Darwin was right to hedge his bets, but today we are pretty certain that all living creatures on this planet are descended from a single ancestor. The evidence, as we saw in Chapter 10, is that the genetic code is universal, all but identical across animals, plants, fungi, bacteria, archaea and viruses. The 64-word dictionary, by which three-letter DNA words are translated into twenty amino acids and one punctuation mark, which means 'start reading here' or 'stop reading here', is the same 64-word dictionary wherever you look in the living kingdoms (with one or two exceptions too minor to undermine the generalization). If, say, some weird, anomalous microbes called the harumscaryotes were discovered, which didn't use DNA at all, or didn't use proteins, or used proteins but strung them together from a different set of amino acids from the familiar twenty, or which used DNA but not a triplet code, or a triplet code but not the same 64-word dictionary - if any of these conditions were met, we might suggest that life had originated twice: once for the harumscaryotes and once for the rest of life. For all Darwin knew - indeed, for all anyone knew before the discovery of DNA - some existing creatures might have had the properties I have here attributed to the harumscaryotes, in which case his 'into a few forms' would have been justified.
Is it possible that two independent origins of life could both have hit upon the same 64-word code? Very unlikely. For that to be plausible, the existing code would have to have strong advantages over alternative codes, and there would have to be a gradual ramp of improvement towards it, a ramp for natural selection to climb up. Both these conditions are improbable. Francis Crick early suggested that the genetic code is a 'frozen accident', which, once in place, was difficult or impossible to change. The reasoning is interesting. Any mutation in the genetic code itself (as opposed to mutations in the genes that it encodes) would have an instantly catastrophic effect, not just in one place but throughout the whole organism. If any word in the 64-word dictionary changed its meaning, so that it came to specify a different amino acid, just about every protein in the body would instantaneously change, probably in many places along its length. Unlike an ordinary mutation, which might, say, slightly lengthen a leg, shorten a wing or darken an eye, a change in the genetic code would change everything at once, all over the body, and this would spell disaster. Various theorists have come up with ingenious suggestions for special ways in which the genetic code might evolve: ways in which, to quote one of their papers, the frozen accident might be 'thawed'. Interesting as these are, I think it is all but certain that every living creature whose genetic code has been looked at is descended from one common ancestor. No matter how elaborate and different the high-level programs that underlie the various life forms, all are, at bottom, written in the same machine language.
Of course we cannot rule out the possibility that other machine languages may have arisen in yet other creatures that are now extinct - the equivalent of my harumscaryotes. And the physicist Paul Davies has made the reasonable point that we haven't actually looked very hard to see if there are any harumscaryotes (he doesn't use the word, of course) that are not extinct but still lurking in some extreme redoubt of our planet. He admits that it is not very likely, but argues - somewhat along the lines of the man who searches for his keys under a street lamp rather than where he lost them - that it is a lot easier and cheaper to look thoroughly on our planet than to travel to other planets and look there. Meanwhile, I don't mind recording my private expectation that Professor Davies won't find anything, and that all surviving life forms on this planet use the same machine code and are all descended from a single ancestor.
Was this article helpful?