Evolution and Arithmetic

Chapter 6: The Monkeys That Didn't

← Believing the HeadlineImpossible Is Not Very Unlikely →

Not a Chance

The third lesson concerns what the more skeptical mathematician quoted in the same coverage actually said, and how it landed. Dr. Ian Stewart, emeritus professor of mathematics at the University of Warwick, was quoted with a flat "Not a chance." The article then summarized what his calculations suggested: that it would take "far, far longer than the age of the Universe" for random typing to produce a flawless copy of the complete works — 3,695,990 characters, all told. And it went on: "along the way there would be untold numbers of attempts with one character wrong; even more with two wrong, and so on. Almost all other books, being shorter, would appear (countless times) before Shakespeare did."

On first read, this sounds like exactly the sober corrective this book keeps asking for. Look closely, though, and it turns out to be its own kind of misleading — not because the "no" is wrong, but because of what "far, far longer than the age of the Universe" actually implies to anyone reading it.

Do the arithmetic yourself and the phrase falls apart. I tried, and my computer blew all its gaskets. Remember the odometer from Chapter 2: a single three-hundred-letter string — one sonnet's worth — needs a three-hundred-drum odometer with the full alphabet on every drum, which works out to the same scale as an ordinary ten-digit counting odometer with four hundred twenty-five drums. Getting our heads around a number that size already required blowing up the universe itself, more than once, and converting the pieces into computers, just to have something to point to. The complete works of Shakespeare isn't a bigger version of that same problem. It's a different problem in kind. Matching all 3,695,990 characters means an odometer running past five million drums — not a harder version of four hundred twenty-five, but a number for which four hundred twenty-five might as well be zero. (The arithmetic: Shakespeare on the Odometer.) Something that takes over five million drums even to count its own possible combinations isn't a thing that merely takes a long time to land on by chance. It's a thing that doesn't happen. Calling that "far, far longer than the age of the Universe" isn't a stretch of the truth. It's a different kind of claim wearing the same clothes.

That's the real problem with the phrase. "Far, far longer than the age of the Universe" sounds like an answer that's still in the same neighborhood as numbers we have some intuitive grip on — big, but nameable, still recognizably a length of time. It isn't. It's a way of describing something outside any physically meaningful timescale using words that make it sound like it's merely outside a very large one. Technically true. Not remotely truthful.

And the second half of that account compounds the problem rather than clarifying it. Saying that shorter books would appear "countless times" before Shakespeare's complete works did is true only in the narrowest, most technical sense — and false in every sense that matters to a reader trying to picture what's actually being described. A slim book might only need, say, a hundred-thousand-drum odometer rather than five million. That sounds like real, meaningful daylight between the two. It isn't. Both numbers are already so far past the point where blowing up the universe stops being enough to describe them that the gap between them carries no real information at all — there's nothing left to point to, no bigger explosion to reach for, that would make one feel any less impossible than the other. Being "less astronomically impossible" doesn't move a hundred thousand drums into the neighborhood of the achievable. Presented without that qualification, the line does real work in the wrong direction: it invites the reader to picture a long but survivable slog through countless completed books on the way to the real prize, rather than a void so complete that not one flawless book of any length would ever actually turn up.

Which raises its own question, aimed less at the professor than at the reporter holding the notebook. Didn't it occur to the reporter to ask the obvious follow-up? If it's really impossible, how exactly did the professor arrive at a number for how long it would take? Was there no time to chase down an answer that might have complicated, or killed, the story already taking shape? Or was something else going on entirely? I don't know which of those it was. I only know the question doesn't seem to have been asked.

Notice, too, that we don't actually know whose words "far, far longer than the age of the Universe" are. The article presents them as a summary of the professor's calculations, not as something he said. Either way, I don't think anyone was being deliberately misleading. The more likely explanation is the plainer one: a rough calculation, never fully worked through, whose actual significance never quite landed before it was turned into a soundbite — by the professor, by the reporter, or by both. That may be exactly what happened to Stephen Hawking too, when he said a horde of monkeys at typewriters would "very occasionally" type out a sonnet of Shakespeare's — technically defensible as a form of words, and completely useless, even actively misleading, as a description of the actual scale of time involved.

Last revised 2026-11-17

← Believing the HeadlineImpossible Is Not Very Unlikely →