If we could somehow rewind time to the earliest days of the Solar System and start over again, would life — and intelligence — reappear? It’s an experiment science fiction authors are able to try, but it defies real world science. Nonetheless, we can make approaches to the problem through the analysis of probabilities. In particular, we can use statistics, and the technique known as Bayesian inference, which weighs probabilities updated by new evidence.

This is a helpful exercise given that so often I hear people referring to the idea that intelligent life must be everywhere because the universe is so vast and there are so many opportunities for it to arise. But does life inevitably emerge on what we might consider habitable worlds?

What if this process of abiogenesis is rare? The question points to the fact that we have absolutely no idea what the likelihood is, and therefore assumptions about intelligent life based solely on numerical opportunity are nothing but speculations.

Enter Columbia University’s David Kipping, whose work has been featured often in these pages. Kipping tackles the question of the likelihood of life and the development of intelligence in a new paper in Proceedings of the National Academy of Sciences. He has a chronology to work with, one involving life’s earliest appearance as found in fossils, the emergence of humans, and the habitability constraints of our planet’s surface conditions, using it to draw inferences on how quickly life can arise, and how unusual intelligence may be.

The scenarios — derived as what in Bayesian terms are called ‘objective priors’ — are relatively straightforward, each of them worth examining in light of the fact that we have no observational evidence for life beyond the Earth. Our planet is our data point, and with all the disadvantages that produces, we can still draw inferences about life elsewhere from the constraints we can establish here. We know that abiogenesis is possible because we are here to write about it. But Kipping applies statistical methods based on Bayesian mathematics to consider the odds.

The first scenario (and the one I favor): Life is common, but rarely develops intelligence. I suspect we’re going to find evidence for simple life all over the Orion Arm as we extend our technologies outward, but little evidence for technologies. But other scenarios exist: Life is common and so is intelligence. And perhaps abiogenesis is rare. In that case, we may find life unusual but intelligence a common consequence when it does happen. Finally, life may be as rare as intelligence.

Image: Are we alone in the universe? A new study uses Bayesian statistics to weigh the likelihood of life and intelligence beyond our solar system. Credit: Shutterstock/Amanda Carden.

Which scenario to choose? Bayesian techniques involve testing a position against new evidence that can be applied to the question, which allows estimates to get better as they are refined. Bayesian mathematical formulae tackle how to model one scenario against another. And I found the result Kipping arrived at encouraging. Let me quote him on the matter:

“In Bayesian inference, prior probability distributions always need to be selected. But a key result here is that when one compares the rare-life versus common-life scenarios, the common-life scenario is always at least nine times more likely than the rare one.”

Drawing on our single data point — Earth — Kipping points out that we know life emerged quickly. We have to factor in the impact with the Mars-sized “Theia” some 4.51 billion years ago (leading to the formation of the Moon), but mineralogical evidence from zircons points to an atmosphere and liquid water present on Earth’s surface by roughly 4.4 billion years ago. The earliest evidence for life is found in 4.1 billion year old zircon deposits in the form of depleted carbon inclusions, a controversial datapoint, but undisputed evidence for life turns up in microfossils found in 3.465 billion year old rocks in western Australia.

We can come up, then, with the length of time for which Earth is expected to persist as habitable for intelligent beings, factoring in the growing luminosity of the Sun and the increased rate of weathering of silicate rocks on Earth and eventual depletion of carbon dioxide in the atmosphere. Kipping arrives at a habitable ‘window’ of 5.304 billion years. That’s from the beginning of life to its likely end, and it shows how significant is the question of how fast abiogenesis happens. If it takes too long, life would never emerge under the conditions most planets would face as their star continued to evolve. 900 million years from now, Earth will be a hostile place indeed.

But back to the key result — and I have to send the reader to the paper for the complex Bayesian mathematics involved — Kipping draws on a 2012 paper from Spiegel and Turner to refine the Bayesian formalism produced there for interpreting life’s early emergence on Earth. He considers it against the broader context of the habitable ‘window.’ From the Kipping paper:

The early emergence of life on Earth is naively interpreted as meaning that if we reran the tape, life would generally reappear quickly. But if the timescale for intelligence is long, then a quick start to life is simply a necessary byproduct of our existence—not evidence for a general rapid abiogenesis rate. Using our objective Bayesian framework, we show that the Bayes factor between a fast versus a slow abiogenesis scenario is at least a factor of 3—irrespective of the prior or the timescale for intelligence evolution. This factor is boosted to 9 when we replace the earliest microfossil evidence… with the more disputed 13C-depleted zircon deposits…

Thus the common life scenario gains odds, and markedly so. As for intelligence, Kipping’s analysis precludes the possibility that it emerges quickly (in less than billions of years), while the idea that intelligence is rare remains viable. Even so, he finds betting odds of only 3:2 that intelligence rarely emerges — this slight preference for rare intelligence is consistent with our lack of SETI results but leaves the question of searching for intelligence elsewhere wide open. Life is likely to emerge on other worlds, in other words, but our one data point — Earth — tells us that intelligence emerges only with time and difficulty. We have no outstanding way to choose here one way or another. Let me quote from the paper on this:

…our work supports an optimistic outlook for future searches for biosignatures…The slight preference for a rare intelligence scenario is consistent with a straightforward resolution to the Fermi paradox. However, our work says nothing about the lifetime of civilizations, and indeed the weight of evidence in favor of this scenario is sufficiently weak that searches for technosignatures should certainly be a component in observational campaigns seeking to resolve this grand mystery.

Life commonly found, the prevalence of intelligence still a mystery. Keep looking, but you now have some insight into where to place your chips in your next trip to the astrobiological casino.

If you’re interested in learning about Bayesian inference and the surprising successes of Bayesian analysis, I recommend Sharon McGrayne’s book The Theory That Would Not Die: How Bayes’ Rule Cracked the Enigma Code, Hunted Down Russian Submarines, and Emerged Triumphant from Two Centuries of Controversy (Yale University Press, 2011) as an excellent backgrounder.

The paper is Kipping, “An objective Bayesian analysis of life’s early start and our late arrival,” Proceedings of the National Academy of Sciences 18 May 2020 (full text). You can see Kipping’s lively video presentation on this work at https://www.youtube.com/watch?v=iLbbpRYRW5Y.