Bet 44

Duration 18 years (02002-02020)

“By 2020, in real or artificial life, sustainable evolutionary progress in a genetically closed system will not have been demonstrated.”

Brig Klyce


Klyce's Argument

'Artificial life' describes computer models used to test evolutionary phenomena that are believed to occur in real life. 'Evolutionary progress' means an increase in organized complexity analogous to that which life appears to undergo as it advances from prokaryotes to [long list of intervening steps] to primates. A similar progress, made possible by human agents, is also evident in the world of personal computers, as [long list of] new features and capabilities become available. In both examples, the length of the necessary instruction set grows as the programs for new features and capabilities are added. 'Sustainable' means capable of continuing indefinitely. The symbiotic joining of two subsystems would not, by itself, exemplify sustainable progress. Neither would the assembly of a puzzle, because, for both examples, when the joining or assembly is complete, the process is over. On the other hand, if these processes are abetted by the invention of programs for new features or capabilities, with no impediment to the invention of other ones, the process appears to be sustainable. Invention is the key; joining and assembly are also necessary, but not sufficient, for evolutionary progress in a closed system. We acknowledge that it is possible to split hairs over the definition of 'progress'. However, as life advances from prokaryotes to primates, progress is intuitively clear. As the wager ages, we hope that wide interest in it will generate discussion that crystalizes the definition of progress. Meanwhile, one can readily imagine demonstrations that would so clearly exemplify progress that hair-splitting would not be necessary. For this reason, the wager is not premature. 'Genetically' means pertaining to genes or their analog, computer programs. Both real and artificial life use linear strings of a few standard symbols to convey instructions that are capable of being executed. A system is genetically closed if, after the experiment begins, no additional instructions in any form -- viruses, bacteria, plasmids; or computer viruses, subroutines, patches, etc. -- are admitted. 'Demonstrated' means clearly shown to occur in a repeatable experiment. Practically speaking, experiments will be much easier to repeat in a reasonable amount of time in artificial life than in real life.

In the standard account of evolution on Earth, programs for new features or capabilities are invented, with no impediment to the invention of other ones, in a closed system. However, the recent emergence of astrobiology alerts us that genetic material may arrive from space, so Earth's biosphere may not be genetically closed. And by horizontal gene transfer, genetic material may be installed far more readily than we used to think. Consequently, the history of life on Earth does not require the invention, on Earth, of programs for new features or capabilities. They might come from elsewhere. Of course, the usual version of the big bang theory requires the whole universe to be a completely closed system that began in a lifeless state a finite time ago. If so, the programs for new features or capabilities had to subsequently get invented in it. But other, non-finite cosmologies have been proposed. In any case, if the programs for new features or capabilities can be invented in a closed system, it should be possible to demonstrate this process. Until this demonstration is made, at least in a computer model, the accepted theory of evolution lacks an important confirmation.

I think that sustainable evolutionary progress in a genetically closed system, in real or artificial life is impossible. Instead, for real life, I like the strong version of panspermia. In it, highly evolved life has always existed, and the history of life on Earth exemplifies the local development of that life. But I am not trying to sell you on strong panspermia. For that see I only want to draw attention to an important test that the standard theory has not yet met. I don't think it ever will.


Klyce was challenged to a bet!

Brig Klyce is negotiating the terms of a bet about this prediction. It will soon be added to Bets on the Record.