The Fermi Paradox: A Very Large, Very Quiet Universe

The story is that the physicist Enrico Fermi, during a lunchtime conversation about extraterrestrial life, asked a short question: where is everybody?

The question is more pointed than it first sounds, and it rests on a straightforward argument.

Our galaxy contains a very large number of stars, most with planets. The galaxy is billions of years old — old enough that any civilisation arising even slightly earlier than ours would have had enormous time to spread or leave traces. Even at speeds far below light, crossing the galaxy takes a few million years, which is brief on this timescale.

So if technological life is anything other than extremely rare, we should see evidence. We see none.

That gap between expectation and observation is the Fermi paradox. It is worth taking seriously not because we can resolve it, but because every candidate resolution says something significant — and some of them are uncomfortable.

What we actually know

The argument has several steps, and our confidence in them varies enormously. Being clear about which is which prevents most of the bad reasoning in this area.

Reasonably well established. Most stars have planets — this was uncertain a few decades ago and is now measured. A substantial number of those planets are in orbits where liquid water is possible. The raw ingredient count is high.

Almost entirely unknown. How often life actually begins on a suitable planet. How often simple life becomes complex. How often complex life becomes technological. How long technological civilisations persist.

The honest position is that these unknowns span many orders of magnitude, and the whole question turns on them. We have exactly one example of life arising, and you cannot estimate a rate from a single observation — particularly when that observation is made by the life in question, which guarantees it happened at least once wherever we are looking from.

This is why estimates of how many civilisations exist vary from essentially zero to millions depending on assumptions that nobody can currently constrain. The famous equation for calculating it is better understood as an organised list of our ignorance than as a calculation.

The candidate explanations

They sort into a few families.

Life is rare. Perhaps the step from chemistry to biology is extraordinarily improbable, or the step from simple cells to complex ones is. Earth took a very long time over that second transition, which some read as evidence it is difficult. If the improbable step is behind us, this is the reassuring answer: we are rare, and we are here.

Technological civilisations do not last. Perhaps they reliably end — through conflict, resource exhaustion, or some hazard that reliably emerges. This is the uncomfortable version, because it implies the filter is ahead of us rather than behind.

They exist and we cannot detect them. Our search has covered a small fraction of the possibilities, for a short time, using assumptions about what to look for that may be wrong. Absence of evidence here is genuinely weak evidence of absence — we have been listening for a few decades in a galaxy that is billions of years old.

They exist and are not interested in expanding or signalling. Expansion is an assumption drawn from our own history, and it may not generalise.

We are early. Star formation continues, and the universe will remain habitable for a very long time. Someone has to be among the first, and there is no rule against it being us.

The Great Filter framing unifies several of these: somewhere in the sequence from lifeless planet to visible civilisation, there is at least one step that almost nothing gets through. The critical unknown is whether that step is in our past or our future — and the same silence is consistent with both.

Why the reasoning is worth studying regardless

Even setting aside the astronomy, the paradox is an unusually clean case study in reasoning under deep uncertainty, and it illustrates several things well.

A conclusion that rests on unknown multiplied factors is not a conclusion. When five unknown probabilities are multiplied together, and each spans several orders of magnitude, the output spans more. Confident answers in either direction reveal assumptions rather than findings.

Absence of evidence depends entirely on how hard you looked. This is where the base rate and search-coverage questions do the real work, and where most casual arguments go wrong in both directions.

Selection effects are unavoidable when you are inside the sample. We could only be observing from a place where the improbable steps happened. That guarantees our own history looks survivable regardless of how rare it is — a structural version of survivorship bias that no amount of care removes.

Silence is data, but weak data. It rules out the most extreme scenarios — a galaxy densely populated with expansionist civilisations broadcasting loudly — and leaves nearly everything else open.

The honest summary is that we do not know, that the range of defensible positions is very wide, and that the most interesting versions of the question are about which step is hard rather than about the final answer. Fermi's question remains open, and the fact that it is still a good question seventy years later is itself informative about how little we have managed to narrow it.