What are the odds that humankind will survive long enough to colonize the Universe?
Katja Grace argues that the odds are low. Stripped of some nuance, her argument comes down to this:
- The fact that we’re around suggests that intelligent life is likely to be common.
- No other intelligent life appears to have colonized the Universe.
- If they haven’t succeeded, why should we?
By coincidence number one, I discovered Katja’s post (via a ringing endorsement from Robin Hanson) just hours after I’d posted yesterday’s entry here on The Big Questions disputing point 1). Of course, if point 1) fails then so does the entire argument.
So in response to Katja’s and Robin’s posts, I think it’s worth quoting a book review from the astrobiologist Charley Lineweaver, who I also quoted yesterday. Here is Lineweaver commenting on the “convergentist hypothesis”—that is, the hypothesis that evolution tends to converge on something like human intelligence:
[The convergentist hypothesis] is an appealing idea, but it has failed a series of exhaustive tests. It disagrees with the best data we have. A series of long-duration, independent and thorough experiments in evolution were set up and left to run. The most straightforward interpretation of the results is that human-like intelligence is not a convergent feature of evolution. There is no “intelligence niche” toward which animal species have a penchant to approach. In the absence of humans, other species do not converge on human-like intelligence as a generic solution, or even a specific solution to life’s challenges. These tests have been universally ignored.
The names of these tests are South America, Australia, North America, Madagascar, and India…For landlocked species, these continents that drifted independently of each other for between 50 and 200 million years were crucial experiments for evolution.
The time scale for tripling the size of the human brain in Africa was about 2-3 million years, while the time scale of the experiments was 50-200 million years. Thus, the experimenters were conservative and ran the tests 10-100 times longer than was necessary.
…
Five continents and millions of species evolving over tens or hundreds of millions of years are yelling at us upwind against our vanity: “…Human intelligence is not a convergent feature of evolution.” Rather, it is a species-specific trait—like the beautiful yellow crest of a sulfur-crested cockatoo.
Now, by coincidence number two, I happen to have had lunch with Robin Hanson yesterday, and I ran this argument by him. Robin is unimpressed. He doubts that five experiments are enough to tell us very much about what might happen on 1022 planets. In Robin’s words, “all this shows us is that the likelihood of evolving intelligence is less than about one in six”. He’s willing to take those odds.
I am sympathetic to Lineweaver’s view that we should no more expect to find extraterrestrial intelligence than extraterrestrial basketball. I am extremely sympathetic to his view that after six experiments, we essentially know for sure that evolution has no tendency to produce intelligence consistently. But I’m also sympathetic to Robin’s view that in the vastness of the Universe, six experiments can’t tell us much about what we might find. What else am I missing?
