Chapter 12: RNA World to the Rescue?
Is RNA a Great Savior?
The RNA World is supposed to solve two separate problems: making protein before DNA exists, and then handing its stored information over to DNA once DNA arrives. Given everything just shown about RNA's apparent inability to do sustained mechanical work, it's worth asking directly whether the first of those two claims actually holds up. Can RNA make protein at all?
Start by being as generous as possible. Set aside, for a moment, any doubts about whether amino acids could have existed on the early Earth in any quantity — grant the RNA World a full supply, sitting ready and waiting.
Even then, the amino acids that form under laboratory conditions meant to mimic the early Earth don't arrive sorted and labeled. They show up mixed together with a great deal of chemical debris — other amino acids no living thing uses, and assorted byproducts alongside them. Nothing about that mixture comes pre-filtered down to the twenty amino acids life actually depends on. Which means the RNA World needs some way of picking the right twenty out of that crowd — and no such mechanism has ever been demonstrated. Researchers working on exactly this question describe attempts at joining an amino acid onto an RNA molecule without any enzyme's help as showing little ability to prefer one amino acid, or even one chemical handedness of an amino acid, over another. As one recent review of the field put it plainly, there is no widely accepted explanation for how this kind of selective pairing could have happened without protein enzymes doing the choosing.
But set that aside too, and grant RNA the full set of correctly chosen amino acids besides. The next question is more basic still: what is actually supposed to do the joining? If the claim is that RNA itself joins one amino acid to the next, over and over, that runs headlong into everything just described about RNA's own toolkit — stringing amino acids together, one after another, in a particular order, is exactly the kind of repeated, precisely sequenced task nothing in RNA's demonstrated abilities suggests it can manage unassisted. And if RNA isn't doing the joining, then it's not clear what the RNA World has actually contributed at that point, since the joining would have to be happening some other way entirely — the very same way it would have had to happen if the RNA World had never been proposed at all.
Suppose, generously again, that some version of joining manages to happen anyway. The amino acids studied under these exact laboratory conditions don't line up into anything — they react with each other more or less at random, producing what the researchers themselves call, without any hedging, random peptides. Nothing is selecting a useful sequence out of that process, because nothing yet exists to recognize one sequence as more useful than another. There's no code linking a particular position on an RNA strand to a particular amino acid — that correspondence is precisely what a genetic code is, and a genetic code is exactly the sort of elaborate, coordinated system this whole story is supposed to explain the origin of, not something it can already assume as a starting tool.
Suppose, past every generous assumption already made, that one of these random chains happened to fold up into something that actually did something useful. What, exactly, has been gained? A single working protein, produced once, by chance, is not a foothold — it's a dead end. RNA, on this telling, has no code yet linking its own sequence to that protein's sequence, no template, no record of what it just made. It cannot make that same protein again, deliberately, tomorrow, and it has no way to pass along whatever accident just produced it. A useful result that can't be repeated or inherited isn't a step toward anything. It's a coincidence, indistinguishable, once it's happened, from having never happened at all.
Last revised 2026-11-18