white.

But then she pointed the telescope at the ocean.

In every world, waves formed.

At a candle, and in every world the flame flickered.

At a tree, and in every world its branches divided.

The details differed, but certain patterns persisted.

She began compiling a catalogue of these invariants.

It eventually filled 600 volumes.

And yet the telescope’s final discovery was not among them.

Late one evening, Ada pointed it at the telescope itself.

Through the eyepiece, she saw a laboratory containing a different telescope, operated by a different scientist.

She looked again.

Another laboratory. Another instrument.

A third time.

A room containing no telescope at all.

She continued observing until dawn.

There were thousands of possible histories in which the telescope had never been built.

Thousands in which it had been constructed differently.

Thousands in which another person had discovered it.

But there was one feature common to every observation.

Somewhere, in every world, something was trying to understand the world it inhabited.

Ada closed her notebook.

For the first time in thirty years, she did not write down the result.