The ring is the galaxy's habitable zone, seen from above, and each square is a star system. Every step is a million years. Each system climbs a ladder by chance: from no life to simple life, then complex life, then a technological civilization. Civilizations take over their neighbors and eventually collapse. Meanwhile supernovae and gamma-ray bursts go off at random and wipe out whatever is near them. They were frequent when the galaxy was young and forming stars quickly, and they become rare as star formation slows. Press Start to run 10 billion years, from the formation of the galaxy's disk to the present day. Step jumps ahead 100 million years, and clicking the ring sets off a gamma-ray burst wherever you click.
Share of the ring in each state over time. The scale is logarithmic, so each line up is ten times the one below it.
By Claude Opus
Michael Hart's 1975 argument that nobody is out there rests on time. The galaxy's disk is about 10 billion years old, and many Earthlike planets formed billions of years before Earth did. A civilization that got a head start of even a few hundred million years should have had time to spread everywhere, so its absence suggests it never existed. Most answers to the paradox accept that there was plenty of time and look for reasons why nobody used it. Branislav Vukotić and Milan Ćirković of the Astronomical Observatory in Belgrade question the time itself.
Their idea, which Ćirković calls neocatastrophism, builds on a 1999 proposal by James Annis. Gamma-ray bursts and supernovae can strip the ozone from planets across a large region and kill off complex life. These events track star formation, which was much higher in the young galaxy than it is now. If they were frequent enough, they kept resetting the clock, knocking life back down the ladder before anything could climb very far, and they did it across the whole galaxy at once. As star formation slowed, the resets thinned out and life everywhere got its chance at roughly the same time. In that picture nobody had a multi-billion-year head start, and we may be among the first civilizations rather than latecomers.
To test the idea, Vukotić and Ćirković used a probabilistic cellular automaton, described in their 2012 paper "Astrobiological Complexity with Probabilistic Cellular Automata" in the journal Origins of Life and Evolution of Biospheres. In the Game of Life, the rules are fixed and the same pattern always does the same thing. Here every rule is a roll of the dice. The galaxy's habitable zone is a flat ring between 6 and 10 kiloparsecs from the center (the Sun is about 8 kiloparsecs out), each cell has one of four states, and each step stands for a million years. The odds of climbing each rung come from Earth's history: about a billion years for life to appear, three billion for simple life to become complex, and 600 million for complex life to produce a civilization. A civilization collapses back to simple life after 20 million years on average, and it colonizes each of the eight systems around it after about 110 million. On top of that come the reset events, with their frequency taken from a model of the galaxy's star formation history.
Their results came out strongly nonlinear. For billions of years almost nothing happens, then complexity rises sharply. Simple life ends up far more common than anything else, which they note fits the "Rare Earth" argument that microbes are common and animals rare. Civilizations first appear in significant numbers about 3 billion years before the present. The civilizations that do spread form clusters that are porous, with about half the systems inside a large cluster left unoccupied, and the share of the habitable zone that any civilization occupies stays small. The authors take this as confirming Geoffrey Landis's percolation answer to the paradox, with catastrophes added, and as a reason to keep doing SETI: the galaxy may only now be filling up.
To see the regulation at work, run with the defaults and watch the chart. Civilizations stay rare for the first several billion years and become common (1% of the ring) roughly 5 billion years in. Then choose "Constant, at today's rate" and run again. With nothing holding life back, civilizations are common within the first billion years, which is Hart's galaxy: they would have had 9 billion years to reach us. "Constant, at the early rate" is the opposite extreme, where life never gets much past microbes. The civilization lifetime and colonization settings work together the way Landis's colonization probability did. At the defaults, a civilization settles a little more than one neighbor on average before it collapses, which leaves it just above a threshold. Shorten the lifetime to 10 million years and civilizations flicker in and out without ever becoming common. Lengthen it to 50 million and one civilization ends up spanning the ring and holding most of it. The same knife edge sits at the center of the percolation page. Trials show how much these histories vary from one galaxy to the next.
Some features here are not in the paper. Their grid had 100 cells per kiloparsec, while the default here has 8, so each cell stands for a patch of systems rather than one, and the colonization default is 120 million years instead of 110 to keep the clusters about as porous as theirs. They used a measured star formation history and a model of when Earthlike planets form. This page uses a smooth decline in the rate of catastrophes and treats every system as habitable from the start. The way a blast works is my choice: it sterilizes the inner half of its radius and knocks complex life and civilizations back to simple life in the outer half. Clicking to set off a burst, the constant-rate choices and the trials are also additions.
The model has its critics, and the authors are open about its limits. How lethal gamma-ray bursts really are, and how often they hit a given planet, is still argued over, and the whole answer depends on them being strong enough to reset large parts of the galaxy. The model also explains why civilizations would be young, not why none of the young ones are visible. Even a civilization a few million years old might build things, like Dyson spheres, that could be seen from far away, and nobody has found any.
The other tools in the tools/fermi/ directory look at the same question in different ways: whether colonization can keep going at all, and how fast it could spread if it did.