The Theory of Everything and the Theory of Everything Else

Stuart Kauffman has a theory of everything else - A conversation with Stuart Kauffman

You told me once that the modern world starts with a clock. That's a strange place to begin a theory of biology.

It's where the trouble starts, though. The first mechanical clock — I think it was at Wallingford Abbey, around 1299, and they built it because the monks kept missing prayers. Let's assume that's roughly right. By 1370 there are clock towers all over Europe, the big hand and the little hand, numbers on the face. And I think that's when we started to think of the world as a machine. I mean, how could you not? You'd never seen a thing like that. Before, the world was an organic whole.

Then you get Copernicus, Kepler, Galileo, Newton — and now the world is a machine. The clockwork universe. And just before Newton there's Bacon, saying stop reading Aristotle, do experiments, wrest from nature what is our due. Underneath that is: the world is a machine, and if we know how it works, we have dominion. Three hundred and fifty years of spectacular science come out of that. I also think it's why we're wrecking the biosphere. But we should do the equation first or none of this lands.

Go on, then. What's the TAP equation actually saying?

A cosmologist friend, Andrew Little, proved a little theorem. Every time you add a new thing to the world, the waiting time until the next new thing is cut in half. Just accept the theorem. At first it takes a thousand years, then five hundred, then two-fifty, then a year, six months, a month and a half. That's the great acceleration. And it's not a metaphor — it's a curve, a hyperbola, crawling along for ages and then going nearly vertical.

And it fits, Luke, that's the thing. It fits the size of the largest human settlement going back three hundred thousand years. With Roger Koppl and Abby Devereaux we fit the growth of global GDP for two thousand years — that's not nothing. Sergi Valverde took three million patents and looked at which inventions came out of which. Theory predicts a power-law slope of about minus 1.3. The data's about minus 1.3.

Why does it speed up, though? What's the actual engine under it?

Well, that's the point exactly. The more things you've got, the more ways you can combine them, and every combination is a candidate for something new. So the adjacent possible — everything you could make next out of what already exists — keeps exploding outward.

It was really hard to invent things forty thousand years ago. It really was. I've got fifteen patents now, or dead patent applications, and it's easy, because there's so much lying around to recombine. My father was born in 1903, the year the Wright brothers flew. I was born in '39 — no helicopter yet, no computer, no jet engine, no chip, no internet, no fertiliser. Think what's happened in my eighty years. Now go back thirty thousand years and live your eighty. Not much happens. Same curve.

That's why we're in the Anthropocene, by the way. We were bound to get here. Pouring more concrete every year, more plastic, degrading the biosphere. It's the same TAP process.

The thing I keep coming back to is the smartphone in my pocket. It isn't one invention - it's a calculator, a computer, a Walkman, a camera, a cine camera, an atlas, a postbox, a library, all collapsed into one slab of glass. A table's worth of separate twentieth-century objects, recombined. It's almost the TAP curve made solid: every one of those had to exist before the phone could.

Exactly - you've got it, that's precisely it. None of it could have been invented in any order. The phone is a recombination of recombinations, and each part was itself somebody's adjacent possible a decade earlier. And here's the bit that should keep you up at night, Luke: you couldn't have listed the phone's uses when the camera was invented, and you can't list them now either. We don't know who people will be, or what they'll want. The adjacent possible the phone opens up is exploding faster than the phone did. That's the curve, going vertical, right there in your hand.

You draw a hard line between machines and living things. Where does it fall?

This is the heart of it, so let me go slowly. A machine doesn't build itself. Your car didn't put together its own engine block - it needs a constructor, hands, cranes, something from outside. Life doesn't.

Picture nine peptides in a loop. Peptide one grabs the two halves of peptide two, holds them next to each other, lowers the barrier, and the bond forms. Then peptide two does the same for peptide three, all the way round. The thing constructs itself. It doesn't describe itself, there's no program, no software, no instruction being followed — it just constructs itself, from the inside. Catalytic closure and constraint closure, joined up.

And - this only hit me a couple of weeks ago - I think that's Aristotle's form. The form of a thing is what makes it specifically what it is, what makes that oak that oak. Life is the form building itself from the inside, with nobody outside doing the building. Bergson's élan vital, except now it's not mysterious. It's just classical physics nobody had looked at. How much magic do you want?

And once a thing builds itself, you say it starts to matter to it whether it lasts.

Yes. And that's not small. A thing that constructs itself gets to keep existing, to persist. So now there's something good for it and something bad for it. There's yuck and there's yum. It's doing thermodynamic work to build itself, so it can do work on the world - it's an agent. I won't say consciousness, not yet. But this is the embodied mind, getting on with getting on.

Feynman said he only understands a thing if he can construct it. Well - Feynman can't construct a bacterial cell. The bacterium constructs the bacterial cell. Isn't that strange?

Take me to the screwdriver. That's where the maths falls over, isn't it?

[laughs] The screwdriver's my favourite, forgive me. What can you use one for? Drive a screw. Open a tin of paint. Tie it to a stick and spearfish. Short an electric circuit. A friend of yours has a coffee table that's an engine block - same idea, bigger.

Here's the theorem: you cannot list all the uses of a screwdriver. You'd have to know everyone who'll ever live, what they'll want, what the world looks like in ten thousand years. And you can't deduce one use from another — knowing it drives screws tells you nothing about shorting a circuit. They're what I call indefinite. Can't be listed, can't be ordered, can't be deduced from one another.

And that kills the equations outright, not just makes them harder.

It kills them, and the physicists hate it. Set theory says two sets are the same if they've got the same members. But I can't list the uses of a screwdriver - so I can't make the set. No set, no number. No number, no equation. The whole apparatus runs on being able to count the members, and the evolving biosphere just won't be counted.

So we don't merely not know what will happen. We don't know what can happen. We can't even write down the space of possibilities, so we can't define a probability. My cousin's a venture capitalist — he says the decision comes from your gut, not a calculation. The biosphere's been doing that for four billion years.

Which puts the physicists' Theory of Everything in some trouble.

It does, and I love it, because we got to the whole thing by accident. My wife made me a t-shirt with the TAP equation on it that says Theory of Everything Else. We did it to be funny. The physicists want a theory of everything - a glorious dream, and they may get it. But if it's supposed to include the evolving biosphere, and to allow deduction the way physics does, there can't be one. The biosphere keeps inventing the categories you'd need to describe it.

So you say it as a joke - the theory of everything and the theory of everything else — and then you realise it's true. There's a domain of entailing law, physics and chemistry. And a domain of no entailing law, the evolving biosphere. The Dao that can be spoken is not the Dao; the Dao is the source of the ten thousand things. Isn't that strange? It's kind of wonderful, actually. You start chuckling at some point.

If you're right, what does it ask of us?

To stop pretending we have dominion. If the biosphere isn't a machine, we can't master it, can't control it, can't know what we'll set off - so the only sane thing is to participate. With care, with concern. Not command.

We're the only species doing this, Luke. Flamingos aren't wrecking the biosphere. Earthworms aren't. Redwoods aren't. So it's plainly possible to live here without wrecking it. We're of nature, not above it. Look out the window - the plants, the beavers, whatever's near you. Four billion years of life building itself into things nobody could have listed in advance.

He says it the way he said it in the forest, eighty years old and gesturing at the trees. We built a civilisation on a clock from 1299, and it has cost us a great deal of the living world to keep believing the clock was the truth. Kauffman thinks the way out is to stop mistaking the machine for the thing it was modelled on. The theory of everything might be wrong, he says — but the theory of everything and the theory of everything else? We've got it knocked. He's chuckling when he says it. He means every word.