Evolutionary Biologist; Emeritus Professor of the Public Understanding of Science, Oxford; Author, The Selfish Gene, The God Delusion, An Appetite For Wonder: The Making of a Scientist
How different could life have been?


Physicists, including several in this group, are fond of asking, “What if the universe had been different?” Are the fundamental constants just numbers we accept as given, but which could have been different? Or is there some deeper rationale, which we shall eventually discover, that renders them unfree to change? Is our universe the way universes have to be? Or is it one of a huge ensemble of universes? Given present company, I would not aspire to this question, fascinating as it is. Mine is its biological little brother. Is the life that we observe the way life has to be? Or could we imagine other kinds of life? Long the stock in trade of science fiction, I want to move it closer to science’s domain. Unfortunately the question is one for a chemist – which I am not. My hope is that chemists will listen, and work on it.

Life as we know it is far more uniform than superficially appears. The differences between an elephant and an amoeba are superficial. Biochemically speaking, we are all playing most of the same tricks. At this level, most of the variation in life is to be found among the bacteria. We large animals and plants have just specialised in a few of the tricks that bacterial R & D developed in the Precambrian.

But all living things, bacteria included, practise the same fundamental tricks. Using the universal DNA code, the one-dimensional sequence of DNA codons specifies the one-dimensional sequence of amino acids in proteins. This determines the proteins’ three-dimensional coiling, which specifies their enzymatic activity, and this, in turn specifies almost everything else. So, I’m not talking about whether living things on other planets will look like us, or will have television aerials sticking on their heads. It is easy to predict that heavy planets with high gravitational fields will breed elephants the size of flies (or flies built like elephants); light planets will grow elephant-sized flies with spindly legs. It is easy to predict that, where there is light, there will be eyes. This is not what I am talking about. I want the answer to a more fundamental question.

My question, which is for chemists, is this. Can you devise a fundamentally different, alternative biochemistry? Given that, as I firmly believe, life all over the universe must have evolved by the differential survival of something corresponding to genes – self-replicating codes whose nature influences their own long-term survival – do they have to be strung along polynucleotides? The genetic code itself almost certainly didn’t have to be the one we actually have – plenty of other codes would have done the job. Ours is a frozen accident which, once crystallised, could not change. But can you think of a completely different kind of molecule, not a polynucleotide at all, perhaps not even organic, which could do the coding? Does it have to be digital like the DNA/RNA code, or could some kind of analogue code be accurate and stable enough to mediate evolution? Does it even have to be a one-dimensional code? And is there any other class of molecules that could step into the shoes of proteins?

Biochemists, please stop focusing exclusively on the way life actually is. Think about how life might have been. Or how life could be on other worlds. Channel your creativity to devising a complete, alternative biochemistry, whose components are radically different from the ones we know, but are at the same time mutually compatible – participants in a wholly consistent system which your chemical calculations show could actually work.

Why should we want this? I wanted to ask the question, “Is there life on other worlds, and how similar is it to the life we know?” But there is no immediate prospect of our receiving direct answers to these questions, and I am pessimistic of our ever doing so. Life has probably arisen more than once, but on islands in space too widely scattered to make a meeting likely. Theoretical calculations may be our best hope, and are certainly our most immediate hope, of at least estimating the probabilities. There’s also the point, which hardly needs making on Edge, that to seek the unfamiliar is a good way to illuminate oneself.

Reply to Paul Davies’s response to John McCarthy

Paul Davies notes that some night-migrating birds navigate by the stars, and asks whether avian DNA contains a map of the sky. "Could a scientist in principle sequence the DNA and reconstruct the constellations?" Alas, no.

Stephen Emlen, of Cornell University, researched the matter in 1975. He placed Indigo Buntings in a circular cage in the centre of a planetarium, and measured their fluttering against different sides of the cage as an indicator of their preferred migratory direction. By manipulating the star patterns in the planetarium, blotting out patches of sky and so on, Emlen showed that the buntings did indeed use Polaris as their North, and they recognized it by the surrounding pattern of constellations.

So far so good. Now comes the interesting part. Is the pattern of stars built into the birds’ DNA, or is there some other, more general way to define the north (or south) pole of the heavens? Put it like that, and the point jumps out at you: the polar position in the sky can be defined as the centre of rotation! It is the hub that stays still, while the rest of the heavens turn. Did the birds use this as a rule for learning?

Emlen reared young buntings in the planetarium, giving them experience of different artificial ‘night skies’. Half of them, the controls, experienced a night sky that rotated about Polaris, as usual. The other half, the experimental birds, experienced a night sky in which the centre of rotation was Betelgeuse. The control birds ended up steering by Polaris, as usual. But the experimental birds, mirabile dictu, came to treat Betelgeuse as though it was due north. Clever, or what?